Glossary

• Allergen: A molecular component with immunogenic potential. Derived from natural kingdoms, they belong to specific allergenic protein families based on their amino acid sequence and structural similarity.

• Allergy: A clinical manifestation triggered by contact with an allergen capable of inducing an inflammatory immune response.

• A-RISC (Allergens’ Relative Identity, Similarity, and Cross-reactivity): An index calculated based on homology, similarity, and identity between two or more allergens. It estimates the probability of cross-reactivity among members of the same family.

• Epitope: The antigenic determinant of a macromolecule recognized by specific immunoglobulins (e.g., IgE).

• Homology: The evolutionary relationship between two sequences. Homologous proteins are commonly recognized by specific IgE. Homology is calculated using the A-RISC index.

• Identity: The degree of exact amino acid match between two different protein sequences.

• Similarity: The percentage of residues with similar physicochemical properties (size, charge, hydrophobicity).

• Cross-reactivity: An immune response induced by an allergen different from the primary sensitizer. Sequence identity ≥70% is commonly indicative of potential cross-reactivity.

Introduction

Food allergies represent a growing public health concern, affecting a significant proportion of the global population, with an estimated prevalence of 6–8% in children and 3–4% in adults, depending on age and geographic region.¹ Within this group, legumes are a frequent cause of allergic sensitization, with soy and peanut being the primary culprits.² In Latin America, an increasing prevalence of food allergies has been reported, with higher sensitization rates to legumes such as lentils, beans, and chickpeas, especially in pediatric populations.³

Besides being a vital source of plant-based proteins, legumes contain allergenic proteins with high structural similarity, which promotes cross-reactivity among species.⁴ These proteins are classified into various superfamilies, including 7S and 11S globulins, 2S albumins, oleosins, and pathogenesis-related proteins (PR-10), which are known for their stability and digestive resistance, characteristics that contribute to their clinical allergenic potential.⁵

Accurate food labeling plays a critical role in preventing adverse reactions among individuals with food allergies. International regulatory agencies, such as the European Food Safety Authority (EFSA) and the US Food and Drug Administration (FDA), mandate the explicit declaration of allergenic ingredients in processed food products.⁶ However, the implementation and enforcement of these labeling standards remain inconsistent, particularly in plant-derived foods and within developing countries, where regulatory oversight and infrastructure may be limited.

Molecular diagnostic strategies have become increasingly important in food allergy characterization. Tools such as the Molecular Allergology User’s Guide (MAUG 2.0)⁷ have improved the identification of specific proteins using microarrays and peptide sequencing, enabling the differentiation between primary sensitization and cross-reactivity.

Considering the importance of legume-related allergies and the need to enhance both diagnostic and labeling strategies, this study employs in silico tools to analyze the sequence identity of allergenic proteins in legumes. The results are expected to contribute to a deeper understanding of cross-reactivity and provide critical insights for clinical management of allergic patients, as well as inform preventive strategies in the food industry.⁸

The A-RISC index is a valuable method for estimating the probability of cross-reactivity between two allergens belonging to the same protein family.

A similarity matrix was constructed using the A-RISC (Allergens’– Relative Identity, Similarity, and Cross-reactivity) index, which reflects structural homology between pairs of proteins. The score ranges from 0 to 1 and is interpreted as follows:

• 0.70–1.00: High risk of IgE cross-reactivity

• 0.50–0.69: Moderate risk of IgE cross-reactivity

• < 0.50: Low risk of IgE cross-reactivity.

Higher scores indicate an increased likelihood of immunological recognition between allergens, with implications for component-resolved diagnosis (CDR) and immunotherapy strategies.

Table 1 Heatmap of A-RISC index among legume allergenic proteins.

According to the index developed by Chruszcz et al. (2018), the A-RISC combines amino acid identity and similarity to estimate the probability that an IgE antibody will fail to distinguish between two allergens from the same protein family. It is based on a weighted average of:

• % Identity: The proportion of exact amino acid matches at aligned positions.

• % Similarity: The proportion of chemically similar amino acids, calculated using the BLOSUM62 substitution matrix.

The resulting A-RISC score ranges from 0 to 1, where:

• 0.00 indicates no structural relationship.

• 1.00 corresponds to identical sequences.

Methods

Allergenic protein sequences were retrieved from specialized databases including the WHO/IUIS Allergen Nomenclature Home Page, AllFam, NCBI, and UniProt. Only full-length, mature protein sequences were considered; signal peptides and propeptides were excluded. For officially recognized allergens, standardized nomenclature was used, including four digits after the period (e.g., Ara h 2.0101). In specific cases where the crystallizable structure of the primary isoform was unavailable, closely related -isoforms (e.g., 0.0101) were accepted.

The main allergens analyzed in this study included:

• Soybean (Glycine max)

• Lupin (Lupinus spp.)

• Pea (Pisum sativum)

• Peanut (Arachis hypogaea)

• Lentil (Lens culinaris)

• Chickpea (Cicer arietinum)

• Common bean (Phaseolus vulgaris)

For the in silico analysis, multiple sequence alignments were performed using Clustal Omega (https://www.ebi.ac.uk/jdispatcher/msa) to identify significant alignment regions across allergenic proteins. Subsequently, SIAS (http://imed.med.ucm.es/Tools/sias.html) was employed with the Blosum62 substitution matrix to calculate pairwise identity and similarity percentages. The resulting similarity matrices were comparable to those used in previous allergen cross-reactivity studies involving aeroallergens, fruits, nuts, shellfish, and fish.

To assess structural homology and potential cross-reactivity risk, A-RISC indices were computed, and corresponding heatmaps were generated for visual representation.^9,10

Additionally, family-specific multiple sequence alignments were visualized using Jalview, applying Blosum62-based coloring and a 90% conservation threshold, to highlight structurally conserved regions or domains.

Results

Protein family classification

Only protein families with at least three allergenic sequences derived from different food sources were included in the analysis.

Families such as profilins, oleosins, defensins, and cyclophilins were excluded because representative allergens were identified in only one or two plant species, limiting the possibility of performing a meaningful comparative analysis of cross-reactivity. The FASTA sequences used for each allergen are available in Appendix 1.

Multiple sequence alignment, homology (A-RISC), and conserved motifs

The percentage identity and similarity calculations performed using SIAS for each protein family are presented in Appendix 2, while the visualization of conserved motifs generated with Jalview is included in Appendix 3.

Bet V 1 – Pathogenesis-related proteins PR-10

Classification: High risk of cross-reactivity.

PR-10 proteins exhibit sequence identity ranging from 67 to 72%, with similarity values close to 75%. These high values elevate the A-RISC index above the 0.70 threshold, suggesting a clinically significant risk of IgE-mediated cross-reactivity between Ara h 8, Gly m 4, and Cic a 4. These proteins are thermolabile and susceptible to digestive degradation, typically inducing mild symptoms such as oral allergy syndrome (OAS). Sensitization often originates from pollen exposure, and foods containing PR-10 proteins may trigger cross-reactive responses.

Several highly conserved regions were identified, such as “VEGNQGPGTIKKL” and “IDEALNGYMVSVVGGAALP” (residues ~40–90), which coincide with functional structural motifs, including a hydrophobic cavity involved in ligand transport and IgE recognition.

NsLTP: Nonspecific lipid transfer proteins

Classification: Low to moderate risk of cross-reactivity among legumes.

The A-RISC index analysis for nsLTPs reveals notable heterogeneity in sequence homology. While some protein pairs showed intermediate values (≥0.45), most comparisons fell below the 0.75 threshold, suggesting a low to moderate risk. Specific pairs, such as Len c 3–Pis s 3 or Cic a 3–Pis s 3, demonstrated moderate homology and potential for partial cross-reactivity. Ara h 9 showed low identity with other legume nsLTPs.

Figure 4. Conserved motif among NsLTP (Len c 3) (Jalview in Appendix 2).

Figure 1 A-RISC among PR-10 protein sequences. A-RISC (Sequence homology index): A-RISC values were calculated to assess the degree of homology among allergens within the PR-10 family.

Figure 2 Conserved motif among PR-10 (Ara h 8). (A) Cartoon representation of the Ara h 8 protein, a member of the PR-10 family, highlighting in blue the conserved motif “VEGNQGPGTIKKL,” located on one of the β-sheets that form part of the characteristic central hydrophobic cavity of the Bet v 1-like fold. (B) Surface representation of the same structure, where the VEGNQGPGTIKKL motif is clearly exposed, reinforcing its role as an immunologically relevant epitope. The structures were generated using UCSF ChimeraX (https://www.rbvi.ucsf.edu/chimerax/) based on the structural model of Ara h 8 (Jalview in Appendix 2).

Figure 3 A-RISC among NsLTP protein sequences. A-RISC (Sequence homology index): A-RISC values were calculated to assess the degree of homology among allergens within the nsLTP family.

Figure 4 Conserved motif among NsLTP (Len c 3). (A) Surface representation of the Len c 3 protein (a lentil nsLTP), highlighting in violet the conserved motifs PC, CC, and LK, which are associated with structurally and immunologically relevant regions. (B) Cartoon representation of the same model, showing the spatial localization of these motifs within the compact fold characteristic of nsLTPs. These motifs correspond to surface-exposed or partially accessible residues that may contribute to structural stability, lipid transport, or immunological reactivity. The structures were generated using UCSF ChimeraX (https://www.rbvi.ucsf.edu/chimerax/) (Jalview in Appendix 2).

Figure 5 A-RISC among 11S globulin protein sequences. A-RISC (Sequence homology index): A-RISC values were calculated to assess the degree of homology among allergens within the 11S globulin.

Figure 6 Conserved motif among 11S globulin (Gly m 6). (A) Surface representation of Gly m 6, an 11S globulin from soybean, highlighting in pink the conserved PHYN motif (Pro372–Asn375). This motif is localized on the surface, potentially accessible to IgE binding. (B) Cartoon representation of the same structure showing the PHYN motif situated within a β-barrel characteristic of the cupin domain, which is a highly conserved structural feature among 11S globulins. The position and exposure of this motif suggest potential immunological relevance. Structures were generated using UCSF ChimeraX (https://www.rbvi.ucsf.edu/chimerax) (Jalview in Appendix 2).

Figure 7 A-RISC among 7S Globulin protein sequences. A-RISC (Sequence homology index): A-RISC values were calculated to assess the degree of homology among allergens within the 7S Globulin.

Three conserved motifs were identified: “PC,” “CC,” and “CLK,” mostly between residues 40–90. These are part of the disulfide-stabilized core structure essential for IgE-binding regions. Additional motifs like “APC,” “YVRG,” “PGKCGV,” and “NPIY” define the immunogenic core of these stable proteins.

Glycinin (legumin, 11S globulin)

Classification: Low to moderate risk of cross-reactivity among legumes.

Homology analysis showed an A-RISC value of 0.57 between Ara h 3 and Gly m 6, suggesting a moderately high risk. Other comparisons showed lower similarity levels.

These proteins have two cupin domains (β-barrels with surrounding α-helices) and form hexamers, conferring strong stability.

In this study, the highly conserved “PHYN” motif was identified in the central regions of the sequences, forming part of the second cupin domain. This conservation, along with the moderate A-RISC values (~0.50–0.57), suggests a shared structural basis that may support the partial cross-reactivity observed clinically among these allergens.

Glycinin (Vicilins, 7S Globulin)

Classification: Low to moderate risk of cross-reactivity.

Vicilins are trimeric cupin superfamily proteins. While recognized as major allergens in peanut and soybean, sequence variability may limit cross-reactivity.

Conserved motifs include “VIV,” “NLR,” “ELVG,” “GFGGINA,” “FLAG,” “DNVI,” and “FPGS.” These motifs are exposed on the trimer surface and are structurally positioned for IgE recognition, enhancing their immunologic relevance.

Table 2 Protein family classification.

AllFam family	Biochemical name	Function in plant	Heat stability	Digestive resistance	Representative allergens
Bet v 1 family	Pathogenesis-related protein, PR-10	Pathogenesis-related proteins involved in plant defense against pathogens.	Low	Low	Gly m 4 Ara h 8 Cic a 4
nsLTP family	Nonspecific lipid transfer protein	Lipid transfer between membranes; defence protein expressed in leaves, fruits, and seeds	High	High	Ara h 9 Ara h 16 Arah17 Lup an 3 Pis s 3 Len c 3 Cic a 3 Pha v 3
Prolamin superfamily	2S Albumin	Seed storage protein; provides sulfur-rich amino acids essential for seed germination.	High	High	Gly m 8 Lup an δ Ara h 2 Ara h 6 Ara h 7
Cupin superfamily	7S Vicilin	Seed storage; involved in nitrogen storage and antimicrobial defense.	High	High	Gly m 5 Ara h 1 Lup an 1 Pis s 1 Len c 1 Cic a 1
	11S Legumin	Major seed storage protein; abundant in legumes, supplies amino acids during germination.	High	High	Gly m 6 Ara h 3 Pis s 2 Cic a 6
Prolamin superfamily	2S Albumin	Seed storage protein; provides sulfur-rich amino acids essential for seed germination.	High	High	Gly m 8 Lup an δ Ara h 2 Ara h 6 Ara h 7
Profilin family	Profilin	Cytoskeletal regulation (actin binding); expressed in all eukaryotic cells	Low	Low	Gly m 3 Ara h 5
Oleosin family	Oleosin	Lipid body stabilization in oil-rich seeds	High	High	Ara h 10 Ara h 11 Ara h 14 Ara h 15
Defensin family	Defensin-like protein	Antimicrobial peptides; plant innate immunity	Variable	Variable	Gly m 2 Ara h 12 Ara h 13
Cyclophilin family	Cyclophilin	Protein folding (isomerase); involved in stress responses	Unknown	Unknown	Ara h 18

Adapted from Pomés et al. (2018)⁸ and Sievers et al. (2011)¹⁰

2S albumin

Classification: Low to moderately low cross-reactivity risk.

The A-RISC index values for 2S albumins among different legumes remain below the clinically relevant threshold of 0.50, indicating a low likelihood of IgE-mediated cross-reactivity between these species.

A conserved “ALQ” motif was located around position 140 in several allergens (Ara h 2, Ara h 6, Ara h 7, Gly m 8, Lup an 6). This motif may contribute to structural stability and surface exposure of IgE-binding sites. Although global homology is low, the preservation of these structural elements may explain clinical co-sensitization among legumes.

Summary of risk estimation

This table synthesizes the main findings of the A-RISC index across all allergenic families and species.

Figure 8 Conserved motif among 7S Globulin (Pis s 1). (A) Surface representation of Pis s 1 showing exposed conserved motifs on the protein surface, including VIV, ELVG, NLR, and FLAG (highlighted in pink). (B) Cartoon representation of the same structure showing both exposed and buried motifs. Motifs GFGGINA, DNVI, and FPGS are located internally and are not visible in the surface model, while VIV, NLR, ELVG, and FLAG are spatially accessible. These motifs may contribute to IgE-binding and structural stability. Structures were generated using UCSF ChimeraX (https://www.rbvi.ucsf.edu/chimerax) based on the crystallographic model of Pis s 1 (Jalview in Appendix 2).

Figure 9 A-RISC among 2S Albumin protein sequences. A-RISC (Sequence homology index): A-RISC values were calculated to assess the degree of homology among allergens within the 2S Albumin.

Figure 10 Conserved motif among 2S Albumin (Ara h 2). (A) Surface representation of Ara h 2 showing the conserved “ALQ” motif (highlighted in bright pink), located on a solvent-accessible region of the protein surface. (B) Cartoon representation of the same structure, illustrating the positioning of the “ALQ” motif within a short β-strand embedded in the core α-helical bundle typical of the prolamin fold. This motif is conserved among several 2S albumins across legume species and may contribute to structural stability and allergenic potential. Visualizations were generated using UCSF ChimeraX (https://www.rbvi.ucsf.edu/chimerax) (Jalview in Appendix 2).

Table 3 Summary of in silico evaluation and estimated IgE cross-reactivity risk (A-RISC) across allergenic protein families from legumes.

Discussion

This study provides a comprehensive in silico comparative analysis of allergenic proteins from legumes, utilizing the A-RISC index and multiple sequence alignment to evaluate potential IgE cross-reactivity. Findings confirm that sequence homology alone does not fully predict cross--reactivity and must be complemented by structural motif analysis and epitope conservation.

Proteins such as Ara h 8 (peanut), Gly m 4 (soybean), and Cic a 4 (chickpea) exhibit high sequence similarity (67–72%), and conserved motifs such as “VEGNGGPGTIKKL,” which overlap with immunogenic regions, are known to elicit IgE responses. PR-10 proteins are thermolabile and susceptible to digestive degradation, typically inducing mild symptoms such as oral allergy syndrome (OAS).⁷ Sensitization commonly originates from pollen exposure, and subsequent ingestion of PR-10-containing foods can trigger cross-reactive responses, even in individuals who have not directly consumed all the legume sources involved.^11,12

Ara h 9, considered a major nsLTP allergen in peanut,¹³ exhibited low homology values with other legume nsLTPs such as lentil, pea, lupin, or bean, indicating a low probability of cross-reactivity between these species. This pattern aligns with previous literature, which has shown that although nsLTPs are highly stable and allergenic proteins, cross-reactivity between them depends on specific structural features and the individual sensitization profile of the patient. Multiple sequence alignments revealed conserved motifs such as “PC,” “CC,” and “CLK,” predominantly located between residues ~40 and 90.¹⁴ These motifs are essential for maintaining four disulfide bridges, which confer conformational stability and resistance to enzymatic digestion.^7,15 These findings align with structural studies by Salcedo et al. and Borges et al., showing that epitopes embedded within these stable structures are commonly recognized by IgE, even among nsLTPs with low overall sequence similarity.^14,15

11S globulins are composed of two cupin-type domains, characterized by β-barrels surrounded by α-helices. Their outer regions include conformational epitopes recognized by IgE. The cupin domain is functionally relevant and forms the structural basis shared across this protein family.⁷ These proteins assemble into highly stable hexamers, which confer thermal and digestive resistance, properties that significantly enhance their allergenic potential.^5,16 In this study, the highly conserved “PHYN” motif was identified in the central regions of the sequences, forming part of the second cupin domain.^16,17 This conservation, along with the moderate A-RISC values (~0.50–0.57), suggests a shared structural basis that may support the partial cross-reactivity observed clinically among these allergens.

7S globulins (vicilins), such as Ara h 1 and Pis s 1, also belong to the cupin superfamily and typically assemble into trimers composed of 45–60 kDa subunits.¹⁸ These proteins feature conserved β-sheet domains forming solvent-exposed epitopes, including motifs like “VIV,” “NLR,” “ELVG,” and “FPGS,” which enhance IgE accessibility.⁵ Their abundance and relative resistance to gastrointestinal digestion, along with the presence of both linear and conformational epitopes.^7,18 reinforce their role as clinically relevant allergens.

Finally, 2S albumins are low molecular weight proteins (12–15 kDa) belonging to the prolamin superfamily. These molecules are rich in cysteine residues and exhibit a compact tertiary structure stabilized by 4–5 disulfide bonds formed by a highly conserved cysteine pattern. These bonds are not typically exposed on the protein surface but instead provide structural support, maintaining the protein’s tertiary conformation—especially the α-helical regions—and conferring exceptional resistance to thermal processing and enzymatic digestion.^19,20 As a result, 2S albumins persist during food processing and digestion, enhancing their allergenic potential in predisposed individuals.^21,22 Our multiple sequence alignment identified a conserved motif, “ALQ,” located around position 140, shared by Ara h 2, Ara h 6, Ara h 7 (peanut), Gly m 8 (soybean), and Lup an 6 (lupin). This motif may play a role in stabilizing the conformational fold while also contributing to the surface exposure of linear IgE-binding epitopes. Although the global A-RISC values for 2S albumins were relatively low, suggesting limited cross-reactivity across legume species, the structural conservation of these motifs may underlie clinically observed co-sensitization, especially in individuals reactive to multiple legumes.^5,7

While the A-RISC index provides a practical and scalable approach to estimate cross-reactivity likelihood among homologous allergens, it does not account for individual sensitization patterns, HLA variability, or patient-specific IgE repertoires. A clinical study by López-Torrejón et al.¹⁶ found that individuals sensitized to lentil (Len c 1) often did not react to other vicilins, despite sequence and structural similarity, reinforcing that in silico predictions must be contextualized with clinical data.

This study is purely in silico in nature and does not introduce new algorithms or experimental validation. Therefore, its findings must not be interpreted as directly translatable to clinical diagnostics or management. Rather, this work serves as a theoretical basis for future in vitro or in vivo studies.

Moreover, the implications extend to food labeling and nutritional recommendations. International regulations, such as those from the EFSA and the US FDA, require the clear identification of allergens in processed foods.⁶ Nevertheless, the implementation of food labeling continues to face significant challenges, particularly for plant-derived foods in developing countries, where allergen-specific data and diagnostic infrastructure may be limited.

Understanding which protein families exhibit higher cross-reactivity risk can inform both clinical practice and regulatory policies. However, further experimental studies and clinical validation are necessary to translate these in silico predictions into practical applications.

In conclusion, this in silico analysis of allergenic legume proteins revealed variable levels of cross-reactivity among structural families. PR-10 proteins showed high risk (A-RISC >0.70), supporting the use of component-resolved diagnostics in pollen-sensitized patients. Moderate risk was observed for 11S and 7S globulins, while nsLTPs and 2S albumins showed globally low A-RISC values but may still contribute to co-sensitization due to structural conservation. These findings highlight the usefulness of the A-RISC index as a predictive tool and support molecular diagnostics to improve clinical interpretation and food labeling, particularly in cases of multiple sensitizations.