Chloroquine Resistance-Conferring Mutations in pfcrt Give Rise to a Chloroquine-Associated H+ Leak from the Malaria Parasite's Digestive Vacuole

Chloroquine resistance in the malaria parasite Plasmodium falciparumis conferred by mutations in the P. falciparum chloroquine resistancetransporter (PfCRT). PfCRT localizes to the membrane of theparasite's internal digestive vacuole, an acidic organelle inwhich chloroquine accumulates to high concentrations and exertsits toxic effect. Mutations in PfCRT are thought to reduce chloroquineaccumulation in this organelle. How they do so is the subjectof ongoing debate. Recently we have shown that in the presenceof chloroquine there is an increased leak of H⁺ from the digestivevacuole in chloroquine-resistant but not chloroquine-sensitiveparasites. Here, using transfectant parasite strains of a singlegenetic background and differing only in their pfcrt allele,we show that chloroquine resistance-conferring PfCRT mutationsare responsible for this chloroquine-associated H⁺ leak. Thisis consistent with the hypothesis that the chloroquine resistance-conferringforms of PfCRT mediate the efflux of chloroquine, in associationwith H⁺, from the malaria parasite's digestive vacuole.

INTRODUCTION

Top

INTRODUCTION

For many years, chloroquine (CQ) was the mainstay of antimalarialchemotherapy. However, the emergence and spread of CQ-resistant(CQR) Plasmodium falciparum parasites has rendered this drugineffective in most regions where malaria is endemic. This hashad dire consequences for the global malaria situation (38).

The acidic digestive vacuole (DV) of the intraerythrocytic malariaparasite is thought to be the site of CQ toxicity. In this organelle,hemoglobin endocytosed from the host erythrocyte cytosol isdigested in a process that liberates monomeric heme (10). Hemehas the potential to be toxic to the parasite, but this is preventedby its incorporation into inert hemozoin crystals (27). UnchargedCQ is membrane permeant and can diffuse freely into the DV interior.However, once inside the acidic DV, CQ becomes trapped in itsprotonated (mostly diprotonated) form (45). CQ binds to hemeand prevents its incorporation into hemozoin. It is thoughtto be the consequent buildup of heme and CQ-heme complexes thatkills the parasite (8, 26).

CQ resistance is associated with a significant reduction inthe amount of CQ accumulated by the parasite (9, 16, 46). Arange of compounds, including the weak base verapamil, increasethe amount of CQ accumulated by CQR parasites and thereby sensitizethe parasites to the drug (16, 23, 39). Given that most of theCQ accumulated by CQ-sensitive (CQS) parasites is within theDV (3) and that isolated DVs from CQR parasites accumulate lessCQ than those from CQS parasites (29), it is likely that a reductionin intravacuolar CQ accumulation is central to the phenomenonof CQ resistance.

The implementation and analysis of a genetic cross between aCQS and a CQR strain of P. falciparum (36, 42, 43) paved theway for the identification of PfCRT as the key determinant ofCQ resistance (7). The transfection of CQS parasites with mutantforms of pfcrt provided conclusive evidence that mutations inthis gene give rise to the verapamil-reversible CQ resistancephenotype (7, 34). PfCRT has been localized to the DV membrane(6, 7), and bioinformatic analyses have placed it in the "drug/metabolitetransporter" superfamily (22, 37).

A number of mutant pfcrt alleles, each possessing a minimumof four point mutations (5), have been identified in the field.Several other CQ resistance-conferring mutant pfcrt alleleshave been generated in the laboratory (6). Virtually all themutant pfcrt alleles found in CQR field isolates possess theK76T mutation, which is predicted to reside near the DV-facingsurface of the membrane in a region of the transporter involvedin substrate specificity (22). It has been shown that reversingthis one mutation in CQR strains restores full sensitivity toCQ (20) and that the mutations preceding residue 76 affect parasitesusceptibility to CQ resistance reversal by verapamil (20).

The precise mechanism by which mutations in PfCRT result ina reduced level of CQ accumulation remains to be established.An alteration in the pH of the DV (pH_DV), which may come aboutif PfCRT has a direct or indirect role in pH_DV regulation, wouldbe expected to affect the "weak base trapping" of CQ in theDV and, hence, parasite susceptibility to the drug (47). Whiletechnical difficulties associated with measuring pH_DV have ledto some controversy (4, 44), three recent independent studiesusing different approaches have reported there to be no significantdifference in pH_DV between CQS and CQR parasites (13, 15, 19).

An alternative hypothesis is that the mutant PfCRT proteinsin CQR parasites mediate CQ efflux from the DV. CQ transportstudies, carried out using intact parasitized erythrocytes,have yielded evidence for the enhanced efflux of CQ from CQRparasites (3, 16, 32). Some have interpreted their findingsin terms of an energy-dependent CQ carrier (30-33), while othershave argued for a "charged drug leak" model, in which protonatedCQ is proposed to diffuse passively down its concentration gradientand out of the DV (3, 41).

The acidity of the malaria parasite's DV is maintained by inwardlydirected H⁺ pumps which counter the (outward) leak of H⁺. Ina recent study in which we investigated the leak of H⁺ fromthe DV by monitoring DV alkalinization following inhibitionof the H⁺-pumping V-type ATPase, we showed that the presenceof CQ gives rise to a substantial outward H⁺ leak from the DVin CQR but not CQS parasites (21). This CQ-associated H⁺ leakis inhibited by the CQ resistance reverser verapamil. Theseobservations provided evidence for the involvement of enhancedCQ:H⁺ efflux from the DV in the phenomenon of CQ resistance.Here, using three P. falciparum strains with the same geneticbackground and differing only in their pfcrt allele, we showthat the verapamil-sensitive CQ-associated H⁺ leak observedfrom the DV of CQR parasites is a consequence of the CQ resistance-conferringmutations in PfCRT. This is consistent with the hypothesis thatmutant PfCRT in CQR parasites mediates CQ efflux from the DV.We also show that mutations in PfCRT give rise to an increasein the rate of DV alkalinization seen following H⁺ pump inhibitionin the absence of CQ. Possible reasons for this and their implicationsare discussed.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

P. falciparum cultures. The pfcrt transfectant lines used in this study were generatedby Sidhu et al. (34) and kindly provided by D. A. Fidock (ColumbiaUniversity, New York). Clones C4_Dd2 and C6_7G8 were generatedby replacing the wild-type pfcrt allele of the CQS strain GC03with pfcrt alleles from the CQR lines Dd2 and 7G8, respectively.Strains Dd2 and 7G8 were isolated in Southeast Asia and SouthAmerica, respectively, and each possesses a different complementof pfcrt mutations. C2_GC03 is a CQS recombinant control thatcontains the two selectable marker cassettes found in C4_Dd2but (because of the location of the recombination events) retainsits wild-type pfcrt allele. The parasites were cultured andsynchronized as described elsewhere (21) and were maintainedin the presence of 5 nM WR99210 (Jacobus Pharmaceuticals, Princeton,NJ) and 5 µM blasticidin (Invitrogen, Australia). Parasitesto be used in experiments were maintained in the absence ofdrug selection for the 4 days immediately prior to the experiment(see below).

Loading of fluorescein-dextran into the DV of the parasite. To enable the monitoring of changes in pH_DV, parasite DVs wereloaded with the membrane-impermeant pH-sensitive dye fluorescein-dextran(pK_a, $~$ 6.4, $~$ 10 x 10³ M_r; Invitrogen, Australia). This was performedby loading uninfected erythrocytes with fluorescein-dextran(55 µM) as described previously (17, 28) and then inoculatingthem with trophozoite-infected erythrocytes. Typically, 1 to2 ml of packed fluorescein-dextran-loaded erythrocytes wereinoculated with an equal volume of infected erythrocyte culture( $~$ 20% parasitemia, 4% hematocrit) and then diluted in mediumto give a total volume of 50 ml. The cells were cultured daily(in the absence of WR99210 and blasticidin) for two completecycles before being used in experiments. At the time of experimentation,most of the trophozoites were growing within fluorescein-dextran-loadederythrocytes and contained fluorescein-dextran in their DVsas a result of endocytosis of the host erythrocyte cytosol.

It has been shown previously that the strain GC03 displays hypergametocytogenesiswhen grown in erythrocytes that are partially depleted of hemoglobin(1). Partial hemoglobin depletion is an unavoidable consequenceof the dye-loading procedure, and we observed that a subpopulationof the parasites (typically 2 to 6% for all three strains) haddifferentiated into morphologically identifiable gametocytesby the time of experimentation.

Fluorescence measurements. Chloroquine diphosphate and verapamil hydrochloride were purchasedfrom Sigma-Aldrich (Australia), and concanamycin A was purchasedfrom MP Biomedicals (Australia). Fluorometry experiments wereperformed as described previously (13, 21, 28) on fluorescein-dextran-loadedmature trophozoite-stage parasites that had been functionallyisolated from their host erythrocytes by brief exposure to 0.05%wt/vol saponin (yielding a 0.005% wt/vol solution of the activeagent sapogenin) under conditions shown not to affect the integrityof the parasite's plasma membrane or its ability to maintainlarge transmembrane ion gradients (35). The parasites were suspendedat a density of $~$ 10⁷ cells/ml in a minimal saline solution at37°C. The ratio of the fluorescence intensity measured at520 nm using two excitation wavelengths (490 nm and 450 nm)was used as an indicator of pH_DV. Microscopic examination atthe conclusion of a number of experiments revealed that theparasites remained intact, with the fluorescence localized exclusivelyto the DV. Furthermore, in our previous study, key results obtainedby using dye-loaded parasites were confirmed by using transfectantstrains containing a pH-sensitive protein localized to the DVby fusion to a native DV protein (15, 21), thereby confirmingthe validity of the method.

Half-times for DV alkalinization were determined by fittinga sigmoidal curve, {F = F₀ + F_max/[1 + (t/t_1/2)^c]}, where Fis the fluorescence ratio, F₀ is the initial fluorescence ratio,t is time, t_1/2 is the half-time for DV alkalinization, F_maxis the maximal change in fluorescence ratio, and c is a fittedconstant, to the data by regression analysis using SigmaPlot2004 (Systat Software, Inc.). F₀ was set to the resting fluorescenceratio averaged over the 20 s immediately prior to the openingof the fluorometer chamber to add concanamycin A.

RESULTS

Top

RESULTS

PfCRT mutations give rise to an increase in the rate of DV alkalinization seen following H⁺ pump inhibition. For each of the three transfectant parasite strains used inthis study (the CQS strain C2_GC03 and the CQR strains C4_Dd2and C6_7G8), the addition of the V-type H⁺-ATPase inhibitorconcanamycin A (100 nM) resulted in an immediate alkalinizationof the DV. Figure 1A shows representative DV alkalinizationtraces for each of the three strains, and Fig. 1B shows averageddata in which the rate of DV alkalinization is expressed asthe inverse of the time taken for half-maximal alkalinizationof the DV (1/half-time).

View larger version (13K):
[in this window]
[in a new window]

FIG. 1. Alkalinization of the DV upon addition of the V-type H⁺-ATPase inhibitor concanamycin A. (A) Representative fluorometer traces for mature trophozoite-stage C2_GC03 (CQS) (dark gray), C4_Dd2 (CQR) (black), and C6_7G8 (CQR) (light gray) parasites. The parasites were isolated by saponin permeabilization of the erythrocyte membrane and suspended in minimal saline solution at pH 7.1 and 37°C. The addition of concanamycin A (100 nM) is indicated by the black arrowhead. The data are representative of those obtained in at least seven independent experiments for each strain, and the fluorescence ratio was normalized (to account for variations in signal intensity between experiments) by dividing by the maximum change in fluorescence ratio. (B) Averaged data for the rate of DV alkalinization following the addition of 100 nM concanamycin A (expressed as the inverse of the half-time for alkalinization) in the CQS strain C2_GC03 and in the CQR strains C4_Dd2 and C6_7G8 in the presence (white bars) and absence (black bars; solvent control) of 50 µM verapamil. The data (shown with the standard error of the mean) are from paired experiments (five for C2_GC03, six for C4_Dd2, and eight for C6_7G8).

As shown in Fig. 1, the rate of DV alkalinization upon the additionof concanamycin A differed between the strains. DV alkalinizationwas significantly slower in CQS C2_GC03 parasites (half-time= 84 s ± 5 s [mean ± standard error of the mean];n = 7) than in CQR C4_Dd2 (half-time = 64 s ± 3 s; n= 7) and C6_7G8 (half-time = 40 s ± 3 s; n = 10) parasites(P $≤$ 0.01; unpaired t tests). Furthermore, there was a significantdifference in the rate of DV alkalinization between the twoCQR strains (P < 0.001; unpaired t test), with the half-timefor strain C4_Dd2 being 1.6-fold greater than that for strainC6_7G8.

We have previously shown that for parasites suspended in theabsence of CQ, 50 µM verapamil caused a slight (but notstatistically significant) slowing of DV alkalinization in threeCQS strains while having no significant effect on the half-timesfor DV alkalinization in three CQR strains (21). In contrast,50 µM verapamil was found to have a significant effecton the half-times for DV alkalinization in CQR C4_Dd2 (P = 0.03;paired t test) and C6_7G8 (P < 0.001; paired t test) parasitesbut not in CQS C2_GC03 parasites (P = 0.2; paired t test). Figure1B shows averaged data from paired experiments for each of thestrains. As can be seen from the results in Fig. 1B, verapamilincreased the rate of DV alkalinization (expressed as 1/half-time)in C4_Dd2 parasites (by 1.3-fold) but decreased the rate ofDV alkalinization in C6_7G8 parasites (by 1.6-fold).

PfCRT mutations give rise to a CQ-associated H⁺ leak from the DV. In the presence of CQ, the rate of DV alkalinization followingH⁺ pump inhibition was increased in CQR C4_Dd2 and C6_7G8 parasitesand decreased in CQS C2_GC03 parasites, consistent with CQ resistance-conferringPfCRT mutations giving rise to the CQ-associated H⁺ leak seenpreviously in three CQR strains (21). The results in Fig. 2show the effect of the length of time of CQ exposure on themagnitude of the CQ-associated increase (for C4_Dd2 and C6_7G8)or decrease (for C2_GC03) in the rate of DV alkalinization (expressedas the inverse of the alkalinization half-time). There was somevariation between the results of independent experiments, buttypically, $~$ 4 min was required for CQ to exert its maximal effect.In recognition of the interexperimental variation, a time-course(such as that shown in Fig. 2) was performed at the beginningof each experiment, and the time point at which CQ had its maximaleffect was used for all remaining traces involving the additionof CQ and/or verapamil.

View larger version (17K):
[in this window]
[in a new window]

FIG. 2. Effect of the length of time of CQ exposure on the rate of concanamycin A-induced DV alkalinization. CQ (2.5 µM) was added to suspensions of isolated mature trophozoite-stage C2_GC03 (CQS) (closed triangles), C4_Dd2 (CQR) (open circles), and C6_7G8 (CQR) (closed circles) parasites at the indicated number of minutes prior to the addition of 100 nM concanamycin A. The data are averaged from the results of four to five independent experiments for C2_GC03, two to five independent experiments for C4_Dd2, and four to seven independent experiments for C6_7G8 and are shown with the standard error of the mean. Where not shown, error bars fall within the symbols.

We next investigated the CQ concentration dependence of theCQ-associated H⁺ leak in the CQR strains C4_Dd2 and C6_7G8.The results in Fig. 3 show the CQ-associated increase in theinitial rate of concanamycin A-induced DV alkalinization asa function of CQ concentration. This was obtained by subtractingthe initial rate of alkalinization measured in the absence ofCQ from that measured in the presence of the different concentrationsof CQ. CQ had a significant effect on the initial rate of DValkalinization at concentrations of (and above) 0.625 µMfor each strain (P $≤$ 0.03; paired t tests). There was no significantdifference between the two CQR strains in the CQ-induced increasein the initial rate of DV alkalinization at any CQ concentration(P > 0.2; unpaired t tests). CQ concentrations above 10 µMwere not tested as they had a marked effect on resting pH_DVthat may affect the determination of DV alkalinization rates(21).

View larger version (16K):
[in this window]
[in a new window]

FIG. 3. CQ concentration dependence of the CQ-associated increase in the initial rate of concanamycin A-induced DV alkalinization in isolated mature trophozoite-stage CQR C4_Dd2 (open circles) and C6_7G8 (closed circles) parasites. Initial rates of DV alkalinization were determined by fitting an exponential curve to the data {[F = F₀ + a(1 – e^–^bt)], where F is the fluorescence ratio, F₀ is the starting fluorescence ratio, t is time, and a and b are fitted constants}. Multiplication of a and b yielded the initial rate of alkalinization. To account for differences in signal intensity between experiments, the fluorescence ratio was normalized prior to determining the initial rate by dividing by the maximum change in fluorescence ratio. Within each experiment, the initial rate of alkalinization for the control trace without CQ was subtracted from the initial rates in the presence of CQ. The data are averaged from the results of five independent experiments for each strain; the lines were drawn using an exponential curve fitted to the data {[y = y₀ + a(1 – e^–^bx)], where y is the initial rate of alkalinization; x is the CQ concentration; and y₀, a, and b are fitted constants}. Error bars show standard errors of the means. For clarity, only positive error bars are shown for C6_7G8 and only negative error bars are shown for C4_Dd2.

The CQ resistance reverser verapamil inhibits the CQ-associated H⁺ leak. In our previous study, we showed that 50 µM verapamilcaused complete inhibition and 25 µM verapamil significant(but only partial) inhibition of the CQ-associated increasein the DV alkalinization rate in CQR K1 and 7G8 parasites whilenot affecting DV alkalinization in the presence of CQ in CQS3D7 or D10 parasites (21). The effects of verapamil alone onthe half-times for DV alkalinization in the two CQR pfcrt transfectantstrains (at the relevant concentrations for inhibition of theCQ-associated H⁺ leak; Fig. 1B) precluded us from making quantitativecomparisons of the relative susceptibilities of these strainsto verapamil-mediated inhibition of the CQ-associated H⁺ leak.However, it was clear that the CQ-associated H⁺ leak was sensitiveto inhibition by verapamil in both CQR strains. Figure 4 showsrepresentative traces of concanamycin A-induced DV alkalinizationfor each of the two CQR transfectant strains in the presenceof 2.5 µM CQ, in the presence of 2.5 µM CQ togetherwith 50 µM verapamil, and in the presence of 50 µMverapamil alone. Verapamil (50 µM) did not have a significanteffect on the half-time for DV alkalinization in the presenceof 2.5 µM CQ in the CQS strain C2_GC03 (P = 0.5; pairedt test; not shown).

View larger version (12K):
[in this window]
[in a new window]

FIG. 4. Representative fluorometer traces showing DV alkalinization in isolated mature trophozoite-stage CQR C4_Dd2 (A) and C6_7G8 (B) parasites in the presence of 2.5 µM CQ (light gray), in the presence of 2.5 µM CQ and 50 µM verapamil (dark gray), and in the presence of 50 µM verapamil alone (black). The data are representative of the results of two independent experiments for C4_Dd2 and three independent experiments for C6_7G8.

DISCUSSION

Top

DISCUSSION

CQ resistance-conferring mutations in PfCRT and the endogenous leak of H⁺ from the DV. Concanamycin A inhibits the V-type H⁺-ATPase that pumps H⁺ intothe parasite's DV against an electrochemical gradient. In thepresence of this inhibitor, H⁺ exiting the DV (down the electrochemicalgradient) via (uncharacterized) "H⁺ leak pathways" can no longerbe replaced by the V-type H⁺-ATPase, and the DV undergoes alkalinization(28).

In this study, we found that CQ resistance-conferring PfCRTmutations increase the rate of concanamycin A-induced DV alkalinization.This is in agreement with the results of previous reports showingthat the DV alkalinization seen following inhibition of theH⁺-ATPase is faster in CQR parasites than in CQS parasites (11,21). The finding is consistent with the PfCRT mutations eitherincreasing the endogenous leak of H⁺ from the DV in CQR parasitesor causing a decrease in the internal buffering power of theDV.

The rate of the concanamycin A-induced DV alkalinization wassignificantly higher in C6_7G8 parasites than in C4_Dd2 parasites(Fig. 1). This difference may be attributable to one or moreof the seven amino acids that differ between the forms of PfCRTfound in the two strains and/or to the reported 30- to 40%-lowerexpression level of the PfCRT protein in C4_Dd2 compared tothe level in C6_7G8 parasites (which express PfCRT at a levelcomparable to that in C2_GC03 parasites) (34). It should benoted that C6_7G8 and C4_Dd2 parasites have very similar 50%inhibitory concentrations for growth inhibition by CQ (34);there is therefore not a direct relationship between the ratesof concanamycin A-induced DV alkalinization and the degree ofCQ resistance in these strains.

Further studies are required to understand the mechanism bywhich mutations in PfCRT increase the rate of DV alkalinizationupon H⁺ pump inhibition. If the increased alkalinization ratedoes result from an increased H⁺ efflux (as opposed to a decreasein buffering capacity), this might be due to the appearanceof an additional H⁺ leak pathway or to the enhancement of aH⁺ leak that is also present in CQS parasites. Furthermore,while it seems likely that PfCRT itself mediates the effluxof H⁺, the possibility that it regulates, rather than forms,a H⁺ leak pathway cannot be excluded. If PfCRT is a H⁺-coupledtransporter (as are other members of the drug/metabolite transportersuperfamily [22]), mutations that alter its H⁺/substrate stoichiometry,substrate specificity, or binding affinity (either for the substrateor H⁺) could conceivably result in an enhanced H⁺ leak fromthe DV. Alternatively, the mutations could uncouple the cotransportof substrate and H⁺ in such a way that PfCRT takes on an additional"H⁺ channel" function (24).

Three recent studies have reported there to be no differencein pH_DV between CQS and CQR parasites (13, 15, 19). For pH_DVto remain unchanged in the presence of a PfCRT-associated increasein H⁺ efflux from the DV, the rate of H⁺ pumping into the DVwould have to be increased (and even more so in the presenceof CQ). The V-type H⁺-ATPase appears to be the only DV H⁺ pumpoperating under the conditions of our experiments (21); however,a DV H⁺-pyrophosphatase activity has also been characterized(28). In a recent study using drug-selected lines generatedin the laboratory, it was shown that the acquisition of mutationsin pfcrt is accompanied by changes in the expression levelsof many genes (14). One gene found to be upregulated in thedrug-selected lines encodes a V-type H⁺-pyrophosphatase (P.falciparum VP2) (14). Although the mutant pfcrt alleles investigatedby Jiang et al. were different from those studied here, thisfinding is consistent with the hypothesis that an increase inthe activity of P. falciparum VP2 may have a role in maintainingpH_DV in parasites with a PfCRT-mediated increase in the endogenousleak of H⁺ from the DV.

Regardless of the specific protein involved, energy is requiredto transport H⁺ ions into the DV against their electrochemicalgradient. An increase in H⁺-pumping activity would thereforebe expected either to increase the overall metabolic rate ofthe cell or to result in the diversion of energy from otherprocesses. In either case, the imposition by PfCRT mutationsof an additional energy demand might result in a cost to parasitefitness (as seen for drug resistance-associated mutations inanother DV membrane protein, Pgh-1 [12]). To date, there hasbeen no detailed investigation into whether PfCRT mutationsgive rise to a fitness cost, although the return of CQS parasitepopulations to certain areas after the cessation of CQ chemotherapy(18, 25, 40) may point to there being such a cost associatedwith CQ resistance.

CQ resistance-conferring mutations in PfCRT underlie the CQ-associated H⁺ leak from the DV. The major finding in this study is that the mutations in PfCRTthat confer CQ resistance give rise to the CQ-associated H⁺leak from the DV observed in CQR (but not CQS) parasites (21).This provides support for the hypothesis that mutant forms ofPfCRT reduce intravacuolar CQ accumulation and, hence, conferCQ resistance by mediating the efflux of CQ (in its protonatedform and/or in symport with protons) from the DV. Based on theresults of bioinformatic and genetic studies, it seems likelythat mutant PfCRT itself transports CQ (2). However, the possibilitythat PfCRT mutations result in CQ efflux through an indirectmechanism cannot be excluded. It should be noted that the CQSstrain GC03 used to construct the pfcrt transfectant strainsinvestigated in this study was derived from a genetic crossbetween a CQS strain (HB3) and a CQR strain (Dd2) (42). Thepfcrt transfectant lines therefore contain an extensive geneticcontribution from Dd2, which may have adapted its genome inresponse to the mutant pfcrt allele. It remains to be seen whetherthe introduction of mutant PfCRT forms into CQS strains withoutthis genetic contribution from a CQR strain would result ina CQ-associated H⁺ efflux from the DV and a CQR phenotype.

The time taken for CQ to exert its maximal effect on the rateof DV alkalinization was $~$ 4 min for each strain (Fig. 2). Thistime is likely to represent that required for the CQ, addedto the extracellular medium, to enter the parasite and equilibrateacross the DV membrane. It is comparable to, although slightlyshorter than, the times required for maximal uptake of radiolabeledCQ by parasitized erythrocytes observed by Sanchez and colleagues(33). As seen previously in three CQS strains (21), CQ causeda slowing of DV alkalinization in the CQS strain C2_GC03. Thismay result from the accumulated CQ exerting a buffering effectin the DV (21).

In both CQR transfectant strains, the initial rate of DV alkalinizationincreased with the CQ concentration up to the highest concentrationtested (10 µM). As the CQ concentration is lowered, sotoo is the amount of CQ-associated H⁺ being translocated acrossthe DV membrane. The lowest CQ concentration at which the CQ-associatedH⁺ leak could still be detected above the constitutive H⁺ leakpathways was 0.625 µM in each strain (Fig. 3), which issomewhat higher than the 50% inhibitory concentrations for growthinhibition for CQ against these strains ( $~$ 0.15 µM) (34).It is important to emphasize that the significance of this worklies not in the magnitude of the CQ-induced H⁺ leak at pharmacologicalCQ concentrations but in the fact that that the data provideevidence that CQ resistance-conferring mutations in PfCRT areassociated with a verapamil-sensitive CQ:H⁺ pathway in the parasiteDV. There was no evidence for these short exposures to suprapharmacologicalCQ concentrations resulting in parasite damage; microscopicexamination at the conclusion of such experiments showed thatthe parasites remained intact and with the fluorescence stilllocalized to their DVs. At each of the CQ concentrations tested,there was no significant difference between the C4_Dd2 and C6_7G8parasites in the CQ-associated increase in the DV alkalinizationrate. This is consistent with the previous finding that bothstrains have very similar CQ sensitivities in parasite proliferationassays (34).

The increase of the initial rate of CQ-associated DV alkalinizationwith CQ concentration was not linear (Fig. 3). As discussedpreviously for another CQR strain (21), the departure from linearitydoes not necessarily reflect a saturation of the CQ efflux systemand could well be a consequence of other factors (e.g., thepermeability of the DV membrane to counterions) limiting therate of DV alkalinization at the higher CQ concentrations tested.However, the data do imply that if the process is saturable,the K_m for CQ must be supramicromolar.

The CQ-associated H⁺ leak in C4_Dd2 and C6_7G8 was sensitiveto verapamil, as was seen in a previous study for the CQR strains7G8, K1, and Dd2-PM2-GFP (21). The verapamil concentration requiredfor complete inhibition of the CQ-associated H⁺ leak (50 µM)was $~$ 50-fold higher than concentrations typically used for CQchemosensitization in parasite proliferation assays. The reasonfor this is not known but may be related to the length of timeof verapamil exposure (<20 min in the case of the transportexperiments performed here versus 2 days for parasite proliferationassays). Another possibility is that verapamil and CQ competefor binding to PfCRT, in which case the need for higher verapamilconcentrations in the transport experiments might be a consequenceof the use in this study of CQ concentrations higher than thoseused in parasite proliferation assays.

The 7G8 and Dd2 PfCRT alleles investigated in this study differat seven amino acid positions. Using an allelic exchange approach,Lakshmanan et al. demonstrated that mutations in PfCRT precedingresidue 76 (at positions 72, 74, and 75, all of which differbetween 7G8 and Dd2) affect the susceptibility of CQR parasitesto chemosensitization by verapamil (20). Consistent with this,verapamil (at 0.8 µM) was shown to have a less-pronouncedeffect on CQ susceptibility in C6_7G8 than in C4_Dd2 in parasiteproliferation assays (34). These strains may therefore be expectedto differ in their sensitivities to verapamil-mediated inhibitionof the CQ-associated H⁺ leak. The effects of verapamil (in theabsence of CQ) on the rate of DV alkalinization in C4_Dd2 andC6_7G8 precluded quantitative comparisons of susceptibilitiesto verapamil-mediated inhibition of the CQ-associated H⁺ leakin the two strains. Nevertheless, the finding that verapamilhad opposing effects in the two CQR transfectant strains isconsistent with there being differences in the interactionsof verapamil with the two mutant forms of PfCRT.

The verapamil-mediated increase in the rate of concanamycinA-induced DV alkalinization in C4_Dd2 raises the possibilitythat the Dd2 mutant form of PfCRT mediates the efflux of verapamil,a monoprotic weak base, from the DV (in its protonated formand/or in symport with protons). No such verapamil-induced H⁺leak was detected in the other CQR transfectant strain, C6_7G8.Whether this reflects the absence of verapamil efflux via the7G8 form of PfCRT or its masking by other effects (such as verapamil-mediatedbuffering of pH_DV or inhibition of a H⁺ leak pathway) remainsto be determined. The absence of an effect of verapamil on DValkalinization in the CQS strain, C2_GC03, may point to verapamilinhibiting the putative additional H⁺ leak pathway formed bythe mutant PfCRT form present in C6_7G8.

Summary. The CQ resistance-conferring Dd2 and 7G8 forms of PfCRT, butnot wild-type PfCRT, give rise to a CQ-associated H⁺ leak fromthe DV, consistent with mutant PfCRT mediating the efflux ofCQ (protonated and/or in symport with H⁺) from the DV. Whilethe CQR strains C4_Dd2 and C6_7G8 appear equivalent in thisregard, they differ in their rates of DV alkalinization followingH⁺ pump inhibition in the absence of CQ and in the extent to(and manner in) which this is modulated by verapamil. The combinationof the pH_DV method and the pfcrt transfectant lines used hereshould prove useful in further studies of the role of pfcrtin influencing parasite susceptibility to a range of antimalarialdrugs.

ACKNOWLEDGMENTS

This work was supported by Australian National Health and MedicalResearch Council grant 418055.

We are grateful to David Fidock for the generous gift of theparasite strains used in this study and to the Canberra Branchof the Australian Red Cross Blood Service for the provisionof blood.

FOOTNOTES

* Corresponding author. Mailing address: Biochemistry and Molecular Biology, School of Biology, The Australian National University, Canberra ACT 0200, Australia. Phone: 61 2 6125 2284. Fax: 61 2 6125 0313. E-mail: Kiaran.Kirk{at}anu.edu.au

Published ahead of print on 13 October 2008.

REFERENCES

Top