Evidence That Mutant PfCRT Facilitates the Transmission to Mosquitoes of Chloroquine-Treated Plasmodium Gametocytes

Abstract

Resistance of the human malarial parasite Plasmodium falciparum to the antimalarial drug chloroquine has rapidly spread from several independent origins and is now widely prevalent throughout the majority of malaria-endemic areas. Field studies have suggested that chloroquine-resistant strains might be more infective to mosquito vectors. To test the hypothesis that the primary chloroquine resistance determinant, mutations in PfCRT, facilitates parasite transmission under drug pressure, we have introduced a mutant or wild-type pfcrt allele into the rodent model malarial parasite Plasmodium berghei. Our results show that mutant PfCRT from the chloroquine-resistant 7G8 strain has no effect on asexual blood stage chloroquine susceptibility in vivo or ex vivo but confers a significant selective advantage in competitive mosquito infections in the presence of this drug, by protecting immature gametocytes from its lethal action. Enhanced infectivity to mosquitoes may have been a key factor driving the worldwide spread of mutant pfcrt.

Acquisition of resistance to chloroquine (CQ) by the human malarial parasite Plasmodium falciparum has been identified as a major cause of earlier documented increases in malaria-related morbidity and mortality [1]. CQ resistance results primarily from point mutations in the pfcrt gene that encodes the P. falciparum chloroquine resistance transporter PfCRT, resident on the digestive vacuole (DV) membrane [2–4]. Mutant forms of PfCRT are thought to efflux CQ out of the DV, thereby preventing this drug from inhibiting the detoxification of iron-bound heme following hemoglobin proteolysis [5–6].

Clinical studies have found a clear association between pfcrt mutations, particularly the K76T polymorphism that is essential for CQ resistance in vitro [4], and CQ treatment failure [7]. Field studies have also suggested that CQ-resistant parasites demonstrate greater infectivity to the mosquito vector. Patients with CQ-resistant infections were more likely than CQ-sensitive infections to carry gametocytes and were more infective to mosquitoes at the time of diagnosis, as well as subsequent to CQ monotherapy [8–10]. This increased infectivity was associated with mutant pfcrt [11]. Importantly, in 2 of these studies patients were gametocyte-negative on presentation, and mosquito feeds were performed 7 days after initiating CQ treatment [10–11]. Taking into account that gametocytogenesis in P. falciparum takes 10–12 days, mature gametocytes present in peripheral blood at the time of mosquito feeding must have survived CQ therapy at earlier stages of development. This suggests that mutant PfCRT might protect immature gametocytes from CQ action, to which they would otherwise be susceptible, presumably because they continue to digest hemoglobin [12].

CQ has also been reported to possess transmission-enhancing properties (reviewed in [13]). In a CQ-resistant Plasmodium yoelii subspecies (originally mistaken for Plasmodium berghei [14]) oocyst numbers were increased up to 2.5-fold when mosquitoes were infected 12 h after a single dose of CQ [15–16]. This CQ-enhanced infectivity to mosquitoes was confirmed in uncloned P. yoelii nigeriensis but lost in CQ-sensitive or CQ-resistant subclones, suggesting that this increased infectivity may not be causally associated with the gene(s) responsible for CQ resistance in P. yoelii [17]. In separate studies, an unidentified host factor or CQ metabolite present in sera for several weeks after CQ treatment was reported to enhance mosquito infectivity of P. falciparum and P. berghei [18]. Together, the protection of immature gametocytes and the transmission-enhancing effects of CQ would selectively increase transmission of parasites carrying mutant pfcrt.

While these findings are strongly suggestive of an important correlation between CQ resistance and enhanced malaria transmission under drug pressure, the non-isogenic, nonclonal nature of infections in the field makes it difficult to distinguish the contribution of mutant pfcrt from unrelated interstrain differences. Lack of involvement of crt mutations in CQ-resistant P. chabaudi [19] and the frequently unstable nature of CQ resistance in other rodent malarias [20–21] also mean that this question cannot be addressed in drug-pressured parasite lines that have earlier been used to investigate mechanistic principles of CQ resistance (e.g., [22–23]).

In this study we have genetically engineered the rodent model malarial parasite P. berghei to express variant forms of pfcrt, in order to test the hypothesis that mutant PfCRT protects gametocytes from CQ action and facilitates transmission in the presence of drug.

METHODS

Parasite Maintenance

Mice were purchased from Charles River Laboratories (Wilmington, Massachusetts). Parasites were propagated by serial passage through intraperitoneal injection. All animal research was performed according to protocols approved by our Institutional Animal Care and Use Committees.

Plasmid Design

Primer sequences are listed in Table 1. Briefly, a 1.0-kb 5′ untranslated region (UTR) sequence from pbcrt (PlasmoDB gene identifier PB000746.03.0) was amplified using primers c1 and c2 and cloned into pLITMUS28 via BglII and BssHII to serve as the 5′ region of homology for recombination. The pfcrt (MAL7P1.27) cDNA from HB3 and 7G8 was generated using primers c5 and c6 and cloned into pBAD-Topo (Invitrogen) to introduce a C-terminal V5 epitope tag (GKPIPNPLLGLDST). Using primers c5 and c7, pfcrt-V5 was amplified from pBAD-Topo/pfcrt and inserted into the pLITMUS28 vector downstream of the pbcrt 5′UTR sequence via AflIII/BssHII (compatible sites) and XhoI. Then .7 kb of pycrt (PY05061) 3′UTR was amplified using primers c8 and c9 and inserted downstream of the pfcrt-V5 sequence via EcoRI and PstI to serve as the 3′ terminator for expression. A 1.0-kb pbcrt 3′UTR sequence was amplified using primers c3 and c4 and cloned via KpnI/StuI to serve as the 3′ region of homology. The transfection plasmid also contained the mutant Toxoplasma gondii dihydrofolate reductase-thymidylate synthase (tgdhfr-ts) selectable marker cassette. A knockout plasmid was constructed in an identical manner to the allelic replacement plasmid except that it lacked the pfcrt coding sequence.

Table 1.

Primer Sequences

Primer Name	Primer Sequence (5' → 3')	Primer Features
Primers for Plasmid Construction
c1	CTTagatctTTTACTATTTTCAATTTCAACGGTGTG	BglII site in lower case
c2	AAGgcgcgcTCCTGTcatatgTAATACACTGATTTAATATATTTAAAAAAATTG	BssHII and NdeI sites in lower case
c3	CTTggtaccCTTAAATGATTTTTGTAAATGCCAC	KpnI site in lower case
c4	AATaggcctGACGTTATGGCGACGTGTTGCC	StuI site in lower case
c5	TTCacgcgtAAAAAGAATAATCAAAAGAATTCAAGCAAAAATGAC	AflIII site in lower case
c6	TTGTGTAATAATTGAATCGACG	penultimate codon of pfcrt underlined
c7	AAGctcgagTTACGTAGAATCGAGACCGAGGAG	XhoI site in lower case; stop codon underlined
c8	CTTgaattcATATTTTTTTTAAATGCCACATAAAG	EcoRI site in lower case
c9	AATctgcagGATATTTCAAAAATCTTAGCATAAGG	PstI site in lower case
Primers for Diagnostic PCRs
p1	CATATGTGATAATTTACTTGCTTGC
p2	AAATATATTAAATGCGCTCCATTT
p3	TTTTTCCCCTTCAATTCACTTTC
p4	AACCTCCACCTAAACGTGAGCC
p5	GCCTTGTCACTTGTTGTGCC
p6	TTATATGCATTTTATAAAATTTTTTATTTATTTATAAGC
Primers for Pyrosequencing
CB10b	CGGATGTTACAAAACTATAGTTACC
CF5c	AATTCAAGCAAAAATGACGAGCG
pyro-2	GTTCTTTTAGCAAAAATT

Generation and Molecular Characterization of pfcrt-Recombinant P. berghei Lines

Transfection of P. berghei ANKA with 5 μg of HpaI/ScaI-linearized plasmid and subsequent selection with pyrimethamine was performed as described [24]. Parasite clones were obtained by limiting dilution. Diagnostic PCRs, Southern blotting, and western blotting with rabbit anti-PfCRT polyclonal [2] or mouse anti-V5 monoclonal antibodies (Invitrogen) were performed as described elsewhere [25].

Immunoelectron Microscopy

Mixed blood stages of P. berghei were fixed in 4% paraformaldehyde (Electron Microscopy Sciences) in .25 mol/L HEPES (pH 7.4) for 1 h at room temperature and then in 8% paraformaldehyde in the same buffer overnight at 4°C. These were infiltrated, frozen, and sectioned as described elsewhere [26]. The sections were immunolabeled with rabbit anti-PfCRT antibodies [2] diluted 1:5 in phosphate-buffered saline (PBS)/1% fish skin gelatin, then with mouse anti-rabbit IgG antibodies, followed directly by 10-nm protein A gold particles (Department of Cell Biology, Utrecht University Medical School, the Netherlands) before examination with a Philips CM120 Electron Microscope (Eindhoven, the Netherlands) under 80 kV.

Ex Vivo Drug Susceptibility Assays

To synchronize P. berghei parasites, blood was collected by cardiac puncture and cultured overnight at 37°C with shaking (65 rpm) in schizont medium (RPMI 1640 with 25 mmol/L HEPES, 3 mmol/L l-glutamine, 24 mmol/L sodium bicarbonate, 367 μmol/L hypoxanthine, 50 U/mL penicillin, and 50 μg/mL streptomycin; pH 7.2) supplemented with 25% fetal bovine serum. The following day mature schizonts were collected by centrifugation (500 × g 10 minutes), and injected intravenously into naive mice. Early ring stages were harvested 3 h post-injection. Ex vivo assays were performed in 24 well plates, with each well containing 500 μl of cell suspension (100–300 μL packed blood in 10 mL), 250 μL [³H]-hypoxanthine (final amount 3 μCi/well), and 250 μL drug at 4× the required concentrations, all in low-hypoxanthine complete medium [27]. Assays were incubated for 24 h at 37°C with shaking. Thereafter the contents were transferred into 96 well plates, lysed by freezing, harvested onto filtermats, and counted using a Wallac MicroBeta scintillation counter (PerkinElmer, Waltham, Massachusetts). IC₅₀ values were calculated by linear regression.

In Vivo Drug Susceptibility Assays (Peters’ 4-Day Test)

Female Swiss Webster mice (∼22–25 g) were infected intraperitoneally with 10⁷ parasites [20]. One hour later, and on days 1–3, mice were injected subcutaneously with PBS or 1 of 3 different concentrations of CQ phosphate (Sigma-Aldrich). Parasitemias were determined from tail blood smears 24 h after the last injections. The ED₅₀ and ED₉₀ values, defined as the drug concentration that reduced parasitemia by 50% or 90% compared with the control mice, were derived by linear regression.

Transmission Experiments

Female CD1 mice were infected intraperitoneally with 5 × 10⁶ parasites of an ∼1:1 mixture of Pb^HB3 and Pb^7G8. Three days later, mice were injected subcutaneously with PBS, 1 mg/kg or 10 mg/kg CQ phosphate. Twelve or 24 hours after drug administration, mice were anesthetized with 80–100 mg/kg ketamine and 10 mg/kg xylazine and exposed to Anopheles stephensi mosquitoes for 20–25 minutes. Blood was collected immediately after mosquito feeds. On days 10–11 of mosquito infection, 10 midguts (containing oocysts) were collected per group. Genomic DNA was prepared using the DNeasy kit (Qiagen, Valencia, California), and PCR products were amplified using primers CB10b and biotinylated CF5c. Allele frequencies at codon 76 (AAA in HB3 → ACA in 7G8) were determined by pyrosequencing with primer pyro-2, using Pyro Gold reagents (Qiagen) on a PSQ 96MA pyrosequencer (Biotage).

Statistical Analysis

Significance was tested using the Mann–Whitney U test.

RESULTS

Generation of pfcrt-Recombinant P. berghei Lines

We implemented an allelic exchange strategy to replace the endogenous pbcrt gene with the P. falciparum ortholog of either wild-type or mutant origin (from the CQ-sensitive HB3 or CQ-resistant 7G8 strains, respectively). The 7G8 pfcrt allele carries 5 point mutations, resulting in amino acid changes C72S, K76T, A220S, N326D, and I356L. This SVMNT (codon 72–76)-type haplotype is common in South America and Papua New Guinea and is occasionally observed in Southeast Asia and India; however it is only rarely found in Africa, where the other major, CVIET-type PfCRT haplotype predominates. We also attempted to introduce a CVIET-type pfcrt allele from the CQ-resistant Dd2 strain, which encodes 8 polymorphisms (M74I, N75E, K76T, A220S, Q271E, N326S, I356T, and R371I), but failed to obtain integration of this allele into the pbcrt locus.

With our double crossover strategy the endogenous pbcrt promoter was retained for expression of the pfcrt transgene (Figure 1A). These transfections produced the recombinant P. berghei lines Pb^HB3 and Pb^7G8. Following cloning of the integrant parasite populations, correct integration of the construct into the pbcrt locus was confirmed by Southern blot and PCR (Figure 1B–1C). Sequencing of diagnostic PCR products confirmed the correct haplotypes in the respective parasite clones. PfCRT expression was demonstrated by western blotting using anti-PfCRT antibodies (Figure 1D). These displayed cross-reactivity with PbCRT in the parental ANKA strain, presumably due to conservation of the 19 amino acid epitope at the PfCRT C-terminus in the P. berghei ortholog (32% identity, 95% similarity). In the recombinant clones, expression was confirmed with antibodies to the V5 epitope tag appended to the PfCRT C-terminus (Figure 1D).

Figure 1.

Generation and molecular characterization of Pb^HB3 and Pb^7G8. (A) Schematic of allelic replacement strategy used to exchange the endogenous pbcrt (dark gray box) with either mutant (7G8) or wild-type (HB3) pfcrt (black box) by double crossover homologous recombination. The targeting cassette was released from the plasmid backbone by HpaI digestion. The 5' and 3' homologous regions (5' and 3' pbcrt UTRs) are indicated in light gray. The pbcrt 5'UTR and pycrt 3'UTR (labeled py3’) serve as the regulatory regions for the expression of pfcrt in the recombinant parasites. The tgdhfr-ts indicates the selectable marker cassette. Fragments generated upon restriction digestion for Southern blotting, the location of the pbcrt 5'UTR probe, and primers used for molecular characterization are indicated. B, BglII; H, HpaI; N, NdeI. (B) Southern hybridization of genomic DNA from P. berghei ANKA, Pb^7G8, and Pb^HB3, digested with BglII/NdeI and hybridized with the pbcrt 5'UTR probe. This yielded the expected bands of 2.5 kb for ANKA and 1.8 kb for the recombinant lines. (C) PCR detection of endogenous pbcrt locus with primers p1+p2 and recombinant loci with primers p3+p4, p5+p2, and p3+p6. Primer locations and expected fragment lengths are shown in (A). M, 1.0-kb marker. (D) Western blotting of protein extracts from ANKA and the recombinant lines. The 1.5 × 10⁶ parasites were loaded per lane on 4–15% gradient sodium dodecyl sulfate-polyacrylamide gels. Top panel: Probed with polyclonal anti-PfCRT antibodies, which cross-react with PbCRT. Bottom panel: Probed with monoclonal anti-V5 antibodies. With the recombinant clones Pb^7G8 and Pb^HB3, a V5 epitope tag was appended to the 3' end of the full-length pfcrt allele. The 7G8 line represents a P. falciparum control.

In earlier studies in P. falciparum, the failure to disrupt pfcrt suggested an essential role of this gene for asexual blood stage viability [28]. To determine whether the P. berghei ortholog is similarly essential we attempted to delete pbcrt. PCR analysis detected a minor subpopulation of pbcrt-deleted parasites in 2 of 4 independent transfection experiments, with the majority of the parasites maintaining the knockout plasmid episomally (data not shown). However, the knockout population could not be cloned, indicating that pbcrt is likely required for productive asexual blood stage development. The fact that P. berghei clones expressing pfcrt in place of pbcrt were readily obtained therefore suggests that pfcrt can functionally substitute for pbcrt. A conserved function is also supported by the high degree of sequence identity (∼64%) between pfcrt and its orthologs in rodent Plasmodia.

PfCRT Traffics Properly to the Digestive Vacuole Membrane in P. berghei

To determine the localization of PfCRT in the recombinant P. berghei lines, immunoelectron microscopy was performed using anti-PfCRT antibodies. In both Pb^HB3 and Pb^7G8, the limiting membrane of the pigment-containing DV was specifically decorated with gold particles (Figure 2; data not shown), confirming that PfCRT localizes correctly in the recombinant P. berghei lines. Additionally, some labeling of intravacuolar membranous structures was observed (Figure 2B). For the parental ANKA strain, the intracellular staining was more variable, with ∼25% of parasites exhibiting labeling on the DV membrane (data not shown). This variability could be due to reduced affinity of the anti-PfCRT antibodies to a partially conserved epitope in PbCRT.

Figure 2.

Ultrastructural detection of PfCRT in pfcrt-expressing P. berghei parasites. Fixed intraerythrocytic Pb^HB3 parasites were immunolabeled with anti-PfCRT antibodies, and the labeling was visualized by incubating with 10-nm protein A gold particles (locations highlighted with arrows). (A–C) PfCRT was localized to the limiting membrane of the digestive vacuole, inside which hemozoin pigment was clearly visible. (B) Some labeling was also associated with intravacuolar membranes (asterisk). (C) Several sections of the DV are visible (see filled triangles), most of which were decorated with gold particles at their periphery (arrows). Similar DV labeling was observed in Pb^7G8 (data not shown). DV, digestive vacuole; ER, endoplasmic reticulum; m, mitochondria. Scale bars are 300 nm.

Mutant PfCRT^7G8 Does Not Confer Ex Vivo Resistance to CQ or Monodesethyl-Amodiaquine (md-AQ) or In Vivo Resistance to CQ in P. berghei Asexual Blood Stages

Introduction of the 7G8 pfcrt allele into the CQ-sensitive P. falciparum GC03 strain was earlier found to result in an ∼6- and ∼17-fold increase in IC₅₀ values for CQ and its metabolite monodesethyl-CQ (md-CQ), respectively [3]. Using ex vivo and in vivo assays, we tested whether the same allele could confer CQ resistance to P. berghei.

Ex vivo assays were set up with synchronized early ring stage parasites, and—because P. berghei schizonts do not rupture ex vivo—the incorporation of [³H]-labeled hypoxanthine was measured in the presence of drug for one developmental cycle. Surprisingly, these assays revealed no statistically significant increase in resistance to CQ of Pb^7G8 compared with the control recombinant Pb^HB3 and the parental, CQ-sensitive ANKA strain (14.8 ± 6.5 nmol/L versus 12.1 ± 2.6 nmol/L and 10.6 ± 1.2 nmol/L, respectively; Figure 3A). We also did not observe any significant differences in md-CQ responses among Pb^7G8, Pb^HB3, and ANKA (35.5 ± 20.7 nmol/L, 15.6 ± 4.9 nmol/L, and 21.4 ± 8.5 nmol/L, respectively; Figure 3B).

Figure 3.

Mutant PfCRT^7G8 does not confer chloroquine or monodesethyl-amodiaquine resistance in P. berghei. Ex vivo drug-susceptibility profiles of parental P. berghei ANKA (white bars), Pb^HB3 (gray bars), and Pb^7G8 (black bars) for (A) chloroquine (CQ), (B) monodesethyl-chloroquine (md-CQ), (C) monodesethyl-amodiaquine (md-AQ), and (D) pyrimethamine (PYR). Bars represent mean + SEM IC₅₀ values from 4–6 (CQ), 3–5 (md-CQ), 3–4 (md-AQ), and 2 (PYR) independent assays. (E) In vivo CQ-susceptibility profiles of P. berghei ANKA (white bars), Pb^HB3 (gray bars), and Pb^7G8 (black bars). Bars represent mean + SEM ED₅₀ values from 3 independent experiments (4 groups of 5 mice per parasite line per experiment).

Recent findings have implicated the 7G8 pfcrt allele in resistance to amodiaquine and its active metabolite monodesethyl-amodiaquine (md-AQ) [29–30]. We therefore also determined md-AQ IC₅₀ values but detected no differences in susceptibility among Pb^7G8, Pb^HB3, and ANKA (17.3 ± 2.4 nmol/L versus 24.5 ± 6.0 nmol/L and 15.1 ± 3.0 nmol/L, respectively; Figure 3C). To validate our assays we determined parasite susceptibility to pyrimethamine, the drug used to select for the tgdhfr-ts marker in Pb^7G8 and Pb^HB3. Consistent with published data [31], both recombinant clones displayed a substantial increase in mean IC₅₀ values compared with ANKA (49,519 ± 17,856 nmol/L and 36,095 ± 1,054 nmol/L versus 61 ± 26 nmol/L, respectively; Figure 3D).

We also assayed CQ responses in vivo following Peters’ 4-day test, which assesses inhibition of growth after 4 days of drug treatment [20]. In agreement with our ex vivo results, we observed no statistically significant increase in the ED₅₀ or ED₉₀ values in Pb^7G8 compared with Pb^HB3 or ANKA. The ED₅₀ and ED₉₀ values were 2.1 ± .3 mg/kg and 3.8 ± .1 mg/kg for Pb^7G8, 2.4 ± .1 mg/kg and 3.5 ± .1 mg/kg for Pb^HB3, and 2.1 ± .3 mg/kg and 4.0 ± .5 mg/kg for ANKA (ED₅₀ data shown in Figure 3E). These values are almost identical to the results for CQ-sensitive P. berghei from a previous study [20] and clearly demonstrate that mutant pfcrt^7G8 is insufficient to confer CQ resistance in P. berghei.

Mutant PfCRT^7G8 Confers a Transmission Advantage to P. berghei in the Presence of CQ

To test our hypothesis that mutant PfCRT might protect gametocytes from CQ, we analyzed the mosquito infectivity of the recombinant P. berghei lines under drug pressure. To control for environmental factors that can influence gametocyte infectivity and that may vary among mice (including drug serum levels, number of schizonts, and hematocrit) [18, 32], we infected mice with mixtures of the 2 pfcrt-recombinant lines. These mixed infections were transmitted to mosquitoes 12 or 24 h following a single dose of PBS, or 1 mg/kg CQ or 10 mg/kg CQ (both single doses are noncurative; data not shown). Allelic frequencies of Pb^HB3 and Pb^7G8 were quantified by pyrosequencing [33], and a mutant pfcrt transmission index (TI) was calculated as the percentage of the 7G8 allele at the oocyst stage divided by its frequency in the starting mixed-blood-stage population.

Importantly, in the PBS-injected control infections, allele frequencies did not change significantly between the blood stage and oocyst populations (TI ≈ 1; Figure 4), consistent with equal fitness of Pb^HB3 and Pb^7G8 at the gametocyte, gamete, ookinete, and oocyst stages in the absence of drug. Similarly, a low dose of CQ did not demonstrate a significant selective effect on the parasite mixture. In contrast, 10 mg/kg CQ strongly selected for the presence of the 7G8 pfcrt allele at the oocyst stage. This effect was apparent when mosquitoes were infected 12 h after CQ treatment (TI = 1.22 ± .27; P = .002) and to an even greater extent at 24 h (TI = 1.46 ± .11; P = .002). Interestingly, no selective effect was observed in 2 competitive transmission experiments with amodiaquine (data not shown). We note that either a relative increase in Pb^7G8 oocyst number or a relative increase in sporozoite number per Pb^7G8 oocyst compared with Pb^HB3 would result in an increased 7G8:HB3 allele ratio. We consider the former more likely, as no difference was observed in oocyst sporozoite numbers between Pb^HB3 and Pb^7G8 after CQ treatment (∼130 midgut sporozoites per oocyst in both lines), but either case would achieve the same net result, that is, facilitated transmission of Pb^7G8 under CQ pressure.

Figure 4.

Mutant PfCRT^7G8 confers a transmission advantage to P. berghei in the presence of chloroquine (CQ). Values are given as the mutant pfcrt transmission index (TI), which represents the ratio of the percentage of the 7G8 allele at the oocyst stage divided by the percentage of this allele in the initial blood stage population of mixed Pb^7G8 and Pb^HB3 parasites. Mosquito feeds were performed either 12 or 24 h after CQ treatment. Mean ± SEM TI values were calculated from 4 independent experiments (6 mice), except for the 12-h treatment with 1 mg/kg CQ (tested in 3 experiments; 4 mice). **, P < .01 (Mann–Whitney U test).

Following treatment with 10 mg/kg CQ, gametocytes frequently harbored a single, large clump of pigment, instead of the dispersed fine granules normally present in these stages. Interestingly, a significantly greater proportion of ANKA and Pb^HB3 gametocytes possessed abnormal pigment compared with Pb^7G8 (Figure 5), suggesting that mutant PfCRT^7G8 protects P. berghei gametocytes from CQ-induced pigment clumping.

Figure 5.

Mutant PfCRT^7G8 protects P. berghei gametocytes from chloroquine (CQ)-induced pigment clumping. Values given are the percentage of P. berghei ANKA (white bars), Pb^HB3 (gray bars), and Pb^7G8 (black bars) gametocytes displaying completely clumped pigment 12 and 24 h after a single dose of 10 mg/kg CQ, as visualized by light microscopy on Giemsa-stained tail blood smears. Bars represent mean + SEM values calculated from 4–13 CQ-treated mice tested in several independent experiments. In PBS-treated controls, <.25% of gametocytes were scored as having clumped pigment (data not shown). *, P < .05 (Mann–Whitney U test). Representative images of gametocytes with normal and with clumped pigment are shown on the right.

DISCUSSION

In this study we provide evidence that the 7G8 allele of pfcrt confers a selective transmission advantage to malarial parasites under CQ pressure compared with wild-type pfcrt. Our findings in isogenic P. berghei lines in competitive mosquito infections corroborate data from the field showing that in mixed infections CQ selects for gametocytes harboring mutant pfcrt [34] and that infections with parasites possessing mutant pfcrt produce significantly higher oocyst numbers compared with parasites with wild-type pfcrt [11]. Importantly, in our recombinant lines this protective effect on mosquito infectivity occurred in the absence of any effect of mutant PfCRT^7G8 on CQ resistance in the asexual blood stages. We propose that this facilitated transmission under drug pressure may have been a key factor in the rapid spread of CQ resistance [35]. It would be interesting to perform a similar analysis with a CVIET-type pfcrt allele, but we were unable to introduce the Dd2 allele, harboring this haplotype, into P. berghei. This may be due to fitness costs associated with this allele, as previously observed in allelic exchange experiments in P. falciparum [25] and supported by field studies that have reported a decline in CVIET-type mutant pfcrt alleles from African parasite populations in the absence of CQ pressure [36–38].

At which point in gametocyte development does this PfCRT-mediated protection act? P. berghei trophozoites become sexually committed during a short period 8–10 hours postinvasion (hr pi) [39]. The first mature gametocytes are observed 26 hr pi and have a survival time of 26 hours in vitro [39]. In our assays, gametocytes that were mature at feeding 12 hours after CQ treatment thus either were already committed immature gametocytes (∼14–25 hr pi) or were already mature (∼26–40 hr pi) when drug was administered. This suggests that selection for Pb^7G8 acts at the level of gametocyte survival or infectivity and not at the level of initiating gametocytogenesis.

At the 24-h time point, infectious gametocytes could also derive from young asexual parasites (∼2–10 hr pi) that were still uncommitted when first exposed to CQ. Therefore, a component of the selection for Pb^7G8 could theoretically also reflect a greater rate of gametocytogenesis induction by CQ in the young Pb^7G8 asexual parasites compared with Pb^HB3. Indeed, CQ has been shown to increase the rate of gametocyte production in P. falciparum and P. chabaudi [40–42]. Nevertheless, we consider preferential gametocytogenesis induction unlikely as asexual blood stages from both lines showed no significant difference in their CQ response. Of note, most gametocytes that were already mature, and thus CQ-insensitive when drug was administered, would have reached the end of their life span 24 h later. Thus, the majority of infectious gametocytes at the 24-h time point would have derived from gametocytes that were immature at the time of CQ administration. This decreasing reservoir of old gametocytes may be responsible for the greater selection that was noted at the 24-h compared with the 12-h time point (although the difference was not statistically significant). We conclude that mutant PfCRT protects immature gametocytes from the lethal action of CQ. This effect may be even greater in P. falciparum due to its extended duration of both gametocytogenesis (10–12 days) and the CQ-sensitive phase of gametocyte development (first ∼6 days) [12].

Mutant PfCRT^7G8 also protected P. berghei gametocytes from CQ-induced pigment clumping, a phenomenon first reported in P. berghei gametocytes [43] and also observed in immature P. falciparum gametocytes [12]. It is tempting to speculate that mutant PfCRT^7G8 reduces the concentration of CQ in the DV of Pb^7G8 gametocytes, thereby preventing pigment clumping. Intriguingly, reduced access of CQ to the DV through mutant PfCRT-mediated CQ efflux is the proposed mechanism underlying CQ resistance in P. falciparum asexual blood stages [5–6, 44]. However, it is unclear if and how the morphological difference in gametocyte pigment translates into functional differences that ultimately result in greater mosquito infectivity of Pb^7G8 gametocytes in the presence of CQ.

In contrast to the protective effect of mutant PfCRT^7G8 on CQ-treated gametocytes, no significant decrease in CQ susceptibility of asexual blood stages could be observed ex vivo or in vivo. Mutant PfCRT^7G8 correctly localized to pigment-containing vesicles, and our data on CQ-induced pigment clumping in gametocytes suggest that it decreased drug access to these compartments. Theoretically, this should decrease parasite CQ susceptibility. The fact that parasites remained CQ sensitive therefore implies that reduced CQ access to the DV is by itself insufficient to mediate resistance in our P. berghei model.

Importantly, pfcrt mutations alter parasite susceptibility to several other antimalarials, including lumefantrine, which inhibits parasites with the pfcrt K76T mutation (in both CVIET and SVMNT haplotypes) more strongly than wild-type parasites [45] and which was recently shown to reduce transmission of P. berghei [46]. Our recombinant P. berghei lines will prove useful in testing whether the transmission of parasites with a mutant pfcrt^7G8 allele is impaired under lumefantrine pressure, an exciting possibility given that artemether-lumefantrine (Coartem) has now become the most widely used first-line antimalarial treatment [47–48].

Funding

This work was supported by R01 AI50234 and an Investigator in Pathogenesis of Infectious Diseases Award from the Burroughs Wellcome Fund to DAF and R01 AI056840 to PS. AE is currently supported by a long-term fellowship from the International Human Frontier Science Program Organization.