Impact of food processing on the structural and allergenic properties of food allergens.

This article reviews recent studies that address one of the major unanswered questions in food allergy research: what attributes of food or food proteins contribute to or enhance food allergenicity?

Evolutionary distance from human homologs reflects allergenicity of animal food proteins.

BACKGROUND: In silico analysis of allergens can identify putative relationships among protein sequence, structure, and allergenic properties. Such systematic analysis reveals that most plant food allergens belong to a restricted number of protein superfamilies, with pollen allergens behaving similarly. OBJECTIVE: We have investigated the structural relationships of animal food allergens and their evolutionary relatedness to human homologs to define how closely a protein must resemble a human counterpart to lose its allergenic potential. METHODS: Profile-based sequence homology methods were used to classify animal food allergens into Pfam families, and in silico analyses of their evolutionary and structural relationships were performed. RESULTS: Animal food allergens could be classified into 3 main families--tropomyosins, EF-hand proteins, and caseins--along with 14 minor families each composed of 1 to 3 allergens. The evolutionary relationships of each of these allergen superfamilies showed that in general, proteins with a sequence identity to a human homolog above approximately 62% were rarely allergenic. Single substitutions in otherwise highly conserved regions containing IgE epitopes in EF-hand parvalbumins may modulate allergenicity. CONCLUSION: These data support the premise that certain protein structures are more allergenic than others. Contrasting with plant food allergens, animal allergens, such as the highly conserved tropomyosins, challenge the capability of the human immune system to discriminate between foreign and self-proteins. Such immune responses run close to becoming autoimmune responses. CLINICAL IMPLICATIONS: Exploiting the closeness between animal allergens and their human homologs in the development of recombinant allergens for immunotherapy will need to consider the potential for developing unanticipated autoimmune responses.

Post-translational modification of barley LTP1b: the lipid adduct lies in the hydrophobic cavity and alters the protein dynamics.

NMR techniques have been used to characterise the effects of a lipid-like post-translational modification on barley lipid transfer protein (LTP1b). NMR chemical shift data indicate that the lipid-like molecule lies in the hydrophobic cavity of LTP1b, with Tyr 79 being displaced to accommodate the ligand in the cavity. The modified protein has a reduced level of backbone amide hydrogen exchange protection, presumably reflecting increased dynamics in the protein. This may result from a loosening of the protein structure and may explain the enhanced surface properties observed for LTP1b.

Heat treatment of bovine alpha-lactalbumin results in partially folded, disulfide bond shuffled states with enhanced surface activity.

Prolonged heating of holo bovine alpha-lactalbumin (BLA) at 80 degrees C in pH 7 phosphate buffer in the absence of a thiol initiator improves the surface activity of the protein at the air:water interface, as determined by surface tension measurements. Samples after 30, 60, and 120 min of heating were analyzed on cooling to room temperature. Size-exclusion chromatography shows sample heterogeneity that increases with the length of heating. After 120 min of heating monomeric, dimeric, and oligomeric forms of BLA are present, with aggregates formed from disulfide bond linked hydrolyzed protein fragments. NMR characterization at pH 7 in the presence of Ca2+ of the monomer species isolated from the sample heated for 120 min showed that it consisted of a mixture of refolded native protein and partially folded protein and that the partially folded protein species had spectral characteristics similar to those of the pH 2 molten globule state of the protein. Circular dichroism spectroscopy showed that the non-native species had approximately 40% of the alpha-helical content of the native state, but lacked persistent tertiary interactions. Proteomic analysis using thermolysin digestion of three predominant non-native monomeric forms isolated by high-pressure liquid chromatography indicated the presence of disulfide shuffled isomers, containing the non-native 61-73 disulfide bond. These partially folded, disulfide shuffled species are largely responsible for the pronounced improvement in surface activity of the protein on heating.

Allergen sequence databases.

A number of specialized databases have been developed to facilitate studies of human allergens. These include molecular databases focused on protein sequences and structures, informational databases focused on clinical, biochemical and epidemiological data related to protein allergens, a database on allergen nomenclature, and other knowledge bases or informational websites that are peripherally-related to research on allergens. Examples of each type of databases are listed and described briefly in this review. Database construction and maintenance and their impact on database quality and usefulness are also discussed.

Food allergy information resources for consumers, industry and regulators.

Through the EU-funded InformAll project a stakeholder consultation was conducted to solicit the views of different stakeholders (allergic consumers, health professionals, retailers, manufacturers, caterers, regulators and risk assessors, and the general public) regarding what information about food allergy is required by these stakeholders. The outcomes of this consultation are presented both generally and specifically from the potential perspectives of different stakeholders. The development of reliable credible resources aimed to begin to meet the needs identified is described, including a database of allergenic food materials which uniquely combines refereed information on the clinical aspects of food allergies with details of individual allergens and web portal with other credible internet resources.

The identification of foam-forming soluble proteins from wheat (Triticum aestivum) dough.

Proteomic methods have been used to identify foam-forming soluble proteins from dough that may play an important role in stabilising gas bubbles in dough, and hence influence the crumb structure of bread. Proteins from a soluble fraction of dough (dough liquor) or dough liquor foam have been separated by two-dimensional gel electrophoresis, and 42 identified using a combination of matrix-assisted laser desorption/ionization-time of flight and quadrupole-time of flight analyses. Major polypeptide components included beta-amylase, tritin and serpins, with members of the alpha-amylase/trypsin inhibitor family being particularly abundant. Neither prolamin seed storage proteins nor the surface-active protein puroindoline were found. Commonly used dough ingredients (NaCl, Na L-ascorbate) had only a minor effect on the 2-DE protein profiles of dough liquor, of which one of the more significant was the loss of 9 kDa nonspecific lipid transfer protein. Many proteins were lost in dough liquor foam, particularly tritin, whilst a number of alpha-amylase inhibitors were more dominant, suggesting that these are amongst the most strongly surface-active proteins in dough liquor. Such proteins may play a role determining the ability of the aqueous phase of doughs, as represented by dough liquor, to form an elastic interface lining the bubbles, and hence maintain their integrity during dough proving.

Structural relatedness of plant food allergens with specific reference to cross-reactive allergens: an in silico analysis.

BACKGROUND: The body of sequence and structural information on allergens and the sequence analysis of whole plant genomes are facilitating the application of bioinformatic approaches to identifying and defining plant allergens. OBJECTIVE: An in silico approach was used to quantify the distribution of plant food allergen sequences across protein families and to develop and apply a novel means of assessing conserved surface features important for IgE cross-reactivity. METHODS: Plant food allergen sequences were classified into Pfam families on the basis of sequence homology. Contact surface areas of selected proteins were calculated with MOLMOL by using a 1.4-A probe, corrected by removing contributions from IgE inaccessible main chains and side chains forming the ligand binding sites. RESULTS: A set of 129 food allergen sequences were classified into only 20 of 3849 possible Pfam families, with 4 families accounting for more than 65% of food allergens. Structural bioinformatic analysis of conserved exterior main chains and amino acid side chains in cross-reactive homologues of Bet v 1 and nonspecific lipid transfer proteins showed higher levels of similarity than shown by simple sequence comparisons. Thus, 75% of the Mal d 1 surface is likely to bind anti-Bet v 1 antibodies, compared with a sequence identity of approximately 56%. CONCLUSION: Most plant food allergens belong to only 4 structural families, indicating that conserved structures and biological activities may play a role in determining or promoting allergenic properties. Structural bioinformatic analysis shows that conservation of 3-dimensional structure should be included in any assessment of potential IgE cross-reactivity in, for example, novel proteins.

Structural, biological, and evolutionary relationships of plant food allergens sensitizing via the gastrointestinal tract.

The recently completed genome sequence of the model plant species Arabidopsis has been estimated to encode over 25,000 proteins, which, on the basis of their function, can be classified into structural and metabolic (the vast majority of plant proteins), protective proteins, which defend a plant against invasion by pathogens or feeding by pests, and storage proteins, which proved a nutrient store to support germination in seeds. It is now clear that almost all plant food allergens are either protective or storage proteins. It is also becoming evident that those proteins that trigger the development of an allergic response through the gastrointestinal tract belong primarily to two large protein superfamilies: (1) The cereal prolamin superfamily, comprising three major groups of plant food allergens, the 2S albumins, lipid transfer proteins, and cereal alpha-amylase/trypsin inhibitors, which have related structures, and are stable to thermal processing and proteolysis. They include major allergens from Brazil nut, peanuts, fruits, such as peaches, and cereals, such as rice and wheat; (2) The cupin superfamily, comprising the major globulin storage proteins from a number of plant species. The globulins have been found to be allergens in plant foods, such as peanuts, soya bean, and walnut; (3) The cyteine protease C1 family, comprising the papain-like proteases from microbes, plants, and animals. This family contains two notable allergens that sensitize via the GI tract, namely actinidin from kiwi fruit and the soybean allergen, Gly m Bd 30k/P34. This study describes the properties, structures, and evolutionary relationships of these protein families, the allergens that belong to them, and discusses them in relation to the role protein structure may play in determining protein allergenicity.

Mass spectrometry and structural characterization of 2S albumin isoforms from Brazil nuts (Bertholletia excelsa).

Proteomic approaches have been used to characterise the main 2S albumin isoforms from Brazil nuts (Bertholletia excelsa). Whilst most isoforms ( approximately 10 discrete protein species) exhibited molecular masses of around 12 kDa with a high amino acid sequence homology, important charge heterogeneity was found, with pIs varying between 4.6 and 6.6, with one >or=7.0. Proteomic analysis showed that these corresponded to a total of six National Center for Biotechnology Information (NCBI) accessions and that three isoforms had been purified to homogeneity corresponding to gi/384327, 112754 and 99609. The latter sequence corresponds to an isoform, previously only identified at the nucleotide sequence level, had a slightly higher molecular weight (13.4 kDa), and with noticeable differences in the primary structure. Proteins corresponding to six different NCBI accessions were identified, the heterogeneity of which had been increased by posttranslational processing. Evidence was found of cyclization of the N-terminal glutamine residue in two isoforms, together with ragged C-termini, indicative of carboxypeptidase activity within the vacuole following posttranslational processing. No evidence of glycosylation was found. Circular dichroism (CD) and Fourier transform-infrared (FT-IR) spectroscopy indicated all the studied isoforms were predominantly alpha-helical in nature, but that the Mr 13400 species was structurally distinct, with a higher proportion of alpha-helical structure.

Thermally induced structural changes in glycinin, the 11S globulin of soya bean (Glycine max)--an in situ spectroscopic study.

The thermal denaturation behaviour of glycinin solutions has been studied in situ as a function of ionic strength using various spectroscopic methods. Changes in secondary structure occurred at temperatures above 60 degrees C, well before the onset of gelation. Even after heating to 95 degrees C, much of the native beta-sheet structure of glycinin was retained, as indicated by the amide I peak maximum at 1635 cm(-1) in the Fourier transformed infrared (FT-IR) spectrum. This was accompanied by an increase in the 1625 cm(-1) band, indicative of the formation of intermolecular beta-sheet associated with protein aggregation. Nuclear magnetic resonance (NMR) spectroscopy confirmed the presence of highly mobile regions in glycinin comprising predominantly of Gln and Glu residues, corresponding to mobile regions previously identified by crystallographic studies. There was also evidence of a hydrogen-bonded structure within this mobile region, which may correspond to an alpha-helical region from Pro(256) to (or just before) Pro(269) in proglycinin. This structure disappeared at 95 degrees C, when heat-set gel formation occurred, as indicated by a sudden broadening and weakening of the NMR signal. Otherwise the NMR spectrum changed little during heating, emphasising the remarkable thermal stability of glycinin. It is proposed that during heating the core beta-barrel structure remains intact, but that the interface between the beta-domains melts, revealing hydrophobic faces which may then form new structures in a gel-network. As Cys(45), which forms the disulfide with Cys(12) linking the acidic and basic polypeptides, is found in this interface, such a rearrangement of the individual beta-domains could be accompanied by cleavage of this disulfide bond, as is observed experimentally. Such information contributes to our understanding the aggregative behaviour of proteins, and hence develops knowledge-based strategies for controlling and manipulating it.