THE SOLUBLE PROTEOME: WHAT DOES THE SCHISTOSOME CYTOSOL COMPRISE?

The first application of proteomics to schistosomes was a comparative anaylsis of the soluble proteins from different life cycle stages, using 2D electrophoresis for separation and quantitation and peptide mass fingerprinting for identification (Curwen et al. Reference Curwen, Ashton, Johnston and Wilson2004). Immunologists had devised a series of soluble antigen preparations derived from cercariae, lung schistosomula, adult worms and eggs (termed SCAP, SLAP, SWAP and SEA, respectively) for probing host humoral and cellular immune reactions to schistosome infection. The unstated assumption was that stage-specific aspects of the immune response could be characterized, an important consideration when vaccine candidates or pathogenic proteins were being sought. In fact, the study revealed that the four preparations had a remarkably similar composition, with more than 70% of their gel spots in common. Expressed in terms of total spot volume, i.e. protein content, the shared spots accounted for an even greater proportion than spot number alone indicates. Thus shared protein content ranged from 66% for SEA to 90·4% for SLAP, suggesting that very few proteins are specific to the lung schistosomulum, reinforcing previous ultrastrucural observations that revealed few unique features (Crabtree and Wilson, Reference Crabtree and Wilson1986). Put another way, approximately 71% of the protein content was shared between the four stages, on the basis of spot volume. Only the egg showed 9% of gel spots that were not shared with another stage. This may reflect both that the egg represents a completely different generation of the life cycle (ovum-miracidium-mother sporocyst), and that its role is to escape from the host, not to enter and persist in it.

Although the actual order of abundance differed between the stages, the common proteins comprised what might be called ‘housekeeping gene products’ involved in functional and structural processes to maintain parasite viability. Among them were glycolytic enzymes (aldolase, enolase, glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase, triose phosphate isomerase), chaperones (14-3-3, cyclophilin, HSP70), cytoskeletal/motor proteins (actin, myosin light chain, dynein light chain, tropomyosin), calcium binding proteins (calreticulin, EF hand proteins Sm 20.8 and Sm 21.7, calponin) and anti-oxidants (glutathione-S-transferases 26 and 28, superoxide dismutase, thioredoxin). Collectively they represent a cross section of the cytosol of the ‘average’ schistosome cell and provide a useful benchmark against which to judge the enrichment of other preparations being analysed. Where we have encountered these proteins in later studies, we largely interpreted their presence as an indicator of cell damage leading to leakage of the cytosol. Sample preparation then becomes crucial to the generation of ‘sensible’ proteomic data, an important consideration given later claims for the composition of schistosome ‘secretomes’ (Cass et al. Reference Cass, Johnson, Califf, Xu, Hernandez, Stadecker, Yates and Williams2007; Liu et al. Reference Liu, Cui, Hu, Feng, Wang and Han2009). It should also be noted that, over the years, several of the ‘housekeeping’ proteins have been proposed as vaccine candidates in spite of their evident intracellular origin (Wilson and Coulson, Reference Wilson and Coulson2006).

WHAT DO INFECTIVE CERCARIAE AND MIGRATING SCHISTOSOMULA SECRETE?

The ideal schistosome vaccine would prevent parasite establishment by blocking the skin invasion process, always a tall order because the immune system would need to be on permanent ‘red alert’. Firstly, each incoming parasite does not provide much of a stimulus to the sensitized host and secondly, there is no time for an anamnestic response during the small window when it is susceptible to antibody-dependent cellular cytotoxicity (ADCC) in vitro (Wilson and Coulson, Reference Wilson and Coulson2006), features neglected by proponents of this effector mechanism. The cercaria uses the contents of its pre-and post-acetabular glands, comprising about 16% of body protein (Harrop and Wilson, Reference Harrop and Wilson1993), to penetrate into host skin. Irrespective of their status as vaccine candidates, characterization of these secretions in two proteomic studies has revealed hitherto unsuspected aspects of the infection process. Note that both investigated only the soluble fraction of the acetabular gland secretions. To date, no-one has characterized the pelletable and sticky components that probably comprise proteoglycans and/or glycoproteins, used by the cercaria to aid adhesion to the stratum corneum (Stirewalt, Reference Stirewalt1959).

The first proteomic analysis of material released when cercariae are transformed into schistosomula in vitro used skin lipid-induced secretion and shearing through a small bore syringe (needle) (Knudsen et al. Reference Knudsen, Medzihradszky, Lim, Hansell and McKerrow2005) as the stimuli. It highlighted three known elastase isoforms plus Sm16 ‘anti-inflammatory protein’ (SPO-1), in part due to the restricted scope provided by searching of the NCBInr database rather than the full transcriptome. However, the vast majority of proteins identified were the internal soluble ‘housekeeping proteins’ listed above, and also included several nuclear histones. To what extent this composition can be attributed to damage incurred during artificial transformation, e.g. by syringe-shearing, or is the result of holocrine secretion of the acetabular gland cell contents, is a matter of conjecture.

The second study of acetabular gland secretions (Curwen et al. Reference Curwen, Ashton, Sundaralingam and Wilson2006) used the method of tail detachment by vortexing of the suspension (Ramalho-Pinto et al. Reference Ramalho-Pinto, Gazzinelli, Howells, Mota-Santos, Figueiredo and Pellegrino1974), known to provide viable schistosomula for in vitro culture, and 70–87% maturation when these are delivered to the portal vasculature of naïve mice (as good as that for ex vivo schistosomula from the lungs, returned to mice; Harrop and Wilson, Reference Harrop and Wilson1993). Two-dimensional electrophoretic separation of the proteins released during the first three hours after transformation revealed a relatively simple pattern of 144 spots on the gel, versus 1314 spots in a comparable separation of soluble cercarial extract (SCAP). The 50 most abundant spots accounted for 79% of the protein content, with 54% of vesicular and 25% of cytosolic origin (Table 1). Collectively the elastase isoforms accounted for the largest proportion (34·4%) with a hitherto unsuspected invadolysin metalloprotease (12·8%) as the second most abundant protein. Also notable were three of the enigmatic SCP (sperm coat protein) domain-containing proteins now classified as SmVALs. The inference must be that these particular VALs exert their functions on the host during or shortly after entry to the skin. Similar venom allergen-like proteins have recently been reported in proteomic analyses of the secretory products of the nematodes Ancylostoma caninum (Mulvenna et al. Reference Mulvenna, Hamilton, Nagaraj, Smyth, Loukas and Gorman2009) and Heligmosomoides polygyrus (Hewitson et al. Reference Hewitson, Harcus, Murray, van Agtmaal, Filbey, Grainger, Bridgett, Blaxter, Ashton, Ashford, Curwen, Wilson, Dowle and Maizels2011). The most abundant cytosolic protein was a serine protease inhibitor (4·9%). The other principal proteins in descending order of concentration were cyclophillin, thioredoxin, glutathione-S-transferase 28, aldolase, triose phosphate isomerase and Sm14-fatty acid binding protein, all typical constituents of the soluble proteome (Curwen et al. Reference Curwen, Ashton, Johnston and Wilson2004); the holocrine secretion of the acetabular glands is the obvious explanation for their presence.

Table 1.

Major proteins in the cercarial acetabular gland secretions

Table 2.

Membrane and membrane-associated tegumental proteins identified by proteomic analysis of the isolated tegument membranes

Data from Braschi et al., Reference Braschi, Borges and Wilson2006a, Reference Braschi, Curwen, Ashton, Verjovski-Almeida and Wilson2006b; Braschi & Wilson Reference Braschi and Wilson2006; Braschi and Wilson, unpublished.

Table 3.

Proteins identified by external biotinylation of live worms

* CD44 was identified in an unfractionated sample not included in the publication

Data from Braschi and Wilson (2006), which used v3 of the genome assembly.

Annotations are updated to version 4/5 in the third column.

Table 4.

Proteins identified by analysis of worm vomitus

Data from Hall et al., Reference Hall, Braschi, Truscott, Mathieson, Cesari and Wilson2011

Table 5.

Secreted proteins identifed after in vitro culture of eggs

* Mathieson et al. 2010; DeMarco et al. 2010.

The secretions of the migrating schistosomula have proved more difficult to characterize, not least because of the small amounts of material released. However, leaving aside possible gut contributions, there are two sources for such material, the head gland, which persists up to the lung stage of migration, and the tegument, which possesses a unique inclusion, the homogeneous body (McLaren et al. Reference McLaren, Hockley, Goldring and Hammond1978). Biosynthetic labelling of cercariae in-snail prior to transformation, or of schistosomula in culture, using radiolabelled methionine established that schistosomula released a dominant protein of ∼ Mr 20 kDa into the external environment (Harrop and Wilson, Reference Harrop and Wilson1993). This protein appeared very relevant to protective immunity, in view of work on the radiation-attenuated cercarial vaccine that had revealed the lung-stage schistosomulum as the target of protective responses in immunised mice (reviewed by Coulson, Reference Coulson1997). However, its identity proved elusive in spite of attempts to clone it and only in 2010 was it identified by proteomics on 2D gel spots as a mixture of two proteins, MEG-3.1 and MEG-3.2, encoded by microexon genes (DeMarco et al. Reference DeMarco, Mathieson, Manuel, Dillon, Curwen, Ashton, Ivens, Berriman, Verjovski-Almeida and Wilson2010 – see below). Furthermore, gene expression was localized to the head gland of the cercaria/schistosomulum by whole mount in situ hybridisation (WISH), the only transcript so far identified from that tissue (DeMarco et al. Reference DeMarco, Mathieson, Manuel, Dillon, Curwen, Ashton, Ivens, Berriman, Verjovski-Almeida and Wilson2010).

It is worth recording, given the description of the adult tegument proteome below, that our attempts to characterize the surface of the lung schistosomulum have proved problematic. The surface does not detach easily, and even a small number of dead or damaged larvae in the medium thwarts the use of surface biotinylation or enzymatic shaving techniques. This is especially disappointing in view of the apparent role the tegument surface in priming the protective response in mice exposed to irradiated cercariae (Riengrojpitak et al. Reference Riengrojpitak, Anderson and Wilson1998). We have therefore had to rely on comparative transcriptome data (Dillon et al. Reference Dillon, Feltwell, Skelton, Ashton, Coulson, Quail, Nikolaidou-Katsaridou, Wilson and Ivens2006; Verjovski-Almeida et al. Reference Verjovski-Almeida, DeMarco, Martins, Guimaraes, Ojopi, Paquola, Piazza, Nishiyama, Kitajima, Adamson, Ashton, Bonaldo, Coulson, Dillon, Farias, Gregorio, Ho, Leite, Malaquias, Marques, Miyasato, Nascimento, Ohlweiler, Reis, Ribeiro, Sa, Stukart, Soares, Gargioni, Kawano, Rodrigues, Madeira, Wilson, Menck, Setubal, Leite and Dias-Neto2003) to infer the likely tegument surface composition of intravascular schistosomula (Castro-Borges et al. Reference Castro-Borges, Dowle, Curwen, Thomas-Oates and Wilson2011a).

WHAT DOES ANALYSIS OF THE TEGUMENT SURFACE REVEAL ABOUT LIFE IN THE BLOODSTREAM AND IMMUNE EVASION?

The external surface of the adult worm that forms its interface with the host bloodstream has attracted attention since the first reports that it was a naked syncytial layer of cytoplasm, termed a tegument, and not an inert cuticle (Morris and Threadgold, Reference Morris and Threadgold1968). A key observation, not apparent in the initial studies, was that when worms were given an additional fixation in uranyl acetate, the apical surface appeared multilaminate (Hockley and McLaren, Reference Hockley and McLaren1973). The multilayered organisation was interpreted as a normal plasma membrane overlain by a secreted bilayer, termed a membranocalyx (Wilson and Barnes, Reference Wilson and Barnes1974), on the basis of detailed ultrastructural analysis and by analogy with the glycocalyx of normal cells. This model for the molecular architecture of the tegument surface is adopted for what follows. The tegument cytoplasm contains two inclusions, multilaminate vesicles and discoid granules, which originate in the cell bodies beneath the musculature (see review by Skelly and Wilson, Reference Skelly and Wilson2006). The fate of the glycan-rich discoid granules appears to be formation of the tegument ground substance (there is no convincing evidence for the external secretion of their contents). The much scarcer multilaminate vesicles, on the other hand, fuse with the plasma membrane at the base of the tegumental pits to release their contents; these unfold and spread out laterally over the surface to form the membranocalyx (Wilson and Barnes, Reference Wilson and Barnes1974, 1977). As this is a continuous albeit slow process, it follows that the membranocalyx must turn over into the external medium. It was suggested that it forms a physical barrier to protect the underlying proteins of the plasma membrane from attack (Wilson and Barnes, Reference Wilson and Barnes1974). Its extrusion over the surface in the first 24 h after transformation to the schistosomulum coincides with the loss of susceptibility to ADCC in vitro (Dessein et al. Reference Dessein, Samuelson, Butterworth, Hogan, Sherry, Vadas and David1981).

Compositional analysis

The approach adopted in York to characterize tegument composition involved an initial freeze/thaw/vortex (f/t/v) step that had the advantage of detaching areas of the surface in large sheets without the complications of detergent extraction (Roberts et al. Reference Roberts, MacGregor, Vojvodic, Wells, Crabtree and Wilson1983). Using alkaline phosphatase as a marker it was possible to obtain a 130-fold enrichment of the surface relative to the worm body and very reproducible patterns of protein separation by electrophoresis. We had to wait 20 years for the advent of proteomics to make any progress in defining tegument surface composition. The first analysis using the f/t/v method to isolate the whole tegument, fractionated the proteins by 1D SDS PAGE, digested in-gel and separated the tryptic peptides by liquid chromatography (LC, van Balkom et al. Reference van Balkom, van Gestel, Brouwers, Krijgsveld, Tielens, Heck and van Hellemond2005). It revealed only 43 out of 1543 distinct proteins (<3%) were unique to the tegument, compared with the stripped bodies, unsurprising given that all tegument proteins are synthesised in the cell bodies that lie below the musculature. The authors drew attention to five previously identified tegument proteins only one of which, the glucose transporter SGTP-4, was a true membrane-spanning protein.

Our initial experience with the tegument recovered after f/t/v and highly enriched by density gradient centrifugation (gradient pellet, GP) was similar, with glycolytic enzymes and cytoskeletal proteins the predominant components identified (Braschi et al. Reference Braschi, Curwen, Ashton, Verjovski-Almeida and Wilson2006b). The GP was therefore subjected to a sequence of extractions with chaotropic agents of increasing potency (1, Tris buffer; 2, urea/thiourea; 3, urea/thiourea/CHAPS/sulfobetaine SB 3.10 (UTCS)) to produce three supernantants and a final insoluble pellet (FP). The first three soluble extracts were subjected to 2D electrophoretic separation before MALDI ToF tandem mass spectrometry. The FP was divided equally and processed by either 1-D electrophoretic separation followed by trypsinisation and fractionation by liquid chromatography (GeLC-MS/MS) or digestion of the sample in solution, followed by two different liquid chromatography steps (multidimensional LC, MDLC-MS/MS). In this procedure, the bulk of glycolytic enzymes and some cytoskeletal material was recovered in the Tris extract. Cytoskeletal proteins predominated in the urea/thiourea extract whilst the UTCS sample contained relatively little novel material. The majority of membrane proteins were detected in the FP. It is clear from these results and later compositional studies (Braschi et al. Reference Braschi, Borges and Wilson2006a and unpublished data) that the tegument surface, rather than being inert, is metabolically very active. Solute and ion transporters with associated ATPases are particularly prominent membrane proteins (Table 2). A number of hydrolytic enzymes has been identified, as well as proteins potentially concerned with immune defence. A series of proteins associated with maintenance of membrane structure (tetraspanins, annexins, ferlins) is evident, the diversity perhaps reflecting the morphological complexity of the highly pitted tegument surface. Surprisingly, no signalling receptors capable of binding external ligands have been found, in spite of reports in the literature that adult schistosomes can respond to host hormones and cytokines such as insulin, TGFβ and TNFα (LoVerde et al. Reference LoVerde, Osman and Hinck2007; Oliveira et al. Reference Oliveira, Carvalho, Venancio, Miyasato, Kawano, DeMarco and Verjovski-Almeida2009; You et al. Reference You, Zhang, Moertel, McManus and Gobert2009, but see review by Wilson, Reference Wilson2011). The only indication of potential signalling pathways has been provided by the presence of a number of small guanosine triphosphate-binding (G) proteins, presumably localized on the cytoplasmic face of the apical tegument membrane. This compositional analysis of the isolated tegument membranes needs to be treated cautiously as it tells us nothing about the precise membranes being analysed or their location in the tegument. The basal invaginations of the tegument are also removed by the f/t/v procedure, so additional evidence is needed to confirm the location of particular membrane proteins at the tegument surface.

Several trematode-specific proteins were identified by direct compositional analysis, known from previous molecular studies only by their molecular weights (Sm200, Sm29, Sm25, Sm13 and LMWP). Remarkably, in view of the apparent invulnerability of the adult worm to attack, one of the most surprising findings was the presence of host antibodies (IgM, IgG1 and IgG3 classes) and components of the complement cascade (C3 and C4) but not those of the membrane attack complex C5-C9. The inference must be that complement fixation is initiated but then inhibited at an early stage. In this context a mouse complement inhibitor (Crry) was also detected (Braschi et al. Reference Braschi, Borges and Wilson2006a) and there is a previous report, based on inhibition assays and immunocytochemistry, of decay accelerating factor on the tegument surface (Pearce et al. Reference Pearce, Hall and Sher1990), suggesting that the parasite may utilise host proteins to inhibit complement. Another surprising observation was the presence of host CD44 in the membrane fractions. Recent observations have confirmed that the entire tegument surface is covered in this protein (Castro-Borges et al. Reference Castro-Borges, Dowle, Curwen, Thomas-Oates and Wilson2011a). In the host, CD44 is found on many cell types, a range of isoforms being generated by alternative splicing. It is decorated by a variety of glycan structures that affect its functions in cell adhesion and migration. The ramifications of its presence on the worm surface are currently being investigated.

Quantitation

The abundance of the major constituents of the tegument surface has been estimated using the QconCAT technique (Castro-Borges et al. Reference Castro-Borges, Simpson, Dowle, Curwen, Thomas-Oates, Beynon and Wilson2011b). An artificial protein was produced that is a concatamer of Quantitation (Q) peptides (hence QconCAT; Rivers et al. Reference Rivers, Simpson, Robertson, Gaskell and Beynon2007), each one having a different mass but otherwise identical physiochemical properties to its target peptide. These were selected from proteins we had already found in the tegument, thus facilitating their simultaneous quantitative estimation. A gene encoding 33 signature peptides was designed, and the concatamer was expressed heterologously in Escherichia coli grown in the presence of [¹³C6] lysine and [¹³C6] arginine (Rivers et al. Reference Rivers, Simpson, Robertson, Gaskell and Beynon2007) to label it with stable isotopes. The labelled QconCAT protein was then purified, quantified and a known amount added to two tegument preparations. Tryptic digestion of the QconCAT-sample mixture released each of the QconCAT peptides in a strict 1:1 stoichiometry and subsequent MS analysis allowed the quantification of each peptide present in the sample. Addition of the concatamer to the gradient pellet preparation of isolated tegument surface (Roberts et al. Reference Roberts, MacGregor, Vojvodic, Wells, Crabtree and Wilson1983) revealed aquaporin to be relatively the most abundant, followed by dysferlin, ATP-diphosphohydrolase, alkaline phosphatase, actin, annexin Smp_077720 and tetraspanin TSP-2 in descending order (Fig. 1). There is a question mark over the dysferlin that will only be resolved by localization studies, as in higher animals it is associated with the plasma membranes of muscle cells. The amounts of IgG1 and IgG3 detected were small both in molar (one third to one half that of the anion channel reference protein) and absolute terms (2·3 and 0·8%, respectively of the total GP). The number of molecules of C3 detected was only one tenth that of the anion channel, i.e. less than a third that of the two IgGs, amounting to 1% of the GP mass. This means that the bound immunoglobulins and complement factors are present at the tegument surface in small amounts, relative to their potential targets. The paucity of complement C3 is further evidence that no amplification of the complement cascade occurs.

Fig. 1.

Relative abundance of 21 S. mansoni tegument proteins in the purified gradient pellet using the QconCAT approach. The mean fold enrichment±S.D is expressed on a log scale relative to the anion channel as unity. The height of the bars is proportional to the number of molecules of each of the proteins, which are grouped according to their molecular function. Data reproduced from Castro-Borges et al. (Reference Castro-Borges, Simpson, Dowle, Curwen, Thomas-Oates, Beynon and Wilson2011b).

When the concatamer was added to the preparation released by PiPLC treatment of live worms, a very different result emerged (Castro-Borges et al. Reference Castro-Borges, Simpson, Dowle, Curwen, Thomas-Oates, Beynon and Wilson2011b). The dominant protein was Sm200 (87%) that represented only 2–3% of the GP in spite of its mass. Alkaline phosphatase (6·8%) and Sm29 (0·44%) were the other two GPI-anchored proteins estimated. It was concluded, on the basis of protein ratios, that Sm200 is present in a different ‘surface compartment’ from Sm29 and alkaline phosphatase. It may simply be that Sm200 is easily detached from the surface due to its size and GPI-anchor, or conceivably that it uniquely resides in the membranocalyx, which separates from the plasma membrane during GP preparation.

A model for the molecular architecture of the tegument surface

We first attempted to put tegument proteins into a molecular model of the tegument surface in 2006, based on the results of the first proteomic analyses (Braschi et al. Reference Braschi, Curwen, Ashton, Verjovski-Almeida and Wilson2006b; Braschi and Wilson, Reference Braschi and Wilson2006). This is very much ‘work in progress’ but we can expand on the earlier observations by adding newly discovered proteins, and also tentatively assign some to the membranocalyx (Fig. 2). It should be noted that the size and shape of the various components, has not been taken into account relative to the thickness of the two lipid bilayers. The most plausible membranocalyx component is Sm200, found in both membrane and cytosolic fractions. We still know little about the lipids that comprise the membranocalyx and give it such tremendous avidity for host proteins. Its permeability to the ions and solutes that must reach the transporters and enzymes of the plasma membrane suggests a relatively fluid organisation, rather than the presence of organised pores, particularly as the enzyme PiPLC (MW, 30 kDa) can apparently access the lipid anchors of proteins inserted in the outer leaflet of the plasma membrane (Castro-Borges et al. Reference Castro-Borges, Dowle, Curwen, Thomas-Oates and Wilson2011a). This fluidity, compared to the underlying plasma membrane, could have many sources. These include shorter chain fatty acids in phospholipids, a preponderance of glycoinositol phospholipids, no cholesterol, fewer membrane structural tetraspanins so an absence of the tetraspanin web, and almost certainly no linkage to the actin cytoskeleton. The way in which the membranocalyx can adhere to the plasma membrane yet also slide laterally over it has lead to the suggestion that there must be some kind of adhesion mechanism, a molecular Velcro, keeping the two layers together. The presence of annexin Smp_077720 external to, but associated with the plasma membrane, coupled with the lipid binding properties of annexins, provides one plausible candidate for the role. This hypothesis could be tested using RNA interference to knock down gene expression and determine the effect of tegument surface structure.

Fig. 2.

Diagram of the putative location of proteins in the membranocalyx and apical plasma membrane of the tegument surface. Note that the proteins and lipid bilayers are not depicted to scale. Host CD44 may be anchored in the plasma membrane, but with its N-terminus protruding above the membranocalyx surface. Data suggest that calpain may be outside the plasma membrane. The location of immunoglobulins and complement factors C3 and C4 are not shown.

In the current model, host CD44 is placed in the membranocalyx because it is readily detected by antibodies on the surface of live worms (Castro-Borges et al. Reference Castro-Borges, Dowle, Curwen, Thomas-Oates and Wilson2011a). However, its apparent absence from the CF pellet of sloughed membranocalyx suggests that it may be anchored in the plasma membrane. This is actually a serious issue because the first estimates of membranocalyx turnover, based on the CF sloughing, gave a t½ of 2–3 h (Wilson and Barnes, Reference Wilson and Barnes1977). This was later revised upwards to 5·4 days on the basis of a solid phase radio-immune assay using antibodies to erythrocyte markers on the worm surface (Saunders et al. Reference Saunders, Wilson and Coulson1987), on the assumption that these were glycolipids inserted into the membranocalyx (Goldring et al. Reference Goldring, Clegg, Smithers and Terry1976, Reference Goldring, Kusel and Smithers1977). If, as now seems likely, the antibodies were directed against CD44 and this host protein is anchored in the plasma membrane, then the turnover values obtained were not for the half-life of the membranocalyx, but for the slow displacement of host CD44.

BLOOD FEEDING: WHAT DOES WORM VOMITUS TELL US ABOUT GUT FUNCTION?

The second major interface with the host, the alimentary tract, is more difficult to investigate than the tegument because of its inaccessibility. At the cercaria/skin schistosomulum stage, a rudimentary gut is present, and actively ingesting material (Thornhill et al. Reference Thornhill, McVeigh, Jurberg and Kusel2010), but erythrocyte ingestion does not begin until the parasite has reached the hepatic portal distributaries of the liver (Clegg, Reference Clegg1965). The gut is blind-ending and in the adult runs the whole length of the body, partly as single and partly as bifurcated caeca. It comprises a single epithelial layer invested with tiny muscle fibres that provide the motive force for peristalsis. The absorptive surface is extended by lamellar plates, rather than more conventional microvilli (Hall et al. Reference Hall, Braschi, Truscott, Mathieson, Cesari and Wilson2011). Immediately anterior to the gut, and ventral to the oesophagus lies an oesophageal gland comprising 20–30 cells, really modified tegument cell bodies, that secretes material into the lumen.

Analysis of gene expression in the gut epithelium by RT-PCR has been performed using laser capture microdissection to recover tissue from sections (Nawaratna et al. Reference Nawaratna, McManus, Moertel, Gobert and Jones2011). This approach identified transcript for several transporters (Ca⁺⁺ATPase, phospholipid transporter, amino acid transporter, multi-drug resistance protein), a small number of lysosomal proteins (LAMP, NPC-2, acid lipase and phospholipase A2) and Cathepsins L, B and D2. However, collection of regurgitated vomitus has also provided valuable insights into gut function (Delcroix et al. Reference Delcroix, Medzihradsky, Caffrey, Fetter and McKerrow2007; Hall et al. Reference Hall, Braschi, Truscott, Mathieson, Cesari and Wilson2011). Fortuitously for tegument studies, as it turns out, schistosomes in vitro are very reluctant to open their mouths. This is in contrast to the liver fluke Fasciola hepatica that regurgitates quantities of vomitus from its proportionally larger gut continuously, playing havoc with tegument analysis (Wilson et al. Reference Wilson, Wright, de Castro Borges, Parker-Manuel, Dowle, Ashton, Young, Gasser and Spithill2011). Giving schistosomes cold shocks or even putting them in distilled water results in preparations highly contaminated with cytosolic and cytoskeletal proteins from the tegument and other tissues (Delcroix et al. Reference Delcroix, Medzihradsky, Caffrey, Fetter and McKerrow2007; Planchart et al. Reference Planchart, Incani and Cesari2007). The cleanest vomitus preparations have been obtained by the culture of ex vivo day 28 juvenile worms in RPMI-1640 medium for 4 days without protein supplements (Hall et al. Reference Hall, Braschi, Truscott, Mathieson, Cesari and Wilson2011). Proteomic analysis of the collected vomit confirmed the earlier biochemical studies (Caffrey et al. Reference Caffrey, McKerrow, Salter and Sajid2004; Delcroix et al. Reference Delcroix, Sajid, Caffrey, Lim, Dvorak, Hsieh, Bahgat, Dissous and McKerrow2006) that described the presence of a series of cathepsins and the pro-enzyme convertase, asparaginyl endopeptidase. Additional hydrolytic enzymes, including several proteases, largely of lysosomal origin, were also identified (Table 4 and Hall et al. Reference Hall, Braschi, Truscott, Mathieson, Cesari and Wilson2011). In the context of the hydrolytic environment, two protease inhibitors, a serpin and alpha-2-macroglobulin were identified, potentially performing a self-protective function to prevent auto-digestion. A series of carrier or binding proteins was also present, likely originating from lysosomes. These included saposins, NPC-2 (Niemann-Pick disease, type C), ferritin isoforms and calumenin, capable of binding lipids, cholesterol, ferric iron and calcium, respectively. The lysosomal markers, lysosome membrane-associated glycoprotein (LAMP) and vesicle associated membrane protein (VAMP), were also detected. In view of the known acid pH of the gut and the presence in vomitus of so many lysosomal proteins it was suggested that digestion of the haemoglobin and plasma proteins that comprise the schistosome diet was effected by secretion of lysosomes from the gut epithelium into the lumen. The long chain acyl-coenzyme thioesterase 1 identified, is normally a constituent of peroxisomes and may indicate that these organelles also make a contribution to luminal digestion.

As a corollary to the above, it was suggested that the function of the carrier proteins identified was to scavenge lipids and inorganic ions from the gut lumen by endocytosis, potentially receptor mediated. All the necessary components for receptor-mediated endocytosis have been identified in the S. mansoni transcriptome or genome, but none have yet been localised to any tissue (Wilson, Reference Wilson2011). Feeding experiments with starved adult worms, using fluorescent space-filling dextran as the probe, have revealed for the first time the occurrence of endocytosis at the luminal surface of the gut epithelium (Hall et al. Reference Hall, Braschi, Truscott, Mathieson, Cesari and Wilson2011). Similar experiments with labelled erythrocytes have also shown their uncoating as they pass through the oesophagus, with transfer of the lipophilic label PKH2 to the oesophageal lining (Hall et al. Reference Hall, Braschi, Truscott, Mathieson, Cesari and Wilson2011). This observation implicates the one known product of the oesophageal gland, the microexon-encoded protein MEG-4.1 (DeMarco et al. Reference DeMarco, Mathieson, Manuel, Dillon, Curwen, Ashton, Ivens, Berriman, Verjovski-Almeida and Wilson2010) originally described as Ag 10.3 (Davis et al. Reference Davis, Davis, Carroll, Rajkovic and Rottman1988), in the uncoating process.

ARE MICRO-EXON GENE PRODUCTS THE SCHISTOSOME'S ANSWER TO LIFE IN THE BLOODSTREAM?

The description of micro exon genes (MEGs), briefly in the S. mansoni genome paper (Berriman et al. Reference Berriman, Haas, LoVerde, Wilson, Dillon, Cerqueira, Mashiyama, Al-Lazikani, Andrade, Ashton, Aslett, Bartholomeu, Blandin, Caffrey, Coghlan, Coulson, Day, Delcher, DeMarco, Djikeng, Eyre, Gamble, Ghedin, Gu, Hertz-Fowler, Hirai, Hirai, Houston, Ivens, Johnston, Lacerda, Macedo, McVeigh, Ning, Oliveira, Overington, Parkhill, Pertea, Pierce, Protasio, Quail, Rajandream, Rogers, Sajid, Salzberg, Stanke, Tivey, White, Williams, Wortman, Wu, Zamanian, Zerlotini, Fraser-Liggett, Barrell and El-Sayed2009) and subsequently as a fuller exposition (DeMarco et al. Reference DeMarco, Mathieson, Manuel, Dillon, Curwen, Ashton, Ivens, Berriman, Verjovski-Almeida and Wilson2010) has added another dimension to our view of the parasite-host interface. MEGs appear to be a unique feature of the schistosome genome. At least 75% of the protein coding region is composed of microexons ranging in length from 6 to 36 base pairs, 92% of which are symmetrical in multiples of three (the commonest frequency being 21). This means that individual microexons can be spliced out without altering the reading frame, potentially forming variant transcripts by exon skipping. Other than a shared gene structure, the 18 MEG families (with up to 23 members) so far identified show no similarity with each other or with annotated genes from outside the Schistosoma genus. They also encode no identifiable motifs or functional domains, apart from a signal peptide, suggesting that they are secreted. Two have already been mentioned in this review (MEG-3; MEG-4.1), but it was research on the secretome of the schistosome egg that first brought them to prominence (Mathieson and Wilson, Reference Mathieson and Wilson2010). In addition, microarray analysis of changes in gene expression associated with infection of, and establishment in the mammalian host has shown that transcripts representing the vast majority of MEG families are highly enriched in the intramolluscan germ balls, cercariae, or day 3 schistosomula (Parker-Manuel et al. Reference Parker-Manuel, Ivens, Dillon and Wilson2011). Indeed, in the schistosomulum, MEGs 4.1, 8, 12, 14, 15, 16, and 17 showed the highest fold increases in expression, indicating the probable importance of the proteins they encode in parasite establishment and subsequent persistence in the mammalian host.

From studies on immune-compromised mice it is apparent that schistosome eggs secrete proteins (Ashton et al. Reference Ashton, Harrop, Shah and Wilson2001b), and recruit the host immune response to aid their passage through the tissues (Doenhoff et al. Reference Doenhoff, Hassounah, Murare, Bain and Lucas1986). The egg is laid undeveloped in an intestinal blood vessel and the miracidium differentiates from the ovum over 5–7 days. A syncytial sub-shell envelope arises but only when the miracidium is mature does this layer begin to release secretions that are crucial for escape (Ashton et al. Reference Ashton, Harrop, Shah and Wilson2001b), as opposed to those proteins that mediate granuloma formation and could originate from any compartment. The first published account of the egg secretome detailed the presence of a large number of cytosolic and cytoskeletal proteins (and even nuclear histones) released from eggs in culture (Cass et al. Reference Cass, Johnson, Califf, Xu, Hernandez, Stadecker, Yates and Williams2007). As many of these lacked a conventional leader sequence, the authors resorted to the use of SecretomeP to justify their presence in the secretions, released via a proposed non-classical secretion pathway. This programme was developed to predict the small number of bacterial and mammalian proteins that do exit cells via a novel route (Bendtsen et al. Reference Bendtsen, Jensen, Blom, Von Heijne and Brunak2004), and its use to justify the presence of cytosolic proteins in schistosome secretions has already been criticised (DeMarco and Verjovski-Almeida, Reference DeMarco and Verjovski-Almeida2009). The authors do not entertain the alternative possibility that a proportion of damaged or dead eggs in their cultures might be leaking the internal proteins (Cass et al. Reference Cass, Johnson, Califf, Xu, Hernandez, Stadecker, Yates and Williams2007). Mathieson and Wilson reported a much smaller number of proteins in the secretions of eggs, in which post-culture viability was >96%, assessed by observation of muscular and flame cell activity in unhatched miracidia (Mathieson and Wilson, Reference Mathieson and Wilson2010). These proteins were dominated by the now well characterized IPSE and Omega-1 (Fitzsimmons et al. Reference Fitzsimmons, Schramm, Jones, Chalmers, Hoffmann, Grevelding, Wuhrer, Hokke, Haas, Doenhoff and Dunne2005; Schramm et al. Reference Schramm, Gronow, Knobloch, Wippersteg, Grevelding, Galle, Fuller, Stanley, Chiodini, Haas and Doenhoff2006). On 2D gel separations IPSE alone accounted for >83% of the released material, while a few minor spots could also be discerned (Mathieson and Wilson, Reference Mathieson and Wilson2010). Using a lectin column to deplete the IPSE and Omega-1 proteins (Schramm et al. Reference Schramm, Gronow, Knobloch, Wippersteg, Grevelding, Galle, Fuller, Stanley, Chiodini, Haas and Doenhoff2006), the background spots were greatly enriched and could be identified by tandem mass spectrometry (DeMarco et al. Reference DeMarco, Mathieson, Manuel, Dillon, Curwen, Ashton, Ivens, Berriman, Verjovski-Almeida and Wilson2010; Mathieson and Wilson, Reference Mathieson and Wilson2010). They proved to be predominantly protein products of the MEG-2 and MEG-3 gene families, each represented by one to several spots on the 2D gel (Table 5). A lack of evidence for post-translational modifications strongly suggests that these are variant protein versions of the variant transcripts found by RT-PCR (DeMarco et al. Reference DeMarco, Mathieson, Manuel, Dillon, Curwen, Ashton, Ivens, Berriman, Verjovski-Almeida and Wilson2010). At least 13 spots were encoded by MEG-3 variants and 10 spots by MEG-2 variants.

The production of variant proteins by other blood dwelling parasites is a well-established phenomenon. The protective variant surface glycoprotein (VSG) coat of Trypanosoma brucei, encoded by >1000 VSG genes that are transcribed one at a time at telomeric expression sites, has been thoroughly documented (Taylor and Rudenko, Reference Taylor and Rudenko2006). An infected host mounts an effective antibody response against the predominant VSG variants, leading to antibody-mediated lysis. However, trypanosomes continually appear that have switched to antigenically different VSGs, leading to a chronic infection that can last for years. Similarly, Plasmodium falciparum possesses a family of clonally variant surface antigens, collectively termed PfEMP1 and expressed on the parasitised erythrocyte surface, encoded by about 60 var genes per haploid genome (Kyes et al. Reference Kyes, Kraemer and Smith2007). Switching of var gene expression allows the parasite to modify the antigenic and functional properties of infected erythrocytes, thereby evading immunity and affecting infection outcome.

The demonstration of the variant proteins in schistosomes raises many questions. The secretion of variant versions of a protein, not all of which can be functional, comes at a metabolic cost to the parasite, and leads to the obvious conclusion that they are involved in immune evasion (DeMarco et al. Reference DeMarco, Mathieson, Manuel, Dillon, Curwen, Ashton, Ivens, Berriman, Verjovski-Almeida and Wilson2010). If so, then schistosomes are unlike populations of the protozoan blood parasites Trypanosoma and Plasmodium, which produce their variants sequentially by switching gene expression, to confuse the immune system. Schistosomes generate their variants simultaneously, so any mechanism of evasion must differ from that of the protozoans. Potentially, the variants could interfere with immune priming of T cells by accessory cells if they increased the range of peptides presented by MHCII molecules, at the same time reducing the frequency of each. This is a question for the future and the most urgent task is to discover the function of one or more of these proteins at the parasite-host interface. An inference must be that if, in the course of schistosome evolution, breaking down exons into small coding cassettes confers a selective advantage then the functions of the encoded proteins must be important for parasite survival. That ties in neatly with their apparent secreted status.

CONCLUSIONS: DOES ANALYSIS OF THE SCHISTOSOME-MAMMALIAN HOST INTERFACE HOLD ANY PROSPECTS FOR DIAGNOSTICS OR VACCINES?

The conclusions drawn from this review need to be prefaced by a major caveat. The identification of schistosome proteins by tandem mass spectrometry is now so easy that researchers need to be severely critical in interpreting their accumulated datasets. Schistosomes obey the normal rules of cell biology, they are not leaky sieves oozing cytosolic proteins across their surface membranes. The first response to finding such proteins in parasite secretions should be to see them as evidence for damage, unless circumstances (such as holocrine secretion) indicate otherwise. The danger is in confusing the ‘necrotome’ with the ‘secretome’. That said, the characterization of proteins expressed on, or secreted from schistosome parasites in the mammalian host should provide opportunities to develop new diagnostics and vaccine candidates. The information obtained is most immediately applicable to the search for diagnostics; its translation to the vaccine sphere requires greater scrutiny.