Molecular and Functional Analysis of Soybean Allergen Proteins with a Focus on Pollen-Food Syndrome

Abstract

Gly m 3 and Gly m 4 are major soybean food allergens, causing birch pollen cross-allergic reactions, particularly with Bet v 1 and Bet v 2. These allergens can mediate anaphylactic reactions; however, the causative factors are still unknown. The goals of this comparative study are to characterize (A) the structural functionality of Gly m allergens of Glycine max and Bet v allergens of Betula pendula form birch, with a focus on their immunological properties, and (B) the molecular mechanisms of cross-allergenicity involved in pollen-food syndrome. This was achieved by extensive analysis using different molecular computer-aided approaches covering (1) physicochemical properties and functional-regulatory motifs, (2) sequence analysis, 2D and 3D structural homology modeling comparative study, (3) conservational and evolutionary analysis, (4) identification of B-cell epitopes based on sequence and structure-docking, while T-cell epitopes were identified by inhibitory concentration and binding score methods. Thus, we found that particular epitopes, in addition to the conserved ones, could be responsible for eliciting cross-reactivity between Bet v 1 and Bet v 2, and their respective homolog allergens proteins found in soybean. Moreover, variable epitopes were present in the Gly m 4 and Gly m 3 structures, which may be also responsible for this causative cross-allergenicity between soybean seed and birch pollen proteins.

1. Introduction

Soybean is a globally recognized dietary staple for its nutritional value, and its allergenic reactions have become a prevalent health concern [1]. The increasing incidence of soybean allergies raises significant worries due to the associated risks of severe allergic reactions, notably anaphylaxis, and the interplay with the pollen-food syndrome, also known as oral allergy syndrome (OAS) [2, 3]. This distinctive condition is characterized by allergic reactions triggered by the cross-reactivity of specific plant allergens, adding complexity to the landscape of soybean allergies [3, 4]. Essentially, individuals experiencing the pollen-food syndrome may develop allergic responses to certain foods, particularly fruits and vegetables, that share proteins with comparable epitopes with pollen allergen proteins [5, 6]. The immune system, sensitized to specific pollen proteins, may recognize similar proteins (epitopes) in certain food proteins, leading to allergic reactions. In the case of soybean allergies, the presence of cross-reactive plant allergen proteins further complicates the understanding and management of allergic responses, making it crucial for affected individuals to navigate their dietary choices with careful consideration [3, 5, 7].

As mentioned, these cross-reactions are attributed to specific protein sequences (epitopes) similarly found in soybeans, such as Gly m 3 (profilin) and Gly m 4 (bet v1-like) proteins [8, 9, 10], which are emerging major allergens. The clinical significance of these allergens is compounded by their propensity to induce birch pollen cross-allergic reactions, particularly with Bet v 1 (bet v1-like) and Bet v 2 (profilin) from Betula pendula (birch) [7, 10, 11]. Understanding the broader context of soybean allergies needs further exploration into the mentioned phenomenon of pollen-food syndrome, as the complex immunological interplay between inhaled pollen allergens and specific plant-derived foods is key to a better management of allergies coming from pollen-food syndrome [7, 12, 13]. This particular syndrome manifests when individuals sensitized to pollen allergen proteins, such as those from birch (Bet v 1 and Bet v 2), exhibit allergic symptoms upon consuming related foods containing soybean seed proteins. The intricate cross-reactivity between specific plant allergens, such as Gly m 3 and Gly m 4 in soybeans, and their counterparts in birch pollen, poses challenges in both clinical management and the formulation of dietary guidelines [2, 7, 9].

In this context, Gly m 3 and Gly m 4, as key soybean allergens, assume a pivotal role in the intricate web of cross-reactivity [7, 10]. While the clinical manifestations such as anaphylaxis due to these soybean allergen proteins are well-documented, the underlying molecular causative factors remain enigmatic. Moreover, the molecular mechanisms that govern the cross-allergenicity between soybeans and birch pollen represent an intriguing area of study. The counterparts in birch pollen allergen proteins, Bet v 1 and Bet v 2, add an additional layer of complexity to this phenomenon, so a comprehensive investigation into their structural and molecular implications is highly needed in this area of research [7, 14, 15, 16, 17].

This study aims to provide a thorough examination of soybean allergies within the broader framework of the pollen-food syndrome and the cross-reactivity of specific plant allergens. Our objectives encompass not only the molecular characterization of Gly m 3 and Gly m 4 but also a comparative analysis with Bet v 1 and Bet v 2 allergen proteins from Betula pendula, with a focus on its structural functionality and also immunological functions. Through advanced molecular computer-aided approaches including the analysis of physicochemical properties and functional-regulatory motifs, sequence analysis, 2D and 3D structural homology modeling, conservational and evolutionary analysis, and the identification of B-cell and T-cell epitopes through sequence-structure docking, we aim to unravel the physicochemical, structural, and immunological dimensions underlying the intricate interplay between soybean and birch pollen proteins. By doing so, we hope to contribute valuable insights that may inform not only our understanding of soybean allergies but also the broader field of pollen-related cross-reactivity reactions.

2. Material and methods

2.1 Bet v1-like (bet v 1) and profilins (bet v 2) sequence database search

Betula pendula (Bet v 1 y Bet v 2) sequences (in the case of Bet v 1, the references according to Uniprot are: O23750, Q0QKW8, Q0QLT4, Q0QLT5, Q0QLT9, Q0QLU7, Q9LEP0, Q9SYW2 and Q9ZS39; in Bet v 2, the references in Uniprot are as follows: A4K9Z8 and P25816) were used as queries to search for pollen and vegetative allergen from the different species of study, against publicly available sequence databases Swiss-Prot/TrEMBL (Uniprot) (https://www.uniprot.org/) and NCBI (https://www.ncbi.nlm.nih.gov/). Other databases were also consulted looking for other allergenic proteins, such as those from soybean allergens and other legume species. Those databases were: Allergome (https://www.allergome.org/); SDAP (http://https//fermi.utmb.edu/); Allergen online (http://www.allergenonline.com/); and FARRP databases (http://farrp.unl.edu/resources/farrp-databases), following the described methods in Refs. [18, 19].

2.2 Allergen proteins comparisons

In order to compare primary sequences of allergen proteins by using alignments, we use the amino acid sequences of (1) 9 Bet v 1 and 15 Gly m 4 sequences; as well as (2) 2 Bet v 2 and 10 Gly m 3 sequences. The alignment procedure was performed using ClustalW tools (https://www.genome.jp/tools-bin/clustalw). The data obtained from the alignments were analyzed to assess the homology of each Bet v protein allergen from Betula pendula with their homologous proteins Gly m from Glycine max.

These alignments were created using the Gonnet protein weight matrix, multiple alignment gap opening/extension penalties of 10/0.5 and pairwise gap opening/extension penalties of 10/0.1. The outputs were manually checked to optimize the alignment using Bioedit v7.2 (https://bioedit.software.informer.com/7.2/) [18, 19].

2.3 Structure analysis of bet v1-like and profilin proteins

With the objective of identifying the secondary structure (2D), we have taken a sequence of each allergen protein from Uniprot and/or NCBI databases to check the amino acids that participate in each secondary structural element. To determine the most important amino acids involved in functional and/or regulatory structural elements of the proteins, we have extracted information from different bibliographic sources, such as the location of those amino acids in the sequence and to which secondary structure it belongs [20, 21, 22].

2.4 Phylogenetic analysis of bet v1-like and profilin families of proteins

To study the phylogenetic relationship between Bet v1-like proteins and profilin protein families, respectively, we generated their phylogenetic trees. To that aim, we use sequences of both, Bet v1-like and profilin families of protein sequences from several legume and grass species. Finally, the phylogenetic trees, enriched with insights into the evolutionary divergence of Bet v1-like and profilin proteins, were visually presented using Treedyn (http://www.treedyn.org/). Treedyn serves as a visualization tool, offering a user-friendly interface for the exploration and interpretation of complex phylogenetic tree structures. The utilization of this software allowed for a comprehensive and intuitive representation of the evolutionary dynamics within the studied protein families across diverse legume and grass species. For further exploration, the generated phylogenetic trees can be accessed and analyzed through Treedyn. The branches were tested with 1000 bootstrap replicates [18, 19].

2.5 B and T-cells epitopes identification from bet v 1-like (bet v 1) and profilin (bet v 2) families of allergen proteins

Stabilization matrix alignment methods were used to identify peptide of Major Histocompatibility Complex (MHC) binding: NetMHCII.2.3 (https://services.healthtech.dtu.dk/services/NetMHCII-2.3/) to identify T-cell epitopes. This method employs binding status scoring qualitative prediction methods and also uses artificial neuron networks.

For the determination of linear (continuous) and conformational (discontinuous) B-cell epitopes, the different protein sequences of Bet v 1, Bet v 2, Gly m 4 and Gly m 3 were submitted to Ellipro (http://tools.iedb.org/ellipro/). This web tool implements the Thornton method, along with a residue clustering algorithm, the MODELLER program and the Jmol viewer, to predict and visualize antibody epitopes in a given protein sequence or structure. This program has an AUC value of 0.732 and its prediction range is in the top 3 for more than 70% of proteins. In our case, we entered the structural template number PDB in Ellipro, being able to change the threshold values in the parameters used for epitope prediction, i.e. the minimum residue score (protrusion index), indicated as S, between 0.5 and 1.0 and the maximum distance, indicated as R, in the range of 4– 8 Å [23].

2.6 Functional interaction meaning of allergen proteins

In order to determine interactions of allergen proteins and their functional meaning, we used the software STRING (https://string-db.org/cgi/input?sessionId=bcUhzvoZtC12&input_page_show_search=on), which determines these associations that are particularly important due to known specificity and adaptability of interactions. This database aims to integrate all known and predicted protein-protein associations, both physical and functional [24].

The network’s nodes depict different proteins, where each node represents proteins produced by a single protein-coding genetic locus. Colored nodes denote query proteins, serving as the initial layer of interactors. The content of a node indicates its representation: an empty node signifies proteins with unknown 3D structure, while filled nodes indicate proteins with either known or predicted 3D structure.

Regarding edges, they symbolize specific and meaningful protein-protein associations, reflecting proteins that jointly contribute to a shared function. This does not necessarily imply a physical connection. Pink edges denote experimentally determined interactions, blue edges represent interactions sourced from selected databases, black edges indicate interactions based on co-expression, and green edges signify interactions identified through text mining.

2.7 3D structure modeling of allergen proteins

Allergen proteins (Bet v 1, Bet v 2, Gly m 3 and Gly m 4) were searched for the protein or its homologs in PDB [25], using a BLAST search [26]. MODELLERR [27] is run to predict the protein 3D structure. The threshold values for BLAST e-value and a number of templates that MODELLER were used as input.

3. Results and discussion

3.1 Bet v 1-like and profilin proteins sequence analysis

The Gly m 4 allergen protein has been identified as a Bet v 1-like, also known as PR-10 (pathogenesis-related proteins). The function of these proteins is to respond to different stress situations such as relevant factors in the sensitization process [28]. Although the structures of some members of the family are known, this information is insufficient to know both their molecular action and their cellular function [29].

The comparison of the primary sequences of Gly m 4 and Bet v 1-like proteins was made using general amino acid alignment, as showed in Figure 1.

Figure 1.

Alignment of Bet v 1 sequences with Gly m 4. Glycines are represented with blue arrows and Tryptophan is represented with yellow arrows. Amino acids involved in RNase activity are represented with pink arrows. The figure shows the alpha-helix in the blue box and the beta strand in the green box. Bet v 1 sequences correspond to the specie Betula pendula; Gly m 4 sequences correspond to Glycine max.

Gly m 4 has an average sequence of 158 amino acids, and its localization in the cell has been determined to be in the cytoplasm as well as in the nucleus [30]. Regarding the molecular functions of this protein, it participates in: abscisic acid binding, a plant hormone that regulates aspects of plant growth [31]; protein phosphatase inhibitor activity, binds to and stops, prevents or reduces the activity of a protein phosphatase [32]; and signaling receptor activity, receiving a signal and transmitting it in the cell to initiate a change in cell activity. This signal is a physical entity or change in state that is used to transfer information in order to trigger a response [32]. We can then conclude that this protein participates in three well-established biological processes: (1) the abscisic acid-activated signaling pathway, involving molecular signals generated through the binding of the plant hormone abscisic acid (ABA) to a receptor, ultimately modulating cellular processes such as transcription [33]); (2) the defense response, characterized by reactions triggered in response to the presence of foreign bodies or injuries, aiming to restrict the damage to the attacked organism or facilitate prevention/recovery from the resulting infection (as mentioned earlier); and (3) the response to biotic stimuli, encompassing any cellular or organismal changes in terms of movement, secretion, enzyme production, gene expression, etc., induced by stimuli originating from living organisms [30]).

The screening of characteristic motifs and patterns indicated that Gly m 4 contains a domain integrated by a pattern corresponding to RNase activity, meaning it catalyzes the hydrolysis of RNA into smaller components, integrated by three amino acids a first Glu at position 96 (E96), a second Glu at position 148 (E148) and a Tyr at position 150 (Y150). This is represented in Figure 1, with pink arrows.

Searching in allergy databases like Allergome or Uniprot for Bet v 1-like allergens, both in pollen and seed proteins, we found sequences from Gly m 4 (Glycine max), Bet v 1 (Betula pendula), Lup an (Lupinus angustifolius) and some more species like Arachis hypogaea and Phaseolus vulgaris.

Sequence alignment comparison has been depicted in Figure 1, showing an identity range between 45 and 55% when comparing Gly m 4 sequences and selected Bet v 1 sequences such as BT092026. In Figure 1, we can notice that few amino acids are strictly conserved in Gly m 4: Gly and Trp, at the positions Gly2 (G2), Gly47 (G47), Gly49 (G49), Gly50 (G50), Gly52 (G52), Gly63 (G63), Trp83 (W83), Gly84 (G84), Gly92 (G92), Gly93 (G93), Gly95 (G95), Gly113 (G113), Gly116 (G116), Gly117 (G117), Gly129 (G129), Gly142 (G142), Gly146 (G146) and Gly155 (G155). Most of them were also conserved in Bet v 1 allergen protein. In addition, the aforementioned molecular pattern with RNase activity appear in this alignment, meaning it catalyzes the hydrolysis of RNA into smaller components [34].

Regarding the 2D structure, the allergen Bet v 1 exhibits four main structural elements in a single domain: two beta strands and two α-helix. The first alpha helix is located between amino acids P13 and V31; the second structure, being the first beta strand, is located between amino acids F36 and G86. To continue, we find the second beta strand lying between L93 and T120; and finally, the second alpha-helix oscillating between Q128 and L149. The figure shows in the blue box the alpha-helix and in the green box the beta strand.

These structures are specific to all Bet v 1-like or PR-10 proteins, determined by the AlphaFold program (https://alphafold.ebi.ac.uk/). Despite the relatively low ID between amino acids of the sequences, both types of proteins are conserved in the 2D structure, as previously described [34]. This study has analyzed the nucleotide sequence of soybean in depth, compared to other species such as Arabidpsis thaliana.

Hofmaier et al. [28] showed in their scientific paper, a similar identity of Gly m 4 with Bet v 1, about 48% in their sequences. In addition, a stress-induced SAM22 protein belonging to the Bet v 1 family, involved in the plant defense response, was identified as Gly m 4 protein. This is a major cause of soybean allergy in patients with birch pollinosis, which through a clinical investigation, where it was revealed that 96% of the patients reacted to Bet v 1, together with Gly m4 specific IgE and 64% of them recognized other soybean proteins [35]. The sequence homology between both proteins is very similar to the one we have found in our study; and the stress-induced SAM22 protein matches also with the results of our study, suggesting a potential existence of this functional activity in Gly m 4.

On the other hand, the allergen protein, Gly m 3, has been identified as a profilin [36]. The function of this protein is to bind actin, affecting the dynamic regulation of the structure of the cytoskeleton. At high concentrations, profilin prevents actin polymerization, while it enhances it at low concentrations [37].

Several species such as Olea europaea, Betula pendula, Phleum pratense and many others are known as important allergens [20]. Gly m 3 has an average sequence of 131 amino acids, and its localization in the cell has been determined to be in the cytoplasm as well as in the cytoskeleton, having a molecular function as actin monomer binding, PIP-lipids binding and proline-rich proteins regulator [37]. Thus, its main function in biological processes is the sequestering of actin monomers when acting in particular intracellular concentrations [37].

Figure 2 shows the alignment of the comparison between Bet v 2 and Gly m 3 proteins. The screening of characteristic motifs or patterns indicated that Gly m 3 presents three domains integrated by three patterns of regulatory function interacting with actin, poly-L-proline proteins and phosphoinositide lipids. The last cell signaling function was the less conserved domain. However, we can describe how changes in the residues within these interacting areas, particularly those interacting with actin, can increase or decrease its relative binding affinity with PIP [20]. In addition, Gly m 3 is located in the C-terminal region and the amino acids that are conserved for this interaction are Leu130 (L130) and Glu131 (E131) [20]. In the case of the interaction with actin, the N-terminal region is not conserved, with a large number of variable residues located near the actin-interacting site on the profilin surface. The amino acids involved in this interaction are His61 (H61) and Gln78 (Q78). In Figure 2, these amino acids are represented by pink arrows. Finally, the function affecting PLP proteins interaction, which are located in the C-terminal region correspond to proline-rich proteins, such interaction is regulated by phosphorylation on the residues Tyr75 (Y75) and Tyr108 (Y108) [20].

Figure 2.

Alignment of Bet v 2 sequences with Gly m 3. Serines are represented with blue arrows, threonines with yellow arrows, cysteines with red arrows and tyrosine with green arrows. Bet v 2 sequences correspond to the species Betula pendula; Gly m 3 sequences correspond to Glycine max.

The exploration of allergy databases for profilin allergens already identified as allergens, both in pollen and seed proteins, includes sequences from Gly m 3 (Glycine max), Bet v 2 (Betula pendula), Lup an (Lupinus angustifolius) and more species such as theses listed in the phylogenetic tree (Figure 3).

Figure 3.

Phylogenetic analysis of Bet v 2 and Gly m 3. Including analysis with other species: Betula pendula (A4K9Z8 and P25816), Glycine max (A7XZJ7, C6SVT2, I1J4G4, I1K601, I1K602, I1KJR5, I1KPV7, I1MCV8, K7KRR8, Q0PPS3, O65809 and O65810), Arachis hypogaea (D3K177), Ambrosia artemisiifolia (Q2KN23), Drosophila melanogaster (Q94JN2), Artemisia vulgaris (Q8H2C8 and Q8H2C9), Phleum pratense (A4KA31, P35079 and O24650), Lupinus albus (FG090100, FG090101 and Q96475), Phaseolus vulgaris (P49231), Olea europaea (A4GCR3, A4GCR5, A4GCR6, A4GCR7, A4GCR8, A4GCS1, A4GD50, A4GD52 and O24169), Hordeum vulgare (F2CT70, F2E5Q1 and P52184), Sorbus torminalis (C5XJ77, C5Z1D7 and C5Z4B6) and Cynodon dactylon (O04725).

Sequence alignment and identity comparison have been developed with Betula pendula, showing a good identity between Gly m3 sequences and some selected Bet v2 sequences. In Figure 2, we can observe that few amino acids are strictly conserved in Gly m 3: Serines, Cysteines, Threonines and Tyrosines, concretely Ser2 (S2), Thr5 (T5), Tyr6 (Y6), Cys13 (C13), Ser25 (S25), Ser33 (S33), Ser38 (S38), Ser39 (S39), Ser40 (S40), Thr50 (T50), Thr65 (T65), Tyr74 (Y74), Ser91 (S91), Thr95 (T95), Thr99 (T99), Tyr108 (Y108), Thr113 (T113), Cys117 (C117) and Tyr126 (Y126) well conserved in comparison with Bet v 2.

Regarding the 2D structural elements, we can appreciate 13 main structures in a single domain of profilins: seven beta strand, four α-helix and two turns. The first alpha helix is located between amino acids W3 and H10; the second alpha helix is located between amino acids D16 and G19 and the first beta strand is located between amino acids A23 and G29. We also found that the second beta strand is lying between V34 and Q37; the third alpha helix is located between amino acids P46 and E57; the first turn is located between amino acids P59 and T65; the third beta strand is located between amino acids L67 and L69; the fourth beta strand is located between amino acids I72 and E80; the second turn is located between amino acids A81 and A83; the fifth beta strand is located between amino acids V84 and K89; the sixth beta strand is located between amino acids G92 and K98, the seventh beta strand is located between amino acids G100 and Y108; and finally, the fourth alpha helix is oscillating between P114 and D130. The figure shows in blue box the alpha-helix, in green box the beta strand and in yellow box the turns [20].

In another paper [38], we can see an example in which two mango profiling proteins have been related to Bet v 2. The length of the protein is determined to be 131 amino acids. In addition, the two mango sequences studied show 88% homology, thus confirming their close relationship; together with the high identities with profilins from other plants, such as fruits (cherry, pear, peach and apple), seed (peanut), verdure (carrot) and pollen (Bet v 2 from Birch).

The comparison of our results with the allergen protein Act d 9, a protein related to Bet v 2 found in kiwi fruit [39], also showed similar amino acid sequence lengths, structural features, and biological functions when compared to the proteins include in our study.

3.2 Phylogenetic analysis of bet v1-like and profilin proteins

In order to analyze the sequence and functional relationships between different profilins from plant species, including those from allergenic proteins in pollen and seeds, we performed a phylogenetic analysis with the aim of identifying different functional groups.

Forty-four sequences of profilins from a wide species representation were aligned and the corresponding groups, or clusters were analyzed. The phylogenetic tree shows four main groups, as we can observe in Figure 3. The blue color cluster is the one with the largest number of Gly m 3 sequences analyzed and named cluster 1. This cluster is also composed of profilin sequences from Lupinus albus, Arachis hypogaea, Phaseolus vulgaris, Sorbus torminalis and Hordeum vulgare.

The following largest group is in red color, named group 2, which included Gly m 3 sequences, two Bet v 2 sequences, Arachis hypogaea and Olea europaea. Group 3, highlighted in green color, is the smallest one, containing sequences of Sorbus torminalis and Cynodon dactylon. Finally, the group with the remaining retrieved sequences corresponding to Hordeum vulgare, Phleum pratense and Zea mays, was called group 4. This analysis determines that Gly m 3 is closely related to profilin proteins from birch.

In order to analyze the relationships between Bet v 1-like proteins from different plant species, including those from already identified allergenic pollen, we performed a phylogenetic analysis to differentiate functional groups. We aligned 40 retrieved sequences from a broad range of species representation of bet v 1-like and analyzed the corresponding groups/clusters, 3 clusters. The red cluster is the one with the majority of Gly m 4 studied, cited as cluster 1. This cluster is also composed of selected bet v1 sequences and bet v1-like proteins from Lupinus albus, Arachis hypogaea and Phaseolus vulgaris. The next largest group is the blue group, called group 2, which includes species such as Paspalum vaginatum, Panicum virgatum, Vigna radiata var. radiata, Oryza sativa, and others. In group 3, green in color and the smallest, there is a theoretical study protein, like those mentioned above, from Vigna radiata var. radiata. This analysis determines that Gly m 4 is closely related to Bet v 1-like proteins (Figure 4).

Figure 4.

Phylogenetic analysis of Bet v 1 and Gly m 4. Including analysis with others species: Betula pendula (AJ00155F, DQ296592, DQ296598, DQ296594, DQ296593, DQ296584, AJ289771, AF124839 and AJ006909), Glycine max (X60043, BT091462, BT091912, BT092026, BT92600, BT092600, BT096508, CM000842, CM0000840 and X60044), Arachys hypogaea (EU514465, EU661964 and DQ813661), Lupinus albus (AJ000108 and AB070618), Phaseolus vulgaris (X61365, X61364 and DQ45598), Vigna radiata (A0A1S3THR8), Oryza sativa (Q7XBY6, Q6EN42, Q5Z8S0, Q6I5C3 and 039831506), Panicum halli (PUZ39982), Triticum turgidum (QKE11155), Paspalum vaginatum (KAJ1296150), Miscanthus lutarioriparius (CAD6204577) and Panicum miliaceum (RLN17232).

3.3 Functional interaction of allergen proteins

In this section, the aim was to study the complexity within cells that arises from functional and regulatory interactions between proteins. Our objective was to collect and integrate protein-protein interactions, both physical and functional, and achieved this using the STRING database [40].

According to Figure 5, we can determine the functional interaction of Gly m 3 with other different proteins.

Figure 5.

Functional interaction of Gly m 3 (Soybean). The query protein is GLYMA08G03650.1 (our protein); GLYMA11G12580.1 (adenylyl cyclase-associated protein), GLYMA12G04790.1 (adenylyl cyclase-associated protein), SAC1 (actin molecules are highly conserved proteins that are involved in various types of cell motility and are ubiquitously expressed in all eukaryotic cells), GLYMA13G41060.1 (uncharacterized protein), GLYMA11G14880.1 (uncharacterized protein), GLYMA08G15480.1, GLYMA02G10170.1 (uncharacterized protein), GLYMA12G06820.3 (uncharacterized protein), ACTIN (uncharacterized protein) and GLYMA18G52780.1 (Uncharacterized protein).

For instance, our Gly m 3, denoted as GLYMA08G03650.1, functions as a profilin binding to actin, influencing cytoskeletal structure. Profilin inhibits actin polymerization at high concentrations and enhances it at low concentrations.

The closest nodes to Gly m 3 are those with the highest homology, in this case, GLYMA11G12580.1 and GLYMA12G04790.1. They are proteins associated with adenylyl cyclases, a protein with known function with signal transduction belonging to the CAP family [41]. Both are first-layer interacting and query proteins. Moreover, their 3D structure is known. On the other hand, this interaction has been also determined through experimental analysis with the query protein, through selected databases (BIOGRID, https://thebiogrid.org/).

The remaining eight nodes are at the same distance, so they possess less homology with Gly m 3. Those nodes are: SAC1, GLYMA13G41060.1, GLYMA11G14880.1, GLYMA08G15480.1, GLYMA02G10170.1, GLYMA12G06820.3, ACTIN and GLYMA18G52780.1. They are uncharacterized proteins belonging to the actin family, except SAC1, which is an actin-1, highly conserved proteins that participate in various types of cell motility and are ubiquitously expressed in all eukaryotic cells. They are all query proteins and participants in the first layer; however, their 3D structures are not known [43]. The interactions with Gly m 3 have been experimentally determined, confirmed by x-ray crystallography assay or experimental interaction detection assay, and documented in databases as REACTOME (https://reactome.org/).

Following this analysis, we have analyzed the functional interaction of the allergen Gly m 4, showing its functional partners in Figure 6.

Figure 6.

Functional interaction of Gly m 4. This protein is soybean allergen, with string entry number C6T3A2. In this scheme we can see how the different bet v 1-like proteins from different species are quite related, depending on their sequence similarities [42]. The query protein Gly m 4 correspond to GLYMA17G03365.1 and the other proteins depicted in the schedule are named as follows: GLYMA17G03365.1, GLYMA04G14361.1, GLYMA01G32310.1, GLYMA20G26600.1, GLYMA19G09990.1, GLYMA08G25520.1, GLYMA15G27660.1, GLYMA20G19000.1, GLYMA19G09810.1, and P21e.

Figure 6 allows us to determine the protein-protein interactions experienced by our main protein (Gly m 4). Gly m 4, identified as GLYMA17G03365.1, is a Bet v 1-like protein. It is the first-layer interactor, with a known 3D structure. The closest nodes, exhibiting greater homology, include GLYMA17G15690.1, an expansion protein, which function is to loosen the cell wall structure so that it can expand and allow cell growth [44], GLYMA04G14361.1 is a chloroplastidic polyphenol oxidase, with the function of catalyzing the oxidation of polyphenols, a group of compounds found in the vacuoles of plant cells [45], GLYMA01G32310.1 (p7 peroxidase, with the function of catalyzing the oxidation of various compounds using hydrogen peroxide as a cofactor [46]), and GLYMA20G26600.1, a precursor protein with Bsp domain, which has a function of broad-spectrum antiviral activity [47]. The identified interaction of these proteins is based on sequence homology.

Nodes with less homology, encompass GLYMA19G09990.1 (germinal-like protein 20, protein with functions like spermatogenesis, oogenesis, fertilization, embryonic development and function in other tissues [48], GLYMA08G25520.1 (precursor of trypsin inhibitor family proteins), GLYMA15G27660.1 (alpha-amylase/subtilisin inhibitor, with functions in inhibit the activity of alpha-amylase and subtilisin) [49], GLYMA20G19000.1 (beta subunit of the voltage-dependent potassium channel, located in cytoplasm function in modulation of channel activation and inactivation, trafficking and assembly and interaction with other proteins) [50], GLYMA19G09810.1 (precursor of Cupin family proteins, a seed storage protein), and P21e (uncharacterized protein). Their nodes indicate their roles as query and first-interaction proteins, with known 3D structures, except for GLYMA19G09990.1.

The next protein to analyze their functional interacting partners is another Gly m 4 isoform from soybean.

Figure 7 allows us to determine the protein-protein interactions experienced by our main protein (Gly m 4). Gly m 4, identified as GLYMA07G37270.1, is a Bet v 1-like protein. Its node reveals its status as a query protein and first-layer interactor, with a known 3D structure. The closest nodes, exhibiting greater homology, include are GLYMA03G36230.1 (uncharacterized protein), GLYMA07G30540.2 (CRISP protein family), GLYMA13G01230.1 (protein precursor containing an extracellular protein domain similar to Scp) and GLYMA17G07350.2 (pathogenesis-related PR-1 protein with function is plant’s defenses system, belongs to the CRISP family) [52]. All are query proteins, first-layer interactors, and possess known 3D structures. Their interaction is based on sequence homology.

Figure 7.

Functional interaction of Gly m 4. This protein is soybean allergen, with string entry number I1L0V3. In this scheme we can see how the different bet v 1-like from different species are related, depending on their sequences similarity [51]. The query protein is GLYMA07G37270.1 and the other proteins are functionally related interacting partners to the query: GLYMA03G36230.1, GLYMA07G30540.2, GLYMA13G01230.1 and GLYMA17G07350.2.

Their functional interaction has been determined by co-expression in Oryza sativa and Zea Mays, as well as by sequence homology, except in GLYMA03G36230.1 which interaction is based only on its sequence homology.

Figure 8 allows us to determine the protein-protein interactions experienced by our main protein (Gly m 4).

Figure 8.

Functional interaction of Gly m 4, string entry number C6T1G1. In this scheme, we can see how the different bet v 1-like proteins of different species are related, depending on their sequence identity similarity) [53]. The query protein is LOC547915 and the other interacting proteins are functionally related to the query protein: HPS, LOC548076.

Gly m 4, identified as LOC547915, is Bet v 1-like protein characterized as a stress-induced SAM22 protein, involved in disease resistance [28]. It consists of 165 amino acids, in comparison to other proteins whose sequence has 180 amino acids.

Its node indicates its role as a query protein and first-layer interactor, with a known 3D structure. It has two closest proteins, HPS (hydrophobic seed) protein precursor, which belongs to the plant LTP family, whose function is related to plant defense against pathogens and environmental stresses [54]) and LOC548076 (2S albumin precursor, 2S seed storage protein). Their interaction with the query protein has been determined by sequence homology.

The remaining proteins described in Figure 8 are profiling proteins, which interact with actin molecules and regulate the structure and dynamics of the cytoskeleton [55].

3.4 T-cell epitopes identification and analysis of Betula and Glycine max allergen sequences

Allergy reactions are well characterized by the interaction of T-cell receptors with lineal epitopes from the allergen proteins [56]. In this study, we have identified and characterized the B- and T-cell epitopes (linear and conformational) in Bet v 1-like and profilin protein families. Table 1 describes the T-cell epitopes in Bet v 2 and Gly m 3 allergen proteins from the profiling family.

Allergen / accession	Epitope 1	Epitope 2	Epitope 3	Epitope 4
Bet v 2 A4K9Z8	VDEHLMCDIDGQGQQVDEHLMCDI	AQSSSFPQFKPQEITFPQFKPQEI	FEEPGHLAPTGLFEEPGHLAP
Bet v 2 P25816	DEHLMCDIDGQYVDEHMCD	AQSSSFPQFKPQEITFPQFKPQEI		GAVIRGKKGSGGITIVIRGKKGSG
Gly m 3 I1J4G4	DDHLMCDIQTYVDDHLM	AQSSSFPQWAQSSSFPQ-----------	FDEPGHLAPTGLHLGLAPTGLHLG	GAVIRGKKGS-G---VIRGKKGSG
Gly m 3 I1KJR5	DDHLMCDIDGTDHLMCDIDG	AQSSSFPQWAQSSSFPQ-----------	FDEPGHLAPTGLHLALAPTGLHLA	GAVIRGKK-GSG--VIRGKKGSG

Table 1.

T-cell epitopes of Bet v 2 and Gly m 3. Four T-cell epitopes were identified and colored as red (T-cell 1), blue (T-cell 2), pink (T-cell 3) and green (T-cell 4). The representative Bet v 2 sequences analyzed are A4K9Z8 and P25816; and for Gly m3 are 1J4G4 and I1KJR5.

T-cell epitopes play a central role in the cellular mechanisms underlying the pathophysiology of different types of pollen and food allergies. Among the sequences, these four epitopes have been identified and are present in most of the sequences analyzed. Four different amino acid sequences namely T1 to T4 were identified among the profilin proteins analyzed (see Table 1).

All four epitopes are present in all sequences except for Epitopes 3 and 4, which are absent in Bet v 2 sequences. Epitope 1 exhibits high polymorphism in the Gly m 3 (I1KJR5) sequence compared to other analyzed Gly m 3 and Bet v 2 sequences; similar comparative polymorphism is observed in Epitope 3. Epitopes 2 and 4 show low polymorphism in both species.

Table 2 describes five different amino acid sequences of allergen proteins, namely T1 to T5 were identified among the Bet v 1-like proteins analyzed (see Table 2). These five epitopes have been identified and are present only in one of the Bet v 1-like sequences. Most of them are missing in the rest of the sequences. This is an indicator of the variability of allergen sequences content in this family of protein allergens.

Allergen / accession	Epitope 1	Epitope 2	Epitope 3	Epitope 4	Epitope 5
Bet v 1 O23750	GNGGPGTIKKISFPEGGPGTIKKI	IEGGPVGDTLEKISNPVGD TLEKI	LKINNKYHTKGDHEVYHTKG DHEV	KVAPQAISSVENI EGISS VENIEG	ISSVENIEGNGGPGTIEGNGGPGT
Bet v 1 Q9ZS39	EGNGGPGTIKKISFPGGPGTIKKI	IEGGPVGDTLEKISNPVGD TLEKI			VENIEGNGGPGTIKKIEGNGGPGT
Gly m 4 C6T3A2	EGNGGPGTIKKITFVPGTIKKITF				EGNGGPGTIKKITFVPGTIKKITF
Gly m 4 I1L0V3	IVEGNGGPGTIKKLT VEGNGGPGT				IVEGNGGPGTIKKLTVEGNGGPGT

Table 2.

T-cell epitopes of Bet v 1 and Gly m 4. Five T-cell epitopes were identified and colored as red (T-cell 1), blue (T-cell 2), pink (T-cell 3), orange (T-cell 4) and green (T-cell 5). The Bet v 1-like sequences references (O23750 and Q9ZS39) and Gly m4 sequences references (C6T3A2 and I1L0V3).

3.5 Lineal and discontinuous B-cell epitopes analysis in Betula and Glycine species

In this study, we have identified and characterized the lineal epitopes in Bet v 1-like and profilin protein families that interact with B-cell receptors.

Table 1 describes the lineal B-cell epitopes in Bet v 1, Bet v 2, Gly m 4 and Gly m 3 allergen proteins from the profilin family. The accuracy parameter (threshold) for this identification was about 0.6 or higher.

We have identified up to 12 potential B-cell epitopes in three sequences. They do not overlap when comparing these in Bet v 2 and Gly m 3. These epitopes are absent in Gly m 4.

Discontinuous B-cell epitopes are determinant for the induction of an efficient immune response. We have analyzed the sequence and 3D localization of B-cell discontinuous epitopes in Bet v 1-like and profilin family of proteins.

As we can see in Table 3, the epitopes of Gly m 3 (5–7) and Bet v 2 (1–4) do not overlap in sequence. These are the same circumstances happening for Gly m 4 and Bet v 1 like proteins.

Allergen (references number)	Epitopes name	Sequences
Bet v2 (A4K9Z8)	1	EEPGHLAPTG
	2	CDIDGQGQQL
	3	EEPVTPGQ
	4	PQFKPQ
Gly m3 (A7XZJ7)	5	AESSQLQH
	6	KEDYDIEVEDENGTKTTKT
	7	QGEESDPNFD
Bet v1 (O23750)	8	HTKGDHEVKAE
	9	SYLLAHSDAYN
	10	VATPDGGSI
	11	FPEGLPFKY
	12	IGDTLE
Gly m4 (C6T1G1)	13	TKGDAEPNQDELK
	14	LAHPDYN
	15	LVAGPNGG
	16	LDSFKSVENVEGNGGPGT
	17	DEANL

Table 3.

Lineal B-cell epitopes of Bet v 1, Gly m 4, Bet v2 and Gly m3.

Localization of conformational epitopes has been made by building the 3D structure of different allergen proteins from the Bet v 1 and profilin families. Partial overlapping between particular epitopes for each homologous allergen has been shown for this type of proteins.

As we can see in Figures 9 and 10, the conformational epitopes of Bet v 2 and Gly m 3 are compared, respectively. We can see how in epitope 5 of Bet v 2 and epitope 8 of Gly m 3, they have shared regions that could give cross-reactivity in organisms. There are also overlaps in epitope 2 of Bet v 2 and epitope 10 of Gly m 3; and, in epitope 4 of Bet v 2 and epitope 11 of Gly m 3 (Figures 11 and 12).

Figure 9.

Localization of the different discontinuous epitopes of Bet v 2 allergen. The figure shows in blue color the epitopes identified in Table 4.

Figure 10.

Localization of the different discontinuous epitopes of Gly m 3 allergen. The figure shows in blue color the epitopes identified in Table 4.

Figure 11.

Localization of the different discontinuous epitopes of Bet v 1 allergen. The figure shows in blue color the epitopes identified in Table 4.

Figure 12.

Localization of the different discontinuous epitopes of Gly m 4 allergen. The figure shows in blue color the epitopes identified in Table 4.

Allergen / accession number	Epitopes	Amino acids that integrate the conformational epitopes
Bet v2; (A4K9Z8)	1	S2, W3, Q4, T5
	2	Y6, E9, H10
	3	K45, P46, Q47, T50
	4	K54, E57, E58, P59, G60, H61, L62, A63, P64, T65, G66, K73
	5	H30, T99, G100, Q101, D126, Y127, I129, D130, Q131, G132, L133
	6	C13, D14, I15, D16, G17, Q18, G19, Q20, Q21, L22, K89, G90, S91, E109, E110, P111, V112, T113, G115, Q116, M119
Gly m3; (A7XZJ7)	7	F112, E113, G114, Y115, K164, A167, E168, S169, S170, Q171, L172, H174
	8	K56, E57, D58, Y59, D60, I61, E62, V63, E64, D65, E66, N67, G68, T69, K70, T71, T72, K73, T74, N86, E87, G88, Y89, A90, P91, D92
	9	S22, Q40, G41, E42, E43, S44, D45, P46, N47, F48, D49
	10	N105, E107, D109, L110, E111, N116
	11	S6, D8, S9, N12, D13, R14
Bet v1; (O23750)	12	G1, V2, F3, I91, G92, D93, T94, E96, H121, T122, K123, G124, D125, H126
	13	E127, V128, K129, A130, E131, K134
	14	E8, I13, P14, A16, R17, V105, A106, T107, P108, D109, G110, G111, I113, S149, Y150, L151, L152, A153, H154, S155, D156, A157, Y158, N159
	15	D75, H76, T77, N78, F79
	16	E42, N43, I44, E45, G46, N47, G48, G49, P50, G51, T52, I53, K55, K65, Y66, E87
Gly m4; (C6T1G1)	17	L152, A153, _:H154, _:P155, _:D156, _:Y157
	18	L35, D36, _S37, _F38, _L59, _E60, _D61, _G62, _E63, _T64, _K65
	19	G2, V3, F4, P92, D93, T94, A95, E96, K97, T122, K123, G124, D125, A126, E127, P128, N129, Q130, D131, E132, L133, K134
	20	V14, A15, L104, V105, A106, G107, P108, N109, G110, G111
	21	E42, N43, V44, E45, G46, N47, G48, G49, P50, G51, T52

Table 4.

Conformational/discontinuous B-cell epitopes of Bet v 1, Gly m 4, Bet v2 and Gly m3.

4. Conclusions

In this study, we present new insights into the structural and functional properties of the soybean allergen Gly m 3, Gly m 4 and their counterparts. Comparison between Gly m 4 and Bet v 1 proteins shows a low degree of conservation, however, individual protein isoforms within each allergen show higher conservation when analyzed separately. In contrast, Gly m 3 and Bet v 2 sequences exhibit greater conservation when compared. Functional and structural analyses revealed that Bet v 1 and Gly m 4 cluster closely in the phylogenetic analysis like Bet v 2 and Gly m 3. Their 2D structural elements are well conserved, and the overall 3D structure is also well conserved between protein isoforms of the same family such as Bet v 1 and Gly m 4, and Bet v 2 and Gly m 3 (profilin).

Analysis of linear and conformational epitopes across the four allergens indicates that T-cell epitopes are predominantly shared within members of the same family, particularly between Bet v 2 and Gly m 3 in terms of sequence and number of epitopes. Differences are more pronounced between Bet v 1 and Gly m 4 in the number of shared epitopes among protein isoforms, though sequence differences are minimal. Additionally, some conformational epitopes are shared among allergen isoforms within the same family. Therefore, cross-reactive epitopes among homologous allergens result from both 3D structural similarity and primary sequence similarity.

Acknowledgments

JCJ-L would like to acknowledge the funding received by the Spanish Ministry of Economy, Industry and Competitiveness (Ramon y Cajal Research Program), grant ref. RYC-2014-16536; the CSIC intramural research program, grant ref. 202240I002; the Spanish Ministry of Science and Innovation, grant ref. number CPP2021-008989.

Conflict of interest

The authors declare no conflict of interest.