The methodology adopted here allowed us to find reproducible alterations of the transcriptome in P. falciparum
parasites under artesunate pressure. To overcome the major hurdle of developmentally regulated genes, we combined the use of a fast acting drug at high dose with limited incubation time at 5 successive but synchronized developmental stages. Most observed alterations were of low amplitude, possibly reflecting particular mechanisms of gene regulation in malaria parasites. This resulted in a relatively broad pertubation, that nevertheless was highly significant and consistent across experiments. We did not detect altered expression of a single specific metabolic pathway or of the putative drug target molecules but several functionally important groups of genes exhibited interesting dynamic changes. Within the intrinsic limitations of all transcriptome studies (in which steady-state RNA levels are studied and not protein quantities, post-translational modifications or biological activities), the overall picture that emerges from our current analysis of the transcriptome changes suggests i) pertubation of the intracellular trafficking and organisation, including changes in chaperones, transporters, remodeling the erythrocyte space beyond the parasitophorous vacuolar membrane up to the membrane of the infected erythrocyte; ii) probable down-phasing of metabolism as deduced from down-regulation of key enzymes/transporters implicated in purine, pyrimidine and isoprenoid synthesis; iii) an altered mitochondrion; iv) an altered redox homeostasis and possibly altered protein turnover. In addition, the transcriptome analysis highlighted numerous genes with unknown function, annotated as coding for hypothetical proteins [see Additional file 2
]. Some display consistent high and early up-regulation and clearly deserve further investigation.
The physiological significance of the observed perturbations need to be assessed in future studies. Transcriptome analysis is the first of a multi-step process, to be followed by scrutiny of the role of the individual genes identified here. In particular, it will be essential to establish whether or not the genes in question are implicated in a rate-limiting process and to analyse the dynamics of the corresponding protein products, including their turnover, post-translational modifications and cellular localization. An important question is to clarifiy which transcriptome alterations are specific to the response to artemisinins and which are related to a parasite lethally injured, as this has numerous implications for drug development.
We will discuss the transcriptome data in the frame of existing knowledge on mechanisms of action of artemisinin derivatives and mechanisms resulting in reduced susceptibility/resistance. We will next discuss possible common changes in steady state RNA levels after different stresses inflicted to the parasite such as exposure to 41°C or treatment with chloroquine, that can cause parasite death.
The mechanism of action of artemisinins is debated. Artemisinin are endoperoxide-containing sesquiterpene lactones. Fe++
-dependent activation of the endoperoxide bridge is required for the drug to be active [49
]. Cleavage of the endoperoxide moiety forms highly reactive oxyl radicals that rearrange to more stable carbon-centered radical intermediates. These in turn form covalent adducts with parasite products thought to be responsible for a pleiotropic effect of the drug. Such pleiotropicity may explain the variety of genes we show here to be affected in their expression. Recent evidence suggest more specific mechanisms, such as inhibition of specific targets, with specific inhibition of the P. falciparum
SERCA-type Ca++ pump (PfATPase6) [50
]. Importantly, field isolates with markedly reduced in vitro
susceptibility to artemether presented a mutant SERCA-type PfATPase6 [1
]. Artemisinin has been shown to inhibit the endoplasmic reticulum – located SERCA-type in Toxoplasma gondii
, a related Apicomplexan parasite as well [52
]. We did not observe altered steady state levels of PfATPase6
mRNA at any time point investigated, indicating that over-expression of this gene is not part of the parasite response to artesunate.
Ultrastructural alterations of the morphology and their kinetics are debated as well. Early-stage alterations of the mitochondrion, the endothelial reticulum and the digestive vacuole were detected in some studies [53
] but not others [54
]. Some of the transcriptome alterations may account for a disorganisation of the digestive vacuole and the mitochondrion. We observed increased levels of Pfmdr1
expression concomitant with decreased mRNA levels of four of the subunits of the H+
vacuolar ATPase, which is implicated in regulating calcium intracellular stores of acidic compartments. Downproduction of the H+
vacuolar ATPase and/or alteration of its subunit ratio is predicted to negatively impact on intracellular calcium homeostasis. This is most probably on the critical path, since disruption of calcium homeostasis is central to cellular death, necrosis endophagy and apoptosis. Our transcriptome analysis also shows evidence that artesunate interferes with expression of genes related to the mitochondrion, in most cases by down-regulating their steady state RNA levels. This calls for additional studies on the mitochondrial activity after artesunate exposure.
Additional factors are to be considered in light of the transcriptome data, in particular the intracellular partitioning of the drug. Several studies have demonstrated the selective uptake of artemisinin derivatives. It has been shown that artemisinin and dihydroartemisinin, two sesquiterpene endoperoxide drugs closely related to artesunate, are transported by the tubulovesicular network (TVN) formed within the erythrocyte cytoplasm for nutrient import to the parasite, and that these drugs disrupted the protein organization of the TVN [55
]. Numerous genes implicated in intracellular space modelling, trafficking, chaperones and transport display altered transcription profiles in artesunate-treated parasites. The markedly bias for such genes was indeed the most salient observation of our transcriptome study. It is tempting to speculate that this results in remodelling the intracellular space, including remodelling of the infected red blood cell membrane and TVN. This could either be a response aimed at correcting drug-inflicted damage, preventing further intake of drug or reflect the dysfunction resulting from such damage. Protein trafficking is all the more important for parasite growth and survival that the parasite develops inside a parasitophorous vacuole within the erythrocyte cytoplasm. Interactions between the parasite and the erythrocyte are multiple, with an increasing number of parasite proteins recognized as being involved (reviewed in [56
]). Parasite-induced modifications of the red blood cell through interactions of parasite proteins exported to the host cell membrane or cytoskeleton play a major role in parasite survival and virulence, through induction of infected red cell cytoadhesive properties or decreased deformability. Altered transcript levels of genes involved in host cell remodelling and intracellular trafficking have been reported in parasites lethally damaged by exposure to febrile temperatures [12
]. This suggests that perturbing intracellular trafficking/remodeling is either on the path to parasite death or an attempt to overcome a lethal injury. This also suggests that alterations of intracellular trafficking by artesunate may be one of the mechanisms through which the drug can be active on a wide range of parasite developmental stages.
Analysis of the proteome of artemether-treated P. falciparum
parasites showed that 30 of the 101 proteins identified using a MALDI-TOF-MS-based analysis [57
], displayed a greater than three-fold expression level compared to untreated control cultures. Two of the over-expressed proteins, namely hsp90 (PF07_0029), PF14_0425 (fructose-biphosphate aldolase) also presented an increased transcript level in our study. Direct comparison of both studies is precluded due to limitations such as different sensitivities of the transcriptome and proteome approaches, more comprehensive coverage of the expressed products in the transcriptome approach, different drug exposure protocols and/or transcription/translation uncoupling, which has been described for approx 50% of the proteins [43
To put the alterations observed under artesunate in perspective with changes induced by unrelated forms of stress, we analyzed the data in the light of the transcriptome alterations reported after chloroquine [8
] treatment or incubation at 41°C [12
], both treatments reported as having led to parasite death. The transcriptome studies are not directly comparable, as they were performed under different experimental conditions (synchronous/asynchronous parasites, glass slide microarrays/Affymetrix chips) and corresponded to different natures of stress (rapid acting vs.slow acting). However, identification of a few genes with altered expression under such different experimental conditions provides interesting insight. Such joint analysis is illustrated in the Venn diagram (Figure see also tables , , , , , , , , , 10 for different functional groups). There were only 4 genes identified as differentially expressed under all 3 types of stress: PFB0095c encoding PfEMP3, PFE1245w, encoding a putative zinc finger protein, and PF14_0151 encoding a putative RNA-binding protein and MAL7P1.171 encoding a hypothetical protein. PfEMP3 and MAL7P1.171 have an export motif. PfEMP3 is exported to the cytoplasmic face of the erythrocyte membrane in the mature stages. It associates with the erythrocyte cytoskeleton, which it destabilizes [58
]. Both PfEMP3 and MAL7P1.171 have been shown to contribute to trafficking of PfEMP1 [59
]. Their role in the response to lethal stresses deserves to be investigated. Characterization of the products encoded by the remaining 2 common genes certainly warrants particular attention, as these genes may play a crucial role in the general response to stress.
Figure 3 Venn diagram of genes differentially expressed under artesunate treatment chloroquine treatment, and heat-shock. Purple circle: 398 genes differentially expressed under artesunate (ART). Orange circle: 30 genes differentially expressed under chloroquine (more ...)
The transcriptome of chloroquine-treated and artesunate-treated parasites shared three additional altered gene profiles, namely PFE1455w encoding a putative sugar transporter, PFI0095c and PFL0040c both encoding putative kinases belonging to the R45-FIKK family. Interestingly again, the three predicted proteins have an export motif. However, since there were large inter-experiment variations in the transcriptome analysis of chloroquine-treated parasites, commonalities are probably largely underestimated.
Ninety four genes were differentially regulated both under heat shock and artesunate, 63 over-expressed and 31 under-expressed under both conditions, amongst which members of the R45-FIKK kinase family, chaperones, molecules involved in lipid metabolism and number of hypothetical proteins -including as discussed above numerous genes coding for proteins with an export motif. Two of the 5 R45-FIKK kinase genes over-expressed under artesunate were also over-expressed under chloroquine pressure [8
]. Of the 3 R45/FIKK kinase genes found over-expressed under heat shock, one was also differentially expressed under chloroquine. These data suggest that expression of genes belonging to the R45-FIKK kinase family reflects the response of the parasite to different types of environmental modifications and that the parasite response to different natures of stress may include common pathways. The same could be said about genes belonging to 2 other families. Eight of the 13 chaperone/chaperone-related encoding genes over-expressed under artesunate were also over-expressed under heat-shock. Of the 30 Hsp40 type P. falciparum
chaperones, 6 were over-expressed under artesunate, 4 in common with heat-shock.