Identification of the PEXEL/VTS (19
advanced our understanding of P. falciparum
-host interactions as it allowed the prediction of more than 300 proteins that constitute the ‘exportome’ (22)
. While some studies have characterised the function of exported proteins that contain a PEXEL (14
, little is known about the machinery involved in their trafficking pathway or how each PEXEL residue contributes to export. The PEXEL was recently identified as a recognition sequence for protease processing with subsequent N acetylation of the cleaved protein; however, the function of the conserved amino acids in this motif has not been addressed (24)
nor has the function of N acetylation in export been characterised. In this work, we have determined the role of each conserved amino acid within the PEXEL with respect to their requirement for protease processing, N acetylation and export to the infected erythrocyte.
Using subcellular fractionation, we have shown that N-terminal processing of chimaeras occurs at the conserved leucine in a PEXEL-dependent manner, within the parasite, rather than the parasitophorous vacuole or during translocation to the erythrocyte cytosol. Identification of PEXEL cleavage at the conserved leucine is in agreement with a recent study (24)
. Their work also showed that brefeldin A (BFA), which inhibits PfSec7 in ER-to-Golgi transport (28)
, does not disrupt PEXEL processing and concluded that this event occurs early in the trafficking pathway within the ER. However, it is possible that treatment with BFA prevented the anterograde transport of the PEXEL protease in those experiments. Using different methodology, we come to a similar conclusion that suggests PEXEL processing occurs within the ER. Therefore, the fate of proteins destined for export is decided early, as previously hypothesised (29)
and involves proteolytic cleavage in a manner dependent on the PEXEL arginine and leucine residues.
It was also shown that N acetylation occurs following PEXEL processing, and it was suggested that this could provide a mechanism to distinguish proteins that are destined to be exported (24)
. A model was proposed where the cleaved PEXEL (xE/Q/D) was required for N acetylation (24)
. Our analysis shows that after cleavage, either at the PEXEL motif or of the signal peptide, N acetylation occurs suggesting that it is common in P. falciparum
. Analysis of the GBP130E>A
chimaera showed that the PEXEL was still cleaved after leucine and N acetylated; however, it was not exported but was trafficked by the default pathway to the parasitophorous vacuole, as shown previously for other GFP reporters that lack a PEXEL (30)
. This demonstrates that the PEXEL glutamic acid is not required for cleavage at the P1 leucine and that it is dispensable for N acetylation. Furthermore, accumulation of the GBP130E>A
chimaera in the parasitophorous vacuole indicates that although PEXEL processing and N acetylation of exported proteins occur both events are insufficient for normal export. This shows that xE/Q/D remaining after PEXEL cleavage plays an important role in export. We also identified an endogenous P. falciparum
protein that was cleaved in the PEXEL at leucine but was not N acetylated. This protein was recovered from the saponin pellet, and thus, we cannot confirm that it was exported; however, this does represent the first example of a protein containing a native PEXEL that was cleaved at the P1 leucine but not N acetylated.
The consequence of PEXEL mutations in reporter proteins has been determined by microscopy to show that the conserved residues are required for efficient export to the infected erythrocyte (19
. We have extended this analysis by determining the effect of mutations on cleavage of the PEXEL as well as export. Mutation of leucine within the PEXEL interferes with processing and consequently blocks export to the erythrocyte. Moreover, chimaeras containing a mutation of arginine in the PEXEL had a similar processing profile to the leucine mutants; thus, processing at the P1 leucine is also dependent on arginine at position P3. This is consistent with the strong conservation of these two residues in all the exported proteins that have been defined (19
. It is noteworthy that we also identified two endogenous proteins that were processed at the PEXEL leucine that contained lysine in place of arginine at position P3. This indicates that lysine is suitable for processing of those proteins and validates the presence of KxLxE/Q/D proteins in the predicted exportome (22)
. It is interesting that the leucine to alanine mutation in both GBP130 and KAHRP still allowed some cleavage, albeit inefficient, and as a consequence, some export was observed. This is perhaps not surprising given the similarity in structure of alanine compared with leucine, as both are nonpolar, aliphatic, hydrophobic molecules often buried in folded proteins. Significantly, when all three conserved PEXEL residues were mutated, cleavage, and consequently, export, was inhibited entirely.
A substantial proportion of the chimeric protein pools were present in the ER, as shown by colocalisation with ERC. We believe that GFP derived from cleaved chimaeras should account for little of this colocalisation, as degradation of reporter proteins to GFP alone is thought to occur in the food vacuole (30)
. It is likely that this fluorescence represents transit through the ER of the chimeric proteins that are exported, or which traffic to the parasitophorous vacuole, and examples of this include the KAHRPwt
chimaeras and a truncated STEVOR-GFP chimaera described recently (32)
. In contrast, the KAHRPR>A
pools that have an intact signal sequence appear to have been retained in the ER membrane, explaining the high level of fluorescence observed there. It was surprising to identify by MS a subpopulation of the KAHRPR>A
chimaera that had retained the signal sequence, and we infer that similar subpopulations of the KAHRPL>A
also retained the signal sequence, the latter of which is of particular interest as there is no PEXEL remaining. One possibility to explain this is that the signal peptidase complex is involved in cleaving the PEXEL, as has been recently suggested (24)
. It is also possible that the PEXEL protease resides in close proximity to the signal peptidase, or even segregates/cleaves this motif beforehand, and that a combination of overexpression and inhibition of PEXEL processing because of mutation is associated with signal peptide retention. Interestingly, we did not specifically observe GBP130 with a signal sequence retained, and a significant proportion of mutant GBP130 chimaeras (particularly RILE>A) were observed in the saponin supernatant, suggesting localisation in the parasitophorous vacuole. Immunofluorescence microscopy also confirmed less GBP130RILE>A
in the ER compared with KAHRPRLQ>A
and this is likely to be because of the retention of the signal sequence for the latter chimaera. These differences between GBP130 and KAHRP chimaeras may be because of a number of factors, including different temporal expression (), possible differences in episomal expression (), the length of the atypical signal sequence of GBP130, which is 35 residues longer than that of KAHRP, or the distance between the predicted SignalP processing site and the PEXEL, which is 5 residues shorter for GBP130 than KAHRP. Although the exact nature of the PEXEL protease is unknown all the current data support the notion that cleavage occurs in the parasite ER (24)
. We consider it unlikely that our PEXEL chimaeras docked at the external leaflet of the ER membrane (29)
, as this would require processing of the PEXEL, which is BFA insensitive (24)
, on the cytoplasmic side of the ER membrane and this would release soluble proteins into the cytoplasm. Furthermore, we detected processing of the signal sequence of KAHRP and GBP130 when PEXEL processing was inhibited, indicating that these mutant chimaeras had access to the ER lumen.
Blockage of PEXEL processing in the KAHRP chimaera allowed determination of the signal peptidase cleavage site, which was identical to that predicted by SignalP (33)
. This is the first confirmation of the signal sequence processing site in KAHRP, and of any P. falciparum
protein, and indicates that signal peptidase specificity is conserved with other eukaryotes. Additionally, processed KAHRP was acetylated at the new N-terminus following signal sequence cleavage, a feature that was also found in other N-terminal peptides in our proteomic analysis. The role of N acetylation following signal sequence processing in P. falciparum
is unknown; however, as the KAHRPR>A
chimaera did not export correctly, it by itself is insufficient for export to the erythrocyte.
We have used soluble exported proteins as models in this study. There are a class of exported proteins that have a signal sequence, PEXEL and up to two putative transmembrane regions that would be inserted into a membrane in the parasite-infected erythrocyte (31)
. In the proteomic analysis, we have identified proteins with putative transmembrane spanning regions that are cleaved after leucine within the PEXEL. It is likely that they are trafficked in a similar manner as soluble proteins, either as soluble complexes or through the ER membrane to the vesicular machinery.
While cleavage of the PEXEL motif at leucine, revealing xE/Q/D at the resulting N-terminus, is a requirement for export to the erythrocyte, the route and machinery involved are unknown. Endomembrane protein transport involves trafficking through protein-coated vesicles to target membranes. Polymerisation of COPII proteins at the transitional ER generates vesicles for transport from the ER and the heptameric COPI complex, along with ARF1, generates vesicles for reterograde transport (reviewed in 34
). To achieve sorting, transmembrane cargo can interact directly with coat proteins; however, soluble cargo requires transmembrane receptors (reviewed in 35
). Alternatively, proteins may traffic indirectly by bulk flow (36)
. Plasmodium falciparum
contains all the COPII subunits (37
and at least four COPI subunits (alpha, beta, delta and epsilon) (42
; however, the exact nature of vesicles that fuse with the parasite membrane is unknown. There are two likely pathways for export of proteins from the ER to the parasite-infected erythrocyte (). In the first model, proteins to be exported are cotranslationally inserted into the ER membrane using the signal sequence, allowing cleavage by signal peptidase and/or recognition of the PEXEL by a specific protease that cleaves this motif. This reveals xE/Q/D at the N-terminus of the protein, which is trafficked to the Golgi and parasitophorous vacuole through the general secretory pathway using the COPII machinery. Once released into the parasitophorous vacuole a translocon, that has been hypothesised previously (2
, would recognise the cleaved motif for export to the erythrocyte. While this model is possible, it would seem less likely as some proteins that are not exported, but may be secreted into the parasitophorous vacuole, would have xE/Q/D revealed after cleavage of the signal sequence (Boddey and Cowman, unpublished), and these mature proteins would be indistinguishable from those destined for export.
Role of the PEXEL in export of Plasmodium falciparum proteins to the infected erythrocyte
In the second model, proteins with xE/Q/D at the N-terminus after PEXEL cleavage in the ER may be sorted by a specific transmembrane PEXEL cargo receptor (soluble PEXEL proteins) or through interaction of the cytoplasmic domain (PEXEL proteins with transmembrane domain) into vesicles by the COPII machinery (). Because an exported chimaera was recently shown to traffic through the Golgi (32)
, it is also possible that sorting occurs there but this again would rely on a minimal sorting sequence after PEXEL cleavage (xE/Q/D). It is unknown if N acetylation combined with xE/Q/D provides the sorting information. Vesicles with cargo to be exported rather than secreted would be functionally differentiated (29)
by vesicle soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptors (v-SNAREs) and would bind to the cognate target SNARE at specific domains at the parasite membrane, releasing the cargo into the parasitophorous vacuole. It has previously been reported that PEXEL-containing chimaeras within the parasitophorous vacuole have the appearance of a necklace of beads that are resistant to recovery after photobleaching, suggesting the presence of subcompartments within this space (7
. These compartments may house a translocon that identifies and translocates proteins trafficked there. Thus, nonexported proteins would traffic through the default secretory pathway, which may involve bulk flow, to compartments in the parasitophorous vacuole that lack the translocon.
There are clearly many gaps in our understanding of the intricacies of the export pathway in P. falciparum. However, this work describes the initial molecular events required for effector protein export and, in defining the role of the PEXEL, significantly enhances our understanding of the export cascade in P. falciparum-infected cells. With malarial resistance to current therapies increasing worldwide, this export pathway represents an excellent, novel target for the design of new antimalarial drugs.