Molecular basis of pollen-related food allergy: identification of a second cross-reactive IgE epitope on Pru av 1, the major cherry (Prunus avium) allergen

Abstract

Birch (Betula verrucosa) pollen-associated food allergy is a well-characterized syndrome, which is due to the cross-reactivity of IgE antibodies to homologous allergens in various foods. One crossreacting area on the major birch pollen allergen Bet v 1 and its homologue in cherry (Prunus avium) Pru av 1 has already been identified. This is the so-called 'P-loop' region, which encompasses amino acid residues around position 45 and is found on the two virtually identical tertiary protein structures. We tried to determine an additional IgE cross-reacting patch on Pru av 1 and Bet v 1. The putative IgE-binding region on Pru av 1 was localized with a mAb (monoclonal antibody) that was generated against Bet v 1, and cross-reacts with several Bet v 1 homologues in food and inhibits the binding of patients' IgE to Pru av 1. mAb reactivity pattern was analysed and amino acid positions 28 and 108 of Pru av 1 were selected and mutated by site-directed mutagenesis. The Pru av 1 mutants were produced as recombinant proteins and characterized for their folding, mAb- and IgE-binding capacity and allergenic potency with a cellular assay using the humanized rat basophilic leukaemia cell line RBL-25/30. Amino acid position 28 is involved in a second major IgE-binding region on Pru av 1 and probably on Bet v 1. The identification of this second major IgE-binding region is an essential prerequisite to understand the phenomenon of cross-reactivity and its clinical consequences, and to produce hypoallergenic proteins for an improved immunotherapy of type I allergy.

Figure 1. Patients' IgE inhibition assay

Inhibition of IgE binding of three different sera from cherry allergic patients (sera no. 2, 5 and 6) to rPru av 1 by purified mAb mP16, as measured by ELISA.

Figure 2. Structure of Pru av 1 (Protein Databank no. 1E09)

Amino acids involved in known and hypothetical IgE-binding regions are highlighted: Glu⁴⁵ (E45), Asn²⁸ (N28), Pro¹⁰⁸ (P108). The structure was modelled with Swiss PDB Viewer 3.7 [33].

Figure 3. Critical amino acids for human IgE binding

Sequence alignment of Pru av 1 with Bet v 1 isoforms Bet v 1a and Bet v 1l, and the homologues Cor a 1.0401, Cor a 1.0101, Api g 1.0101, Api g 1.0201 and Dau c 1.0104. Reactivity pattern of mAb mP16 was used for the alignment to determine sequence differences between Bet v 1 homologues that reacted with mP16 (Pru av 1, Bet v 1a and Cor a 1.0401) and that did not react with mP16 (Bet v 1l, Cor a 1.0101, Api g 1.0101, Api g 1.0201 and Dau c 1.0104). Alignments are divided into two groups for selection of amino acids critical for putative IgE-binding region (group 1, sequence identity 58–65%; group 2, sequence identity 39–42% to Pru av 1). Group 1 (A, amino acid position Asn²⁸): Pru av 1, Cor a 1.0401 and Bet v 1a have an asparagine residue in position 28, whereas Bet v 1l and Cor a 1.0101 have a lysine residue. Group 2 (B, amino acid position Pro¹⁰⁸): Dau c 1.0104 and Api g 1.0101 have an alanine residue in position 108, instead of proline as in Pru av 1. The output of MultAlin 5.4.1 [34] was edited with ESPript 3.02. White residues on black background, strict identity; bold black residues, similarity within a group; boxed residues, similarity across groups; arrow, mutation sites.

Figure 4. CD spectra of Pru av 1 wt and its mutants

The spectra overlay showed that the mutants possess a nearly identical distribution of secondary structure elements in comparison with Pru av 1 wt.

Figure 5. Direct binding of mP16 Pru av 1 wt and to the mutants Asn²⁸→Lys (N28K), Pro¹⁰⁸→Ala (P108A) and Asn²⁸→Lys/Pro¹⁰⁸→Ala (N28K, P108A) measured by ELISA

Figure 6. Surface plasmon measurement of the interactions between immobilized mAb mP16 with Pru av 1 wt (A), Bet v 1a (B) and Pru av 1 Asn²⁸→Lys mutant (C)

Analyte concentrations: 111 nM, 37 nM, 12.3 nM. The broken lines are theoretically calculated.

Figure 7. Comparison of IgE-binding capacity of rPru av 1 wt and mutants

Patients' sera (n=25) were tested for their binding to Pru av 1 wt and its mutants by EAST. A statistically significant reduction in IgE reactivity was found for the mutants Asn²⁸→Lys (N28K) and Asn²⁸→Lys/Pro¹⁰⁸→Ala (N28K, P108A). U/ml, units/ml.

Figure 8. Inhibition of IgE binding to Pru av 1 wt measured by EAST

The pooled serum of 4 patients was pre-incubated with rPru av 1 and mutants as inhibitors. Pru av 1 Asn²⁸→Lys (N28K) and Asn²⁸→Lys/Pro¹⁰⁸→Ala (N28K, P108A) mutants showed a clearly reduced inhibition capacity compared with Pru av 1 wt.

Figure 9. Mediator release of Pru av 1 wt, mutants and mAbs as inhibitors with humanized RBL 30/25 cell line

Cells were sensitized with sera from individual cherry allergic patients and stimulated (i) with Pru av 1 wt and the three Pru av 1 mutants (A, B), and (ii) with Pru av 1 wt pre-incubated with mAb mP16 or G4a cell culture supernatant (C, D). The mediator release was clearly reduced for the Pru av 1 Asn²⁸→Lys (N28K) and the Asn²⁸→Lys/Pro¹⁰⁸→Ala (N28K, P108A) mutants in comparison with the wt protein (A, B). Pre-incubation of mAb mP16 with Pru av 1 wt led to almost total reduction of release, whereas mAb G4a did not greatly reduce the mediator release (C, D). Release of Pru av 1 (%) is reduced in (C) and (D) compared with (A) and (B), because colour intensity of Pru av 1 reference solution was adjusted to mAb cell culture supernatant colour. 1e-5, 1×10⁻⁵ etc.