The ntpJ Gene in the Enterococcus hirae ntp Operon Encodes a Component of KtrII Potassium Transport System Functionally Independent of Vacuolar Na (cid:49) -ATPase*

The ntpJ gene, the tail end in the vacuolar type Na (cid:49) ATPase ( ntp ) operon of Enterococcus hirae , encodes a putative 49-kDa hydrophobic protein resembling K (cid:49) transporter protein in Saccharomyces cerevisiae (Takase, K., Kakinuma, S., Yamato, I., Konishi, K., Igarashi, K., and Kakinuma, Y. (1994) J. Biol. Chem. 269, 11037–11044). Northern blotting experiment revealed that the ntpJ gene was transcribed as a cistron in the ntp operon. We constructed an Enterococcus strain in which the ntpJ gene was disrupted by cassette mutagen- esis with erythromycin resistance gene. The growth of this mutant was normal at low pH. However, the mutant did not grow at high pH in K (cid:49) -limited medium (less than 1 m M ), while the wild type strain grew well; the internal K (cid:49) concentration of this mutant was as low as 7% of that of the wild type strain, suggesting that the K (cid:49) accumulation at high pH was inactivated by disruption of the ntpJ gene. Potassium uptake activity via the KtrII system, which had been proposed as the proton potential- independent, Na (cid:49) -ATPase-coupled system working at high pH (Kakinuma,

All living cells show Na ϩ circulation across the cell membrane. This circulation is driven by active transport systems, which extrude sodium ions and maintain the Na ϩ concentration gradient directed inward (1)(2)(3). In animal cells, the familiar Na ϩ ,K ϩ -ATPase expels sodium ions, in which K ϩ uptake is tightly coupled. Bacteria have evolved diverse mechanisms for active sodium extrusion. Secondary Na ϩ /H ϩ antiporters are widely distributed (4), and some bacteria have been found to produce primary sodium pumps coupled with chemical reactions such as decarboxylation (5), electron transport (6), and ATP hydrolysis (7). Na ϩ reenters the cells via the Na ϩ gradient consumer, such as Na ϩ -coupled secondary co-transport systems, as the widespread route (8). Furthermore, Na ϩ -motive flagellar motor and the Na ϩ potential-driven ATP synthase are known for their physiological importance in some bacteria (9,10).
The Gram-positive bacterium Enterococcus hirae lacks the respiratory chain; the proton motive force is generated by proton expulsion via the F 0 F 1 , H ϩ -translocating ATPase (11). This bacterium has two sodium extrusion systems: Na ϩ /H ϩ antiporter (12,13) and an ATP-driven primary pump, Na ϩ -translocating ATPase (7). There has been no clear observation that suggests the Na ϩ gradient-consuming systems in this organism. The physiological role of sodium extrusion systems may be the elimination of sodium ions from cytoplasm and making room for K ϩ accumulation (14,15).
Our recent biochemical and molecular biological work on E. hirae Na ϩ -ATPase (16 -23) has suggested that this enzyme is the vacuolar type ATPase distributed in various eukaryotic endomembrane systems and archaebacteria (24 -26). This enzyme is encoded by a gene cluster (ntp operon) consisting of 11 ntp genes: ntpF, -I, -K, -E, -C, -G, -A, -B, -D, -H, and -J (22). In addition to the homologous counterparts of eukaryotic V-ATPases, 1 A, B, and K (16-kDa proteolipid) subunits, we found that six other Ntp proteins (F, I, E, C, G, and D) were similar counterparts of V-ATPase subunits in eukaryotes (22,(27)(28)(29). Thus, the expected molecular structure of E. hirae Na ϩ -ATPase resembles those of the eukaryotic vacuolar type H ϩ -ATPase complex.
On the other hand, one decade ago, Kakinuma and Harold (30) reported a peculiar feature of E. hirae Na ϩ -ATPase. They examined the proton potential-independent K ϩ transport activity (KtrII system) in this bacterium (31) and found that this activity depended on the activity of Na ϩ -ATPase. An apparent equimolar exchange of the internal Na ϩ for the external K ϩ was observed in the absence of the proton potential. They proposed the mechanism of the KtrII system as the direct Na ϩ /K ϩ exchange by the Na ϩ -ATPase. Since the molecular mechanism of this enzyme had not been elucidated, it was the simplest explanation (32). In this connection, we paid attention to the function of the ntpJ gene of the ntp gene cluster. This gene encoded a putative 49-kDa hydrophobic protein, which resembles those of K ϩ transport systems of Saccharomyces cerevisiae (Trk1 and Trk2) and of Escherichia coli (Trk) (22) and has not been assigned so far to other V-ATPase subunits. Thus, we thought that the NtpJ protein was the K ϩ transporting component for the KtrII activity in the Na ϩ -ATPase.
In this work we disrupted the ntpJ gene by cassette mutagenesis and examined the properties of this mutant. Although the Na ϩ -ATPase was alive in this mutant, the KtrII K ϩ uptake activity was deficient, suggesting that the NtpJ protein is a membraneous component of this K ϩ uptake system but not the essential one of the Na ϩ -ATPase complex. The KtrII K ϩ transport system is important for this organism in K ϩ -limited medium at high pH.

EXPERIMENTAL PROCEDURES
Strains and Growth Conditions-E. hirae strains used were ATCC 9790 (wild type strain), obtained from the American Type Culture Collection, and the mutants derived from it. Strain Nak1 is the mutant defective in the Na ϩ -ATPase (33), and JEM2 is the strain in which the ntpJ gene is disrupted as reported in this study. Cells were grown in the following standard media: KTY (1% Bacto tryptone, 0.5% Bacto yeast extract, 1% glucose, and 1% K 2 HPO 4 ) and NaTY (1% Bacto tryptone, 0.5% Bacto yeast extract, 1% glucose, and 0.85% Na 2 HPO 4 ). In some experiments, the concentration of yeast extract was decreased to 0.05% in order to decrease the K ϩ content. When required, K 2 CO 3 or Na 2 CO 3 was added to these media so as to alkalinize the medium pH. For culture of mutant JEM2, erythromycin (10 g/ml) was added to the media. The cell growth was monitored by a photometer at 600 nm.
Plasmids and Recombinant DNA Techniques-Plasmid isolation, gel electrophoresis, ligation, restriction enzyme analysis, and transformation of Escherichia coli were done as described by Ausubel et al. (34). The preparation of E. hirae cells competent for electrotransformation and the transformation were performed as described elsewhere (35).
Northern Blotting-Total RNA was extracted as described elsewhere (34) from E. hirae cells grown in various media, fractionated through formaldehyde gels, and transferred to nitrocellulose membranes. DNA fragments labeled by the random priming method were used as probes ( Fig. 1).
Disruption of the ntpJ Gene-The chromosomal locus of the ntpJ gene was disrupted by insertion of an erythromycin resistance gene as shown in Fig. 2A. First, the 2.4-kb HindIII-HindIII fragment covering from the part of ntpD to ORFX in the ntp operon was subcloned into an E. coli vector pUC119 (5.6-kb pKAZ132). The 1.8-kb AvaI-HindIII fragment of pVA838 (36), which contains erythromycin resistance gene (erm), was cut out and blunt-ended by the Klenow fill-in procedure. This erythromycin gene cassette was introduced into the blunt-ended SphI site of the ntpJ gene of pKAZ132 (7.4-kb pJEM2) ( Fig. 2A). Finally, the 4.2-kb ntpJ::erm fragment was liberated from pJEM2 by HindIII digestion and electroporated into strain 9790. Three erythromycin-resistant (Erm R ) transformants were obtained. The chromosomal DNA was isolated from strain 9790, mutant JEM2 (one of the Erm R transformants) was digested with HindIII, and Southern hybridization was performed with pJEM2 as a probe. pJEM2 cut with HindIII yielded two fragments of 4.2 and 3.2 kb. As a control, the probe was hybridized to the fragments with the expected sizes (Fig. 2B, lane 3). Then this probe was hybridized to 2.4-and 4.2-kb HindIII-HindIII genomic fragments from the wild type and JEM2, respectively (Fig. 2B, lanes 1 and 2). The same hybridization pattern as JEM2 was observed for two other Erm R strains. Hybridization was also performed with DNA fragments digested with other restriction enzymes. All these results of Southern blotting analysis confirmed that the ntpJ gene was disrupted by insertion.
Southern Hybridization-Chromosomal DNA, prepared by the method described by Ausubel et al. (34), was cut with restriction enzymes, and the resultant DNA fragments were separated on agarose gels. The resolved fragments were transferred to positively charged nylon membranes and fixed by UV irradiation. Labeled DNA probes were prepared according to the manufacturer's instructions for the BcaBEST labeling kit (Takara Shuzo Co.). Hybridization was carried out overnight at 55°C, and the filters were washed with 1 ϫ SSPE and 0.5% sodium dodecyl sulfate for 15 min at room temperature and 2 h at 60°C.
Measurement of the Internal Na ϩ and K ϩ -The cellular contents of K ϩ and Na ϩ in growing cells were determined by flame photometry (37). Samples of cell suspension (10 ml at OD 600 ϭ 0.2) were spun at 12,000 ϫ g in a Hitachi centrifuge through a layer of silicon oil and mineral oil (1.8:0.3). Part of the supernatant was removed, and the remainder was carefully removed by suction; the tip of the centrifuge tube, containing the pellet, was cut off and extracted with hot 5% trichloroacetic acid. Aliquots were analyzed for K ϩ and Na ϩ by a flame photometer. The cytoplasmic water space was taken to be 1.75 l/mg of cells (37).
Transport Assays-To measure the KtrII activity (30), the cells were loaded with Na ϩ as described by Bakker and Harold (37) and suspended in 50 mM Na ϩ -CHES buffer, pH 9.0, at a density of 1 mg (dry weight)/ ml. After incubation with 10 mM glucose and 20 M tetrachlorosalicylanilide for 10 min, the reaction was initiated by the addition of 1 mM KCl. Cell samples were collected by filtration on membrane filters (pore size, 0.4 m; nucleopore polycarbonate, Costar Scientific Co., Cambridge, MA) and washed with 2 mM MgSO 4 . Sodium and potassium contents were determined by flame photometry after extraction of the cells with hot 5% trichloroacetic acid. Sodium extrusion was monitored with 22 Na ϩ as described previously (12). Cells harvested in the late log phase were used directly as K ϩ -loaded cells. Washed cells were suspended at 4 mg (dry weight)/ml in 50 mM K ϩ -HEPES buffer (pH 7.0) containing 100 mM maleate-KOH with 20 mM 22 NaCl (2.315 MBq/ mmol) and incubated at 25°C for 60 min. At intervals, samples (0.2 ml) were filtrated through membrane filters (pore size, 0.45 m; Toyo Roshi Co., Tokyo) and washed with the same buffer, and the radioactivity was measured with a liquid scintillation counter.
Miscellaneous Methods-Western blotting was performed as described elsewhere (23) and visualized by using goat anti-rabbit IgG conjugated to alkaline phosphatase. The cell membranes were prepared by the standard procedure as described previously (17) and, if necessary, stored frozen at Ϫ80°C. The Na ϩ -stimulated ATPase activity of the membranes was determined at pH 8.5 in the presence or absence of 25 mM NaCl as described previously (30). Denatured polyacrylamide gel electrophoresis was carried out using the system of Laemmli with 10% polyacrylamide (38). Protein was determined by the method of Lowry et al. (39) with bovine serum albumin as a standard. The membrane potential of intact cells was calculated on the basis of the distribution of [ 3 H]tetraphenylphophonium ion as described previously (30).
Materials-Enzymes for recombinant DNA techniques were purchased from TOYOBO (Tokyo) and Takara Shuzo Co. (Kyoto).
[␣-32 P]dCTP (111 TBq/mmol) was purchased from Amersham (United Kingdom) and 22 NaCl was from Daiichi Pure Chemical Co. (Tokyo). All reagents used were commercial products of analytical grade.

RESULTS
The ntpJ Gene Is a Cistron in the ntp Operon- Fig. 1 shows the structure of the ntp operon encoding the Na ϩ -ATPase and its neighboring genes; the ntpR and -X genes are in the opposite direction of the operon. Although nine ntp gene products, from F to D, in this operon are found to be the homologues of vacuolar type ATPase subunits (22,(27)(28)(29), we could not so far identify any vacuolar ATPase subunits corresponding to the ntpH and ntpJ gene products. As there is no potential sequence for a ribosomal binding site preceding the start codon of ntpH, it is likely that the ntpH gene is not the reading frame. In addition, two sequences able to form stem-loop (or palindrome) structures lie in the sequence between the ntpD and ntpJ genes, and a possible promoter sequence (Ϫ35 and Ϫ10 boxes) is observed in the region preceding the ntpJ (22). It is, therefore, possible that the ntp operon is terminated at the end of ntpD and that the ntpJ gene is in another operon. Indeed, the primary sequence of the ntpJ gene product resembles K ϩ transporter (22) but not the V-ATPase subunits. Thus, we first examined the transcripts of the ntp genes by northern hybridization. Fig. 3 shows the Northern blotting experiment with three different probes for ntp genes (Fig. 1). E. hirae Na ϩ -ATPase is not constitutive; this enzyme is highly induced in culture conditions with high Na ϩ and/or at high pH (16,40). In this experiment, total RNA was prepared from strain 9790 cultured in various media: KTY medium at pH 7.5 (Fig. 3, lane  1), NaTY medium containing 0.4 M NaCl at pH 7.5 (Fig. 3, lane  2), or NaTY medium containing 0.4 M NaCl at pH 10.0 (Fig. 3,  lane 3). The probe I, the 2.5-kb HindIII-HindIII DNA fragment corresponding to the portion from the ntpA to ntpD gene (Fig.  1), hybridized the RNA longer than 10-kb in the RNA extract prepared from the cells grown in high Na ϩ (Fig. 3A, lane 2). As the spot of hybridized RNA was broad, the precise size of mRNA was difficult to determine from the data. This mRNA was hardly detected in the RNA extract from the cells grown in limited Na ϩ medium (Fig. 3A, lane 1) and, by contrast, significantly increased in that from cells grown under high Na ϩ and high pH conditions (Fig. 3A, lane 3). When the probe II, 0.8-kb PvuII-PvuII fragment corresponding to only a part of the ntpJ gene (Fig. 1), was used, the RNA longer than 10-kb was likewise hybridized (Fig. 3B). The amount of this mRNA was diminished in Na ϩ -limited conditions (Fig. 3B, lane 1) but significantly increased under high Na ϩ and high pH conditions (Fig. 3B, lane 3). When the probe III, 1.2-kb PvuII-PvuII fragment covering the genes of ntpJ and ORFX (Fig. 1) was used, we observed two spots: one for mRNA of the ntp operon whose amount is altered according to the culture conditions (Fig. 3C,  lanes 1, 2, and 3) and the other for ORFX. These results indicate that the ntpJ gene is cotranscribed with other ntp genes. The Northern blotting experiments described above and the Western blotting experiment with anti-V 1 serum (data not shown; Ref. 23) have showed that a change in the amounts of Na ϩ -ATPase in these media corresponds to the change in the amounts of mRNA, suggesting that expression of the ntp operon is regulated at transcriptional level. 2 Thus, the function of the NtpJ protein resembling K ϩ transporter is probably coordinated with that of the vacuolar type Na ϩ -ATPase complex in this organism.
Growth Properties of the ntpJ-disrupted Mutant-To elucidate the physiological role of the NtpJ protein, we examined the phenotype of cells lacking the ntpJ gene product. The chromosomal locus of the ntpJ gene was disrupted by insertion of an erythromycin resistance gene as described under "Experimental Procedures" (Fig. 2). Fig. 4 shows the growth of the parental strain 9790 and the ntpJ-disrupted mutant JEM2 at different pHs. First, cells were grown at pH 7.5 in NaTY medium containing 1 mM K ϩ and 130 mM Na ϩ . In this medium, mutant JEM2 grew well as strain 9790 did (Fig. 4, A and B, open symbols). The intracellular concentrations of K ϩ and Na ϩ of these cells were calculated to be 650 Ϯ 26 mM and 10 Ϯ 5 mM for strain 9790, and 620 Ϯ 30 mM and 60 Ϯ 25 mM for mutant JEM2, respectively. Even when the medium pH was shifted from 7.5 to 10.0 by the addition of 80 mM Na 2 CO 3 , the wild type strain grew well, with nearly the same doubling time as that at pH 7.5 (Fig. 4A, closed circles). By contrast, the pH shift to 10.0 inhibited the growth of mutant JEM2 (Fig. 4B, closed circles). Table I shows the internal ion concentrations of these strains growing at pH 10.0. At pH 10.0, the high K ϩ concentration was maintained in strain 9790, although the amount of K ϩ decreased to about one-third of the value at pH 7.5. The internal Na ϩ concentration was kept lower than the external one in strain 9790. However, the internal K ϩ concentration of the mutant JEM2 in this medium was only 17 mM. Most of the internal K ϩ was replaced with Na ϩ . These features of cell growth and the cation contents for mutant JEM2 were similar to those for a Na ϩ -ATPase mutant Nak1 reported previously (33). When passive K ϩ efflux was measured with K ϩ -loaded JEM2 cells, the K ϩ efflux rate was the same as that of the parent strain (data not shown). These results suggest that the ntpJ mutant is defective in K ϩ accumulation of this bacterium at high pH. The mutant JEM2 grew well at low pH, such as 6.0, as did the wild type strain. Even when the medium osmolarity was increased by the addition of 0.5 M sorbitol, the cell growth of strain 9790 or JEM2 was not affected (data not shown). The defect of the ntpJ mutant may be specific for this bacterium's physiology at high pH. Fig. 4C shows the growth of mutant JEM2 in NaTY medium where 10 mM KCl was also added. It is noteworthy that the mutant JEM2 grew well in this medium even after shifting the medium pH to 10.0 (Fig. 4C, closed circles). The internal concentrations of K ϩ and Na ϩ of JEM2 growing at pH 10.0 in this medium were 120 and 80 mM, respectively (Table I). It is important to point out here that the Na ϩ -ATPase mutant, Nak1, did not grow in the same medium: NaTY medium supplemented with additional 10 mM KCl (33). These results suggest that E. hirae grows at high pH even where the internal K ϩ concentration was relatively moderate (120 mM; Table I) and that K ϩ is accumulated in the cells at high pH in an NtpJindependent manner where the external K ϩ concentration is more than 12 mM. The expected mechanism for K ϩ transport in the ntpJ-disrupted strain is described under "Discussion." An ntpJ-disrupted Mutant Lacks the KtrII Activity-It is reported that the apparent K m value for K ϩ of the KtrII system was 0.5 mM (30,31); this system should work for K ϩ accumulation under K ϩ -limited growth conditions (32). Therefore, it is likely that the NtpJ protein is related with the activity of K ϩ uptake via the KtrII system. Fig. 5 shows the KtrII activities of strain 9790, the Na ϩ -ATPase mutant, Nak1 (33), and mutant JEM2. In this experiment, cells were grown in NaTY medium, and the KtrII activity was assayed at pH 9.0, where the proton motive force was dissipated by addition of the protonophore tetrachlorosalicylanilide. The KtrII activity that exchanged nearly equivalently Na ϩ for K ϩ was observed in strain 9790 (Fig. 5A) but not in Nak1 (Fig. 5B). In contrast, in JEM2, K ϩ uptake activity was limited, although efflux of the internal Na ϩ was normal (Fig. 5C); Na ϩ efflux and K ϩ influx were not equimolar in this mutant. These results suggest that NtpJ protein is indeed essential for K ϩ uptake via the KtrII system. Furthermore, Na ϩ extrusion does not seem to be obligatorily coupled with K ϩ uptake.
The ntpJ-disrupted Mutant Has Normal Activity of Vacuolar Na ϩ -ATPase-Glucose-dependent downhill Na ϩ efflux observed in the Na ϩ -loaded cells, as shown in Fig. 5, has been ascribed to the activity of Na ϩ -ATPase (30). Therefore, the result shown in Fig. 5C suggests that the Na ϩ -ATPase activity is not damaged by disruption of the ntpJ gene. The activity of the Na ϩ -ATPase was examined. Fig. 6 shows active sodium extrusion via the Na ϩ -ATPase by the whole cells. In this experiment, cells were cultured in high Na ϩ medium (NaTY medium containing 0.5 M NaCl) so as to induce the Na ϩ -ATPase (16). The proton potential-independent active 22 Na ϩ extrusion activity was observed by the parent strain (Fig. 6A); this activity was not observed by Nak1 (33). Active 22 Na ϩ extrusion was also observed by JEM2 cells (Fig. 6B). Further-more, a Western blotting experiment using anti V 1 serum was performed. The molecular size of the purified V 1 moiety of Na ϩ -ATPase was about 400 kDa, consisting of polypeptides of 69 kDa (A), 52 kDa (B), and 29 kDa (D) with a probable stoichiometry of 3:3:1 (23). Antiserum raised against this enzyme reacted intensely with the A subunit (Fig. 7, lane 1). Immunoblotting revealed that the amount of Na ϩ -ATPase in the membranes of JEM2 was increased when grown under high Na ϩ and high pH conditions, as was the case for strain 9790 (Fig. 7). Western blotting of total cell lysates showed similar results (data not shown). The Na ϩ -stimulated ATP hydrolytic activity of the cell membranes prepared from strain 9790 and JEM2 grown in the same high Na ϩ medium (Fig. 6) were 0.13 and 0.13 units/mg protein, respectively. These results suggest that the NtpJ-dependent K ϩ uptake, KtrII, is not directly coupled with the reaction of the Na ϩ -ATPase and that the NtpJ protein is not essential for induction of the Na ϩ -ATPase and its subunit assembly.

DISCUSSION
In bacteria, the transports of K ϩ and Na ϩ are mediated by separate transport systems usually linked to the chemiosmotic proton circulation (1-3). In E. hirae, in addition to two Na ϩ extrusion systems (Na ϩ -ATPase and Na ϩ /H ϩ antiporter), two distinct potassium uptake systems have been recognized. The major one, KtrI, is thought to be constitutive and resembles the Trk system of E. coli (37), dependent on the proton motive force and ATP. The second one, KtrII, which was internal Na ϩ -dependent, but not dependent on the electrochemical potentials of either H ϩ or Na ϩ , required ATP (30,31). This system stoichiometrically exchanged Na ϩ for K ϩ . The mechanism of the KtrII system has not been clearly understood. However, since (i) the mutant that lacked the Na ϩ -ATPase also lacked KtrII and (ii) FIG. 5. The KtrII activities of various E. hirae strains. Cells were grown in NaTY medium (pH 7.5), loaded with Na ϩ , and suspended in 50 mM Na ϩ -CHES buffer (pH 9.0) at a cell density of 1 mg (dry weight)/ml. The suspension was supplemented with 20 M tetrachlorosalicylanilide and with (closed symbols) or without (open symbols) 10 mM glucose at 0 min; K ϩ uptake was initiated by the addition of 1 mM KCl at 10 min as indicated by arrows. The cellular contents of K ϩ (circles) and Na ϩ (squares) were determined by flame photometry. A, strain 9790; B, mutant Nak1; C, mutant JEM2.

TABLE I
The internal concentrations of K ϩ and Na ϩ in E. hirae growing at high pH Cells were grown on NaTY medium with or without added 10 mM KCl as shown in Fig. 4. At A 600 ϭ 0.1, the medium pH was shifted to 10 by the addition of 80 mM Na 2 CO 3 . After 1.5 h, aliquots of cell suspensions were collected, and the K ϩ and Na ϩ amounts in media and cells were determined by flame photometry. The interna water space was estimated as 1.75 l/mg of cells. Standard deviations were from duplicate experiments. the KtrII and the Na ϩ -ATPase were induced in parallel when cells were grown in media rich in Na ϩ , particularly under the conditions that limit the generation of the proton motive force, Kakinuma and Harold proposed the simplest hypothesis, that the sodium ATPase is KtrII itself (30). Therefore, we expected a new Na ϩ ,K ϩ -ATPase in this bacterium before finding out that the molecular structure of this Na ϩ -ATPase belonged to that of the vacuolar type ATPase, an electrogenic proton pump (24 -26). This speculation should be now withdrawn and replaced by the following. First, the Na ϩ -ATPase transports Na ϩ electrogenically, not obligatorily linked with potassium ion transport (41); even in the absence of the potassium ion, the membrane potential was generated via the electrogenic Na ϩ flux by the Na ϩ -ATPase. In this context, it is noteworthy that the 16-kDa proteolipid, whose amino acid sequence is homologous to those of several eukaryotic V-ATPase proteolipids, is the subunit of E. hirae Na ϩ -ATPase complex, and the DCCDreactive glutamic acid residue (Glu 139 ) is present in the fourth membrane-spanning region (21). The Na ϩ -ATPase activity was inhibited by DCCD, 3 suggesting that the 16-kDa proteolipid is the electrogenic Na ϩ pathway. Second, in the ntpJ-disrupted mutant in which the KtrII activity was negative (Fig. 5C), the activity of Na ϩ -ATPase was normal (Figs. 6 and 7). In other words, NtpJ protein is essential for the KtrII but not function-ally essential for the Na ϩ -ATPase reaction. Finally, the ATP hydrolytic activity of this enzyme is activated by Na ϩ but not K ϩ . The activity was not synergistically activated by Na ϩ and K ϩ (30). Although both Na ϩ -ATPase and NtpJ protein are essential for the KtrII activity, these results suggest that the KtrII activity is not a part of the Na ϩ -ATPase activity. The NtpJ protein is probably an integral membrane protein having at least 10 membrane-spanning domains, whose amino acid sequence is similar to those of the presumptive K ϩ /H ϩ symporter proteins such as Trk1 and Trk2 of S. cerevisiae (22,42). Therefore, it is likely that the NtpJ protein is the cotransporter itself. The K ϩ gradient (in 3 out) of 230 was generated in strain 9790 growing at pH 10.0 in NaTY medium containing 1 mM K ϩ (Table I), where the membrane potential of about Ϫ70 mV (inside negative) was generated; the intracellular pH of these cells should be acidified to about 8.2 (43). Since the Na ϩ gradient (out 3 in) of 10 was generated in these cells (Table I), the magnitude of the Na ϩ electrochemical potential of Ϫ130 mV, but not the H ϩ gradient, is sufficient to generate the K ϩ gradient shown in Table I, suggesting that the NtpJ protein may be the Na ϩ /K ϩ symporter. However, in the KtrII assay (Fig. 5A), the K ϩ gradient of about 200 was generated in strain 9790 where the sodium potential was negligible; the Na ϩ gradient (in 3 out) of 5 and the membrane potential of less than Ϫ70 mV across the cell membrane were generated. In a previous paper (30), the K ϩ gradient of at least 800 was generated by the H ϩ -ATPase mutant under the same assay, where the size of the Na ϩ potential was negligible. These in vitro results are not consistent with the simple secondary co-transport mechanism and suggest the possibility of the primary pump mechanism; in this case, there should be gene(s) encoding the other component(s), such as the ATPase catalytic subunit(s) of KtrII, whose expression should also be regulated by the same signal for the ntp operon. At the moment, it is hard to propose the molecular mechanism of the NtpJ-dependent KtrII K ϩ uptake system. To know more details of the KtrII system, we are now working toward the isolation of the NtpJ-independent K ϩ uptake mutant and the functional expression of the ntpJ gene in the K ϩ transport mutant of E. coli.
It is important to point out that the KtrII may not be the only route of K ϩ uptake at high pH. In growing JEM2, the K ϩ concentration gradient of about 10 -17 was generated whether or not the medium was supplemented with additional 10 mM K ϩ (Table I). This NtpJ-independent K ϩ accumulation is enough for the growth of this bacterium at high pH where the external K ϩ is moderate (Fig. 4C). The action of the Na ϩ -ATPase is important for this K ϩ uptake, because the Na ϩ -ATPase mutant did not grow under the same growth conditions as described above. We think that the membrane potential generated by the Na ϩ -ATPase is the driving force for the NtpJindependent K ϩ accumulation, since the membrane potential of about Ϫ70 mV was generated in JEM2; K ϩ accumulation by JEM2 was limited in the KtrII assay (Fig. 5C).
Thus, the physiological function of E. hirae Na ϩ -ATPase is to extrude Na ϩ from cytoplasm and generate the sodium potential, which drives active K ϩ transport systems at high pH where the proton motive force is minimal. It is notable that two important genes for the cation homeostasis of E. hirae at high pH form an operon.  4), respectively. Ten micrograms of each membrane and 0.5 g of purified V 1 -ATPase (lane 1) were electrophoresed and immunoblotted with antiserum against purified V 1 -ATPase (dilution 1:3000) and visualized by using the alkaline phosphatase system.