|Home | About | Journals | Submit | Contact Us | Français|
Development of malaria parasites within vertebrate erythrocytes requires nutrient uptake at the host cell membrane. The plasmodial surface anion channel (PSAC) mediates this transport and is an antimalarial target, but despite its importance, its molecular basis has been unknown. We now report a parasite gene family responsible for PSAC activity. We performed high-throughput screening to find transport inhibitors specific for distinct lines of the human pathogen P. falciparum. One inhibitor, 800-fold more active against PSAC from the Dd2 line than from HB3 parasites, was used with a genetic cross to map a single parasite locus on chromosome 3. DNA transfection and in vitro selections indicate that PSAC-inhibitor interactions are determined by two clag genes previously assumed to function in cytoadherence. These genes are conserved in plasmodia, exhibit expression switching, and encode an integral protein on the host membrane, as predicted by functional studies. This protein establishes novel ion channel activity on the erythrocyte surface.
Successful intracellular development of malaria parasites depends on directed remodeling of the host erythrocyte cytosol and plasma membrane. For example, P. falciparum and all studied malaria parasites increase erythrocyte permeability to numerous solutes including nutrients required for intracellular parasite growth. Sugars, amino acids, purines, vitamins, and the cation choline are needed for in vitro propagation and have increased permeability after infection (Homewood and Neame, 1974; Ginsburg et al., 1985; Upston and Gero, 1995; Saliba et al., 1998; Staines et al., 2000). Anions and antimalarial toxins also have increased uptake (Kirk et al., 1994; Hill et al., 2007; Lisk et al., 2008). Intriguingly, the parasite nevertheless maintains a low Na+ permeability at the host membrane to avoid osmotic lysis in the bloodstream (Cohn et al., 2003).
New antimalarial drugs are desperately needed. Conservation of erythrocyte permeability changes in divergent malaria parasites suggests that the responsible mechanism may be a suitable drug target (Lisk and Desai, 2005). Interest has been further increased by the recognition that the parasite can alter the permeability to acquire resistance to antimalarials (Hill et al., 2007; Lisk et al., 2008).
Despite its central role in parasite biology and therapeutic potential, the origin and precise nature of the uptake mechanism have been controversial. Electrophysiological studies over the past decade have implicated transmembrane transport through one or more ion channels on the erythrocyte surface (Desai et al., 2000; Staines et al., 2007). While there is agreement that these channels are anion-selective and that they appear on the host membrane some hours after parasite invasion, the genes responsible for each proposed ion channel remain unknown. Some studies have proposed activation of quiescent human channels intrinsic to the erythrocyte (Decherf et al., 2004; Duranton et al., 2004; Verloo et al., 2004), while others support one or more parasite-encoded proteins trafficked to the host membrane. The plasmodial surface anion channel (PSAC) is the primary candidate for a parasite-encoded ion channel (Alkhalil et al., 2004), but computational analyses of completed parasite genomes have not revealed orthologs of classical anion channels.
We have now used a genetic approach to identify a parasite gene family involved in PSAC activity. Our studies combine high-throughput screens for novel channel inhibitors, electrophysiology revealing direct action on PSAC, linkage analysis in a P. falciparum genetic cross, DNA transfection experiments, in vitro selections, and mutant analysis to demonstrate the role of two paralogous genes in PSAC activity. We propose that these genes contribute to a novel microbial ion channel for the uptake of diverse nutrient solutes.
We searched for small molecule inhibitors with differing efficacies against channels induced by divergent parasite lines. Such inhibitors presumably bind to one or more variable sites on the channel, which may result either from polymorphisms in a parasite channel gene or from differing activation of human channels. To find these inhibitors, we used a transmittance-based assay that tracks osmotic lysis of infected cells in sorbitol, a sugar alcohol with increased permeability after infection (Wagner et al., 2003). This assay has previously been adapted to 384-well format and used to find high affinity PSAC inhibitors (Pillai et al., 2010). Here, we used this format to screen a library of compounds against erythrocytes infected with the HB3 and Dd2 P. falciparum lines. To maximize detection of hits, we chose a low stringency in our screens by using library compounds at a high concentration (10 μM) and by reading each microplate at multiple timepoints (Pillai et al., 2010). Figure 1A tallies the all-points histogram for compound activity averaged from screens of both lines. 8% of compounds met or exceeded the threshold of 50% normalized block at 2 h (Eq. S2), consistent with a low screening stringency. We then defined a weighted difference statistic (WDS) that normalizes measured differences in efficacy against HB3 and Dd2 channels to the standard deviation of positive control wells in each microplate (Eq. S3). An all-points histogram of this parameter revealed that 86% of all compounds produced indistinguishable effects on the two parasite lines (WDS ≤ 1.0, Figure 1B). Thus, most inhibitor binding sites are conserved.
Nevertheless, a small number of compounds produced significantly differing activities in the two screens. One such inhibitor, named ISPA-28 (for isolate-specific PSAC antagonist based on studies described below, Figure 1C), was reproducibly more effective at inhibiting sorbitol uptake by Dd2- than HB3-infected cells. Secondary studies with ISPA-28 revealed an ~ 800-fold difference in half-maximal affinities (Figures 1D and 1E, K0.5 values of 56 ± 5 nM vs. 43 ± 2 μM for Dd2 and HB3, respectively; P < 10−10).
We also examined ISPA-28 effects on uptake of the amino acids alanine and proline as well as the organic cation phenyl-trimethylammonium (PhTMA), solutes with known increases in permeability after infection (Ginsburg et al., 1985; Bokhari et al., 2008). Each solute’s permeability was inhibited with dose responses matching those for sorbitol (Figure 1E), providing strong evidence for a single shared transport mechanism used by these diverse solutes.
We next tested 22 different laboratory parasite lines and found significant transport inhibition with only Dd2 and W2 (Figure S1). Because Dd2 was generated by prolonged drug selections starting with W2 (Wellems et al., 1988), their channels’ distinctive ISPA-28 affinities suggest a stable heritable element in the parasite genome.
To explore the mechanism of ISPA-28 block, we performed patch-clamp of infected erythrocytes. Using the whole-cell configuration, we observed similar currents on HB3- and Dd2-infected cells in experiments without known inhibitors (Figures 2A and 2B). These currents exhibited inward rectification. Previous studies have determined that they are carried primarily by anions with a permeability rank order of SCN− > I− > Br− > Cl− (Desai et al., 2000). 10 μM ISPA-28 reduced their magnitudes, but had a significantly greater effect on Dd2-infected cells (P < 10−9, n = 22-28 cells each). In the cell-attached configuration with 1.1 M Cl− as the charge carrier, ion channel activity characteristic of PSAC was detected on both lines (~20 pS slope conductance with fast flickering gating, Alkhalil et al., 2004); without inhibitor, channels from the two lines were indistinguishable (Figure 2C). However, recordings with 10 μM ISPA-28 revealed a marked difference as Dd2 channels were near-fully inhibited whereas HB3 channels were largely unaffected (Figures 2D and 2F; n = 6-7 channel molecules each; P < 0.0002). Thus, this compound’s effects on single PSAC recordings parallel those on uptake of sorbitol and other organic solutes.
We analyzed closed durations from extended recordings and determined that ISPA-28 imposes a distinct population of long block events, but only in recordings on Dd2-infected cells (Figure S2). At the same time, intrinsic channel closings, which occur in the absence of inhibitor, were conserved on both parasites and were not affected by ISPA-28 (events of < 20 ms duration in Figure S2).
We next examined ISPA-28 efficacy against PSAC activity on red blood cells infected with recombinant progeny clones from the Dd2 × HB3 genetic cross (Wellems et al., 1990). For each clone, we examined sorbitol uptake in the absence and presence of 7 μM ISPA-28, a concentration that optimally distinguishes the parental channel phenotypes, and quantified inhibition (Eq. S2). Although a few of the 34 independent progeny clones exhibited intermediate channel inhibition, most resembled one or the other parent (Figure 3A). Quantitative trait locus (QTL) analysis was used to search for associations between ISPA-28 efficacy and inheritance of available microsatellite markers. A primary scan identified a single significant peak having a logarithm of odds (LOD) score of 12.6 at the proximal end of chromosome 3 (Figure 3B). A secondary scan for residual effects did not find additional peaks reaching statistical significance (Figure S3).
The mapped locus contains 42 predicted genes (Table S1). Although none have homology to classical ion channels from other organisms, many are conserved in other plasmodia, as expected for the responsible gene(s) from conservation of PSAC activity in malaria parasites (Lisk and Desai, 2005). The mapped region is enriched in genes encoding proteins destined for export to host cytosol (P < 10−4 by simulation), as typical of apicomplexan subtelomeric regions. Some of the encoded proteins have one or more predicted transmembrane domains as usually involved in channel pore formation, but this criterion may miss some transport proteins. The PEXEL motif, which directs parasite proteins to the host cell (Marti et al., 2004), was present in some genes, but this module is not universally required for export (Spielmann and Gilberger, 2010). Thus, computational analyses suggested several candidates, but could not specifically implicate any as ion channel components.
We therefore selected a DNA transfection approach and chose piggyBac transposase to complement Dd2 parasites with the HB3 allele of individual candidate genes (Balu et al., 2005). This method has the disadvantage that successfully transfected parasites will carry both parental alleles and therefore be merodiploid for candidate genes. Nevertheless, we reasoned that the marked difference in ISPA-28 efficacy between the parental lines should produce a detectable change in transport phenotype upon complementation with the responsible gene. Importantly, the high efficiency of random integration conferred by piggyBac permits rapid examination of many genes (Balu et al., 2009).
We cloned 14 genes with their endogenous 5′ and 3′ UTR regions from the HB3 parent into the pXL-BacII-DHFR plasmid; a 15th construct containing a conserved but not annotated open reading frame (ORF 147 kb) was also prepared. Each was transfected individually along with a helper plasmid encoding the transposase into Dd2 parasites (Figure 4A). Selection for hDHFR expression yielded parasites that stably carry both Dd2 and HB3 alleles for each candidate. Because an altered channel phenotype presumably requires expression of the HB3 allele, we used reverse transcriptase PCR to amplify polymorphic regions of each gene and sequenced the amplicons to determine if both parental alleles were transcribed; this approach confirmed expression of 12 candidates (Figure 4B, asterisks). We then performed ISPA-28 dose responses for inhibition of sorbitol uptake by erythrocytes infected with each transfectant (Figures 4B and 4C). Two transfectants, expressing HB3 alleles for PFC0110w and PFC0120w, produced significant changes in ISPA-28 efficacy with K0.5 values between those of Dd2 and HB3, as expected for cells carrying channels from both parental lines (P = 0.01 and P < 10−7 in comparison to Dd2, respectively). Limiting dilution cloning of the PFC0120w transfectant yielded a clone, Dd2-pB120w, which had undergone at least one integration event (Figure S4); its ISPA-28 K0.5 was indistinguishable from the transfection pool (not shown). For both genes, quantitative analyses suggest relatively low level expression of the HB3 allele because the transfectant K0.5 values (95 ± 8 and 140 ± 12 nM) are closer to those of Dd2 than of HB3. Expression levels of the two parental alleles may be influenced by the genomic environment of the integration site, relative promoter efficiencies, and a gene silencing mechanism examined below.
PFC0120w and PFC0110w are members of the clag multigene family conserved in P. falciparum and all examined plasmodia; the family’s name is based on proposed roles as cytoadherence-linked antigens. These paralogs are known as clag3.1 and clag3.2, respectively, based on their location on chromosome 3; P. falciparum carries three additional paralogs on chromosomes 2, 8, and 9.
To examine the unexpected possibility that clag3 products contribute to PSAC activity, we used an allelic exchange strategy designed to transfer potent ISPA-28 block from the Dd2 line to HB3 parasites. Because Dd2 parasites express clag3.1 but not clag3.2 (Kaneko et al., 2005; see also next section), their clag3.1 gene presumably encodes high ISPA-28 affinity. We therefore constructed a transfection plasmid carrying a 3.2 kb fragment from the 3′ end of the Dd2 clag3.1 allele, an in-frame C-terminal FLAG tag followed by a stop codon, and the fragment gene’s 3′ untranslated region (pHD22Y-120w-flag-PG1, Figure 5A). Because this plasmid carries only a gene fragment and lacks a leader sequence to drive expression, an altered transport phenotype requires recombination into the parasite genome. We transfected HB3 with this plasmid and used PCR to screen for integration into each of the five endogenous clag genes. This approach detected recombination into the HB3 clag3.2 gene; limiting dilution cloning yielded HB33rec, a clone carrying a single site integration event without residual episomal plasmid (Figures S5A and S5B). DNA sequencing indicated recombination between single nucleotide polymorphisms at 3718 and 4011 bp from the HB3 clag3.2 start codon. This recombination site corresponds to successful transfer of downstream polymorphisms including a recognized hypervariable region at 4266-4415 bp; contamination with other laboratory parasite lines was excluded by fingerprinting (Figure S5C).
PSAC activity on HB33rec exhibited a marked increase in ISPA-28 efficacy (Figure 5B), further supporting a role for clag3 genes in sorbitol and nutrient uptake. Although this allelic exchange strategy yielded a gene replacement in contrast to the complementations achieved with piggyBac, the channel’s ISPA-28 affinity was again intermediate between those of HB3 and Dd2 (Figure 5C). Several mechanisms may contribute to the quantitatively incomplete transfer of inhibitor affinity. First, two or more polymorphic sites on the protein might contribute to ISPA-28 binding. If some of these sites are upstream from the recombination event, the resulting chimeric protein may have functional properties distinct from those of either parental line. Second, the channel may contain additional unidentified subunits; here, transfection to replace each contributing HB3 gene with Dd2 alleles might be required to match the ISPA-28 affinity of Dd2. Finally, in addition to the chimeric clag3.2HB3-3.1Dd2 gene produced by transfection, HB33rec also carries the clag3.1 gene endogenous to HB3 parasites. Expression of both paralogs could also produce an intermediate ISPA-28 affinity.
To explore these possibilities, we performed cell-attached patch-clamp on HB33rec-infected cells. Individual channel molecules exhibiting ISPA-28 potencies matching those of each parental line were identified (Figure S5, panels D-F). These recordings exclude scenarios that require a homogenous population of channels with intermediate ISPA-28 affinity but are consistent with several other mechanisms.
In addition to the complex behavior of HB33rec, we noticed that certain progeny from the genetic cross had lower ISPA-28 affinity than Dd2 despite inheriting the mapped chromosome 3 locus fully from the Dd2 parent (notably 7C20, 7C12, and CH361 in Figure 3A and dose responses in Figure 6A). Because subtelomeric multigene families in P. falciparum are susceptible to recombination and frequent gene conversion events (Freitas-Junior et al., 2000), we sequenced both clag3 paralogs and neighboring genomic DNA from 7C20 and Dd2 but did not find DNA-level differences. We therefore considered epigenetic mechanisms that may influence ISPA-28 affinity. clag3.1 and clag3.2 have been reported to undergo mutually exclusive expression (Cortes et al., 2007). Monoallelic expression and switching, also documented for other gene families in P. falciparum (Chen et al., 1998; Lavazec et al., 2007), allows individual parasites to express a single member of a multigene family. Daughter parasites resulting from asexual reproduction continue exclusive expression of the same gene through incompletely understood epigenetic mechanisms (Howitt et al., 2009). After a few generations, some daughters may switch to expression of another member of the gene family, affording diversity that contributes to immune evasion (Scherf et al., 2008).
We performed reverse transcriptase PCR and found that Dd2 expresses clag3.1 almost exclusively while the three discordant progeny express clag3.2 at measurable levels (Figure 6B), suggesting epigenetic regulation. We therefore applied selective pressure to progeny cultures with osmotic lysis in sorbitol solutions containing ISPA-28. Inclusion of ISPA-28 preferentially spares infected cells whose channels have high inhibitor affinity: these cells incur less sorbitol uptake and do not lyse (Figure 1D). These selections, applied on multiple consecutive days, yielded marked reductions in parasitemia. Surviving parasites exhibited improved ISPA-28 affinity quantitatively matching that of the Dd2 parent (Figure 6C). Identical selections applied to HB3 and three progeny inheriting its chromosome 3 locus did not change ISPA-28 affinity (Figure 6D, n = 3-4 separate attempts each; data for progeny clones QC13, SC01, and7C421 not shown), excluding effects of the selections on unrelated genomic sites.
Real time qPCR using primers specific for each of the 5 clag genes revealed that selection with sorbitol and ISPA-28 reproducibly increased clag3.1 expression while decreasing that of clag3.2 in progeny inheriting the Dd2 locus (Figure 6E; expression ratios in Figure 6F). Selections applied to the parental HB3 line were without effect (Figure 6F), consistent with its unchanged inhibitor affinity. Although these selections did not alter relative expression of other paralogs (clag2, clag8, and clag9), we cannot exclude their possible contributions to PSAC activity.
Selections were also applied to HB33rec, which carries a chimeric clag3.2HB3-3.1Dd2 transgene and the clag3.1 gene native to HB3. In contrast to the lack of effect on the isogenic HB3 line, these synchronizations increased the transfectant’s ISPA-28 affinity to a K0.5 of 51 ± 9 nM, matching that of Dd2 channels (Figures 6D and 6G). This change in channel phenotype correlated with a near exclusive expression of the transgene (Figure 6H, first timepoint), confirming that expression of HB3 clag3.1 by a subset of cells accounts for the intermediate ISPA-28 affinity observed in Figure 5E. These findings also delimit the determinants of ISPA-28 binding to polymorphic sites within the Dd2 clag3.1 gene fragment transferred to HB33rec.
Expression switching in P. falciparum multigene families occurs over several generations and should lead to a drift in population phenotype. After selection of the chimeric gene in HB33rec, continued in vitro propagation yielded a gradual decay in ISPA-28 affinity that correlated with decreasing transgene expression (Figures 6G and 6H). As with other multigene families (Lavazec et al., 2007), several factors may affect the steady-state ISPA-28 affinity and relative expression levels for the two clag3 genes upon continued culture without selective pressure.
We next sought a PSAC inhibitor with reversed specificity for the two Dd2 clag3 products. To this end, we surveyed hits from our high-throughput screen using the progeny clone 7C20 before and after selection for clag3.1 expression. This secondary screen identified ISPA-43 as a PSAC inhibitor with an allele specificity opposite that of ISPA-28 (Figures (Figures6I6I and S6B, K0.5 of 32 and 3.9 μM for channels associated with clag3.1 and clag3.2 genes from Dd2, respectively). We then applied sorbitol synchronizations with 4 μM ISPA-43 to the clag3.1-expressing 7C20 culture and achieved robust reverse selection: the surviving parasites exhibited both low ISPA-28 affinity and a reversed clag3 expression profile (Figures 6J and 6K). Thus, inhibitors can be used in purifying selections of either clag3 gene. Because ISPA-28 affinity can be reduced either through drift without selective pressure or by selection for the alternate paralog with an inhibitor having reversed specificity, these studies mitigate concerns about indirect effects of exposure to sorbitol or individual inhibitors.
A stable parasite mutant with altered PSAC selectivity, gating, and pharmacology was recently generated by in vitro selection of HB3 with leupeptin (Lisk et al., 2008). We sequenced clag3 genes from this mutant, HB3-leuR1, and identified a point mutation within its clag3.2 gene that changes the conserved A1210 to a threonine (Figures (Figures7A7A and S7A), consistent with a central role of clag3 genes in solute uptake. HB3-leuR1 silences its unmodified clag3.1 and preferentially expresses the mutated clag3.2 (expression ratio of 19.2 ± 1.5, determined as in Figure 6F), as required for a direct effect on PSAC behavior. This mutation is within a predicted transmembrane domain and may directly account for the observed changes in channel gating and selectivity. Because HB3-leuR1 may carry additional genome-level changes, further studies will be needed to determine the precise role of the A1210T mutation in leupeptin resistance.
To directly contribute to PSAC activity, at least some of the clag3 product must associate with the host membrane, presumably as an integral membrane protein. We therefore raised polyclonal antibodies to a carboxy-terminal recombinant fragment conserved between the two clag3 products. Confocal microscopy with this antibody confirmed reports localizing these proteins to the host cytosol and possibly the erythrocyte membrane as well as within rhoptries of invasive merozoites (Vincensini et al., 2008; Figure S7B). To obtain more conclusive evidence, we used immunoblotting to examine susceptibility of these proteins to extracellular protease. Without protease treatment, a single ~160 kDa band was detected in whole-cell lysates, consistent with the expected size of clag3 products (Figure 7B). Treatment with pronase E under conditions designed to prevent digestion of intracellular proteins reduced the amount of the full-length protein and revealed a 35 kDa hydrolysis fragment. In contrast, a monoclonal antibody against KAHRP, a parasite protein that interacts with the host membrane cytoskeleton but is not exposed (Kilejian et al., 1991), confirmed that intracellular proteins are resistant to hydrolysis under our conditions (Figure 7C). As reported for another protease (Baumeister et al., 2006), pronase E treatment significantly reduced PSAC-mediated sorbitol uptake (Figure 7D); this effect was sensitive to protease inhibitors, suggesting that proteolysis at one or more exposed sites interferes with transport.
Ultracentrifugation of infected cell lysates revealed that the clag3 product is fully membrane-associated (Figure 7E); a fraction could however be liberated by treatment with Na2CO3, which strips membranes of peripheral proteins (Fujiki et al., 1982). Because this fraction was protease insensitive, it reflects an intracellular pool of clag3 product loosely associated with membranes. The C-terminal hydrolysis fragment was present only in the carbonate-resistant insoluble fraction, indicating an integral membrane protein.
Because our polyclonal antibodies might cross-react with clag products from other chromosomes, we next examined protease sensitivity in HB33rec, whose chimeric clag3 transgene encodes a C-terminal FLAG tag. Anti-FLAG antibody recognized a single integral membrane protein in HB33rec and no proteins from the parental HB3 line, indicating specificity for the recombinant gene product (Figure 7F). Treatment with pronase E prior to cell lysis and fractionation revealed a hydrolysis fragment indistinguishable from that seen with the antibody raised against the native protein’s C-terminus.
We have used high-throughput screening to identify ISPA-28, an inhibitor with remarkable specificity for solute uptake by erythrocytes infected with Dd2 parasites. Patch-clamp studies revealed that this compound acts specifically on PSAC. Because its activity is restricted to one parasite lineage (Figure S1), ISPA-28 addresses previous concerns regarding inhibitor specificity and implicates a central role for PSAC in anion and nutrient transport at the host membrane. We then tracked inheritance of ISPA-28 efficacy in the Dd2 × HB3 genetic cross and identified a single parasite genomic locus. Gene complementation, allelic exchange, in vitro selections using ISPA-28, and a point mutation in a known PSAC mutant implicate two clag genes from this locus.
A transport role for members of the clag gene family is unexpected because the encoded proteins already have proposed roles unrelated to transport. The first clag gene was identified via a chromosome 9 deletion event linked to loss of infected erythrocyte binding to melanoma cells (Schmidt et al., 1982; Day et al., 1993). clag9 is present in the deletion locus and appears to function in infected cell cytoadherence to the endothelial receptor CD36 (Trenholme et al., 2000). Other clag genes, identified through the P. falciparum genome sequencing project, were presumed to serve related functions. Proposals have included infected cell binding to various endothelial receptors (Holt et al., 1999), merozoite invasion of erythrocytes (Rungruang et al., 2005; Kaneko, 2007), or as a chaperone for protein export (Craig, 2000; Vincensini et al., 2008; Goel et al., 2010). The novel role for clag3 products in solute transport requires a broader view for functions of this gene family; this role is compatible with two previous observations. First, isolated disruption of clag9 abolishes melanoma cell binding (Trenholme et al., 2000), suggesting that other clag products cannot compensate for loss of that member. Second, mutually exclusive expression of clag3.1 and clag3.2 despite constitutive expression of other family members has been reported (Cortes et al., 2007), suggesting that clag3 products may serve roles requiring greater protection from immune challenge. Our transport-based in vitro selections confirm this atypical gene regulation: while purifying selection for one or the other clag3 gene was readily achieved, significant changes in expression of other members were not detected (Figure 6).
The encoded proteins share a conserved alpha-helical CLAG domain also present in RON2, a rhoptry neck protein implicated in erythrocyte invasion (Richard et al., 2010; Anantharaman et al., 2007). While clag genes are present only in plasmodia, RON2 is more widely distributed in apicomplexan parasites and appears to be the ancestral member (Figure S6A). Phylogenetic analysis suggests that after divergence from RON2, clag genes underwent expansion and split into two distinct groups early in the Plasmodium lineage. One group, typified by clag9 in P. falciparum, contains a single member in each plasmodium, while the second group has from one to four members in each species. Interestingly, both RON2 and the clag products are packaged into rhoptries, organelles present in the extra-erythrocytic merozoite (Kaneko et al., 2001); they are then secreted into the host cell upon invasion (Ling et al., 2004; Vincensini et al., 2008).
Recognizing that clag3 encodes an integral protein at the host erythrocyte membrane, we envision two models for how the protein may contribute to channel activity. In one, the clag3 product alone forms a novel microbial nutrient and ion channel (Figure 7G, left diagram). In the other, the clag3 product interacts with one or more other proteins to form functional PSAC (Figure 7G, right diagram). Multi-subunit channels are well established with individual components contributing to solute recognition and permeation, channel gating, inhibitor or ligand binding, and/or other regulatory functions. If there are additional subunits, they might be either parasite-encoded or resident host proteins; the identities of the additional proteins as well as the roles served by each would need to be determined.
Several findings are relevant to the discussion of these models. First, clag genes are present in all plasmodial species but absent from other apicomplexan genera, paralleling transport studies that show PSAC activity in only plasmodia (Alkhalil et al., 2007). Second, a role for clag3 product in pore formation is suggested by the A1210T mutation in a leupeptin-resistant channel mutant: point mutations in other ion channel genes often yield altered gating and selectivity. Third, the extracellular loop 35 kD from the protein’s C-terminus, identified through protease susceptibility studies, exhibits marked variability between P. falciparum lines, presumably due to positive selection by host immunity (Figure S6C; Iriko et al., 2008). Because the Dd2 clag3.1 gene carries a unique sequence here, this site may define the ISPA-28 binding pocket. Although conventional membrane-spanning regions are missing, the Phobius algorithm detects two conserved hydrophobic segments with high probability of forming transmembrane domains (residues 1000-1062 and residues 1208-1231 in the clag3.1 product, Figure S7C). These atypical transmembrane domains along with the lack of homology to known ion channels may contribute to PSAC’s unusual selectivity and conductance properties, as established by functional studies (Ginsburg and Stein, 1988; Cohn et al., 2003; Bokhari et al., 2008).
Our findings open several new directions for future research. Most importantly, they permit molecular studies into the mechanisms of PSAC-inhibitor interaction and the process of solute recognition and permeation through this unusual channel. A second direction will be to explore the trafficking of parasite-encoded channel components to the host membrane. Surprisingly, the clag products are synthesized many hours before channel activity is detectable. As described above, these proteins are packaged into rhoptries, secreted into the host erythrocyte upon invasion, and trafficked across the parasitophorous vacuolar membrane prior to reaching the host membrane. Why is this circuitous route taken when there is a more direct mechanism involving synthesis in the intracellular parasite and signal sequence-mediated export (Boddey et al., 2010; Russo et al., 2010)?
By providing a definitive protein target, our study will also stimulate drug and vaccine development against PSAC (Pillai et al., 2010). Should enthusiasm be tempered by the presence of two paralogous genes, selection for diversity at exposed sites, and variable inhibitor affinities for parasite isolates? While polymorphic sites on the channel have permitted linkage analysis and gene identification, most of the coding region is highly conserved in P. falciparum and other malaria parasites, consistent with conserved selectivity and transport biophysics in all infected cells (Lisk and Desai, 2005). Because most inhibitors do not appear to interact with polymorphic sites, targeting critical conserved parts of the protein for drug development will be facilitated by the high-throughput screening approaches used here.
Infected cell osmotic lysis in permeant solutes was used to quantify solute permeabilities and was performed as described (Wagner et al., 2003). High-throughput inhibitor screening was based on this transmittance assay and was performed in 384-well format as described (Pillai et al., 2010). Details are provided in Extended Experimental Procedures.
Whole-cell and single channel currents were recorded under voltage-clamp conditions using previously described patch-clamp methods (Alkhalil et al., 2004). Details are provided in the Extended Experimental Procedures.
All plasmids were confirmed by restriction digestion and sequencing, propagated in E. coli strain XL-10 gold (Strategene), and prepared with EndoFree Plasmid Mega Kit (QIAGEN). Percoll-enriched trophozoites were mixed with plasmid-loaded erythrocytes and allowed to grow for two cycles before application of 5 nM WR99210. Transformed cells were typically detected by microscopic examination within 3-8 weeks.
We thank John Adams, Alan Cowman, and Thomas Wellems for providing reagents and Mehreen Baakza, Kent Barbian, Yasmin El-Sayed, Juraj Kabat, Glenn Nardone, Son Nguyen, Norton Peet, Steve Porcella, Jose Riberio, Lattha Souvannaseng, Kurt Wollenberg, and Albert Youn for technical help and the NIAID Scientific Director’s office for providing required temporary laboratory space. We thank Edwin McCleskey and Thomas Wellems for critical reading of the manuscript. S. Desai is a named inventor on a provisional patent application describing the PSAC antimalarial target identified in this article. B. Turpin is an employee and shareholder of National Instruments, Inc., which markets data analysis software used in this study. This research was supported by the Medicines for Malaria Venture (MMV) and the Intramural Research Program of the National Institutes of Health, National Institute of Allergy and Infectious Diseases and the National Library of Medicine.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Supplemental information includes Extended Experimental Procedures, seven figures, and one table.