One of the major goals of our study was the characterisation of A. viteae chitinase genes. This involves the description of the genomic structures as well as the characterisation of different transcripts of the individual stages of the life cycle of A. viteae.

According to our studies, the genome of A. viteae contains three different chitinase genes.

Southern blot analysis showed, particularly for the Pvu II digest, that there are at least two different chitinase genes. However, the different band intensities of especially the Pvu II digest could be explained by the fact that the probe bound at least twice more to close gene copies on a single DNA fragment. This indicates that there is a cluster with of a minimum of two chitinase genes. A similar suggestion was made by Arnold and colleagues (1996), who investigated gDNA from B. malayi for chitinase genes. The variation in the intensity of individual bands was also explained by the existence of a gene cluster. Further analysis of nine genomic sequences from a gene library of A. viteae worms confirmed the existence of three independent chitinase gene sequences. The sequences of two genomic clones (1 and 9) had a cluster of two genes as predicted from Southern hybridisation results, and an identical overlapping region with some minor differences. The corresponding genomic region of two inserts comprises a length of 6,845 bp, of which 93% was identical between two pairs of genes. The main differences arose from the beginning of insert 1, the end of insert 9, a repeating sequence of 25 nucleotides and a non-matching region in the intergenic region of both sequences. An additional argument for the identity of both sequences is the fact that the introns of both pairs of chitinase genes were 100% identical. This strongly indicates that the two pair of sequences are identical because intron sequences should show a higher mutation rate (Neafsey et al., 2005) since they are not under evolutionary pressure as exon sequences coding for particular amino acids responsible for the function of a protein. This hypothesis was confirmed by sequencing two independent clones (13 and 4) with similar restriction maps like clones 1 and 9. Analysis showed that the incomplete gene sequences at the beginning of clone 1 and at the end of clone 9 were truncated versions of gene sequences II and III, respectively. The discrepancies initially observed could have been due to sequencing errors or to artefacts inherent in the construction of the genomic libraries.

In consequence, there were three independent genomic chitinase sequences that were very similar in the structure, and showed the highest degree of variation in exons for the serine threonine rich domain in the 3’ part of the genes. A comparison with the structure of the B. malayi MF 1 chitinase gene showed that the whole structure of the genes is identical to the A. viteae genes, with the only difference being in the exons coding for the serine threonine rich domain. There are three other annotated chitinase genes in B. malayi having a broad range of organisations with some domains and exons fully, partially represented or absent (). The absence of some exons in these genomic sequences could be seen as a footprint of unequal recombination occurring between exons of homologous and nonallelic chitinase genes (Maeda and Smithies, 1986). Interestingly, two of these chitinase genes are found in tandem on the same genomic assembly. Taken into account that a similar situation was demonstrated for A. viteae, it may be surmised that these genes exist in a cluster in these filarial parasites and are consequently a hot spot for recombination.

A. viteae chitinase gene sequence II, in contrast to sequences I and III, had no first exon with a start ATG and probably no functional transcript. A comparable sequence is found in the genome of B. malayi. Both sequences lack a first exon, and have an irregular second exon, a situation which may have arose by unequal crossing over resulting to the loss of the 5’ exon with the start ATG (Maeda and Smithies, 1986). While A. viteae gene II has a regular stop codon and polyadenylation site, and may thus be considered as part of a true gene, the B. malayi sequence is definitely a pseudogene due to the presence of proretrovirus-like elements after the 10^th exon (Felder et al., 1994), and to an alteration of the reading frame.

The gene structure for the chitinase gene of the free living nematode C. elegans has a different distribution of exons coding for the enzymatic domain and for the serine- threonine rich linker structure. An unusually long exon codes for the glycosyl hydrolase domain responsible for the catalytic activity of the deduced protein, and for most of the serine threonine rich linker region. The exons coding for the secretory signal sequence and the chitin binding domain correspond to those in the filariae. The C. elegans chitinase is an orthologue of the filarial chitinases, despite the lack of similar introns and conservation of all intron positions and phases. This is congruent with the clustering of all of these sequences in a phylogenetic tree, and to the fact that no other C. elegans chitinase has a better match to filarial chitinases than this C. elegans chitinase. A similar observation was made by comparing an 83 kb synthenic region between B. malayi and C. elegans and demonstrating that corresponding orthologous genes clearly had differences in gene structure (Guiliano et al., 2002). Apart of this chitinase, there are some other 30 expressed chitinase genes in C. elegans, whose implication and function have not been addressed (www.wormbase.org; Popovici et al., 1999).

It is not known why C. elegans has many more chitinase genes than the filarial nematodes. The C. elegans chitinase gene family is made up of many orthologues and homologues that may have arisen by duplication and diversion of a common ancestor gene (Popovici et al., 1999).Though the function of all the members of this gene family is not yet known, diversion may have led to new copies with different functions as well as tissue or substrate specificities. Some of the genes are very identical with up to 10 nonallelic copies clustered on one chromosome, while others have very low identities and are widely dispersed on different chromosomes.A general conclusion drawn from the C. elegans chitinases would be that there may be more members of a gene family than one could suspect on the basis of protein analysis and by the use of probes to fish out identical genes of the same family. This is supported by the fact that many of the chitinase genes share common motifs, but have very low sequence identities. In line with this hypothesis, there may be several other chitinase genes in the filarial nematodes which have not been characterised based on the low identities they share with known family members. Support for this hypothesis comes from the B. malayi irradiated L3 chitinase EST that has a bare 32 % identity to the published sequence for B. malayi chitinase. Another observation showing that such a phenomenon is present in the filarial genome is the description of two chitin synthases in Brugiawitha bare 27% identity (Harris et al., 2000; Foster et al., 2005).The existence of orthologous genes with the same function in a parasite makes it difficult to effectively use these genes as vaccine or drug targets. Thus, the further characterisation of filarial chitinases as drug or vaccine targets will entail the identification and/or exclusion of further functional chitinase genes in the nematode of interest.

In keeping with the identification of three chitinase genes in A. viteae it was hypothesised that there could be several stage-specific A. viteae chitinase genes which may have different substrate specificities or functions. There are two published A. viteae chitinase cDNA sequences(Adam et al., and Wu et al., 1996) that differ only in the outmost 5’ ends. The Wu et al., sequence begins with a start methionine and three additional amino acids which are absent in the Adam et al., sequence. Within the frame work of this study, data from cDNA and genomic sequences could confirmthat the Wu et al., sequence is the variant that is expressed in vivo. The exons of A. viteae genomic gene I are 100% identical to the sequence of Wu and colleagues, particularly with regards to the 5’ end. Transcripts could actually be found for this sequence in uterine microfilaria, blood microfilariae, L3 and L4, while no transcripts could be found for gene sequence III. It would however not be excluded that other A. viteae chitinase-like transcripts could be present in A. viteae, since western blot data show the existence of four molecules reacting with antibodies to A. viteae chitinase. It was shown that sera of Meriones protected by vaccination with irradiated L3 were mainly directed against L3 molecules of 205 kDa, 68 kDa and 17 KDa (Lucius et al., 1991). The 205 and 68 KDa molecules were characterised as chitinases (Adam et al., 1996). In addition, monoclonal antibodies produced from sera of animals immunised with irradiated attenuated L3 and Excretory secretory products (ESP) could recognise an array of molecules in different filarial stages in immunoblots (Adam et al., 1996). Amongst others, a 68 kDa chitinase could be shown for L3 and L4, while 220 kDa, 205 kDa and 140 kDa molecules were found in female worms, L3 and microfilariae, respectively. The authors proposed that the 205 kDa L3 molecule could be as a result of oligomers formed from the 68 kDa monomers, or post-synthetic modification of the monomer. The antigens from female worms were shown to be from uterine microfilariae (Adam et al., 1996). The implication of such a finding could be that there are possibly different stage-specific chitinases in microfilariae and L3 of A. viteae. While we have demonstrated that the same chitinase molecule (68 kDa) is expressed in mf, L3 and L4, it remains to be shown that there are other chitinase messages in different parasite stages.

Filarial chitinases had been exclusively attributed to the L3 and mf stages of A. viteae and O. volvulus (Wu et al., 1996; Adam et al., 1996) or to the microfilarial stages of Brugia (Fuhrman et al., 1992; Arnold et al., 1996). However, Brugia EST data point to the existence of a novel chitinase transcript in B. malayi L3, a stage from which no chitinase gene had previously been characterised. In this study, we showed chitinase transcripts in uterine microfilaria and L3 and L4, but not in the cDNA library from adult female worms, even though the uterine microfilariae were derived from gravid female worms. A similar difficulty was found in the isolation of B. malayi chitinase from microfilaria (Southworth et al., 1996) and O. volvulus chitinase (Perler, personal communication) from a female cDNA library. In both cases, the reasons advanced were the low representation of chitinase transcripts in the total transcript population of the filarial stages looked into, or suboptimal isolation of targeted transcripts from these filarial stages (Southworth et al., 1996; Perler, personal communication).

A comparison of the deduced amino acid sequences from the three A. viteae chitinase gene sequences revealed a very high rate of identity except for the serine threonine rich linker region. The variation in the serine threonine rich linker regionof A. viteae chitinase is also found when different filarial chitinase sequences are compared. It is not known whether the size and composition of the serine threonine rich linker are of particular relevance to the function of the protein. However, this region has several consensus sequences for glycosylation and when glycosylated will influence protein solubility, secretion and the resistance of the protein against proteases as has been demonstrated for insect chitinases (Arakane et al., 2003). Such properties could be of particular relevance to the host-specificity of parasites and their chitinase molecules.

The function of chitinases in filarial worms is not yet known. However, in protozoan parasites like Plasmodium and Leishmania, chitinase has been shown to be important for transmission by their role in lysis of the arthropod peritrophic matrix (Langer and Vinetz, 2001; Romalho-Ortigao and Traub-Cseko, 2003). While Simulium vectors of O. volvulus have been reported to have peritrophic membranes (Ramos et al., 1994; Demanou et al., 2003; Soumbey-Alley et al., 2004), such membranes have not been demonstrated in tick vectors of A. viteae (Sonenshine, 1991). It is therefore not known whether similar mechanisms operate in these vectors. However, L3 chitinases of O. volvulus and A. viteae could have a role in the egress of these parasites from their arthropod vectors since their expression coincides with the transmission of the parasites (Wu et al., 1996).

If chitin is the unique substrate for filarial chitinases, it may be speculated that chitinases play a role in parasite transmission based on two observations. Firstly, Brugia microfilariae express chitinase; secondly, it has been shown that Brugia microfilariae not expressing chitinase could not penetrate the arthropod vector (Fuhrman et al., 1987). It is hypothesised that lectins or agglutinins in the arthropod midgut binds to the microfilariae and reduces their infectivity. Chitinase promotes infectivity by digesting its substrate and releasing N-acetylglucosamine (GlcNAc) that saturates the lectins and reduces their interaction with microfilariae. Such a hypothesis has been proven in other parasite systems by supplementing the infection blood meal with N-acetylglucosamine, the monomers of chitin, which in turn increases infectivity (Welburn et al., 1993). Such a hypothesis can be tested in the filariae by carrying out similar experiments.

Parallel to the role of insect chitinases in ecdysis (Kramer and Muthukrishnan, 1997), filarial chitinases are speculated to be involved in moulting and hatching, since filarial eggs and adult sheath have been shown to contain chitin (Harris et al., 2000; Brydon et al., 1987). Such a role in ecdysis was suspected for A. viteae chitinase by the expression of chitinase on the surface of microfilaria, just before they hatch from their modified eggshell, and the absence of chitinase in blood-borne microfilaria (Adam et al., 1996).

There are three possibilities to control filarial infections: vector control, chemotherapeutic control, and immune control through vaccines. Chitinase is an attractive target for vaccine and drug control, since it is an enzyme that is expressed in two key stages (microfilariae and L3s) in the life cycle of filariae. In addition, knock out experiments with the C. elegans chitinase orthologue show that this gene is important for the development of larvae that are the equivalent of microfilariae in the filarial nematodes (). Moreover, knowledge on the structure of filarial chitinases will help towards the development of selective inhibitors against these molecules. Chitinase inhibitors might at the same time also target other chitin-containing pathogens like fungi and arthropods. For these reasons, it was necessary to produce large amounts of soluble and active A. viteae chitinase.

A bacterial expression system was chosen for the expression of A. viteae chitinase. The chitinase gene was cloned into the pET 22 b (+) vector (Novagen) and could be expressed in a suitable E. coli host. Expression and purification of large amounts of soluble full-length molecule was not possible, and so the N-terminal portion corresponding to the glycosyl hydrolase domain of the molecule was expressed.Failure to obtain a full-length molecule could have been due to difficulties in disulfide bond in the substrate-binding domain of the molecule that contains 6 cysteins. Amongst other things, the expression of soluble active chitinases from filarial paraistes (Southworth et al., 1996; Drabner et al., 2001) and other parasites like Leishmania (Razek-Desoukey et al., 2001) has always been difficult due to the 6 cysteins in the chitin binding domain, which disturbs proper folding. To overcome these problems of expressing soluble proteins in heterologous genes in E. coli, different strategies are used like reduction in induction temperature (Schein and Noteborn, 1989; Razek-Desoukey, 2001), using different bacteria expression hosts (Vinetz et al., 1999), fusion with a solubility-enhancer tag (Davis et al., 1998), export of the protein into the periplasmatic space (Southworth et al., 1996), and expression of truncated forms or independent domains of the molecule (Southworth et al., 1996).

The glycosyl hydrolase domain of chitinases folds independently of the substrate-binding domain (Synstad et al., 2004; Fusetti et al., 2003), and both domains are independently stable and active (Henrissat and Bairoch, 1993, 1996; Arakane et al., 2003) (http://afmb.cnrs-mrsfr/CAZY)). In this perspective, studies on chitinase activity and inhibitor binding could be made with the N-terminal glycosyl hydrolase domain of the molecule. For this reason, the N-terminal glycosyl hydrolase domain of A. viteae chitinase was cloned into the vector pET 22 b (+) and expressed in BL 21 (DE3) E. coli cells. Most of the protein was insoluble and could be purified only under denaturing conditions. There was a marginal increase in protein solubility when expression was done at lower temperatures. Export into the periplasmatic space produced little amounts of soluble material, which was 16 times more active compared to the cytoplasmic protein. A comparable 12-fold increase in the activity of a B. malayi MBP fusion chitinase was obtained upon export of the protein into the periplasmatic space (Southworth et al., 1996; Venegas et al., 1996). The increased specific activity of chitinase exported into the periplasma could be attributed to the better folding conditions in the periplasm as opposed to the cytoplasm (Rietsch et al., 1996; Raina and Missiakas, 1997; Sone et al., 1997).

Since most of the chitinase was expressed as insoluble material in the cytoplasm, bacteria hosts that enhance protein solubility in the cytoplasm were used for expression. Bacteria strains like Origami B (DE3) carry the trxB and gor mutations that enhance the formation of disulfide bonds in the cytoplasm (Derman et al., 1993; Aslund et al., 1999) and increase protein solubility. The use of this strain led to a 4-fold increase in chitinase solubility and a 13-fold increase in activity as compared to chitinase purified under denaturing conditions.

In order to obtain much more active A. viteae chitinase, current efforts are directed towards a eukaryotic expression system like insect cells. The reasons for this are two-fold. Firstly, glycosylation may lead to an increase in filarial chitinase activity (Fuhrman et al., 1992; 1995). Secondly, a commercially available eukaryotically expressed O. volvulus chitinase was shown to be 4-fold active compared to the A. viteae chitinase in this study. The eukaryotic expression system offers better folding conditions and the possibility of post-synthetic modifications, which are absent in the prokaryotic expression system (Vialard et al., 1995).

Several independent observations and experiments hint to the development of protective immunity against filariae. In natural endemic populations, there are putatively immune (PIs) individuals (Elson et al., 1994; Turaga et al., 2000) with no current or past evidence of filarial infections. In natural and surrogate model systems for filariasis, immunisation with irradiated infective stage larvae led to near-sterile immunity (Lucius et al., 1991; Prince et al., 1992; Lange et al., 1993; Taylor et al., 1994; Johnson et al., 1998; Trees et al., 2000). However, the expression of identified filarial vaccine candidates as E. coli expressed recombinant proteins or as DNA vaccines in various models led to varying rates of protection that did not match up to the almost complete protection rates achieved by the use of irradiated third-stage larvae (Lustigman et al., 2002; Cook et al., 2001; Abraham et al., 2002). This failure to mimic vaccination with attenuated parasites can be due amongst others to the absence of non-protein moieties, like in phosphorylcholine and carbohydrates in the recombinant vaccines (Nutman, 2002). In addition, native-protective antigens may have a different tertiary structure and thus a greater capacity to induce protective immunity as opposed to recombinant proteins. This is consistent with observations in the cestodes for which the initial use of native vaccine antigens has led to the development of successful vaccines (Lightowlers et al., 2003). The paucity of starting material however limits such an approach in the filarial nematodes. Thus the use of heterologous nematode protein expression systems may lead to the faithful expression of native proteins.

An understanding of the protective mechanisms in PIs and the irradiated L3 vaccination models will help in targeted development of potential vaccines. The protective mechanisms in PIs is not clearly associated with any of the arms of immunity: antibody responses (Stewart et al., 1995; Boyer et al., 1991) as well as Th1 and Th2 responses (Turaga et al., 2000) have been described. On the other hand, protective immunity associated with vaccination using irradiated larvae was clearly dependent on Th2 responses and involved antibodies and eosinophils (Le Goff et al., 2000; Lange et al., 1994; Bleiss et al., 2002). A role for antibodies in natural systems is provided by the elevated levels of IgG1/ IgG3 in protected subjects (Stewart et al., 1995) and the demonstration that sera from both infected and protected subjects mediate adherence and killing of microfilariae (Brattig et al., 1991) and infective larvae (Johnson et al., 1994).

Filarial chitinases could be considered as targets for immune attack based on two observations: the sera of PIs differentially recognise filarial chitinase molecules as opposed to clinically infected individuals (Rhagavan et al., 1994); and the sera of rodents immunised with irradiated A. viteae L3s dominantly recognised L3 chitinase in blots (Adam et al., 1996). Moreover, a monoclonal antibody, that mediated microfilarial clearance upon transfer, recognised B. malayi microfilaria chitinases in western blot (Canlas et al., 1984a; Canlas et al., 1984b).

In addition to these observations, confirmatory vaccination studies were done with filarial chitinases from O. volvulus leading to a 53% significant reduction in worm burden (Harrison et al., 1999).

It may therefore be surmised that eliciting the right antibody response against chitinase would contribute to immune protection.

The catalytic and substrate binding domains of chitinase have been shown to have different immunological properties (Venegas et al., 1996; Wu et al., 2001). We hypothesized that an increased antibody response to the glycosyl hydrolase domain and/or the active site of A. viteae chitinase will lead to the selective inhibition of chitinase activity and function, and an inhibition in the development of stages associated with this enzyme. The use of antibodies to epitopes or peptides from active sites of enzymes has been shown to be a successful tool for vaccination in cases where the enzyme appears at critical points in the life cycle of a parasite. Passive transfer of monoclonal antibodies binding to the active site of glutathione S-transferase of the nematode Schistosoma bovis impaired egg development, and this property was in turn correlatated with the enzyme inhibition (Da Costa et al., 1999). Using a similar strategy, it has been shown that vaccination of rabbits with synthetic peptides from the active site of Helicobacter pylori urease produced neutralising antibodies, and was a useful vaccination tool (Hirota et al., 2001). Synthetic peptides can be obtained from active sites of enzymes by homology modelling using other well characterised enzymes (Fujii et al., 2004).

The N-terminal catalytic domain of chitinase, as well as synthetic peptides designed from the active site of A. viteae chitinase 3-D model, was used in immunisation studies. Immunisation with the whole catalytic domain with STP as adjuvant led to a significant reduction in worm burden in a unique experiment. When alum was used as adjuvant, there was no tendency towards reduction in worm burden.

Despite the fact that STP was shown to be the better adjuvant for immunisation, alum was used for immunisation studies involving synthetic peptides, since our premise was to induce neutralising antibodies to the active site of chitinase. Alum is an adjuvant that directs the immune resonse to the Th2 direction /humoral arm. Immunisation with the synthetic peptides showed a tendency towards an overall decrease in worm burden in two independent experiments. In these same experiments, the group vaccinated with peptide 1 had a consistent significant reduction in microfilaria load, while the group vaccinated with peptide 2 had a significant reduction in microfilaria load in just one experiment.

Immunisation experiments were done just once with recombinant protein, and would therefore have to be repeated to confirm the trend observed. On the other hand, two independent vaccination studies were done using the synthetic peptides. The trend towards reduction in worm / mf burden was inconsistent when both peptides were used. It can therefore not be definitely concluded that vaccination with either synthetic peptides or recombinant protein conferred protection to challenge infection. Inconsistencies in the protective potential of highly immunogenic filarial proteins have been observed before (Peralta et al., 1999). In their study, Peralta and colleagues used differential screening with sera of a diverse group of filariasis patients to isolate highly immunogenic proteins from a B. malayi expression library. However, independent vaccination experiments with the same proteins produced inconsistent protection results (Peralta et al., 1999).

Vaccination with peptide 1 could consistently reduce the microfilaria burden. Vaccination with this peptide may interfere with hatching of eggs to microfilaria or the development and moulting of microfilaria and L3. Such a hypothesis will be consistent with the finding that chitinase is present on the cuticle of post-infective L3 of unsheathed filarial (Harrison et al., 1999) and on the surface of uterine microfilaria (Adam et al., 1996). If chitinase has a role in the the development of filarial worms,any process that interferes with enzyme activity could also distort worm development. Such a case could be expected by the production of antibodies to the active site of the enzyme by vaccination with synthetic peptides.Thus, these molecules may therefore have a role in the development of uterine microfilariae and post-infective L3 which is distorted by vaccination with chitinase. The possible mechanism for reduction in microfilaria burden is not clear, and should be understood by demonstrating the role of antibodies in the observed tendency towards reduction in worm burden.

The prospects for the future are three-fold: determination of the biological function of A. viteae chitinase, and its possible influence on host physiology; development of inhibitors to worm chitinases, and use of different immunisation approaches.

The function of A. viteae chitinase is not well defined. Therefore, future studies would be aimed at determining the biological function and possible influence on the host physiology. Reverse genetic studies like RNAi would be used to clarify the function of this gene in A. viteae. Three chitinase genes were described in this study. Analysis demonstrated transcripts in microfilariae, L3 and L4 larvae of A. viteae for gene I only, but not for gene III. To verify whether functional transcripts could actually be produced from gene III, an in vitro transcription / translation system would have to be used. In addition, other A. viteae stages will have to be screened for chitinase genes.

Soluble active A. viteae is required for the development of inhibitors to worm chitinases, and for crystallisation studies. The expression and purification of soluble active chitinase in prokaryotic systems was difficult since most of the protein was found in inclusion bodies. The expression of A. viteae chitinase in eukaryotic systems like yeast or insect cells would provide better alternatives. Moreover, the proteins will be post-synthetically modified in these systems. A comparison of the modification pattern with that of native A. viteae chitinase will illustrate the similarities in post-synthetic modification.

Vaccination with prokaryotically expressed A. viteae chitinase did not lead to significant reductions in worm burden. Since, native proteins have been shown to be better vaccines in cestodes (Lightowlers et al., 2003), A. viteae chitinase expressed in eukaryotic systems would be tested in immunisation experiments along side native and prokaryotically expressed A. viteae chitinase. This will give a clue as to whether the modified proteins could have better protective capacities as compared to non-modified proteins.