Report on the potential allergenicity of genetically modified organisms and their products

Assessing protein digestibility in allergenicity risk assessment: A comparison of in silico and high throughput in vitro gastric digestion assays

The susceptibility of a novel food protein to digestion in the pepsin resistance test is widely used to inform the allergenicity risk assessment process. However, it does not model the variation in the intragastric environment found in vivo. Consequently a 96-well plate format in vitro gastric digestion protocol has been developed with a high and low pepsin activity test executed at pH 1.2, 2.5, 5.5 and 6.5. It was used to analyse seven allergens (from milk, egg, peach and peanut) and two non-allergens (cytochrome c and zein). Digestion was monitored using SDS-PAGE and densitometry. In silico predictions were not confirmed experimentally for most of the proteins studied. Proteins were ranked according to half-life and showed susceptibility to digestion was related to the stability of protein structure and protein solubility rather than allergenicity per se. Highly digestible proteins, such as β-casein and Ara h 1, generated abundant resistant fragments Mr > 3.5 kDa in the low pepsin activity test which could be immunologically significant within the context of allergenicity risk assessment for susceptible groups such as infants. The high- and low pepsin activity tests used in this study provided complementary data to support allergenicity risk assessment and used only 10 mg protein.

The relevance of a digestibility evaluation in the allergenicity risk assessment of novel proteins. Opinion of a joint initiative of COST action ImpARAS and COST action INFOGEST

The current allergenicity assessment of novel proteins is based on the EFSA GMO guidance. Recently, EFSA launched a new guidance document on allergenicity assessment of GM plants (2017). This document describes, amongst other topics, the new scientific and regulatory developments on in vitro protein digestibility tests. The EFSA GMO Panel stated that for in vitro protein digestibility tests, additional investigations are needed before any additional recommendation in the form of guidance can be provided. To this end, an interim phase is considered necessary to evaluate the revisions to the in vitro gastrointestinal digestion test, proposed by EFSA. This prompted the establishment of a joint workshop through two COST Action networks: COST Action ImpARAS and COST Acton INFOGEST. In 2017, a workshop was organised to discuss the relevance of digestion in allergenicity risk assessment and how to potentially improve the current methods and readouts. The outcome of the workshop is that there is no rationale for a clear readout that is predictive for allergenicity and we suggest to omit the digestion test from the allergenicity assessment strategy for now, and put an effort into filling the knowledge gaps as summarized in this paper first.

Food Allergens: Is There a Correlation between Stability to Digestion and Allergenicity?

Food allergy is a major health problem in the Western countries, affecting 3-8% of the population. It has not yet been established what makes a dietary protein a food allergen. Several characteristics have been proposed to be shared by food allergens. One of these is resistance to digestion. This paper reviews data from digestibility studies on purified food allergens and evaluates the predictive value of digestibility tests on the allergenic potential. We point out that food allergens do not necessarily resist digestion. We discuss how the choice of in vitro digestibility assay condition and the method used for detection of residual intact protein as well as fragments hereof may greatly influence the outcome as well as the interpretation of results. The finding that digests from food allergens may retain allergenicity, stresses the importance of using immunological assays for evaluating the allergenic potential of food allergen digestion products. Studies assessing the allergenicity of digestion products, by either IgE-binding, elicitation or sensitizing capacity, shows that digestion may abolish, decrease, have no effect, or even increase the allergenicity of food allergens. Therefore, the predictive value of the pepsin resistance test for assessing the allergenic potential of novel proteins can be questioned.

Evaluating pH-induced gastrointestinal aggregation of Arachis hypogaea 1 fragments as potential components of peanut allergy

The seed storage glycoprotein Arachis hypogaea (Ara h) 1 is a major allergen found in peanuts. The biochemical resistance of food proteins to protease digestion contributes to their allergenicity. The rapid proteolysis of Ara h 1 under gastric conditions challenges this model. Biophysical and in vitro digestion experiments were carried out to identify how Ara h 1 epitopes might survive digestion, despite their facile degradation. The bicupin core of Ara h 1 can be unfolded at low pH and reversibly folded at higher pH. Additionally, peptide fragments from simulated gastric digestion predominantly form noncovalent aggregates when transferred to base. Disulfide cross-links within these aggregates occur as intermediates in relatively low amounts only at early times and play no role in shielding peptides from degradation. It is proposed that peptide fragments which survive gastric conditions form large aggregates in basic environments such as the small intestine, making epitopes available for triggering an allergic response.

Characterization of the allergenic potential of proteins: an assessment of the kiwifruit allergen actinidin

Assessment of the potential allergenicity (IgE-inducing properties) of novel proteins is an important challenge in the overall safety assessment of foods. Resistance to digestion with pepsin is commonly measured to characterize allergenicity, although the association is not absolute. We have previously shown that specific IgE antibody production induced by systemic [intraperitoneal (i.p.)] exposure of BALB/c strain mice to a range of proteins correlates with allergenic potential for known allergens. The purpose of the present study was to explore further the utility of these approaches using the food allergen, actinidin. Recently, kiwifruit has become an important allergenic foodstuff, coincident with its increased consumption, particularly as a weaning food. The ability of the kiwifruit allergen actinidin to stimulate antibody responses has been compared with the reference allergen ovalbumin, and with the non-allergen bovine haemoglobin. Haemoglobin was rapidly digested by pepsin whereas actinidin was resistant unless subjected to prior chemical reduction (reflecting intracellular digestion conditions). Haemoglobin stimulated detectable IgG antibody production at relatively high doses (10%), but failed to provoke detectable IgE. In contrast, actinidin was both immunogenic and allergenic at relatively low doses (0.25% to 1%). Vigorous IgG and IgG1 antibody and high titre IgE antibody responses were recorded, similar to those provoked by ovalbumin. Thus, actinidin displays a marked ability to provoke IgE, consistent with allergenic potential. These data provide further encouragement that in tandem with analysis of pepsin stability, the induction of IgE after systemic exposure of BALB/c strain mice provides a useful approach for the prospective identification of protein allergens.

Allergic reactions compared between BN and Wistar rats after oral exposure to ovalbumin

There is currently no validated animal model for evaluating the potential allergenicity of food proteins. This study aimed to compare the allergic reactions between BN and Wistar rats after oral exposure to ovalbumin (OVA) by studying immune responses and clinical manifestations. Female BN and Wistar rats were orally exposed to OVA on days 1 and 14, and thereafter daily from day 15 to day 42. Sera and plasma were screened for OVA-specific antibodies and histamine. On day 49, all the OVA-sensitized animals were orally challenged with OVA before blood pressure was measured. One day later (on day 50), histopathology and differential cell counts were performed. The results indicate that oral exposure of BN rats to OVA yielded IgE, IgG, and IgG(2a) antibody responses that were generally of higher levels than those observed in Wistar rats (p < 0.05). However, the Wistar rats presented with more serious clinical manifestations and histopathologic changes that could have serious implications for any OVA-induced anaphylaxis. The studies here proved that OVA-sensitized BN and Wistar rats evinced different immune responses and clinical manifestations; these outcomes suggested that the two rat strains might differ in their immunologic mechanisms of allergy and that there was no correlation between immune responses and the severity of clinical symptoms. To be clear, the data from these studies should be viewed as 'preliminary', as only a single protein allergen was examined. Accordingly, further studies are needed to compare the allergic reactions between BN and Wistar rats by using purified strong-, weak-, and non-allergenic proteins based on the experiments reported here.

Influence of protein expression system on elicitation of IgE antibody responses: experience with lactoferrin

With increased interest in genetically modified (GM) crop plants there is an important need to understand the properties that contribute to the ability of such novel proteins to provoke immune and/or allergic responses. One characteristic that may be relevant is glycosylation, particularly as novel expression systems (e.g. bacterial to plant) will impact on the protein glycoprofile. The allergenicity (IgE inducing) and immunogenicity (IgG inducing) properties of wild type native human lactoferrin (NLF) from human milk (hm) and neutrophil granules (n) and a recombinant molecule produced in rice (RLF) have been assessed. These forms of lactoferrin have identical amino acid sequences, but different glycosylation patterns: hmNLF and nNLF have complex glycoprofiles including Lewis (Le)(x) structures, with particularly high levels of Le(x) expressed by nNLF, whereas RLF is simpler and rich in mannose residues. Antibody responses induced in BALB/c strain mice by intraperitoneal exposure to the different forms of lactoferrin were characterised. Immunisation with both forms of NLF stimulated substantial IgG and IgE antibody responses. In contrast, the recombinant molecule was considerably less immunogenic and failed to stimulate detectable IgE, irrespective of endotoxin and iron content. The glycans did not contribute to epitope formation, with equivalent IgE and IgG binding recorded for high titre anti-NLF antisera regardless of whether the immunising NLF or the recombinant molecule were used substrates in the analyses. These data demonstrate that differential glycosylation profiles can have a profound impact on protein allergenicity and immunogenicity, with mannose and Le(x) exhibiting opposing effects. These results have clear relevance for characterising the allergenic hazards of novel proteins in GM crops.

Digested Ara h 1 loses sensitizing capacity when separated into fractions

The major peanut allergen Ara h 1 is an easily digestible protein under physiological conditions. The present study revealed that pepsin digestion products of Ara h 1 retained the sensitizing potential in a Brown Norway rat model, while this sensitizing capacity was lost by separating the digest into fractions by gel permeation chromatography. Protein chemical analysis showed that the peptide composition as well as the aggregation profiles of the fractions of Ara h 1 digest differed from that of the whole pool. These results indicate that the sensitizing capacity of digested Ara h 1 is a consequence of the peptides being in an aggregated state resembling the intact molecule or that most peptides of the digests need to be present in the same solution, having a synergistic or adjuvant effect and thereby augmenting the immune response against other peptides.

Evaluating the effect of food processing on the potential human allergenicity of novel proteins: International workshop report

The ILSI Health and Environmental Sciences Institute Protein Allergenicity Technical Committee organized an international workshop in June 2006 in Estoril, Portugal, co-sponsored by the ILSI Research Foundation, ILSI International Food Biotechnology Committee and ILSI Europe. The objective was to discuss the effects of food processing on the allergenic potential of proteins and foods. The impact of food processing on the sensitization/induction phases of food allergy, and the bioavailability of allergens to the immune system were presented. Studies evaluating the stability, digestibility, and allergenicity of processed food allergens were identified, and their complexity and limitations discussed. Participants agreed that investigating food allergy mechanisms, validating appropriate methods for identifying allergenic proteins, and refining strategies to assess and manage the risks from food allergy were important before processing considerations are integrated into public-health decision-making for novel proteins. Other factors may also play a role in food allergy and include: food matrix; multiplicity of epitopes; geographic variation in patterns/prevalence of food allergies; and genetic factors, but required further exploration. Food processing may increase or decrease the intrinsic allergenicity of a protein, but current data do not facilitate the identification of specific variables that could be used to reliably determine how processing will influence protein allergenicity.

A mouse model for food allergy using intraperitoneal sensitization

Food allergy is an important health issue. With the increasing interest in novel foods derived from transgenic crop plants, there is a growing need for the development of approaches for the characterization of the allergenic potential of proteins. Although most foreign proteins are immunogenic (able to induce IgG antibody responses), relatively few are important food allergens with the capacity to provoke IgE antibody production. There is currently no validated animal model for the determination of allergenic potential of food proteins. One approach that appears to show some promise is outlined in the current chapter. BALB/c strain mice are immunized by intraperitoneal injection and the potential to cause allergenicity assessed as a function of the induction of specific IgE antibody, measured by homologous passive cutaneous anaphylaxis. Progress to date with this method is summarized, and comparisons are made with other experimental models, including considerations of route of exposure, use of adjuvants and selection of appropriate end points.

Evaluating the allergic risk of genetically modified soybean

Genetically modified (GM) soybean (carrying the EPSPS transgene) is the most common GM food in Korea. In order to assess whether genetic modification increases the allergenic risk of soybeans, the allergenicity and IgE-reactive components of wild-type and GM soybean extracts were compared in allergic adults who had been sensitized to soybeans. We enrolled 1,716 adult allergy patients and 40 healthy, non-atopic controls. Skin prick tests and IgE enzyme linked immunosorbent assays (ELISAs) were performed using wild-type and GM soybean extracts, along with other common inhaled allergens. The specificities of serum IgE antibodies from allergic patients and the identities of the IgE-reactive components of the soybean extracts were compared using ELISA inhibition testing, 2-dimensional gel electrophoresis, and IgE immunoblotting. To evaluate the effects of digestive enzymes and heat treatment, the soybean extracts were heated or pre- incubated with or without simulated gastric and intestinal fluids. The IgE sensitization rates to wild-type and GM soybeans were identical (3.8% of allergic adults), and circulating IgE antibodies specific for the two extracts were comparable. The results of the ELISA inhibition test, SDS-PAGE, and IgE immunoblotting showed a similar composition of IgE-binding components within the wild-type and GM extracts, which was confirmed using two-dimensional gel electrophoresis, IgE immunoblotting, and amino acid sequencing. None of the subjects had a positive response to purified EPSPS protein in the skin prick test, ELISA, or IgE immunoblot analysis. These findings suggest that the IgE sensitization rate to GM soybean extracts is identical to that of wild-type soybean extracts in adult allergy patients. In addition, based on both in vivo and in vitro methods, the allergenicity of wild type and GM soybean extracts was identical.

Characterisation of immune responses to food allergens in mice

There is considerable interest in the development and evaluation of approaches for the safety assessment of novel foods, and in particular in methods for characterisation of allergenic potential. One strategy that has found favour is a tiered approach in which the potential of novel proteins to induce allergic sensitisation is assessed based on considerations of stability of the protein in a simulated gastric juice and homology with, or structural similarity to, known allergens. Linked to such an approach may be evaluation of serological identity with proteins known to cause allergic disease. With the aim of supplementing such approaches with a more direct measurement of potential allergenic activity, attempts have been made to characterise the quality of immune responses elicited in BALB/c strain mice. Such evaluations comprise measurement of IgG and IgE antibody production and (to a lesser extent) of induced cytokine expression patterns. Investigations to date suggest that in mice proteins provoke variable immune responses, those with the potential to cause allergic sensitisation stimulating IgE (and IgG) antibody production. In contrast, non-allergenic, but nevertheless immunogenic, proteins are associated with IgG antibody responses in the absence of marked IgE production. Consistent with the selective activation of selective type 2 T lymphocyte responses, exposure of mice to allergenic protein is associated with preferential expression of IL-4, -5, -10 and -13. Collectively these data suggest that characterisation of the nature of immune response induced in mice by proteins may provide a useful adjunct or alternative to current strategies for the assessment of allergenic potential.

Intradermal exposure of BALB/c strain mice to peanut protein elicits a type 2 cytokine response

There is a growing need for the development of methods to characterize the allergenic properties of novel proteins, particularly those expressed by transgenic crop plants. Hence, there is considerable interest in the development of suitable animal models for this purpose. The production of specific IgE antibody has been reported following sensitization with food allergen via oral or systemic (intraperitoneal) routes of exposure. We have characterized cytokine profiles induced by intradermal treatment of BALB/c strain mice with a purified peanut allergen, Arachis hypogea lectin. Mice were exposed to peanut lectin by intradermal administration and the cytokine responses in the lymph node draining the site of exposure analyzed at the secreted protein level by enyzme-linked immunosorbent assay (ELISA) and cytokine mRNA level by ribonuclease protection assay (RPA). Exposure to peanut lectin, under conditions that induced robust IgE antibody titers, was found to be associated with a T helper 2 (Th2)-type cytokine expression profile at both the mRNA and secreted protein levels. Culture of naïve lymph node cells with peanut lectin failed to stimulate marked proliferation or cytokine production, confirming this protein is not mitogenic for mouse lymphocytes. Furthermore, the expression of Th2 cytokines was associated with the effector/memory CD62L- cell population. Similar treatment with a non-allergenic protein, potato acid phosphatase, failed to induce Th2 cytokine expression. These data demonstrate that exposure of mice to peanut allergen results in the selective stimulation of a Th2-type response.

A multi-laboratory evaluation of a common in vitro pepsin digestion assay protocol used in assessing the safety of novel proteins

Rationale. Evaluation of the potential allergenicity of proteins derived from genetically modified foods has involved a weight of evidence approach that incorporates an evaluation of protein digestibility in pepsin. Currently, there is no standardized protocol to assess the digestibility of proteins using simulated gastric fluid. Potential variations in assay parameters include: pH, pepsin purity, pepsin to target protein ratio, target protein purity, and method of detection. The objective was to assess the digestibility of a common set of proteins in nine independent laboratories to determine the reproducibility of the assay when performed using a common protocol. Methods. A single lot of each test protein and pepsin was obtained and distributed to each laboratory. The test proteins consisted of Ara h 2 (a peanut conglutin-like protein), beta-lactoglobulin, bovine serum albumin, concanavalin A, horseradish peroxidase, ovalbumin, ovomucoid, phosphinothricin acetyltransferase, ribulose diphosphate carboxylase, and soybean trypsin inhibitor. A ratio of 10U of pepsin activity/microg test protein was selected for all tests (3:1 pepsin to protein, w:w). Digestions were performed at pH 1.2 and 2.0, with sampling at 0.5, 2, 5, 10, 20, 30, and 60min. Protein digestibility was assessed from stained gels following SDS-PAGE of digestion samples and controls. Results. Results were relatively consistent across laboratories for the full-length proteins. The identification of proteolytic fragments was less consistent, being affected by different fixation and staining methods. Overall, assay pH did not influence the time to disappearance of the full-length protein or protein fragments, however, results across laboratories were more consistent at pH 1.2 (91% agreement) than pH 2.0 (77%). Conclusions. These data demonstrate that this common protocol for evaluating the in vitro digestibility of proteins is reproducible and yields consistent results when performed using the same proteins at different laboratories.

Statistical Evaluation of Local Alignment Features Predicting Allergenicity Using Supervised Classification Algorithms

BACKGROUND

Recently, two promising alignment-based features predicting food allergenicity using the k nearest neighbor (kNN) classifier were reported. These features are the alignment score and alignment length of the best local alignment obtained in a database of known allergen sequences.

METHODS

In the work reported here a much more comprehensive statistical evaluation of the potential of these features was performed, this time for the prediction of allergenicity in general. The evaluation consisted of the following four key components. (1) A new high quality database consisting of 318 carefully selected, non-redundant allergens and 1,007 sequences carefully selected to be non-allergens. (2) Three different supervised algorithms: the kNN classifier, the Bayesian linear Gaussian classifier, and the Bayesian quadratic Gaussian classifier. (3) A large set of local alignment procedures defined using the FASTA3 alignment program by means of a wide range of different parameter settings. (4) Novel performance curves, alternative to conventional receiver-operating characteristic curves, to display not only average behaviors but also statistical variations due to small data sets.

RESULTS

The linear Gaussian classifier proved most useful among the tested supervised machine learning algorithms, closely followed by the quadratic Gaussian equivalent and kNN. The overall best classification results were obtained with a novel feature vector consisting of the combined alignment scores derived from local alignment procedures using different substitution matrices.

CONCLUSIONS

The models reported here should be useful as a part of an integrated assessment scheme for potential protein allergenicity and for future comparisons with alternative bioinformatic approaches.

Evaluation of protein allergenic potential in mice: dose-response analyses

BACKGROUND

With the increasing interest in novel foods derived from transgenic crop plants, there is a growing need for the development of approaches for the characterization of the allergenic potential of proteins. Whereas immunogenicity is a common property of foreign proteins, including food proteins, relatively few are significant dietary allergens with the inherent capacity to provoke IgE antibody production and immediate-type hypersensitivity responses.

OBJECTIVE

In order to evaluate an approach for the measurement of the allergenic potential of proteins, detailed dose-response analyses of humoral immune responses induced following systemic exposure of BALB/c strain mice to proteins known to differ in terms of sensitizing activity have been conducted.

RESULTS

Mice were exposed to a range of concentrations of ovalbumin, a major allergenic constituent of hen's egg, a purified peanut allergen, Arachis hypogea agglutinin, or to the milk allergen bovine serum albumin, and to materials considered to lack significant allergenicity: a crude potato protein extract and a purified potato protein, potato agglutinin. The specific IgE antibody was measured by homologous passive cutaneous anaphylaxis assay, and the specific IgG antibody was measured by enzyme-linked immunosorbent assay. Each of the five proteins was immunogenic in mice, inducing IgG antibody responses at all doses tested, although there was some variation with respect to the vigour of IgG responses. Marked differences in the capacity of these proteins to induce IgE responses were observed, however, with relatively high-titre IgE antibody provoked by all three allergens over the dose ranges examined, whereas the potato proteins stimulated low-titre IgE antibody at the highest dose (10%) only. Importantly, differences in IgE antibody production have been observed against a background of equivalent immunogenicity (IgG antibody responses).

CONCLUSION

The data presented here suggest that the measurement of antibody (IgE) responses in BALB/c mice appears to identify allergens accurately and to distinguish them from those materials that apparently lack allergenicity.

Induction of IgE antibody responses by protein allergens: inter-laboratory comparisons

There is a growing interest in the development of methods for the evaluation of the allergenic potential of novel proteins. One approach is the measurement of specific IgE antibody production stimulated by systemic (intraperitoneal; i.p.) exposure of BALB/c strain mice. In the current investigations, inter-laboratory comparisons have been performed of IgE antibody production induced in mice by food proteins of differing sensitizing potential. Female BALB/c strain mice (n=5) were exposed to 0.1% peanut agglutinin, an allergenic constituent of peanuts, to 2% ovalbumin (OVA), a major allergenic constituent of hens' egg, or to a protein considered to lack significant allergenicity, potato agglutinin (5%). Specific IgE antibody was measured by homologous passive cutaneous anaphylaxis assay and IgG and IgG1 antibody production was analysed by enzyme-linked immunosorbent assay (ELISA). Two independent experiments were conducted in each laboratory, but with all serological analyses conducted in one of the laboratories. Each of the proteins induced vigorous IgG and IgG1 antibody responses, with no statistically significant differences in titres recorded between laboratories. Furthermore, OVA and potato agglutinin induced responses of equivalent immunogenicity with respect to both IgG and IgG1 antibody titres. Administration of peanut agglutinin and OVA each stimulated marked IgE antibody responses in every experiment. In the two laboratories, titres ranged from 1:32 and 1:64 for peanut agglutinin, and from 1:8 and 1:32 for OVA. In contrast, exposure to potato agglutinin failed to induce vigorous IgE production, with no detectable IgE (negative with neat serum), or titres of 1 (positive with neat serum only) recorded. These data demonstrate that the induction of IgE antibody by food proteins of differing allergenic potential is a relatively robust phenomenon and transferable between laboratories. Furthermore, these results provide additional evidence that the measurement of antibody (IgE) responses in BALB/c mice may allow discrimination between allergens and those materials that apparently lack allergenicity.

The major Platanus acerifolia pollen allergen Pla a 1 has sequence homology to invertase inhibitors

BACKGROUND

Sycamores or plane trees are an important source of airborne allergens in many cities of the United States and Western Europe. Pla a 1 has been described as a major allergen from Platanus acerifolia (London plane tree).

OBJECTIVE

To clone and characterize the cDNA for Pla a 1 and to express the recombinant protein.

METHODS

Pla a 1 was isolated by cationic exchange, gel filtration, and reverse-phase chromato-graphies. Pla a 1 cDNA was cloned by reverse transcription followed by polymerase chain reaction, using amino acid sequences from tryptic peptides of the allergen. The Pla a 1 encoding sequence has been subcloned into the pKN172 expression vector and expressed in Escherichia coli as a non-fusion protein. Purified recombinant protein has been tested for its IgE-binding capacity in immunoblot, immunoblot inhibition, and ELISA.

RESULTS

Pla a 1 reacted with serum IgE from 35 of the 42 (83.3%) Platanus-allergic patients studied and represented 60% of the total IgE-binding capacity of the P. acerifolia pollen extract. The allergen displayed 43% sequence identity to a grape invertase inhibitor and showed a predicted secondary structure characteristic of all-alpha proteins. Serological analysis revealed that both natural and recombinant forms of Pla a 1 displayed similar IgE-binding capacity.

CONCLUSIONS

Pla a 1 belongs to a new class of allergens related to proteinaceous invertase inhibitors. Recombinant Pla a 1 binds IgE in vitro like its natural counterpart and, therefore, it can be useful for specific diagnosis and structural studies.

Plant-based heterologous expression of Mal d 2, a thaumatin-like protein and allergen of apple (Malus domestica), and its characterization as an antifungal protein

Mal d 2 is a thaumatin-like protein and important allergen of apple fruits that is associated with IgE-mediated symptoms in apple allergic individuals. We obtained a full-length cDNA clone of Mal d 2 from RNA isolated from ripe apple (Malus domestica cv. Golden Delicious). The cDNA's open reading frame encodes a protein of 246 amino acid residues including a signal peptide of 24 residues and two putative glycosylation sites. The deduced amino acid sequence of the mature Mal d 2 protein results in a predicted molecular mass of 23,210.9Da and a calculated pI of 4.55. Sequence comparisons and molecular modeling place Mal d 2 among those pathogenesis-related thaumatin-like proteins that contain a conserved acidic cleft. In order to ensure the correct formation of the protein's eight conserved disulfide bridges we expressed Mal d 2 in Nicotiana benthamiana plants by the use of a tobacco mosaic viral vector. Transfected N.benthamiana plants accumulated Mal d 2 to levels of at least 2% of total soluble protein. MALDI-TOF mass spectrometric analyses of the recombinant Mal d 2 and its proteolytic fragments showed that the apple-specific leader peptide was correctly cleaved off by the host plant and that the mature recombinant protein was intact and not glycosylated. Purified recombinant Mal d 2 displayed the ability to bind IgE from apple-allergic individuals equivalent to natural Mal d 2. In addition, the recombinant thaumatin-like Mal d 2 exhibited antifungal activity against Fusarium oxysporum and Penicillium expansum, implying a function in plant defense against fungal pathogens.