The Lysosomal Proton Pump Is Electrogenic*

Lysosomes were purified approximately 40-fold from rat kidney cortex by differential and Percoll density gradient centrifugation. In a sucrose medium, the lysosomes quenched the fluorescence of the potential sensitive dye diS-CJ-(5) (3,3’-dipropylthiocarbo- cyanine iodide) in a time-dependent manner, indicating that the dye accumulates within the lysosomal interior. After treatment of the lysosomes with valino- mycin, the dye fluorescence displayed a logarithmic dependence upon the external K+ concentration; thus, the fluorescence signal provides a semiquantitative measure of the lysosomal membrane potential (A*). In the absence of valinomycin, lysosomal quenching of diS-C3-(5) fluorescence was partially reversed by agents which collapse the lysosomal pH gradient (am-monium sulfate, chloroquine, and K nigericin), suggesting that the proton gradient across the lysosomal membrane contributes to A*. A rapid increase in diS-C3-(5) fluorescence, indicative of an increase in A*, was observed upon the addition of Mg-ATP to the lysosomes. The ATP-dependent fluorescence change was inhibited by protonophores, K valinomycin, permeable anions, and N-ethylmaleimide, but was un- affected by ammonium sulfate, K nigericin, or sodium vanadate. Oligomycin had no effect at concentrations below 2 pg/ml; at higher concentrations, oligomycin partially inhibited the fluorescence response to Mg- ATP, but it also inhibited the fluorescence response to K valinomycin, Car-negie-Mellon Miscellaneous Assays-The activities of the following enzymes were determined as markers for various subcellular organelles: 5"nucleo- tidase for plasma membranes (13); succinate-p-iodonitrotetrazolium violet reductase (13) and cytochrome c oxidase (14) for mitochondria; glucose 6-phosphatase (13) for endoplasmic reticulum; and arylsulfatase (15) and N-acetyl-0-glucosaminidase (15) for lysosomes. La- tency measurements were carried out as described earlier (16) using 0.1% Triton X-100 in the assay mixture for determining total activity. Protein was measured by the method of Lowry et al. (17).

hand, experiments with plant vacuolysosomes (8,9) suggest that activation of the proton pump by Mg-ATP causes an increase in the potential of the lysosomal interior of approximately 60 mV. Moreover, Ohkuma et al. (7) have stated that the lysosomal proton pump can generate a membrane potential under certain conditions, although the data supporting this conclusion are as yet unpublished.
In the present paper, we describe a new procedure for the isolation of lysosomes in high purity and good yield from rat kidney cortex. This procedure avoids the use of density altering agents such as Triton WR-1339 which could modify the permeability properties of the lysosomal membrane. Using the membrane potential sensitive fluorescent dye diS-C3-(5)' as well as measurements of permeable ion distributions, we have obtained evidence that the activation of the lysosomal proton pump with Mg-ATP causes a rapid increase in the lysosomal membrane potential. This ATP-dependent change in membrane potential is blocked by protonophores and by N-ethylmaleimide and is reduced in magnitude by the presence of permeable anions. The results provide strong evidence that the lysosomal proton pump is electrogenic.

MATERIALS AND METHODS
Preparation of Lysosomes-Male albino rats (Charles River CD) weighing 350-450 g were used in these studies; in general, the best yields of lysosomes were obtained with older rats. Rats were killed by decapitation, and the kidneys were dissected and chilled in cold 0.3 M sucrose containing 1 mM EDTA adjusted to pH 7.0 with Tris (sucrose/EDTA buffer). Renal cortex was separated with a razor blade, weighed, minced, and homogenized in sucrose/EDTA buffer with four strokes in a Potter-Elvehjem glass homogenizer with a Teflon pestle rotating at approximately 1000 rpm. The volume of the homogenate was adjusted to 10 ml/g of wet tissue and centrifuged at 270 X g (maximum) for 5 min. The sediment was discarded, and the supernatant was spun at 3000 X g (maximum) for 10 min in a Sorvall SS-34 rotor. The sediment thus obtained consisted of a dark greenish pellet of lysosomes surrounded by red blood cells and overlaid with a brown zone which contained mitochondria, as well as some nuclear debris and unbroken cells. The supernatant was removed by aspiration and discarded along with the loosely packed upper portion of the pellet, which slid down the sides of the tube during aspiration. The pellet was then resuspended in a large volume (20-fold or more) of sucrose/EDTA buffer and subjected to two consecutive spins at 480 X g (maximum) for 5 min so as to remove the erythrocytes which sedimented at this speed. The remaining supernatant was centrifuged at 3000 X g (maximum) for 10 min to yield the crude lysosomal preparation (heavy lysosomes). The pellet was resuspended in a minimum volume (300 pl/g of tissue) of sucrose/EDTA buffer and fractionated by density gradient centrifugation as described below.
A preformed density gradient of Percoll was prepared by spinning 14 ml of a gradient mixture consisting of 75% Percoll, 15% 2 M sucrose, and 10% 0.1 M Mops buffered to pH 7.4 with Tris at 48,000 X g (maximum) for 60 min in a glass centrifuge tube (33-34 rotor). The heavy lysosome fraction was layered on the gradient and centrifuged for 30 mi, at 48,000 X g (maximum). After centrifugation, the densest zone of turbidity ( Fig. 1, fraction 4 ) was collected by aspiration, suspended in a 10-fold excess of sucrose/EDTA buffer, and centrifuged at 3,000 X g (maximum) for 10 min to sediment the lysosomes. The pellet was resuspended in sucrose/EDTA buffer and centrifuged once again at 3,000 X g (maximum) for 10 min. Most of the Percoll was removed by these centrifugation steps because it did not sediment at such low speeds. The final pellet, which was dark green in color, was resuspended in sucrose/EDTA buffer at a final protein concentration of 5-10 mgjml. All the solutions used in this out at 4 "C.
procedure were chilled on ice, and all centrifugation steps were carried ion Distribution and ApH Measurements-A novel procedure was developed for the ion distribution measurements using ion exchange resins to trap the labeled ion in the lysosomal extracts. The advantage of this approach is that the total water space, the external water space (measured with [14CJsucrose), and the labeled ion content of the lysosomal pellets can all be determined with the same sample. Lysosomes (50 p1) were mixed with 50 p1 of 0.3 M sucrose, 50 mM Mops adjusted to pH 7.0 with Tris which contained 100 pCi/ml of 3H20, 40 pCi/ml of [U-'4C]sucrose, and any one of the following: 20 pCi/ml of ['4C]methylamine hydrochloride (56 mCi/mmol), 20 pCi/ ml of potassium thio[I4C]cyanate (56 mCi/mmol) 2-8 pCi/ml of ffiRbC1 ("0.1 mM), and 1-2 pCi/ml of [3H]tetraphenylphosphonium bromide (2.5 Ci/mmol). When =RbC1 was used, 0.25 pl of 2 mM valinomycin in dimethyl sulfoxide was added after the addition of the lysosomes. This mixture was incubated a t 25 "C for 10 min and then transferred to Eppendorf pipette tips which were heat-sealed at the narrow end. The tips were centrifuged in Fisher model 235A microcentrifuge for 2 min. The tube was cut transversely with a razor blade through the middle of the pellet, and the lower portion of the tube was transferred to 0.5 ml of 1% Nonidet P-40 (or, in some experiments, 1% Triton X-100). An aliquot (1 p l ) of the supernatant was transferred to a second tube containing 0.5 ml of the detergent solution.
Small columns (5-8-mm length) of Dowex 50-X (hydrogen form) or Dowex 1 (chloride form, used only for experiments with KS["C] CN) were constructed in 14.6-cm Pasteur pipettes plugged with cotton. The Nonidet P-40 extracts of the lysosomal supernatant and pellet fractions were applied to separate columns, and the ["C]sucrose and 3H,0 associated with the fractions were eluted with three 500-pl washes of distilled water; all washings were collected in a scintillation vial. Control experiments showed that three washes were enough to remove more than 98% of the 3H20 and [14C]sucrose applied to column; any labeled methylamine, SCN-, Rb', or TPP+ remained bound to the column under these conditions. Following the washing procedure, the tip of the Pasteur pipette was broken off and the ion exchange resin was washed into a second scintillation vial with two 0.5-mI aliquots of 2 M KCI. To each scintillation vial was added 10 ml of Aquasol (New England Nuclear), and the samples were counted on a Beckman LS 7800 scintillation spectrometer. All radioactive compounds were obtained from Amersham Corp. with the exception of [3H]TPP' which was prepared by the Isotope Synthesis Group of Hoffmann-LaRoche, Inc. under the direction of Dr. Arnold Liebman.
The number of counts for ['4C]sucrose and the labeled ion (methylamine, etc.) in the pellet extracts were divided by the number of counts for 3H20, and these values were divided in turn by the corresponding ratios for the supernatant to yield the pellet to supernatant distribution factor ( T ) , e. The accumulation factor ( F ) for the ion in question, i.e. the ratio of the intralysosomal to extralysosomal concentrations, was calculated as illustrated for methylamine (12): F(MA) = ( r M Ars)/(lrs) where rs is the pellet to supernatant distribution factor for ['4C]sucrose. The average (fS.D.) value for rs for 25 separate measurements was 0.47 f 0.17. The values for ApH were calculated as ApH = 60 log F(MA), and the membrane potential (A@) was calculated as A@ = -60 log F (TPP' or Rb') or A@ = 60 log F(SCN). The membrane potential is defined as the electrical potential of the lysosomal interior, with respect to the external medium.
Fluorescence Measurements-All fluorescence measurements were conducted at 25 "C at excitation and emission wavelengths of 622 and 670 nm, respectively, in a Perkin-Elmer MPF-4 fluorescence spectrometer. Lysosomes (3-5 pl) were added to quartz cuvettes containing 2 ml of 0.5 p M diS-C3-(5) in 0.3 M sucrose, 50 mM Mops buffered to pH 7.0 with Tris. The dye was added to the cuvettes as a 1 mM solution in ethanol. We are grateful to Dr. Alan S. Waggoner, Carnegie-Mellon Institute, for providing the diS-Cs-(5).
Miscellaneous Assays-The activities of the following enzymes were determined as markers for various subcellular organelles: 5"nucleotidase for plasma membranes (13); succinate-p-iodonitrotetrazolium violet reductase (13) and cytochrome c oxidase (14) for mitochondria; glucose 6-phosphatase (13) for endoplasmic reticulum; and arylsulfatase (15) and N-acetyl-0-glucosaminidase (15) for lysosomes. Latency measurements were carried out as described earlier (16) using 0.1% Triton X-100 in the assay mixture for determining total activity. Protein was measured by the method of Lowry et al. (17).

RESULTS
Preparation of Rat Kidney Lysosomes-As described in detail under "Materials and Methods," a crude lysosome suspension was obtained from the dark green bottom layer of the 3000 X g pellet of rat renal cortex homogenates. The lysosomes were separated from contaminating mitochondria on a preformed gradient prepared from 75% Percoll in 0.3 M sucrose. As indicated diagrammatically in Fig. 1, 97% of the mitochondrial marker enzymes present in the crude lysosomal preparations banded at a density of 1.098 g/cm3, whereas 76% of the lysosomal enzymes banded at a density greater than 1.145 g/cm3. The lysosomes from this region were centrifuged and resuspended several times in 0.3 M sucrose, 1 mM EDTA, pH 7.0, to remove Percoll and finally resuspended in the above buffer at 5-10 mg of proteinjml. Data on the purity of these lysosomal preparations are presented in Table I which shows the activities of marker enzymes for lysosomes (arylsulfatase), mitochondria (succinate-p-iodonitrotetrazolium violet reductase), endoplasmic reticulum (glucose 6-phosphatase), and plasma membranes (5'-nucIeotidase). As shown, the average (+S.D.) purification factor for arylsulfatase is  The values given are the average ( S . D . ) specific activities of n separate lysosomal preparations. Values given in parentheses are the percentage yield of total enzymatic activity in the lysosomal preparations. Arylsulfatase was assayed in the presence of 0.1% Triton Xnitrotetrazolium phatase 37.8 k 6.8 (n = 7 ) , with an overall recovery of 5.7 f 2.1% of the total activity in the crude homogenate. Similar values were obtained for N-acetyl-0-glucosaminidase, another lysosomal marker enzyme (data not shown). The activities of these enzymes were 80-90% latent, indicating that the lysosomes were isolated as intact organelles. In contrast, the specific activities of the other marker enzymes were considerably reduced in the lysosomal preparations compared to the crude homogenate.
These preparations showed only low levels of contamination by mitochondria, a reflection of the excellent separation of mitochondrial and lysosomal markers on the Percoll gradients ( Fig. 1). Thus, the specific activity of the mitochondrial enzyme succinate-p-iodonitrotetrazolium violet reductase was only 4% of its specific activity in the crude homogenate (Table  I). Since mitochondria comprise only a fraction of the total protein of the crude homogenate, it follows that mitochondrial contamination of the lysosomal preparations must be considerably less than 4%. In some preparations, cytochrome oxidase activity was also measured as a mitochondrial marker, with results similar to those obtained for succinate-p-iodonitrotetrazolium violet reductase (data not shown). The relative absence of mitochondria in these preparations was confirmed by electron microscopic observations (data not shown). The vast majority of the particles consisted of spherical membrane-bound electron-dense bodies with diameters ranging from 0.5 to 1.4 cm; mitochondrial profiles were observed with a frequency of one/l60 lysosomes. The results indicate that lysosomes can be obtained from rat kidney cortex in high purity and good yield following the procedure described under "Materials and Methods." Potential Sensitive Dye Fluorescence Measurements-The carbocyanine dye diS-Cs-(5) has been widely used to monitor the membrane potential in a variety of cellular and subcellular systems (18,19). The positively charged dye accumulates within compartments that are electrically negative with respect to their surroundings, a process that leads to quenching of the dye fluorescence. As shown in Figs. 2 and 3, when lysosomes are added to a solution of 300 mM sucrose, 50 mM Mops/Tris, pH 7.0, containing 0.5 p~ diS-C,-(5), 60-70% of the dye fluorescence is quenched in a time-dependent manner.
Experiments at more rapid chart speeds have shown that 15-25% of the total quenching occurs rapidly, within the experimental mixing time (2-3 s), while the remainder occurs more gradually. Lysosomes which have been frozen and thawed or exposed to the nonionic detergent Nonidet P-40 (&I%), procedures which destroy the integrity of the lysosomal membrane, show the rapid phase of fluorescence quenching but not the time-dependent phase. The rapid phase of quenching may represent the effects of the pigmented constituents of the lysosomes, whereas the time-dependent phase probably reflects the gradual accumulation of the dye within the lysosomal interior. This latter phase of fluorescence quenching could be due to either the presence of an internal negative membrane potential or the interaction of the dye with internal binding sites, or both.
Regardless of the precise mechanism of dye accumulation, manipulation of the lysosomal membrane potential with ionophores leads to changes in the fluorescence signal which are consonant with the expected changes in membrane potential. Thus, as shown in Fig. 2A (upper trace), the addition of the K ionophore valinomycin leads to an additional quenching of the dye fluorescence which is reversed upon the subsequent addition of KC1 to the external medium. As shown in Fig. 2 B, the magnitude of the fluorescence increase over that observed prior to valinomycin addition is a logarithmic function of the external K' concentration. In the absence of valinomycin ( Fig. 2A, lower truce) the effects of KC1 addition are markedly reduced in magnitude although some increase in fluorescence is still apparent. Somewhat larger effects of external K' in the absence of valinomycin are observed if the salt of an impermeable anion (e.g. K,SO,) is used (cf. Fig. 3C and Fig.   4C). The results suggest that the lysosomal membrane exhibits a low, but significant, permeability toward K'.
More importantly, the logarithmic K dependence of the fluorescence changes observed in the presence of valinomycin (Fig. 2 B ) indicate that changes in the dye fluorescence can be used to monitor changes in the lysosomal membrane potential. Extrapolation of the line shown in Fig. 2B to zero fluorescence change yields an intercept of 0.3 mM KCl. This value presumably represents the point a t which KJKo = exp(-A\k,/RT), where A*, is the value of the membrane potential prior to the addition of valinomycin, R is the gas constant, and T is the absolute temperature. Since neither K, nor A\ko is known with certainty, the fluorescence signals

Lysosomal Proton
Pump cannot be calibrated in terms of the magnitude of A*. Nevertheless, the signals clearly provide a semiquantitative indication of changes in the membrane potential. Relationship between A* and ApH-The average pH difference across the lysosomal membrane ( ApH), as determined by the distribution of methylamine, is 1.2 k 0.3 ( n = 3) pH units (acid inside) at an external pH of 7.0. What is the contribution of this proton gradient to the apparent lysosomal membrane potential? This issue can be approached in a qualitative manner by assessing the effects on diS-C3-(5) fluorescence of agents which collapse the pH gradient. As shown in Fig. 3A, when a weak base (5 mM ammonium sulfate) is added to the lysosomes in the presence of the dye, a rapid increase in fluorescence is observed. Similar results were obtained with 0.1 mM chloroquine and, as described later in connection with Fig. 4C, with the combination of nigericin and external K' . The results suggest that the lysosomal membrane exhibits a significant permeability toward protons and that the pH gradient contributes to the apparent membrane potential (negative inside) observed under these conditions.
This conclusion is reinforced by the effects of protonophores on the fluorescence signal. As shown in Fig. 3B, when the lysosomes are treated with 1 p~ FCCP, only a small degree of additional quenching of dye fluorescence is observed. Moreover, the magnitude of the effect of ammonium sulfate on dye fluorescence is similar to that seen in the absence of FCCP. The results suggest that protons are not far from equilibrium, even in the absence of FCCP, and that the membrane potential is approximately equal to the equilibrium potential for protons, i.e. A\k GZ EH = -RT In (H,/Ho). A different pattern of results is obtained, however, if the sucrose in the medium is replaced by K gluconate. In this case (Fig.   3C), the amount of fluorescence quenching is reduced in comparison to that observed in the sucrose medium (Fig. 3, A  and B ) , suggesting that the higher external K+ concentration has caused the membrane potential to become more positive inside. Furthermore, when FCCP is added to the lysosomes in the K gluconate medium, a dramatic decrease in fluorescence is observed. The FCCP-induced fluorescence decrease is only partially reversed by agents which collapse the pH gradient such as 5 mM ammonium sulfate or nigericin plus external K' (data not shown); indeed, the magnitude of the fluorescence increase produced by such agents under these 2. SUCROSE conditions is similar to that observed in the sucrose medium (cf. traces A and B in Fig. 3). The results suggest that in the K gluconate medium, protons are not in equilibrium with the lysosomal membrane potential and that increasing the membrane proton conductivity with FCCP shifts A@ to a more negative (inside) value, presumably one that approximates EH. The fact that the original level of fluorescence quenching is not restored by ammonium sulfate and K nigericin (agents that shift E H toward zero) suggests that A* > 0 in the K gluconate medium prior to the addition of the protonophores. The results are consistent with the idea, expressed previously (6, 7 , lo), that the lysosomal membrane shows only a limited permeability toward both protons and potassium ions.
Effect of ATP on Aq-As shown in Fig. 4 A ,  effects of the ionophores described above indicate that Mg-ATP induces a shift (positive inside) in the lysosomal membrane potential, presumably by activating the lysosomal proton pump.
The effect of Mg-ATP on A* is also evident from measurements of the distribution of permeable ions (Table 11). Three different ionic probes were used for these determinations: the lipophilic cation [3H]TPP' (20, 21), %Rb+ in the presence of valinomycin, and the permeable anion S[14C]CN-. As shown in Table 11, each probe gave different values for A* in the absence of ATP, ranging from -120 mV for TPP' to -9 mV  Table I1 provide strong confirmation of the conclusions reached in the fluorescence experiments, i.e. Mg-ATP causes the lysosomal membrane potential to become more positive inside.
Inhibitors-As shown in Fig. 5A, oligomycin (6.2 pg/ml) produced a 50% decrease in the magnitude of the fluorescence change induced by Mg-ATP. Other lysosome preparations frequently showed lesser degrees of inhibition, but in these instances, higher concentrations of oligomycin invariably brought on the inhibitory response. Oligomycin concentrations less than 2 pg/ml had no effect on the ATP-dependent fluorescence response. These concentrations of oligomycin are approximately 50-fold higher than those required to suppress completely the effects of Mg-ATP on diS-Ca-(5) fluorescence in kidney mitochondria (data not shown). Thus, it seems unlikely that the graded effects of oligomycin within such a high range of concentrations could be attributed to contamination of the lysosome preparations by submitochondrial particles.' This conclusion is supported by the data shown in Fig. 5B, which indicate that oligomycin also inhibited by 50% the increase in fluorescence produced by establishing an inwardly directed K gradient in the presence of valinomycin. Therefore, oligomycin either interferes with the response of the dye to a change in membrane potential or alters the permeability properties of the lysosomal membrane such that potentials generated by either Mg-ATP or K valinomycin are reduced in magnitude.
AS shown in Fig. 5, C and D, DCCD (50 p~) markedly inhibited the effects of Mg-ATP on diS-Cs-(5) fluorescence, and like oligomycin, it altered the response to K valinomycin. In this case, the fluorescence change caused by the addition of 10 mM KC1 (prior to Valinomycin) is dramatically increased * In intact mitochondria, Mg-ATP induces a negative shift in A*, and thus it produces an additional quenching of diS-C3-(5) fluorescence rather than an increase in fluorescence as observed with lysosomes. Submitochondrial particles, which have an inverted topology, would be expected to exhibit an increase in fluorescence (increase in A*) upon addition of Mg-ATP. over that in the control trace (Fig. 5D). This suggests that DCCD produced a increase in the permeability of the lysosomal membrane toward K+ and perhaps other ions as well.
Oligomycin produced similar effects although they were not as highly developed as in the case of DCCD (compare B and D of Fig. 5). Therefore, although both agents inhibited the ATP-dependent fluorescence increase, some or all of this inhibition probably reflects an alteration in membrane permeability rather than a direct effect on the lysosomal proton Pump.
NEM, on the other hand, nearly abolished the effect of ATP on diS-C3-(5) fluorescence (Fig. 5E) at concentrations (100 p~) that had no effect on the K valinomycin response (Fig. 5F); moreover, this concentration of NEM had no effect on the fluorescence response of mitochondria to Mg-ATP (data not shown). Another possible inhibitor, sodium vana-

Effect of Mg-ATP on A*
Membrane potentials (fS.E.) were calculated from the accumulation factor for each ion as described under "Material and Methods." The numbers in parentheses refer to the number of lysosome preparations assayed. For values designated +Mg-ATP, the medium included 1 mM MgSO, plus 1 mM Na2ATP (final concentration).  date (1 mM), had no effect on the fluorescence change induced by Mg-ATP (Fig. 5G).
Permeable Anions-The data in Fig. 7 show the effects of 20 mM concentrations of the potassium salts of various anions on the ATP-induced fluorescence changes. As shown, the effect of ATP declines in the following order: phosphate > C1-> NO, > SCN-. This order is the inverse of that for the anion permeability of the lysosomal membrane (22). The effect of anion permeability can also be seen in the response of dye fluorescence to the addition of the various salts prior to adding ATP (Fig. 7). Potassium phosphate produced a gradual increase in fluorescence, whereas C1-, NO; and SCNeach produced an initial decrease, followed by a gradual increase, in fluorescence. The magnitude of the initial decrease in fluorescence increased with increasing anion permeability (Cl-c NO: < SCN-), whereas the opposite relation was observed for the second phase of the fluorescence response.

DISCUSSION
Lysosomes from rat kidney were first isolated by Straus (23)(24)(25)(26) as large (0.1-5.0 PM) "droplets" which sedimented at relatively low centrifugal forces and were further characterized by Shibko and Tappel (27) and by Maunsbach (28,29). The procedure described here can be carried out rapidly without specialized equipment and yields lysosomes that are approximately $0-fold purified over the crude homogenate (Table I). Its success depends upon the fact that, unlike liver lysosomes, kidney lysosomes are quite dense and can readily be separated from mitochondria and other organelles by density gradient centrifugation. This feature of kidney lysosomes has been previously described by Maunsbach (28). The present procedure differs from that of Maunsbach in that a relatively low centrifugation speed was used to obtain the initial crude lysosomal preparation from the homogenate and that subsequent purification of the lysosomes was accomplished using isosmotic gradients of Percoll rather than sucrose gradients. The latter factor is particularly important because it avoids the osmotic stress that sucrose gradients of the required density would place upon the lysosomes. An added advantage of the present procedure is that densityaltering agents such as Triton WR-1339 have not been employed, thereby minimizing the risk of inducing nonphysiological changes in the permeability properties of the lysosomal membrane.
In a sucrose medium buffered at pH 7.0, the pH gradient across the membrane of the isolated kidney lysosomes is approximately 1.2 pH units, as determined from the distribution of ['4C]methylarnine. The relatively minor influence of protonophores on the fluorescence of diS-Ca-(5) and the apparent increase in AS/ caused by agents which collapse the lysosomal pH gradient (Figs. 3 and 4) suggest that protons are nearly in equilibrium under these conditions. This means that AS/ is approximately equal to the equilibrium potential for protons, given by EH = -60 ApH = -72 mV. This value is considerably more negative than that obtained from the distribution of S[14C]CN-and more positive than that obtained from [3H]TPP' distribution (Table 11). These discrepancies probably reflect the binding of SCN-and TPP' to sites within the lysosomes. On the other hand, EH is only slightly more positive than the value of AS/ obtained from the distribution of ffiRb' in the presence of valinomycin. Indeed, the difference between these two values may reflect the fact that under the conditions of these measurements, valinomycin causes a negative shift in AS/ because of the presence of K' within the lysosomes (Fig. 2A). As discussed in connection with Fig. 2 B  Although protons may be in equilibrium in a buffered sucrose medium, this does not appear to be the case in potassium gluconate. Under these conditions, AS/ reflects the high external K+ concentration and may well attain a positive value (see "Results"). In any event, AS/ is considerably more positive than EH, as shown by the large decrease in diS-Cs-(5) fluorescence when the lysosomal membrane is made permeable to protons by treatment with FCCP (Fig. 3C). This suggests that in a high K' medium such as the cell cytoplasm, the intralysosomal acidity represents a distinctly nonequilibrium distribution of protons. If this is the case, then a Donnan equilibrium can play little or no role in the development or maintenance of the lysosomal pH gradient (cf. Ref. 10).
When Mg-ATP is added to the lysosomes in the presence of diS-C3-(5), an increase in fluorescence is observed. This represents a positive shift in the membrane potential generated by the activity of the lysosomal proton pump. This conclusion is based on the following lines of evidence. (a) The ATP-dependent fluorescence shift is blocked by protonophores and by K valinomycin, agents which increase the conductivity of the lysosomal membrane and thereby short circuit the current generated by the proton pump (Fig. 4). A similar inhibitory effect is produced by the presence of permeable anions in the external medium, the degree of inhibition increasing with the permeability of the lysosomal membrane to the anion in question (Fig. 7). (b) Since the effect of ATP is not blocked by agents that collapse the lysosomal pH gradient (NH: and K nigericin) (Fig. 4), it cannot be attributed to a secondary effect of lysosomal acidification and must therefore represent the electrogenic operation of the pump itself. (c) The characteristics of the ATP-dependent fluorescence response are similar to the known characteristics of the lysosomal proton pump. Thus, the activity is inhibited by NEM, it is insensitive to vanadate, and it displays a nucleotide specificity similar to that reported for lysosomal acidification (1,6, 7 ) . ( d ) Finally, the activity cannot be attributed to other organelles, most notably mitochondria or submitochondrial particles' that might contaminate the lysosomal preparation. This is evident from the fact that the ATP-dependent fluorescence response represents a reversal of quenching and that the quenching itself is clearly of lysosomal origin. Thus, mitochondrial membranes, which represent only a minute fraction of the total protein present in the lysosomal preparations (Table I), could not possibly account for the magnitude of the fluorescence response to Mg-ATP. Furthermore, the fluorescence response shows little sensitivity to the potent mitochondrial ATPase inhibitor oligomycin; some inhibition is observed at high concentrations of oligomycin, but this probably reflects an effect of the agent on the lysosomal membrane permeability rather than a direct effect on the ATPase (Fig. 5B). Finally, NEM (0.1 mM) completey blocks the effect of Mg-ATP on diS-C3-(5) fluorescence of lysosomes (Fig. 5E) although it has no effect on the mitochondrial fluorescence response to ATP (data not shown).
The present findings are in agreement with those of Marin et al. (8) and Cretin (9) who observed that Mg-ATP produced a positive shift in membrane potential in plant vacuolysosomes (lutoids) obtained from the rubber tree.
Moreover, electrogenicity is also a feature of proton pumps in other intracellular organelles, including chromaffin granules, platelet granules, neurophypophyseal granules, insulin secretory granules, sperm acrosomes, and yeast vacuoles (reviewed in Refs. 30 and 31). Schneider (4, 11) has claimed that the proton pump of rat liver lysosomes is not electrogenic and has suggested that this is because phosphate ions are cotransported with protons during ATP-dependent acidification. However, as shown in Fig. 7A, the presence of phosphate in the external medium does not eliminate the ATP-dependent shift in A*; therefore, if phosphate accumulates within lysosomes during ATP-dependent acidification as indicated by Schneider's data (ll), it does so by some mechanism other than electroneutral proton-phosphate co-transport. Other evidence cited by Schneider (4, 11) for an electroneutral proton pump is that Mg-ATP failed to alter significantly the distribution of S[14C]CNin rat liver lysosomes. However, the data in Table I1 demonstrate that in rat kidney lysosomes, Mg-ATP produces a marked shift in S[14C]CN-distribution, consistent with a change in A* of more than 40 mV. As pointed out by Ohkuma et al. ( 7 ) , Schneider's results may have been affected by the presence of high concentrations of KC1 (100 mM) in his experiments; this concentration of a relatively permeable anion such as C1-would greatly reduce the magnitude of any membrane potential generated by the proton pump. This is evident from Fig. 7, which shows that permeable anions (including C1-at 20 mM) do indeed reduce the magnitude of the shift in A* produced by Mg-ATP.
The electrogenicity of the lysosomal proton pump may be important in promoting the osmotic stability of lysosomes (cf. Ref. 10). I n uitro, the presence of ATP exerts a stabilizing effect on lysosomes in the presence of electrolytes; this effect may be due in part to the positive shift in membrane potential induced by Mg-ATP, which would retard the diffusion of salt across the lysosomal membrane by reducing the rate of influx of the cationic species (cf. Ref. 32). I n situ, the osmotic stability of lysosomes may be critically dependent upon maintaining a relatively low internal K' concentration, and the positive potential generated by the proton pump may well provide the mechanism by which this is accomplished. Of course, such a potential woul also result in the accumulation of permeable anions within the lysosome, but this probably would not create a major osmotic stress because of the low concentration of such ions (e.g. C1-) within the cytoplasm of most mammalian cells. A decline in the magnitude of the lysosomal membrane potential may be involved in the lysosomal alterations which occur during the development of pathological states, such as ischemia, that are accompanied by a fall in cellular ATP levels. For example, swelling of intracellular lysosomes is observed within 15 min of the onset of ischemia in cardiac tissue; after 30 min, there is a progressive loss of lysosomal integrity resulting in the release of lysosomal enzymes into the cell cytoplasm (reviewed in Refs. 33 and 34). Although many factors are undoubtedly involved in this complex process, it seems likely that the gradual influx of cytoplasmic K' into the lysosomes, in response to a decline in the lysosomal membrane potential, may play a role in the swelling and eventual disruption of these organelles.