Possessing a system that experimentally controls gene expression has been a Holy Grail in molecular malaria research. Several strategies to control gene expression at different levels have been developed; the controlled step can range from transcription initiation to post-translational modification and/or protein degradation. Strategies successfully developed in model organisms and adapted to the malaria parasite can be classified into four categories aimed at the conditional control of (i) gene deletion, (ii) gene transcription, (iii) mRNA translation, and (iv) protein stability. Here, I intend to describe the various strategies available and compare and contrast their advantages and limitations. In the absence of a unique, ubiquitous solution, it is instrumental to utilize a variety of approaches that can respond to the particular needs of each gene.
Malaria causes more than one million deaths annually, primarily in children. The rapid emergence of parasites resistant to a limited repertoire of drugs, and the absence of an efficient vaccine are responsible for this inextricable situation. The development of novel intervention strategies against malaria is imperative and critically depends upon a robust technology to study gene function in Plasmodium to validate a novel arsenal of potential drug and vaccine targets. Frequently, the validation of a gene product as a drug target requires demonstration of its essentiality for parasite survival. The most direct way to demonstrate essentiality is to specifically inactivate the gene of interest and examine the resulting phenotype. Plasmodium is amenable to genetic manipulation by electroporation during the erythrocytic cycle, when the parasite genome is haploid. Therefore, conditional strategies are necessary to study genes essential for the intra-erythrocytic stages.
Recently, the ‘genetic toolbox’ aimed at ectopically controlling malaria gene products has been considerably expanded (Limenitakis and Soldati-Favre, 2011) and has culminated in the adaptation of Plasmodium falciparum to conditional gene excision based on the DiCre-Lox system (Collins et al., 2013).
This important step forward offers an opportunity to recapitulate the various strategies available and compare and contrast their advantages and limitations. In the absence of a unique, ubiquitous solution, it is instrumental to utilize a variety of approaches that can respond to the particular needs of each gene.
Gene function can be controlled at different levels, ranging from transcription initiation to post-translational modification and/or protein degradation. Strategies successfully developed in model organisms and adapted to the malaria parasite can be classified into four categories aimed at the conditional control of (i) gene deletion, (ii) gene transcription, (iii) mRNA translation and (iv) protein stability.
Temporal control of gene deletion
Gene excision by site-specific recombinases is the most radical and irreversible way to disrupt gene function; however, there are challenges associated with this method including the need to preserve maximal recombination efficiency while controlling the recombinase activity in a timely and tightly regulated fashion. Such a gene deletion approach based on site-specific recombination has been applied to Plasmodium berghei and P. falciparum using different recombinases. In P. berghei, the first system described to conditionally disrupt an essential gene involved a flippase (FLP) recombinase that recognizes a pair of FLP recombinase target (FRT) sequences that flank a genomic region of interest (Carvalho et al., 2004; Combe et al., 2009). The stage-specific control of FLP expression and the use of a temperature sensitive recombinase allowed the successful removal of genes essential in mosquito stages and sporozoites. This excision is by definition irreversible and therefore is of limited relevance to the study of genes essential for erythrocytic stages but well suited to the study of mosquitoes and hepatic stages (Combe et al., 2009). In P. falciparum erythrocytic stages, FLP recombinase has been used to remove selection cassettes from the genome of transgenic parasites but not to generate conditional excision of essential genes (van Schaijk et al., 2010; O'Neill et al., 2011).
Cre-lox technology has been used since the 1980s in a variety of organisms and cell types as the most efficient recombinase, which recognizes pairs of loxP sites. In Toxoplasma gondii, the use of a stage specific promoter or even the absence of promoter were not controlled tightly enough to prevent the few molecules of Cre from generating efficient excision (Brecht et al., 1999). This drawback was elegantly solved by the development of a regulatable fragment complementation system for Cre termed DiCre (Jullien et al., 2003). This system is based on splitting Cre into two inactive moieties and fusing them to either FKBP12 (FK506-binding protein) or FRB (binding domain of the FKBP12-rapamycin associated protein). These domains can subsequently be efficiently hetero-dimerized by rapamycin, leading to the reinstatement of Cre activity. This system has been successfully applied for the deletion of essential genes for survival of T. gondii (Andenmatten et al., 2013). In their recent study, Collins et al. designed a vector to recombine in the region coding for the C-terminus of the gene of interest (Pfsera5) by single homologous recombination and in a second step excised the 3′ UTR of Pfsera5 in erythrocytic stages (Fig. 1).
SERA5 is a member of the SERA gene family and previous work has shown it to be refractory to disruption using conventional targeted homologous recombination, suggesting that SERA5 protein is indispensable in asexual blood stages of the parasite life cycle. SERA5 is thought to play a role in schizont rupture (egress) and/or erythrocyte invasion by released merozoites (Blackman, 2008). The removal of the 3′ UTR failed to reduce the level of SERA5 protein level due to the existence of alternative cryptic polyadenylation sites within the modified locus. Nonetheless, this work is important since (i) it presents the first application of the DiCre system in P. falciparum, (ii) it shows a very high efficiency of excision (close to 100%), (iii) it illustrates the considerable potential of this technology to remove ‘essential’ genes from the genome and examine the consequence for the erythrocytic stages, and (iv) it led to the generation of a valuable strain that possesses a tightly rapamycin-regulated DiCre expression cassette integrated into the genome (Collins et al., 2013). This approach is clearly attractive in a parasite where double homologous recombination is very difficult to obtain such as P. falciparum.
Additional improvements are required to render this technology more efficient and versatile. These include the addition of a fluorescent marker expressed upon excision of the gene, which would facilitate the purification and subsequent analysis of the mutant parasites (Andenmatten et al., 2013). Instead of excising the 3′ UTR, which might have limited impact on the downregulation of gene expression, it is conceivable to place the loxP site in an intron or upstream of the open reading frame to delete part of or the whole coding sequence of the gene respectively (Fig. 1). This approach is the most unambiguous way to establish gene essentiality but it reaches its limits when the gene of interest is truly non-dispensable. In such a case, parasite mutants cannot be cloned and it becomes crucial to obtain a high percentage of excision in order to perform in depth, phenotypical analysis on the pool of parasites. In this regard it is very encouraging to see the efficiency of recombination reported by Collins et al.
Further work is required to validate this approach with the excision of a gene critical for parasite survival.
Although not a conditional system, engineered zinc-finger nucleases (ZFN) may constitute a strategy of choice for in depth analysis of important but not essential genes, in spite of their cost. ZFNs bind specific DNA sequences and generate double strand breaks allowing efficient gene disruption, replacement and site-specific editing (Straimer et al., 2012).
Control of gene transcription
When studying proteins expressed in exo-erythrocytic stages, a simple promoter swap can be sufficient to generate stage-specific gene knock-downs (Siden-Kiamos et al., 2011). This strategy has been successfully applied to PbMyoA, by exchanging the endogenous promoter with the PbAMA1 promoter that is inactive in ookinetes. This strategy is straightforward but limited to the study of genes expressed in mosquito and hepatic stages.
The tetracycline-controlled transcriptional activation, or Tet system, is another widely used technology to control gene transcription. It relies on a transactivator composed of the tetracycline repressor (TetRep) fused to an activating domain. The transctivator binds via the TetRep to tet operator sequences (TetO) placed in front of a minimal promoter and activates transcription. In the presence of tetracycline or its derivative anhydrotetracycline (ATc), the affinity of the TetRep for the TetO is dramatically reduced and transcription is turned off. This system has been widely exploited to tightly control gene expression in eukaryotes (Bujard, 1999) and a modified version of the transactivator was developed for the Apicomplexa (Meissner et al., 2002; 2005) (Fig. 1). Recently, additional trans-activating domains (TRAD) derived from ApiAP2 transcription factors allowed the establishment of a robust Tet-repressible transactivator system, well suited for the study of genes essential for the erythrocytic stages of P. berghei development (Pino et al., 2012). This system has the potential to be applied to the study of vector and liver stages, however the accessibility of ATc to the parasite in the mosquito and the possible leakiness of the system are among the limiting factors. The TRAD-based transactivators are functional in P. falciparum asexual blood stages, however further work is required to investigate if any of the TRADs are robust enough when integrated as a single copy into the genome of the human parasite.
Control of mRNA translation
The application of RNAi-based regulation in the Plasmodium sp. is not broadly applicable, due to the absence of a complete RNAi machinery (Kolev et al., 2011; Barnes et al., 2012).
An alternative strategy is the use of long double-stranded RNA (dsRNA) to interfere with messenger expression (Gissot et al., 2005) (Fig. 1). Overall, the use of dsRNA to modulate gene expression in Plasmodium spp. needs further validation.
Recently, a strategy involving the targeting of ribonuclease P (RNase P) activity to a specific mRNA has been adapted to P. falciparum (Augagneur et al., 2012). This principle capitalizes on the capacity of selectively designed peptide-morpholino oligomer conjugates to bind and promote the cleavage of specific mRNAs as a way to inhibit gene function (Fig. 1). This approach has strong potential despite possible shortcomings such as the issue of specificity and the need for an efficient delivery system for the peptide-morpholino oligomer into the parasite.
Approaches using autocatalytic RNA (riboswitches) have been reported for different organisms. The use of the cis-acting hammerhead ribozyme N9 in the vicinity of the translation start allowed the downregulation of reporter genes in T. gondii and P. falciparum (Agop-Nersesian et al., 2008).
The use of protein-binding RNA aptamers has recently been reported to be functional in P. falciparum (J. C. Niles, MAM 2012, Lorne). This strategy relies on genetically encoded TetR-binding RNA elements (aptamers) in the 5′ UTR of the mRNA making translation of the downstream coding sequence under the control of TetR and tetracycline analogues (Goldfless et al., 2012).
Control of protein stability
Ultimately, controlling gene expression at the level of protein stability can be viewed as the fastest responding system. Control at the level of protein stability can be achieved via the fusion of the protein of interest with the ligand-regulatable FKBP protein destabilization domain (ddFKBP), which is unstructured and consequently targeted for degradation in the proteasome (Banaszynski et al., 2006). Stabilization of the protein occurs in the presence of a rapamycin-derived ligand called shield (Shld-1), which specifically interacts with the ddFKBP, stabilizes it and prevents its degradation (Fig. 1). This methodology has been successfully transferred to T. gondii and P. falciparum with several reports of effective applications (Armstrong and Goldberg, 2007; Herm-Gotz et al., 2007). In contrast, the destabilization appears to be rather ineffective in the rodent malaria models and the use of Shld-1 in vivo is expensive. Ideally, this system would be best suited to C-terminal tagging involving only single homologous recombination despite the risk of reversion (Fig. 1). This approach appears highly suitable to regulate trans-dominant versions of a protein of interest or generating conditional overexpression mutants. However, as the destabilization domain is inserted by knock-in into the gene of interest, the selection process must be done in the presence of Shld-1. Such a long exposure to Shld-1 often results in transgenic parasites unresponsive to Shld-1 removal (P. Pino et al., unpubl. data). In addition, this strategy is not suitable for proteins where tagging affects function, and its potential when targeting secreted proteins remains unclear.
Recently, several mutants of ddFKBP have been assessed for the best regulation at the C-terminus of proteins including one that is exported (de Azevedo et al., 2012).
An alternative, cheaper approach based on the same principle using the DHFR degradation domain (DDD) Escherichia coli dihydrofolate reductase (DHFR) enzyme was reported (Muralidharan et al., 2011). A DDD combined with GFP and HA tags can be stabilized by inexpensive folate analogues such as trimethoprim (TMP) and allows researchers to probe aspects of protein biology such as localization and interacting partners.
The panoply of technologies available to target essential genes for malaria parasite survival has significantly increased in the last few years, but less than ten genes have been conditionally disrupted, highlighting the extreme difficulty of the challenges offered by the malaria parasite. All the technologies listed above have advantages and drawbacks and so far none has imposed itself as a ‘gold standard’. The recent addition of the DiCre system to this repertoire constitutes a significant step forward, as irreversible gene deletion is the most decisive way to disrupt gene function. We can anticipate a significant boost in functional investigations leading to the validation of potential vaccine and drug targets to ultimately reduce the burden of malaria.
We are grateful to Dominique Soldati-favre, Hayley Elise Bullen and Damien Jaquot for critical reading of the manuscript. P.P. is supported by an EVIMalaR fellowship.