An Acid-loading Chloride Transport Pathway in the Intraerythrocytic Malaria Parasite, Plasmodium falciparum*

The intraerythrocytic malaria parasite exerts tight control over its ionic composition. In this study, a combination of fluorescent ion indicators and 36Cl− flux measurements was used to investigate the transport of Cl− and the Cl−-dependent transport of “H+-equivalents” in mature (trophozoite stage) parasites, isolated from their host erythrocytes. Removal of extracellular Cl−, resulting in an outward [Cl−] gradient, gave rise to a cytosolic alkalinization (i.e. a net efflux of H+-equivalents). This was reversed on restoration of extracellular Cl−. The flux of H+-equivalents was inhibited by 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid and, when measured in ATP-depleted parasites, showed a pronounced dependence on the pH of the parasite cytosol; the flux was low at cytosolic pH values < 7.2 but increased steeply with cytosolic pH at values > 7.2. 36Cl− influx measurements revealed the presence of a Cl− uptake mechanism with characteristics similar to those of the Cl−-dependent H+-equivalent flux. The intracellular concentration of Cl− in the parasite was estimated to be ∼48 mm in situ. The data are consistent with the intraerythrocytic parasite having in its plasma membrane a 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid-sensitive transporter that, under physiological conditions, imports Cl− together with H+-equivalents, resulting in an intracellular Cl− concentration well above that which would occur if Cl− ions were distributed passively in accordance with the parasite's large, inwardly negative membrane potential.

Malaria is an infectious disease caused by a unicellular eukaryote (genus Plasmodium) that, in the course of its complex lifecycle, invades the erythrocytes of its host. The intraerythrocytic parasite has, in its plasma membrane, a V-type H ϩ -ATPase, which extrudes H ϩ , thereby playing a key role in the regulation of cytoplasmic pH (pH i ) 2 (1)(2)(3), as well as being the source of the parasite's large, inwardly negative, membrane potential (4). H ϩ ions are also extruded from the parasite via a H ϩ :monocarboxylate symporter, which provides a route for the efflux of lactic acid originating from glycolysis (5)(6)(7).
In many cell types, acid extrusion mechanisms such as these act in concert with "acid-loading" mechanisms that mediate the uptake of "H ϩ -equivalents," usually via the export of OH Ϫ or HCO 3 Ϫ , often in exchange for Cl Ϫ ions. Acid-loading and acidextruding mechanisms are typically regulated in a coordinated fashion, together ensuring that pH i is maintained within a narrow range (8). There is little known about acid-loading mechanisms in the malaria parasite. Nor have there been any studies of the transport of Cl Ϫ across the parasite plasma membrane. An x-ray microanalysis study of the elemental composition of the cytosol of mature, intraerythrocytic Plasmodium falciparum trophozoites estimated a [K ϩ ]:[Na ϩ ]:[Cl Ϫ ] ratio of 1:0.13: 0.28 (9). If it is assumed that, as in other cells, the combined concentration of K ϩ and Na ϩ is of the order of 150 mM, this gives an estimated cytosolic [Cl Ϫ ] ([Cl Ϫ ] i ) of ϳ40 mM.
A normal, uninfected human erythrocyte has a [Cl Ϫ ] i of 73-95 mM (10). This may be reduced slightly (by an estimated 12% (9)) in erythrocytes infected with mature, trophozoitestage parasites. The parasitophorous vacuole membrane in which the intracellular parasite is enclosed is thought to be freely permeable to Cl Ϫ and other monovalent ions (11). The [Cl Ϫ ] at the extracellular surface of the parasite is therefore likely to be in the range 65-85 mM. If Cl Ϫ were to distribute passively across the parasite plasma membrane, in accordance with the membrane potential (estimated as Ϫ95 mV under physiological conditions (4)), then [Cl Ϫ ] i is predicted (by the Nernst equation) to be below 3 mM. The fact, that the x-ray microanalysis data indicate that the [Cl Ϫ ] i is more than an order of magnitude higher than this, is consistent with Cl Ϫ being taken up, into the parasite, across its plasma membrane, via a process that maintains [Cl Ϫ ] i away from electrochemical equilibrium. However, the presence and properties of any such system remain to be demonstrated.
Herein we report the functional characteristics of a Cl Ϫ -dependent acid-loading mechanism in the plasma membrane of the intraerythrocytic malaria parasite. The physiological characteristics of the system match those of the pathway or pathways mediating Cl Ϫ transport into the parasite. The data are consistent with this system serving as a route for the uptake of Cl Ϫ into the intracellular parasite, coupling the influx of Cl Ϫ to the inward H ϩ electrochemical gradient.
Parasite Culture-P. falciparum parasites (3D7 strain, except where specified otherwise) were maintained in continuous suspension (12), in synchronous cultures in Group O, Rhϩ erythrocytes as described previously (4). All experiments were conducted on mature trophozoite-stage parasites (36 -40 h post invasion). In the majority of experiments, trophozoites were "isolated" from their host erythrocytes by treatment of parasitized erythrocytes with the plant detergent saponin, as described elsewhere (13). Saponin renders the erythrocyte plasma membrane and parasitophorous vacuole membrane permeable to macromolecules (14) but leaves the parasite plasma membrane intact and able to generate and maintain transmembrane ion gradients (2,3,15) and a large transmembrane potential (4).
Solutions-Unless specified otherwise, cells were suspended in a HEPES-and MES-buffered saline containing 125 mM NaCl, 5 mM KCl, 20 mM glucose, 20 mM HEPES, 20 mM MES, and 1 mM MgCl 2 , at pH 7.4. For "Cl Ϫ -free saline" NaCl and KCl were replaced isosmotically either with the equivalent gluconate salts (125 mM sodium gluconate, 5 mM potassium gluconate), sulfate salts (100 mM Na 2 SO 4 , 4 mM K 2 SO 4 ) or sucrose (260 mM); in each case MgCl 2 was replaced with MgSO 4 . pH i Measurements Using BCECF: Confocal Laser Scanning Microscopy-The cytosolic pH of BCECF-loaded parasites within intact P. falciparum-infected erythrocytes was monitored using confocal laser scanning microscopy. Suspensions of parasitized erythrocyte cultures (1% hematocrit, Ͼ10% parasitemia) were loaded with BCECF by incubation with 5 M of its acetoxymethyl ester for 10 min at 37°C. The cells were washed by centrifugation (1800 ϫ g, 5 min), resuspended in the HEPES-and MES-buffered saline, then immobilized on polylysine-coated coverslips in a Bioptechs FCS2 perfusion chamber and maintained in the HEPES-and MESbuffered saline at 22°C.
The fluorescence signals from parasitized erythrocytes were collected on a Zeiss Pascal confocal laser scanning microscope through a Plan-Apochromat 63 ϫ 1.2 numerical aperture water objective. Excitation of BCECF was performed using an argon ion laser at 488 nm. Emitted light was collected through a 560 nm long pass filter from a 543 nm dichroic mirror. Photobleaching (the irreversible damage of BCECF, producing a less fluorescent species) was assessed by continuous exposure (5 min) of loaded cells to laser illumination. For each experiment, the laser illumination and microscope settings that gave no reduction in signal were used. Data capture and extraction were carried out with Zeiss Pascal software and Photoshop. Measurements of pH i were calibrated in situ using the nigericin/ high-K ϩ method (16) as described previously for malaria parasites (2,17,18). pH i Measurements Using BCECF: Spectrofluorometric Measurement-The pH i of parasites isolated from their host cells by saponin permeabilization of the erythrocyte and parasitophorous vacuole membranes was measured at 22°C using BCECF in a PerkinElmer Life Sciences LS-50B spectrofluorometer as described previously (2). The parasites were suspended at a cell density of 1-3 ϫ 10 7 cells/ml. Measurements of pH i were calibrated using the nigericin/high-K ϩ method (2).
For experiments in which the anion transport inhibitor 4,4Јdiisothiocyanostilbene-2,2Ј-disulfonic acid (DIDS) was added to isolated parasites, pH i was calibrated in both the presence and absence of the inhibitor. In initial experiments it was found that the effect of DIDS varied with cell density; subsequent experiments with this reagent were therefore conducted using cell suspensions with a density close to 2 ϫ 10 7 cells/ml.
In a number of experiments parasites were suspended in a glucose-free medium to deplete them of ATP, thereby halting metabolism and inactivating the plasma membrane H ϩ -ATPase (2). Glucose was replaced with 10 mM Na ϩ -salt to maintain the osmolality, and the parasites were incubated under these conditions for at least 15 min at 37°C prior to beginning the experiment.
In one series of experiments, the effect on pH i of removal of extracellular Cl Ϫ was investigated under conditions in which the initial pH i was set to a range of different values. BCECFloaded parasites, de-energized by preincubation in glucose-free medium, were equilibrated for 15 min in a weakly buffered saline (155 mM NaCl/5 mM KCl/2 mM HEPES/2 mM Tris/2 mM MES/1 mM MgCl 2 ) in which the pH ranged from 6.5 to 8.0. Under these conditions pH i approached pH o . The cells were then diluted 1:100 into a well buffered Cl Ϫ -free saline (125 mM sodium gluconate, 5 mM potassium gluconate, 20 mM glucose, 20 mM HEPES, 20 mM MES, and 1 mM MgSO 4 ) at pH 7.1, giving a final extracellular Cl Ϫ concentration of 1.62 mM.
The ability of the parasite to respond to either an intracellular alkali load or an intracellular acid load was investigated by the addition of either 40 mM NH 4 ϩ (added as 40 mM NH 4 Cl or 20 mM (NH 4 ) 2 SO 4 (2)) or 40 mM sodium lactate (7), respectively.
The buffering capacity of the parasite cytosol at 22°C was determined using parasites that were preincubated in the absence of Cl Ϫ and glucose. The cells were preincubated for at least 15 min at a range of pH o values (as above), then 2.5 mM (NH 4 ) 2 SO 4 was added. Intracellular buffering power (␤ i ) was calculated by determining the difference between the observed change in pH induced by (NH 4 ) 2 SO 4 and that predicted to occur in the absence of intracellular buffering (19). ␤ i was plotted as a function of pH i , and the data were fitted by the equation, ␤ i ϭ 19 ϩ 56 (pH i Ϫ7.4) 2 mM/pH unit, which was subsequently used for the calculation of H ϩ -equivalent transport rates from the measured rate of change of pH i .
Monitoring the pH of the Parasite-digestive Vacuole-In one series of experiments the pH of the digestive vacuole of the parasite was monitored qualitatively (i.e. without calibration) using suspensions of saponin-isolated parasites in which the digestive vacuole had been loaded with the membrane-impermeant fluorescent pH indicator, fluorescein-dextran, as described previously (20).
Measurements of 36 Cl Ϫ Uptake-36 Cl Ϫ uptake into isolated parasites was measured at 22°C using techniques described previously for other solutes (7,13). Briefly, saponin-isolated trophozoite-stage parasites were suspended in a HEPES-buffered saline (125 mM NaCl/5 mM KCl/20 mM glucose/25 mM HEPES/1 mM MgCl 2 , pH 7.1). Aliquots (200 l) of this cell suspension (typically 10 ϫ 10 7 cells/ml) were mixed gently with an equal volume of HEPES-buffered saline containing ϳ1 Ci/ml 36 Cl Ϫ , over a 300-l oil layer (5 parts dibutyl phthalate:4 parts dioctyl phthalate) in a 1.5-ml microcentrifuge tube. After an appropriate incubation period the tubes were centrifuged (15,800 ϫ g, 2 min) to sediment the parasites through the oil layer, thereby terminating the uptake of 36 Cl Ϫ . The cell pellets were processed for scintillation counting as described elsewhere (13).
In uptake was calculated from the total radioactivity present in the cell pellet by subtracting the amount of radioactivity trapped in the extracellular space between the pelleted cells. The latter was estimated from the amount of radioactivity associated with pellets centrifuged from suspension immediately after (i.e. within 1-2 s of) combining the isolated parasites and 36 Cl Ϫ .
[ 36 Cl Ϫ ] i was calculated from the 36 Cl Ϫ uptake and the total intracellular water space of the cell pellet. The latter was determined from the parasite number and the water space of a single saponin-isolated parasite, estimated previously to be 28 fl (13).
[Cl Ϫ ] i Measurements Using MQAE-Isolated trophozoites were loaded with the fluorescent Cl Ϫ indicator MQAE by incubation with a 5 mM concentration of the dye for 45 min at 37°C in RPMI. The cells were then washed three times at 22°C and suspended in HEPES-buffered saline, pH 7.1. The cell suspension was transferred to a cuvette, and Cl Ϫ activity was measured at 22°C in a PerkinElmer Life Sciences LS-50B spectrofluorometer (excitation 350 nm, emission 460 nm). Calibration was performed using a double ionophore technique described previously (21). Cells were suspended in standard [Cl Ϫ ] solutions containing 5 M nigericin and 10 M tributyltin chloride. The Cl Ϫ standards were prepared by mixing, in varying ratios, two K ϩ -rich solutions, one containing Cl Ϫ (130 mM KCl/20 mM glucose/25 mM HEPES/1 mM MgCl 2 , pH 7.1) and the other Cl Ϫ -free (130 mM potassium gluconate/20 mM glucose/25 mM HEPES/1 mM MgSO 4 , pH 7.1).
MQAE is able to permeate cell membranes, and a slight reduction in fluorescence intensity of MQAE-loaded trophozoites was observed over time as the dye leaked slowly from the cells. The rate of the fluorescence change associated with this dye loss remained approximately constant throughout each experiment. In experiments in which MQAE fluorescence was monitored over a sustained period of time, specifically those experiments giving rise to Fig. 13B, the reduction in fluorescence intensity associated with dye leakage was subtracted from the data prior to the final conversion of MQAE fluorescence to [Cl Ϫ ] i .

RESULTS
The Effect on pH i of the Removal of Extracellular Cl Ϫ -Initial experiments investigating the effect of removal of extracellular Cl Ϫ (Cl Ϫ o ) on pH i of mature parasites within intact erythrocytes were carried out using confocal laser scanning microscopy. Fig. 1A shows the distribution of BCECF fluorescence in an infected erythrocyte. Fluorescence was distributed uniformly in the parasite cytosol but was largely absent from the parasite's internal digestive vacuole. For infected erythrocytes bathed in a HEPES-and MES-buffered, Cl Ϫ -containing medium (pH 7.4) the resting pH i of trophozoites was estimated as 7.44 Ϯ 0.01 (mean Ϯ S.E., n ϭ 4). Upon replacement of Cl Ϫ o with gluconate, there was a rapid alkalinization of the parasite cytosol, with pH i reaching a value of 7.81 Ϯ 0.12 (mean Ϯ S.E., n ϭ 4) within 2 min (Fig. 1B). Restoration of Cl Ϫ (132 mM) to the extracellular medium resulted in an acidification, with pH i returning to close to its original value within 2 min.
In erythrocytes, replacement of extracellular Cl Ϫ with an impermeant anion (such as gluconate) results in a rapid alka- linization of the erythrocyte cytosol (22,23); the alkalinization seen in the intraerythrocytic parasite is therefore likely to be due, at least in part, to an increase in the pH of the host cell cytosol. To study the regulation of pH i in the parasite, independently of the host cell, all further pH measurements were made in parasites effectively isolated from their host cells by saponin permeabilization of the erythrocyte and parasitophorous vacuole membranes. Fig. 2A shows the effect on pH i of the removal of Cl Ϫ o from a suspension of isolated trophozoites at an extracellular pH (pH o ) of 7.1. The initial pH i was 7.32 Ϯ 0.02 (mean Ϯ S.E., n ϭ 8). Upon replacement of extracellular Cl Ϫ with gluconate, pH i increased to 7.91 Ϯ 0.03 (mean Ϯ S.E., n ϭ 8) within 2 min. Upon restoration of Cl Ϫ (40 mM) to the extracellular medium the pH i returned to close to its original value within 2 min. The initial rate of this re-acidification corresponded to a H ϩ -equivalent flux of 63 Ϯ 15 mmol H ϩ /(L cell H 2 O⅐min) (mean Ϯ S.E., n ϭ 3). A similar alkalinization of the parasite cytosol was observed when the extracellular Cl Ϫ was replaced with SO 4 2Ϫ (Fig. 2B) or when the NaCl and KCl in the extracellular medium was replaced isosmotically with sucrose ( Fig. 2C).
In some cell types, removal of extracellular Cl Ϫ results in cell shrinkage (e.g. Refs. 23,24). To assess whether the increase in pH i that follows the removal of extracellular Cl Ϫ might be due to a reduction in parasite volume two series of experiments were carried out. In the first, suspensions of isolated parasites (in Cl Ϫ -containing media) were made hypertonic by the addition of increasing concentrations of either NaCl, sodium gluconate, Na 2 SO 4 , or sucrose, with the aim of inducing osmotic shrinkage of the parasites. As can be seen from Fig.  3A increasing the concentration of either NaCl or sodium gluconate by up to 150 mM (thereby increasing the extracellular osmolarity by up to 290 mosM, approximately double the original osmolarity, in each case) had no significant effect on pH i . Addition to the medium of up to 300 mM sucrose (thereby increasing the extracellular osmolarity by up to 315 mosM) caused a slight decrease in pH i , whereas addition of up to 150 mM Na 2 SO 4 (thereby increasing the extracellular osmolarity by up to 347 mosM) caused a slight increase in pH i . Together, these data indicate that cell shrinkage per se has little if any effect on pH i in the parasite.
In the second series of experiments, the alkalinization seen on  . Effect of osmolarity on the pH i of saponin-isolated trophozoites and the pH i response of parasites transferred to a Cl ؊ -free medium. A, a suspension of BCECF-loaded parasites (in Cl Ϫ -containing medium) was made hypertonic by the sequential addition of aliquots of concentrated (1 M) NaCl, sodium gluconate, sucrose, or Na 2 SO 4 (in water), with the aim of inducing osmotic shrinkage of the parasites. In each "series" the left-hand trace is the one obtained before the first addition was made. The approximate osmolarity of the suspension following each addition is shown above the traces. replacement of extracellular Cl Ϫ isosmotically with gluconate (as in Fig. 2) was compared with that seen when the Cl Ϫ solution was replaced with a hypotonic gluconate solution (containing 100 mM gluconate rather than the usual 125 mM gluconate). Reduction of the osmolarity of the medium by ϳ50 mosM will result in cell swelling, thereby countering, in part or in whole, any cell shrinkage that might have resulted from the removal of extracellular Cl Ϫ . As can be seen from Fig. 3B, the trace observed in the hypotonic (ϳ273 mosM) gluconate medium was very similar to that observed in the higher osmolarity (322 mosM) medium. These data are again consistent with cell volume changes playing little if any role in the pronounced changes in pH i seen upon removal and re-addition of extracellular Cl Ϫ .
The 3D7 parasites used throughout this study are chloroquine-sensitive. In the course of this work one other chloroquine-sensitive strain (D10) and two chloroquine-resistant strains (7G8 and K1) of P. falciparum were tested for the effects on pH i of the removal then re-addition of extracellular Cl Ϫ (as in Fig. 2). For all four strains, removal of extracellular Cl Ϫ resulted in a pronounced alkalinization; the subsequent addition of 40 mM NaCl resulted in pH i returning to its original resting value (as in Fig. 2; data not shown). The phenomenon is therefore present in a range of parasite strains and is unrelated to the chloroquine resistance status of the parasites.
In another series of experiments the digestive vacuoles of 7G8 and K1 parasites were loaded with a membrane-impermeant fluorescent pH indicator (fluorescein-dextran) and the pH of the digestive vacuoles thereby monitored as described previously (20). For saponin-isolated parasites the alkalinization of the cytosol seen upon removal of extracellular Cl Ϫ was accompanied by an alkalinization of the digestive vacuole (not shown). This phenomenon was not studied in any detail; however, it does rule out the possibility that the cytosolic alkalinization is a consequence of the uptake of H ϩ by the digestive vacuole as any such uptake would have generated an acidification, rather than an alkalinization, of the digestive vacuole.
The Effect of ATP Depletion-To investigate whether the alkalinization of the parasite cytosol following the removal of Cl Ϫ from the extracellular medium was influenced by the energy status of the cell, the experiment was repeated with parasites preincubated and suspended in glucose-free medium. In the absence of glucose the parasite is rapidly depleted of ATP and is therefore unable to support H ϩ -ATPase activity (1,2). Under these conditions pH i decreased to a value close to pH o (pH 7.1 (Fig. 4A)). When Cl Ϫ o was replaced with gluconate under glucose-free conditions, there was, again, an alkalinization (Fig. 4A). The alkalinization was substantially slower than that seen for cells in glucose-replete medium (Fig. 4B), with pH i taking 20 min to reach a maximum. Furthermore, the trace for the ATP-depleted parasites appeared sigmoidal, with the rate of alkalinization initially increasing as pH i increased, then slowing as the pH i plateaued at a value (7.91 Ϯ 0.01; mean Ϯ S.E., n ϭ 8) similar to the value observed in ATPreplete cells (p ϭ 0.97, Student's t test). Upon addition of 40 mM NaCl to the (alkalinized) cells in the Cl Ϫ -free medium a re-acidification was observed (Fig. 4A) with an initial rate equivalent to a H ϩ -equivalent flux of 48 Ϯ 3 mmol H ϩ /(L cell H 2 O⅐min) (mean Ϯ S.E., n ϭ 3). This value was slightly, but not significantly, lower than the re-acidification rate measured in glucose-replete cells (p ϭ 0.423, Student's t test).
In one series of experiments ATP-depleted parasites, which had alkalinized upon replacement of extracellular Cl Ϫ with gluconate and then re-acidified in response to the addition of 40 mM NaCl (as in Fig. 4A), were subjected to a second round of Cl Ϫ replacement. Removal of Cl Ϫ from cells that had alkalinized then re-acidified previously caused the cells to undergo a second alkalinization, of similar magnitude to the first. The second alkalinization occurred more rapidly than the first (the time taken was reduced by a factor of up to ϳ3; data not shown) but was still much slower than the very rapid alkalinization observed upon removal of Cl Ϫ from ATP-replete cells (Fig. 4B). The increased rate of alkalinization observed in ATP-depleted parasites subjected to two rounds of Cl Ϫ removal was not investigated further.
The Effect of Transport Inhibitors-The parasite's plasma membrane V-type H ϩ -ATPase should be inactive in the ATPdepleted parasites (2). Nevertheless, to rule out a role for the H ϩ -pump in the alkalinization observed upon removal of Cl Ϫ o , ATP-depleted cells were pre-treated with the V-type H ϩ -ATPase inhibitor concanamycin A (75 nM) prior to replacement of Cl Ϫ o . The inhibitor had no observable effect on the rate of alkalinization following removal of Cl Ϫ o , nor on the rate of re-acidification following restoration of Cl Ϫ o (data not shown).
The Cl Ϫ transport inhibitor DIDS was tested for its effect on the alkalinization induced by removal of Cl Ϫ o . Isolated BCECFloaded trophozoites were incubated with 500 M DIDS for 5 min prior to removal of Cl Ϫ o . Fig. 5 shows representative pH i traces obtained with ATP-replete cells (Fig. 5A) and ATP-depleted cells (Fig. 5B). In ATP-replete (but not ATP-depleted) cells addition of 500 M DIDS caused a slight but significant increase in resting pH i from 7.36 Ϯ 0.02 to 7.43 Ϯ 0.01 over 5 min (mean Ϯ S.E., n ϭ 5; p ϭ 0.048, Student's t test). In both ATP-replete and ATP-depleted cells, 500 M DIDS abolished completely the alkalinization seen upon removal of Cl Ϫ o . In control cells treated with an equivalent volume of the solvent DMSO, the alkalinization observed on Cl Ϫ removal proceeded as normal.
In a similar experiment, ATP-replete and ATP-depleted trophozoites were alkalinized by removal of Cl Ϫ o , and then treated with 500 M DIDS for 5 min before 40 mM NaCl was added back to the cells. Representative pH i traces are shown in Fig. 5C for ATP-replete cells and Fig. 5D for ATP-depleted cells. In the DIDS-treated cells, recovery of pH i on re-addition of Cl Ϫ o was abolished, whereas in control (solvent-treated) cells, pH i reacidified as normal. The finding, that both the alkalinization observed upon removal of Cl Ϫ o , and the recovery of pH i observed on re-addition of Cl Ϫ o , were both inhibited by DIDS, is consistent with the same pathway underlying the two processes.
The Relationship between [Cl Ϫ ] o and pH i -The relationship between extracellular chloride concentration ([Cl Ϫ ] o ) and parasite pH i was investigated in more detail by measuring pH i in isolated BCECF-loaded trophozoites that had been incubated in solutions of varying [Cl Ϫ ] (0 -100 mM). The pH i of these cells was measured only once it had reached a maximum value (pH i,max ). This took ϳ5 min for ATP-replete cells and ϳ20 min for ATP-depleted cells (cf. Fig. 4, A and B). The data indicate that, although the ATP status of the parasite influences both the resting pH i and the rate at which pH i increases upon removal of extracellular Cl Ϫ (Fig. 5), the relationship between pH i and [Cl Ϫ ] o is similar in both ATP-depleted and ATP-replete cells. To avoid potential complications arising from the operation of the parasite's H ϩ pump(s) and/or the generation or utilization of H ϩ -equivalents by metabolism, all subsequent pH i measurements, aside from those represented below in Fig. 13, were carried out using ATP-depleted parasites.
pH i Dependence of the Rate of Alkalinization following Removal of Cl Ϫ o -The dependence on pH i of the rate of alkalinization following removal of extracellular Cl Ϫ was investigated in more detail using a protocol in which pH i was set to a range of different starting values prior to the removal of Cl Ϫ . The starting pH i was manipulated by pre-equilibrating BCECF-loaded, ATP-depleted parasites in weakly buffered (Cl Ϫ -containing) saline in which the pH ranged from ϳ6.9 to 7.7. Under these conditions pH i approached pH o (2). At "time zero" the cells were diluted 1:100 into a strongly buffered Cl Ϫ -free saline at pH 7.1, and pH i was monitored. Fig. 7A shows the resulting pH i traces.  Curves were fitted to the pH i traces, allowing an estimate of the rate of alkalinization immediately following the dilution of the cells in Cl Ϫ -free saline. This "initial slope" was then multiplied by ␤ i to give the initial H ϩ -equivalent flux, which was plotted as a function of the initial pH i (Fig. 7B). The H ϩ -equivalent efflux displayed a sigmoidal dependence on pH i . At starting pH i values Ͻ 7.2 the rate of alkalinization observed on depletion of extracellular Cl Ϫ was low and approximately constant; at pH i values Ͼ 7.2 the rate of alkalinization increased with increasing pH i .
The Kinetics of the Cl Ϫ -dependent Acidification-To investigate the kinetics of the acid-loading mechanism responsible for the decrease in pH i seen on addition of Cl Ϫ to parasites suspended in an initially Cl Ϫ -free medium, isolated ATP-depleted parasites were first suspended in a Cl Ϫ -free (gluconate-containing, glucose-free) medium, then NaCl was added to the extracellular solution at concentrations ranging from 2.5 to 200 mM. Upon addition of NaCl to the extracellular solution the parasite cytosol underwent an acidification, the rate and magnitude of which increased with [NaCl]. The corresponding pH i traces are shown in Fig. 8A.
In a control experiment equivalent concentrations of Na ϩ gluconate (2.5-200 mM) were added to parasites suspended in the Cl Ϫ -free solution. Concentrations up to 30 mM had no significant effect on pH i (data not shown). The addition of Na ϩ gluconate at concentrations Ͼ30 mM caused a slight decrease in pH i (data not shown); this change was subtracted from the recorded pH i changes induced by [NaCl] Ͼ 30 mM.
The time course for the acidification induced by each NaCl concentration (Fig. 8B) was fitted to a first order exponential equation, and the initial rate of acidification was thereby estimated. This was then multiplied by ␤ i to give the initial H ϩequivalent flux.
H ϩ -equivalent flux showed a non-linear dependence on [Cl Ϫ ] o (Fig. 8C). The data were fitted to the Michaelis-Menten equation, yielding an apparent K m of 20 Ϯ 4 mM Cl Ϫ , and a V max of 52 Ϯ 7 mmol H ϩ /(L cell H 2 O⅐min) (mean Ϯ S.E., n ϭ 5).
The Anion-selectivity Profile of the Acidification Mechanism-To investigate the anion selectivity of the acid-loading mechanism, various different anions, each at a concentration of 10 mM, were added to ATP-depleted cells that had been alkalinized by suspension in a Cl Ϫ -free (gluconate-containing) medium. Those anions able to substitute for Cl Ϫ induced a re-acidification of the parasite cytosol. The initial H ϩ -equivalent flux induced by addition of the anions was estimated as above, and the results are shown in Fig. 9. NO 3 Ϫ , Br Ϫ , and I Ϫ all induced acidification, with initial rates in the order NO 3 Ϫ Ͼ Br Ϫ Ϸ Cl Ϫ Ͼ I Ϫ . The addition of 10 mM phosphate, SO 4 2Ϫ , or gluconate, had no significant effect on pH i of cells in a Cl Ϫ -free medium; i.e. the cells remained alkaline.
In a control experiment NO 3 Ϫ , Br Ϫ , Cl Ϫ , or I Ϫ (each at 10 mM) were added to ATP-depleted parasites suspended in medium containing a normal Cl Ϫ concentration (132 mM) and found to have no significant effect on pH i (data not shown). The traces shown are from a single representative experiment. B, the initial slope of each curve in A was calculated and multiplied by ␤ i to give the initial rate of H ϩ -equivalent flux induced by Cl Ϫ replacement. These values were then plotted as a function of the initial pH i . The three different symbols correspond to data from three independent experiments. The line was drawn using a fiveparameter sigmoidal curve fitted to the data (R 2 ϭ 0.94).
Cl Ϫ Transport across the Parasite Plasma Membrane-The results presented so far provide evidence for an anion-dependent flux of H ϩ -equivalents across the parasite plasma membrane. The transport characteristics of Cl Ϫ itself were investigated by measuring the uptake of 36 Cl Ϫ into saponin-isolated parasites.
The resting [Cl Ϫ ] i of isolated ATP-replete trophozoites was also estimated using the Cl Ϫ -sensitive fluorescent indicator MQAE. The relationship between [Cl Ϫ ] and fluorescence intensity of MQAE is described by the Stern-Volmer  Fig. 11A shows traces from a representative calibration experiment and Fig.  11B shows the Stern-Volmer calibration curve, from which K sv was estimated to be 35 Ϯ 1 mM Ϫ1 (mean Ϯ S.E., n ϭ 5). The MQAE fluorescence in isolated parasites suspended in a HEPES-buffered saline at pH 7.1 yielded a [Cl Ϫ ] i of 70 Ϯ 6 mM (mean Ϯ S.E., n ϭ 5); i.e. a value very similar to that estimated on the basis of 36 Cl Ϫ distribution.
For isolated parasites suspended in the HEPES-buffered saline the extracellular [Cl Ϫ ] was 132 mM, significantly higher than that in the infected erythrocyte cytosol (estimated to be in the order of 65-85 mM (9, 10)) and therefore higher than that in the external environment to which the intracellular parasite is exposed in vivo. We therefore estimated parasite [Cl Ϫ ] i over a range of [Cl Ϫ ] o that encompassed the estimated range of [Cl Ϫ ] in the erythrocyte cytosol. Fig. 11C  ] to which the intraerythrocytic parasite is exposed, indicated by the gray shading in Fig. 11C) [Cl Ϫ ] i was estimated to be in the range of 45-58 mM.
The pH Dependence of 36 Cl Ϫ Influx-As shown in Fig. 12A, reducing pH o from 7.1 to 6.1 had little effect on the influx of 36 Cl Ϫ into isolated, ATP-replete, P. falciparum trophozoites. By contrast, 36 Cl Ϫ influx showed a marked dependence on pH i . As shown in Fig. 12B, alkalinization of the parasite cytosol by the addition of 40 mM NH 4 Cl (at the time of addition of 36 Cl Ϫ (Fig. 12B, inset)), resulted in a marked increase in 36 Cl Ϫ influx (p ϭ 0.019, Student's t test). The initial rate of 36 Cl Ϫ influx in these cells was 242 Ϯ 39 mmol Cl Ϫ /(L cell H 2 O⅐min), compared   ). B, DIDS inhibition of 36 Cl Ϫ uptake into ATP-replete saponin-isolated trophozoites. 36 Cl Ϫ uptake was measured into isolated trophozoite suspensions that were treated with 500 M DIDS (for 5 min prior to, and throughout the time course). The rate of 36 Cl Ϫ influx was calculated from the initial slope of the time course and is shown as a % of the rate of influx in control (solvent-treated) cells. The data are averaged from four independent experiments (error bar denotes ϮS.E.).
The Role of Cl Ϫ in the Recovery from an NH 4 ϩ -induced Alkalinization-The data of Figs. 1-9 show effects on pH i of manipulating the extracellular [Cl Ϫ ]. The data are consistent with the presence of a pathway that, under conditions in which cells have been alkalinized by removal of Cl Ϫ from the medium then re-exposed to extracellular Cl Ϫ , mediates the import of H ϩ -equivalents. The question remains as to whether this pathway might play a role in the recovery of the parasite from an intracellular alkalinization under conditions of normal extra- To test this, ATP-replete parasites in a Cl Ϫ -containing medium were alkalinized by the addition to the medium of (NH 4 ) 2 SO 4 (20 mM). Fig. 13A shows the response of parasite pH i following the addition; there was a rapid alkalinization, followed by a slower recovery, which, over a period of several minutes, brought pH i back down to close to its resting value. Fig. 13B shows the corresponding changes in [Cl Ϫ ] i (measured with MQAE) under the same conditions. Upon addition of (NH 4 ) 2 SO 4 there was an increase in [Cl Ϫ ] i that rose (on a timescale of several minutes) to ϳ15 mM above the initial resting value. The increase in [Cl Ϫ ] i coincided with the recovery of pH i following the initial alkalinization, as well as with the increase in 36 Cl Ϫ influx, relative to control cells, seen in parasites alkalinized with NH 4 Cl (Fig. 12B). The increase in [Cl Ϫ ] i may be attributed to an increased influx of Cl Ϫ , activated in response to the alkali load.
The initial rate of recovery of pH i following the alkalinization induced by the addition of (NH 4 ) 2 SO 4 to isolated parasites in Cl Ϫ -containing medium was inhibited to 17 Ϯ 3% (mean Ϯ S.E., n ϭ 4) of its control rate by DIDS (500 M (Fig. 13C)).
To investigate the Cl Ϫ dependence of pH i recovery following intracellular alkalinization, the experiment was repeated in cells suspended in the absence of extracellular Cl Ϫ for 30 min prior to the addition of (NH 4 ) 2 SO 4 . The results are shown in Fig. 13D. The initial pH i of the cells in the Cl Ϫ -free medium was, as expected, higher than that of control cells. Nevertheless, upon addition of (NH 4 ) 2 SO 4 there was a significant further alkalinization; the rate of recovery from which was only 5 Ϯ 3% (mean Ϯ S.E., n ϭ 3) the rate of recovery measured for cells in Cl Ϫ -containing medium (as in Fig. 13A).

DISCUSSION
The Flux of H ϩ -equivalents following the Removal and Restoration of Extracellular Cl Ϫ -Upon removal of Cl Ϫ from the extracellular medium (with the consequent creation of an outward Cl Ϫ concentration gradient) the malaria parasite underwent a marked cytosolic alkalinization (Figs. [1][2][3][4][5]. The data are consistent with the presence in the parasite plasma membrane of a Cl Ϫ -dependent pathway which, following the imposition of an outward Cl Ϫ gradient, exports H ϩ -equivalents to the external medium. Upon restoration of Cl Ϫ to the extracellular medium pH i returned to close to its original value. Both the initial alkalinization and the subsequent re-acidification were blocked by the anion transport blocker DIDS (Fig. 5), consistent with the same pathway being involved in the two processes.
As is illustrated schematically in Fig. 14 (A and B), the pathway may be a OH Ϫ /Cl Ϫ antiporter, or a H ϩ /Cl Ϫ symporter. These alternatives are thermodynamically equivalent. Fig. 14A shows the postulated mode of action under physiological con-  ditions (in which the system mediates the uptake of Cl Ϫ and the import of H ϩ -equivalents); Fig. 14B shows the reversal of the transporter in Cl Ϫ -free media. Upon removal of extracellular Cl Ϫ there is a net efflux of Cl Ϫ (down its gradient) and a coupled influx of OH Ϫ or efflux of H ϩ (as in Fig. 14B). Upon restoration of extracellular Cl Ϫ the situation is reversed; there is a net influx of Cl Ϫ and a coupled efflux of OH Ϫ or influx of H ϩ (as in Fig.  14A). The [Cl Ϫ ] o dependence of the rate of re-acidification following addition of Cl Ϫ to cells in a Cl Ϫ -free medium (Fig. 8) is consistent with the putative transporter being saturable, with a K m for Cl Ϫ of 20 Ϯ 4 mM. The data of Fig. 9 are consistent with the transporter having the selectivity sequence NO 3 Ϫ Ͼ Br Ϫ Ϸ Cl Ϫ Ͼ I Ϫ (Fig. 9). The [Cl Ϫ ] o dependence of the maximum pH i attained following a decrease in [Cl Ϫ ] o (Fig. 6) is consistent with a stoichiometry of 3 H ϩ -equivalents per Cl Ϫ (from the initial slopes of the fitted curves), raising the possibility that there is a charge translocation involved in the transport step and that transport may therefore be influenced by the membrane potential; however this requires further investigation.
The ATP Dependence of the pH i Response-The cytosolic alkalinization upon removal of extracellular Cl Ϫ was seen in both energized (ATP-replete) and de-energized (ATP-depleted) parasites, with the same final pH i (ϳ7.9) being reached in both cases. However the characteristics of the pH i response varied significantly with the energy status of the cell.
In ATP-replete parasites the increase was immediate and rapid (Fig. 4B). By contrast, in ATP-depleted parasites the trace was sigmoidal; the rate at which pH i increased was low at first, but increased as pH i increased (Fig. 4A). The sigmoidal shape of the alkalinization time course seen in ATP-depleted parasites may be attributed, at least in part, to the pH i dependence of the system involved. As illustrated in Fig. 7 the system underlying the alkalinization of ATP-depleted parasites showed low (albeit non-zero) activity at pH i Ͻ 7.2; however the activity increased with increasing pH i at pH i values Ͼ 7.2. The increased rate of alkalinization with increasing intracellular pH i (despite there being a decreased thermodynamic driving force for the efflux of H ϩ -equivalents) is consistent with the pathway involved being subject to pH i -dependent kinetic control, with the pH i of 7.2 representing an "activation threshold" for the system. This is directly analogous to the behavior of the mammalian Na ϩ /H ϩ exchanger, the activity of which shows a sigmoidal dependence on pH i (albeit in the opposite direction to that seen here (26)), as well as to the behavior of Cl Ϫ /OH Ϫ and Cl Ϫ /HCO 3 Ϫ exchangers in a number of mammalian cell types (27)(28)(29).
In the present study the resting pH i of ATP-depleted parasites was ϳ7.1, significantly lower than that in ATP-replete parasites (ϳ7.3) and below the pH i activation threshold. The alkalinization seen upon removal of extracellular Cl Ϫ from ATP-depleted cells was therefore very slow until the pH i reached the activation threshold, after which the rate increased, giving rise to the sigmoidal appearance of the trace. Nevertheless, the alkalinization rate seen (at pH i values Ͼ 7.3) in ATPdepleted parasites was well below that seen at equivalent pH i values in ATP-replete cells. By contrast, the rate of acidification  results in a large outward Cl Ϫ gradient that forces the transporter to operate in the reverse direction, effluxing Cl Ϫ either in exchange for OH Ϫ or together with H ϩ , thereby resulting in an alkalinization. The results of this study do not allow us to distinguish between Cl Ϫ /OH Ϫ exchange and H ϩ /Cl Ϫ symport. The "n" prefix indicates the uncertain stoichiometry of the system. In C the dashed line is based on the data from Fig. 7, obtained in ATP-depleted parasites. The postulated shift of the pH i dependence curve under ATP-depleted conditions, away from the normal resting pH i , mirrors the behavior of the mammalian Na ϩ /H ϩ exchanger (26). The black circle indicates the activity of the system in ATP-replete parasites at the normal resting pH i (ϳ7.3). The white circle indicates the activity of the system in ATP-depleted parasites suspended at a pH o of 7.1, under which conditions pH i Ϸ 7.1. The squares indicate the activity of the system in both ATP-replete (black square) and ATP-depleted (white square) parasites under Cl Ϫ -free conditions (pH i Ϸ 7.9). The postulated rightward shift of the activation curve in ATP-depleted parasites could account both for: (i) the observation that the rate of alkalinization seen upon removal of extracellular Cl Ϫ is many-fold higher under ATP-replete conditions than under ATP-depleted conditions (cf. the relative activities of the system at the points indicated by the black and white circles, respectively) and (ii) the observation that the initial rate of re-acidification following the addition of Cl Ϫ to cells in Cl Ϫ -free medium is only slightly faster in ATP-replete cells than in ATP-depleted cells (cf. the relative activities of the system at the points indicated by the black and white squares, respectively). seen on restoration of extracellular Cl Ϫ to ATP-depleted cells was only slightly lower than that seen in ATP-replete cells.
These observations might be accounted for by the hypothesis (represented in Fig. 14C) that the pH i dependence of the system, like that of the mammalian Na ϩ /H ϩ exchanger (26), shifts under conditions of ATP depletion. For the mammalian Na ϩ /H ϩ exchanger, ATP depletion causes the curve describing the pH i dependence of the exchanger to shift such that a larger perturbation of pH i is required to activate the system than is the case under ATP-replete conditions (26). If the same were true for the system described here then, as illustrated schematically in Fig. 14C, this might explain both: (i) the observation that the rate of alkalinization seen upon removal of extracellular Cl Ϫ is much higher under ATP-replete conditions than under ATPdepleted conditions (as represented in Fig. 14C, the activity of the system in ATP-replete cells would be much higher than that in ATP-depleted cells, both at the initial resting pH i values (ϳ7.3 in ATP-replete cells, represented by the closed circle, and ϳ7.1 in ATP-depleted cells, represented by the open circle), and at higher pH i values) and (ii) the observation that the rate of re-acidification following the addition of Cl Ϫ to the medium is only slightly faster in ATP-replete cells than in ATP-depleted cells (at the higher pH i from which this occurs the activity of the system is only slightly greater under ATP-replete conditions, represented by the closed square, than under ATP-depleted conditions, represented by the open square).
Ideally the hypothesis that the pH i dependence of the system of interest here shifts under conditions of ATP depletion might be tested by comparing the pH i dependence of the system in ATP-depleted cells (as in Fig. 7) with that in ATP-replete parasites. However, in ATP-replete parasites the cytosol is exposed to a strong acid load (resulting from metabolism and/or the import of acid-equivalents), which is normally countered by the plasma membrane V-type H ϩ -ATPase (1-3). Although it is possible to inhibit the V-type ATPase, the significant acidification that results (2) makes it difficult to investigate the pH i dependence of the system of interest here under ATP-replete conditions.
One observation that the hypothesis represented in Fig. 14C does not readily account for is the fact that there was a greater discrepancy between the rate of alkalinization and the rate of the subsequent recovery in ATP-depleted cells than in ATP replete cells. The finding, that in ATP-depleted cells undergoing a second cycle of exposure to a Cl Ϫ -free medium, the rate of alkalinization was significantly increased (data not shown), implies that there are as yet unidentified factors regulating the activity of the system(s) involved. Elucidating these factors is beyond the scope of the present study.
Uptake of Cl Ϫ by the Parasite-In this study [Cl Ϫ ] i was estimated using two independent methods. For isolated parasites suspended in medium containing 132 mM Cl Ϫ , measurements of the equilibration of 36 Cl Ϫ yielded an estimate of 69 Ϯ 12 mM. Fluorescence measurements using the Cl Ϫ -sensitive indicator MQAE yielded a very similar estimate of 70 Ϯ 6 mM. As is shown in Fig. 11C (9)), the [Cl Ϫ ] i in the parasite was estimated (using MQAE) as 48 Ϯ 7 mM (mean Ϯ S.E., n ϭ 5), close to the x-ray microanalysis estimate of 40 mM (9).
The results obtained here confirm that the [Cl Ϫ ] i of trophozoites is substantially higher than the Ͻ3 mM expected if Cl Ϫ were simply distributed passively across the parasite plasma membrane in accordance with the (Ϫ95 mV) membrane potential. The parasite must therefore take up Cl Ϫ from its external environment (i.e. the erythrocyte) via one or more systems that maintain [Cl Ϫ ] i away from electrochemical equilibrium.
The influx of 36 Cl Ϫ into isolated parasites was slowed some 4-fold by ATP depletion (Fig. 10A). In glucose-replete cells, 36 Cl Ϫ influx increased in response to intracellular alkalinization and decreased in response to intracellular acidification (Fig. 12B). The pH i dependence of 36 Cl Ϫ influx was similar to that of the efflux of H ϩ -equivalents following the removal of extracellular Cl Ϫ in glucose-depleted cells (Fig. 7). In addition, both 36 Cl Ϫ uptake and the Cl Ϫ -dependent flux of H ϩ -equivalents were inhibited by DIDS (Figs. 10B and 5, respectively). The available data are therefore consistent with the uptake of Cl Ϫ , and the Cl Ϫ -dependent flux of H ϩ -equivalents, occurring through a common pathway that couples the transport of Cl Ϫ to the transport of H ϩ -equivalents. This might be achieved through an OH Ϫ /Cl Ϫ antiporter or an H ϩ /Cl Ϫ symporter (as represented schematically in Fig. 14, A and B). Such a system would provide a mechanism by which, under physiological conditions, the inward H ϩ electrochemical gradient (1, 2) could energize the accumulation of Cl Ϫ , as well as mediating a recovery of pH i in the event of a cytosolic alkalinization.