Am. J. Trop. Med. Hvg.. 5 I(S). 1994. pp. 645—658 Copyright 0 1994 by The American Society of Tropical Medicine and Hygiene

A STRONG ASSOCIATION BETWEEN MEFLOQUINE AND HALOFANTRINE RESISTANCE AND AMPLIFICATION, OVEREXPRESSION, AND MUTATION IN THE P-GLYCOPROTEIN GENE HOMOLOG (pfindr) OF PLASMODIUM FALCIPARUM IN VITRO

SHEILA A. PEEL, PATRICIA BRIGHT, BOYD YOUNT, JEAN HANDY, ANDRALPH S. BARIC Department of Epidemiology, School of Public Health, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina: Department of Microbiology and Immunology. School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina

Abstract. Stepwise selection for increased mefloquine resistance in a line of Plasmo dium falciparum in vitro resulted in increased resistance to halofantrine and quinine, in creased sensitivity to chloroquine, and amplification and overexpression of the P-glyco protein gene homolog (pfmdrl). A point mutation (tynosine to phenylalanine) noted at amino acid 86 in pfmdrl in the mefloquine-nesistant line W2mef was amplified in more resistant lines derived from it by in vitro selection pressure with mefloquine. Conversely, lines selected for increased chloroquine resistance exhibited a nevertant phenotype that was sensitive to mefloquine and halofantrine. These lines also demonstrated increased sensitiv ity to quinine, loss of amplification of pfindrl. loss of the mefloquine/halofantrine phe nylalanine-86 mutation, and selection for a tyrosine-86 mutation previously associated with chlonoquine resistance. These findings provide strong evidence for pfindrl mediating cross resistance to halofantrine and metloquine in P. falciparum in vitro.

Plasmodium falciparum, the most lethal hu en quinoline-containing compounds.5 @‘Sensitive man malaria parasite, has become increasingly and resistant parasites accumulate chloroquine at difficult to control since strains have emerged similar rates while chlonoquine-resistant para that are resistant not only to established anti sites eliminate drug 40—50 times more rapidly.7 malanials such as chloroquine and quinine, but It is not known whether mefloquine is rapidly also to newer therapeutic agents such as meflo eliminated from the parasite since efflux studies quine, halofantnine, and artemisinin.' 2 Although with this extremely lipophilic drug have not yet the mechanism(s) of resistance to these drugs been successful. Agents that reverse the MDR has not been elucidated, resistance may be sim resistant phenotype have also been successfully ilar to the multidrug-resistant (MDR) phenotype used to inhibit drug efflux from chloroquine-re observed in mammalian tumor cell lines, which sistant parasites, but not from chlonoquine-sen is characterized by decreased intracellular drug sitive ones.8'° Penflunidol, a neuroleptic drug, accumulation, increased cellular drug efflux, and reversed mefloquine and halofantrine resistance cross-resistance to a variety of structurally di in P.falciparum in vitro―'2 and mefloquine re verse lipophilic cyclic compounds.3 In MDR tu sistance in P. yoelii in vivo.'3 No compound, mon cells, this phenotype is mediated by the en however, has been identified that modulates re hanced expression of a 120—180 kD plasma sistance to all the quinoline-containing antima membrane glycoprotein, termed P-glycoprotein, larials; thus, the mechanisms may be indepen which functions as a putative ATP-dependent ef dent. flux pump.3 Drug efflux is inhibited by calcium Two MDR-like P-glycoprotein gene homo antagonists, tricyclic antidepressants, and anti logs, pfmdrl and pfmdr2, have been identified histaminics.4 in P. falciparum, 4. 5 and they share predicted Although the classic MDR phenotype has not structural motifs associated with the superfamily been demonstrated in P. falciparum, resistance of AlP-binding cassette (ABC) transporter pro derived to one quinoline-containing antimalarial teins.'6 Gene amplification and ovenexpression in vitro resulted in altered susceptibilities to oth of pfmdrl have been demonstrated in some chlo

648 GENETIC BASIS OF DRUG RESISTANCE IN P. FALCIPARUM 649 roquine-resistant isolates,'4 7 in a mefloquine-re tory concentration measurements were statisti sistant line in vitro,'5 and in MDR isolates from cally analyzed and are presented as the mean ± Thailand;'7 however, direct linkage of pfmdrl to standard deviation. Statistical signi ficance of resistance to these antimalarials has not been measurements was determined by a one-way proven. Point mutations in pfmdrl have also analysis of variance with a post hoc contrast to been associated with chloroquine and mefloqui compare group means. ne resistance,hl 8 but it has not been determined Nucleic acid extraction. Parasites were iso whether these mediate resistance since results lated from asynchronous cultures by saponin ly from different laboratories have been contradic sis, and genomic DNA was extracted essentially tony. Neither chloroquine nor mefloquine nesis as described.23 Stage-specific parasites were en tance was linked to the pfmdrl gene following nched by sedimentation through Percoll gradi restriction fragment length polymorphism anal ents,24 and total RNA was extracted by acid ysis of a laboratory cross of chloroquine-sensi guanidi nium thiocyanate phenol-chloroform tive and -resistant parasites;'9 20 rather, chlono methods.25 quine resistance was mapped to a locus on DNA probes. Gene probes were generated by chromosome 7 that was independent of the polymerase chain reaction (PCR) methodology pfmdrl allele.2° In support of these findings, se using the following deoxyoligonucleotide prim lection for increased chloroquine resistance in an ens derived from known plasmodial sequences: MDR parasite containing three copies of the 1) CCAGGAATACCAGTGC and CTTTACC pfmdrl allele resulted in loss of amplification of TTCACGACC for PCR of the circumsporozoite this allele. Increased sensitivity to mefloquine protein (CSP);26 2) GGAGGATFATATCCCG was also noted, suggesting a relationship of and CAGCTGCAACAATI'GG for pfmdrl; and pfindrl with mefloquine resistance.2' In this pa 3) GCTGAAGGAGAAGATACG and GGTTT pen, we have examined the role of pfmdrl in the TCCAGATTCTTC for the glycophorin binding acquisition of mefloquine resistance in a series protein (GBP) gene.27 Radiolabeled probes spe of highly resistant/sensitive parasite lines. cific for either the CSP, GBP, on pfmdrl genes were obtained using Qiagen gel extraction pro

MATERIALS AND METHODS tocols (Qiagen Inc., Chatsworth, CA). The DNA was extracted with phenol:chloroform:isoamyl Parasite lines. The parasites W2 and W2mef alcohol, precipitated, and radiolabeled using the were derived in our laboratory from the Indo Random Primers DNA Labeling System (Gibco china III strain,5 and strain FCR3-Gambia was BRL, Gaithensburg, MD). Free nucleotides were obtained from Dr. William Tragen (Rockefeller removed by spin-column chromatography.28 University, New York, NY). The lines Mef 55, Analysis of gene copy number. Genomic Mef 100, Mef 300, and Mef 2.4 were derived DNA was restricted with the appropriate restnic from W2mef by high-dose relapse drug pressure tion endonuclease, electrophoresed on 0.7% aga and were grown using standard methods of in nose gels, then transferred to nitrocellulose vitro cultivation as previously described.6 Selec (Schleicher and Schuell, Keene, NH).29 Blots tion protocols for derivation of increased chlo were hybridized at 42°Cwith 2 X l0@ cpm/ml noquine resistance from W2mef were as pnevi of probe in 50% formamide (Sigma), 5X SSPE ously described.6 (0.9 M NaC1, 50 mM sodium phosphate, pH 7.7, Drug susceptibility profiles. The index of an 5 mM EDTA), 5 x Denhardt's solution (0.1% timalarial activity was assessed using 3H-hypo Ficoll, 0. 1% bovine serum albumin, 0. 1% po xanthine incorporation microdilution assays as lyvinylpyrolidone), I00 pg/ml of polyadenylic previously described.6 22 The 50% inhibitory acid (Sigma), and 0.2% Tween 20 (Sigma). The concentrations (IC50) of chloroquine diphosphate nonspecific background was removed by wash (Sigma, St. Louis, MO), mefloquine hydrochlo ing the blots with 0. 1 X SSPE, and 0. 1% sodium ride (Walter Reed Army Institute Research dodecyl sulfate at 65°C.Filters were exposed to [WRAIR], Washington, DC), quinine sulfate XAR film (Eastman Kodak, Rochester, NY) at (Sigma), halofantrine hydrochloride (WRAIR), —70°C.Gene copy number was computed using and artemisinin (WRAIR) were determined by AMBIS@, a radioanalytical imaging system nonlinear regression analysis of dose-response (AMBIS Systems, Inc., San Diego, CA). Blots data for 5—Il independent assays.22 All inhibi were counted over a I2—24-hr period, windows 650 PEEL AND OTHERS

for pfmdrl were counted, and background signal (chloroquin. r..i.tant) was subtracted. Blots were then stripped of ‘if probe by immersion in 10 mM Tris, pH 8.0, 1 w2@f (Nsfloquine and thl3roquin. rs.i.tant) mM EDTA, 0. 1x Denhardt's solution for 2 hr F at 75°C,rinsed with 0. 1 X SSPE at room tem @F55 N.p/r too peratune, drained, and then exposed to XAR film =@ p@F 100 @ to ensure removal of probe. The blots were re ,I I probed with the CSP gene probe, scanned with III, 300 Ng.P/cQ AMBIS, and the ratios of signal for pfindrl to pem0mL ‘I MEF 1.2 csP werecalculatedfor eachline and normal ized to the ratios obtained for W2, which dem NRF 2.4 onstrates a single copy of pfindrl. DNA cloning and sequencing. The PCR am FIGURE 1. Selection protocols for the derivation of Plasmodium falciparum lines resistant to mefloquine plified gene fragments of pfindrl (5' primer: hydrochloride (MEF) and chloroquine diphosphate 497-514; 3' primer: 1002-983; 5' primer: 497- (CQ) (ng/ml). Parasites were selected for increasing 514; 3' primer: 1138-1121) were ligated into resistance to MEF and CQ as described previously@ either PCR 1000 (Invitrogen, San Diego, CA) or and in the text. Parasite lines shown were stabilated pGEM-T (Promega, Madison, WI) using the TA and used in the current study. cloning systems (Invitrogen and Promega),l4 transformed into INVaF' (PCRI000) (Invitro gen) or DH5-a (pGEM-T), and selected with ei selection for increased mefloquine resistance ther kanamycin or ampicillin (Sigma). Double also resulted in lines exhibiting increased sen stranded sequencing was performed using the sitivity to chloroquine,6 we reasoned that selec dideoxy chain termination method with Sequen tion for increased chloroquine resistance in ase Version 2.0 (United States Biochemical, W2mef may result in the isolation of revertant Cleveland, OH). The sequence of both strands lines, providing a model to study the molecular was obtained using M13 forward and reverse basis of drug resistance. The selection protocol primers. Direct sequencing of the PCR product initiated with chloroquine is shown in Figure 1. was performed as follows. Amplified product Continuous drug pressure was initiated with 100 was concentrated using Centricon 30 concentra ng/ml of chloroquine, and concentrations were tors (Amicon, Inc., Beverly, MA), 4 p.1 of sam then increased in 50 ng/ml increments until ple was used as template in an asymmetric PCR lines that were resistant to 100 ng/ml with a primer ratio of I .0 pmol of limiting prim (W2mefCQlOO), 150 ng/ml (W2mefCQl 50), er: 50 pmol of excess primer, 5 U of Taq poly and 200 ng/ml (W2mefCQ200) of chloroquine merase (Promega), and 1 p.M of dNTP. Asym were derived over a I2-month period. metric product was purified from excess primer Microdilution assays based on 3H-hypoxan and dNTPs by PCR spin-column chromatogra thine incorporation into parasite DNA were used phy (Qiagen), and dideoxy sequencing was per to determine the drug susceptibility of formed using 8 p.1 of purified product with 10 W2mefCQ200 to chloroquine, mefloquine, hal pmol of primer (the opposite primer used in the ofantrine, quinine, and artemisinin (Table 1). asymmetric reaction). As expected, the IC50 of chloroquine for Expression levels of RNA and Northern W2mefCQ200 (I 02.34 ±9.69 ng/ml) was sig analysis. The mRNA expression levels were de nificantly increased above that of W2mef (62.61 termined by slot-blot hybridization analysis,@° ±5. 18 ng/ml) (Table 1; P < 0.01). Interestingly, and RNA transcription analysis was determined W2mefCQ200 was now extremely susceptible by formaldehyde gel electrophoresis.28 to mefloquine with an IC@ of 2.06 ± 1.18 ng/ ml as compared with 14.49 ± 3.49 ng/ml for RESULTS W2mef (Table I ; P < 0.001), and was signifi cantly more sensitive to mefloquine than W2, Selection for increased chioroquine resis the progenitor clone of W2mef at the IC@ (3.08 tance in W2mef. Previously, we isolated a se ±0.16 ng/ml versus 11.61 ±0.67 ng/ml; P < ries of mefloquine-resistant lines from the mul 0.001 ), IC95 (4.34 ±0.49 ng/ml versus 15.74 ± tidrug-resistant line W2mef (Figure l).6 Since 1.46 ng/ml; P < 0.001), and IC99(9.35 ±2.33 GENETIC BASIS OF DRUG RESISTANCE IN P. FALCIPARUM 651

TABLE I Drug susceptibility profiles

Clone/lineIC,@,*ChloroquineMefloquineHalofantrineQuinineArtemisininW2 ±8.14 ±0.89 ±0.01 ±5.43 ±0.46 W2mef 62.61 ±5.18 14.49 ±3.49 1.48 ±0.19 56.32 ±4.17 3.46 ±0.38 Mef 2.4 48.35 ±4.57 20. 17 ±7.21 3.42 ± 1.86 97.54 ±9.49 3.59 ±0.45 W2mefCQ200 102.34 ±9.69 2.06 ± 1.18 0.18 ±0.05 48.88 ±8.24 3.20 ±0.40 Sensitivet91.70 <10.04.35 <8.00.37

ng/ml versus 33.76 ± 5.22 ng/ml; P < 0.001) pressed as the ratio of pfindrl to CSP and reflect levels. The IC@ of halofantrine was also signif at least two separate experiments. In agreement icantly reduced, 0. 18 ± 0.05 ng/ml for with previous findings, pfmdrl was amplified W2mefCQ200 as compared with 1.48 ± 0.19 three-fold in the W2-mef clone.lS The Mef 55 ng/ml for W2mef (P < 0.001), indicating in line, selected with continuous pressure protocols creased sensitivity to this antimalarial as well to 55 ng/ml, also demonstrated a three-fold am (Table 1). The IC@ values for W2mefCQ200 to plification of the pfindrl allele. Interestingly, an quinine indicated a slight decrease (P < 0.05 eight-fold amplification of pfmdrl was noted in versus W2mefl, while repeated assays (n = 5) the Mef 100, Mef 300, and Mef 2.4 lines, which with Mef 2.4 indicated a slight increase in the were selected by discontinuous drug pressure IC@ as compared with W2 and W2mef (P < with 100, 300, and 2,400 ng/ml of mefloquine, 0.01; Table 1). The IC@ to artemisinin for respectively (Figures 1 and 2 and Table 2).6 The W2mefCQ200 had not significantly changed mefloquine-halofantrmne—sensitive revertant line from that of W2mef. W2mefCQ200 contained a single copy of Amplification of pfmdrl in highly meflo pfmdrl, losing the three-fold amplification of quine-resistant lines. To determine pfindrl gene pfindrl that had been acquired during the induc copy number, Southern blot hybridization anal tion of mefloquine resistance in W2mef. As ex ysis was performed on W2 and W2mef, the me pected, a single copy of the pfindrl gene was floquine passage lines Mef 55, Mef 100, Mef noted in the chloroquine-resistant, mefloquine 300, Mef 2.4, and the mefloquine-halofantnine sensitive parasites W2 and FCR3 (Figure 2 and revertant line W2mefCQ200. A 390-basepair Table 2). (bp) pfindrl DNA fragment was used to probe Expression of mRNA and Northern analy Southern blots of a Bgl H digest of genomic sis. Previous studies have demonstrated enhanced DNA. In vitro selection for increasing meflo pfindrl transcript prevalence in mefloquine-resis quine resistance resulted in a corresponding am tant clones and [email protected] hybridiza plification of the pfindrl gene in the mefloquine tion analysis was used to determine mRNA ex passage lines (Figures 2A and B) in comparison pression levels of pfindrl in W2, W2mef, Mef with the known single copy CSP gene (Figures 2.4, and W2mefCQ200 (Figure 3). Direct scan 2C and D). Dihydrofolate reductase was also ning of nitrocellulose membranes indicated that found to be single copy in W2, Mef 300, and equivalent amounts of total RNA were analyzed Mef 2.4. The gene copy number of each of these as indicated by hybridization with the GBP gene lines (Table 2) was obtained following Southern control. An approximate three-fold increased blot analysis by direct scanning of the nitrocel pfmdrl mRNA expression was detected in lulose membranes using a radioanalytical quan W2Mef, and an eight-fold enhanced expression tification imaging system since the linear capac was detected in Mef 2.4 as compared with low ity of radiographic film is rapidly exceeded, level expression in the mefloquine-sensitive clone making densitometric determinations of copy W2 and the mefloquine-halofantrine revertant number difficult. The reported values are ex line W2mefCQ200 (Figure 3 and Table 2). North 652 PEEL AND OTHERS A B @ 1 2 3 4 6 7 8 12,21*

.7,12*

C. D

FIGURE 2. Southern blot analysis of the P-glycoprotein gene homolog (pfmdrl) in the Plasmodiumfalciparum passage lines. Four micrograms of genomic DNA of FCR3 (lane I), W2 (lane 2), W2mef (lane 3), Mef 55 (lane 4), Mef 100 (lane 5), Mef 300 (lane 6), Mef 2.4 (lane 7), and W2mefCQ200 (lane 8) was digested with Bgl II, separated by electrophoresis on 0.7% agarose gels, transferred to nitrocellulose, and hybridized with either a 32P-labeled pfindrl (A, 24-hr exposure; B, 36-hr exposure) or a circumsporozoite gene (C, 24-hr exposure; D, 36-hr exposure) DNA probe as described in the experimental procedures. Values on the left side of A are in basepairs.

em analysis of agarose formaldehyde gels of late resulting in an asparagine (AAT) to tyrosine @ stage trophozoites of W2, W2mef, and Mef 2.4 (TAT) change at amino acid 86. Since a tyro also demonstrated the two major transcripts of the sine (TAT) to phenylalanine (ITT) change at po expected size that were significantly increased in sition I 84 has also been reported,'7 oligonucle amount in W2-mef and Mef 2.4 as compared otide primers were constructed to amplify a 502- with W2 (Figure 4). or 63 l-bp fragment spanning the point mutations Sequence analysis of pfmdrl. One to four at either nucleotides 754 or 1049, respectively. amino acid changes have been correlated with Extensive sequence analysis was performed on the chlonoquine resistance phenotype. The pu the PCR product and genomic subclones from tative mutation reported for W2 and W2mef in W2, W2mef, Mef2.4, and W2mefCQ200. Direct pfindrl, referred to as the KI allele, was an A sequencing of the asymmetric PCR product re to T point mutation at nucleotide position 754, vealed the expected TAT at position 86 in W2 GENETIC BASIS OF DRUG RESISTANCE IN P. FALCIPARUM 653

TABLE 2 Gene copy number, mRNA expression levels, and Se 1 2 3 quence comparisons in the P-glycoprotein gene ho molog (pfmdrl) gene among Plasmodium falcipa rum clones and lines

acid copy/ at position at position mRNA l84fW2Parasite lineGene levels'Nucleotide75Sf l@9tAmino 86t ,— _. TAT Tyr W2mefI/I 3/3TAT TAT TAT Tyr Tyr PheMef 1T1'Tyr 55 ND ND Mef 100 8/ND ND ND ND ND NDMefMef 3003/ND 8/NDNDND NDND ND 2.4 TAT Tyr W2mefCQ2008/8 I/I‘ITI'TAT TATPhe Tyr Tyr S All values are rounded to the nearest integer. Gene copy number is expressed as a ratio of lpfmdrl/circumsporozoite proteini and mRNA Icy els as a ratio of lpfmdrl/glycophorin binding proteini. f Chloroquine-sensitivelinesencodeAAT(Asn)atnucleotide755.and TAT (Tyr) at nucleotide 049.―

(Figure 5A); however, an additional A to T mu tation at nucleotide 755 (liT) was also noted in W2mef and Mef 2.4 (Figures SB and C). Close inspection of the results revealed a faint band in the A lane of W2mef (see arrow, Figure SB), which indicated a mixed sequence population at this locus. Since quantitative PCR was not used to obtain gene fragments for DNA sequencing, and since both W2mef and Mef 2.4 contained

FIGURE 4. Northern analysis of the P-glycoprotein 5.Oug O.5ug gene homolog (pfmdrl) in Plasmodium falciparum lines W2 (lane 1), W2mef (lane 2), and Mef 2.4 (lane 3). Five micrograms of total cellular RNA from tro phozoite stage parasites was applied to Immobilon N 1 (Millipore Corp., Bedford, MA) and hybridized to a pfmdrl gene probe.

2 — multiple copies of pfmdrl, sequence analysis 3 was performed on plasmid subclones of the PCR product isolated from these lines. The putative chloroquine-associated allele, TAT, was noted in 4 10 of 10 subclones of W2; however, analysis of this region in W2mef demonstrated only four of 16 subclones with the expected TAT sequence (Figure SE), while the remainder had mutated to FIGURE 3. Slot-blot analysis of the P-glycoprotein gene homolog (pfmdrl) in Plasmodium falciparum Tfl' (Figure SD), resulting in an amino acid lines W2 (lane 1), W2mef (lane 2), Mef 2.4 (lane 3), change from tyrosine (TAT) to phenylalanine and W2mefCQ200 (lane 4). Total cellular RNA was (TIT). This mixed locus in W2mef was con isolated by the guanidinium isothiocyanate method, firmed in clones that had been maintained in fro and S and 0.5 jig of total RNA was applied to Im mobilon N (Millipore Corp., Bedford, MA) and hy zen storage since 1984, while sequence analysis bnidized to a pfindr! or a glycophonin binding protein of 24 plasmid subclones of Mef 2.4 demonstrat gene probe. ed only 1T@I'at this position. If the pfmdrl gene, 654 PEEL AND OTHERS

A B C GA T C

D E F GA TC GAT C GAT C

4- 4-

FIGuRE5. Variant allele in the P-glycoprotein gene homolog (pfmdrl) gene in the Plasmodium falciparum lines. Genomic DNA was amplified across a 502-basepair region encompassing the putative chloroquine (CQ)— specific allele at nucleotide position 754. Direct sequencing of the polymerase chain reaction product of A, W2; B, W2mef; and C, Mef 2.4, and double-stranded sequencing of representative independent plasmid clones of D and E, W2mef; and F, W2mefCQ200 demonstrate the variability at position 754/5. The arrows indicate the variant allele TAT or TIT The gels are read from the bottom, G (lane 1), A (lane 2), T (lane 3), and C (lane 4).

and in particular this region of the gene, is under DISCUSSION strong selective pressure and confers resistance to mefloquine, W2mefCQ200, which was sen Selection for enhanced mefloquine resistance sitive to mefloquine and halofantrine, should re from W2 to W2mef to Mef 2.4 resulted in lines flect sequence changes at this allele. A 502-bp demonstrating increased halofantrine resistance fragment encompassing this region was ampli and increased chloroquine sensitivity. The slight fled by PCR, subcloned, and eight of eight plas increase in quinine resistance noted for these mid subclones were sequenced and found to lines, and its decrease (albeit slight) following contain the putative chloroquine mutation (TAT) selection for chloroquine resistance, suggests at position 86 (Figure 5). that selection for high levels of resistance to me Sequence analysis of the 63l-bp PCR product floquine may also modulate levels of quinine re and eight of eight genomic subclones of W2, sistance. Interestingly, increased mefloquine and W2mef, Mef 2.4, and W2mefCQ200 across nu quinine resistance were noted following selec cleotide 1049 revealed tyrosine (TAT) at this po tion for halofantrine resistance in vitro in both sition (Table 2). Thus changes at this position chloroquine-sensitive and -resistant parasites.32 could not be correlated with mefloquine or hal Current observations with clinical isolates from ofantrmne resistance in our lines. Africa and Thailand also indicate that responses GENETIC BASIS OF DRUG RESISTANCE IN P. FALCIPARUM 655 to mefloquine and halofantrine demonstrate a derscores the importance of sequencing individ positive correlation with each other and an in ual clones from genes that are amplified. verse correlation with chloroquine resistance.3@ That stepwise selection with mefloquine re 36 Whether or not quinine susceptibility varies in suIts in the specific amplification ofpfindrl with any consistent way with responses to mefloquine only the mefloquine-associated mutation (86- and halofantrine is less clear; however, in vivo Phe) in Mef 2.4 is significant and not without studies suggest that quinine demonstrates a pos precedent. Cross-resistance in human KB carci itive correlation with these drugs, and an inverse noma cells was found to result from a cluster of correlation with chloroquine resistance.33 @“Se point mutations in the MDRI gene that resulted lection for increased chloroquine resistance in in a Gly-185 to Val-l85 substitution in the P W2mef resulted in a line demonstrating a never glycoprotein.38 The mutations were identified in tant sensitive phenotype to mefloquine and hal a cell line where the wild-type MDR1 was am ofantrine. These results are consistent with those plified, yet selection for the next level of resis reporting a two-fold decrease in the IC@ of me tance resulted in all amplified copies of the gene floquine following selection for increased chlo carrying the same mutations.39 While the IC@ of roquine resistance in W2mef,2l although the ab chloroquine has decreased in Mef 2.4, the par solute values of IC50 cannot be compared asite remains resistant to chloroquine, suggest because of the different assay methodology. Sus ing that either 86-Tyr is not critical for resis ceptibility to halofantrine for these lines was not tance, or that the 86-Phe change remains a reported. The apparent lack of modulation of ar competent allele for chloroquine resistance. temisinin resistance by quinoline-containing The loss of amplified genes carrying the me methanols suggests that pfmdrl may not mediate floquine-halofantrine—associated mutation, 86- a truly classic MDR phenotype in P. falciparum. Phe (TV!'), subsequent reversion to a mefloqui ne-halofantrine—sensitive phenotype, and Increased resistance to mefloquine and halo selection for an allele (TAT) associated with fantrine correlated with amplification and over chloroquine resistance in W2mefCQ200 suggest expression of pfrndrl; however, lines selected that this area of the gene is under heavy selec with mefloquine concentrations of 100 ng/ml or tive pressure from quinoline-containing corn greater did not demonstrate proportionately pounds. While appearing to support a role for higher levels of pfmdrl amplification and over the KI allele in chloroquine resistance,18 these expression. Analysis of the W2mefCQ200 line results seem to contradict those that suggested revealed that the three-fold amplification of that there was no link between pfindrl and chlo pfindrl noted in the progenitor line, W2mef, had roquine resistance, but implicated an allele on been lost. The W2mefCQ200 line now con chromosome 7.@°In support of these findings, tamed a single copy of the gene that was not chloroquine-resistant isolates from Thailand did overexpressed. These results are consistent with not demonstrate the putative chloroquine nesis studies that indicate that amplification of pfmdrl tance—associated KI allele (TAT).17 In both stud does not correlate with increased chloroquine re ies, however, the clones and isolates used con sistance,'7' 2l.37as well as with studies suggesting tamed multiple copies of the pfmdrl gene. For that pfindrl amplification correlates with in example, although it is clear that chloroquine re creased mefloquine and halofantrine resistance)7 sistance maps to a locus on chromosome 7,@°the Examination of the intragenic allele at amino resistant clone D2d, derived from W2rnef, likely acid 86, previously associated with chloroquine contained two alleles, one previously associated resistance,18 suggested that a specific hydropho with chloroquine resistance (86@Tyr),lR and the bic domain mutation was derived depending newly identified mutation (86-Phe) associated upon the antimalarial used in the selection pro with mefloquine and halofantrine resistance. tocol. The identification in W2mef of a mixed Since the genotype of all pfindrl copies of the sequence population with TAT at position 754, parents or progeny of the genetic cross has not and a previously unidentified mutation TTF at been determined, pfindrl cannot be excluded as position 755, emphasizes a potential problem as a genetic determinant modulating mefloquine sociated with the current technique of consensus and/or chloroquine resistance. sequencing of the PCR product since all se Although the 86-Tyr to 86-Phe mutation lo quence changes may not be identified. This un cated between transmembrane (TM) domains 1 656 PEEL AND OTHERS

and 2 in pfmdrl in our passage lines is the has been localized principally to the surface of strongest link, to date, of a membrane-associated the digestive vacuole of the parasite,37 the pro @ mutation with mefloquine and halofantrine resis posed site of action of these drugs.45 Although tance, the high gene copy number of Mef 2.4 the protein may be positioned to directly affect has made mutational analysis difficult. Clearly, drug levels within this vacuole, its actual ori additional mutations in pfmdrl on yet unrecog entation (export versus import) within the vac nized genes may also function in resistance. uole membrane is not known. If pfmdrl func Studies using clinical isolates from Thailand in tions in a manner similar to P-glycopnotein in dicated that mefloquine resistance may be as MDR cells, mefloquine and halofantnine, which sociated with an A to T mutation at nucleotide are more tightly associated with membranes than 1049 resulting in a Tyr to Phe change between chloroquine or quinine, may be better substrates TM2 and 317 That this mutation was also noted for transport. With only minor amounts of Pghl in the lone mefloquine/halofantrine sensitive line localized to the plasma membrane of the para (TM335) suggests that this mutation does not site,37 however, the mechanism by which these necessarily correlate with resistance to these drugs could then move from the digestive vac drugs. While our data support this interpretation, uole, through the parasite plasma membrane, the it should be noted that mutations within TM do parasitophonous vacuole membrane, and finally mains of other ABC gene family members, such through the cytoplasm and plasma membrane of as TMII of rodent mdrl and mdr3 and TMI, the erythrocyte, remains unknown. TM6, and TM 10 of the human cystic fibrosis Alternatively, studies with MDR tumor cells transmembrane conductance regulator protein suggest that P-glycopnotein may function as a (CFTR), have also been correlated with specific proton/chloride pump.3 Since quinoline-contain ity to various compounds.@°4' Although analysis ing antimalarials as weak bases are concentrated of chimenic MDRI/MDR2 genes has indicated by the acidic digestive vacuole,4546 regulation of that neither point mutations nor specific amino pH of this organelle has also been suggested as acid sequences encoded drug-resistance profiles, a possible function of Pgh 1.“@In support of these drug binding and recognition seemed to involve studies, Bray and others have suggested that higher-order structure affected by multiple, in changes in chloroquine accumulation (presum dependent parts of the molecule.42 Recent stud ably due to increased vacuole pH), and not ef ies have indicated that the first cytoplasmic loop flux from the food vacuole, may determine re of MDRI , and in particular residues 165, 166, sistance.48 168, and 169, appear to confer functional mul Analysis of our progressively more drug-re tidrug transporter activity.43 If drug recognition and binding involves the participation of mem sistant parasite lines indicated that a single mu brane domains from both halves of the pfindrl tation in Pghl could not alone be responsible for protein product, P-glycoprotein homolog I high levels of resistance since the Mef 100, Mef (Pgh I ), in P. falciparum, mutations in on near 300, and Mef 2.4 lines contained equivalently one or more tnansmembnane domains may be amplified and overexpressed pfmdrl genes, and critical determinants of resistance to mefloquine yet Mef 2.4 is significantly more resistant.6 In and halofantnine. addition, selection for the revertant line with Although the results presented provide strong chloroquine resulted in hypersensitivity to me support for pfmdrl mediating cross-resistance to floquine and halofantnine and decreased quinine mefloquine, halofantnine, and perhaps quinine in resistance as compared with the progenitor clone P. falciparum, it is also not yet clear how Pghl W2. Since this revertant line retained the puta could modulate resistance to these drugs. In tive chloroquine-associated allele (TAT) found MDR cells, the most important determinant of in pfmdrl of W2, yet was more sensitive than transport by the putative export pump is a com W2 to these compounds, additional genetic pound's relative hydrophobicity.3 Compounds changes must have occurred. The availability of concentrated in plasma membranes are thought an in vitro model with lines demonstrating in to be detected and expelled as they partition into creased resistance, as well as revertant lines, the membrane in a manner similar to a hydro should allow us not only to begin to dissect the phobic vacuum cleaner,@ as well as through the complex mechanisms of drug resistance in P. channel of the transporter.43 In contrast, Pghl falciparum, but also to examine the biological GENETIC BASIS OF DRUG RESISTANCE IN P. FALCIPARUM 657 function of Pgh 1 in the parasite and its potential 10. Peters W, Ekong R, Robinson BL, Warhurst DC, role in mediating resistance to antimalarials. Pan XQ, 1989. Antihistaminic drugs that re verse chloroquine resistance in Plasmodium fal ciparum. Lancet ii: 334—335. Acknowledgments: We thank Dr. Lorraine K. Alex ander for help with the statistical analyses. We are I 1. Kyle DE, Oduola AM J, Martin 5K, Milhous WK, 1990. Plasmodium falciparum: modulation by grateful to the North Carolina branch of the American Red Cross for the provision of blood and plasma. calcium antagonists of resistance to chloroquine, desethylchloroquine, quinine, and quinidine in Financial support: This work was supported by re vitro. Trans R Soc Trop Med Hyg 84: 474—478. search grants from the John Reed Trust to Sheila A. 12. Oduola AMJ, Omitowoju GO, Gerena L, Kyle Peel, and the North Carolina Biotechnology Center DE, Milhous WK, Sowunmi A, Salako LA, (92 13-ARG-0405) to Ralph S. Baric and Jean Handy. I993. Reversal of mefloquine resistance with Ralph S. Baric is a recipient of an established inves penfluridol in isolates of Plasmodium falcipa tigator award from the American Heart Association rum from south-west Nigeria. Trans R Soc Trop (AHA-89-0l93). Med Hvg 87: 8 1—83. Authors' addresses: Sheila A. Peel, Patricia Bright, 13. Peters W, Robinson BL, 1991 . The chemotherapy Boyd Yount, and Ralph S. Baric, Department of Epi of rodent malaria. XLVI. Reversal of mefloqui demiology, CB 7400, McGavran-Greenberg Hall, Uni ne resistance in rodent Plasmodium. Ann Trop versity of North Carolina at Chapel Hill, Chapel Hill, Med Parasitol 85: 5—10. NC, 27599-7400. Jean Handy, CB 7600, Clinical Mi 14. Foote SJ, Thompson JK, Cowman AF, Kemp Di, crobiology and Immunology Laboratories, UNC Hos 1989. Amplification of the multidrug resistance pitals, Chapel Hill, NC 27599-7600. gene in some chloroquine-resistant isolates of P. falciparum. Cell 57: 92 1—930. 15. Wilson CM, Serrano AE, Wasley AM, Bogen REFERENCES schutz MR Shankar AH, Wirth DF, 1989. Am plification of a gene related to mammalian mdr genes in drug resistant Plasmodium falciparum. I . Peters W, 1985. The problem of drug resistance Science 244: 1184—1186. in malaria. Parasitologv 90: 705—7 15. 2. Oakes SC, Mitchel VS. Pearson GW, Carpenter 16. Hyde SC, 1990. Structural model of ATP-binding CCJ, 1991. Malaria. Obstacles and opportuni proteins associated with cystic fibrosis, multi ties. A Report of the Committee for the Stud's' on drug resistance, and bacterial transport. Nature 346: 362—365. Malaria Prevention and Control: Status Review and Alternative Strategies. Washington, DC: Di 17. Wilson CM, Volkerman 5K, Thaithong 5, Martin vision of International Health, Institute of Mcd RK, Kyle DE, Milhous WK, Wirth DF, 1993. icine, National Academy Press, 144—166. Amplification of pfmdrl associated with meflo 3. Gottesman MM, Pastan I, 1993. Biochemistry of quine and halofantnine resistance in Plasmodium multidrug resistance mediated by the multidrug falciparum from Thailand. Mo! Biochem Par transporter. Annu Rev Biochem 62: 385—427. asitol 57: 151—160. 4. Endicott JA, Ling V 1989. The biochemistry of 18. Foote Si, Kyle DE, Martin RK, Oduola AMJ, For P-glycoprotein-mediated multidrug resistance. syth K, Kemp Di, Cowman AF, 1990. Several Anna Rev Biochem 58: 137—171. alleles of the multidrug resistance gene are 5. Oduola AJ, Milhous WK, Weatherly NF, Bowdre closely linked to chloroquine resistance in P/as JH, Desjardins RE, 1988. Plasmodium fa/cipa modium falciparum. Nature 345: 255—258. rum: induction of resistance to melloquine in 19. Wellems TE, Panton U, Gluzman IY, do Rosario cloned strains by continuous drug exposure in VE, Gwadz RW, Walker-Jonah A, Krogstad Di, vitro. Exp Parasitol 67: 354—360. 1990. Chloroquine resistance not linked to mdr 6. Peel SA, Merritt SC, Handy J, Baric RS, 1993. like genes in a Plasmodium falciparum cross. Derivation of highly metloquine-resistant lines Nature 345: 253—255. from Plasmodium falciparum in vitro. Am J 20. Wellems TE, Walker-Jonah A, Panton Li, 1991. Trop Med. Hyg 48: 385—397. Genetic mapping of the chloroquine resistant lo 7. Krogstad DJ, Gluzman IY, Kyle DF, Oduola AMJ, cus on Plasmodium falciparum chromosome 7. Martin 5K, Milhous WK, Schlesinger PH, 1987. Proc Nat! Acad Sci USA 88: 3382—3386. Efflux of chioroquine from P. falciparum: 21. Barnes DA, Foote Si, Galatis D, Kemp Di, Cow mechanisms of chloroquine resistance. Science man AF, 1992. Selection for high-level chloro 238: 1283—1285. quine resistance results in deamplification of the 8. Martin SK, Oduola AMJ, Milhous WK, 1987. Re pfmdrl gene and increased sensitivity to meflo versal of chloroquine resistance in Plasmodium quine in Plasmodium falciparum. EMBO J 11: falciparum by verapamil. Science 235: 899— 3067—3075. 901. 22. Desjardins RE, Canfield Ci, Haynes JD, Chulay 9. Bitonti AJ, Sjoerdsma A, McCann PP. Kyle DE, JD, 1979. Quantitative assessment of antima Oduola AMJ, Rossan RN, Milhous K, Davidson lanaI activity in vitro by a semiautomated mi DE, 1988. Reversal of chloroquine resistance in crodilution techn ique. Antimicrob Agents the malarial parasite Plasmodium falciparum by Chemother 16: 710—718. desipramine. Science 242: 1301—1303. 23. Jobson KJH, Jennings MW, 1991. The structure 658 PEEL AND OTHERS

of the calmodulin gene of Plasnwdium falcipa relates of mefloquine resistance of Plasmodium rum. Mo! Biochem Parasito! 46: 19—34. fa!ciparum in eastern Thailand. Am J Trop Med 24. Kunter 5, Breuer WV, Ginsburg H, Aley SB, Ca Hyg 44: 553—559. bantchik ZI, 1985. Characterization of perme 37. Cowman AF, Karcz 5, Galatis D, Culvenor JO, ation pathways in the plasma membrane of hu 1991. A P-glycoprotein homologue of Plasmo man erythrocytes infected with early stages of diumfalciparum is located on the digestive vac Plasmodium fa!ciparwn: association with pam uole. J Ce!! Biol 113: 1033—1042. site development. J Ce!! Physiol 125: 52 1—527. 38. Choi K, Chen C, Kriegler M, Roninson lB. 1988. 25. Chomczynski P. Sacchi N, 1987. Single-step An altered pattern of cross-resistance in multi method of RNA isolation by acid guanidinium drug resistant human cells results from sponta thiocyanate-phenol-chloroform extraction. Anal neous mutations in the mdrl (P-glycoprotein) Biochem 162: 156—159. gene. Ce!! 53: 5 19—529. 26. Lockyer MJ, Schwartz RT, 1987. Strain variations 39. Roninson lB. 1992. The role of the MDR1 (P in the circumsporozoite protein gene of P/as glycoprotein) gene in multidrug resistance in modiumfalciparum. Mol Biochem Parasito! 22: vitro and in vivo. Biochem Pharmaco! 43: 95— 101—108. 102. 27. Kochan J, Perkins M, Ravetch JV, 1986. A tan 40. Gros P, Dhir R, Croop J, Talbot F, 1991. A single demly repeated sequence determines the binding amino acid substitution strongly modulates the domain for an erythrocyte receptor binding pro activity and substrate specificity of the mouse tein of P. falciparum. Cell 44: 689—696. mdrl and mdr3 efflux pumps. Proc Nat! Acad 28. Sambrook J, Fritsch EF, Maniatis T, 1989. Molec Sd USA 88:7289—7293. u/ar Cloning: A Laboratory Manual. Second 41. Anderson MP, Gregory Ri, Thompson S. Souza Edition. Cold Spring Harbor, NY: Cold Spring DW, Sucharita P. Mulligan RC, Smith AE, Harbor Press. Welsh Mi, 1991. Demonstration that CFTR is 29. Southern EM, 1975. Detection of specific se a chloride channel by alternation of its anion quences among DNA fragments separated by specificity. Science 253: 202—205. gel electrophoresis. J Mo! Biol 98: @03—517. 42. Dhir R, Gros P. 1992. Functional analysis of chi 30. Kafatos FC, Jones WC, Efstratiadis A, 1979. De meric proteins constructed by exchanging ho termination of nucleic acid sequence homolo mologous domains of two P-glycoprotein con gies and relative concentrations by dot blot hy ferring distinct drug resistance profiles. bridization procedure. Nucleic Acids Res 7: Biochemtry 31: 6103—6110. 1541—1552. 43. Curnier SJ, Kane SE, Willingham MC, Cardarelli 31. Volkerman 5K, Wilson CM, Wirth DF, 1993. CO. Pastan I, Gottesman MM, 1992. Identifi Stage specific transcripts of the Plasmodiumfal cation of residues in the first cytoplasmic loop of P-glycoprotein involved in the function of ciparum pfindrl gene. Mo! Biochem Parasito! chimeric human MDRI-MDR2 transporters. J 57: 203—212. Biol C/oem267: 25153—25159. 32. Nateghpour M, Ward SA, Howells RE, 1993. De 44. Raviv Y, Pollard RB, Bruggemann EP, Pastan I, velopment of halofantrine resistance and deter Gottesman MM, 1990. Photosensitized labeling mination of cross-resistance patterns in Plas of a functional multidrug transporter in living modium falciparum. Antimicrob Agents drug resistant tumor cells. J Biol Chem 265: Chemother 37: 2337—2343. 3975—3980. 33. Webster HK, Thaithong 5, Pavanand K, Yongvan 45. Yayon A, Cabantchik ZI, Ginsburg H, 1984. Iden itchit K, Pinswasdi C, Boudreau EF, 1985. tification of the acidic compartment of Plas Cloning and characterization of mefloquine-re modium falciparum-infected human erythro sistant Plasmodium fakiparum from Thailand. cytes as the target of the antimalarial drug Am J Trop Med Hyg 34: 1022—1027. chloroquine. EMBO J 3: 2695—2700. 34. Webster HK, Boudreau EF, Pavanand K, Yongvan 46. Krogstad DJ, Schlesinger PH, Gluzman IY, 1985. itchit K, Pang LW, 1985. Antimalarial drug sus Antimalarials increase vesicle pH in Plasmodi ceptibility testing of Plasmodium falciparum in urn falciparum. J Ce!! Biol 101: 2302—2309. Thailand using a microdilution radioisotope 47. Ginsberg H, Krugliak M, 1992. Quinoline-con method. Am J Trop Med Hyg 34: 228—235. taming antimalarials-mode of action, drug resis 35. Basco LK, Le Bras J, 1992. In vitro activity of tance and reversal. An update with unresolved halofantrine and its relationship to other stan puzzles. Biochem Pharmaco! 43: 63—70. dard antimalarial drugs against African isolates 48. Bray P0, Howells RE, Ritchie GY, Ward SA, and clones of Plasmodium falciparum. Am J 1992. Rapid chloroquine efflux phenotype in Trop Med Hyg 47: @21—527. both chloroquine-sensitive and chloroquine-re 36. Childs GE, Boudreau EF, Wimonwattratee T, Pang sistant Plasmodium fa!ciparum@ Biochem Phar L, Milhous WK, 1991. In vitro and clinical cor macol 44: 1317—1324.