Defining the Role of Mutations in Plasmodium vivax Dihydrofolate Reductase-Thymidylate Synthase Gene Using an Episomal Plasmodium falciparum Transfection System

Abstract

Plasmodium vivax resistance to antifolates is prevalent throughout Australasia and is caused by point mutations within the parasite dihydrofolate reductase (DHFR)-thymidylate synthase. Several unique mutations have been reported in P. vivax DHFR, and their roles in resistance to classic and novel antifolates are not entirely clear due, in part, to the inability to culture P. vivax in vitro. In this study, we use a homologous system to episomally express both wild-type and various mutant P. vivax dhfr (pvdhfr) alleles in an antifolate-sensitive line of P. falciparum and to assess their influences on the susceptibility of the recipient P. falciparum line to commonly used and new antifolate drugs. Although the wild-type pvdhfr-transfected P. falciparum line was as susceptible to antifolate drugs as the P. falciparum parent line, the single (117N), double (57L/117T and 58R/117T), and quadruple (57L/58R/61M/117T) mutant pvdhfr alleles conferred a marked reduction in their susceptibilities to antifolates. The resistance index increased with the number of mutations in these alleles, indicating that these mutations contribute to antifolate resistance directly. In contrast, the triple mutant allele (58R/61M/117T) significantly reversed the resistance to all antifolates, indicating that 61M may be a compensatory mutation. These findings help elucidate the mechanism of antifolate resistance and the effect of existing mutations in the parasite population on the current and new generation of antifolate drugs. It also demonstrates that the episomal transfection system has the potential to provide a rapid screening system for drug development and for studying drug resistance mechanisms in P. vivax.

Of the four species of Plasmodium that commonly cause malaria in humans, Plasmodium vivax is the most widely distributed and can account for up to 80 million cases annually (25). Although P. vivax infections cause less mortality than P. falciparum, they do cause a debilitating disease that contributes to significant morbidity and economic loss in many regions where P. vivax is endemic (25). This is compounded by frequent relapses that can occur many times and for many months after the initial infection (6, 16, 22).

Plasmodium vivax parasites are susceptible to most antimalarial drugs. However, over the last 20 years there have been many reports that highlight the significant increase in resistance of P. vivax malaria to chloroquine, the recommended first-line treatment of P. vivax, and/or sulfadoxine-pyrimethamine (SP) (1, 7, 9, 13, 29, 31, 41, 42). Although the determinant(s) for chloroquine resistance in P. vivax remains elusive, a genetic basis for antifolate resistance in P. vivax has been identified as polymorphisms in the P. vivax dihydrofolate reductase (PvDHFR) active site, the same mechanism observed in antifolate resistance in P. falciparum (2, 11, 12, 17, 18, 20, 21, 24, 39).

In P. falciparum, point mutations within DHFR have been determined as the cause for antifolate resistance, such as pyrimethamine and cycloguanil resistance (4, 5, 8, 30, 32-36, 44, 45). It has been shown that resistance to antifolates results from the accumulation of mutations in the P. falciparum DHFR, principally A16V, N51I, C59R, S108N or S108T, and I164L.

A large number of point mutations have been identified in PvDHFR, and some were reported to be prevalent in many areas of P. vivax endemicity (2, 12, 18, 20, 21, 39). A particular set of mutations (F57L + S58R + T61M + S117T) within the P. vivax DHFR was shown to correlate with SP treatment failures (18, 39) and to confer significant antifolate resistance when transfected into antifolate-sensitive P. falciparum (28). Resistance is due to alterations to the pyrimethamine binding site of PvDHFR that reduces parasite-drug interactions (23, 28). However, the contribution of these PvDHFR mutations to resistance to a variety of antifolates drugs is not clear.

Due to the difficulty in maintaining P. vivax in in vitro cultures, most studies of the mutations within PvDHFR have been limited to surrogate biological systems such as yeast and Escherichia coli (17, 19, 24, 38). Although these systems have obvious experimental utility, they are different from Plasmodium in many biological aspects, particularly in membrane structures and transporters that can potentially affect the susceptibility to drugs. A recent publication (28) showed that the pvdhfr-ts quadruple mutant allele (57L + 58R + 61M + 117T) that is episomally expressed in P. falciparum provides significant protection against antifolates. This demonstrated an excellent potential of using P. falciparum as a biological system for the transgenic expression of pvdhfr-ts alleles to assess DHFR-TS interactions with antifolates.

We report here the use of this P. falciparum expression system to assess the effect that specific mutations within the P. vivax DHFR have on conventional and new-generation antifolate drugs. The findings improve our understanding of the effect of various mutant pvdhfr alleles observed in the field on the parasite responses to current and new generations of antifolates, improve the prediction of malaria drug treatment outcome, and provide a useful tool for drug development.

MATERIALS AND METHODS

Parasite lines.

The P. falciparum lines D6 and TM91c235 were used in the present study. D6 is susceptible to pyrimethamine, cycloguanil, and WR99210 (27) and was used as a recipient of various pvdhfr alleles. TM91c235 has quadruple mutations (51I/59R/108N/164L) in its DHFR and has been shown to be susceptible to WR99210 but resistant to cycloguanil and pyrimethamine (26). This line was used as an antifolate-resistant control. Parasite lines were cultured continuously in a 4% hematocrit blood-LPLF RPMI 1640 (consisting of 0.0005 mg of para-aminobenzoic acid/liter, 0.01 mg of folate/liter, and l-glutamine; Gibco, Grand Island, NY) supplemented with 10% human plasma, 25 mmol of HEPES buffer/liter, 25 mmol of NaHCO₃/liter, and gentamicin (25 mg/liter) and were grown at 37°C in tissue culture flasks gassed with 90% N₂, 5% O₂, and 5% CO₂ (40).

Plasmid construction.

The open reading frames (ORFs) from P. vivax dhfr-ts (Table 1) were amplified by PCR using primers and conditions described previously (2, 39). Genomic DNA from P. vivax isolates containing various mutant pvdhfr alleles (2) were used as templates. The pvdhfr triple mutant (58R/61M/117T) allele was kindly provided by Carol Sibley (University of Washington, Seattle). The plasmid construction was essentially performed as described earlier (28) with minor modifications. The pvdhfr PCR product was inserted between the P. falciparum dhfr-ts promoter and the P. falciparum thymidylate synthase gene, followed by an hrp2 termination sequence. All plasmids except for the pvdhfr quadruple mutant (57L/58R/61M/117T) had the 0.4-kb Saccharomyces cerevisiae bsd ORF inserted, in frame, between the promoter and dhfr-ts ORF as described previously (28) (Fig. 1).

TABLE 1.

P. vivax dhfr ORFs transfected into D6 parasites^a

Plasmid	Amino acid mutation at position:							Country of origin
	33	50	57	58	61	117	173
pvdhfr wild-type allele	P	N	F	S	T	S	I	East Timor
pvdhfr single mutant allele (117N)	P	N	F	S	T	N	I	East Timor
pvdhfr double mutant allele (58R/117T)	P	N	F	R	T	T	I	East Timor
pvdhfr double mutant allele (57L/117T)	P	N	L	S	T	T	I	Papua New Guinea
pvdhfr triple mutant allele (58R/61M/117T)	P	N	F	R	M	T	I	Indonesia
pvdhfr quadruple mutant allele (57L/58R/61M/117T)	P	N	L	R	M	T	I	Papua New Guinea

^aAny changes from the wild-type allele are indicated in boldface. All isolates were obtained as previously described (2).

FIG. 1.

Schematic illustration of the organization of P. falciparum expression plasmid vector (28). The sizes of the boxes are not proportional to the lengths of the genes.

Transfections.

Red blood cells were electroporated (320 V, 950 μF, 200 Ω) with 300 μg of plasmid DNA and then added to synchronized D6 schizont-staged parasites (10, 28). D6 parasites transfected with bsd-pvdhfr were selected with 2.5 ng/μl of blasticidin (Invitrogen Corp.). D6 parasites transfected with the pvdhfr quadruple mutant (57L/58R/61M/117T) were selected with 100 ng/μl of pyrimethamine (Hoffmann-La Roche, Switzerland). Parasites were first observed by Giemsa-stained thin smears between 15 and 30 days, with growth rate and morphology being indistinguishable from those of nontransfected parent D6 by day 40. Transfected parasites remained under blasticidin or pyrimethamine pressure except when being assayed by DHFR inhibitors. Plasmid DNA was recovered from transfected parasites during each drug assay, and the plasmid-bound dhfr was sequenced (data not shown) to ensure that additional mutations were not introduced.

Copy number determination.

The pvdhfr gene copy number for the pvdhfr quadruple mutant (57L/58R/61M/117T) and the bsd copy number for the other plasmids were estimated by a quantitative real-time PCR assay using the MX3000P multiplex quantitative PCR system (Stratagene) according to procedures similar to those described elsewhere (37). The primers used to quantify pvdhfr were those used to amplify the pvdhfr quadruple mutant (57L/58R/61M/117T) described previously (3). A single-copy gene encoding P. falciparum EBA175 (GenBank accession no. MAL7P1.176) was used as a reference (normalizer) gene for estimating either the pvdhfr or the bsd copy number. Primers used to amplify fragments of bsd and the eba175 genes are listed in Table 2. The copy numbers of the plasmids were determined based on the threshold cycle (C_T) values of pvdhfr and bsd using the ΔΔC_T method (37). The difference in the copy numbers between different transfectant lines was determined and tested (P value) by using a nonparametric comparison (Wilcoxon signed-rank test).

TABLE 2.

Primers used in the quantitative PCR to determine the pvdhfr copy number

Primer	Primer set	Sequence (5′-3′)
BSD	BSD F	GCGACGGCCGCATCT
	BSD R	ACAAGGTCCCCCAGTAAAATGA
EB175	EB175 F	CATCTGAAATGTCGCATAATAGTAGT
	EB175 R	GGTTCCCAAATCACCAACAGTT
Pvdhfr	dhfrS (3)	TCTGGGCAATAAGGGGACT
	dhfrA (3)	AGTTTCTACTTAGGCATTCCCTAT

In vitro susceptibility studies.

WR99210 was kindly supplied by Jacobus Pharmaceutical Company (Princeton, NJ). Cycloguanil and clociguanil were supplied by ICI Pharmaceuticals (Macclesfield, United Kingdom), and pyrimethamine was supplied by Hoffmann-La Roche (Switzerland). All compounds were initially dissolved in dimethyl sulfoxide and diluted in complete LPLF RPMI 1640. The susceptibility assays were essentially performed as described previously (14) with minor modifications. The parasite suspension (90 μl/well) consisted of infected erythrocytes (1.5% hematocrit, 0.7% parasitemia) in complete LPLF RPMI 1640. Parasites were challenged with dilutions of the drugs for 72 h. [³H]hypoxanthine (Perkin-Elmer) was added after 48 h when parasites had already reinvaded the red blood cells and were at ring stages, and the parasites were harvested after 72 h. Incorporation of ³H (measured in cpm) was recorded for each well, and the concentrations of drugs that inhibited 50% of parasite growth (IC₅₀s) were determined (28). All assays were performed in duplicate on at least three different occasions. The IC₅₀ was determined using nonlinear regression analysis provided by Sigma Plot for Windows, version 11 (Systat Software, Inc., San Jose, CA).

RESULTS

D6 parasites episomally expressing P. vivax dhfr-ts and their susceptibility to antifolates.

D6 parasites transfected with wild-type and various mutant pvdhfr alleles were assayed to determine the direct effect that these alleles have on the parasites' susceptibility to pyrimethamine, cycloguanil, clociguanil, and WR99210 (Table 3 and Fig. 2).

TABLE 3.

Susceptibility of parasites and transfected parasite lines to pyrimethamine, cycloguanil, clociguanil, and WR99210 and copy number of episomes per parasite^a

Origin of dhfr	Pyrimethamine		Cycloguanil		Clociguanil		WR99210		Copy no.
	IC50 (±SD)	RR	IC50 (±SD)	RR	IC50 (±SD)	RR	IC50 (±SD)	RR
Native P. falciparum lines
D6	0.028 (±0.016), 0.112 (±0.064)	1	0.118 (±0.032), 0.468 (±0.127)	0	0.021 (±0.02), 0.066 (±0.066)	1	0.166 (±0.1), 0.384 (±0.23)	2
TM91c235	142.61 (±4.161), 573.42 (±16.73)	2,592	116.4 (±50), 462.45 (±198)	50	3.16 (±1.58), 9.99 (±4.99)	144	0.339 (±0.12), 0.784 (±0.27)	5
D6 transfected with various pvdhfr alleles
Wild type	0.055 (±0.0007), 0.221 (±0.0003)	1	0.429 (±0.1), 1.7 (±0.4)	1	0.022 (±0.007), 0.069 (±0.022)	1	0.067 (±0.03), 0.155 (±0.069)	1	0.8-7
Single mutant (117N)	2.5 (±0.01), 10.05 (±0.04)	46	0.81 (±0.1), 3.22 (±0.4)	2	0.067 (±0.016), 0.211 (±0.05)	6	0.427 (±0.17), 0.988 (±0.39)	6	1.83-1.85
Double mutant (57L/117T)	3.66 (±2.64), 14.72 (±10.61)	67	2.77 (±2.5), 11.05 (±9.9)	6	0.219 (±0.033), 0.692 (±0.104)	10	1.5 (±0.04), 3.47 (±0.092)	22	0.7-9
Double mutant (58R/117T)	6.26 (±0.834), 25.17 (±3.35)	114	1.699 (±0.94), 6.75 (±3.73)	4	0.582 (±0.051), 1.84 (±0.161)	26	0.69 (±0.095), 1.6 (±0.22)	10	0.5-6
Triple mutant (58R/61M/117T)	2.81 (±1.9), 11.29 (±7.64)	51	0.587 (±0.25), 2.332 (±0.99)	1	0.073 (±0.04), 0.23 (±0.126)	3	0.105 (±0.025), 0.243 (±0.058)	2	0.8-4.5
Quadruple mutant (57L/58R/61M/117T)	467.34 (±90), 1,879 (±361)	8,497	320 (±22.93), 1,271.4 (±91.1)	746	56.44 (±29), 178.5 (±91.7)	2565	2.93 (±0.36), 6.78 (±0.833)	44	1.19-1.46

^aIC₅₀s are means of triplicate data points and are expressed as: mean nM ± SD (first value), mean ng/ml ± SD (second value). Values reflect the means of triplicate data points. The relative resistance index (RR) was determined in comparison to D6 transfected with the pvdhfr wild-type allele. The copy number indicates the range of pvdhfr copy numbers determined at three different time points.

FIG. 2.

In vitro susceptibility of P. falciparum D6 and D6 transfectant parasite lines episomally expressing various pvdhfr mutants to DHFR inhibitors: pyrimethamine (top left panel), cycloguanil (top right panel), clociguanil (bottom left panel), and WR99210 (bottom right panel). The symbols represent the means of triplicate data points.

D6 transfected with the P. vivax wild-type dhfr-ts (WT) had a susceptibility profile similar to that of the parental D6 strain, as observed earlier (28). This demonstrated that the WT was at least as equally susceptible to antifolates as the wild-type P. falciparum dhfr-ts. In contrast, D6 transfected with most of the mutant P. vivax dhfr alleles were less susceptible to antifolates than was the WT. D6 transfected with the P. vivax single mutant pvdhfr (117N) was 46-fold less susceptible to pyrimethamine, 6-fold less susceptible to clociguanil, and 6-fold less susceptible to WR99210 than the WT and D6.

The D6 strains transfected with two double mutant pvdhfr alleles—57L/117T and 58R/117T—were less susceptible to all of the antifolates tested than the 117N allele. The 58R/117T allele was more resistant to pyrimethamine and clociguanil but more susceptible to cycloguanil and WR99210 than was the 57L/117T allele.

D6 transfected with the triple mutant pvdhfr allele (58R/61M/117T) was more susceptible to the antifolates than D6 transfected with either single or double mutant alleles. The quadruple mutant pvdhfr allele (57L/58R/61M/117T) conferred a greater resistance to all of the antifolates than all of the other pvdhfr mutant alleles and the native P. falciparum quadruple mutant TM91c235. The 57L/58R/61M/117T mutant was 8,497-fold more resistant to pyrimethamine than the WT and >3-fold more resistant than TM91c235.

Copy number of episomes versus antifolate susceptibility.

The copy numbers of the episomally expressed pvdhfr in the transfected D6 parasites were determined by quantitative PCR after each drug assay was performed. As shown in Table 3, the average copy number within the parasite population was found to range between 0.5 and 9 copies per parasite over three different time points. There was no significant difference in the copy numbers between different transfectant parasite lines at the time of the assay (P = 0.576). Before the susceptibility assays were conducted, the transfected parasites were taken off selective drug pressure. A trend was observed that the longer the transfected parasites were without drug pressure the lower the copy number of the episomes per parasite (data not shown).

DISCUSSION

This study demonstrates, for the first time, the effect that specific point mutations within the P. vivax dhfr have on the susceptibility to several DHFR inhibitors in a homologous system by the transfection and episomal expression of the various wild-type and mutant pvdhfr alleles in P. falciparum. It provides direct evidence that commonly seen mutations in pvdhfr confer resistance to antifolate drugs, with higher numbers of mutations often causing higher levels of resistance. It also demonstrates that various mutations generate different resistance profiles.

We observed that P. falciparum transfected with the wild-type pvdhfr was at least as susceptible to antifolates as the parent strain D6, which is consistent with our earlier findings (28). We further demonstrated here that this susceptibility to antifolates was not affected by the copy number of the episomes carrying the wild-type pvdhfr per parasite since the WT-transfected parasite copy numbers ranged between 0.8 and 7 copies per parasite over three different time points without significant changes in their susceptibility profiles. This may result from a tight regulation of the parasite DHFR-TS expression level (43).

The in vitro susceptibility of the parasites transfected with mutant pvdhfr alleles to antifolate drugs reduced significantly. The single mutation of 117N conferred a marked (46-fold) increase in resistance to pyrimethamine compared to the WT pvdhfr-transfected line and moderate levels of reduced susceptibility to cycloguanil, clociguanil, and WR99210. This effect is consistent with that observed in P. falciparum resulting from the homologous S108T mutation (15, 23), which was suggested to be due to steric clashes between the pyrimethamine and 108N mutation (23).

The two pvdhfr alleles with double mutations (58R/117T and 57L/117T) that share one common mutation provided an opportunity to analyze the roles of the mutations that differ between the two alleles. Both alleles conferred higher resistance to all four antifolates than the WT- and single mutant 117N-transfected lines. The double mutant pvdhfr 58R/117T-transfected line showed a similar susceptibility to cycloguanil but a reduced susceptibility to pyrimethamine and clociguanil compared to the 57L/117T-transfected line. The 58R/117T transfectant line was more susceptible to WR99210 than was the 57L/117T transfectant line. Since both of these double mutants have the same 117T mutation, we hypothesize that the 57L mutation is responsible for the greater level of resistance to cycloguanil and particularly to WR99210. This result is supported by an earlier report (28) of a structural study that the 57L mutation clashes sterically with cycloguanil and WR99210 more than with pyrimethamine because this drug has a more flexible -CH₂-CH₃ group in the same position.

Interestingly, we observed that the triple pvdhfr mutant (58R/61M/117T) conferred much higher susceptibility to all of the antifolates tested than the double and single mutant pvdhfr alleles, almost reversing the susceptibility close to that of the parasite transfected with wild-type pvdhfr. This mutation is specific to P. vivax DHFR. This suggests that the T61M mutation may be a compensatory mutation that occurred after the development of resistance mutations (58R, 117T) to balance the damaging effect of resistant mutations on the enzyme activity. Although T61M may not directly affect the binding of the antifolate drugs from binding to the enzyme, this change may be essential before parasites developing additional mutations can become quadruple and quintuple mutants.

The quadruple pvdhfr mutant (57L/58R/61M/117T) showed the greatest level of resistance to the antifolates tested, and clinically this combination of mutations has been linked to treatment failure with SP (17, 18, 39). Molecular modeling and energy minimization studies (23, 28) have provided supporting evidence that the combination of these specific point mutations 57L, 58R, 61M, and 117T greatly hinders the binding of antifolates to the parasite enzyme.

In summary, we identified here a strong correlation between specific point mutations within pvdhfr and resistance to various antifolate drugs and provided novel insights into the role of these mutations: some contributing directly to resistance and some to enhancing the fitness of the mutant enzyme. These findings help elucidate the mechanism of antifolate resistance and the effect of existing mutations in the parasite population on current and new generations of antifolate drugs. For instance, the system can be used directly to assess the action and effect of the highly potent WR99210 series of antifolates on various existing P. vivax DHFR mutant alleles and thus help to speed up the drug development process. A similar approach could be used to identify and confirm other potential drug targets in P. vivax. Therefore, the episomal transfection system has the potential to provide a rapid screening system for drug development and for studying drug resistance mechanisms in P. vivax in the absence of a stable long-term in vitro culture system for P. vivax.