OUP user menu

Adaptive response to the antimalarial drug artesunate in yeast involves Pdr1p/Pdr3p-mediated transcriptional activation of the resistance determinants TPO1 and PDR5

Marta Alenquer, Sandra Tenreiro, Isabel Sá-Correia
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00095.x 1130-1139 First published online: 1 December 2006


The expression of the transcription regulator Pdr1p and its target genes PDR5 and TPO1 is required for Saccharomyces cerevisiae adaptation and resistance to artesunate, a promising antimalarial drug, also active against tumour cells and viruses. PDR5 and TPO1 encode plasma membrane multidrug transporters of the ATP-binding cassette and the major facilitator superfamilies, respectively. The transcriptional activation of TPO1 (10-fold) and PDR5 (13-fold) was registered after 30 min of exposure of the unadapted yeast population to acute artesunate-induced stress, being significantly reduced in the absence of Pdr1p and abolished in the absence of Pdr1p and Pdr3p. Maximum TPO1 mRNA levels were rapidly reduced to basal values following adaptation of the yeast population to artesunate, while high PDR5 levels were maintained during drug-stressed exponential growth.

  • artesunate
  • multidrug resistance transporters
  • stress response
  • Saccharomyces cerevisiae
  • TPO1
  • PDR5
  • Pdr1p/Pdr3p


The therapeutic potential of drugs is seriously limited by the manifestation of cellular drug resistance. The multidrug efflux transporters, found in the plasma membrane of all living cells, recognize a wide variety of structurally dissimilar hydrophobic organic compounds and actively extrude them from the cytoplasm to the outer medium, providing protection from chemotherapeutic agents. Artesunate ((3R,5aS,6R,8aS,9R,10S,12R,12aR)-decahydro-3,6,9-trimethyl-3,12-epoxy-12H-pyrano[4,3-j]-1,2-benzodioxepin-10-ol, hydrogen succinate; Fig. 1) is a semisynthetic derivative of the sesquiterpene artemisinin, the active principle of Artemisia annua herb. Artesunate and other artemisinin derivatives are very powerful drugs in the treatment of malaria, with very few side effects (Woodrow, 2005). In addition to its remarkable activity against multidrug-resistant Plasmodium strains, artesunate has a profound cytotoxic action on a variety of cancer cell lines (Efferth, 2001) and possesses antiviral activity, in particular against ganciclovir-resistant human cytomegaloviruses (Efferth, 2002). Clinically relevant artesunate resistance in Plasmodium falciparum has not been demonstrated (Woodrow, 2005). However, it is likely to occur, since decreased susceptibility to artemisinin compounds has been seen in the laboratory (Walker, 2000). In this work, we searched for determinants and mechanisms of resistance to artesunate in the yeast Saccharomyces cerevisiae. Although this drug acts on specific targets of the malaria parasite or other infectious agents that may not exist in yeast, many of the mechanisms underlying multidrug resistance (MDR), in particular those involving a multiplicity of mutidrug transporters, are apparently conserved among phylogenetically distant organisms. Therefore, the characterization of determinants and mechanisms of resistance to this drug in the experimental eukaryotic model S. cerevisiae may contribute to the understanding of the underlying mechanisms in the more complex and less genetically accessible P. falciparum.


Chemical structure of the antimalarial drug artesunate (3R,5aS,6R,8aS,9R,10S,12R,12aR)-decahydro-3,6,9-trimethyl-3,12-epoxy-12H-pyrano[4,3-j]-1,2-benzodioxepin-10-ol, hydrogen succinate).

Among the genes of the so-called pleotropic drug resistance (PDR) network in yeast, we proved that the transcription regulators Pdr1p and Pdr3p and their target genes PDR5 and TPO1, encoding two multidrug transporters of the ABC (ATP-binding cassette) superfamily and of the major facilitator superfamily (MFS), respectively, are required for the adaptive response and resistance to artesunate. Pdr1p and Pdr3p are Zn(II)2Cys6 zinc finger transcription factors, sharing an overall amino acid identity of 36%. These major regulators of the PDR network in yeast have been the subject of intense research, including several genome-wide expression studies that led to the identification of many putative target genes (DeRisi, 2000; Devaux, 2001). These genes mainly encode multidrug transporters and other permeases and proteins involved in lipid and cell wall metabolism (DeRisi, 2000; Devaux, 2001). Tpo1p is a plasma membrane MDR transporter of the MFS required for yeast resistance to polyamines and thought to catalyse polyamine excretion (Albertsen, 2003). This MDR determinant also confers resistance to cycloheximide and quinidine (do Valle Matta, 2001), mycophenolic acid (Desmoucelles, 2002), caspofungin (Markovich, 2004) and the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) (Teixeira & Sá-Correia, 2002). Pdr5p is one of the best characterized yeast ABC transporters, conferring resistance to a very broad range of amphiphilic substrates (Kolaczkowski, 1996). Remarkably, Pdr5p and mammalian ABC P-glycoproteins (P-gp), implicated in MDR phenotypes of tumour cells towards anticancer drugs, are functional homologues sharing numerous drug substrates (Wolfger, 2001). In this work we present results indicating that the yeast response to artesunate stress involves the Pdr1p- and Pdr3p-mediated transcriptional activation of TPO1 and PDR5 and we characterize the adaptive response exhibited by an unadapted population throughout the different phases of adaptation and growth, following sudden exposure to the drug.

Materials and methods

Strains, basal growth media and general methods

The S. cerevisiae strains used in this work are listed in Table 1. Cells were batch-cultured at 30°C, with orbital agitation (250 rev min−1) in differently supplemented basal growth media (pH 4.5) with the following composition (per litre): 1.7 g of yeast nitrogen base without amino acids or NH4+ (Difco), 20 g of glucose and 2.65 g of (NH4)2SO4 (Sigma). For strain BY4741 and the derived mutants, the growth medium MM4 (pH 4.5) comprised basal medium supplemented with 20 mg of methionine, 20 mg of histidine, 60 mg of leucine and 20 mg of uracil (all from Sigma). For strain FY1679-28C and the derived mutants, the growth medium MM6 (pH 4.5) comprised basal medium supplemented with 20 mg of histidine, 60 mg of leucine, 40 mg of tryptophan (Sigma) and 20 mg of uracil. To maintain selective pressure on recombinant cells, MM4 and MM6 media without uracil, MM4-U and MM6-U, respectively, were used. Solid MM4 medium contained, besides the above indicated ingredients, 20 g of agar per litre (Iberagar).

View this table:

Saccharomyces cerevisiae strains used in this study and comparative levels of susceptibility to artesunate determined by spot assays

StrainGenotype/descriptionArtesunate susceptibilitySource/Reference
BY4741MATa, his3Δ1, leu2Δ0, met15Δ0, ura3Δ0Euroscarf
Transcription regulators
BY4741Δpdr1BY4741 pdr1::kanMX4+++Euroscarf
BY4741Δpdr3BY4741 pdr3::kanMX4Euroscarf
BY4741Δpdr8BY4741 pdr8::kanMX4Euroscarf
BY4741Δyrr1BY4741 yrr1::kanMX4Euroscarf
BY4741Δyrm1BY4741 yrm1::kanMX4Euroscarf
BY4741Δmsn2BY4741 msn2::kanMX4+Euroscarf
BY4741Δmsn4BY4741 msn4::kanMX4Euroscarf
BY4741Δyap1BY4741 yap1::kanMX4+Euroscarf
BY4741Δwar1BY4741 war1::kanMX4Euroscarf
BY4741Δskn7BY4741 skn7::kanMX4Euroscarf
MFS–MDR transporters
BY4741Δaqr1BY4741 aqr1::kanMX4Euroscarf
BY4741Δatr1BY4741 atr1::kanMX4Euroscarf
BY4741Δazr1BY4741 azr1::kanMX4Euroscarf
BY4741Δdtr1BY4741 dtr1::kanMX4Euroscarf
BY4741Δflr1BY4741 flr1 ::kanMX4Euroscarf
BY4741Δqdr1BY4741 qdr1 ::kanMX4Euroscarf
BY4741Δqdr2BY4741 qdr2::kanMX4Euroscarf
BY4741Δqdr3BY4741 qdr3::kanMX4Euroscarf
BY4741Δsge1BY4741 sge1::kanMX4Euroscarf
BY4741Δtpo1BY4741 tpo1::kanMX4++Euroscarf
BY4741Δtpo2BY4741 tpo2::kanMX4Euroscarf
BY4741Δtpo3BY4741 tpo3::kanMX4Euroscarf
BY4741Δtpo4BY4741 tpo4::kanMX4Euroscarf
BY4741Δyhr048wBY4741 yhr048w::kanMX4Euroscarf
ABC superfamily transporters
BY4741Δbpt1BY4741 bpt1::kanMX4Euroscarf
BY4741Δpdr5BY4741 pdr5::kanMX4++Euroscarf
BY4741Δpdr10BY4741 pdr10::kanMX4Euroscarf
BY4741Δpdr11BY4741 pdr11::kanMX4Euroscarf
BY4741Δpdr12BY4741 pdr12::kanMX4Euroscarf
BY4741Δpdr15BY4741 pdr15::kanMX4Euroscarf
BY4741Δsnq2BY4741 snq2::kanMX4Euroscarf
BY4741Δybt1BY4741 ybt1::kanMX4Euroscarf
BY4741_Δycf1BY4741 ycf1::kanMX4Euroscarf
BY4741_Δyor1BY4741 yor1::kanMX4Euroscarf
FY 1679-28CMATa, ura3-52, leu2-Δ1, trp1-Δ63, his3Δ200, GAL2+NTA. Goffeau
FY 1679-28C/ECFY 1679-28, pdr1-Δ2::TRP1NTC. Jacq (Delaveau, 1994)
FY 1679-28c/TDECFY 1679-28C, pdr1-Δ2::TRP1, pdr3Δ::HIS3NTC. Jacq (Delaveau, 1994)
  • * –, Susceptibility identical to that of wild-type strain.

  • +,++ and +++, increased susceptibility levels compared to wild-type strain BY4741; NT, not tested.

Cloning procedures and transformation of yeast cells were carried out using standard methods or as described previously (Brôco, 1999).

Artesunate susceptibility assays

The susceptibility of the deletion mutants to artesunate was first compared with wild-type susceptibility by spot assays. Artesunate was kindly provided by Dafra Pharma (Belgium). The product carries a free carboxylic group (COOH) (Fig. 1), the suffix ‘ate’ in artesunate referring to the ester function (information obtained from Dafra Pharma). The drug was first dissolved in ethanol 70% v/v. Ethanol concentration in the growth media (including the control medium lacking the artesunate) was kept below 0.7% (v/v) to avoid growth inhibition due to the solvent.

Cell suspensions used to inoculate the agar plates were mid-exponential cells grown without the drug, until culture OD600 nm=0.4±0.02 was reached, and then diluted in sterile water to obtain suspensions with OD600 nm=0.025±0.005. These cell suspensions and the subsequent dilutions (1 : 5; 1 : 10) were applied as spots (4 μL) onto the surface of the agarized media (at pH 4.5), supplemented with adequate concentrations of artesunate.

Susceptibility assays were also carried out in liquid MM4 minimal medium (at pH 4.5), supplemented with adequate concentrations of artesunate, at 30°C, with orbital agitation (250 r.p.m.). Growth media were inoculated with mid-exponential cells grown in the absence of artesunate. Cell growth was followed by measuring culture OD600 nm. The concentration of viable cells during batch cultivation was assessed as the number of CFU on YPD (20 g of glucose, 20 g of yeast extract, 10 g of Bacto peptone per litre) rich medium agar plates, incubated at 30°C for 2 days.

Northern blot analysis

The effect of yeast exposure to artesunate on TPO1 and PDR5 transcription was examined based on Northern blot experiments. RNA extraction from yeast cells and Northern blot hybridizations were carried out as described previously (Tenreiro, 2001). The total RNA in each sample used for Northern blotting was approximately constant (15 μg, based on A260 nm). The specific DNA probes for TPO1 and PDR5 transcripts were prepared by PCR amplification using the following primers: 5′-TCGGATCATTCTCCCATTTC-3′ and 5′-GAAGAACACATAGCAATGGCA-3′ for TPO1; and 5′-CATACCTTAGAATGAAATCCA-3′ and 5′-ACTTTTGTTGGGTGCCGATT-3′ for PDR5. The first probe consists of a 500-bp fragment of the TPO1 coding region, showing no significant homology with other regions of the yeast genome. The second probe, corresponding to the PDR5 transcript, includes 200 bp of the gene's promoter region and the first 200 bp of the coding region, showing no significant homology to the rest of the genome. The ACT1 mRNA level was used as an internal control; the ACT1 probe was prepared as described before (Tenreiro, 2001). The specificity of the probes was confirmed using RNA extracts from Δtpo1 and Δpdr5 deletion mutant cells. Hyperfilm MP (Amersham Biosciences) films were exposed to nitrocellulose membranes and incubated with an intensifying screen at −70°C. The relative intensities of the hybridization signals in the autoradiograms were quantified by densitometry (Image Scanner and Image Master 1D Elite; Amersham Biosciences).

Construction of the TPO1–lacZ expression fusion and β-galactosidase expression assays

Levels of TPO1 and PDR5 expression were also estimated based on β-galactosidase activity from TPO1–lacZ or PDR5–lacZ fusion plasmids, respectively, present in the cells of BY4741, FY1679-28C and derived deletion mutants. The PDR5lacZ (pKV2 plasmid), SNQ2lacZ (pEAE8 plasmid) and YOR1lacZ (pSM109 plasmid) fusions were prepared previously by others (Katzmann, 1994, 1995; Decottignies, 1995) and were kindly provided by A. Goffeau.

The TPO1–lacZ expression fusion was prepared in this work in the centromeric plasmid YCpAJ152, used previously to prepare other lacZ fusions (Brôco, 1999; Tenreiro, 2002), and a PCR product with the TPO1 promoter region, extending from position −1000 to +38. This PCR product was obtained using two hybrid primers (5′-GTCTTCAAGAATTCTCATGTTTGACAGCTTATCATCGATAAGCTTCAATAATGACTAAAAGGTGA-3′ and 5′-GTAACGCCAGGGTTTTCCCAGTCACGACGTTGTAAAACGGGATCCTGATTTTCCTTATTAGAAAT-3′), which introduced, respectively, a 5′-region complementary to the plasmid region preceding the lacZ gene and a 3′-region complementary to the beginning of the lacZ gene (underlined sequences). Cells of S. cerevisiae FY1769-28c were cotransformed with the YCpAJ152 plasmid, previously linearized with BamHI, and the PCR amplification product. The URA3+ clones resulting from the homologous recombination between the two DNA fragments were selected and the harboured plasmid was rescued and sequenced to confirm that the first codons of the TPO1 coding region were in frame with the lacZ gene and that no mutation was present in the promoter regions.

Cells transformed with TPO1–lacZ, PDR5–lacZ, SNQ2–lacZ and YOR1–lacZ were cultivated in artesunate-supplemented and -unsupplemented selective media, and growth was followed by measuring culture OD600 nm. Culture samples were harvested at adequate time intervals and assays were based on the method of Miller as previously described (Brôco, 1999). The enzyme specific activity units (U) – Miller units – were defined as the increase in A420 min−1 (OD600 nm)−1× 1000. The lacZ expression profiles shown are representative of more than two independent growth experiments carried out for each condition.


Pdr1p transcription factor and their target genes TPO1 and PDR5 are determinants of yeast resistance to artesunate

This study began with a comparison of the susceptibility to artesunate of the wild-type strain BY4741 and 29 single-deletion mutants for genes encoding the main transcription regulators of the PDR phenomenon, Pdr1p and its homologues Pdr3p, Pdr8p, Yrr1p and Yrm1p, and several MDR transporters of the ABC superfamily and the MFS. All tested mutants and the results of these susceptibility assays, based on spot tests, are shown in Table 1. These spot assays were followed by a comparison of the inhibitory effect of artesunate in the growth curves of the deletion mutants compared to the wild-type strain in the same liquid media. Among the genes encoding the above-mentioned transcription factors, only PDR1 proved to be a determinant of yeast resistance to artesunate by spot assays (Table 1 and Fig. 2). However, in liquid medium, PDR3 expression exerted some protection, although the protective effect of PDR1 expression was much higher (results not shown). Among the genes coding for MDR transporters whose protective effect was tested by spot assays, only TPO1, encoding a transporter of the MFS, and PDR5, encoding an ABC transporter, are required for yeast resistance to artesunate (Table 1 and Fig. 2). Interestingly, both TPO1 and PDR5 are known targets of Pdr1p and Pdr3p. The registered growth curves and cell viability counts of Δtpo1 and Δpdr5 cell populations, during growth in liquid medium supplemented with 0.6 g L−1 artesunate, revealed significant differences between the susceptibility of the different populations to the drug (Fig. 2). For 0.6 g L−1 artesunate, after an initial loss of cell viability that was more pronounced for the Δtpo1 population than for the wild-type population, the adapted cell population with TPO1 deleted resumed exponential growth under artesunate stress, although with a specific growth rate below the one exhibited by the wild-type population (Fig. 2). For the same artesunate concentration, the Δpdr5 population exhibited a more prominent initial loss of cell viability, followed by a short period of cell division that stopped prematurely when cell culture density was still far from reaching the cell concentration obtained with the wild-type population or the Δtpo1 population (Fig. 2). These results suggest that, under the experimental conditions used in this study, the level of protection provided by PDR5 expression is above that associated with TPO1 expression. A different conclusion was first suggested by results from spot assays (Fig. 2 and results not shown), but the reason behind this behaviour is unknown.


PDR1, PDR5 and TPO1 expression is required for artesunate resistance in yeast. Comparison of the susceptibility to artesunate of Saccharomyces cerevisiae BY4741 parental strain and of the deletion mutants Δpdr1, Δpdr3, Δtpo1 and Δpdr5 by spot assays in MM4 solid medium. The effects of artesunate on the growth curves of the wild-type strain (●,○) and Δtpo1 (▪,◻) and Δpdr5 (▲,△) were also compared in MM4 medium (black symbols) or in this medium supplemented with 0.6 g L−1 artesunate (open symbols). Cell suspensions used to prepare the spots in (b) and (c) were 1 : 5 and 1 : 10 dilutions of the cell suspension used in (a), respectively. Growth curves were followed by measuring culture OD600 nm (a, b) and cell viability (c, d) and are representative of at least three independent growth experiments leading to similar results. Cell suspensions used as inocula were exponential-phase cells cultivated in the absence of drug.

Transcriptional activation of PDR5 and TPO1 during yeast growth in the presence of artesunate

The levels of transcripts from PDR5 and TPO1 were monitored, by Northern blot assays, following sudden exposure of an unadapted exponentially growing cell population of the parental strain BY4741 to a moderate stress induced by artesunate and compared with those of the unstressed population (Fig. 3). A significant increase in the levels of mRNA from TPO1 (up to 10-fold) and PDR5 (up to 13-fold) was registered during acute artesunate stress after 30 min of incubation with the drug compared with the unstressed cell population. This was the maximum level registered for TPO1 mRNA under the experimental conditions used. In fact, following the initial period of adaptation of the yeast cell population to the drug, with the resumption of exponential growth with a reduced growth rate, TPO1 mRNA levels were rapidly reduced to basal values, close to those exhibited by unstressed cells. However, maximum PDR5 mRNA levels (17-fold the basal levels determined in unstressed cells) were reached only after 2 h of drug exposure, and the high mRNA levels were maintained during a significant part of the exponential phase of artesunate-stressed growth.


Early activation of TPO1 and PDR5 transcription by artesunate. Results from Northern blot hybridization of total RNA using TPO1 or PDR5 as probes, and the actin-encoding gene ACT1 as internal control. Total RNA was extracted from cells of the Saccharomyces cerevisiae BY4741 parental strain, harvested during cultivation in MM4 liquid medium (black symbols) or in this medium supplemented with 0.5 g L−1 artesunate (open symbols), at the indicated time of cultivation. Growth was followed by measuring culture OD600 nm (●,○). Relative values of the mRNA TPO1/mRNA ACT1 (▲, △) and mRNA PDR5/mRNA ACT1 (▪,◻) were obtained by densitometry of the autoradiograms. The relative mRNA value for each gene immediately before exposure to artesunate (0 h) was set as 1.

The expression patterns of PDR5 and TPO1 during yeast cultivation in the presence or absence of artesunate were also compared by measuring the β-galactosidase activity of cells of parental strain BY4741 transformed with plasmids containing TPO1–lacZ or PDR5–lacZ fusions (Fig. 4). Maximal β-galactosidase activity in yeast cells challenged with artesunate was much higher for PDR5 (up to 18-fold, compared to 3-fold in TPO1). Differences observed in the profiles of mRNA levels and β-galactosidase activities of TPO1 and PDR5 in yeast cells during artesunate-stressed cultivation suggest that the stability of the two transcripts may be different and/or that the time-dependent transcription patterns for the two genes following artesunate stress do not coincide. Two other lacZ fusions, corresponding to two other Pdr1p/Pdr3p target genes, SNQ2 and YOR1, were also tested, and artesunate was proved to activate the expression of these genes by approximately 3- and 10-fold, respectively (results not shown). However, no detectable effect in yeast resistance to artesunate could be associated with the expression of SNQ2 or YOR1 (Table 1).


Effect of artesunate on TPO1 and PDR5 expression based on lacZ expression fusions. Growth curves (●,○) (a) and ß-galactosidase activity (b, c) of cells of Saccharomyces cerevisiae BY4741 (wild-type) harbouring a TPO1–lacZ (▲, △) (b) or a PDR5–lacZ (▪,◻) (c) fusion plasmid, grown in MM4-U medium in the absence of drug (black symbols) or in the presence of 0.5 g L−1 artesunate (open symbols). β-Galactosidase activity is expressed in Miller units (U), defined as the increase in A420 nm min−1 (OD600 nm)−1.

Expression activation of TPO1 and PDR5 under artesunate stress depends on Pdr1p and Pdr3p

To identify the transcription activators behind PDR5 and TPO1 transcription activation following yeast cell exposure to artesunate, we searched for the presence of regulatory elements for transcription factors known to be involved in the yeast response to drugs and other chemical stresses, within their promoter regions (Fig. 5). This analysis revealed a number of candidate regulatory elements, including known binding sites for Pdr1p and Pdr3p, Pdr8p, Yrr1p, Msn2p/Msn4p, Yap1p, War1p or Skn7p (Fig. 5). To test these candidate transcription regulators, we compared the β-galactosidase activity from lacZ fusions for TPO1 and PDR5 during growth of parental strain BY4741 under artesunate stress and mutants with each of the transcription factors individually deleted, transformed with TPO1–lacZ or PDR5–lacZ fusion plasmids (results not shown and Fig. 6). Artesunate-induced activation of the expression of both TPO1 and PDR5 was found to be dependent on the product of PDR1, since the elimination of PDR1 severely reduced artesunate-induced activation of PDR5 (by 90%) and significantly affected TPO1 activation (by 60%) (Fig. 6). The elimination of PDR3 caused a more subtle reduction of artesunate-induced activation of TPO1 and had no detectable effect on PDR5 activation, when cells harvested at the same early or mid-exponential phase of artesunate-stressed growth under identical experimental conditions were examined. The disruption of the genes encoding the other above-mentioned candidate transcription factors had no detectable effect on artesunate-induced activation of both genes (results not shown), although Msn2p and Yap1p expression was found to provide some protection against the drug (Table 1).


Putative regulatory elements in TPO1 and PDR5 promoter regions. Coordinates are relative to the translation start site. The database YEASTRACT (http://www.yeastract.com; Teixeira, 2006b) was used in this search.


Effect of Pdr1p and Pdr3p on the expression activation level of TPO1 and PDR5 by artesunate. Comparison of ß-galactosidase activity in cells of Saccharomyces cerevisiae BY4741 (wild-type), and Δpdr1 or Δpdr3 mutants harbouring a TPO1–lacZ (a) or a PDR5–lacZ fusion plasmid (b), during exponential growth in MM4-U medium in the absence (control) or in the presence of 0.5 g L−1 artesunate.

Although severely reducing artesunate-induced activation of TPO1 and PDR5, PDR1 deletion did not lead to the complete elimination of TPO1 and PDR5 activation by artesunate. To test if the remaining activation was due to Pdr3p, TPO1–lacZ and PDR5–lacZ fusion plasmids were inserted into cells of strain FY1679-28C and the levels of activation of TPO1 and PDR5 during growth of parental FY1679-28C under artesunate stress were compared with those in the isogenic derivatives carrying the Δpdr1 mutation or the Δpdr1Δpdr3 double mutation (Fig. 7). To perform this experiment, artesunate concentration was reduced to 0.25 g L−1 because the double mutant Δpdr1Δpdr3 is highly sensitive to artesunate and could not grow at the concentration used in other experiments. Results show that artesunate-induced activation of TPO1 expression was practically abolished in the Δpdr1Δpdr3 double-deletion strain, either in the presence or in the absence of the antimalarial drug, while PDR5 expression was abolished, indicating that the remaining expression activation in the Δpdr1 mutant depends on the presence of Pdr3p.


The activation of PDR5 and TPO1 by artesunate is abolished in the absence of Pdr1p and Pdr3p. Growth curves (a, b) and ß-galactosidase activity of cells of Saccharomyces cerevisiae FY1679_28c (wild-type) (●), Δpdr1 (▪) and Δpdr1Δpdr3 (▲) mutants harbouring a TPO1–lacZ fusion (c, d) or a PDR5–lacZ fusion (e, f), grown in MM6-U medium in the absence of drug (a, c, e) or in the presence of 0.25 g L−1 artesunate (b, d, f).


In this study, we used a yeast model to search for mechanisms of adaptation and determinants of resistance to artesunate. The Zn(II)2Cys6 zinc finger regulator Pdr1p, contributing to PDR in yeast, was implicated in resistance to artesunate, showing that this drug is one of the range of toxic compounds for which PDR1 expression exerts protection in yeast. The susceptibility to artesunate toxicity of yeast deletion mutants for six Pdr1-regulated genes encoding known or predicted MDR transporters, PDR10, PDR15, PDR5, SNQ2, TPO1 and YOR1 (Table 1), indicates that only PDR5 and TPO1 exert a protective role under the experimental conditions used. The genes PDR5 and TPO1 were involved in the early yeast response to acute artesunate-induced stress, since the transcript levels rapidly increased following exposure of an unadapted yeast population to the drug. The time-dependent transcription activation pattern of PDR5 and TPO1 in response to sudden cell stress with artesunate suggests that TPO1 is responsible for part of the early response to the drug, while PDR5 transcription activation is also observed in artesunate-adapted cells, during exponential growth in the presence of the drug, consistent with the notion that this gene has a role in cellular detoxification during exponential phase (Mamnun, 2004). Transcriptional activation of both genes by artesunate is significantly dependent on the presence of Pdr1, while the elimination of the highly homologous transcription factor Pdr3p has a slight effect on TPO1 activation by artesunate.

The expression of the multidrug transporters from the ABC superfamily SNQ2 and YOR1, which are known Pdr1/Pdr3 targets, was also found to be activated by artesunate. However, the level of expression of these genes had no apparent influence on yeast resistance to artesunate, under the standard conditions tested in this study. Similar observations were reported before in response to other chemical stress situations, based on microarray experiments. Indeed, it has been demonstrated that increased transcription of a particular gene in response to chemical stress does not imply that it has a role as a determinant of resistance to this specific stress. Significant examples include the yeast response to stress induced by the acid herbicide 2,4-D (Teixeira, 2006a), citric acid (Lawrence, 2004) or cisplatin (Birrell, 2002). Although intriguing, these results are consistent with the notion that greater cell flexibility in the response to a wide range of stress situations, through a broad transcriptional response via a few signalling mechanisms, may be more cost-efficient than a large number of more specific regulatory components.

The involvement of both TPO1 and PDR5 in the adaptive response and resistance of the yeast cell to the herbicide 2,4-D, another weak acid of high lipophilicity, has been observed previously in our laboratory (Teixeira & Sá-Correia, 2002). Like 2,4-D, artesunate, also called artesunic acid, is a weak acid with a free carboxylic group (pKa=4.6 (Augustijns, 1996)). However, artesunate is much less lipophilic than other artemisinin derivatives, such as artemether and arteether (octanol – water partition coefficient=0.31; Augustijns, 1996). As proposed for other weak acids, the toxic form of artesunate is the liposoluble undissociated form, present in the incubation medium at pH 4.5. Following the entrance of the undissociated form into the cell by passive diffusion, its dissociation in the approximately neutral cytoplasm leads to a decrease in intracellular pH (pHi) and the accumulation of the counterion, which cannot cross plasma membrane lipids but may be actively exported through specific inducible transporters, presumably Pdr5p and Tpo1p. The results of susceptibility and expression assays shown in this work suggest that these ABC and MFS multidrug transporters may have a role as active exporters of the artesunate counterion, as hypothesized before for the acid herbicide 2,4-D (Teixeira & Sá-Correia, 2002). Remarkably, none of the multidrug transporters required for resistance to the less lipophilic short-chain monocarboxylates (C2–C6) or benzoate, in particular Pdr12 (Piper, 1998), Azr1 (Tenreiro, 2000), Aqr1 (Tenreiro, 2002), Tpo2 and Tpo3 (Fernandes, 2005), are involved in artesunate resistance. As hypothesized before for 2,4-D (Teixeira & Sá-Correia, 2002), the mechanism behind artesunate toxicity and tolerance cannot be reduced to its action as a weak acid, independently of the nature of the R group of the acid (R-COOH).

As it is known that many of the mechanisms of MDR are apparently conserved among phylogenetically distinct organisms, we searched for Pdr5p- and Tpo1p-like proteins in the P. falciparum genome (http://www.ncbi.nlm.nih.gov/BLAST/). Compared to the genomes of free-living eukaryotic microorganisms, the P. falciparum genome encodes fewer transporters and its predicted transport capabilities resemble those of obligate intracellular prokaryotic parasites (Paulsen, 2000). Plasmodium falciparum presents a very limited repertoire of membrane transporters, with only six members of the MFS and one member of the amino acid/polyamine/choline (APC) family (Gardner, 2002). Among the predicted MFS transporter proteins, it was not possible to identify any homologue of Tpo1p. The same result was obtained when the genomes of Plasmodium yoelli and Plasmodium vivax (http://plasmodb.org/) were examined. However, a BLAST comparison between the amino acid sequence of Pdr5p and the annotated proteins of the P. falciparum genome revealed the existence of one possible homologous protein, encoded by ORF PF14_0244, annotated as coding for a putative ABC transporter. The analysis of the predicted structural domains of the PF14_0244 protein revealed a single nucleotide-binding domain (NBD) that precedes the six putative transmembrane segments (TMS6) (http://smart.embl-heidelberg.de/), typical of the so-called ABC ‘half-transporters’ (Rogers, 2001; Borges-Walmsley, 2003). The first half of the Pdr5p protein shares 22% identity and 41% similarity with 90% of the entire PF14_0244 protein sequence, while the second half of the Pdr5p protein shares 35% identity and 58% similarity with 37.5% (corresponding to the NBD) of the whole PF14_0244 protein. Although it is a half-transporter, this candidate protein may be a functional homologue of Pdr5p, since, in human cells, the multidrug-resistant phenotype associated with the expression of the ABC half-transporter ABCG2 partially overlaps the phenotype associated with expression of P-gp, despite the existing differences in domain organization (Ejendal & Hrycyna, 2002). Therefore, it would be of interest to examine the eventual involvement of ORF PF14_0244 in the P. falciparum response to artesunate. Indeed, only a small number of P. falciparum genes, predicted to encode members of the ABC superfamily of transporter proteins capable of conferring drug resistance, have been demonstrated to confer resistance to antimalarials, in particular to quinoline ring-containing drugs.


We thank A. Goffeau (FYSA, UCL, Belgium) and C. Jacq (ENS, CNRS, France) for gifts of plasmids and strains. We thank Dafra Pharma, Turnhout, Belgium for providing artesunate and H. Jansen (Dafra Pharma) for helpful information on the product. This research was supported by FEDER, ‘Fundação para a Ciência e Tecnologia’ (FCT) and the POCTI Programme (contracts POCTI/BIO/38115/2001 and POCTI/BME/46526/2002, and grants to M. Alenquer (SFRH/BD/17260/04) and S. Tenreiro (BPD/5649/01)).


  • Editor: André Goffeau


View Abstract