The importance of the 2S albumins for allergenicity and cross-reactivity of peanuts, tree nuts and sesame seeds

Abstract

Allergies to peanuts, tree nuts, and sesame seeds are among the most important food related causes of anaphylaxis. Important clinical questions include: why is there a variable occurrence of co-allergy among these foods and is this immunologically mediated? Clinical and immunological data summarized here suggest an immunologic basis for these co-allergies based on similarities among the 2S-albumins. Data from component resolved diagnostics have highlighted the relationship between IgE binding to these allergens and the presence of IgE-mediated food allergy. Furthermore, in vitro and in vivo experiments provide strong evidence that the 2S albumins are the most important allergens in peanuts for inducing an allergic effector response. Although the 2S albumins are diverse, they have a common disulfide linked core with similar physicochemical properties that make them prime candidates to explain much of the observed co-allergy among peanuts, tree nuts and sesame seeds. The well-established frequency of cashew and pistachio nut co-allergy (64–100%) highlights how the structural similarities among their 2S albumins may account for observed clinical cross-reactivity. A complete understanding of the physicochemical properties of the 2S albumins in peanuts, tree nuts and sesame seeds will enhance our ability to diagnose, treat and ultimately prevent these allergies.

Introduction

IgE-mediated food allergy is a major health problem with allergy to peanuts, tree nuts or sesame seeds affecting approximately 3% of children and up to 2% of adults in the US, the UK and Europe.^1–3 This is an important area of investigation because most children do not outgrow allergies to peanuts, tree nuts and sesame seeds, these foods account for most severe allergic reactions to foods, and co-allergy to these foods is relatively common.^4–9 Among the tree nuts, the most commonly reported allergies are walnut and cashew in the US, Brazil nut, walnut, hazelnut in Europe, and almond, Brazil nut and walnut in the UK.¹ These regional variations in the occurrence of allergies to peanuts, tree nuts and sesame seeds may reflect local diets, foods fed during infancy and levels of urbanization.^{6, 10–15}

Peanuts are legumes that grow underground and are phylogenetically different from nuts that grow on trees.¹⁶ However, peanuts, tree nuts and sesame seeds are all seeds in that they are fertilized ovules of their respective plants, so it is not surprising that these foods have similar physiology and biochemistry.¹⁷ Several common families of protein allergens have been identified in peanuts, tree nuts and sesame seeds including 2S albumins (the subject of this review; Table E1, available at www.jacionline.org), the cupin superfamily, which includes vicilins (7S globulins) and legumins (11S globulins), profilins, pathogen related (PR)-10 proteins and non-specific lipid transfer proteins (nsLTPs) that are also called PR-14 allergens.^18–21 Although we argue that the 2S albumins are the most important family of cross-reactive allergens, IgE binding to these other classes of allergens contribute to allergic reactivity and potentially to cross-reactivity.^{16, 18, 19, 22, 23} This is particularly true for IgE cross-reactivity to the PR-14 allergens that have been implicated in the oral allergy syndrome.^{16, 24}

Modern long read sequencing (e.g. MinION and other Next Generation Sequencing methods) and structural biology techniques (e.g. crystallography and in silico methods, see below) have enabled our groups and others to identify common IgE binding sequences within families of allergens.^{16, 25–29} However, due to the low overall sequence identity among allergens from the different sources, even within the same protein family (PFAM), relating the amino acid sequences of IgE epitopes to either cross-sensitization or IgE-mediated food allergy has been challenging. The goals of this manuscript are to review studies establishing that clinical cross-reactivity among peanuts, tree nuts and sesame seeds is prevalent, to review the evidence that the 2S albumins appear to be the most important allergens of peanuts, tree nuts and sesame seeds and to demonstrate how the structures of 2S albumins likely contribute to clinical cross-reactivity among these foods.

Co-allergy among peanuts, tree nuts and sesame seeds

As described in US and European guidelines, the presence of specific IgE (sIgE) either in serum or plasma or by demonstration of a positive percutaneous skin test (PST) is evidence of “sensitization” and not necessarily of food allergy.^{30, 31} “IgE-mediated food allergy” requires both the presence of sensitization and the report of clinical signs and symptoms on exposure to that food.^{30, 32} A “confirmed food allergy” requires a positive oral food challenge (OFC) or a double blind placebo controlled food challenge (DBPCFC).^{31, 33, 34} ”Co-allergy” indicates a positive OFC or DBPCFC to two or more allergenic foods such as peanuts, tree nuts or sesame seeds.³⁵ This may occur by either cross-reactivity or independent sensitization to each food. “IgE cross-reactivity” refers to the ability of one allergen to interfere with IgE binding to another as a consequence of the presence of shared IgE antibody binding sites among allergens from these different sources.^{16, 36}

Tree nut allergy in peanut allergic subjects.

Maloney and colleagues studied 324 US subjects (ages 2.4 months to 40 years) with suspected peanut, tree nut and/or sesame seed allergy. Of these, 234 had IgE-mediated allergy to peanuts⁴ and 201 (86%) were sensitized (sIgE of ≥0.35 kAU/L or a positive SPT) to at least one tree nut.⁴ Among those with IgE-mediated peanut allergy who were tree nut sensitized, the rate of IgE-mediated allergy to tree nuts was relatively low, varying from 6% (cashew) to 28% (walnut) (Table 1). Andorf and colleagues described 36 US children (ages 4–15 years) with confirmed allergy to peanuts who were enrolled due to a history of multiple food allergies and therefore had a higher likelihood of having a second confirmed food allergy, 34 of 36 (94%) were sensitized (sIgE>4kU_A/L and/or SPT>6mm) to at least one tree nut and/or sesame seeds.⁸ Among those with confirmed peanut allergy, the rate of confirmed allergy to tree nuts varied from 17% (almonds or sesame seeds) to 67% (pistachio) and 69% (cashew) (Table 1).

TABLE 1.

Tree nut or sesame seed allergy in peanut allergic subjects.

First author and reference	Location (ages)	Number with peanut allergy	Number (%) sensitized* to tree nuts or sesame seeds	Of those with peanut allergy and sensitized to tree nuts or sesame seeds, number (%) with IgE-mediated allergy (Maloney4) or with confirmed allergy (Andorf8)	Of those with peanut allergy, % with IgE-mediated allergy (Maloney4) or with confirmed allergy (Andorf8) to the specified tree nuts or sesame seeds
Maloney4	US (2.4 mo - 40 years)	234 (IgE-mediated food allergy)	Cashew: 152 (65)	Cashew: 9 (4)	6
			Almond: 126 (54)	Almond: 26 (11)	21
			Walnut: 101 (43)	Walnut: 28 (12)	28
			Pecans: 117 (50)	Pecan: 14 (6)	12
			Pistachio: 178 (76)	Pistachio: 14 (6)	8
			Sesame seed: ND	Sesame seed: ND
Andorf8	US (4–15 years)	36 (OFC)	Cashew: 26 (72)	Cashew: 25 (69)	96
			Almond: 13 (36)	Almond: 6 (17)	46
			Walnut: 21 (58)	Walnut: 19 (53)	90
			Hazelnut: 25 (69)	Hazelnut: 17 (47)	68
			Pecan: 18 (50)	Pecan: 17 (47)	94
			Pistachio : 24 (67)	Pistachio : 24 (67)	100
			Sesame seeds: 9 (25)	Sesame seeds: 6 (67)	67

^*In Maloney⁴, sensitization was defined as a positive skin test (size not specified) or increased sIgE (≥0.35kAU/L). In Andorf⁸, sensitization was defined as SPT>6mm and/or sIgE>4kU_A/L (Immunocap® and only sensitized subjects were challenged. ND, not determined.

Other food allergies (not tree nuts or sesame seeds) in peanut allergic subjects.

In the Andorf study, of 273 positive DBPCFC that were performed in 60 US children (ages 4–15 years) recruited based on their history of allergic reactions to multiple foods, 236 (86%) challenges demonstrated allergy to peanut, tree nut or sesame compared to 19 (7%) for egg,14 (5%) for milk and 2 (0.7%) for both soy and wheat (p<0.0001; data derived from Andorf et al.).⁸

Peanut allergy in tree nut allergic subjects.

Andorf and colleagues also reported that in 55 US children (ages 4–15 years old) with confirmed tree nut allergy who were enrolled due to multiple food allergies, the presence of confirmed tree nut allergy was associated with a 54–60% incidence of confirmed peanut allergy. No single confirmed tree nut allergy was more likely than another to be associated with confirmed peanut allergy.⁸

High rates of co-allergy between cashews and pistachios and between walnuts and pecans have been observed in multiple studies.

Savvatianos and colleagues reported 100 Greek children (ages 3.5–10 years) with suspected IgE-mediated allergy to either cashews or pistachios. Of these, 25 were identified by OFC to have either confirmed pistachio or cashew allergy. All 25 (100%) had confirmed allergy to both.³⁷ Similar high rates for co-allergy between pecan and walnut were seen by Elizur and colleagues in 34 Israeli subjects with confirmed pecan allergy; 100% of confirmed pecan allergic subjects were confirmed to be allergic to walnuts.³⁵ However, in the Elizur study, only 64% of those confirmed to be walnut allergic were found, on OFC, to be allergic to pecans showing that there is not perfect occurrence of co-allergy.³⁵ In the previously mentioned US study by Andorf and colleagues of subjects with multiple food allergies, there were 46 subjects with confirmed cashew allergy. Of these, 42 (91%) were confirmed to be allergic to pistachios.⁸ These 42 were all the subjects allergic to pistachios in this study, so 100% had confirmed allergy to cashews. In the same study, of 32 subjects with confirmed walnut allergy, 29 (81%) had confirmed allergy to pecans. All 29 subjects with confirmed allergy to pecans also had confirmed allergy to walnuts (100%).⁸

Cross-desensitization is possible for certain tree nut pairs.

In a single center prospective study of 46 Israeli patients with IgE-mediated or confirmed walnut/peanut co-allergy, Elizur and colleagues found that all who were successfully desensitized to walnut (maximum dose of 4,000 mg) tolerated 2,500 mg of pecan protein.³⁸ Andorf and colleagues have also reported two US clinical trials of oral immunotherapy to multiple foods. In these studies a subset of participants with confirmed allergy to cashew and pistachio or walnut and pecan were given OIT with cashew but not pistachio or walnut but not pecan.^{39, 40} In the first study, 20 of the 24 (83%) participants who were successfully desensitized (passed 2g DBPCFC) to cashew by cashew-OIT, also passed a 2g pistachio DBPCFC.³⁹ All 17 participants who were successfully desensitized to walnut were also desensitized to pecan without receiving pecan in their OIT.³⁹ Similarly, in the second study, 3 of 4 (75%) participants with confirmed allergy to walnut and pecan were cross-desensitized to pecan, and 8 of 8 (100%) participants with confirmed allergy to cashew and pistachio were cross-desensitized to pistachio.⁴⁰

Immunochemical assays

Relationship of IgE binding to clinically observed co-allergies.

In the previously mentioned study by Maloney and colleagues, of 234 subjects with suspected peanut, tree nut or sesame seed allergies, Spearman rank correlation coefficients of sIgE values between peanuts and tree nuts varied from 0.4 to 0.53 whereas among the tree nuts, the coefficients were 0.40 (almond and walnuts), 0.84 (almond and hazelnuts), 0.96 (pecans and walnuts) and 0.95 (pistachios and cashews).⁴ For peanuts and sesame seeds, the correlation coefficient was 0.34.⁴ In the US study of 60 children with multiple food allergies, IgE binding to pistachios significantly correlated with sIgE levels to Ana o 3 (cashews) (r=0.85) and IgE binding to pecan significantly correlated with IgE binding to Jug r 1 (walnuts) (r = 0.84).⁸

Inhibition assays.

Published data have been inconclusive, with reports that inhibition assays with extracts of peanuts and specific tree nuts were either moderately correlative^{23, 41}, weakly correlative^42–44 or not correlative²² with clinical histories of co-allergy (Table E2; available at www.jacionline.org). The first 3 studies shown in Table E2 examined very few sera and are shown for completeness.^42–44 Masthoff and colleagues employed ImmunoCAP inhibition and found no significant inhibition of IgE binding to rCor a 14 by nAra h 2 or by nAra h 6 but did find nearly complete inhibition of IgE binding to nCor a 9 by nAra h 3 in 2 of 13 subjects. Overall, inhibition experiments with the 11S globulins were highly variable among subjects.²² Villata and colleagues also used ImmunoCAP inhibition and reported that preincubation of sera with whole walnut extract reduced binding to both Cor a 9 and Cor a 14, the 11S globulin and 2S albumin respectively of hazelnut.²³ Elizur and colleagues used ELISA inhibition and found that, in patients with walnut and pecan co-allergy, walnut pretreatment completely blocked IgE binding to pecan but pecan incubation only partially blocked IgE binding to walnut.⁴¹ Thus, based on published data, there is not consistent evidence that co-allergy is due to immunologic cross-reactivity. The explanation for these variable findings may be selection of subjects, the use of recombinant proteins that may not have correct tertiary structure or that the presence of co-allergies is due to the random occurrence of these different food allergies in persons predisposed to food allergies. In the subsequent sections of this manuscript, we will review data indicating that further elucidation of IgE binding to linear and conformational epitopes of the 2S albumins of peanuts, tree nuts and sesame seeds holds the key to understanding co-allergy among these foods.

2S albumins appear to be the most important allergen family for peanuts, tree nuts and sesame seeds

IgE binding to 2S albumins is highly related to the presence of IgE-mediated food allergy.

Among all the allergens identified in tree nuts, peanuts and sesame seeds, it is remarkable that for peanut, cashews, hazelnuts, walnuts and sesame seeds, IgE binding to the 2S albumins is highly related to the presence of IgE-mediated allergy to peanut or to a specific tree nut (Table E3; available at www.jacionline.org).^{45, 46} For example, the sensitivity/specificity of IgE binding to Ara h 2 and Ara h 6 in relation to confirmed allergy to peanut is better than IgE binding to crude peanut extract or other peanut allergens.^{34, 45, 47–49} Serum IgE binding to both Ara h 2 and Ara h 6 correlates better with clinical symptoms than measuring either independently.^{50, 51} Measuring sIgE to Ana o 3, Jug r 1 or Ses I 1+2 has the highest relationship with the presence of IgE-mediated allergy to cashews, walnuts or sesame seed, respectively.^52–54 IgE binding to Cor a 9 (11S globulin) and Cor a 14 (2S albumin) are equivalently related to the presence of IgE-mediated allergy to hazelnuts (Table E3).^{36, 55} IgE to the 2S albumin, Pin p 1, from pine nut, is highly related to IgE-mediated allergy to pine nuts, but this has not been correlated with outcomes of OFC and therefore Pin p 1 is not included in Table 2.⁵⁶ Furthermore, the ability of IgE binding to any of these allergens to predict the presence of IgE-mediated or confirmed allergy has not been confirmed in a prospective study.

TABLE 2. The most conserved area of the 2S albumin allergens contains a similar pattern of cysteine residues, separated by variable residues with similar PCPs^a.

The sequences are listed according to their similarity to Ara h 2. See Fig 1 for a graphical presentation of their similarities according to their property distances (PD).

^aStarting from the most conserved areas of Ara h 2 (26 amino acids, first line), the peptide similarity search tool in SDAP was used to identify other tree nut and sesame seed 2S albumin allergens based on their physicochemical property distance (PD). PD (Property Distance). The table is organized according to decreasing similarity to the Ara h 2 as indicated by Z scores of the found hits. Z-PDmin uses in the calculation of the Z-score the lowest PD values in each allergen sequence found in a search in SDAP. PDall is based on the PD values of all windows of identical length in all sequences in SDAP. The known linear IgE epitope in Ara h 2 is underlined. Cysteines are shown in a larger font to aid in alignment.

^*The sequence of Pin p 1 is taken from¹¹⁸.

IgE binding to the 2S albumins is associated with relevant clinical outcomes.

High levels of IgE against the 2S albumins, Ana o 3 (cashew), Jug r 1 (walnut) and Cor a 14 (hazelnut) were significantly associated with GI reactions during DBPCFC.⁸ IgE binding to specific linear epitopes of Ara h 2 may distinguish peanut allergic from peanut sensitized subjects and, in infants, has been associated with severity of clinical history, sensitivity to challenge and with likelihood of achieving sustained responsiveness following OIT.^57–59

The 2S albumins are the most potent peanut allergens for inducing an allergic effector response.

Ara h 2 and Ara h 6 are the most potent peanut allergens in basophil assays and for PST.^60–62 Koppelman and colleagues performed basophil histamine release and intracutaneous testing with purified native Ara h 1, Ara h 2 and Ara h 3. In both assays, Ara h 2 was 100-fold more potent on a weight basis than Ara h 1 or Ara h 3.⁶⁰ Palmer and colleagues used the rat basophil leukemia cell assay, and reported, on a molar basis, that native Ara h 2 was 50-fold more potent than native Ara h 1.⁶¹ Peeters and colleagues performed SPT with purified native Ara h 1, Ara h 2, Ara h 3 and Ara h 6. Ara h 2 and Ara h 6 resulted in positive SPT at 0.1 μg/ml whereas Ara h 1 was positive at 10 μg/ml and Ara h 3 at 100 μg/ml.⁶². In addition, Ara h 2 and Ara h 6 account for much of the effector activity in crude peanut extracts.^63–66 Chen and colleagues examined the potency of Ara h 2 and Ara h 6 by selectively removing these 2S albumins from a crude extract of peanuts using an immunoaffinity column and found that this reduced the activity of the extract by 10-fold.⁶³ Kulis and colleagues reported that the 2S albumins⁶³ alone were sufficient to desensitize peanut allergic mice to challenge with whole peanut extract.⁶⁴ Monoclonal antibodies directed against Ara h 2⁶⁵ and nanoparticles presenting specific linear epitopes of Ara h 2 and Ara h 6⁶⁶ can block basophil activation elicited by crude peanut extracts by as much as 80%. A recently developed hypoallergenic version of Ara h 2 may be useful for immunotherapy.⁶⁷

The sequences and structures of the 2S albumins are key to understanding allergic reactivity to peanuts, tree nuts and sesame seeds

The biology of the 2S albumins.

The conserved presence of homologues of the 2S albumins in the seeds of plants that diverged over 300 million years ago⁶⁸ suggest they may play some functional role. The 2S albumins contain 2–4% methionine (M), 5–6% cysteine (C), 10–12% arginine (R) and ~25% of either glutamine (Q) and glutamate (E), suggesting that they may serve as a reserve source of sulfur and/or nitrogen (compared to other food proteins) for cellular metabolism.^{19, 69} Many of the IgE binding epitopes for Ara h 2 are glutamine (Q)-rich, with several short segments containing a preponderance of Q (e.g. RQLQQQEQIK in Ana o 3, KRQQQQGQFREGIR in Pis v 1, RQQQQEGQFER in Car i 1, RRQQQQQGLR in Jug r 1 and RLQGRQQEQQ in Ara h 2 (Fig E1 (available at www.jacionline.org) and not shown). Similar Q-rich regions are characteristic of ω gliadins from wheat, which are important allergens⁷⁰ whereas the 2S albumins of soybeans, which are relatively non-allergenic, do not contain Q-rich regions.⁷¹ Q-rich regions are important for transcription factor activity as well as other cellular functions and are also found in a wide variety of viruses, including mumps, herpes and influenza.⁷² Although the high content of C residues in the 2S albumins are for the most part involved in disulfide bonds, additional free C residues may serve as metal binders.⁷³ Conserved areas of the 2S albumins may also play a role in resistance to fungi.^{19, 74, 75}

The 2S albumins are characterized by their small size (< 150 amino acids). They are synthesized as a single chain of 18–21 kDa and are post-translationally processed in the endoplasmic reticulum: 4–5 disulfide bonds are formed and in most cases proteolysis leads to the formation of two subunits, held together by disulfide bonds.⁷⁶ Under reducing conditions, such as for example on gel electrophoresis, the post-translationally cleaved 2S albumins appear as a small N-terminal subunit (3–5 kDa) and a larger C-terminal subunit (8–10 kDa).^{19 ,77, 78} The most conserved area of the 2S albumins is a conserved 8-C motif (..C…C/..CC..CxC…C..C).⁷⁹ The intervening residues have similar physicochemical properties (PCPs) according to the peptide similarity search implemented in the Structural Database of Allergenic Proteins (SDAP).⁸⁰ This program finds peptides with a low “property distance” (PD) to a user supplied peptide (up to 30 AA long) in the allergenic sequences included in SDAP (Table E4; available at www.jacionline.org).^81–83 Table 2 illustrates the similarity of the 2S albumin allergens identified from a search, starting with a core sequence from Ara h 2, GSSQHQERCCNELNEFENNQRCMCEAA, in terms of bit and E-scores (for description of these terms, see the footnote of Table E4). Note that outside of the absolutely conserved array of C, the sequences appear quite diverse. However, they are all hydrophilic, with many Q and charged residues. An E-score <0.01 indicates that all of these proteins will have a similar fold. Their diversity is reflected in the low bit scores. Peptides from this area with low PD value to known epitopes have been shown to be IgE binding epitopes. The central part of this peptide, NELNEFENNQR, was identified as a linear IgE binding epitope for Ara h 2.^{58, 84, 85} This region is 73% identical to the related region in Ara h 6 and 82% in Ara h 7.

A new program, PD-graph relates the PD values between the sequences to their likelihood to cause cross-reactions (Fig 1).^{86, 87} Consistent with the phylogenetic distance of their producing species^68,88 and with clinical studies, the 2S-albumins of walnut/pecan (Jug r 1, Jug n 1, Car i 1) and pistachio/cashew (Pis v 1, Ana o 3) occur near one another in the plot, suggesting these protein pairs would be most likely to cross-react. The 3 peanut allergens are most similar to each other, with greater distance to all of the tree nuts and sesame seeds. Pin p 1 of pine nut, from a conifer, is an outlier, consistent with the low number of cross reactions to other tree nuts (which are angiosperms) and peanut⁵⁶, perhaps due to their evolutionary distance to this gymnosperm derived allergen.

FIG 1. PD-graph of the central region of 2S albumins from peanut, tree nuts and sesame seeds.

PD-graph is a program that automatically calculates and graphically portrays the pairwise property distances (PD) between unaligned sequences supplied in Fasta format as a text file.⁸⁷ Lower PD indicates higher sequence similarity. The program is available from Github (https://github.com/bjmnbraun/DGraph).¹³⁴ Consistent with observed clinical studies, walnut/pecan (Jug r 1, Jug n 1, Car i 1) and pistachio/cashew (Pis v 1, Ana o 3) share extensive common sequences. Pine nut (Pin p 1) is an outlier, consistent with the low number of cross-reactions observed. The 3 peanut allergens (Ara h 2, Ara h 6 and Ara h 7) are most similar to each other, with greater distance to all of the tree nuts and sesame seeds. See Table 4 for allergen source and sequence. Allergen names have been shortened to comply with FASTA header format, which does not permit spaces. The scale (left) relates the line density to the property distance (PD) between pairs (lower PD= higher similarity in PCPs).

Structural stability of 2S albumins is related to immunogenicity.

The disulfide bonded core plays an important role in maintaining the structure of 2S albumins. Disrupting disulfide bonds by chemical reduction and alkylation leads to a change in structure and concomitant change in biochemical and immunological characteristics.^89–91 Compared to Ara h 1 and Ara h 3, the exceptional stability of Ara h 2 and Ara h 6 to gastric acid and digestive enzymes may enhance their allergenicity.^{74, 92–96} Two digestion products of Ara h 2 and Ara h 6 (5 and 9 kDa), held together by disulfide bonds, with intact secondary and tertiary structure are detected in circulation and breastmilk of non-peanut allergic human volunteers after peanut consumption.^97–99 In vitro, these are as potent as the undigested allergens in binding IgE.¹⁰⁰

Relating secondary and tertiary structure to IgE binding.

In view of the importance of conformational epitopes^{25, 82}, structure determination of the 2S albumins was an early priority for molecular allergologists. Ara h 2 was difficult to crystalize, probably due to the flexibility of a variable size internal loop. Thus, the first structure determination for the peanut 2S albumins was completed for the core region of Ara h 6, which proved more stable than that of Ara h 2. A reliable model structure of Ara h 2 was based on the Ara h 6 NMR structure.¹⁰¹ A crystal structure, albeit incomplete, was finally achieved by fusing Ara h 2 to maltose binding protein.¹⁰² The 3 dimensional fold of Ara h 2 is shown in Fig 2A and compared to the fold of another 2S albumin allergen, Ber e 1 (Fig 2B). These figures illustrate the common fold of the 2S albumins with five α-helices and a C-terminal loop that is stabilized by 4–5 disulfide bonds.⁷⁴ Between α-helix H2 and H3, there is an exposed and highly antigenic hypervariable region⁷⁴, which may or may not be cleaved. This loop region containing several copies of the highly IgE-reactive motif DPYS was only partially discernable in the electron density map, indicating this area is “disordered”.

FIG 2. The structural folds of Ara h 2 and Ber e 1.

A) Crystal structure of Ara h 2 (PDB id: 3OB4) and B) NMR structure of Ber e 1 (PDB id. 2LVF) showing the secondary structure elements (Helices H1-H5) conserved in the 2S albumins. The unstructured (disordered) loop characteristic of Ara h 2 is shown as a dotted line to the right connecting H2 and H3. The four disulphide bonds connecting helix H1/H4, H3/H5, H3/H4 and H4 and the C-terminal of the protein are shown in yellow.¹⁰⁶

IgE binding epitopes of the 2S albumins are both linear and conformational.

The linear sequences of the 2S albumins from different peanuts, tree nuts and sesame seeds, show low similarities (29–60%) (Table E5;available at www.jacionline.org) with the exception of the known dyads of walnut/pecan (Jug n/r 1 and Car i 1; 88%, both from trees classified as Juglandaceae) and cashew/pistachio (Ana o 3 and Pis v 1; 70%, both from trees classified as Anacardiaceae) (Fig 1 and Fig E1).¹⁶ As discussed above, these pairs come from plants classified within the same family and show significant co-allergy. Even though the peanut-2S albumins, Ara h 2, Ara h 6 and Ara h 7, are only 49–55% identical¹⁶, their PCPs are similar.⁸⁷ As the PD-graph of Fig 1 shows, they cluster together and are quite distinct from the sesame and tree nut 2S albumins. Evidence that indicates the most important epitopes of the 2S albumins are conformational, includes 1) reduced and alkylated preparations of Ara h 2 have greatly reduced ability to bind IgE compared with native molecules^{89–91, 103} and 2) IgE binding mimotopes identified by screening a random phage library with affinity purified anti-Ara h 2 and Ara h 6 IgE do not relate to linear sequences.¹⁰⁴ As noted above, ”unstructured” and even ”disordered” areas of the proteins are also important. Peptides representing the DPYSP^OH (P^OH is hydroxyproline) motif of Ara h 2.01, which is duplicated in Ara h 2.02, are particularly active in both IgE binding and basophil activation assays.^{66, 103} The hydroxylation of specific prolines in this region significantly enhances IgE binding.¹⁰³ Compared to other linear regions, the DPYSP^OH motif of Ara h 2 is the most reactive IgE binding motif yet, following reduction and alkylation of Ara h 2 which should not affect the structure of this region, most IgE binding is lost. It is possible, that although this is the most potent linear region for binding IgE, the majority of the binding of IgE to native Ara h 2 is to conformational regions. A quantitative study to assess this has not, to our knowledge, been published.

Relating sequence and 3D structures of the walnut/pecan and cashew/pistachio 2S-albumins to cross-reactivities.

Novel structural bioinformatics tools, such as Cross-React finds surface patches on 3D structures of potential cross-reacting allergens with amino acid compositions similar to an epitope of a known allergen.¹⁰⁵ The Pearson correlation coefficient (PCC) is used to rank accessible surface patches of the target allergens to the amino acid composition of the query epitope.¹⁰⁵ Conformational epitopes can also be identified based on models, as for example the clinically cross-reactive allergens cashew/pistachio and walnut/pecan. The multiple sequence alignment (Fig E1) shows that the sequence identity is 88% for Jug r 1 (walnut)/Car i 1 (pecan), while that of Ana o 3 (cashew) and Pis v 1 (pistachio) is only 70%. In Fig 3, Ana o 3 and Pis v 1 were modeled as 5-helix bundles, based on the NMR solution structure of the BN allergen, Ber e 1.¹⁰⁶ Linear epitopes of Ana o 3 are shown in Fig 3A and conserved residues between Ana o 3 and Pis v 1, as identified from the sequence alignment (Fig E1) are shown for Ana o 3 in Fig 3B. The corresponding figure for Pis v 1 is very similar to Fig 3B and not shown. Mapping conserved residues of Ana o 3 and Pis v 1 on the models identified a large surface exposed area (blue), indicated by a dashed oval in Fig 3B, which overlaps with the linear epitopes in helix 1 and helix 3 of Ana o 3.¹⁰⁶ The identified common surface patches of these allergens overlap with the known linear epitopes of Ana o 3 (Fig 3B).^106–108 The Cross-React program also predicts this surface as a potential common conformational epitope of Ana o 3 and Pis v 1 with a high score (Fig 3C and D). The patches contain residues from the linear epitope 1 of Ana o 3, such as Q15, and residues from epitope 3 (E44, Q47, E48, Q50 and E 51). These observations, coupled with the results for the linear epitopes, suggest that this surface patch could account for the clinically observed cross-reactivity between Ana o 3 and Pis v 1.

FIG 3. Potential conformational epitopes shared by Ana o 3 and Pis v 1.

A) Mapping of four linear epitopes on the Ana o 3 structure. The 3D model of Ana o 3 was generated using the NMR structure Ber e 1¹⁰⁶ as a template. The four linear IgE epitopes of Ana o 3 as identified in^{107, 108} are highlighted in colors. B) Conserved residues between Ana o 3 and Pis v 1, identified in the alignment of the sequences (Fig E1) are mapped on the surface of the 3D model of Ana o 3. This representation shows a large surface exposed area (blue), indicated by a dashed oval. This conserved region overlaps to a large extent with the linear epitopes in helix 1 and helix 3 of Ana o 3. C) and D) Structurally similar surface patches of Ana o 3 (red) and Pis v 1 (green). A common surface patch of Ana o 3 and Pis v 1 was found by Cross-react with a high similarity score of 0.90. The surface patch in Ana o 3 is centered at residue Q47; its best counterpart in Pis v 1 was found at the equivalent residue Q54 of Pis v 1. Residue Q47 is also labeled in A). This surface patch could be part of a conformational epitope responsible for the clinical observed cross-reactivity between CN and Pis v 1.

Conclusions

Co-allergy to peanuts, tree nuts and/or sesame seeds may be due to immunologic cross-reactivity¹⁶ or may be a random event occurring in patients who are prone to producing IgE.²² However, the extensive clinical, immunochemical and structural data summarized here indicates that finding IgE-mediated allergy or confirmed allergy in the same patient to some combination of peanut, tree nuts or sesame seed is more likely related to specific immunologic responses to regions of common PCPs of the allergenic proteins. In addition, although understanding the structures of all known allergens is a worthwhile goal, we argue that the most important allergenic proteins in the context of understanding co-allergy, are the 2S albumins. First, confirmed allergies to peanut, tree nuts or sesame seeds appear to coexist at a higher rate compared with confirmed allergy to non-seed animal based foods (e.g. milk and eggs) and non-seed vegetables (e.g. soy and wheat). Second, the physicochemical similarities (e.g. resistance to gastric acid) and structural similarities (e.g. similar core sequences and 3D structures) of the 2S albumins supports the concept that coexistence of these allergies has a structural basis. Third, pairs of tree nuts (e.g. walnuts/pecans and cashews/pistachios), where, as we show here, the 2S albumins are very similar (Figs. 2, 3), have particularly high frequencies of IgE-mediated co-allergy. Finally, the 2S albumins from these foods share key IgE binding regions that likely contribute to surface epitopes and these, in turn, may explain the co-existence of IgE-mediated allergy to peanuts, tree nuts and sesame seeds.

Understanding cross-reactivity among peanuts, tree nuts and sesame seeds will require a detailed view of the 3D structure of the 2S albumins combined with extensive experimental studies of IgE binding to potentially cross-reactive linear and conformational epitopes. Better identification of conformational IgE binding epitopes and design of hypoallergenic proteins can further improve understanding of the molecular basis for co-allergy among peanuts tree nuts and sesame seeds. This approach promises to greatly enhance our ability to diagnose, treat and ultimately prevent these allergies.

Abbreviations:

C

cysteine

DBPCFC

double blind placebo controlled food challenge

M

methionine

PCC

Pearson correlation coefficient

PCP

physicochemical properties

PD

property distance (a measure of similarity)

PFAM

protein families

P^OH

hydroxyproline

Q

glutamine

R

arginine

SPT

skin prick tests

S

Svedberg sedimentation coefficient

SDAP

Structural Database of Allergenic Proteins

sIgE

specific IgE