We have studied in parallel three P. falciparum clones, one MQ sensitive (W2) and two (3D7 and FCB) with decreased sensitivity. The pfmdr1, pfcrt, pfmrp1 and pfmrp2 ORF were fully sequenced to determine the genetic variability between the clones. The three clones, show two different main pfmdr1 polymorphisms: a single nucleotide polymorphism at amino acid position 86 (N86 in 3D7, 86Y in W2 and FCB) and gene copy number variation (W2 and 3D7 harbouring one copy [25], with FCB carrying two copies [23]). Numerous single nucleotide polymorphism differences were seen in the three remaining genes (GenBank accession numbers: GU797309, GU797310, GU797311, GU797312, GU797315, GU797316).

In vitro MQ 50% inhibitory concentration (IC₅₀) of W2, 3D7 and FCB clones were in average from 6 independent assays 10±3.3nM, 50±16.9nM and 50±14.1nM respectively. As for IC₉₉ the average results was 44±13.9nM, 146±36.1nM, 146±50.6nM for W2, 3D7 and FCB respectively. These concentrations were used for challenging the clones in the assay performed.

Exposure to MQ induced an unexpected and unambiguous delay in the cell cycle progression of the three tested parasites, as evaluated by Giemsa staining under light microscopy for the four counted stages (Figure 1). The degree of delay effect was clone dependent.

At MQ IC₉₉ a pronounced morphology arrest was seen in the three clones remaining essentially at the experimental initial ring stage with picnotic features. The quantification of end point RNA indicated the presence of viable cells, an observation supported for all tested parasites by their growth recovery 7–10 days after drug withdrawal (Figure S1).

Each gene has its own stage morphology expression profile, which once induced by drugs, it was previously shown that the expression is strictly correlated with the cell morphology [26], [27]. The MQ induced cell morphology delay was also detectable through the analysis of the transcript accumulation patterns (Figure 2, green and red curves). Applying the described non-linear regression model to the cycle (figure 3), significant differences in the degree of cell morphology progression were seen with MQ IC₉₉ exposure in the three clones (p<0.05; figure 4: red lines). The calculated rate at IC₅₀, compared to control was clone dependent. The effect was less pronounced in W2 parasites with approximately 15% delay (p = 0.23) vs. ∼40% for the MQ tolerant clones FCB and 3D7 (p<0.01) (Figure 4 and Table 1). Similar cell cycle morphology delay events were also observed upon quinine exposure (IC₅₀ and IC₉₀) in exploratory 12 hour follow up assay (Figure S2).

The pfmdr1, pfcrt, pfmrp1 and pfmrp2 cell cycle patterns of relative transcript abundance from control (i.e. non drug exposed) experiments for the reference 3D7 clone were similar to the available data at PlasmoDB [28] (Figure 2, black curves), using as endogenous control seryl-tRNA synthetase gene. In brief, pfmdr1, pfcrt and pfmrp2 displayed the lowest levels of expression at approximately 24 hours (middle/late trophozoite stage) and highest amplitudes around 42–48 hours post invasion. pfmrp1 revealed the previously described clearly different expression profile [29], [30], with a peak of expression during the development of trophozoites (12–24 hours post-invasion).

Comparing relative transcript abundance (R±SD) between the three clones at onset (Figure 2, time-point 0h), pfmdr1 mRNAs showed higher levels in the FCB clone (18.15±2.84), as compared to W2 (14.18±0.99) and 3D7 (4.90±0.21) with a fold difference of 1.3 and 3.7 respectively. Pfcrt showed to be less expressed in the W2 (3.86±0.15) compared with 3D7 (7.40±1.90) and FCB (6.10±0.88) with a fold difference of 0.5 and 0.63 respectively. As for the mrp genes, pfmrp2 gene showed higher expression for 3D7 (8.40±1.95) compared with W2 (5.06±0.13) and FCB (3.81±1.03) with a fold difference of 1.7 and 2.2 respectively while pfmrp1 at time-point 0 the results was 3.5 fold difference between FCB (0.007±9^E-4) and W2 (0.002±1^E-4) and 1.75 fold comparing with 3D7 (0.004±6^E-4).

The fact that the drug exposure led to a proliferative delay in the parasites stage created a new challenge on how to differentiate gene transcription levels affected by cell morphology from the direct action of the drug on the transcript accumulation of the studied genes. We solved this challenge through two approaches: (1) From the few possible comparative time-points (control vs. MQ exposure time-points that had statistically non distinguishable parasite morphology proportion differences (p>0.05)), transcript abundance values were compared. The levels of significant gene expression modulation, within each clone upon MQ exposure, ranged from 0.6 to 5.8 fold: pfmdr1 (0.6–1.5 fold), pfcrt (0.8–1.3 fold), pfmrp2 (0.8–2.7 fold) and pfmrp1 with the highest MQ driven induction (0.7–5.8) (Table 2). (2) To complement this limited comparative point by point verification of drug direct action, which gave rise to few analyzable points, another analytical approach based on a non-linear regression analysis of gene expression over time was conducted for the MQ IC₅₀ experiments. The calculated fold difference confirmed that the drug exposure was generally associated with mild changes in the expression of the genes (figure 5). Considering each gene at IC₅₀ exposure: (a) The pfmdr1 gene, was significantly more induced in the less MQ sensitivity clones 3D7 (1.35 fold, p = 0.029) and FCB (1.49 fold, p<0.001), than in the more sensitive W2 (1.23 fold). Comparisons between the two less sensitive parasites showed that FCB can induce this gene significantly more than 3D7 (p = 0.003). (b) As for pfmrp1, our data show no statistical difference between the clones. (c) Concerning pfcrt gene, the W2 and 3D7 parasites appear to lack the capacity to significantly induce pfcrt upon MQ exposure (1.06 fold and 0.92 fold) compared to FCB (1.25 fold). (d) Pfmrp2, appears to significantly induce the less MQ sensitivity clones 3D7 (1.32 fold, p = 0.003) and FCB (1.46 fold, p = 0.023) compared with W2 (0.96 fold).

P. falciparum 3D7 clone was obtained from Prof. D. Walliker (Department of Animal and population genetics, University of Edinburgh, UK), while parasite clones W2 (MRA-157) and FCB (MRA-309) were obtained from Malaria Research and Reference Reagent Resource Center (MR4, ATCC Massanas Virginia). The parasite lines were selected on the basis of their sensitivity to mefloquine hydrochloride (Sigma-Aldrich®, St.Louis, MO, USA) and their genetic background (sequences submitted to GenBank accession numbers: GU797309, GU797310, GU797311, GU797312, GU797313, GU797314, GU797315, GU797316).

The inhibitory concentration of 50 and 99% for the three clones was assessed using an Histidine-Rich Protein 2 based Double-Site Sandwich Enzyme-Linked Immunosorbent Assay [38] followed by nonlinear regression analysis (http://malaria.farch.net).

The parasites were kept in culture using conventional methods [39] in human O⁺ RBCs and Malaria Culture Medium, containing RPMI 1640 culture medium supplemented with 10% L-glutamine, 0.05% gentamicine (Gibco®/Invitrogen™, Carlsbad, CA, USA) and 10% human AB⁺ serum. Large scale culture of each clone was produced with 3% parasitaemia in 2.5% hematocrite and synchronized in early ring stage at least twice by applying the MACS® system (Miltenyi Biotec, Bergisch Gladbach, Germany). The batch was then divided for three experiments: one control culture (no drug added) and two cultures with MQ IC₅₀ and IC₉₉.concentration followed by spreading it in culture plates (2mL/well). At the time of reinvasion, early rings (time-point 0), two Giemsa-stained smears were made for examination of parasites stages, while three replicate wells of parasite culture were harvested for RNA extraction. This procedure was followed for the three experiments, in six hour time-point intervals up to 48 hours (total of 9 time-points analyses including time 0). The smears were fixed with methanol and stained with 5% Giemsa and visualization under light microscopy (1000× magnification). Their examination included counting 100 parasites, twice per slide (2 slides per time point), distinguishing four main morphological stages: early rings, late rings/early trophozoites, trophozoites and schizonts. A total count of 400 parasites per time-point, per experiment was performed. The experimental design was equally applied for the three clones analyzed (W2, 3D7 and FCB).

Parasite cultures collected for RNA extraction were centrifuged at 0.8 g-force for 2 min. 1.85 mL of supernatant were removed and 150 µL of PBS plus 300 µL of Lysis buffer (Applied Biosystems™, Fresno, CA, USA) were added. The mix was kept at −20°C until RNA extraction. RNA extraction was carried out using an ABIPRISM®6100 Nucleic Acid PrepStation® (Applied Biosystems™, Fresno, CA,USA) according to the recommendations of the manufacturer. Total RNA quality and quantity was measured using the AgilentRNA 6000 Pico total RNA assay in an Agilent2100 Bioanalyser™ (Santa Clara, CA, USA). Sample concentrations with RNA Integrity Number >5 were normalized for each clone before cDNA synthesis (High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Fresno, CA, USA)). Real-time analyses were performed in an ABIPRISM® 7900HT Sequence Detection System (Applied Biosystems™, Fresno, CA, USA) with gene specific MGB TaqMan® probes and primers. TaqMan® probes and primers for target gene pfmdr1 were as previously published [22]. New designs were developed for pfcrt, pfmrp1, pfmrp2 and for endogenous control gene [30], [40] seryl-tRNA synthetase (PF07_0073) (Table S1). Amplification reactions were done in quadruplicate in 384 well plates with 10µl containing TaqMan® Gene Expression Mastermix (Applied Biosystems™, Fresno, CA, USA), 300nM of each forward and reverse primer, 100nM of TaqMan®probe and 2 µL of amount-normalized cDNA. The thermal cycle profile was 50°C for 2 min, 95°C for 10 min and forty cycles of 95°C for 15 s and 60°C for 1 min. The amplification efficiency (E) was estimated for each gene and used as a correction factor for relative gene expression quantification. All experimental threshold cycle values (Ct) were first transformed to adjust the RNA concentration adding to the Ct value the log₂ RNA concentration of each clone. Relative gene expression was calculated as the ratio (R) between the transformed Ct values of target gene and internal control reference gene (PF07_0073), taking in account the amplification efficiency for each gene [41]. Calculations were conducted using the SAS 9.1 system for Windows™.

In order to distinguish the changes in cDNA abundance associated with the parasite cell cycle progression from the ones associated with drug exposure, two different approaches were performed: