Critical role of a K⁺ channel in Plasmodium berghei transmission revealed by targeted gene disruption

Abstract

Regulated K⁺ transport across the plasma membrane is of vital importance for the survival of most cells. Two K⁺ channels have been identified in the Plasmodium falciparum genome; however, their functional significance during parasite life cycle in the vertebrate host and during transmission through the mosquito vector remains unknown. We hypothesize that these two K⁺ channels mediate the transport of K⁺ in the parasites, and thus are important for parasite survival. To test this hypothesis, we identified the orthologue of one of the P. falciparum K⁺ channels, PfKch1, in the rodent malaria parasite P. berghei (PbKch1) and examined the biological role by performing a targeted disruption of the gene encoding PbKch1. The deduced amino acid sequence of the six transmembrane domains of PfKch1 and PbKch1 share 82% identity, and in particular the pore regions are completely identical. The PbKch1-null parasites were viable despite a marked reduction in the uptake of the K⁺ congener ⁸⁶Rb⁺, and mice infected with PbKch1-null parasites survived slightly longer than mice infected with WT parasites. However, the most striking feature of the phenotype was the virtually complete inhibition of the development of PbKch1-null parasites in Anopheles stephensi mosquitoes. In conclusion, these studies demonstrate that PbKch1 contributes to the transport of K⁺ in P. berghei parasites and supports the growth of the parasites, in particular the development of oocysts in the mosquito midgut. K⁺ channels therefore may constitute a potential antimalarial drug target.

The mosquito-borne parasite Plasmodium falciparum is the causative agent of the deadliest form of malaria, claiming an estimated 1 million to 2 million deaths annually. The spread of resistance to almost all of the currently available antimalarial drugs necessitates the development of new drugs. Putative drug targets in P. falciparum are the parasite-encoded transport proteins, which mediate the uptake of nutrients and disposal of waste products across the parasite's plasma membrane (1). Several of the parasite's transport proteins have been cloned and functionally characterized (2–6), and some of them may be potential drug targets (7–9).

K⁺ channels constitute the largest and most diverse of ion channel families and are involved in K⁺ transport, cell volume control, and regulation of membrane potential. Two putative K⁺ channel-encoding genes have been found in the P. falciparum genome (10–11), but it remains to be established whether they are functional and whether they are important for parasite survival. So far attempts to express and functionally characterize Plasmodium K⁺ channels in heterologous cell systems have been unsuccessful. Targeted gene disruption by homologous recombination has provided a valuable approach for functional characterization of gene products in the Plasmodium genus. In contrast to P. falciparum, the Plasmodium berghei genome can be manipulated with relative ease (12), and in that regard P. berghei serves as an excellent model organism for the study of gene function in malaria parasites. We identified the orthologue of the P. falciparum K⁺ channel PfKch1 in the P. berghei genome by BLAST analysis, and we generated PbKch1-null parasites. Physiological and functional studies with these null parasites suggest that PbKch1 not only mediates K⁺ uptake in the erythrocytic stages of the parasite but also is critical for the development of the mosquito midgut oocyst stage of the parasite, thus directly implicating its functional significance during the malaria transmission process.

Results and Discussion

Identification of the Orthologue of PfKch1 in P. berghei.

PbKch1, the P. berghei orthologue of PfKch1, was identified by BLAST search of the P. berghei genome database (www.plasmodb.org). The amino acid identity was 82% between the putative six transmembrane domains of the two orthologues (90% similarity), and in particular the pore loops were completely identical (amino acid sequence N-DFVYFGVITMSTVGYGDYTP-C) (Fig. 1A). K⁺ channel proteins with identical pore-loop sequences were identified by BLAST search in several other Plasmodium species, and they all shared ≈90% amino acid sequence similarity with respect to the putative six transmembrane domains of PfKch1 (data not shown). There were no close homologues to PfKch1 outside of the Plasmodium genus, which indicates that this particular K⁺ channel is unique to Plasmodium species.

Fig. 1.

Sequence analysis of the Plasmodium K⁺ channels. (A) Alignment of the P. berghei putative K⁺ channel PbKch1 and the orthologue P. falciparum K⁺ channel PfKch1. The ORFs of PbKch1 and PfKch1 encode 986 and 1,940 aa, respectively. Identical or similar amino acids are colored red. The channels belong to the superfamily of six transmembrane-segment K⁺ channels. Except for the S4 segment, the membrane topology of the channel was predicted by the TMHMM algorithm. The S4 segment of six transmembrane-segment K⁺ channels is notoriously difficult to predict because of a number of charged residues, usually arginines, and this particular segment was therefore identified by eye. The amino acid identity in the parts of the proteins comprising the six transmembrane domains (indicated by gray boxes) is 82%. The amino acids of the pore regions are identical. (B) Alignment of the S5–S6 linker of PfKch1 and PbKch1 with corresponding linkers from different hypothetical and known K⁺ channels. The pore loop contains the canonical K⁺ channel signature sequence ((T/S)XXTXGYG). B. bovis, C. hominis, and T. annulata: Hypothetical proteins from Babesia bovis [Protein Data Bank (PDB) ID code XP_001610013], Cryptosporidium hominis (PDB ID code XP_668687), and Theileria annulata (PDB ID code XP_952459). nBK, Caenorhabditis elegans BK channel (PDB ID code NP_001024261); hBK, human BK channel (PDB ID code NP_001014797); hShaker, human voltage-gated shaker-related K⁺ channel (PDB ID code NP_002223); hKCNQ1, human cardiac KCNQ channel (PDB ID code CAO03369). Amino acids in the K⁺ channel signature sequence are highlighted in bold; amino acid identity with PfKch1 is colored gray.

Many K⁺ channel blockers bind in or near the outer pore mouth, which is mainly contributed by the pore loop. The amino acid composition of the pore loop, therefore, constitutes a major molecular determinant for pharmacological K⁺ channel inhibition. The pore loop of PfKch1 shared 70–80% amino acid sequence identity with pore loops of putative K⁺ channels from other primitive eukaryotic organisms (Fig. 1B). The pore loops of different human K⁺ channels all shared <60% sequence identity with the pore loop of PfKch1, which suggests that it may be possible to identify specific pharmacological inhibitors of the Plasmodium K⁺ channel 1 homologue.

Targeted Disruption of the PbKch1 Gene.

To study the biological role of PbKch1, we generated P. berghei parasites in which the PbKch1 locus was disrupted by homologous recombination (Fig. 2A). Two fragments from the 5′ and 3′ flanking regions of the putative six transmembrane domains of PbKch1 were cloned on either side of the Toxoplasma gondii dihydrofolate reductase (TgDHFR) cassette in the vector pB3D. The resultant plasmid (pB3DPbKch1) was introduced into mature schizonts of P. berghei by the Amaxa electroporation procedure. Transfected parasites were immediately introduced into mice, and the mice were treated with pyrimethamine to select for recombinant parasites, which were further cloned by limiting dilution.

Fig. 2.

Disruption strategy for generation of PbKch1-null parasites. (A) (Upper) The genomic locus. (Lower) The disrupted gene. The black line represents the locus (PDB ID code PB_RP3798) containing the gene encoding PbKch1 (bar). The deduced six transmembrane domains (indicated by the small interconnected bars below the gene) of PbKch1 are encoded by the core region. The 5′ and 3′ regions are the recombination sites used to replace, by double crossover, the core region of the gene with the Tg-DHFR ORF (gray color). The Tg-DHFR confers resistance to the compound pyrimethamine, which can be used to select PbKch1-null parasites. (B) Validation of the PbKch1-null genotype. Gene-specific primers amplify different PCR products from DNA purified from WT P. berghei and PbKch1-null P. berghei. Fragments 1 and 2 (same as in A) can be amplified only from WT parasites, not from PbKch1-null parasites. The primers used to generate fragments 3 and 4 (same as in A) target the Tg-DHFR ORF from outside the recombination sites. These fragments can be amplified only from PbKch1-null parasites, not from WT parasites, nor from parasites containing circular (nongenomic) copies of the targeting plasmid. Lanes 1–4: PCR products amplified from WT parasite DNA. The presence of fragments 1 (1,022 bp) and 2 (667 bp), but not 3 and 4, confirms the undisrupted genotype. Lanes 5–8: PCR products amplified from PbKch1-null parasite DNA. The presence of fragments 3 (968 bp) and 4 (625 bp), but not 1 and 2, confirms that the core region has been correctly replaced by the Tg-DHFR ORF, and that there is no contamination from WT parasite DNA. Standard (St) lanes contain BstEII-digested Escherichia coli λ-phage DNA with base-pair length of selected markers shown to the right. (C) RT-PCR with RNA from WT parasites (lanes 1–3) and PbKch1-null parasites (lanes 4–6). Primer pairs were the same as those used above to assess the presence of the uninterrupted PbKch1 gene and the interrupted gene where the core region of PbKch1 has been replaced by the Tg-DHFR ORF. Fragment 2 (667 bp) could be amplified from WT RNA (lane 1), but not from PbKch1-null RNA (lane 4). Fragment 3 (968 bp) could be amplified from PbKch1-null RNA (lane 5), but not from WT RNA (lane 2). PCR amplification not preceded by the reverse-transcriptase step gave no bands from WT RNA (lane 3) or PbKch1-null RNA (lane 6), which confirms the absence of genomic DNA contamination in the RNA preparations. Standard (St) marker is as mentioned in B.

To determine whether the introduced plasmid had integrated into the parasite genome, genomic DNA from the WT parasite and the cloned transfectant parasite were analyzed by PCR using integration-specific primers. Primers targeting the undisrupted ORF produced bands with WT DNA, but no bands were visible with DNA from the PbKch1-null parasites (Fig. 2B). On the other hand, primer pairs that targeted the DHFR ORF and either the 5′ or the 3′ flanking regions produced bands with DNA from the PbKch1-null parasites, but not from WT parasites. These findings confirmed replacement of the targeted sequence of the PbKch1 ORF by insertion of the DHFR ORF in PbKch1-null parasites DNA. Lack of expression of PbchK1 in the null parasites was also confirmed by RT-PCR using purified RNA from WT and PbKch1-null parasites (Fig. 2C).

K⁺ Uptake Kinetics in WT and PbKch1-Null Parasites.

In most eukaryotic organisms, K⁺ channels contribute to the maintenance of the resting membrane potential. In Plasmodium parasites, the membrane potential is probably created through the action of the plasma membrane proton pump, which extrudes protons from the parasite's cytoplasm (13, 14). The thereby generated highly negative membrane potential likely facilitates a net K⁺ uptake into the parasites (15), and we hypothesized that PbKch1, at least in part, mediates this K⁺ uptake.

To investigate whether disruption of PbKch1 affected K⁺ transport, we compared the uptake of the K⁺ congener ⁸⁶Rb⁺ in PbKch1-null parasites with the uptake in WT parasites. ⁸⁶Rb⁺ added to a suspension of Plasmodium-infected red blood cells enters the red blood cell cytoplasm via the parasite-induced, broad-specificity, new permeation pathways (16), and subsequently accumulates in the parasites because of their high negative membrane potential (15, 17). In this study, infected erythrocytes were treated with saponin to allow functional access to the parasite's plasma membrane (18). The saponin-isolated parasites were incubated with ⁸⁶Rb⁺, and the intracellular accumulation was followed for up to 30 min. The uptake data were fitted to a one-phase exponential association, and our results show that the uptake in PbKch1-null parasites was significantly decreased compared with the uptake in WT parasites (Fig. 3A). K⁺-uptake rates for the initial 1½ min were calculated from ⁸⁶Rb⁺ uptake, and for WT parasites it was 5.5 ± 1.2 × 10⁻⁹ mol K⁺/(10⁹ cell × min) [95% confidence interval (CI): 4.3–6.7 × 10⁻⁹]. This value is lower than an earlier reported K⁺-uptake rate for intact P. falciparum-infected erythrocytes (16). The difference is most likely because the values reported here were determined for a mixed population of parasites, whereas the values reported in ref. 16 were assessed for late-stage parasites only. Erythrocytes infected with late-stage parasites are known to have a higher K⁺-uptake rate than erythrocytes infected with early-stage parasites (16). The K⁺-uptake rate for PbKch1-null parasites was 1.6 ± 2.5 × 10⁻⁹ mol K⁺/(10⁹ cell × min) (95% CI: 0–4.1 × 10⁻⁹), three to four times lower than for WT parasites. Theoretically, a lower uptake rate in PbKch1-null parasites could be caused by a maturation arrest in those parasites, resulting in early immature stage parasites with low uptake rates only. However, successful production of PbKch1-null parasites clearly demonstrates that they developed as erythrocytic asexual and sexual stages (see below) similar to those in the WT parasites.

Fig. 3.

Characterization of ⁸⁶Rb⁺ transport in the PbKch1-null parasite. (A) Uptake of ⁸⁶Rb⁺ in isolated P. berghei parasites. Erythrocytes from mice infected with WT or PbKch1-null parasites were treated with saponin, thereby allowing functional access to the parasite's plasma membrane. The saponin-treated parasites were incubated with trace amounts of ⁸⁶Rb⁺, and samples were taken at the indicated time points. PbKch1-null parasites (triangles) showed decreased uptake compared to WT parasites (squares). Each point represents mean value ± SEM (n = 3). (B) Pharmacological characterization of ⁸⁶Rb⁺ uptake in isolated parasites. Uptake was assessed at a fixed time point (t = 10 min) in the presence of either suspending buffer alone (control), 10 mM Ba²⁺, 20 mM TEA, or 0.5 mM quinine. The height of each bar represents mean value ± SEM (n = 4). When denoted by *, ⁸⁶Rb⁺ uptake was statistically different from ⁸⁶Rb⁺ uptake in untreated WT parasites (P < 0.05, unpaired t test).

The pharmacological profile of the parasites ⁸⁶Rb⁺ uptake was further investigated with K⁺ channel blockers. Isolated parasites were incubated in the presence of K⁺ channel blockers before the addition of ⁸⁶Rb⁺, and parasite samples were analyzed at a fixed time point (t = 10 min). Prior treatment with 10 mM Ba²⁺ or 0.5 mM quinine resulted in ≈80% inhibition of uptake of ⁸⁶Rb⁺ in WT parasites as compared with parasites pretreated with solvents only (control) (Fig. 3B). When tested in parallel, PbKch1-null parasites revealed ⁸⁶Rb⁺ uptake to the level detected as residual uptake in WT parasites upon treatment with Ba²⁺ or quinine. Furthermore, this already reduced uptake of ⁸⁶Rb⁺ in PbKch1-null parasites was insensitive to further inhibition by Ba²⁺ and quinine. Taken together, these results strongly suggest that PbKch1 is the predominant K⁺ channel in the erythrocytic stages of P. berghei parasites. The residual uptake seen in WT parasites in the presence of Ba²⁺ or quinine, and in the PbKch1-null parasites, is most likely mediated by other membrane transporters, e.g., cotransporters, present in the plasma membrane of the parasites. In addition, another putative K⁺ channel has been identified in P. falciparum (11), and this channel is highly conserved among members of the Plasmodium genus (unpublished observation). It cannot, a priori, be excluded that this channel could mediate Ba²⁺- and quinine-insensitive ⁸⁶Rb⁺ fluxes. The nonspecific K⁺ channel blocker tetraethylammonium (TEA; 20 mM) had a small and nonsignificant inhibitory effect on ⁸⁶Rb⁺ uptake in WT parasites, suggesting that PbKch1 is insensitive to TEA. In a recent study (19), it has been shown that exposure of P. berghei sporozoites to intracellular concentrations of K⁺ enhances their infectivity by 8–10 times. This effect of a high K⁺ concentration is significantly attenuated by K⁺ channel blockers, including TEA. The latter observation is not consistent with the lack of TEA inhibition found in the present study, and it cannot be excluded that the effect of TEA reported in ref. 19 can be caused by other actions of TEA than K⁺ channel blockage.

Comparison of Blood-Stage Parasite Growth Kinetics and Virulence.

Next, we investigated in vivo asexual growth kinetics of PbKch1-null and WT parasites. Naïve mice were infected with an equal inoculum of WT and PbKch1-null parasites, and parasitaemia was determined daily. Both groups of mice rapidly progressed in parasitaemia from day 4. In contrast to WT-infected mice, however, parasitaemia in PbKch1-null-infected mice leveled off at a plateau 9 days after infection, before approaching the parasitaemia of WT-infected mice at day 14 (Fig. 4A). This delay in parasitaemia was not statistically significant, but it may provide an explanation for the slightly prolonged survival of PbKch1-null-infected mice compared with WT-infected mice (Fig. 4B). In addition to the asexual parasite stages, normally appearing gametocytes were observed in the blood taken from both groups of mice (data not shown).

Fig. 4.

Phenotype characterization of the PbKch1-null parasite. (A) Growth kinetics of the PbKch1-null parasite. Parasitemia of mice infected with WT parasites (squares) or PbKch1-null parasites (triangles) is shown. Each point represents mean value ± SEM. (B) Survival of mice infected with WT parasites (stippled curve) or PbKch1-null parasites (continuous curve). Median survival time was 15 days for mice infected with WT parasites and 21 days for mice infected with PbKch1-null parasites (P < 0.001, χ² test).

Evaluation of Malaria Transmission to Anopheles Mosquitoes.

In our quest for functional significance of PbKch1, we extended our analysis to evaluate transmission of malaria parasites from infected vertebrate hosts to anopheline mosquito vector. Adult mosquitoes were fed on PbKch1-null- or WT-infected mice, and 10–12 days after feeding the number of oocysts per mosquito midgut was assessed (Fig. 5). Mosquitoes infected (81 of 89 dissected) with WT parasites revealed a median oocyst burden of 115 per mosquito (25 and 75 percentiles, oocyst numbers 37 and 251). In sharp contrast, only two mosquitoes of 112 fed on PbKch1-null-infected mice had oocysts in their midguts, thus demonstrating a 98% reduction in the infectivity of PbKch1-null parasites during transmission through mosquitoes. Approximately 10 mosquitoes from each group were dissected 24 h after the blood meal, and the presence of ookinetes was validated by light microscopy (data not shown). By this semiquantitative method, no gross difference in ookinete count could be detected between mosquitoes fed on WT- and PbKch1-null-infected mice, thus pointing to a compromised maturation of ookinetes into mature oocysts. Because of the deficient oocyst development we were not able to generate sporozoites from PbKch1-null parasites and directly evaluate liver cell infection by PbKch1-null sporozoites. Others have shown that infectivity of Plasmodium sporozoites can be attenuated by K⁺ channel blockers (19), and we suggest that this effect is most likely caused by the inhibition of PbKch1.

Fig. 5.

Transmission results of the WT and PbKch1-null parasite. Mosquitoes were fed on PbKch1-null infected mice or WT-infetced mice. Of 89 mosquitoes fed on WT-infected mice (n = 3), 81 contained oocysts (115 median oocyst per mosquito). Of 112 mosquitoes fed on PbKch1-null infected mice (n = 3) only two contained oocysts (1 and 22 oocysts, respectively).

Conclusion

Data in the present article show that PbKch1 functions as a K⁺ channel and it mediates K⁺ transport in the intraerythrocytic stages of P. berghei parasites. The function of PbKch1 was not critical for asexual replication of P. berghei. However, the dependency of sexual replication in mosquitoes on PbKch1 implies that PfKch1 may be used as a potential drug target in P. falciparum parasites.

Materials and Methods

Identification and Cloning of PbKch1.

The P. berghei orthologue of PfKch1 (GenBank accession no. NP_701625), PbKch1 (GenBank accession no. XP_676033), was identified by BLAST search of the P. berghei genome database (www.sanger.ac.uk). The alignment was made with the Dialign algorithm (http://bioweb.pasteur.fr/seqanal/interfaces/dialign2-simple.html).

Targeted Disruption of the PbKch1 Gene.

Plasmid pB3DPbKch1 was constructed by cloning fragments from the 5′ and 3′ flanking regions of the deduced six transmembrane domains-encoding part of the PbKch1 gene. The 5′ fragment was PCR-amplified from genomic DNA of P. berghei by using the forward primer B1.L.1548–66 (5′-GGTACCGTAAGAAAGGCAATCAACC-3′) and the reverse primer B1.R.2032–11 (5′-AAGCTTGTTATCTGTTTTTCTTTTATCG-3′). The 3′ fragment was PCR-amplified from genomic DNA of P. berghei by using the forward primer B1.L.3149–70 (GGATCCGAATCCATATTTTTATTTACCC-3′) and the reverse primer B1.R.3602–3585 (GCGGCCGCTTGATCATCCTTTTCCC-3′). The fragments contained sites for KpnI and HindIII at the 5′ and 3′ ends of the 5′ flanking fragment and sites for BamHI and NotI at the 5′ and 3′ ends of the 3′ flanking fragment. These sites were used for cloning of the fragments into the pB3D vector (kindly provided by Andrew Waters, University of Glasgow, Scotland) on either side of the Toxoplasma gondii dihydrofolate reductase (Tg-DHFR) cassette. Before transfection, pB3DPbKch1 was linearized with three restriction enzymes (KpnI, NotI, and ScaI). Schizont-stage parasites of P. berghei 2.34 ANKA strain were transfected by electroporation with ≈25 μg of linearized targeting plasmid by using the Amaxa Nucleofactor device, program T-001 (13). Transfected parasites were immediately injected into female Swiss–Webster mice (4 weeks old), and the mice were treated with pyrimethamine in the drinking water (70 mg/liter) starting 24 h after transfection. Pyrimethamine-resistant parasite populations were detected 7 days after infection in mice. Integration of the plasmid into the PbKch1 locus was confirmed by PCR analysis. Parasite clones were obtained by the method of limiting dilution. Expression of the uninterrupted PbKch1 gene in WT parasites and lack of expression in PbKch1-null parasites was assessed by RT-PCR using poly(A)-selected, DNase-treated RNA from WT parasites and PbKch1-null parasites. As a control for genomic DNA contamination in the RNA preparations, PCR amplification without the reverse-transcriptase step was performed in parallel.

⁸⁶Rb⁺ Transport in PbKch1-Null P. berghei Parasites.

The alkali ion ⁸⁶Rb⁺ was used as a tracer for K⁺ in an uptake protocol. To detect differences in ⁸⁶Rb⁺ transport between the PbKch1-null parasites and WT parasites, parasites were isolated from their red blood cells by saponin treatment. Saponin is a plant detergent that permeabilizes the cholesterol-containing red blood cell membrane and parasitophorous vacuole membrane, but leaves the parasite's plasma membrane intact (18, 20). Saponin-treated infected red blood cells are referred to as isolated parasites hereafter.

Infected red blood cells were treated with 0.05% saponin in RPMI medium 1640 and immediately spun at 2,000 × g for 5 min. The remaining parasite pellet was washed twice at room temperature in RPMI medium 1640, thereby removing hemoglobin and most of the red cell ghosts. Parasites were resuspended in RPMI medium1640 and placed in a water bath at 37°C. Because of the insolubility of BaSO₄, Hepes-buffered saline (125 mM NaCl, 25 mM Hepes, 10 mM glucose, 5.4 mM KCl, 0.4 mM MgCl₂, pH 7.4) was used in the pharmacological studies, which included the presence of Ba²⁺. Cell number was counted in a Neubauer chamber (cell concentration 0.7–4.5 × 10⁸ cells per ml). Microscopy revealed free parasites of all stages, and, only occasionally, erythrocyte ghosts. Trace amounts of ⁸⁶RbCl was added (0.8–2 × 10⁵ Bq/ml, specific activity 2.497 × 10¹⁵ Bq/mol), and at various time points (0–30 min), 200-μl samples were spun (16,000 × g for 20 s) through 0.25-ml oil cushions, consisting of a 5:4 mixture of dibutyl/dioctyl phthalate, overlaid with 0.8 ml of isotonic saline. The cell pellets were lysed with 0.1% Triton X-100, and the protein was precipitated with 5% trichloroacetic acid. After centrifugation (10,000 × g for 10 min), supernatants were transferred to plastic vials and counted in a β-scintillation counter. K⁺-uptake rates were calculated from ⁸⁶Rb⁺ uptake by using the specific activity of the isotope, the calculated ratio [⁸⁶Rb⁺]/[K⁺] in the uptake buffer (≈8.3 × 10⁻⁴), and the counting efficiency of the β-scintillation counter (0.30 cpm/dpm).

Asexual Growth Kinetics and Virulence of PbKch1-Null Parasites.

To detect differences in parasite growth during blood-stage development between the PbKch1-null parasites and WT parasites, 4-week-old female Swiss–Webster mice were infected i.p. with either 10⁵ PbKch1-null parasites or WT parasites. Parasite growth, monitored by Giemsa-stained blood smears, and survival of the mice, was assessed daily.

Mosquito Transmission Differences.

Starved adult (5–6 days old) A. stephensi mosquitoes were allowed to feed on mice infected for 4 days with equal numbers of WT or PbKch1-null parasites (three mice in each group). Blood-fed mosquitoes were then maintained at 20°C in insectary chambers at 70–80% relative humidity. Ten to 12 days after blood feeding, mosquito midguts were dissected and oocyst numbers were enumerated after staining with 0.1% mercurochrome.