A critical role for PfCRT K76T in Plasmodium falciparum verapamil-reversible chloroquine resistance

Abstract

Chloroquine resistance (CQR) in Plasmodium falciparum is associated with mutations in the digestive vacuole transmembrane protein PfCRT. However, the contribution of individual pfcrt mutations has not been clarified and other genes have been postulated to play a substantial role. Using allelic exchange, we show that removal of the single PfCRT amino-acid change K76T from resistant strains leads to wild-type levels of CQ susceptibility, increased binding of CQ to its target ferriprotoporphyrin IX in the digestive vacuole and loss of verapamil reversibility of CQ and quinine resistance. Our data also indicate that PfCRT mutations preceding residue 76 modulate the degree of verapamil reversibility in CQ-resistant lines. The K76T mutation accounts for earlier observations that CQR can be overcome by subtly altering the CQ side-chain length. Together, these findings establish PfCRT K76T as a critical component of CQR and suggest that CQ access to ferriprotoporphyrin IX is determined by drug-protein interactions involving this mutant residue.

Figure 1

Allelic exchange strategy and molecular characterization of recombinant clones. (A) Schematic depicting integration of the pK76 plasmid into the endogenous pfcrt locus by homologous recombination and single-site crossover. The recombinant downstream locus contained a full-length functional pfcrt gene, transcriptionally controlled by a shortened pfcrt promoter and the endogenous 3′UTR. A truncated pfcrt remnant (with the first 5 of 13 exons), the bsd selectable marker and pBluescript (pBS) sequence were located upstream. Square brackets delineate the plasmid sequence that could integrate as multiple tandem copies (n). Fragments obtained upon restriction digestion, and pfcrt or bsd probe locations, are indicated. B, BglII; S, StuI; X, XbaI. (B) PCR detection of recombinant functional (downstream) pfcrt with primers p3/p4 (top panel), the truncated (upstream) remnant with primers p5/p6 (middle panel) or endogenous pfcrt with primers p5/p4 (bottom panel). Lanes: 1: Dd2; 2: T76K-1^Dd2; 3: C-1^Dd2; 4: 7G8; 5: T76K-1^7G8; 6: C-1^7G8; 7: TMD1-1; M: 1 kb DNA ladder. Lane identities are maintained throughout the figure. (C, D) Southern hybridization of genomic DNA digested with BglII/StuI/XbaI (C) or BglII/StuI (D), and probed with pfcrt or bsd. (E) RT–PCR assays with primers p7/p9 revealed a 1.1 kb transcription product from the functional downstream locus (recombinant lines) and the endogenous locus (parental lines). Primers p7/p6 gave a 1.0 kb transcription product from the upstream truncated pfcrt remnant (in the recombinant but not parental lines). (F) Northern hybridization of total RNA probed with pfcrt or ef-1α. The RNA gel used for blotting shows equivalent loading. Data shown are representative of four separate Northern analyses performed on synchronized parasites. (G) Western blot of protein samples from recombinant and parental lines probed with antibodies to PfCRT or the Golgi marker PfERD2. Total protein amounts were loaded in two-fold dilutions, with three sets of dilutions per line (see Supplementary data).

Figure 2

Antimalarial susceptibility profiles of pfcrt-modified lines on the Dd2 genetic background, showing a graphical representation of IC₅₀ values (mean±s.e.m.) (values indicated in Supplementary Table I). Note that the m-dCQ graphs have a two-segment Y-axis to adequately represent the range of values. For graphs with left and right Y-axes (with different scales), the left side (blacks bars) corresponds to QD or ADQ, whereas the right side (gray bars) corresponds to MFQ or AMT. Values are expressed in nM for all drugs, except AMT for which the unit is μM. VP was included at 0.8 μM. Each mean value was calculated from 3–14 assays performed in duplicate. Determinations of statistical significance used unpaired two-tailed t-tests, with the P-value reporting the lesser of the significant values obtained when comparing back-mutant lines against each of the two control recombinant lines. ^*P<0.05; ^**P<0.01; ^***P<0.001. The haplotypes of all lines are listed in Table I.

Figure 3

Antimalarial susceptibility profiles of pfcrt-modified lines on the 7G8 genetic background. Data are presented as described in Figure 2 legend.

Figure 4

Effect of PfCRT mutations on parasite susceptibility to CQ side-chain analogs. Shown are the mean±s.e.m. IC₅₀ values (in nM) for recombinant and nontransfected lines tested with CQ and the side-chain diaminoalkane analogs AQ13, AQ26, AQ33 and AQ40 (De et al, 1996). For each line, statistical comparisons between individual analogs and CQ were performed using unpaired, two-tailed t-tests (see Supplementary Table II). ^*P<0.05; ^**P<0.01; ^***P<0.001.

Figure 5

CQ accumulation at equilibrium in recombinant and parental lines. For each line, the curve of best fit is shown for the saturable uptake of [³H]CQ in infected erythrocytes ([CQ]_int), expressed as a function of varying extracellular concentrations of unlabeled CQ. Mean±s.e.m. data (from 6 independent experiments) and apparent K_d values are shown for (A) Dd2, (B) 7G8 and (C) GC03 sets of recombinant and parental lines. (D) Apparent K_d values were plotted against CQ IC₅₀ values. Log₁₀-transformed data, line of best fit and correlation coefficient are indicated.

Figure 6

CQ–FP binding in recombinant and parental lines. Levels of [³H]CQ binding to its FP target, expressed as picomoles of CQ per micromole of FP, are shown for (A) Dd2, (B) 7G8 and (C) GC03 sets of recombinant and parental lines. Data represent means±s.e.m., calculated from 6–8 independent experiments. (D) CQ–FP binding values were plotted against CQ IC₅₀ values calculated in the absence or presence of VP. Log₁₀-transformed data, line of best fit and correlation coefficient are indicated.