View Report Details

Do protein structural features determine the allergenicity of the alpha-class family of plant proteins?

Project Code: T07019

28/02/2004

University of Nottingham
Alcocer, M

Background

For reasons that are not understood, normal physiological processes are disrupted in allergic individuals leading to a continual IgE response towards a specific group of compounds, generally proteins. Contrary to common knowledge, few families of proteins are involved in IgE specific responses. In the entire plant kingdom for instance, there are only 19 protein families regarded as allergens (Pfam results not published). These observations suggest that the protein allergenic families might share common “intrinsic” attributes able to evoke an allergic response. Despite much effort, so far the search for such a common “intrinsic” attribute has proven illusive.

In the absence of a common factor, official agencies around the world have relied on parameters such as the sequence homology to other allergens and stability of the query protein to intestinal track digestion as assessment of the allergenic potential of unknown genetically modified foods.

The fact that not all proteins stable to digestion are allergenic, and not all individuals that consume “allergenic” proteins become allergic, suggest that the basic factors involved in the assessment of the allergenic potential of unknown proteins need to be improved.

In the present project we have concentrated our efforts on a model protein family that is wide spread in the plant kingdom, and has historically been implicated in severe cases of allergy to brazilnut. Distinct from the complex peanut allergy model currently in use, the model protein (Ber e 1) used in this study is a small protein (13 kDa), non glycosylated, with probably few IgE epitopes and that has been solely implicated as the major allergen in several clinical challenges. In order to be able to critically analyse our results, we have used, as negative control, a widely consumed protein from sunflower seed (SFA8) that has not been implicated in cases of allergy in Europe and possess similar 3D structure to our proposed model allergen.

Rationale and Objectives

We have proposed to use the well-known 2S albumin family of food allergens from brazilnut (Ber e 1) and from sunflower (SFA8) as our model protein system. We have cloned, isolated, purified and analysed the characteristics of these proteins regarding their stability, immunogenicity and potential allergenicity.

As the preparation of pure allergenic proteins is always problematic, several critical questions have been addressed throughout the project:

1. Can large quantities of recombinant Ber e 1 and SFA8 be produced?
2. Which levels of purity can be achieved?
3. Are they properly folded when compared with native proteins found in the seed?
4. Are the recombinant proteins heat, chemically and proteolytically stable?
5. Are they as stable as the native seed proteins?
6. Is the allergenic protein (Ber e 1) more stable to in vitro digestion than SFA8?
7. In order to map the allergenic determinant, can properly folded hybrids containing the 2 proteins be produced?
8. Do IgEs from subjects allergic to the seed protein recognise the recombinant proteins?
9. If so, is there any part of the molecule that is immunodominant?
10. Can the final 3D structure of the allergen in solution be calculated?
11. Do the recombinant proteins have the ability to induce IgE antibodies in vivo?
12. Which features of the protein molecules are important for the “intrinsic” allergenicity?

Approach and findings

We have cloned and expressed genes encoding the allergenic brazil nut 2S albumin (Ber e 1) and the sunflower albumin 8 (SFA8) in the methylotrophic yeast Pichia pastoris. We show that both proteins were secreted at high levels and that the purified proteins were properly folded. We also showed that Ber e 1 is glycosylated during secretion and that the glycan does not interfere with the folding or immunoreactivity. The disulphide map of the Ber e 1 protein was experimentally established and is in agreement with the conserved disulphide structure of other members of the 2S albumin family. A model three-dimensional structure of the allergen was generated. During the expression studies and through mutation we have also shown that alteration of the sequences around the Kex2 endoproteolytic processing site in the expressed fusion protein can compromise the secretion by targeting part of the protein for possible degradation.

In digestibility studies where the stability was measured in simulated gastric fluid followed by SDS-PAGE we have demonstrated for the first time that the clinically more important allergen from brazilnut (Ber e 1) is also the water-soluble protein least susceptible to digestion. The results corroborate the suggestion that abundance and exposure of intact protein to the gut immune system might have a role to play in allergenicity.

The stability of the proteins were further analysed by Circular Dichroism spectra. The proteolytic stability of Ber e 1 was reflected not only in the thermal stability but also in the resistance to unfolding displayed when the protein was exposed to low pH and chemical denaturant. Equally high resistance to proteolytic digestion, pH and chemical denaturant was observed with sunflower seed SFA8. The recombinant albumins mirrored their native counterparts in all stability assays. Interestingly, none of the proteins tested were fully denatured even at 95 C and all regained their original structure upon cooling. These results are not only relevant to the inherent stability of 2S albumins but also to food processing techniques possibly employed to destroy allergens.

Immunoblot and ELISA analysis carried out with sera from brazilnut allergic individuals have established that immunoreactivity of the recombinant proteins, as measured by humanspecific IgE, closely reflected those of corresponding native proteins.

Chimeric proteins in which loops and helix-loop-helix regions of the SFA8 proteins were replaced by the equivalent structures from Ber e 1 were assembled and expressed in Pichia pastoris. The folding of resulting chimeric proteins was confirmed by CD analysis. In order to be able to screen intact protein structures with a large number of sera, a protein microarray chip was then developed. Initial testes indicated that the loop region contained in helix-turn 3-helix of the Ber e 1 is the structural IgE immunodominant region for this allergen. The results are in agreement with similar epitope region been identified in three other 2 S albumin allergens Jug r 1 [1], Bra J 1e and Sin a 1 [2, 3] using different methodology.

Labelled ¹⁵N and ¹³C Ber e 1 was produced for structure analysis. The main part of the backbone was assigned using standard triple resonance NMR experiments. This work is still ongoing.

In animal experiments Ber e 1 can induce de novo IgE antibodies, although with a muchreduced titre than the control ovalbumin. An unexpected result from this experiment is that the purified protein alone seems not to induce IgE production. We have conclusively shown that some specific lipid-protein complexes are responsible for the IgE inducing ability of Ber e 1.

Outcome/Key Results Obtained

􀂉 We have shown that the allergenic Ber e 1 and SFA8 2S albumins were secreted at high levels from P. pastoris. Through mutation, we also showed that altered sequence around the Kex2 processing site can compromise secretion probably by targeting the fusion protein to destruction.

􀂉 We also showed that both proteins were properly folded and that the recombinant Ber e 1 is glycosylated, probably by O-linked short saccharides.

􀂉 The disulphide map of the recombinant Ber e 1 brazil nut protein was established and is in agreement with the map determined for other members of the 2S albumin family.

􀂉 The recombinant versions of the 2S albumins Ber e 1 and SFA8 produced in Pichia closely reflect the corresponding native proteins with respect to structure, immunogenicity and thermal/chemical stability.

􀂉 Ber e 1 is the most proteolytic resistant protein present in brazilnut seed. The stability of the recombinant version reflects the native protein and there is no significant proteolytic resistance difference between Ber e 1 and SFA8.

􀂉 In agreement with recent literature, our results indicate that stability to digestion should not be used as a predictive of selective immune responses.

􀂉 A protein microarray containing structural epitopes was developed.

􀂉 The structural epitope recognised by the human IgE antibody was mapped to a single helix-loop-helix region and is consistent with the results found in the literature for other 2S albumins.

􀂉 2S albumin chimeric proteins with structures similar to that of the native proteins provide unique tools to study antigen presentation.

􀂉 The determination of the 3D structure in solution of Ber e 1 it is in the final stages of calculations

􀂉 The purified “allergenic” protein Ber e 1, although immunogenic, is not able to induce IgE production in the mouse animal model system. Other molecules associated with the antigen might be essential to its allergenicity.

Main File

Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: