ATP-driven Exchange of Na+ and K+ Ions by Streptococcus faecalis"

We describe the characterization of KtrII, a novel potassium transport system of Streptococcus fuecdis, first discovered by H. Kobayashi ((1982) J. Bucteriol. 150, 506-511). KtriI requires sodium ions and me- diates the stoichiometric exchange of internal Na+ for external K+. Potassium accumulation is not energized by the electrochemical potentials of either H+ or Na"; the energy source is probably ATP. Two lines of evi- dence indicate that KtrII is a manifestation of the sodium-stimulated ATPase reported earlier (Heefner, D. L., and Harold, F. M. (1982) Proc. Nutl. Acud. Sci. U. S. A. 79, 2798-2802). (i) Mutants that lack the ATPase also lack KtrII, and revertants recover both in parallel. (ii) KtrII and the Na+-ATPase are induced in parallel when cells are grown on media rich in sodium, particularly under conditions that limit the generation of a proton potential. KtrlI is not induced in response to K' deprivation. We propose that the Na+-ATPase exchanges Na+ for K+ ions. and

-10 0 nisms they were diluted into 2KTY medium, pH 6.0, and those that grew were discarded.
Transport Experiments-Most of our experiments were done with sodium-loaded cells. Cells harvested from overnight cultures were washed twice with 2 mM MgS04, resuspended in 50 mM sodium maleate (pH 8.0), and incubated at 37 "C with the ionophore monactin (2 pg/ml; 15 min). The cells were collected by filtration and washed at least six times with sodium maleate buffer. They were then washed with the buffer to be used in the experiment and suspended in the latter at a density of 1 mg (dry weight)/ml.
A few experiments required cells containing neither Na+ nor K' ions. These were prepared by incubating sodium-loaded cells in buffer containing 50 mM Tris chloride, 400 mM choline chloride, and 10 mM glucose, pH 8.5 (70 min, room temperature). It iS likely that choline is the chief intracellular cation of these cells.
Analytical Procedures-Cell samples were collected by filtration on Millipore filters (pore size 0.45 pm) and were washed with 2 mM MgSO,. Sodium and potassium contents were determined by flame photometry after extraction of the cells with hot 5% trichloroacetic acid (Harold and Papineau, 1972;. The membrane potential was calculated from the uptake of ['HITPP'' (Bakker and Harold, 1980). The cellular water space was taken to be 1.75 pl/mg, dry weight. Turnover of the potassium pool was examined by the use of 42K as described by Slayman and Tatum (1965). scribed by Kinoshita et al. (1984) from cells harvested in the early Unless otherwise noted, membrane vesicles were prepared as destationary phase. The sodium-stimulated ATPase was assayed by the procedure of Kinoshita et al. (1984) in the presence of 0.5 mM DCCD.

RESULTS
Potassium Accumulation by Mutant AS25"According to Kobayashi and Unemoto (1980), mutant AS25 has a lesion in the KFo-ATPase that impairs proton extrusion and the generation of an electrochemical proton potential. In consequence, KtrI should be inoperative. Cells of this mutant nevertheless accumulate K' ions under certain conditions; our experiments corroborate  conclusion that K' uptake by AS25 is the work of a novel potassium transport system distinct from KtrI. Fig. 1A shows a typical experiment. Mutant AS25 was grown on medium NaTY; the cells were then loaded with sodium ions and allowed to glycolyze in Na'-Tricine buffer, pH 8.5. Upon addition of 0.5 mM KC1 the cells exchanged part of their sodium complement for potassium, establishing Time (mid and also by the protonophore TCS (data not shown). However, neither reagent inhibited the initial rate of K+ uptake as measured with 42K+, suggesting that the site of inhibition is not the uptake of K' into the cells (see below), Even in the presence of both DCCD and TCS the cells established a K' concentration gradient of nearly 400 (Fig. IC). Potassium uptake by AS25 was maximal at pH 8.5 to 9, and Rb+ ions were accumulated only to a small degree (data not shown).
The alkaline pH optimum of K' uptake, the discrimination against Rb' , and the relative indifference to DCCD and to proton conductors distinguish K+ uptake by AS25 from that by the parent strain, 9790 (Bakker and Harold, 1980) and warrant its attribution to a second transport system. Operationally, we recognize KtrII by its capacity for DCCD-resistant uptake of K+ ions (but not of Rb+ ions) at pH 8.5. A second criterion, an absolute dependence on sodium ions, will be described below. In our experience, K' uptake by AS25 must be attributed entirely to KtrII.
Conditions for the Expression of KtrII-As a matter of convenience most of our experiments were done with mutant AS25 which effectively lacks KtrI. In the parent strain, 9790, significant activity of KtrII was observed only when the cells were grown under special conditions . Table   I summarizes a series of experiments in which AS25 and 9790 were grown on various media; the activity of KtrIJ was then assayed by the accumulation of K' under a standard set of

TABLE I
Conditions for the expression of KtdZ Cells were grown on the media listed below and harvested in the early stationary phase. KtrII activity was assayed with sodium-loaded cells suspendd in 50 mM Na+-Tricine buffer, pH 8.5, in the presence of glucose, 0.2 mM DCCD, and 0.5 mM KCI. Special media contained tryptone, yeast extract, and glucose in the usual proportions plus the following salts: ZKTY, 20 g of K,HP04; KNaTY, 10 g of KzHP04 and 8.5 g of NazHP04; TrisTY, 0.2 M Tris base and 0.05 M H3P0,. The pH was brought to 7.5 with KOH, NaOH, or Tris. conditions specified in the legend. We interpret the results to mean that KtrII is induced when the cells are grown in media rich in sodium and under conditions that reduce the proton motive force, whether by mutation or for other reasons. Cells of 9790 grown on NaTY medium had low but significant KtrII activity. Note that growth on potassium-deficient medium was neither necessary nor sufficient to elicit expression of KtrII and that the presence of high K+ concentrations did not repress it. System KtrII Mediates Exchange of Intracellular Na' for Extracellular K+-We mentioned above that KtrII was induced only when the cells were grown on sodium-rich media, and  reported that Na' ions were required for its operation. Fig. 2 documents that K' uptake by KtrII specifically requires intracellular Na' ions. Mutant AS25 was grown on NaTY medium, and the cells were then depleted of both K+ and Na' ions as described under "Experimental Procedures." Such cells glycolyzed well (30-40 nmol of H+/ mg cells-min) but took up K+ only when the medium was supplemented with Na+ ( Fig. 2A). By contrast, sodium-loaded cells took up K+ in the absence of external Na' (Fig. 2B). We would emphasize that the total amount of Na+ present in the latter cells if released into the medium would bring the external Na' concentration to 1 mM; this level of Na+ supported limited K+ uptake and at a reduced rate. We infer that K' uptake via KtrII requires intracellular Na+ ions. Fig. 3 documents the complementary conclusion that expulsion of Na+ ions from AS25 is markedly stimulated by external K' . This is an important point. Under the conditions of the present experiments (50 mM Na+-Tricine, 2 mM KC1 or RbCl, pH 8.2, DCCD) the wild-type strain 9790 extruded only a small part of its sodium complement by exchange for K' or Rb+ (Fig. 3A). The reason, we believe, is that inhibition of the FIFo-ATPase by DCCD largely abolished the activity of KtrI, while KtrII activity is minimal (Table I). Cells grown on KTY did not extrude Na' ions at all (data not shown); this suggests that the minimal sodium extrusion seen in Fig.  3A is due to KtrII rather than to residual KtrI activity. By contrast with the parent strain, AS25 slowly but steadily exchanged internal Na' for external K' ; Rb' was a poor substitute (Fig. 3B). We attribute this sodium extrusion to the capacity of KtrII to exchange Na' for K+ ions. Cells of AS25 grown on 2KTY medium, which does not induce KtrII, were unable to expel Na+ ions (Fig. 3C).
The Sodium-ATPase Is Part of KtrII-We know from earlier work that sodium extrusion by S. faecalis 9790 is a primary process, probably mediated by a sodium-translocating ATPase Harold 1980, 1982). Very recently, Kinoshita et al. (1984) discovered that mutant AS25 produces elevated levels of the Na+-stimulated ATPase when grown on NaTY medium and so does the parent strain 9790 when grown on NaTY medium plus CCCP. Since these conditions also induce KtrII which, moreover, requires internal Na' ions, we asked whether the Na+-stimulated ATPase plays a role in that potassium transport system.
Mutant 7683 lacks the capacity to extrude Na+ ions and also lacks the Na+-stimulated ATPase (Heefner and Harold, 1982). When grown on medium KNaTY plus CCCP, conditions that induce KtrII in the parent strain, no DCCDresistant potassium uptake was observed (Table 11). Two revertants of this mutant were also examined. R-I, which has recovered only limited capacity for sodium extrusion (possibly attributable to a Na'/H+ antiporter) and does not contain Na+-stimulated ATPase, again lacked KtrII activity. By contrast R-11, which has recovered the sodium-stimulated AT-Pase, contained KtrII activity as well (Table 11). These observations suggest that KtrII is a manifestation of the sodiumstimulated ATPase.
The Steady-State K' Leuel-Our observations, like those of , indicate that KtrII is a primary transport process whose energy source is ATP. We must then account for the partial inhibition of K' accumulation by DCCD and TCS, which do not inhibit the Na'-ATPase (Heefner and Harold, 1982;. The following experiments suggest that the K' concentration gradient attained by KtrII is a steady state in which rapid influx is balanced by rapid efflux and that the inhibitors selectively accelerate efflux. Given that the potassium pool represents a steady state, does DCCD reduce entry of K+ or stimulate its exit? Fig. 5 shows that the turnover of the K+ pool was more rapid in the presence of DCCD than in its absence; the rate constant for influx was almost unaffected, but that for efflux was elevated 3-fold (legend). Separate experiments confirmed that the initial rate of net K' uptake was almost unaffected by DCCD or by TCS: 25 nmol/min. mg cells in the controls, 20 nmol in the presence of DCCD or TCS. We conclude that the inhibition of K+ accumulation by DCCD (and also by TCS; data not shown) is primarily due to enhanced K+ efflux.
We do not know how DCCD and TCS enhance K' efflux from AS25, but it may be pertinent that the inhibitors reduced the electrical potential difference across the plasma membrane. AS25 is thought to have a defective F1Fo-ATPase, and its capacity to extrude protons is limited (Kobayashi and Unemoto, 1980). Nevertheless, cells of AS25 (grown on NaTY) glycolyzing in Na+-Tricine buffer maintained a membrane potential of approximately -50 mV as judged by [3H] TPP+ uptake. In the presence of DCCD (0.2 mM) or of TCS (10 PM) this was reduced to approximately -30 mV. A number of other reagents that lower the steady-state K' level, including the lipophilic cations triphenylmethylphosphonium ion (1 mM) and dibenzyldimethylammonium ion (5 mM), may also exert their effects by modulating the membrane potential. , were grown on the media listed below; the cells were loaded with Na+ and resuspended in 50 mM Na+-Tricine buffer, pH 8.2. All suspensions received DCCD (0.2 mM) at -15 min. The sodium content of the initial cells was designated as 100%; absolute Na+ levels are listed below. A, parent strain 9790 grown on NaTY medium, 1.05 pmol of Na+/mg of cells; B, AS25 grown on NaTY, 0.85 pmol of Na+/mg of cells; C, AS25 grown 2KTY medium, 0.95 pmol of Na+/mg of cells. Aliquots of each suspension received the following additions: 0, none; 0, plus 10 mM glucose at -5 min. B, plus glucose, followed by 2 mM RbCl at 0 min (arrow). A, plus glucose, followed by 2 mM KC1 at 0 min (arrow).

TABLE I1 Corretation between KtrII activity and the Na+-stirnuhted ATPase
Cells grown on the media listed below were harvested in the early stationary phase. KtrII activity was assayed with sodium-loaded cells suspended in 50 mM Na+-Tricine buffer, pH 8.5, in the presence of glucose, 0.2 mM DCCD, and 0.5 mM KCI. Na+-ATPase activity was assayed as described under "Experimental Procedures," with and without 25 mM NaC1. The number in parentheses is the basal activity in the absence of Na+. Please see legend to Table I  NaTY 13 800 * These membrane vesicles were prepared as described by Heefner and Harold (1982

DISCUSSION
On the basis of Kobayashi's paper (1982) and of the present study, the characteristics of KtrII can be summarized as follows. (i) KtrII selects K+ over Rb+ but has a relatively low affinity (K, approximately 0.5 mM); the optimal pH is about 8.5. (ii) Potassium uptake requires Na+ ions to be present in the cytoplasm; overall, KtrII mediates equimolar exchange of Na+ for K+ ions. (iii) Potassium uptake by KtrII requires the cells to generate ATP but does not depend on a proton potential. (iv) KtrII is expressed by cells growing in sodiumrich media, particularly under conditions that limit the generation of a normal proton potential; it is not produced in response to potassium deprivation nor is it repressed by excess potassium. We infer from these characteristics that KtrII is quite unlike the Kdp system of E. coli, functionally as well as mechanistically. KtrII cannot scavenge traces of K+ from a deficient medium; its function is to permit growth under conditions that render KtrI inoperative. The conditions that induce KtrII expression suggest that a rise in the cytoplasmic Na+ level, rather than a lowered level of K+, is the effective signal.
We began these studies in the expectation that KtrII was a porter of some kind, possibly a Na+/K+ symporter that allows the cells to accumulate K+ in response to a sodium gradient established by the sodium pump. The data presented above are clearly incompatible with this notion. Fig. 1A shows that the cells maintain a gradient of potassium concentration, [K+]i/[K+],, of 5000, while [Na+]i/[Na+], is about 10; the sodium gradient is in the wrong direction to contribute to the driving force upon a symporter. A Na+/K+ antiport mechanism might be consistent with Fig. 1 but would not explain K+ accumulation under the conditions of Fig. 2A in which [Na+Ii is lower than [Na+], (data not shown). The electrical potential is too small to rescue the hypothesis that KtrII is a o 20 40 60 80 100 120

14;
Time (mid FIG. 4. Net K+ efflux from AS25. Organisms were grown on NaTY, loaded with Na+, and resuspended in 50 mM Na-Tricine buffer, pH 8.5, at 1 mg of cells/ml. Glucose (10 mM) was added at -20 min, 0.5 mM KC1 at 0 min. Subsequent treatments were as follows: 0, control, no additions; H, plus 0.2 mM DCCD at 50 min; 0, at 50 min, diluted the cell suspension 10-fold with Na-Tricine buffer plus glucose; A, a parallel cell suspension in Na+-Tricine buffer plus glucose and 0.05 mM KCI, at a density of 0.1 mg of cells/ml. At this time (0 min), "K+ of high specific activity was added to both suspensions, and turnover of the pool was monitored; the cells' K+ content remained constant for the duration of the experiment. The following rate constants were calculated control, influx 5.4 X lo-* min", efflux 2 X rnin"; plus DCCD, influx 8.6 X lo-' min-', efflux 6 X lo-* rnin".
porter. Glycolyzing cells of AS25 grown on NaTY do generate a membrane potential of about -50 mV under certain conditions, by an unknown mechanism. Addition of DCCD and/or TCS reduced this to -30 mV, a potential gradient too small to account even for the reduced K+ accumulation shown in Fig. 1, B and C. Our experiments gave no hint of the participation of chloride or of other ions. If K+ accumulation is to be attributed to either symport or antiport with Na+ ions or protons one must postulate either a strange stoichiometry or else a localized ion current. On balance, we conclude that KtrII is not a porter but a primary transport system in which a chemical reaction supplies the driving force to exchange Na+ for K+.
The obvious candidate for such a reaction is the sodiumstimulated ATPase discovered by Heefner and Harold (1982). Two lines of argument support the identification of KtrII with this sodium pump. (i) Growth conditions that lead to the expression of KtrII consistently induced enhanced levels of the sodium-stimulated ATPase (Table I; Kinoshita et al. (1984). (ii) Mutant 7683, which lacks the sodium-stimulated ATPase, lacks KtrII; and revertants that recover the former also possess the latter. There is also a qualitative correlation between the level of Na+-ATPase and the initial rate of K+ uptake (Table 11). Previous papers from this laboratory suggested that the sodium-stimulated ATPase mediates exchange of Na+ for H+ Heefner and Harold, 1982), but the evidence bearing on the identity of the counterion was never strong.The present results suggest that, at least in AS25, the enzyme exchanges Na+ for K'. We cannot distinguish between two possible formulations. (i) The Na+-ATPase has relatively low specificity with respect to its counterion, accepting H+ in place of K+ under some conditions. For all we know, even Tris may be a low-affinity substrate. On this view, what we call KtrII is simply the expression of high levels of the Na+-ATPase. (ii) It remains possible that KtrII is a modified form of the Na+-ATPase that exchanges Na+ for K+ rather than for H+. We prefer the former hypothesis because it is simpler.
If it were true that the Na+-stimulated ATPase of AS25 or 9790 catalyzes exchange of Na+ for K', one should expect cells grown under the appropriate conditions (Table I) to contain a membrane-bound ATPase that is stimulated synergistically by Na+ and K+ ions. We have confirmed the finding of Kinoshita et al. (1984) that the level of the ATPase is markedly elevated in AS25 but observed little or no stimulation of its activity by K+ ions. This must be counted as an objection to our hypothesis, but not necessarily a fatal one. Under the assay conditions presently employed, even Na+ ions stimulate the rate of ATP hydrolysis by less than 2-fold, and there is a high basal activity that requires neither Na+ nor K+. Whether the in vitro activity of the Na+-ATPase is representative of that in the intact cells remains open to doubt. We suspect that the coupling between K+ and Na+ ions is less direct than it is in the ouabain-sensitive ATPase of animal cells but will refrain from pursuing this speculation. We note in passing that our conclusions concerning the nature of KtrII are virtually identical to those drawn by Benyoucef et al. (1982aBenyoucef et al. ( , 1982b regarding the uptake of K+ ions by Mycoplasma myeoides; in these organisms, also, exchange of Na+ for K+ appears to be mediated by an ATPase that is stimulated only by Na+ ions. This mutant has a defective FIFo-ATPase (1). KtrI (2) is assumed to be a potassium/proton symporter that functions only when it is phosphorylated. In the parent strain, 9790, the FIFo-ATPase generates the proton potential that supports K+ accumulation via KtrI. AS25 cannot sustain a large proton potential; potassium uptake by this strain depends on KtrII (3), the Na+-stimulated ATPase that exchanges cytoplasmic Na+ for external K+. It is not clear whether the Na+-ATPase is specific for K+ or can also accept H+ and other ions. KtrI functions chiefly in the direction of K+ efflux, setting a limit to the potassium concentration gradient. There must be at least one route for Na+ ions to enter the cell (4); the nature of this pathway is not known.

in Streptococci 2091
Another doubtful matter is the relationship between the KtrII, the vanadate-insensitive Na+-ATPase, and the exquisitely sensitive K+-stimulated ATPase isolated by Hugentobler et al. (1983). When intact streptococci were loaded with vanadate by incubation with 0.1 mM vanadate plus glucose, KtrI and KtrII activities were both unaffected, but calcium extrusion was blocked? In our view, the ATPase of Hugentobler et al. (1983) corresponds to neither of the two potassium transport systems known to exist in S. faecalis.
Returning now to the physiology of the cell, what is the nature of the K+ efflux pathway that sets a limit to the K' concentration gradient attained by KtrII? The existence of an efflux system was reported by Bakker and Harold (1980), who noted the rapid loss of K+ from cells glycolyzing in the presence of DCCD or of proton-conducting ionophores. Much earlier, Harold et al. (1967) isolated a mutant of S. faecalis that was leaky to K+; yet the only known defect in that mutant is a reduced ability to extrude protons (Harold and Papineau, 1972). A simple hypothesis that accommodates all these observations is depicted in Fig. 6. It shows KtrI as a K+/H+ symporter that responds to the proton potential but is regulated by phosphorylation, as several authors have suggested. Under conditions that depolarize the membrane but allow continued ATP generation, KtrI would catalyze K+ efflux and thus work against KtrII. It must be added that there is no direct evidence that KtrI catalyzes K+ efflux under our conditions or even that it mediates symport. Fig. 6 is intended to depict the minimal number of entities required to explain the results.
To account for K+ accumulation by KtrII we must invoke an additional element, a pathway for Na+ ions to enter the cells. The existence of at least one such pathway in S. faecal& is established , but its nature is unclear; Fig. 6 shows a dashed arrow to acknowledge this fact. Ill-defined pathways for Na+ entry have also been postulated in other bacteria (Padan et al., 1981;Krulwich, 1983;Booth and Kroll, 1983); they seem likely to play a major role in the homeostatic functions of ion transport.