Review on-line at: www.actabp.pl Definition of immunogenic carbohydrate epitopes �

Carbohydrates are known as sources of immunological cross-reactivity of allergenic significance. In celery and in cypress pollen, the major allergens Api g 5 and Cup a 1 are recognised by antisera raised against anti-horseradish peroxidase and by patients' IgE which apparently bind carbohydrate epitopes; mass spectrometric analysis of the tryptic peptides and of their N-glycans showed the presence of oligosaccharides carrying both xylose and core alpha1,3-fucose residues. Core alpha1,3-fucose residues are also a feature of invertebrates: genetic and biochemical studies on the fruitfly Drosophila melanogaster, the parasitic trematode Schistosoma mansoni and the nematode worm Caenorhabditis elegans indicate that these organisms possess core alpha1,3-fucosyltransferases. Various experiments have shown that fucosyltransferases from both fly and worm are responsible in vivo and in vitro for the synthesis of N-glycans which cross-react with anti-horseradish peroxidase; thus, we can consider these enzymes as useful tools in generating standard compounds for testing cross-reactive carbohydrate epitopes of allergenic interest.

The role of carbohydrates in allergy has been much discussed in recent years and is still controversial.It is not disputed that patients with pollen, food and venom allergies oĞen have IgE recognising N-glycan epitopes (Fötisch & Vieths, 2001); however, the clinical significance of these antibodies has been doubted (van Ree, 2002).Such lack of clarity only serves to show that suitable reagents are highly necessary in order to study the role of carbohydrates in allergy.Monovalent glycoconjugates, such as free oligosaccharides or glycoproteins such as pineapple bromelain which carry only a single glycan, may be suitable for testing binding of IgE, but are not suitable for studying histamine release, since they are not capable of cross-linking IgE receptors.
The IgE-binding carbohydrates are oĞen Nglycans carrying core xylose and/or core α1,3-fucose; these are features absent in mammals (see Fig. 1) and are, therefore, immunogenic (Wilson, 2002).Core α1,3-fucose is also a feature of the Nglycans of many invertebrates, and antisera raised against N-glycosylated plant and invertebrate proteins, such as horseradish peroxidase, carrot fruc-tosidase or bee venom, contain antibody species recognising these 'foreign' elements.Also there is immunological cross-reactivity between plant and insect glycoproteins; i.e., an antibody raised against a plant glycoprotein (e.g., horseradish peroxidase) can recognise insect glycoproteins (e.g., bee venom) and vice versa (Prenner et al., 1992;Wilson et al., 1998).Anti-horseradish peroxidase is probably the most widely used antiserum for detecting the presence of core xylose and/or core α1,3-fucose and for about twenty years it has been used to track neuronal pathways in invertebrates (Jan & Jan, 1982;Snow et al., 1987;Siddiqui & CuloĴi, 1991;Haase et al., 2001); however, only our recent studies have shown directly the role of core α1,3-fucose as the key cross-reactive determinant in the case of Drosophila (Fabini et al., 2001).
It has also become clear that carbohydrates are a factor in parasitic diseases; nematode N-glycans and glycolipids sometimes carry phosphorylcholine, a component also present on the surfaces of some pathogenic bacteria, and appear to be immunomodulatory (HarneĴ & HarneĴ, 2001) -in this case, the phosphorylcholine may be acting as 630 2005 K. Paschinger and others a "Tarnhelm", which, as in the Nibelungen legend, could confer invisibility on its bearer.Interestingly, too, insect glycolipids carry a related structure, phosphorylethanolamine (Seppo et al., 2000), which is also present in some bacteria; however, although antibodies can be raised against this structure, no other immunological data is available.Furthermore, IgE recognising core α1,3-fucose has been detected in parasite-infected sheep (van Die et al., 1999), whereas fucose-containing glycans may have a role in the TH2 response towards nematode proteins (Tawill et al., 2004).
The biotechnological repercussions of the immunogenicity of these structures are many-fold.First, the use of insect cell or whole plant systems for the expression of therapeutic glycopharmaceuticals needs to be carefully considered, so that systems that either naturally do not express core xylose and/or core α1,3-fucose are chosen, or that the ability to express these epitopes is repressed.Secondly, the use of recombinant enzymes allows us for the first time to generate defined reagents with properties suitable for the adequate testing of antibody binding and immunological activity, as has been performed with recombinant Arabidopsis core xylosyl-and fucosyltransferases (Bencúrová et al., 2004).Thirdly, using genetic approaches, we can now start to consider the role of core α1,3-fucose in invertebrate biology.
In our studies, we have examined core α1,3fucosylation in plants and invertebrates using a variety of techniques: use of neoglycoconjugates, glycan analysis, expression of recombinant enzymes and DNA/RNA-based manipulation of fucosylation.These results pave the way for a beĴer understanding of the biological and immunological significance of glycosylation.

ANTI-CARBOHYDRATE EPITOPES OF PLANT GLYCOPROTEINS
Celery is considered to be one of the major sources of food allergy and a number of celery proteins have been shown to bind IgE from patients' sera.Some of this binding can be inhibited by preoxidising the proteins with periodate, an indication that carbohydrate chains may be the epitopes for patients ' IgE (Bublin et al., 2003).To study this more specifically, patients' sera were preincubated with neoglycoconjugates: either bovine serum albumin carrying bromelain glycopeptides (called BSA-MUXF 3 ) or bovine serum albumin carrying the pentasaccharide core structure common to eukaryotic N-glycans (so-called BSA-MM).Indeed, the former, and not the laĴer, glycoconjugate prevented binding of IgE, in a manner similar to that mediated by intact native Api g 5 allergen.This suggested that Api g 5 may carry N-glycans with the MUXF 3 structure.Indeed, MALDI-TOF MS analysis verified that celery Api g 5 carries N-glycans with the probable compositions MUXF 3 and MMXF 3 (Bublin et al., 2003).It was therefore concluded that the binding of IgE to Api g 5 is due to the presence of glycans carrying core xylose and α1,3-fucose.
In another study, the glycans of a cypress pollen allergen were investigated.Allergy to cypress pollen is common in Mediterranean countries and the binding of patients' IgE is oĞen also sensitive to prior periodate oxidation (Afferni et al., 1999).Purified Cup a 1 was examined by Western bloĴing, using anti-horseradish peroxidase as the primary antibody.The antibody was preincubated in the presence and absence of various concentrations of the aforementioned BSA-MUXF 3 as well as Immunogenic carbohydrate epitopes a defucosylated variant BSA-MUX (Fig. 2).The results showed a strong inhibition, particularly with the BSA-MUXF 3 conjugate, suggesting that the allergen carries core α1,3-fucosylated N-glycans; this was compatible with a previously-published MALDI-TOF MS N-glycan analysis (Alisi et al., 2001).

ANTI-CARBOHYDRATE EPITOPES OF INVERTEBRATE GLYCOPROTEINS
Using similar methods, N-glycan analysis and antibody inhibition assays, we began to examine the origin of anti-horseradish peroxidase cross-reactivity towards invertebrate glycoproteins.Glycan analyses of both Caenorhabditis and Drosophila indicated the presence of difucosylated N-glycans with probable MMF 3 F 6 -type structures (Fabini et al., 2001;Haslam & Dell, 2003).Indeed, in the case of the Drosophila N-glycans, the presence of these glycans was also determined by RP-HPLC in conjunction with exoglycosidase digestion.Interestingly, the nematode has a far more complicated N-glycan spectrum than the fruitfly, even though one would expect a fly to be more complex than a small 'worm' which, however, has many more genes.Preincubation of antihorseradish peroxidase with BSA-MUXF 3 inhibited binding to Caenorhabditis extracts in Western blots or to Drosophila embryonal neural tissue (Fabini et al., 2001).BSA-MUX, however, was relatively ineffective as an inhibitor, a result compatible with the presence of core α1,3-fucose, and the absence of core xylose, in these species.
Our work with Caenorhabditis was expedited by the availability of fucosyltransferase mutants: we were able to find one mutant which did not display anti-horseradish peroxidase reactivity and which lacked certain fucosylated N-glycans (Paschinger et al., 2004).This mutant (VC378) has a deletion in the fut-1 (K08F8.3)gene.Therefore, we expressed the cDNA in both Pichia pastoris and in Drosophila Schneider S2 cells.In the case of the yeast expression system, the activity of the recombinant enzyme was tested with various N-glycan substrates (Fig. 3) and only activity towards MM was detected; with the Schneider cells system, it was shown that transfection of fut-1 cDNA conferred ectopic expression of the anti-horseradish peroxidase epitope.The recombinant FUT-1 can also be used to re-create the antihorseradish peroxidase epitope in vitro (Paschinger et al., 2004).Similar studies with Drosophila suggested that FucTA (CG6869) gene was the fucosyltransferase responsible for the biosynthesis of this epitope in the fruitfly (Fabini et al., 2001).
In contrast to the nematode and the fly, the trematode worm Schistosoma mansoni expresses glycoconjugates containing both core xylose and core α1,3-fucose (Khoo et al., 1997).A previous study showed that we could detect both xylosyl-and fucosyltransferase activities that modify N-glycan substrates in schistosome egg extracts (Faveeuw et al., 2003).Therefore, it was of no surprise to see that anti-horseradish peroxidase binding to glycoproteins in schistosome extracts was inhibited by both BSA-MUXF 3 and BSA-MUX (K.Paschinger, unpublished).However, the genes responsible for the presence of  Recombinant yeast-expressed FUT-1 was incubated for 17 h at room temperature with a variety of dabsyl-glycopeptide substrates in the presence of GDP-fucose.The incubation was then analysed by MALDI-TOF MS.Only the MM substrate was converted to its fucosylated derivative.For an explanation of the N-glycan nomenclature, see Fig.

CONCLUSION
In our studies on plant and invertebrate glycosylation, we have dissected to some extent the origins of inter-species immunological cross-reactivity.A mixture of biochemical, analytical and geneticbased tools have been used and we are now at the threshold of being able to pursue the use of these in studies to uncover the wider biological and allergological significance of carbohydrate epitopes; results which will also have a bearing on the use of nonmammalian expression systems in biotechnology.

Figure 1 .
Figure 1.N-glycans referred to in this review.

Figure 3 .
Figure 3. Assay of recombinant Caenorhabditis FUT-1 with N-glycans.Recombinant yeast-expressed FUT-1 was incubated for 17 h at room temperature with a variety of dabsyl-glycopeptide substrates in the presence of GDP-fucose.The incubation was then analysed by MALDI-TOF MS.Only the MM substrate was converted to its fucosylated derivative.For an explanation of the N-glycan nomenclature, see Fig.1.
Figure 3. Assay of recombinant Caenorhabditis FUT-1 with N-glycans.Recombinant yeast-expressed FUT-1 was incubated for 17 h at room temperature with a variety of dabsyl-glycopeptide substrates in the presence of GDP-fucose.The incubation was then analysed by MALDI-TOF MS.Only the MM substrate was converted to its fucosylated derivative.For an explanation of the N-glycan nomenclature, see Fig.1.
Figure 3. Assay of recombinant Caenorhabditis FUT-1 with N-glycans.Recombinant yeast-expressed FUT-1 was incubated for 17 h at room temperature with a variety of dabsyl-glycopeptide substrates in the presence of GDP-fucose.The incubation was then analysed by MALDI-TOF MS.Only the MM substrate was converted to its fucosylated derivative.For an explanation of the N-glycan nomenclature, see Fig.1.
this parasite are yet to be identified.