Proteomics of Eimeria : a focus on host cell invasion

Liz Bromley and Fiona Tomley
Institute for Animal Health, Compton, Berkshire, RG20 7NN, UK

Apicomplexan parasites are serious pathogens of man and domestic livestock for which there is an urgent need to develop novel, sustainable therapies based on new drugs or vaccines. The genomes of many species have been sequenced, or are nearing completion, including those of Plasmodium falciparum, Cryptosporidium parvum, Theileria parva, Theileria annulata, Toxoplasma gondii and Eimeria tenella. Detailed analysis of these genomes will be invaluable for developing an understanding of the biology and biochemistry of the Apicomplexa and for guiding the selection of novel, effective targets for drug and vaccine design. However, whilst genome sequences give good indications of the genes that are present within each parasite, they provide no clues about the expression or localisation of gene products, for example whether they are expressed at all, targeted to specific sub-cellular location, switched on or off at particular points during the parasite lifecycles or whether their level of expression is modulated by certain conditions. It is precisely these types of questions that proteomics, defined as the study of the full protein content of an organism, or an organelle, is designed to answer.

For obligate intracellular parasites such as the Apicomplexa, the molecular interactions between parasite and host-cell surfaces define uniquely each host-parasite relationship. The essential nature of host cell recognition, attachment and invasion, the repeated rounds of invasion that occur during the course of infections and the accessibility of the extracellular parasite make these interactions priority targets for intervention. The process of invasion is more or less conserved between most apicomplexans and consists broadly of (1) contact of parasite with host cell (2) reorientation of the parasite to make apical contact (3) tight attachment of the parasite apex to the host cell plasma membrane (PM) (4) rapid invasion of the parasite accompanied by deformation of the host PM to form a parasitophorous vacuole (5) pinching off of the PV from the PM. The process is driven by a parasite actinomyosin contractile motor and because host cell surface proteins are excluded from the PV membrane, parasites within the PV remain isolated from the host cell endocytic pathway. In recent years, a working hypothesis of apicomplexan invasion has evolved in which four classes of parasite proteins, in addition to the motor proteins, are implicated. These are GPI-linked surface antigens and a variety of soluble and membrane-bound proteins derived from the microneme (MIC), rhoptry (ROP) and dense granule (GRA) secretory organelles. However, as with the process of invasion itself there are many paradoxes that remain to be explored concerning apicomplexan secretory organelles. Chief amongst these is the fact that the number and type of organelles varies enormously between parasites, and between different developmental stages of the same parasite, and may even be absent. Thus micronemes range from a handful (Plasmodium merozoites and sporozoites) through to many hundred (Eimeria merozoites) per cell, rhoptries from zero (Plasmodium ookinetes) or one (Cryptosporidium sporozoites) to around a dozen (Eimeria sporozoites) and dense granules from possibly zero (Eimeria sporozoites) to around a dozen or more (Toxoplasma tachyzoites).

We have focused attention in our laboratory on defining the proteomes of the microneme and rhoptry organelles of Eimeria tenella with a view to understanding the precise function of these organelles. Using sequences of 36 characterised microneme, rhoptry and dense granule proteins from T. gondii and 21 from Plasmodium spp., we initially screened E. tenella genome and EST databases in silico to look for orthologues and homologues. Whilst we found some degree of conservation between the microneme proteins of the three species, largely due to the possession of common adhesive domains, there was little evidence for conservation of either rhoptry or dense granule proteins (for example, see Table below).

Comparison of organelle proteins between T. gondii and E. tenella

Using proteomics tools we have now characterised the protein content of gradient-purified preparations of micronemes and rhoptries from E.tenella sporozoites. The methods for purifying these organelles is long-established in our laboratory (Kawazoe et al, 1992, Parasitology 104, 1-9) and whilst there is some debate about whether dense granules are present in these stages of Eimeria, these organelles would most likely be isolated in the same fraction as the rhoptries. Organellar proteins were separated by two dimensional gel electrophoresis, which fractionates proteins on the basis of both their pI and molecular mass. After staining with a modified silver stain, protein spots were excised, digested with trypsin and analysed by various methods including matrix-assisted laser desorption/ionisation time-of-flight (MALDI-TOF) mass spectrometry (MS), and tandem electrospray MS. Interpretation of the mass spectra and comparison of these data with in silico protein predictions derived from the Eimeria genome and other databases using Mascot software resulted in identification of a large proportion of the protein spots.

Whilst 2D gels have the advantage over 1D gels of allowing differentiation between several proteins of the same molecular mass, they are acknowledged to be unsuitable for the separation of certain types of protein, including very large or low abundance proteins, hydrophobic proteins and proteins of extreme pI. Each large format 2D gel was loaded with parasite organelle protein purified from 5-10 infected chickens. However even when the technique was fully optimised with help from 2D gel experts it was clear that the amounts of proteins visible on the 2D gels did not reflect the quantities of protein loaded, particularly for the rhoptry fractions. Therefore we also separated batches rhoptry proteins by 1D SDS PAGE and found that many more of the larger proteins were separated so these were also analysed by Tandem MS.

The proteins identified using these approaches fell into three major groups. In the first group were exogenous non-Eimeria contaminants, such as trypsin inhibitor, that were introduced during subcellular fractionation. Second were Eimeria proteins that from their homologies to other known proteins are not expected to localise to the apical organelles and were most likely structural contaminants. The third group consisted of Eimeria proteins that most likely encode novel organellar proteins. This final group included 8 putative novel microneme proteins and 28 putative novel rhoptry proteins. We have selected a number of these novel proteins for further study, as demonstrated by some examples outlined below.

ROPF is a protein which shares some limited homology to hypothetical proteins from other apicomplexan parasites so may represent a novel class of organellar proteins from these parasites. ROPF is an 200kDa protein excised from a SDS PAGE gel and the sequences of tryptic peptides from this band all matched to a single contiguous genome sequence and to hypothetical proteins from two species of Plasmodium. Interestingly, peptides from proteins migrating above and below ROPF on SDS PAGE in the range 180-220kDa were also found to hit the same region of the genome Due to the overlap of the peptides from different proteins along the length of the gene we have concluded that these proteins are differently modified forms of expressed gene product. Sequencing the full-length cDNA of ROPF has revealed that it occupies >9kb of the genome and is organised over 22 exons. We have found evidence of extensive alternative splicing, which accounts for several of the isoforms and are currently sequencing clones from a ROPF mini-cDNA library to determine the exact sequence differences between the isoforms.

ROPJ is a putative rhoptry protein that shares homology with another set of Plasmodium hypothetical proteins, as well as scoring hits onto the genome and/ or ESTs of T. gondii, N. caninum and T. annulata, suggesting that this protein belongs to a well-conserved family of rhoptry proteins from across the whole apicomplexan phylum. We have sequenced the full-length cDNA of ROPJ, revealing that the gene is spread over 5kb of the genome, over 12 exons one of which is alternatively spliced between different cDNAs. ROPJ contains multiple predicted transmembrane domains and a putative C terminal tyrosine-based rhoptry sorting signal.

EtSUB is a subtlisin, a type of serine protease named after its similarity to a protease which was found to be secreted by the bacterium Bacillus subtilis. The active site of subtilisins consists of a characteristic catalytic triad of residues: aspartic acid, histidine and serine. The catalytic region of EtSUB has homology with T. gondii subtilisin TgSUB2, which itself is homologous to the P. falciparum subtilisin PfSUB2. The apicomplexan subtilisins are of interest to us as they are variously localised to all three apical organelles: the microneme (TgSUB1), the rhoptry (TgSUB2) and the dense granules (PfSUB1 and PfSUB2). Whilst EtSUB appears to be most similar to the rhoptry enzyme, TgSUB2 we have not yet definitively localised the enzyme to these organelles. We have sequenced the full length cDNA of EtSUB, which has revealed a transcript of approximately 1.1kb encoding an enzyme of ~114kDa. Again the gene is complex, being organised over 16 exons with one region of alternative splicing.

These three examples of putative rhoptry proteins highlight some of the characteristic features of E. tenella genes. They are typically complex, being organised over multiple exons and employing alternative splicing as a mechanism for gene expression (see http://www.tigr.org/tigr-scripts/tgi/splnotes.pl?species=e_tenella for other examples of alternative splicing).

In conclusion, many novel proteins of E. tenella have been identified through proteomic analyses of subcellular fractions of parasites enriched for either microneme or rhoptry secretory organelles. The relatively complex nature of these organelles is readily apparent and the analysis of organelle-specific proteomes of one parasite can lead to the identification of previously unknown homologues in other parasites. It is clear that an understanding of the molecular structure, processing, cellular context and precise interactions of target molecules will be critical to the rational development of effective intervention strategies against apicomplexan parasites. Also, significant insights into the molecular evolution of this complex and important group of pathogens can be obtained by comparative analysis of their organelle-specific proteomes. In collaboration with experts on Plasmodium and Toxoplasma we are now expanding work in this area with the intention of (i) using our expertise in cellular fractionation to purify organelles from other parasites where organelles are less abundant (ii) generating organellar proteomes from important apicomplexan parasites to build a comprehensive picture of the evolution, diversity and function of secretory organelles across the phylum and (iii) exploiting the more tractable reverse-genetic systems of Toxoplasma gondii and Plasmodium berghei to begin functional analysis of organellar proteins that are conserved between different genera, focusing initially on the genes that we have already identified in the current work. We are also currently undertaking a much more comprehensive, high-throughput proteomics study in which we aim to define whole cell proteomes for several different developmental stages of E. tenella using both 2DGE and liquid-chromatography separations, a process that will not only help to build up knowledge regarding stage-specific gene regulation in this parasite but will also served to help and support the mammoth task of accurate and comprehensive genome annotation.

We would like to thank the large number of people who have contributed to this work including Rich Oakes, Pierre Rivailler and others from the parasitology group at IAH; Mike Dunn and colleagues at Proteome Sciences, London; Jonathan Wastling and colleagues from University of Liverpool, Bob Sinden and colleagues from Imperial College and by no means least the entire Eimeria genome consortium for providing the essential sequence data that underpins the work.