Potassium transport into plant vacuoles energized directly by a proton-pumping inorganic pyrophosphatase

Proc. Natl. Acad. Sci. USA Vol. 89, pp. 11701-11705, December 1992 Cell Biology

Potassium transport into plant vacuoles energized directly by a proton-pumping inorganic pyrophosphatase (potassium accumulation/tonoplast energization/patch clamp)

JULIA M. DAVIES*, RONALD J. POOLEt, PHILIP A. REAf, AND DALE SANDERS*§ *Biology Department, University of York, York YO1 SDD, United Kingdom; tBiology Department, McGill University, 1205 Avenue Dr. Penfield, Montreal,

Quebec H3A iB1, Canada; and tPlant Science Institute, Department of Biology, University of Pennsylvania, Philadelphia, PA 19104

Communicated by Emanuel Epstein, August 24, 1992

ABSTRACT Potassium is accumulated in plant vacuoles against an inside-positive membrane potential. The mechanism facilitating energized K+ transport has remained obscure. However, electrogenic activity of the inorganic pyrophosphatase (H+-PPase) at the vacuolar membrane is dependent on cytoplasmic K+, raising the possibility that the enzyme translocates K+ into the vacuole. Membrane currents generated by the H+-PPase were measured (using a patch clamp technique) in intact vacuoles isolated from Beta vulgaris storage tissue. A significant orthophosphate-dependent outward current mediated by the enzyme in reverse mode is evoked only when potassium is present at the vacuolar face of the tonoplast, suggesting that potassium is a translocated ion. Furthermore, current-voltage analysis of the effects of extravacuolar potassium and pH on the reversal potential of the H+-PPasegenerated current points to direct translocation of K+ and H+ by the enzyme. Thus the H+-PPase represents a distinct class of eukaryote translocase and could facilitate vacuolar K+ accumulation in vivo.

Potassium is a major nutrient for all plants. Although cytoplasmic K+ is stable at around 100 mM, accumulation of the ion (as an osmotic regulator) in the main storage compartment of plant cells-the vacuole-frequently results in vacuolar concentrations as great as 200 mM in K+-replete tissues (1) and possibly 500 mM in open stomatal guard cells (2). Vacuolar K+ accumulation is opposed by an inside-positive membrane potential (AT) of about +20 mV (3), thus necessitating energized K+ transport across the vacuolar membrane. H+-coupled transport systems that energize accumulation of Na+ (4, 5) and Ca2+ (6) by the vacuole have been characterized but the mechanism for energized K+ transport remains obscure. Two primary H+ pumps reside at the vacuolar membrane-an ATPase (H+-ATPase; EC 3.6.1.3) and an inorganic pyrophosphatase (H+-PPase; EC 3.6.1.1) (7). The transport activity of the H+-PPase is dependent on the presence of K+ ions (8) at its cytosolic face (9), which raises the possibility that the enzyme translocates K+ into the vacuole. If the H+-PPase were to translocate K+ into the vacuole according to the generalized reaction:

n[H ]c + m[K+]c + PPi

-

n[H+]v + m[K+]v + 2Pi, [Il

then the reversal potential of the enzyme-generated current (Erev: the point at which enzymatic activity is poised at equilibrium) is given as: RT

Erev= (n + mOF I

Kpp [PPi][H+]'[K+]m

where [] signifies chemical activity; PP1 is inorganic pyrophosphate; P1 is inorganic orthophosphate; subscripts c and v refer to cytoplasmic and vacuolar compartments, respectively; n and m are, respectively, the stoichiometric ratios of H+ and K+ translocated per PP1 hydrolyzed; Kpp is the equilibrium constant for PP1 hydrolysis, and R, T, and F have their usual meanings. Erv is therefore a function of the prevailing trans-tonoplast K+ gradient, if the enzyme translocates that ion. Here, we present a thermodynamic analysis of H+-PPase-generated membrane currents (measured by the patch clamp technique), which demonstrates that the enzyme may be directly responsible for K+ accumulation in plant vacuoles.

MATERIALS AND METHODS Vacuole Preparation and Patch Clamp Media. Sugar beet plants (Beta vulgaris var. Regina M49) were grown and storage tissue vacuoles were isolated as described (9). Where stated, commercially grown red beets were obtained and stored at 4°C for up to 5 days before use. The basal bathing medium comprised (in mM) 1 MgCl2, 0.1 EGTA, 20 Tris-Mes (pH 8) with additions as follows: In Pi-reversal experiments,

100 choline chloride (Bi), 100 choline chloride/10 H3PO4 (B2); in experiments on the effect of extravacuolar K+ on Erev, 30 KCl (B3), 13.35 KC1/6.65 K2HPO4/3.35 KH2PO4/0.1 Tris-PPi/16.65 Trizma-Cl (B4), 100 KCl (B5), 83.35 KC1/6.65 K2HPO4/3.35 KH2PO4/0.1 Tris-PPi/16.65 Trizma-Cl (B6); in experiments on the effect of extravacuolar Cl- on Ercv, KCI in solutions B3-B6 was replaced by potassium gluconate, thus giving solutions B7-B10; in experiments on the effect of extravacuolar pH on Erev, solutions B7 and B8 were employed and, after adjustment to pH 7.5, as solutions B11 and B12. Vacuoles were isolated in control (PP, and Pi-free) solutions. The basal pipette solution comprised (in mM) 1 MgCl2, 0.1 EGTA, 20 Tris-Mes (pH 6) with the following additions: in Pi-reversal experiments, 100 choline chloride (P1) or 100 KCl (P2); in K+-Erev experiments, 30 KCl (P3) or 100 KCl (P2); in "pH" or "Cl-" experiments, 30 potassium gluconate (P4) or 100 potassium gluconate (P5). D-Sorbitol (>99%o purity, Fluka) was added to the bathing and pipette solutions to raise their osmolarities 100-150 mosM above that of an expressed cell sap sample. The pH values of bath and pipette solutions were recorded after osmotic adjustment. The concentrations of PP, and Pi (0.1 and 10 mM, respectively) were selected empirically to yield a value of Erev within measurable range. Fresh vacuoles were prepared for each experiment. Experiments were performed at room temperature (20-25°C). Abbreviations: H+-PPase, H+-translocating inorganic pyrophosphatase; H+-ATPase, H+-translocating adenosine triphosphatase; AT, membrane potential; PP1, inorganic pyrophosphate; Kpp., equilibrium constant for PPi hydrolysis; Kef, reference equilibrium constant; Erev, reversal potential; Ej, liquid junction potential; ( ), cytoplasmic concentration; ()v, vacuolar concentration; [ ]c, cytoplasmic activity; ([ 1, vacuolar activity; I-V, current-voltage. §To whom reprint requests should be addressed.

Pi]2[H+]n[K~l ]

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

11701

11702

Proc. Nadl. Acad. Sci. USA 89 (1992)

Cell Biology: Davies et al.

A

B C

4

V,mV

FIG. 1. Effect of(K+), on the Pi-dependentI-Vrelationships inred beet vacuoles. (A) Whole membrane I-Vrelationships obtained in solution B1 (o) and then solution B2 (o)-i.e., 100 mM choline chloride and then 100 mM choline chloride plus 10 mM Pi. The patch pipette contained solution P2 (100 mM KCI). I-V relationships were adjusted by the following values to compensate for liquid junction potentials (mV; values in parentheses are the ±f50%Yo error values, respectively): o, -7 (-6.5, -7); *, -7 (-7, -7.5). First-order polynomials were used to fit the data. (B) Derived I-V difference relationship from A for (K+)v = 100 mM (--- -). The derived I-V difference relationship is also indicated for a separate experiment (-) where (KW)v = 0 (solution P1, 100 mM choline chloride) but bathing solutions were as given for A. In the latter case, whole membrane I-V curves were adjusted by 0 mV (0, 0; solution B1) and -0.5 mV (-0.5, -0.6; solution B2).

Patch Clamp Protocol. Whole vacuole [analogous to '"whole cell" (10)] current-voltage (I-V) relationships were measured first in the absence, and then in the presence, of PPi and Pi (or Pi alone). Where applicable, the entire procedure was then repeated on the same vacuole after changing the K+ concentration orpH of the bathing medium. The patch clamp protocol used to obtain the desired whole-vacuole configuration and the perfusion system used were as described (9). The bath reference electrode (List Electronics, Darmstadt, F.R.G.) was filled with the pipette solution. The patch pipette potential was adjusted for zero current flow prior to the establishment of a seal. The vacuole was perfused for 4 min in the appropriate medium prior to voltage clamping. Clamping pulses were delivered from a patch clamp amplifier (EPC7, List Electronics) in 10-mV increments as a bipolar staircase from a 0-mV baseline, using commissioned software or PClamp (Axon Instruments, Foster City, CA). Pulse duration was 4 s; current was sampled from the last 1 s and filtered at 3 kHz. All I-V recordings were performed during perfusion. Analysis of I-V Relationships. Whole membrane I-V relationships were adjusted to compensate for the uncompensated liquid junction potential (Ej) at the reference electrode, according to ref. 11, using the generalized Henderson equation. Mobility constants were taken from ref. 12, derived from limiting conductance data (13, 14), or estimated from conductance data on the basis of carbon chain length (Tris, Mes, gluconate). Activity coefficients for inorganic ions were taken from ref. 15. For the organic ions EGTA, gluconate, Mes, and Tris, activity coefficients were taken to be those of inorganic ions of the same valence. The impact of error on the Ej compensation estimates was assessed by calculating ±50+% error values for each estimate. These were calculated by

assuming a simultaneous ±50%o error in all activities (except H+) and estimated mobility constants. Empirical nonlinear least square fits (16) using a third-order polynomial (unless otherwise stated) were made for each I-V relationship. The line fitted to the control I-V was subtracted from that of the test (+PPj/P1 or +Pj) to yield the H+-PPase (difference) I-V relationship. The E.,, was thus the interpolated zero current intercept. This subtraction procedure obviated the effect of changes in K+ on channel currents. For difference I-V relationships any errors in the estimates of Ej were eliminated with the exception of the effects of PP1 and Pi on Ej. Estitio

of the EquilIbrium Constant for PP1 Hydrolysis

(Kp,). Individual values of Kpp; were calculated for each ionic

condition by defining initially a reference equilibrium, as described in refs. 17 and 18. The reference reaction was: HP207 = 2HPO2- + HW.

[3]

Evaluation of the equilibrium constant for the reference reaction (Kj enabled calculation of the apparent Kpp. for any given condition of pH, free Mg2+, free K+, etc. K:f was calculated from the thermodynamic data of ref. 19 (table II, cases 12 and 13) by computing the activities of the two complexes from the association constants of PP1 and Pi, specified in refs. 20-22. The mean derived value of Kf was 7770 M. Apparent values for Kpp were then determined for any given condition by substitution into the relationship:

[HP20-]/[MPPjd iallPo24-]/[pj)2

Table 1. Effects of ionic gradients on the Em, of the vacuolar H+-PPase Major (K+)C, (K+)c, (K+)v, AEo,, mV mM mM mM anion pHc pHc pHv Ere, mV ± 2 (6) ± +21 8.30 100 +1 4 (6) 30 8.30 30 6.23 Cl-18 ± 2 (6) 30 8.20 +7 ± 3 (6) 6.25 100 8.16 100 Cl-22 ± 2 (4) 8.17 +4 ± 4(4) 30 6.25 100 8.22 100 Gn30 7.56 +20 ± 2 (4) 6.20 30 8.09 -13 ± 10 (4) 100 GnThe effects of changes in cytosolic K+ concentration and pH-(K+)c and pH¢, respectively-on the Er,,v of the H+-PPase are indicated. The first value of En,,v is absolute and the second (AEres) shows the mean change in response to the new ionic conditions for that set of vacuoles. Enrv is cited in the form mean ± SEM (number of vacuoles sampled). Individual results were from separate sugar beet preparations. Gn-, gluconate.

[4]

Cell Biology: Davies

et

A

where E specifies the total PP1 and Pi present regardless of ionic form. Molar fractions were again computed for each experimental condition. The extreme lower and upper estimates for Kref derived from ref. 19 in similar ionic conditions to those deployed in the present study were 6881 and 8659 M, respectively; these values generated ±11% uncertainty in the derived values of Kpp. Resulting values of Kpp were substituted into Eq. 2 to yield predicted Erev values. K+ activities were estimated using activity coefficients from ref. 15 at 25TC and appropriate ionic strength. The accuracy of Erev estimates was ± 1 mV (through uncertainty in derived Kpp. values) for models invoking symport of K+ and H+ and ±3 mV for uniport models. Predicted values of AE,,v in response to a change in ionic background conditions were not affected by uncertainties in Krf.

RESULTS Reversal of the H+-PPase by Pi. With 100 mM K+ present at the vacuolar face of the tonoplast (solution P2) but 100 mM choline chloride at the cytoplasmic face (solution Bi), the addition of 10 mM Pi to the bathing medium (solution B2) evoked an outwardly directed current from red beet vacuoles. Fig. 1A illustrates the I-V relationships typical for this set of ionic conditions. No Er,- for the Pi current was evident in trials on four individual vacuoles (Fig. 1B). The mean Pi-dependent current, for an individual vacuole, over a voltage clamp of -50 to +50 mV ranged from +1.8 ± 0.1 (±SEM) to 15 ± 1 mAm-2 (calculated as the mean of the Pi-dependent current value at each discrete clamped voltage). The overall mean current from the four trials was +6.9 ± 1.2 mA m-2. All currents (as scalar values) are within the range reported for

converse, hydrolytic mode H+-PPase activity in B. vulgaris-i.e., with selective addition of PPi and K+ present at the cytoplasmic face (9). However, the currents in the two sets of experiments are of opposite polarity, indicating that the ion-translocating H+-PPase is kinetically reversible. The K+ dependence of the Pi-dependent outward current was investigated by substituting vacuolar K+ with choline. With 100 mM choline chloride at the vacuolar face (solution P1), addition of 10 mM Pi to the cytoplasmic face (solution B2) of red beet vacuoles evoked a far smaller outwardly directed current (Fig. 1B). The mean current over a -SO to +50 mV clamp ranged from +0.1 ± 0.01 to +0.5 ± 0.2 mAm-2, with an overall mean of +0.21 ± 0.2 mAm-2 (n = 4). It is suggested that this current represents either the passive transport of Pi into the vacuole or an effect of Pi on the liquid junction potential of the reference electrode. In either case, the current represents only a minor component of that observed in the presence of vacuolar K+, thereby confirming that the reverse mode (outward) current through the H+-PPase is dependent on the presence of vacuolar K+. These results therefore complement those demonstrating an obligatory dependence of the PPi-elicited inward current on cytoplasmic K+ and are anticipated if the H+-PPase does indeed translocate K+. Influence of the Trans-Tonoplast K+ Gradient on Er,. A definitive investigation of whether the H+-PPase translocates K+ must address the influence of K+ on the thermodynamic properties of the enzyme rather than simply its kinetic properties. Therefore we investigated the effect of ionic gradients on the Erev of the H+-PPase. Typical whole vacuole I-V relationships for a vacuolar concentration- (K+),-of 30 mM are shown in Fig. 2 A and B. The difference relationships in response to joint addition of PPi and Pi are derived in Fig. 2C and demonstrate a clear positive shift in Et, in response to an increase in K+ concentration in the bathing medium(K+)c-from 30 to 100 mM. The results of six independent trials confirming this result are summarized in Table 1 (line 1). Table 1 also shows that the measured E& was dependent on

Proc. Natl. Acad. Sci. USA 89 (1992)

11703

the K+ gradient (as anticipated by Eq. 2), not the absolute levels of K+. Thus with (K+)X and (K+)c set to 100 mM, the value of E,,, was not significantly different from that observed in the presence of equimolar 30 mM K+. Moreover, A

Whole membrane 40 EE

30 K

pH 8.3

0K H 6.2

E

I20 -

(PPi+pi) Control

-60

v

-29

20

40

V

60

(mV)

-20 -

-40 -

B

40-

E E

100 K

20 -

pH 8.3

(PP,+P')

°C ont rolI

20-

-60

-40 I-

20

.

40

60

V (mV)

-20 -

-40

C

-

Difference - 20 E

E

V (mV)

-20

FIG. 2. Effect of (K+), on (PPi plus P0-dependent I-V difference relationships of a single sugar beet vacuole. The patch pipette contained solution P3 (30 mM KCI). (A) Whole membrane I-V curves were measured in bathing solution B3 (v) and then solution B4 (v)-i.e., 30 mM KCI and then 30 mM KCI plus (PPi/P?, respectively. (B) Whole membrane I-V curves were then measured from the same vacuole in bathing solution B5 (O) followed by solution B6 (M)-i.e., 100 mM KCI and then 100 mM KCI plus (PPi/PO, respectively. (C) PPi-dependent I-V relationships for (K+)c = 30 mM(-- -) and (K+)c = 100 mM (-) were derived as given in the text. Whole membrane I-Vrelationships were adjusted (mV) as follows; v, 0 (-O, +0); v, -3 (-4, -3); O, 0 (0, +5); *, -2 (-2, +4) (values in parentheses are the ±-50% error values, respectively). The abscissa intersects are the Ere, values of the H+-PPase; there is a 28-mV positive shift as extravacuolar K+ increases.

11704

Cell Biology: Davies et al.

when (K+)c was set to 30 mM, E,.v became -11 ± 2 mV: the change in Erev is therefore negative rather than positive, though the absolute magnitude of the shift is similar to the value obtained for this change in (K+)c when (KX)v is 30 mM. These effects of K+ on the Ercv of the H+-PPase therefore suggest independently that K+ is translocated by the enzyme. The possibility that the Erv shifts were due to the concomitant changes in (Cl-)c with (K+)c is refuted by the results of experiments in which KCl was replaced by potassium gluconate. Table 1, line 3, shows that for (K+)v = 100 mM, gluconate recordings yielded a good agreement with the analogous chloride experiments for the absolute Erv value in equimolar K+ and the magnitude of the Erev change when (K+)c = 30 mM. Influence of the Trans-Tonoplast pH Gradient on Erm. H+ translocation by the H+-PPase requires that the Ev of the enzyme also responds in a thermodynamically consistent manner to a change in transmembrane pH difference. Fig. 3 A and B illustrate the effect on the whole membrane I-V relationship of lowering the pH of the bathing medium (pHJ) by about 0.5 unit. Again, as anticipated by Eq. 2, Erv moved positive as [H+]c was increased. In Fig. 3C there is a 26-mV positive shift in Erev as extravacuolar pH decreases. Table 1, line 4, summarizes the results from four independent determinations: for the conditions pHc 8.0, (K+)v = 100 mM, (K+)c = 30 mM, the absolute Er., value is in agreement with that observed from the analogous experiments in which (K+)c was shifted. The mean change in Erev was 20 ± 2 mV. Thus, the enzyme appears to translocate K+ and H+. Stoichiometric Translocation Ratios. Quantitative stoichiometric translocation ratios for K+ and H+ by the H+-PPase can, in principle, be obtained from the absolute values of Erev and from the magnitude of the shift in Erv as K+ and H+ are selectively changed. Substitution of the parameter values into Eq. 2 for various integer values of the translocation ratios m and n, for K+ and H+, respectively, between 1 and 2 fails to yield any consistent prediction of the experimental results (Table 2). In particular, the observed values of E&v fall well outside the range predicted for sole translocation of either K+ or H+. However, the absolute values of Erv are embraced by the theoretical possibilities m = 1, n = 2 and m = 2, n = 1. Therefore we investigated the possibility that translocation occurs with non-integer stoichiometries specifically for K+ and H+ (albeit with an integral stoichiometry of 3 for overall ion translocation). Excellent predictions of all the experimental data result if the stoichiometric translocation ratio with respect to PP1 hydrolysis is taken as 1.3HI:1.7K+ (Table 2). If uncertainties in the bona fide value of Kpp. are taken into account (Table 2, legend), stoichiometries of 1.2H+:1.8K+ and 1.4H+:1.6K+ are equally in accord with our measurements, although ratios outside this range are not feasible for the particular ionic conditions deployed in the present study.

DISCUSSION The hydrolysis of PP1 and the inward translocation of positive charge by the HI-PPase are known to depend specifically on the presence of K+ at the cytoplasmic face of the vacuole (9) and are independent of the vacuolar K+ status. A strong indication that K+ is translocated directly by the H+-PPase emerges from the observation that a P-dependent outward current of similar magnitude to that generated by the enzyme in its hydrolytic mode could be obtained only when K+ was present at the vacuolar tonoplast face. Rejection of this interpretation of the P1 effect must result in the adoption of less incisive, more complex interpretations that invoke the existence of such transporters as a Kv-dependent Pi translocator. The role of the H+-PPase in K+ translocation is confirmed by the observation that manipulation of the transmembrane K+ gradient has a direct impact on the E,,, of the ion pump.

Proc. Nati. Acad. Sci. USA 89 (1992)

A

Whole membrane

(PP, +Pj)

30 K pH 8.09

40 60 V (mv)

20

-40

B

E

30 K

pH 7.56

-60

40-

E -20 20 -

0l

K pH 6.2

A

-40

(Pp1+p1) - - - a~Control

20

60 V (mV)

40

-20 -

-40 -

C

Difference C, 20E pHC = 8.09

E

pHC = 7.56

E reV". -40

,

E rev

20

,

40

V

-20

(mV)

-

FIG. 3. Effect of bathing medium pH on the (PPi plus Pi)dependent l-V difference relationships in a single suga beet vacuole. The patch pipette contained solution P5 (100 mM potassium gluind u conate). (A) Whole membrane I-V relationships were bathing solution B7 (i) and then solution B8 ().e., 30 M potassium gluconate (pH 8.09) and then 30 mM poassum glcoae plus (PPi/PJ) (pH 8.09). (B) Whole membrane I-V relti were then obtained from the same vacuole in bathing solution B11 (A) followed by solution B12 (v)-i.e., 30 mM potassium gnate (pH (pH 7.56). 7.56) and then 30 mM potassium gluconate plus (PP (pH) I-V relationships were adjusted by the following values to compensate for liquid junction potentials (mV; values in parentheses are the ±50%o error values, respectively): o, -10 (-5, -16); , -10 (-6, -16); &, -10 (-5, -16); A, -10 (-6, -16). (C) Derived I-V difference relationships for 30 mM bathing K+, pH 8.09 (--- -) andpH 7.56 (-). There is a 26-mV positive shift in Em,, as extravacuolar pH decreases.

Previous studies have shown that the H+-PPase is elec-

trogenic (23) and mediates intravacuolar H+ accmulation (24). With the enzyme pumping K+, the possibility arises that

Cell Biology: Davies et al.

Proc.

Nadl. Acad. Sci. USA 89 (1992)

11705

Table 2. Determination of transport stoichiometry of the vacuolar H+-PPase by comparison of calculated with observed (Obs) values of Er Emv or AErev, mV Calculated for n:m ratios Ionic composition E 1:0 0:1 1:1 1:2 2:1 2:2 1.3:1.7 Obs Chloride -2 +1 (K+)v = 30, (K+)c = 30 Erev +31 +153 +15 +10 -31 -23 +6 +21 +38 +19 +23 +13 +17 +19 (K+)c= 100 AErev +7 +7 (K+)v = 100, (K+)c = 100 Emv +56 +169 +28 +19 -19 -14 -18 (K+)c = 30 Mr, -17 -40 -14 -24 -15 -19 -21 Gluconate +4 +4 (K+)v = 100, (K+)c = 100 Erev +47 +163 +23 +16 -23 -18 -1 -33 -15 -20 -9 -14 -17 -22 AEM, (K+)c = 30 -2 -30 -30 -11 -13 pHv = 6.20, pHc = 8.09 Erev +52 +134 +11 +24 +21 pHc = 7.56 AEev +62 +31 +31 +20 +31 +24 Calculated Emv values for a (nH+, mK+)-PPase are given for the eight experimental conditions of Table 1. Calculations are based on Eq. 2 and details of ionic conditions (where not specified) are given in the legend to Table 1. Kpp, values (in M) used were, respectively, by line in the table, as follows: 386, 529, 713, 479, 574, 497, 593, 1940. Note that since Kpp1 increases steeply as a function of decreasing pH, the predicted changes in E&, (bottom line) are considerably larger than would be anticipated on the basis of considerations of transmembrane pH gradient alone.

PPi-dependent H+ accumulation is indirect, the result of enhancement of a hypothetical K+/H+ exchange. That possibility may be discounted because Er, responds to a change in extravacuolar pH as it does to a change in K+. If, as suggested here, K+ and H+ are translocated by the PPase, then the observed values of Ere, must correspond to the predictions of Eq. 2. Although the observed AEO, values agree well with those predicted for a simple 1H+:lK+ stoichiometry, the observed absolute E,,, values are far more negative than predicted. If PP1 or Pi were to elicit an outward leak current, the latter would cause any absolute Erev to appear more negative than its true value. AEr, values would remain unchanged. However, neither PP1 (9) nor Pi (Fig. 1) has been shown to induce a significant leak current. Good agreement between experimental and predicted values of Erey arises if the stoichiometric translocation ratio is taken to be 1.3H+:1.7K+. These non-integer values need not, of themselves, invoke any special mechanisms for ion translocation. Though the 3-fold changes in transmembrane [K+] and [H+] gradients used in the present work preclude a confident appraisal of the catalytic mechanism, a scheme involving competition between K+ and H+ for binding sites and translocation of a constant number-3-of positive charges per catalytic cycle might be envisaged. On the basis of experiments monitoring net H+ translocation by the H+-PPase, the physiological poise of the enzyme is thought to be in the direction of PP1 hydrolysis (24)-a point reinforced by the finding that this enzyme probably represents the sole mechanism for disposal of cytosolically produced PP, in the plant cell (25). These claims can be reevaluated for the stoichiometric ratio of K+ and H+ translocation proposed here. Taking cytosolic free Mg2+ as 0.4 mM (26) and cytosolic pH as 7.2 (27) for a typical plant cell, Kpp, can be calculated as 7367 M. Thus if cytosolic [PP,], [Pd, and [K+] are 0.25 mM (25), 5 mM (28), and 100 mM (1), respectively, an in vivo value of Erev can be estimated (from Eq. 2) as +42 mV for n = 1.3 and m = 1.7 when pHv is 5.5 (29) and [K+]v is 200 mM. For n = 1 and m = 1, E,,v would become +84 mV. At a typical transmembrane electrical potential of +20 mV, the H+-PPase would clearly operate in a hydrolytic mode as a K+-H+ translocator and so facilitate vacuolar K+ accumulation. Though a vacuolar electroneutral K+-H+ antiporter has been reported (30), its activity in vitro is inhibited at K+ concentrations of >25 mM and so its

physiological significance remains in doubt. Thus, as the H+-PPase and H+-ATPase appear to catalyze different transport reactions, the long-standing question on the coresidence of two primary pumps in the plant vacuolar membrane (7) may be resolved. We thank Colin Abbott for plant cultivation, Ian Jennings and Neil Mackay for software development, and David White for provision of additional software. Financial support came from the Agricultural and Food Research Council (U.K.) and in part from the Natural Science and Engineering Research Council (Canada). 1. 2. 3. 4. 5.

6. 7. 8.

9. 10. 11. 12. 13. 14.

15. 16. 17. 18. 19. 20.

21. 22. 23. 24.

25.

Leigh, R. A. & Wyn Jones, R. G. (1984) New Phytol. 97, 1-14. Wilmer, C. M. (1983) Stomata (Longman, London). Spanswick, R. M. & Williams, E. J. (1964) J. Exp. Bot. 15, 193-200. Blumwald, E. & Poole, R. J. (1985) Plant Physiol. 78, 63-67. Fan, T. W.-M., Higashi, R. M., Norlyn, J. & Epstein, E. (1989) Proc. Nad. Acad. Sci. USA 86, 9856-9860. Blackford, S., Rea, P. A. & Sanders, D. (1990) J. Biol. Chem. 256, 9615-9620. Rea, P. A. & Sanders, D. (1987) Physiol. Plant 71, 131-141. Wang, Y., Leigh, R. A., Kaestner, K. H. & Sze, H. (1986) Plant Physiol. 81, 497-502. Davies, J. M., Rea, P. A. & Sanders, D. (1991) FEBS Lett. 278, 66-68. Hamill, 0. P., Marty, A., Neher, E., Sakmann, B. & Sigworth, F. J. (1981) Pfluegers Arch. 391, 85-100. Barry, P. H. & Lynch, J. W. (1991) J. Membr. Biol. 121, 101-117. Amman, D. (1986) Ion-Selective Electrodes (Springer, Berlin). Dean, J. A. (1985) Lange's Handbook of Chemistry (McGraw-Hill, New York). Robinson, R. A. & Stokes, R. H. (1965) Electrolyte Solutions (Butterworths, London). Keilland, J. (1937) J. Am. Chem. Soc. 59, 1675-1678. Marquardt, D. W. (1963) J. Soc. Ind. Appl. Math. 11, 431-441. Rosing, J. & Slater, E. C. (1972) Biochim. Biophys. Acta 267, 275-290. Davies, J. M., Poole, R. J. & Sanders, D. (1992) Biochim. Biophys. Acta, in press. Flodgaard, H. & Hleron, P. (1974) J. Biol. Chem. 249, 3465-3474. Smith, R. M. & Martell, A. E. (1975) Critical Stability Constants (Plenum, New York), Vol. 2, pp. 281-283. Smith, R. M. & Martell, A. E. (1976) Critical Stability Constants (Plenum, New York), Vol. 4, pp. 56-59. Marteli, A. E. & Smith, R. M. (1982) Critical Stability Constants (Plenum, New York), Vol. 5, pp. 260-261, 407-408. Hedrich, R., Kurkdjian, A., Guern, J. & Flugge, U. I. (1989) EMBO J. 8, 2835-2842. Johannes, E. & Felle, H. (1990) Plant Physiol. 93, 412-417. Weiner, H., Stitt, M. & Heldt, H. W. (1987) Biochim. Biophys. Acta 893,

19-21. 26. Yazaki, Y., Asukagawa, N., Ishikawa, Y., Ohta, E. & Sakata, M. (1988) Plant Cell Physiol. 29, 919-924. 27. Felle, H. & Bertl, A. (1986) J. Erp. Bot. 37, 1416-1428. 28. Rebeille, F., Bligny, R. & Douce, R. (1984) Plant Physiol. 74, 355-359. 29. Fox, G. G. & Ratcliffe, R. G. (1990) Plant Physiol. 93, 512-521. 30. Cooper, S., Lerner, H. R. & Reinhold, L. (1991) Plant Physiol. 97, 1212-1220.