Introduction

Top Summary Introduction Materials & Methods Results Discussion References

Current theory suggests that chloroquine kills parasites by preventing the detoxification of hematin inside the parasite food vacuole (Slater, 1993). Heme, which is released as a byproduct of hemoglobin digestion, is oxidized rapidly to hematin and sequestered into hemozoin or malarial pigment by an autocatalytic mechanism (Dorn et al., 1995). Chloroquine inhibits the polymerization process in vitro and is proposed to do the same in vivo, causing a build up of free hematin or hematin/chloroquine complex that would ultimately kill the parasite (Slater, 1993; Sullivan et al., 1996). Alternatively, it has been proposed that weakly basic chloroquine accumulates to high levels in the acid food vacuole by a proton-trapping mechanism (Yayon et al., 1985). Thus, chloroquine could kill parasites in its own right by direct inhibition of vacuolar enzymes such as phospholipase (Ginsburg and Geary, 1987) or proteinase (Vander Jagt et al., 1987).

A proportion of the total chloroquine taken up by Plasmodium falciparum is saturable at nanomolar drug concentrations, and early studies hinted that this saturable component is important for drug activity (Fitch, 1970). There also is a component of chloroquine uptake that is nonsaturable in the submicromolar range (Fitch, 1970). Recent evidence suggests that nonsaturable uptake can be attributed to low affinity chloroquine binding to plentiful cytosolic proteins (Menting et al., 1997; Dorn et al., 1998). The contribution of both saturable and nonsaturable uptake to the antimalarial activity of chloroquine is unknown.

Chloroquine readily forms a complex with free hematin in vitro (Chou et al., 1980), and it was originally suggested that saturable uptake into cells is due to the binding of chloroquine to hematin inside the food vacuole. Chloroquine/hematin binding clearly occurs to some extent in situ. Radiolabeled quinolines are found to be associated with hemozoin after prolonged incubation of infected cells with sublethal concentrations of drug (Sullivan et al., 1996), and intracellular quinoline/hematin interactions have been detected by photoacoustic spectroscopy (Balasubramanian et al., 1984). However, the contribution of chloroquine/hematin binding to the saturable cellular uptake and the importance of this binding for the antimalarial activity of the drug have not been determined. Accordingly, the hematin binding hypothesis has been challenged on quantitative grounds, and it was suggested that the uptake of chloroquine is determined mainly by the titration of protons inside the food vacuole (Ginsburg and Geary, 1987; Yayon et al., 1985). In this case, the observed saturation kinetics would reflect saturation of the vacuolar proton pump rather than hematin binding (Ginsburg and Stein, 1991). More recently, the saturation kinetics of chloroquine uptake have been attributed to a carrier-mediated chloroquine transport process: Based on the ability of 5-(N-ethyl-N-isopropyl)amiloride, a specific inhibitor of Na⁺/H⁺ antiport, to competitively inhibit chloroquine uptake, it was suggested that chloroquine is directly transported by the parasite plasma membrane Na⁺/H⁺ exchanger, in place of sodium and in exchange of protons (Sanchez et al., 1997). In a modification of this theory, it was later suggested that chloroquine is carried through the Na⁺/H⁺ exchanger in a burst of self-stimulated sodium/proton exchange (Wünsch et al., 1998).

Using a two-component model, we have shown that the saturable component of chloroquine uptake is solely responsible for the antimalarial activity of the drug. Further studies have focused on the nature of saturable chloroquine uptake in the context of the theories outlined above. We used Ro 40-4388, a potent and specific inhibitor of plasmepsin I (Moon et al., 1997), to assess the importance of chloroquine/hematin binding. Plasmepsin I acts on native hemoglobin, cutting the Phe33---Leu34 bond of the  chain, which unfolds the hemoglobin tetramer and allows the release of heme (Gluzman et al., 1994). We believe that Ro 40-4388 allows chloroquine/hematin binding to be differentiated from alternative uptake mechanisms that may operate inside intact cells. This inhibitor will inhibit the release of heme by blocking the first step in the degradation of hemoglobin but is not likely to affect the Na⁺/H⁺ exchanger or the proton gradient. Binding parameters have been measured in chloroquine-susceptible and -resistant strains in the presence or absence of resistance modulators such as verapamil. The differences in the cellular accumulation and activity of chloroquine in susceptible and resistant strains, together with the effects of verapamil, can be fully described by the chloroquine-hematin binding parameters. Analysis of the data indicates that there is no change in the rate of hemoglobin digestion or heme sequestration in resistant strains. Instead, the chloroquine-resistant parasite has evolved a mechanism to reduce the access of chloroquine to hematin.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Parasites. The chloroquine susceptible 3D7, T9-96, and HB3 clones were obtained from Prof. D. Walliker, Edinburgh University (Edinburgh, UK). The K1 chloroquine-resistant strain (cloned in-house) was obtained from Dr, D. C. Warhurst, London School of Hygiene and Tropical Medicine (London, UK), and the TM5 and TM6 chloroquine-resistant clones were obtained from Dr. P. Tan-Ariya, Mahidol University (Bangkok, Thailand). Parasites were maintained in continuous culture using standard techniques (Desjardins et al., 1979).

Sensitivity assays. Sensitivity of all the strains to chloroquine, daunomycin, primaquine, and verapamil was determined by measuring the ability of serial dilutions of drugs to inhibit the incorporation of radiolabeled [³H]hypoxanthine into parasite nucleic acids (Desjardins et al., 1979). Chloroquine assays were performed in the absence or presence of 5 µM verapamil or 1.5 µM primaquine, at the lowest practicable inoculum size (0.5% parasitemia, 1% hematocrit). IC₅₀ values were calculated for each assay using the four-parameter logistic method (Grafit; Erithacus Software, Kent, UK), and values presented are the mean of five sensitivity assays, unless otherwise stated.

Accumulation of radiolabeled drugs. Radiolabeled [³H]chloroquine (specific activity, 50.4 Ci/mmol), [³H]verapamil (specific activity, 66 Ci/mmol), and [³H]daunomycin (specific activity, 4 Ci/mmol) were purchased from DuPont-New England Nuclear (Boston, MA). Radiolabeled [³H]amodiaquine (specific activity, 106.5 mCi/mmol) was synthesized as described previously (Hawley et al., 1996a). Radiolabeled [¹⁴C]primaquine (specific activity, 85 mCi/mmol) was a gift from Prof. W. Peters.

Modeling saturable and nonsaturable chloroquine uptake and the relationship with antimalarial activity. Total uptake versus external chloroquine concentration data were fitted by computer to eq. 1 using an iterative procedure (Grafit). The slope of the nonsaturable component was determined graphically and checked by using this value as an initial estimate for iteration. The amount of nonsaturable uptake at each concentration was calculated by multiplying the external concentration by the slope of the nonsaturable uptake component. Saturable uptake at each concentration was calculated by subtracting nonaturable uptake from total uptake at each concentration.

(1)

(2)

(3)

(4)

Association of preloaded bound chloroquine with hematin-containing cell debris after cell lysis. Cells infected with HB3 or K1 strain parasites (5% parasitemia, 2% hematocrit), synchronized at the trophozoite stage were loaded with [³H]chloroquine in complete medium, under gas, for 30 min at 37°. The loaded cells then were separated from the medium by centrifugation, and the cell pellet was lysed by quick freezing in liquid nitrogen, followed by thawing at 5-7°C. This procedure was repeated, and the cell debris was diluted into 50 volumes of distilled water buffered to pH 7.4 with 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (temperature, 5-7°C) containing antimalarial drugs or chemosensitizers at the appropriate concentrations and shaken vigorously for 20 sec. Samples were centrifuged in a microcentrifuge (12,000 × g for 5 min at 4°); a sample of the supernatant was taken for measurement, and the remainder was discarded. The cell debris was processed for scintillation counting as described previously (Bray et al., 1994). The water space associated with the cell debris was estimated by using ¹⁴C-inulin, and the pellet counts were corrected accordingly.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

Chloroquine is not transported as a simple sodium substitute by the parasite Na⁺/H⁺ exchanger. Based on the inhibition of saturable chloroquine uptake by EIPA (a specific blocker of the plasma membrane Na⁺/H⁺ exchanger), it recently has been proposed that chloroquine is actively imported into the parasite via the Na⁺ binding domain of the Na⁺/H⁺ exchanger, instead of sodium and in exchange of protons (Sanchez et al., 1997). If this hypothesis is correct, removal of competing substrate (sodium) should increase chloroquine accumulation. In fact, chloroquine accumulation is decreased significantly when sodium in the medium is replaced by choline or N-methyl-D-glucamine (Fig. 1), suggesting that chloroquine is not simply substituted for sodium as a substrate for proton exchange. Furthermore, phenotypic differences in the accumulation of chloroquine by chloroquine-susceptible and -resistant clones are largely maintained in sodium-free buffer. For example, the HB3 chloroquine-susceptible clone takes up three to four times more chloroquine than the K1 chloroquine-resistant clone in sodium-free buffer, and clone-specific differences in chloroquine accumulation due to chemosensitizers such as verapamil and daunomycin are maintained in sodium-free buffer (Fig. 1).

View larger version (47K): [in this window] [in a new window]   Fig. 1.   Effect of medium sodium on steady state chloroquine accumulation. Steady state cellular accumulation ratio of [3H]chloroquine in the HB3 clone (top) or the K1 clone (bottom), determined in medium containing sodium, choline, or N-methyl-D-glucamine in the absence of chemosensitizers (light shading) or the presence of 10 µM verapamil (intermediate shading) or 5 µM daunomycin (dark shading). Data represent mean ± standard deviation values of single observations from 15 separate experiments.

Saturable chloroquine binding characteristics of resistant and susceptible strains. We found that all six strains tested exhibited both saturable and nonsaturable components of chloroquine accumulation. Fig. 2, top, shows representative derived saturable accumulation data from experiments using the K1 chloroquine-resistant clone and the HB3 chloroquine-susceptible clone.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Characterization of the saturable equilibrium binding of chloroquine. Top, equilibrium binding of [3H]chloroquine was determined as a function of the external chloroquine concentration. Data are presented for the HB3 (CQS, ) and the K1 (CQR, ) clones. Also plotted are data for the K1 clone in the presence of 5 µM verapamil (). Data represent mean ± standard deviation values of observations from three to five separate experiments. Bottom, Hill plot of the data presented in Fig. 2, top.

The relationship of chloroquine accumulation and activity. Fig. 3, top, depicts the relationship between total cellular chloroquine accumulation (measured at external concentrations equivalent to the IC₅₀) and activity (taken as the IC₅₀) for all the strains in the presence or absence of verapamil. The relationship is nonlinear and can be fitted to a double-exponential curve as illustrated. It is clear that neither the degree of resistance nor the reversal of resistance can be adequately explained by differences in the simple total cellular accumulation of chloroquine.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   The activity of chloroquine is linearly related to the apparent receptor Kd rather than total cellular chloroquine accumulation. Top, total cellular accumulation ratio at IC50 was interpolated from graphs of chloroquine accumulation versus external chloroquine concentration. Line, fit of the data to the double exponential equation: y = I1·e + I2·e, where I1 = initial value 1 (at x = 0) = 335, k1 = rate 1 = 0.0001, I2 = initial value 2 (at x = 0) = 1606, and k2 = rate 2 = 0.0416. This equation has no theoretical justification but has a suitable mathematical form to allow a smooth curve to run through the data. IC50 values are mean values of five independent experiments. Bottom, apparent affinity data were taken from Table 1. Line, least-squares linear regression fit to the data.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Validation of the two-component model. Top, data from three CQS strains and three CQR strains with and without 5 µM verapamil were fitted to eq. 4. Line, least-squares linear regression fit to the data. Bottom, data from three CQS strains and three CQR strains with and without 5 µM verapamil were fitted to eq. 3. Line, least-squares linear regression fit to the data.

Displacement of transiently bound chloroquine. After lysis and centrifugation, pellets of cell debris were dark brown (implying the presence of hemozoin/hematin), contained no hemoglobin, and by light microscopy contained no intact erythrocytes or parasites. Cell debris from parasitized erythrocytes contained ~50% of the radioactivity associated with saturable chloroquine binding to intact infected erythrocytes, with concentrations several hundredfold greater than can be explained by the water space. In contrast, cell debris from preloaded uninfected erythrocytes contained no more chloroquine than that due to the water space. The nonsaturable component of chloroquine accumulation seen with intact infected cells was completely lost when using the cell debris. The radiolabeled chloroquine recovered reflected only the saturable component of uptake into intact cells (Fig. 5). The apparent K_d value of the chloroquine bound to the cell debris (21. 56 nM) was very similar to the apparent K_d value of chloroquine bound to intact cells (21.0 nM, Table 1), although the capacity was reduced to about half of that measured with intact cells (17.3 ± 2.9 µM versus 34.0 ± 12 µM). The inability to recover all the saturable radioactivity is due to the transient nature of binding, which decreases progressively with time (data not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Saturable chloroquine binding associated with cell debris. Cells infected with the HB3, CQS clone were preloaded with [3H]chloroquine for 1 hr, at 37°, in the presence of increasing concentrations of nonradiolabeled chloroquine as indicated. Cells were lysed, and the associated radioactivity was counted as described in Experimental Procedures. Values are mean ± standard deviation of duplicate observations from three separate experiments.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Displacement of chloroquine bound to cell debris from cells preloaded with [3H]chloroquine. Cell debris was exposed to verapamil (), primaquine (), daunomycin (), amodiaquine (), or nonradiolabeled chloroquine ( at the concentrations indicated. Top, data for the HB3, CQS clone. Bottom, data for the K1, CQR clone. Results represent mean ± standard deviations of three independent experiments.

Chloroquine accumulation is reduced by an inhibitor of plasmepsin I. Data presented in Fig. 7 for the HB3 clone show that chloroquine accumulation by intact parasitized erythrocytes is markedly reduced by the plasmepsin I inhibitor Ro 40-4388 in a concentration-dependent manner. A similar effect was seen with the K1 chloroquine-resistant clone (data not shown). In contrast, leupeptin, a nonspecific cysteine proteinase inhibitor, did not affect chloroquine accumulation. These results suggest that a large proportion of chloroquine accumulation by CQS strains could be due to the binding of chloroquine to hematin because plasmepsin I is the enzyme responsible for the initial cleavage of hemoglobin, so the inhibition of this enzyme also would be expected to inhibit the supply of hematin.

View larger version (23K): [in this window] [in a new window]   Fig. 7.   The effect of increasing concentrations of the protease inhibitors leupeptin and Ro 40-4388 on the accumulation of chloroquine by the HB3 clone of Plasmodium falciparum. Drug accumulation was measured over 1 hr after a 5-min preincubation with the protease inhibitors leupeptin () or Ro 40-4388 (). Values are mean ± standard deviation of single observations from five separate experiments.

View larger version (27K): [in this window] [in a new window]   Fig. 8.   The effect of Ro 40-4388 is specific for weak base drugs that bind to hematin. Accumulation of radiolabeled drugs was measured over 1 hr after a 5-min preincubation with 10 µM Ro 40-4388. Histogram bars, mean ± standard deviation values of single observations from five separate experiments.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Scatchard plot of saturable equilibrium chloroquine binding to hematin. Binding was measured in the presence () or absence () of 5 µM Ro 40-4388 using the HB3 chloroquine-susceptible clone. Saturable chloroquine binding was determined as described in Experimental Procedures. Values are mean of single observations from three separate experiments. Lines, least-squares linear regression fit to the data points.

View larger version (21K): [in this window] [in a new window]   Fig. 10.   Ro 40-4388 inhibits the verapamil effect. The K1 clone parasites were preincubated in the presence or absence of 10 µM Ro 40-4388 for 5 min. Samples from each group were incubated for 1 hr with 1 nM [3H]chloroquine in the presence (dark shading) or absence (light shading) of 5 µM verapamil. Histogram bars, mean ± standard deviation of single observations from five separate experiments.

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

Data presented in this report provide evidence that the saturable component of chloroquine accumulation is responsible for the antimalarial activity of the drug [i.e., the K_d value of saturable uptake and the IC₅₀ are well correlated (Fig. 3, bottom)]. Although the binding of chloroquine to hematin is an extremely persuasive explanation for saturable chloroquine accumulation, there are at least two other theories that deserve serious consideration.

First, we considered that the proton gradient into the food vacuole drives the uptake of weakly basic antimalarials (Yayon et al., 1985). It has been amply demonstrated that the vacuolar pH gradient is important for the accumulation of chloroquine and related drugs (Yayon et al., 1985; Geary et al., 1990; Bray et al., 1992a; Martiney et al., 1995; Hawley et al., 1996c). It also has been demonstrated that these drugs titrate protons inside the food vacuole and raise the vacuolar pH (Krogstad et al., 1985; Yayon et al., 1985; Ginsburg et al., 1989). Therefore, it is quite possible that the saturable chloroquine accumulation observed with intact cells could stem from a saturation of the vacuolar proton pumping capacity (Ginsburg and Stein, 1991) rather than a saturation of chloroquine-hematin binding, and this possibility was considered in the current work.

We also have considered the possibility that both the saturability and the specificity of chloroquine accumulation are attributable to a drug importer or permease in the parasite (Warhurst, 1986). In support of this theory, it was proposed recently that the parasite Na⁺/H⁺ exchanger is responsible for saturable chloroquine uptake (Sanchez et al., 1997). Two transport mechanisms have been proposed: (1) chloroquine could be an alternative substrate for the exchanger, substituting for sodium in direct exchange for protons (Sanchez et al., 1997), or (2) chloroquine could be carried through the exchanger in a burst of sodium/proton exchange stimulated by chloroquine itself (Wünsch et al., 1998).

Sodium-free buffer produces a rapid and reversible acidification of the parasite cytosol by 0.5 pH unit (Bosia et al., 1993). The cytosolic pH of the parasite in sodium buffer is at most 0.2 pH unit higher than the pH of the host cell cytosol (Wünsch et al., 1998). Therefore, in sodium-free buffer containing 1 nM chloroquine, there is a huge molar excess of protons available to drive chloroquine uptake by a proton exchange mechanism (from values for cytosolic buffer capacity given in Wünsch et al., 1998, a gradient of 0.3 pH unit will provide ~50 mmol of protons for every nmol of chloroquine). In a situation of direct chloroquine/proton exchange, this would predict 5 × 10⁷-fold drug accumulation. In fact, chloroquine accumulation in sodium-free medium is >4 orders of magnitude lower than this prediction (Fig. 1). Furthermore, instead of being enhanced by sodium-free buffer compared with sodium buffer (as might be expected if a competing substrate is removed), chloroquine accumulation actually is reduced significantly (Fig. 1). These data indicate that chloroquine is not a substrate for direct stoichiometric proton exchange. Nevertheless, a significant amount of chloroquine is taken up in the absence of sodium by both chloroquine-susceptible and -resistant parasites (Fig. 1). This uptake is saturable, stimulated by chemosensitizers, and uniquely dependent on the availability of hematin (Bray PG, Mugthin M, and Ward SA. Verapamil-sensitive chloroquine resistance is not determined by the Na⁺/H⁺ exchanger in Plasmodium falciparum, manuscript in preparation). With regard to the second proposed active uptake mechanism outlined above, it is difficult to see how this saturable chloroquine uptake could in any way be attributed to a burst of rapid sodium/hydrogen exchange on stimulation of the Na⁺/H⁺ exchanger. Finally, the specific chloroquine accumulation/chemosensitization phenotype of chloroquine-susceptible and -resistant clones is preserved in sodium-free medium (Fig. 1), suggesting that the chloroquine-resistance phenotype may be attributable to mutation of proteins other than the Na⁺/H⁺ exchanger. We believe that these data cast doubt on the involvement of direct chloroquine transport through the parasite Na⁺/H⁺ exchanger in both the uptake of chloroquine by the parasite and in the mechanism of chloroquine resistance. However, in keeping with the large body of evidence showing that pH gradients influence the accumulation of chloroquine (see above), we acknowledge that any of the pH regulation mechanisms of the parasite (including the Na⁺/H⁺ exchanger) may be indirectly involved in the uptake of chloroquine.

The techniques and new inhibitor used in this study have, for the first time, allowed discrimination between the proposed mechanisms of drug accumulation in P. falciparum. The substantial amount of saturable chloroquine accumulation associated with the cell debris (Figs. 5 and 6) can only reflect chloroquine binding to hematin rather than proton trapping or active uptake, both of which require a high degree of membrane integrity. Much of the heme/chloroquine complex undoubtedly will be present in the supernatant as well as the cell debris, but our results are in line with the demonstrated ability of hematin and hematin aggregates to bind chloroquine while adsorbed to cell debris (Chou et al., 1980).

Ro 40-4388 is a potent and specific inhibitor of plasmepsin I (Moon et al., 1997), the P. falciparum aspartic proteinase that is responsible for the initial cleavage of hemoglobin and release of free heme (Francis et al., 1994). Therefore, our proteinase inhibitor studies have established that the saturable chloroquine accumulation is dependent on the initial cleavage of hemoglobin and release of heme (Figs. 7 and 9). The lack of effect of this inhibitor on the accumulation of other weak base drugs that do not bind hematin argues against a nonspecific effect on the vacuolar pH (Fig. 8). Finally, it is difficult to see how this specific aspartic proteinase inhibitor could reduce the capacity of the proposed chloroquine importer. Although we acknowledge the importance of the pH gradient for the overall process of chloroquine accumulation, we propose that the saturable chloroquine accumulation that is responsible for the specific activity of chloroquine (Fig. 3) is attributable solely to chloroquine-hematin binding.

Free hematin exists only transiently in situ, and the amount available for drug binding is difficult to measure. Accordingly, the hematin binding hypothesis has been questioned on quantitative grounds when the rate of hemoglobin digestion was found to be insufficient to account for the large total amount of chloroquine accumulated by infected cells (Ginsburg and Geary, 1987). These calculations were made using measurements of whole-cell accumulation. No attempt was made to distinguish saturable accumulation from nonsaturable accumulation. This distinction is critical as the major proportion of chloroquine accumulation in resistant strains is nonsaturable (i.e. not bound to hematin) (Fig. 4, bottom). In fact, using published estimates of the rate of hemoglobin digestion (Ginsburg and Geary, 1987), we calculate that only ~4% of the heme turnover would need to be available to bind chloroquine to account for our measurements of saturable uptake, provided we assume that most of the hemoglobin digestion occurs in the mid to late trophozoites. These data are in agreement with the small effect of chloroquine on the polymerization of hematin that is observed in situ as the parasites are killed (Meshnick, 1996). The important question remains as to how interaction of chloroquine with hematin leads to parasite death. Our data are consistent with a suggestion that the formation of chloroquine/hematin complexes, rather than a build-up of free hematin, is the prime cause of cell killing and that hematin polymerization is only a secondary consequence of this. We are currently investigating a hypothesis that enhanced lipid solubility of the chloroquine/hematin complex (Ward SA, unpublished observations) allows it to escape from the food vacuole along a concentration gradient. Once the vacuolar membrane has been crossed, the elevated pH and lower hematin concentrations in the cytosol would make polymerization much less likely, and the hematin or drug/hematin complex then could interact with a wide range of vital cellular targets.

In terms of the mechanism of chloroquine resistance, the capacity of chloroquine-hematin binding is the same for both chloroquine-susceptible and -resistant strains and is not altered by modulators of chloroquine resistance (Table 1). These observations are incompatible with chloroquine resistance mediated by a mechanism of reduced production of heme, accelerated sequestration of hematin, or peroxide-mediated decomposition of heme (Fitch, 1989). Thus, several theories to explain chloroquine resistance can now be ruled out. Our data demonstrate that chloroquine resistance is associated with reduced apparent affinity of chloroquine-hematin binding rather than changes in capacity (Table 1). It is difficult to envisage a mechanism by which the true affinity of hematin for chloroquine could be altered by the parasite. Consequently, our data suggest that resistant parasites have evolved a mechanism to reduce the accessibility of hematin rather than alter its structure (Fig. 6).

It was largely due to the publication of many data demonstrating pH-dependent uptake and proton trapping that the chloroquine-hematin binding hypothesis has fallen from favor (Ginsburg and Geary, 1987). However, because the pH gradient plays a role in concentrating the amount of chloroquine at the site of hemoglobin degradation, we believe that the two hypotheses are compatible. The apparent affinity of chloroquine binding can be readily altered by manipulating the pH gradient from outside or inside. For example, the apparent affinity is more than doubled by increasing the medium pH by 0.3 pH unit (Table 1). A widely supported model of chloroquine resistance predicts that resistant parasites have an elevated vacuolar pH (Ginsburg and Stein, 1991). Mathematical models suggest that increased vacuolar pH will reduce the rate of uptake of chloroquine (Ferrari and Cutler, 1991; Ginsburg and Stein, 1991). In support of this, it would be fair to say that the majority of kinetic studies measure a reduced rate of chloroquine accumulation into chloroquine-resistant compared with chloroquine-susceptible strains (Ginsburg and Stein, 1991; Bray et al., 1992a, 1994, 1996; Martiney et al., 1995; Sanchez et al., 1997; Wünsch et al., 1998). Such observations are entirely compatible with our interpretation of the data in this report. Because heme is being released and hematin is sequestered at a fixed rate, any process that reduces the rate of uptake of chloroquine would reduce the relative amount of drug available to bind hematin per unit time. This would reduce the apparent affinity and extent of saturable drug accumulation at steady state.

The "resistance reversing" effects of verapamil pose problems for all the current theories of chloroquine resistance. First, it is difficult to see how verapamil could decrease vacuolar pH as required by the pH model, and second, the increased cellular chloroquine accumulation brought about by verapamil is much too small to explain resistance reversal by simple inhibition of rapid efflux (Bray et al., 1994). The lack of an unifying hypothesis to account for both chloroquine resistance and resistance reversal prompted the proposal that verapamil-sensitive chloroquine resistance may be multifactorial (Ginsburg and Stein, 1991; Bray et al., 1994; Sanchez et al., 1997). Our data correlate both chloroquine resistance and the verapamil effect with the apparent affinity of chloroquine/hematin binding, suggesting a common mechanism (Fig. 3). Although our data do not rule out the involvement of multiple genes in the process of chloroquine resistance, our findings are in accord with a single genetic locus for the verapamil effect and chloroquine resistance phenotype, as demonstrated in a genetic cross (Wellems et al., 1991). At the molecular level, it is possible that the candidate resistance protein CG2 (Su et al., 1997) alters the cellular drug distribution directly by pumping drug out of the infected cell (Krogstad et al., 1987). It is more likely however, given the data presented here, that such a protein acts specifically at the target site, either by altering trans-vacuolar ion gradients or possibly by binding to free hematin itself and reducing the accessibility to chloroquine.

The chloroquine-hematin binding hypothesis was first proposed in the 1960s by Macomber et al., (1967) and later championed by Fitch (1989). Although compelling, the hypothesis was never proved, and many other target molecules were proposed (Ginsburg and Geary, 1987). The data presented here finally provide good evidence in support of the original hypothesis. The unparalleled success of chloroquine, before the evolution and spread of resistance, is testimony to what can be achieved by targeting the process of hemoglobin catabolism in the parasite. Our data suggest that P. falciparum is uniquely vulnerable to hematin binding drugs in that it cannot alter the quantity or the nature of the hematin target. Instead, the parasite has evolved in a way to reduce the accessibility of hematin to some of these drugs. Regardless of how such resistance is achieved at the molecular level, it is evident that the resistance mechanism can be overcome by relatively simple alterations of the basic 4-aminoquinoline structure (Bray et al., 1996; De et al., 1996; Hawley et al., 1996b; Ridley et al., 1996). For these reasons, we believe that great priority must be given to the development of novel hematin-binding agents.