Resistance to chloroquine of malaria strains is known to be associated with a parasite protein named PfCRT, the mutated form of which is able to reduce chloroquine accumulation in the digestive vacuole of the pathogen. Whether the protein mediates extrusion of the drug acting as a channel or as a carrier and which is the protonation state of its chloroquine substrate is the subject of a scientific debate. We present here an analytical approach that explores which combination of hypotheses on the mechanism of transport and the protonation state of chloroquine are consistent with available equilibrium experimental data. We show that the available experimental data are not, by themselves, sufficient to conclude whether the protein acts as a channel or as a transporter, which explains the origin of their different interpretation by different authors. Interestingly, though, each of the two models is only consistent with a subset of hypotheses on the protonation state of the transported molecule. The combination of these results with a sequence and structure analysis of PfCRT, which strongly suggests that the molecule is a carrier, indicates that the transported species is either or both the mono and di-protonated forms of chloroquine. We believe that our results, besides shedding light on the mechanism of chloroquine resistance in P. falciparum, have implications for the development of novel therapies against resistant malaria strains and demonstrate the usefulness of an approach combining systems biology strategies with structural bioinformatics and experimental data.
Citation: Chinappi M, Via A, Marcatili P, Tramontano A (2010) On the Mechanism of Chloroquine Resistance in Plasmodium falciparum. PLoS ONE 5(11): e14064. https://doi.org/10.1371/journal.pone.0014064
Editor: Vladimir Brusic, Dana-Farber Cancer Institute, United States of America
Received: June 16, 2010; Accepted: October 28, 2010; Published: November 19, 2010
Copyright: © 2010 Chinappi et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by King Abdullah University of Science and Technology (KAUST) - Award No. KUK-I1-012-43 (http://www.kaust.edu.sa/), Human Frontier Science Program (HFSP) - RGP0054/2006-C grant (http://www.hfsp.org/), and Fondazione Roma (http://www.fondazioneroma.it/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
In the last decades, due to its effectiveness and reasonable cost, chloroquine has represented the best and more widely used antimalarial drug. Unfortunately, within a decade of its introduction, P. falciparum parasite resistance to chloroquine was observed in most of the malaria-endemic countries. Nowadays, insurgence of resistance against chloroquine is a considerable hurdle for malaria control .
In its erythrocyte stage, P. falciparum invades the red blood cells where it forms a lysosomal isolated acidic compartment known as the digestive vacuole (DV). In the erythrocyte, the parasite grows by ingesting haemoglobin from the host cell cytosol and depositing it in the DV, where the protein is degraded to its component peptides and heme, which is incorporated into the inert and harmless crystalline polymer hemozoin .
Chloroquine is a diprotic weak base and, at physiological pH (∼7.4), can be found in its un-protonated (CQ), mono-protonated (CQ+) and di-protonated (CQ++) forms. The uncharged chloroquine is the only membrane permeable form of the molecule and it freely diffuses into the erythrocyte up to the DV. In this compartment, chloroquine molecules become protonated and, since membranes are not permeable to charged species, the drug accumulates into the acidic digestive vacuole ,  where it is believed to bind haematin, a toxic byproduct of the haemoglobin proteolysis , , preventing its incorporation into the haemozoin crystal , , , , . The free haematin seems to interfere with the parasite detoxification processes and thereby damage the plasmodium membranes .
Chloroquine sensitive parasites (CQS) accumulate much more chloroquine in the DV than chloroquine resistant strains (CQR) , , . Recent studies have associated the reduced chloroquine accumulation observed in the parasite vacuole of resistant strains  with point mutations in the gene encoding for the P. falciparum chloroquine resistance transporter (PfCRT) protein (for a review see , ). PfCRT is localized in the digestive vacuole membrane and contains 10 predicted membrane-spanning domains , . CQR phenotype isolates have all been found to carry the PfCRT critical charge-loss mutation K76T or, in two single cases, K76N or K76I , , , . Another mutation, S163R, restores the chloroquine sensitivity of CQR parasites , . The K76T amino acid mutation might allow the interaction of PfCRT with the positively charged chloroquine (CQ+ or CQ++) and allow its exit from the vacuole, with the net result of decreasing the chloroquine concentration within the DV , . The single amino acid change S163R, by reintroducing a positive charge, is thought to block the leak of charged chloroquine from the DV, thus restoring chloroquine sensitivity , . In a recent work, Martin and collaborators  were able to express both wild-type and resistant forms of PfCRT on the surface of Xenopus laevis oocytes and clearly demonstrated that chloroquine resistance is due to the direct transport of a protonated form of the drug out of the parasite vacuole via the K76T PfCRT mutant. Interestingly, they also showed that the introduction of the K76T single mutation in PfCRT of CQS parasites is necessary but not sufficient for the transport of chloroquine via PfCRT. These evidences are however compatible with two alternative models for PfCRT : (1) the channel model (i.e. a passive channel that enables charged chloroquine to leak out of the food vacuole down its electrochemical gradient) or (2) the carrier model (i.e. an active efflux carrier extruding chloroquine from the food vacuole).
Several experimental set-ups have been used to answer the question of whether PfCRT is a channel or a carrier, namely measures of chloroquine accumulation, trans-stimulation and measures of chloroquine efflux. However the available data have been interpreted in different ways by different authors and the debate about the nature of PfCRT is still ongoing.
Sanchez and colleagues showed that chloroquine accumulation is energy dependent in both CQR and CQS . These authors monitored the time course of labeled chloroquine uptake in the absence and in the presence of glucose. Glucose was added 20 min after choloroquine addition (i.e. when the stationary state was reached). They found that, after glucose addition, the time courses of choroquine uptake were markedly different in CQS and CQR: chloroquine accumulated to an increased extent in the CQS strain, but decreased in the CQR strain. A similar experiment was repeated by the same authors in 2004  using a broader range of different antimalarial drugs. The authors concluded that the data are compatible with most models that attempt to account for chloroquine resistance and that some energy-dependent mechanism leads to loss of chloroquine from CQR cells and to its accumulation in CQS cells.
Bray et al , in 2006, measured the Cellular Accumulation Ratio (CAR) of chloroquine in six experimental conditions, namely in sensitive and resistant strains, in the absence and presence of carbonylcyanide p-trifluoromethoxyphenylhydrazone (FCCP), a ionophoric uncoupling agent, and in the absence and presence of glucose. In particular they found that, in absence of glucose, chloroquine the CAR is equal in CQS and CQR strains (∼700), reaching a level that is approximately intermediate between that observed in CQS (∼1200) and CQR (∼350) strains in the presence of glucose. They used several different Plasmodium strains and showed that, in the absence of FCCP, i.e. when the pH of the vacuole is lower than the external pH, the chloroquine CAR is three to four times higher (about 1200 versus about 350) in sensitive strains with respect to resistant strains, while addition of FCCP abolishes the differences leading to a CAR value of about 700 in both cases. They also demonstrate that, in the absence of glucose, the CAR is identical to that obtained in the presence of FCCP suggesting that the energy provided by the glucose is needed to maintain the pH difference between the cytoplasm and the DV. According to the authors, the hypothesis that PfCRT is an active efflux carrier does not appear to fully explain their findings. In this hypothesis, in fact, a single mutation would transform an energy-dependent chloroquine uptake process in an energy-dependent chloroquine efflux process. Therefore they favor the hypothesis that the chloroquine movement through PfCRT is not an active process.
Trans-stimulation of labeled chloroquine ([3H]-CQ)) uptake after the parasites were pre-loaded with increasing concentrations of unlabelled chloroquine , , ,  was observed in CQR strains and not in CQS isolates. Sanchez and collaborators conclude that the trans-stimulation phenomenon is unequivocally characteristic of saturable, carrier-mediated transport systems . On the other hand, Bray and colleagues  propose that the trans-stimulation data reported by Sanchez et al cannot by themselves be used to conclude whether the chloroquine transport is in the inward or outward direction: stimulation of [3H]-CQ uptake could indeed be due to acceleration of the transporter cycle by the outgoing unlabelled chloroquine or, as Sanchez et al assert, it could result from reduced efflux of [3H]-CQ due to the carrier competitive inhibition from the pre-loaded unlabelled CQ. In the latter model, labeled and unlabelled chloroquine should be on the same side of the membrane when they interact with the carrier, i.e. they are mixed together. In order to verify this hypothesis, Bray et al  incubated CQR lines with premixed chloroquine (labeled and unlabelled), but did not observe trans-stimulation of chloroquine uptake, thus suggesting that labeled and unlabelled chloroquine must be on opposite sides of the membrane for the trans-stimulation effect to take place, i.e. transport of unlabelled chloroquine via the carrier would be in the outward direction while labeled chloroquine transport would occur in the inward direction; in other words, mutant PfCRT would act as a bidirectional carrier, which is not compatible with an active efflux pump. In particular, these authors conjecture that trans-stimulation results might also be explained in terms of a gated channel.
Many authors measured chloroquine efflux from CQS and CQR isolates , , ,  under different conditions: in presence and absence of glucose, with or without proton gradient uncoupling, with or without Verapamil, an L-type calcium channel blocker of the Phenylalkylamine class. The results of these experiments have been interpreted in different ways by different authors and did not lead to a consensus view about the nature of PfCRT.
Sanchez et al  studied the kinetics of chloroquine efflux in ‘reverse varying-trans’ conditions  from CQR and CQS isolates. This procedure investigates whether extracellular unlabelled chloroquine would stimulate the release of pre-loaded [3H]-CQ. These authors expected that, in the presence of an active carrier, trans chloroquine should increase the initial efflux rate. They found an increasing initial efflux rate for both CQR and CQS lines and accordingly proposed that both CQR and CQS parasites possess a carrier of chloroquine with different transport properties.
It should appear clear from this survey that qualitative interpretations of the experimental findings are insufficient to draw conclusions about the nature of PfCRT and that more quantitative analyses are required.
A quantitative model cannot be derived from transient experiments because kinetic parameters, such as the rate of the vacuole pH equilibration during chloroquine uptake and the kinetic constants of chloroquine-hemozoine binding, are unknown and impossible to extract from the available data in an unambiguous way. The only data that can be used to derive a quantitative model without making an unreasonable number of hypotheses on the unknown parameters are those measured at equilibrium.
As we will show, the analytical model that we developed and used here indicates that equilibrium data are compatible with both the carrier and channel model for PfCRT, which explains why they could be interpreted differently by different authors. On the other hand, the carrier and channel hypotheses are only compatible with specific assumptions on the protonation state of the transported species and of the species binding to haeme of haeme-related molecules in the vacuole. For example, a carrier model is only compatible with the data if the transported molecule is protonated.
Another route to understand the nature of the macromolecule is to study its evolutionary relationship with other proteins of known function. Also in this case, data interpretation is controversial. Previous computational analyses of PfCRT ,  suggested that PfCRT belongs to the drug/metabolite trasporter (DMT) superfamily, whereas other studies proposed that it resembles to ClC chloride channels .
Here, we use state-of-the-art bioinformatics tools to identify PfCRT homology relationships and provide evidence that it is indeed a member of the DMT superfamily This finding is also supported by the observation that a three-dimensional model of the protein based on a DMT-like fold is consistent with experimental data about the mutations involved in insurgence and reversion of CQ resistance in Plasmodium.
By combining this latter conclusion with the results of the analytical method, we propose that PfCRT is a carrier of CQ+, CQ++ or both and that either all chloroquine species or only the uncharged one can bind hame of hame related species inside the vacuole.
We selected to use the data from the Cellular Accumulation Ratio (CAR) experiments described by  because, as mentioned above, the experimental conditions (stable and controlled pH values) allow the model to be built using a number of parameters comparable with the number of observations.
Our approach consists in testing the consistency of all plausible hypotheses about the binding mode of the drug to the heme related species inside the vacuole and the mechanism of action of PfCRT with experimental data. Our reasoning does not require any assumption on which heme form or heme related molecule binds to chloroquine and, for this reason, we refer to the heme related species bound by chloroquine inside the vacuole as HM.
Two mechanisms are thought to be involved in chloroquine accumulation into the P. falciparum vacuole: acidic trapping due to low vacuolar pH and chloroquine binding to heme or heme related species. It is reasonable to assume that PfCRT does not directly affect the molecular mechanism of chloroquine-HM binding. Recent studies , ,  indicate that the vacuolar pH of CQS and CQR strains are similar, hence the reduction of chloroquine accumulation in resistant strains cannot be explained in terms of different acidic trapping and PfCRT must be directly involved in releasing chloroquine out of the vacuole. A proof of this hypothesis has been recently provided by Martin and collaborators .
The analytical approach
In this study we formulate an analytical model describing different combinations of the two sets of hypotheses described below, the first related to the mode of chloroquine binding to HM, the second to the mechanism of action of the mutated PfCRT. We test all possible combinations of these hypotheses for their consistency with the available experimental data provided by  and summarized in Table 1. Abbreviations used throughout the manuscript are reported in Table 2.
Concerning the chloroquine binding to HM we consider two possibilities:
H1) The concentration of the complex chloroquine:HM inside the vacuole, [CQ∶HM]DV, linearly increases with the concentration of the binding form of chloroquine in the vacuole.
According to which chloroquine species reacts with HM, this can be expressed as [CQ∶HM]DV = α [CQ]DV or [CQ∶HM]DV = α [CQ+]DV, or [CQ∶HM]DV = α [CQ++]DV or [CQ∶HM]DV = α [CQTOT]DV, where [CQTOT] = [CQ]+[CQ+]+[CQ++]. The latter representing the case of chloroquine binding HM regardless of its protonation state.
H2) The concentration of the complex chloroquine:HM increases non linearly with the concentration of the binding chloroquine form in the vacuole and reaches saturation at concentrations of chloroquine above a given threshold.
If we define the threshold concentrations as tHM,CQ, tHM,CQ+, tHM,CQ++ or tHM,CQtot or simply tHM when the species is clear from the context, the above hypothesis can be expressed as: [CQ∶HM]DV = f([CQ]DV) that, for CQDV>tHM,CQ, reads [CQ∶HM]DV = constant. Similar expressions hold for [CQ∶HM]DV = f([CQ+]DV), [CQ∶HM]DV = f([CQ++]DV) and [CQ∶HM]DV = f([CQTOT]DV).
As far as PfCRT is concerned, the two possible cases are:
J1) PfCRT acts as a passive channel for CQ+ (or CQ++, or both).
and thereby the outward flux of chloroquine across the vacuole membrane due to PfCRT, JPfCRT, only depends upon the difference in concentration on the two sides of the vacuolar membrane and on the membrane potential. For instance, if the membrane potential is zero and the channel allows CQ+ to move out of the DV, JPfCRT = f([CQ+]DV−[CQ+]e), the suffix “e” indicating the plasmodium cytoplasm. As shown in the Text S1, section S1, being the vacuole membrane freely permeable to un-protonated chloroquine CQ, the hypothesis that PfCRT acts as a channel for CQ alone can be immediately discarded since it is not consistent with the experimentally observed differences between CQR and CQS.
J2) PfCRT is a carrier for CQ, (or CQ+, or CQ++).
In this case, it is reasonable to assume that the flux through the channel is a linear function of the CQ (or CQ+, or CQ++) concentration for concentration values below a given threshold and a constant above the threshold concentration. Let us define the threshold concentrations as tPfCRT,CQ, tPfCRT,CQ+, or tPfCRT,CQ++ or simply tPfCRT when the species is clear from the context. The above hypotheses can be expressed as:similar expressions hold for JPfCRT = f([CQ+]DV), JPfCRT = f([CQ++])DV.
The procedure adopted to test all combinations of the above hypotheses for their consistency with the available experimental data makes use of the analytical expressions for the membrane equilibrium, for the Cellular Accumulation Ratio (CAR) of chloroquine, which is the quantity measured in the experiments described in  and for the basis dissociation equilibrium.
The membrane equilibrium equation is:(1)where Pcq is the membrane permeability to unprotonated chloroquine; equation (1) has been obtained taking into account that the system has reached a steady state in the analysed experimental conditions. Consequently, the net chloroquine flux across the membranes is zero. The only form of chloroquine for which the erythrocyte and the external plasmodium membranes are permeable is the un-protonated one and this implies that CQ concentrations are the same on the two sides of these membranes.
The expression for CAR is:(2)where the total concentration of chloroquine inside the vacuole is:(3)being [CQ∶HM]DV the concentration of chloroquine bound to HM.(4)is the concentration of chloroquine outside the vacuole, [C]in is the average chloroquine concentration in the infected erythrocyte, Vin is the volume of the infected erythrocyte and VDV is the volume of the digestive vacuole. The analytical derivation of eqs (1) and (2) is detailed in Text S1, sections S2 and S3, respectively. The proof of the first equality of eq (4) ([C]e = [C]out) is reported in Text S1, section S4.
Finally, the relationships between the concentrations of the three forms of chloroquine given by the two-base dissociation equilibrium are:(5)(6)with 1/k′ = (1/k2+1/k2) and k″ = k1k2 and k1 and k2 are the two dissociation constants.
Experimentally determined values for the constants used in Eqs (2) to (6) are reported in Table 3.
Figure 1 summarizes the various hypotheses tested here. The columns refer to the mode of binding of the chloroquine to HM, assuming either that the experimental data have been obtained in non saturation conditions (linear regime, hypothesis H1) or that the concentration of the chloroquine species is above the (unknown) saturation concentration for HM (saturation conditions, hypothesis H2). The rows are related to assumptions about the mutated PfCRT function considering the possibility that the latter is a channel (first row, hypothesis J1) or an active carrier (remaining rows, hypotheses J2) and, in the case of active carrier, considering the possibility that the considered chloroquine species concentration into the vacuole is or is not above the (unknown) threshold needed to saturate PfCRT.
The columns refer to the mode of binding of the chloroquine to HM, the rows to mutated PfCRT, considering the possibility that the latter is a channel or an active carrier. Shaded cells correspond to combination of hypotheses inconsistent with the analysed data.
In the following we will show the procedure used to test the hypotheses summarized in Figure 1. We use the symbol CQ* in equations that hold for all four forms of chloroquine (CQ, CQ+, CQ++ and CQTOT).
We examined the consequences of each of the hypotheses reported in Figure 1. As a general strategy, we used the values of CAR in experiments A, B and C reported in Table 1 (CARA, CARB and CARC, respectively) to calibrate the model, inferred a CAR value for experiment D (CARD) and compared it with the experimental one. In some cases, the value of CARA and CARB was used to infer the value of the vacuolar pH (pHDV) subsequently compared with the experimental one. When the results show that a hypothesis is inconsistent with the experimental data and has to be discarded, the corresponding cell in Figure 1 is shaded. Inferred values of CARD and their agreement or disagreement with experimental values are shown in Figure 2B.
A) Computed values for the pH of the vacuole in the hypothesis that the concentration of the complex of HM with the indicated CQ species linearly increases with the concentration of the ligand (first four columns) or that the system is at saturation (last column). ALL refers to either CQ, CQ+, CQ++, CQTOT. Orange shaded area indicates the range of experimental values and their uncertainty measured by different authors in different experiments. As described in the text, the calculation only uses the CAR for sensitive strains (experiments A and B in Table 1) and therefore does not require any hypothesis on the nature of the mutated PfCRT. B) Computed values of the CAR compared with observed values obtained in experiment D (see Table 1). The different hypotheses are shown in the table at the bottom. The first row refers to the cells of the table in Figure 1. The second row reports the tested hypotheses on the mechanism of PfCRT (channel or carrier), on whether the binding of PfCRT with the indicated species is in the linear regime (linear) or at saturation (saturated) and on the protonation state of the transported molecule (CQ: neutral; CQ+: mono-protonated; CQ++: di-protonated). The third row refers to the tested hypotheses for the binding of the listed chloroquine species (CQ, CQ+, CQ++ or CQTOT; ALL refers to either CQ, CQ+, CQ++, CQTOT.) with HM in the hypotheses that the latter complex is at saturation or in the linear regime. Grey shaded cells indicate that the computed values are incompatible with the experimental ones. The orange shaded area indicates the range of experimental values and their experimental error. Dashed arrows indicate that the model provides an upper limit for the value. Question marks indicate that no conclusion can be derived using the indicated combination of hypotheses.
Test of hypothesis H1 ([CQ∶HM]DV = α [CQ*]DV).
Here, we use the values of CARA and CARB to infer the pHDV value. Notice that, since CARA and CARB refer to CQS strains, here we are not making any assumption on PfCRT.
Let us consider the hypothesis that the concentration of the complex CQ*∶HM linearly increases with the vacuolar concentration of CQ*, i.e. hypothesis H1 corresponding to the twenty cells C1–5,5–8 in Figure 1 in the columns labeled as “Linear regime”. We can have four different expressions for the total chloroquine accumulated in the vacuole [C]DV depending on which chloroquine form reacts with HM, namely(7)(8)(9)(10)corresponding to columns 5, 6, 7 and 8 of Figure 1, respectively.
Combining these equations with (2), (3), (4), (5), and (6) an expression for CAR as a function of the binding constant α and of pHDV can be obtained. For instance, in case 1 (eq. (7)), we have:(11)where Ve = Vin−VDV.
Let us consider the behavior of CQS strains. In this case, [CQ]DV = [CQ]e ((eq-S3) Text S1, section S2) and, in the presence of FCCP, [H+]DV is equal to [H+]e. Hence, by substituting the known value of CAR for sensitive strains (experiment A in Table 1), we can compute the parameter α. Knowing α and using the experimental value of CAR in the absence of FCCP (experiment B in Table 1) we can solve the second order equation (11) in pHDV, which is found to have only one positive solution. The same procedure can be applied to equations (8–10) to derive the equivalent of (11) for cases 2–4 (see also Text S1, section S5). The results for the four cases are:
Only the pHDV value obtained in case 1 is compatible with the experimental data for pHDV (Figure 2A and Table 3) and therefore we can conclude that, if the data are obtained in non saturating conditions for HM binding, the non protonated form of chloroquine binds to HM, regardless of the mechanism of action of the mutated PfCRT. The C1–5,6–8 cells in Figure 1 are therefore shaded since they correspond to cases not compatible with the experimental data.
Test of hypotheses J1 (PfCRT is a channel) and J2 (PfCRT is a carrier) [cells C1–5,5 Figure 1 and Figure 2B].
Let us now consider the data obtained for resistant strains (C and D in Table 1). We can use the values of pHDV and α to calculate the concentration of the un-protonated chloroquine inside the vacuole for CQR strains in the absence of FCCP (experiment C Table 1) using equation (11). We call this value [CQ]DV,C to indicate that it is related to experiment C and use eq (1) to calculate the outward flux due to PfCRT, i.e.obtaining and .
Next, we examine the specific hypotheses for PfCRT.
If PfCRT is a channel and the DV membrane potential is zero, the transport is only driven by the difference in concentration of the transported chloroquine form.
As a paradigm of our procedure, we describe the case where the transported form is CQ+. A similar reasoning can be applied if the transported species is CQ++ or a combination of CQ+ and CQ++.
We have(12)The function f(x) is increasing and it is zero when its argument is zero. The presence of FCCP in experiment D (Table 1) implies that pHDV = pHe, i.e. that the the concentrations of the three chloroquine forms are the same inside and outside the vacuole. Combining this observation with eq. (12) and the membrane balance equation (eq. (1)), we obtain that [CQ]DV,D = [CQ]e, [CQ+]DV,D = [CQ+]e and [CQ++]DV,D = [CQ++]e, i.e. both the chloroquine diffusive flux through the vacuolar membrane and through PfCRT (eq (12)) are zero (for detailed calculations see Text S1, section S6). These findings imply that CARD = CARA, which is consistent with the experimental data (Figure 2B). Accordingly, cell C1,5 of Figure 1 and Figure 2B is not shaded.
Notice that, as a by-product of our calculations, the JPfCRT permeability can be explicitly calculated in the hypothesis that f(x) is a linear function of ([CQ+]DV−[CQ+]e) (see Text S1, section S7).
The argument used to demonstrate that CARD = CARA when the membrane potential is zero also holds if the membrane potential is different from zero. According to ,  we assume that the DV membrane potential is mainly due to a proton gradient; therefore, the presence of the proton un-coupler FCCP in experiments A and D (Table 1) lowers the membrane potential to a negligible value. On the other hand, a non zero membrane potential should only be taken into account when analyzing the results of experiments B and C, whose data are not used here to demonstrate that CARD = CARA. In other words, the presence of a non zero membrane potential, that would require the addition of a term in equation (12), does not affect experiments A and D. Consequently it would have no effect on eq (12) and on our whole reasoning.
For testing the hypothesis that PfCRT is a carrier, corresponding to cells C2–5,5 in Figure 1, we need to consider the case of the CQ* concentration inside the vacuole being lower (or not) than the threshold value tPfCRT, i.e. being insufficient (or sufficient) to saturate the carrier (hypotheses J2a and J2b).
In the first case, corresponding to cells C3–5,5 in Figure 1, the outward flux is a linear function of the concentration of the transported chloroquine form CQ*, i.e.(13)As shown in Text S1, section S8, the CARD values for each of the chloroquine forms can be computed from eqs. (1), (11) and (13). In particular:Corresponding to cells C3–5,5 in Figure 1, respectively. The only cases where the results are consistent with the data reported in Table 1 are that PfCRT is an active carrier for either CQ+ or CQ++ in a linear regime (Figure 2B). This implies that only cell C3,5 is shaded in Figure 1 and Figure 2B.
A similar line of reasoning allows the CARD value to be computed in the hypothesis that the CQ* concentration inside the vacuole is not lower than the threshold value tPfCRT. In this case (corresponding to cell C2,5 in Figure 1) we have(14)In particular, JPfCRT,C = JPfCRT.D where the suffix C and D refer to the experiments shown in Table 1. Eq. (14) and the CQR membrane balance equation (eq (1)) implies [CQ]DV,C = [CQ]DV,D. By substituting the values for [CQ]DV,D and α in equation (11) and remembering that pHDV = pHe in the presence of FCCP, we obtain CARD = 204. This value significantly differs from the experimental one indicating that hypothesis J2b is not compatible with the data (Figure 2B). Accordingly, the corresponding cell C2,5 in Figure 1 and Figure 2B is shaded.
Test of the hypothesis H2 ([CQ∶HM]DV = constant).
We first use the values of CARA and CARB to infer the pHDV value and verify its consistency with the experimentally known one. Notice that, since CARA and CARB refer to CQS strains, we are not initially making assumptions on PfCRT.
Let us test the possibility that [CQ*] is above the concentration tHM,CQ needed to saturate the HM binding sites in the experiment with sensitive strains in the presence of FCCP (Experiment A in Table 1). As a first step, we need to understand whether this is the case also for the experiment performed on sensitive strains in the absence of FCCP (Experiment B in Table 1).
Because of the presence of FCCP in experiment A, we have:and, consequently, [CQ]DV,A = [CQ]DV,B, [CQ+]DV,A<[CQ+]DV,B and [CQ++]DV,A<[CQ++]DV,B, which implies that, in B, the concentration of CQ* is either the same as in A or higher and, therefore, [CQ∶HM]DV,B must be in the saturation regime, where [CQ∶HM]DV = constant, as well. It follows that:Similarly to the linear regime hypothesis for HM binding (hypothesis H1), it is possible to use equations (2), (3), and (4) to rewrite the expression for CAR as follows:(15)where [CQTOT]x (with x being either DV or e) is the sum of the three free chloroquine forms, namely [CQTOT]x = [CQ]x+[CQ+]x+[CQ++]x = [CQ]x (1+[H+]x/k′+[H+]2x/k″). Note that equation (15) is a general expression for CAR that, in the hypothesis H1, reduces to (11). The value of [CQ∶HM]DV,A can now be calculated using the experimental value of CARA, while pHDV,B can be obtained from equation (15) (where the dependence on pHDV is embedded in the [CQTOT]DV term). The only positive solution admitted by the equation is pHDV,B = 5.36, which is consistent with the experimental data (Figure 2B and Table 3). This finding, obtained without making any assumption on the nature of PfCRT, implies that hypothesis H2 is plausible, regardless of the chloroquine species involved in the chloroquine:HM binding and of the mechanism of action of the mutated PfCRT. Consequently, none of the 20 cells C1–5,1–4 can be shaded at this stage.
Test of hypotheses J1 (PfCRT is a channel) and J2 (PfCRT is a carrier) [cells C1–5,1–4 Figure 1].
As shown before, if experiment A has been performed in saturation conditions, the same is true for experiment B. We now need to infer what are the conditions of experiment C, i.e. whether(16)Notice that in experiment A there is no effect of PfCRT (being this the case of a sensitive strain) or of pH differences (because of the presence of FCCP), therefore all the chloroquine accumulation is due to the chloroquine bound to HM. In particular, it is apparent from eq. (15) that, without the [CQ∶HM]DV contribution, CARA would be 1. Equation (16) implies that the [CQ∶HM]DV contribution to CARC is equal to the [CQ∶HM]DV contribution to CARA hence, due to the additional contribution of pH to chloroquine accumulation in C, we would have CARC>CARA. Since the experimental data show that CARC<CARA, we can reject the hypothesis represented by equation (16) and conclude that the concentration of chloroquine in experiment C is not in the saturation region (detailed numeric calculation can be found in Text S1, section S9). In conclusion, if experiment A was performed at saturating chloroquine:HM concentration, the same holds for experiment B, but not for experiment C. This implies that, in experiment C, we have [CQ∶HM]DV = f(CQ*), where f(CQ*) is a monotonic increasing function of one of the chloroquine species.
In the hypothesis that PfCRT is a channel, the reasoning is identical to that of section 1.1.1 for cell C1,5 of Figure 1 and leads to the same conclusions: we cannot exclude that PfCRT is a channel (Figure 2B) therefore, cells C1,1–4 in Figure 1 are not shaded. Notice that no hypothesis on the chloroquine:HM binding is required in the reasoning of section 1.1.1
We now discuss the case when the transported form is CQ+, i.e. PfCRT is an active carrier that transports CQ+ in the linear regime (cell C4,1–4 Figure 1). A similar reasoning can be applied if the transported species are CQ or CQ++; the detailed calculations for these other cases are reported in Text S1 (sections S11 and S12).
In the hypothesis that PfCRT transports CQ+ and that the dependency of the JPfCRT flux is linear in the concentration of vacuolar CQ+ (JPfCRT = λ[CQ+]DV) we need to consider the following cases for chloroquine-HM binding:(17a)(17b)(17c)(17d)corresponding to cells C4,1–4 of Figure 1, respectively.
Equation (15) clearly shows that the difference between CARC and CARD is due to its second and third terms. Combining the expression for the PfCRT flux JPfCRT = λ([CQ+]DV) with that of the membrane equilibrium (eq. (1)) we have that [CQ]DV = [CQ]e/(1+[H+]DV λ/(k′ Pcq) ), and hence(18)that is, for any value of the unknown parameter λ, an increasing function of [H+]DV in the interval 0<pHDV<9.15, which is the interesting one for plasmodium metabolism (the proof is reported in Text S1, section S10). Being [H+]DV,C>[H+]DV,D, we have [CQTOT]DV,C>[CQTOT]DV,D; therefore the second member of equation (15) is larger for case C than for case D. As far as the third member of the equation is concerned, we have to distinguish the four cases (17a–d). Detailed calculations relating to each single hypothesis for the chloroquine:HM binding (cases (17a–d)) are reported in Text S1, section S13. Taking into account that [H+]DV,C>[H+]DV,D and that f(CQ*) is monotonic, we eventually obtain (Figure 2B):While in the first case no conclusion can be drawn on the relationship between CARC and CARD (indicated by a question mark in Figure 2B), cases (17b), (17c) and (17d) can be excluded and the corresponding cells C4,2, C4,3 and C4,4 can be shaded in Figure 1 and Figure 2B.
If JPfCRT = constant (cells C2,1–4 in Figure 1), this expression can be used in the vacuolar membrane balance equation for CQR strains (eq. (1)). Following the reasoning reported in Text S1, section S14, we obtain that CARD<CARC, which is inconsistent with the experimental data (Figure 2B). This implies that hypothesis H2 in conjunction with the hypothesis that PfCRT is a saturated chloroquine carrier is not plausible. Accordingly, cells C2,1–4 (Figure 1 and Figure 2B) are shaded.
We tested how stable our model is with respect to reasonable variations of the CAR values. We repeated the whole procedure using values of CAR derived for other parasite lines and for variation within the experimental error obtaining the same conclusions (data not shown).
In conclusion, ten of the forty cells of Figure 1, each corresponding to a different assumption about the mode of binding of chloroquine to HM in conjunction with a specific form of PfCRT, are not shaded, i.e. they correspond in principle to hypotheses consistent with the experimental data.
Sequence and structure analysis of PfCRT
Previous bioinformatics analyses of the PfCRT protein were devoted at identifying the functional role of the PfCRT protein, leading to different conclusions. On one hand several authors ,  assigned PfCRT to the drug/metabolite transporter (DMT) superfamily and, among the previously defined protein families, reported that PfCRT has the highest similarity with the drug/metabolite exporter (DME) family. On this basis, the authors concluded that PfCRT is likely to function as an exporter of metabolites in symport with H+.
Other studies proposed that PfCRT might share significant sequence similarity with the ClC chloride channels of other organisms , , which would reinforce the hypothesis that the protein acts as a gated aqueous pore.
As already mentioned by other authors , data supporting the ClC similarity hypothesis are not available and were impossible to reproduce using any available sensitive and updated sequence analysis tool. On the contrary, as detailed later, several different and effective methods strongly support the hypothesis that PfCRT is an active carrier.
The HHsearch tool ,  identifies three protein families in the Pfam database  with the highest similarity to PfCRT: PF06027 (DUF914, E-value 1.2×10−34), PF08449 (UAA transporters, E-value 5.6×10−24), PF04142 (nucleotide-sugar transporters, E-value 7.8×10−23). These all belong to the Drug/Metabolite Transporter clan CL0184, which consists of several families of different secondary carriers, among which drug, sugar and nucleotide antiporters have the highest similarity with PfCRT. The alignment of PfCRT homologous with the closest Pfam family (DUF914) is shown in Figure 3.
Profile-profile alignment of PfCRT homologs (see Materials and Methods) and the Pfam family DUF914, a member of the DMY clan. Residues are colored according to their physico-chemical properties (green = hydrophobic, red = negatively charged, blue = positively charged, pink = polar, white = small, yellow = cysteine, gray = proline, orange = histidine). The image was generated using the MPI Bioinformatics Toolkit .
Two amino acids of PfCRT are known to be involved in chloroquine resistance, 163 and 76, and therefore expected to be involved in the transport mechanism. To verify whether this is indeed the case, we built a homology model of the protein.
By searching the PDB database  with HHsearch, we retrieved two significant hits (e-value<0.1). The significant matches correspond to two solved structures (pdb codes 3B5D and 2I68) of the same protein, EmrE. This protein is a small E.coli multidrug transporter belonging to the same drug/metabolite transporter family described above that acts by exchanging various positively charged aromatic drugs across the plasma membrane with protons. EmrE is a homodimer, each monomer being composed by four membrane spanning helices. Two different regions of PfCRT sequence match three helices of each EmrE monomer, consistently with the presence of an internal sequence symmetry in PfCRT. We also searched for homologous proteins spanning the complete sequence of the protein to use as templates. Three different and sensitive methods (Shrimp , Phyre  and PROCAIN ) all identified the Glycerol-3-phosphate transporter GlpT from E.Coli (PDB code 1PW4) as a homolog of known structure with a significant sequence similarity spanning the whole sequence (e-value<10−4 in all cases). The protein was ranked first in the searches performed with Shrimp and PROCAIN and second by Phyre. The top ranking protein identified by Phyre (LacY) has the same fold and belongs to the same superfamily as the Glycerol-3-phosphate transporter.
GlpT was therefore selected as the template for the model. It is an active transporter presenting a 12-helix trans-membrane helix protein with an internal sequence symmetry.
Residues K76 (left side) and S163 (right side) are shown in red using a stick representation.
Mapping of residues 163 and 76 in the three-dimensional model of the protein (Figure 4) shows that they face each other and line the path of the transported molecule. This is consistent with the experimental observations and it appears plausible that the K76T and S163R mutations can permit positive charged species to be transported or not, respectively.
In conclusion, sequence analysis strongly suggests that PfCRT is an active carrier belonging to the Drug/Metabolite transporter superfamily and, furthermore, a three-dimensional model based on the basis of this evolutionary relationship is consistent with the experimental data on the protein.
The ongoing debate on PfCRT, the molecule responsible for chloroquine resistance in Plasmodium, has not yet provided a conclusive answer to the question of whether PfCRT is a channel or a carrier.
The qualitative analysis of chloroquine accumulation, trans-stimulation and chloroquine efflux data has been used to support both the channel and the carrier hypotheses by different authors , , , , . This is not surprising since we can analytically demonstrate that, at least as far as the choloroquine accumulation ratio experiments are concerned, the data are consistent with both hypotheses. A similarly rigorous approach cannot be used for the other experiment types since the problem is underdetermined.
This notwithstanding, we are able to show here that, if PfCRT is a carrier, it can only transport protonated or di-protonated chloroquine molecules and that choloroquine can either bind heme or heme related species in the digestive vacuole regardless of its charge or in its neutral form.
On the other hand, a reassessment of the evolutionary relationships of the protein with state-of-the-art methods and updated databases strongly suggests that the protein is indeed a member of the Drug/Metabolite transporter clan. This conclusion is further substantiated by the observation that a model based on this detected evolutionary relationship is consistent with experimental data.
Taking together the results of these two interdisciplinary approaches allow us to conclude that the chloroquine species are transported out of the vacuole through PfCRT in their protonated form, in agreement with studies such those presented by Lehane et al ,  who provided evidence that the presence of chloroquine increases the leak of H+ from the vacuole.
We would like to emphasize that our analytical approach does not require ad-hoc hypotheses as it would be the case if we were to model data coming from trans-stimulation and chloroquine efflux experiments where parameters such as the rate of the vacuole pH equilibration during chloroquine uptake or the kinetic constants of chloroquine-hemozoine binding are unknown. On the other hand, the model can be used effectively to interpret the results of stationary experiments. As an example, recently Martin and collaborators  set up a system in which PfCRT is expressed at the surface of Xenopus leavis oocytes and measured the chloroquine uptake which, in this system, is not influenced by chloroquine-HM binding. The available data have been obtained in the pre-stationary state, but the same system could in principle be used to measure chloroquine uptake at equilibrium in a pH-controlled experiment. In this case a simplified version of our analytical model could be employed to derive the systems parameters and provide more detailed information about PfCRT.
Remarkably, our analysis conclusively demonstrates that experimental data on the chloroquine accumulation ratio at equilibrium are consistent with both hypotheses that the mutant molecule is an active or a passive carrier of the drug and therefore insufficient to distinguish between the two mechanisms, while at the same time they can be used to restrict hypotheses on the nature of the transported and HM-binding species in the two cases.
Our computational analysis of the sequence and a structural bioinformatics approach strongly suggests that the mutated protein acts as an active carrier of chloroquine and, in this assumption, we can conclude that the PfCRT mutated protein confers resistance by carrying either the mono or di-protonated chloroquine out of the vacuole. It is tempting to speculate that the mechanism is that the mutated PfCRT uses the H+ gradient to expel the protonated form of chloroquine from the vacuole. Cationic transport inhibitors could be tested to further support this hypothesis and, perhaps, as starting point for developing novel therapies against resistant malaria strains.
We hope that the example of the power of systems and computational biology analysis of the data presented here will convince the Plasmodium community to take advantage of our results.
Materials and Methods
Ethics Statement: N/A
We performed sequence similarity searches on the nr database  release of October 4th, 2010 using three iterations of CS-Blast . We selected the first 500 hits, all having an e-value lower than 1e-5, and realigned their sequences using the multiple sequence alignment tool MUSCLE . Such alignment was then used to scan the Pfam  and PDB  databases on October 5th, 2010 with the Hidden Markov Model based search method HHpred , the results of which are reported in Table S1. Three distant homology recognition tools, Phyre , Shrimp  and PROCAIN , were used to identify templates spanning the complete PfCRT sequence (see Table S2, Table S3 and S4). In all the cases, a statistically significant similarity of the PfCRT sequence with that of the Glycerol-3-Phosphate transporter GlpT from E.Coli (PDB code 1PW4) spanning the whole sequence of both target and template proteins was detected.
Details of the calculations presented in the main manuscript.
(1.24 MB PDF)
HHpred results on Pfam and PDB databases.
(0.02 MB PDF)
Conceived and designed the experiments: MC AV PM AT. Analyzed the data: MC AV PM. Wrote the paper: AV AT.
- 1. Wellems TE, Miller LH (2003) Two worlds of malaria. N Engl J Med 349: 1496–1498.TE WellemsLH Miller2003Two worlds of malaria.N Engl J Med34914961498
- 2. Slater AF, Swiggard WJ, Orton BR, Flitter WD, Goldberg DE, et al. (1991) An iron-carboxylate bond links the heme units of malaria pigment. Proc Natl Acad Sci U S A 88: 325–329.AF SlaterWJ SwiggardBR OrtonWD FlitterDE Goldberg1991An iron-carboxylate bond links the heme units of malaria pigment.Proc Natl Acad Sci U S A88325329
- 3. Homewood CA, Warhurst DC, Peters W, Baggaley VC (1972) Lysosomes, pH and the anti-malarial action of chloroquine. Nature 235: 50–52.CA HomewoodDC WarhurstW. PetersVC Baggaley1972Lysosomes, pH and the anti-malarial action of chloroquine.Nature2355052
- 4. Yayon A, Cabantchik ZI, Ginsburg H (1984) Identification of the acidic compartment of Plasmodium falciparum-infected human erythrocytes as the target of the antimalarial drug chloroquine. EMBO J 3: 2695–2700.A. YayonZI CabantchikH. Ginsburg1984Identification of the acidic compartment of Plasmodium falciparum-infected human erythrocytes as the target of the antimalarial drug chloroquine.EMBO J326952700
- 5. Bray PG, Janneh O, Raynes KJ, Mungthin M, Ginsburg H, et al. (1999) Cellular uptake of chloroquine is dependent on binding to ferriprotoporphyrin IX and is independent of NHE activity in Plasmodium falciparum. J Cell Biol 145: 363–376.PG BrayO. JannehKJ RaynesM. MungthinH. Ginsburg1999Cellular uptake of chloroquine is dependent on binding to ferriprotoporphyrin IX and is independent of NHE activity in Plasmodium falciparum.J Cell Biol145363376
- 6. Bray PG, Mungthin M, Ridley RG, Ward SA (1998) Access to hematin: the basis of chloroquine resistance. Mol Pharmacol 54: 170–179.PG BrayM. MungthinRG RidleySA Ward1998Access to hematin: the basis of chloroquine resistance.Mol Pharmacol54170179
- 7. Pagola S, Stephens PW, Bohle DS, Kosar AD, Madsen SK (2000) The structure of malaria pigment beta-haematin. Nature 404: 307–310.S. PagolaPW StephensDS BohleAD KosarSK Madsen2000The structure of malaria pigment beta-haematin.Nature404307310
- 8. Slater AF (1993) Chloroquine: mechanism of drug action and resistance in Plasmodium falciparum. Pharmacol Ther 57: 203–235.AF Slater1993Chloroquine: mechanism of drug action and resistance in Plasmodium falciparum.Pharmacol Ther57203235
- 9. Sullivan DJ Jr, Matile H, Ridley RG, Goldberg DE (1998) A common mechanism for blockade of heme polymerization by antimalarial quinolines. J Biol Chem 273: 31103–31107.DJ Sullivan JrH. MatileRG RidleyDE Goldberg1998A common mechanism for blockade of heme polymerization by antimalarial quinolines.J Biol Chem2733110331107
- 10. Sullivan DJ Jr, Gluzman IY, Russell DG, Goldberg DE (1996) On the molecular mechanism of chloroquine's antimalarial action. Proc Natl Acad Sci U S A 93: 11865–11870.DJ Sullivan JrIY GluzmanDG RussellDE Goldberg1996On the molecular mechanism of chloroquine's antimalarial action.Proc Natl Acad Sci U S A931186511870
- 11. Sugioka K, Shimosegawa Y, Nakano M (1987) Estrogens as natural antioxidants of membrane phospholipid peroxidation. FEBS Lett 210: 37–39.K. SugiokaY. ShimosegawaM. Nakano1987Estrogens as natural antioxidants of membrane phospholipid peroxidation.FEBS Lett2103739
- 12. Saliba KJ, Folb PI, Smith PJ (1998) Role for the plasmodium falciparum digestive vacuole in chloroquine resistance. Biochem Pharmacol 56: 313–320.KJ SalibaPI FolbPJ Smith1998Role for the plasmodium falciparum digestive vacuole in chloroquine resistance.Biochem Pharmacol56313320
- 13. Fitch CD (1970) Plasmodium falciparum in owl monkeys: drug resistance and chloroquine binding capacity. Science 169: 289–290.CD Fitch1970Plasmodium falciparum in owl monkeys: drug resistance and chloroquine binding capacity.Science169289290
- 14. Bray PG, Martin RE, Tilley L, Ward SA, Kirk K, et al. (2005) Defining the role of PfCRT in Plasmodium falciparum chloroquine resistance. Mol Microbiol 56: 323–333.PG BrayRE MartinL. TilleySA WardK. Kirk2005Defining the role of PfCRT in Plasmodium falciparum chloroquine resistance.Mol Microbiol56323333
- 15. Wellems TE, Plowe CV (2001) Chloroquine-resistant malaria. J Infect Dis 184: 770–776.TE WellemsCV Plowe2001Chloroquine-resistant malaria.J Infect Dis184770776
- 16. Martin RE, Kirk K (2004) The malaria parasite's chloroquine resistance transporter is a member of the drug/metabolite transporter superfamily. Mol Biol Evol 21: 1938–1949.RE MartinK. Kirk2004The malaria parasite's chloroquine resistance transporter is a member of the drug/metabolite transporter superfamily.Mol Biol Evol2119381949
- 17. Cooper RA, Lane KD, Deng B, Mu J, Patel JJ, et al. (2007) Mutations in transmembrane domains 1, 4 and 9 of the Plasmodium falciparum chloroquine resistance transporter alter susceptibility to chloroquine, quinine and quinidine. Mol Microbiol 63: 270–282.RA CooperKD LaneB. DengJ. MuJJ Patel2007Mutations in transmembrane domains 1, 4 and 9 of the Plasmodium falciparum chloroquine resistance transporter alter susceptibility to chloroquine, quinine and quinidine.Mol Microbiol63270282
- 18. Huaman MC, Yoshinaga K, Suryanatha A, Suarsana N, Kanbara H (2004) Short report: polymorphisms in the chloroquine resistance transporter gene in Plasmodium falciparum isolates from Lombok, Indonesia. Am J Trop Med Hyg 71: 40–42.MC HuamanK. YoshinagaA. SuryanathaN. SuarsanaH. Kanbara2004Short report: polymorphisms in the chloroquine resistance transporter gene in Plasmodium falciparum isolates from Lombok, Indonesia.Am J Trop Med Hyg714042
- 19. Cooper RA, Hartwig CL, Ferdig MT (2005) pfcrt is more than the Plasmodium falciparum chloroquine resistance gene: a functional and evolutionary perspective. Acta Trop 94: 170–180.RA CooperCL HartwigMT Ferdig2005pfcrt is more than the Plasmodium falciparum chloroquine resistance gene: a functional and evolutionary perspective.Acta Trop94170180
- 20. Fidock DA, Nomura T, Talley AK, Cooper RA, Dzekunov SM, et al. (2000) Mutations in the P. falciparum digestive vacuole transmembrane protein PfCRT and evidence for their role in chloroquine resistance. Mol Cell 6: 861–871.DA FidockT. NomuraAK TalleyRA CooperSM Dzekunov2000Mutations in the P. falciparum digestive vacuole transmembrane protein PfCRT and evidence for their role in chloroquine resistance.Mol Cell6861871
- 21. Cooper RA, Ferdig MT, Su XZ, Ursos LM, Mu J, et al. (2002) Alternative mutations at position 76 of the vacuolar transmembrane protein PfCRT are associated with chloroquine resistance and unique stereospecific quinine and quinidine responses in Plasmodium falciparum. Mol Pharmacol 61: 35–42.RA CooperMT FerdigXZ SuLM UrsosJ. Mu2002Alternative mutations at position 76 of the vacuolar transmembrane protein PfCRT are associated with chloroquine resistance and unique stereospecific quinine and quinidine responses in Plasmodium falciparum.Mol Pharmacol613542
- 22. Johnson DJ, Fidock DA, Mungthin M, Lakshmanan V, Sidhu AB, et al. (2004) Evidence for a central role for PfCRT in conferring Plasmodium falciparum resistance to diverse antimalarial agents. Mol Cell 15: 867–877.DJ JohnsonDA FidockM. MungthinV. LakshmananAB Sidhu2004Evidence for a central role for PfCRT in conferring Plasmodium falciparum resistance to diverse antimalarial agents.Mol Cell15867877
- 23. Wellems TE (2004) Transporter of a malaria catastrophe. Nat Med 10: 1169–1171.TE Wellems2004Transporter of a malaria catastrophe.Nat Med1011691171
- 24. Lehane AM, Kirk K (2008) Chloroquine resistance-conferring mutations in pfcrt give rise to a chloroquine-associated H+ leak from the malaria parasite's digestive vacuole. Antimicrob Agents Chemother 52: 4374–4380.AM LehaneK. Kirk2008Chloroquine resistance-conferring mutations in pfcrt give rise to a chloroquine-associated H+ leak from the malaria parasite's digestive vacuole.Antimicrob Agents Chemother5243744380
- 25. Martin RE, Marchetti RV, Cowan AI, Howitt SM, Broer S, et al. (2009) Chloroquine transport via the malaria parasite's chloroquine resistance transporter. Science 325: 1680–1682.RE MartinRV MarchettiAI CowanSM HowittS. Broer2009Chloroquine transport via the malaria parasite's chloroquine resistance transporter.Science32516801682
- 26. Sanchez CP, Stein WD, Lanzer M (2007) Is PfCRT a channel or a carrier? Two competing models explaining chloroquine resistance in Plasmodium falciparum. Trends Parasitol 23: 332–339.CP SanchezWD SteinM. Lanzer2007Is PfCRT a channel or a carrier? Two competing models explaining chloroquine resistance in Plasmodium falciparum.Trends Parasitol23332339
- 27. Sanchez CP, Stein W, Lanzer M (2003) Trans stimulation provides evidence for a drug efflux carrier as the mechanism of chloroquine resistance in Plasmodium falciparum. Biochemistry 42: 9383–9394.CP SanchezW. SteinM. Lanzer2003Trans stimulation provides evidence for a drug efflux carrier as the mechanism of chloroquine resistance in Plasmodium falciparum.Biochemistry4293839394
- 28. Sanchez CP, McLean JE, Stein W, Lanzer M (2004) Evidence for a substrate specific and inhibitable drug efflux system in chloroquine resistant Plasmodium falciparum strains. Biochemistry 43: 16365–16373.CP SanchezJE McLeanW. SteinM. Lanzer2004Evidence for a substrate specific and inhibitable drug efflux system in chloroquine resistant Plasmodium falciparum strains.Biochemistry431636516373
- 29. Bray PG, Mungthin M, Hastings IM, Biagini GA, Saidu DK, et al. (2006) PfCRT and the trans-vacuolar proton electrochemical gradient: regulating the access of chloroquine to ferriprotoporphyrin IX. Mol Microbiol 62: 238–251.PG BrayM. MungthinIM HastingsGA BiaginiDK Saidu2006PfCRT and the trans-vacuolar proton electrochemical gradient: regulating the access of chloroquine to ferriprotoporphyrin IX.Mol Microbiol62238251
- 30. Sanchez CP, McLean JE, Rohrbach P, Fidock DA, Stein WD, et al. (2005) Evidence for a pfcrt-associated chloroquine efflux system in the human malarial parasite Plasmodium falciparum. Biochemistry 44: 9862–9870.CP SanchezJE McLeanP. RohrbachDA FidockWD Stein2005Evidence for a pfcrt-associated chloroquine efflux system in the human malarial parasite Plasmodium falciparum.Biochemistry4498629870
- 31. Stein WD (1986) Transport and diffusion across cell membranes. New York: Academic Press. WD Stein1986Transport and diffusion across cell membranesNew YorkAcademic Press
- 32. Bray PG, Howells RE, Ritchie GY, Ward SA (1992) Rapid chloroquine efflux phenotype in both chloroquine-sensitive and chloroquine-resistant Plasmodium falciparum. A correlation of chloroquine sensitivity with energy-dependent drug accumulation. Biochem Pharmacol 44: 1317–1324.PG BrayRE HowellsGY RitchieSA Ward1992Rapid chloroquine efflux phenotype in both chloroquine-sensitive and chloroquine-resistant Plasmodium falciparum. A correlation of chloroquine sensitivity with energy-dependent drug accumulation.Biochem Pharmacol4413171324
- 33. Krogstad DJ, Gluzman IY, Kyle DE, Oduola AM, Martin SK, et al. (1987) Efflux of chloroquine from Plasmodium falciparum: mechanism of chloroquine resistance. Science 238: 1283–1285.DJ KrogstadIY GluzmanDE KyleAM OduolaSK Martin1987Efflux of chloroquine from Plasmodium falciparum: mechanism of chloroquine resistance.Science23812831285
- 34. Sanchez CP, Rohrbach P, McLean JE, Fidock DA, Stein WD, et al. (2007) Differences in trans-stimulated chloroquine efflux kinetics are linked to PfCRT in Plasmodium falciparum. Mol Microbiol 64: 407–420.CP SanchezP. RohrbachJE McLeanDA FidockWD Stein2007Differences in trans-stimulated chloroquine efflux kinetics are linked to PfCRT in Plasmodium falciparum.Mol Microbiol64407420
- 35. Warhurst DC, Craig JC, Adagu IS (2002) Lysosomes and drug resistance in malaria. Lancet 360: 1527–1529.DC WarhurstJC CraigIS Adagu2002Lysosomes and drug resistance in malaria.Lancet36015271529
- 36. Hayward R, Saliba KJ, Kirk K (2006) The pH of the digestive vacuole of Plasmodium falciparum is not associated with chloroquine resistance. J Cell Sci 119: 1016–1025.R. HaywardKJ SalibaK. Kirk2006The pH of the digestive vacuole of Plasmodium falciparum is not associated with chloroquine resistance.J Cell Sci11910161025
- 37. Kuhn Y, Rohrbach P, Lanzer M (2007) Quantitative pH measurements in Plasmodium falciparum-infected erythrocytes using pHluorin. Cell Microbiol 9: 1004–1013.Y. KuhnP. RohrbachM. Lanzer2007Quantitative pH measurements in Plasmodium falciparum-infected erythrocytes using pHluorin.Cell Microbiol910041013
- 38. Klonis N, Tan O, Jackson K, Goldberg D, Klemba M, et al. (2007) Evaluation of pH during cytostomal endocytosis and vacuolar catabolism of haemoglobin in Plasmodium falciparum. Biochem J 407: 343–354.N. KlonisO. TanK. JacksonD. GoldbergM. Klemba2007Evaluation of pH during cytostomal endocytosis and vacuolar catabolism of haemoglobin in Plasmodium falciparum.Biochem J407343354
- 39. Hildebrand A, Remmert M, Biegert A, Soding J (2009) Fast and accurate automatic structure prediction with HHpred. Proteins 77: Suppl 9128–132.A. HildebrandM. RemmertA. BiegertJ. Soding2009Fast and accurate automatic structure prediction with HHpred.Proteins77Suppl 9128132
- 40. Soding J (2005) Protein homology detection by HMM-HMM comparison. Bioinformatics 21: 951–960.J. Soding2005Protein homology detection by HMM-HMM comparison.Bioinformatics21951960
- 41. Sonnhammer EL, Eddy SR, Durbin R (1997) Pfam: a comprehensive database of protein domain families based on seed alignments. Proteins 28: 405–420.EL SonnhammerSR EddyR. Durbin1997Pfam: a comprehensive database of protein domain families based on seed alignments.Proteins28405420
- 42. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, et al. (2000) The Protein Data Bank. Nucleic Acids Res 28: 235–242.HM BermanJ. WestbrookZ. FengG. GillilandTN Bhat2000The Protein Data Bank.Nucleic Acids Res28235242
- 43. Bernsel A, Viklund H, Elofsson A (2008) Remote homology detection of integral membrane proteins using conserved sequence features. Proteins 71: 1387–1399.A. BernselH. ViklundA. Elofsson2008Remote homology detection of integral membrane proteins using conserved sequence features.Proteins7113871399
- 44. Kelley LA, Sternberg MJ (2009) Protein structure prediction on the Web: a case study using the Phyre server. Nat Protoc 4: 363–371.LA KelleyMJ Sternberg2009Protein structure prediction on the Web: a case study using the Phyre server.Nat Protoc4363371
- 45. Wang Y, Sadreyev RI, Grishin NV (2009) PROCAIN server for remote protein sequence similarity search. Bioinformatics 25: 2076–2077.Y. WangRI SadreyevNV Grishin2009PROCAIN server for remote protein sequence similarity search.Bioinformatics2520762077
- 46. Castrignano T, De Meo PD, Cozzetto D, Talamo IG, Tramontano A (2006) The PMDB Protein Model Database. Nucleic Acids Res 34: D306–309.T. CastrignanoPD De MeoD. CozzettoIG TalamoA. Tramontano2006The PMDB Protein Model Database.Nucleic Acids Res34D306309
- 47. Lehane AM, Hayward R, Saliba KJ, Kirk K (2008) A verapamil-sensitive chloroquine-associated H+ leak from the digestive vacuole in chloroquine-resistant malaria parasites. J Cell Sci 121: 1624–1632.AM LehaneR. HaywardKJ SalibaK. Kirk2008A verapamil-sensitive chloroquine-associated H+ leak from the digestive vacuole in chloroquine-resistant malaria parasites.J Cell Sci12116241632
- 48. Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Sayers EW (2009) GenBank. Nucleic Acids Res 37: D26–31.DA BensonI. Karsch-MizrachiDJ LipmanJ. OstellEW Sayers2009GenBank.Nucleic Acids Res37D2631
- 49. Biegert A, Soding J (2009) Sequence context-specific profiles for homology searching. Proc Natl Acad Sci U S A 106: 3770–3775.A. BiegertJ. Soding2009Sequence context-specific profiles for homology searching.Proc Natl Acad Sci U S A10637703775
- 50. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32: 1792–1797.RC Edgar2004MUSCLE: multiple sequence alignment with high accuracy and high throughput.Nucleic Acids Res3217921797
- 51. Bennett-Lovsey RM, Herbert AD, Sternberg MJ, Kelley LA (2008) Exploring the extremes of sequence/structure space with ensemble fold recognition in the program Phyre. Proteins 70: 611–625.RM Bennett-LovseyAD HerbertMJ SternbergLA Kelley2008Exploring the extremes of sequence/structure space with ensemble fold recognition in the program Phyre.Proteins70611625
- 52. Biegert A, Mayer C, Remmert M, Soding J, Lupas AN (2006) The MPI Bioinformatics Toolkit for protein sequence analysis. Nucleic Acids Res 34: W335–339.A. BiegertC. MayerM. RemmertJ. SodingAN Lupas2006The MPI Bioinformatics Toolkit for protein sequence analysis.Nucleic Acids Res34W335339
- 53. Perrin DD (1965) Dissociation Constants of Organic Bases in Aqueous Solution. London: Butterworth. DD Perrin1965Dissociation Constants of Organic Bases in Aqueous SolutionLondonButterworth
- 54. Ferrari V, Cutler DJ (1991) Kinetics and thermodynamics of chloroquine and hydroxychloroquine transport across the human erythrocyte membrane. Biochem Pharmacol 41: 23–30.V. FerrariDJ Cutler1991Kinetics and thermodynamics of chloroquine and hydroxychloroquine transport across the human erythrocyte membrane.Biochem Pharmacol412330
- 55. Ferrari V, Cutler DJ (1991) Simulation of kinetic data on the influx and efflux of chloroquine by erythrocytes infected with Plasmodium falciparum. Evidence for a drug-importer in chloroquine-sensitive strains. Biochem Pharmacol 42: SupplS167–179.V. FerrariDJ Cutler1991Simulation of kinetic data on the influx and efflux of chloroquine by erythrocytes infected with Plasmodium falciparum. Evidence for a drug-importer in chloroquine-sensitive strains.Biochem Pharmacol42SupplS167179