ABOLICIÓN DE LA ESCLAVITUD

Protein targeting in Plasmodium parasites is a very complex process, and multiple desti- nations to target proteins exist. The following sections describe briefly what is known so far about protein targeting to various destinations in Plasmodium parasites.

1.6.1. Protein targeting to the host cell

During the intraerythrocytic development cycle, the parasite remodels the erythrocyte, a process which requires protein trafficking. It was shown that proteins targeted to the host cell contain the Plasmodium export element/vacuolar targeting signal (PEXEL/VTS) for translocation across the parasitophorous vacuole membrane (Hiller et al., 2004, Marti et al., 2004). However, how the proteins get across this membrane is not known. Two routes for transport to the erythrocyte membrane have been described. Proteins targeted to the host cell like KAHRP or PfEMP-3 contain a typical signal peptide at their N- terminus that allows co-translational import into the secretory pathway (Wickham et al., 2001, Lopez-Estraño et al., 2003). The proteins are transported to the parasitophorous vacuole and the PEXEL/VTS signal allows transport across the parasitophorous vacuole membrane. In the erythrocyte cytosol, the proteins interact with the Maurer’s clefts, but how they are transported to the cytoplasmatic side of the erythrocyte membrane is unclear. Transport of PfEMP-1 to the erythrocyte membrane requires a different mechanism, since PfEMP-1 does not contain a classical signal peptide and thus its entry into the secretory system is not clear. Data suggest that PfEMP-1 enters the ER via its putative transmem-

brane domain and is then transported to the parasitophorous vacuole membrane (Knuepfer et al., 2005). It has been suggested that PfEMP-1 is packaged into the Maurer’s clefts while they are forming by budding off the parasitophorous vacuolar membrane (Spycher et al., 2006), and recently it was shown that P. falciparum skeleton binding protein 1 (PfSBP-1) is involved in this process (Maier et al., 2007). It seems that PfSBP-1 is also required for PfEMP-1 trafficking to the erythrocyte surface (Maier et al., 2007). Further- more, exported proteins were identified without an obvious PEXEL/VTS motif (Blisnick et al., 2000, Spycher et al., 2003, Spielmann et al., 2006), suggesting yet another potential trafficking route for these proteins. Potentially, these proteins are also exported while the Maurer’s cleft are formed, but further analyses are required to support this hypothesis.

1.6.2. Protein targeting to the food vacuole

Host cell haemoglobin is endocytosed by the cytostome and is subsequently transported to the food vacuole. Here, haemoglobin is digested by several proteases, which are targeted to the food vacuole via the secretory system, requiring ER and Golgi transport (Tonkin et al., 2006a). However, a food vacuole targeting peptide has not yet been described, although recently it was shown that the first 120 amino acids of the cysteine protease falcipain-2 are sufficient to target GFP to the food vacuole (Dasaradhi et al., 2007). These data suggest that trafficking to the food vacuole might be mediated by a potential N- terminal targeting peptide, but further analyses are required to support this finding. After entry into the secretory system, two pathways for trafficking to the food vacuole were described. The first was observed for the aspartic protease plasmepsin II, which is in- volved in the first steps of haemoglobin degradation (Klemba et al., 2004a). The protein contains a transmembrane domain and is transported through the ER as a membrane pro- tein. It is trafficked to the cytostome and from there it is delivered to the food vacuole, possibly through vesicular transport. In the food vacuole, the transmembrane domain is cleaved releasing plasmepsin II into the food vacuole (Klemba et al., 2004a). A similar trafficking mechanism was proposed for falcipain-2, which also contains a transmem- brane domain (Dasaradhi et al., 2007). Interestingly, no classical signal peptide is found at the N-terminus of this protein. But as already suggested for PfEMP-1 import into the secretory system, the hydrophobic transmembrane domain might facilitate the import into the ER. Another route of trafficking was described for dipeptide aminopeptidase I, which contains no transmembrane domain (Klemba et al., 2004b). It was shown that the protease enters the endomembrane system and targets to the food vacuole via the parasitophorous vacuole, where it is taken up again during endocytosis and transported to the food vacuole.

1.6.3. Protein targeting to the apical organelles

Protein trafficking to the apical organelles is not very well understood in P. falciparum. It is known that targeting to the rhoptries in P. falciparum and T. gondii is via the ER and Golgi (Howard and Schmidt, 1995), thus the proteins possess a classical N-terminal signal peptide. Within the secretory system a second targeting step is required and for T. gondii it has been shown that this is accomplished by either a tyrosin-based motif at the C-terminus of the proteins, or by the interaction with other proteins. Indeed, targeting of membrane proteins to the micronemes also requires a tyrosin-based and acidic amino acid-based mo- tif in T. gondii (Cristina et al., 2000, Hoppe et al., 2000). These features are also found in some Plasmodium proteins, however, for the P. falciparum rhoptry-associated protein 1 (RAP-1) and for the micronemal erythrocyte binding antigen 175 (EBA-175), it was shown that the C-terminal tyrosin-based motifs are not required for correct targeting to the organelles (Baldi et al., 2000, Gilberger et al., 2003). It was suggested that the proteins potentially are transported in concert with other proteins, which escort them to the correct organelle. Recently, Treeck and colleagues demonstrated that a luminal cysteine rich re- gion at the C-terminus of EBA-175 together with correct timing of expression is crucial for micronemal targeting of this protein (Treeck et al., 2006). Although this cysteine rich region is highly conserved in the EBL-superfamily, it is not found in other micronemal proteins like apical membrane antigen 1 (AMA-1) or subtilisin like protease (SUB2), sug- gesting that multiple mechanisms facilitate micronemal protein targeting (Treeck et al., 2006). Not very much is known about protein targeting to the dense granules. The only requirement shown so far for correct targeting in P. falciparum was correct timing of ex- pression. A GFP fusion protein of the ring infected erythrocyte antigen (RESA) was only targeted to the dense granules when under control of its endogenous promoter (Rug et al., 2004).

1.6.4. Protein targeting to the apicoplast

Since most of the proteins originally encoded by the plastid genome were transferred into the nuclear genome, a machinery to import apicoplast proteins encoded by the nucleus had to be developed. In plants, which contain a primary plastid, proteins designated for the plastid contain a N-terminal transit peptide for translocation into the plastid (Vothknecht and Soll, 2000, Jackson-Constan and Keegstra, 2001). The transit peptide is rich in basic amino acids and thus contains a positive net-charge, and additionally contains a chaperone binding site. The transit peptide interacts with protein complexes within the plastid mem- branes, which facilitate the post-translational protein import. These protein complexes are

known as Toc and Tic (Translocon at the outer/inner membrane of chloroplasts) (Soll and Schleiff, 2004). During protein import, the transit peptide is cleaved once it has passed the inner membrane, catalysed by a stromal processing peptidase producing the mature protein (Richter and Lamppa, 1998). The situation in apicomplexan parasites is somewhat more complicated than in plants, because the apicoplast is a secondary plastid surrounded by four membranes rather than two. Thus, proteins targeted to the apicoplast possess, in addition to the transit peptide, a signal peptide that allows import into the secretory path- way (van Dooren et al., 2001). Nuclear encoded proteins that are directed to the apicoplast possess a bipartite leader sequence that firstly imports the protein co-translationally into the secretory system and secondly directs it to the apicoplast (Waller et al., 2000). It was also suggested that apicoplast targeting through the secretory pathway is Golgi indepen- dent and diverts straight from the ER to the apicoplast (DeRocher et al., 2005, Tonkin et al., 2006b). Toc and Tic homologues were identified in the P. falciparum genome, and it is speculated that they span the four apicoplast membranes to promote protein import into the apicoplast (Mullin et al., 2006).

The likelihood of an apicoplast localisation of a protein can be estimated on the basis of the bipartite leader sequence using several prediction programs including SignalP and the

Plasmodium specific PATS (prediction of apicoplast targeted sequences) (Nielsen et al.,

1997, Zuegge et al., 2001).

1.6.5. Protein targeting to the mitochondrion

Most of the mitochondrial proteins are encoded in the nucleus and thus need to be im- ported into the mitochondrion. Mitochondrial protein import occurs post-translationally, and targeting is facilitated via different pathways. The most widespread method for mi- tochondrial protein import is via a N-terminal transit peptide, but internal signals and transmembrane domains are also able to facilitate mitochondrial protein import (Pfanner and Geissler, 2001). Mitochondrial protein import in P. falciparum has so far been shown to be accomplished by N-terminal transit peptides, which if fused to green fluorescent protein (GFP) direct GFP into the mitochondrion (Sato et al., 2003). Import of the pro- tein across the two mitochondrial membranes is achieved by the protein complexes Tom and Tim (Translocase of the outer/inner membrane) (Pfanner and Geissler, 2001). Ho- mologues of the Tom and Tim complexes were identified in the P. falciparum genome, suggesting that mitochondrial protein import in Plasmodium is similar to mitochondrial protein import in other organisms (van Dooren et al., 2006).

nature of the transit peptide using programs like MitoProt, Predotar or the P. falciparum specific PlasMit (Claros and Vincens, 1996, Bender et al., 2003, Small et al., 2004).

1.6.6. Dual protein targeting

In plants it has been shown that some proteins have multiple destinations and that these proteins are dually targeted to different organelles (Silva-Filho, 2003). Different mecha- nisms can be involved in the regulation of this process. Alternative splicing is one possi- bility or the duplication of the gene with the acquisition of different 5’ ends can determine the location of the protein. Alternative translation initiation or translation initiation from non-AUG start codons are other ways of regulation. Post-translational modifications as well as ambiguous targeting sequences can also play important roles in determining the final localisation of a protein (Silva-Filho, 2003).

Multiple targeting has also been observed in P. falciparum. The zinc metalloprotease fal- cilysin was shown to be targeted to the food vacuole, to the apicoplast and also likely to the mitochondrion (Ponpuak et al., 2007). In the food vacuole, falcilysin participates in the degradation of haemoglobin, whereas in the apicoplast and mitochondrion it was sug- gested that it may be involved in transit peptide degradation. Different mechanisms that might be involved in the multiple targeting of falcilysin were discussed by Ralph (2007) and include potential post-translational modifications, alternative splicing and/or alterna- tive translation initiation. Recently, dual targeting was also demonstrated in T. gondii. Pino and colleagues showed dual targeting of proteins to the apicoplast and mitochon- drion (Pino et al., 2007). The functions of the dually targeted proteins in T. gondii are required by both organelles and the authors suggest that the main mechanism involved is bimodal targeting. Generally, with the bimodal mechanism, dual localisation of a protein is achieved by the different "organelle import machineries" recognising different parts of the protein. In apicomplexan parasites one crucial step for dual targeting to the apicoplast and the mitochondrion is the import into the secretory pathway. Bimodal targeting sug- gests that the signal peptide of these dual targeted proteins is weak and is not recognised by the signal recognition particle (SRP) at all times, resulting in some proteins trafficking to the apicoplast via the secretory system and some proteins being post-translationally imported into the mitochondrion. However, Pino and colleagues have also shown that different translation initiation and differential splicing can affect the protein localisation, and are likely to play a role in dual protein targeting (Pino et al., 2007).

Dual protein targeting in Plasmodium and Toxoplasma might allow the parasites to adapt quickly to changes in their environment, but the precise mechanisms controlling dual tar-

geting remain elusive and require further analyses.