Patients allergic to mustard

Yellow mustard seeds extract, purified Sin a 2 and rabbit anti-Sin a 2 serum

Yellow mustard seeds, almond, hazelnut, pistachio, walnut and roasted peanut protein extracts were obtained as previously described for yellow mustard seeds [12]. The allergenic 11S globulin Sin a 2 was purified from yellow mustard seeds extract as described [12].

The specific anti-Sin a 2 serum was prepared by immunizing a New Zealand white rabbit with purified Sin a 2 by weekly injection of the protein in complete Freund’s adjuvant. After 21 days of treatment the serum was obtained by centrifugation of the blood.

Skin-prick tests

SPT were performed in all patients according to standard procedures [36]. The panel of commercial food extracts included peach, chestnut, soy, sunflower seed, almond, hazelnut, peanut, walnut, pine nut, pistachio, chickpea, lentil and bean. SPT with kiwi and apple were performed by the prick-prick method as described [37]. SPT with home-made mustard (Sinapis alba L.) extract (50 μg/ml) and with purified natural Sin a 2 (10 μg/ml) were performed. Histamine dihydrochloride (10 mg/ml) and physiologic saline solutions were used as the positive and negative controls, respectively. A wheal area <7 mm² was considered as negative. We included almond as a nut (instead of a Rosaceae fruit), and considered peanut separately.

Electrophoresis and immunoblotting

SDS-PAGE was performed in 15% polyacrylamide gels. Proteins (0.5 μg/lane of purified proteins or 50 μg/lane of protein extracts) were visualized by Coomassie Blue or alternatively transferred to nitrocellulose membranes (Amersham, Buckinghamshire, United Kingdom). The protein concentration was determined using the method of bicinchoninic acid (Pierce Chemical Co, Rockford, Ill, USA).

Immunodetection of proteins in membranes was achieved as described [14] by using different pool of sera from patients allergic to mustard, (diluted 1/5), or rabbit specific anti-Sin a 2 serum (diluted 1/100000). The binding of human IgE was detected by mouse anti-human IgE antibodies, provided by ALK-Abelló (Madrid, Spain), diluted 1/5000, followed by horseradish peroxidase-labelled goat anti-mouse IgG (diluted 1/5000; Pierce). Reaction to anti-Sin a 2 serum was detected by horseradish peroxidase-labelled goat anti-rabbit IgG (diluted 1/3000; BioRad, Richmond, CA). The signal was developed by using the ECL-Western blotting reagent (Amersham). For the IgG and IgE-inhibition experiments in immunoblotting, the pools of sera (diluted 1/5) or the rabbit anti-Sin a 2 serum (diluted 1/100000) were preadsorbed with 1 mg/ml of yellow mustard seeds, almond, hazelnut, walnut, pistachio or peanut extracts or with 20 μg/ml of purified Sin a 2 for 2h prior to membrane incubation as described [38]. Bovine serum albumin (BSA) was used as negative controls of inhibition. Volummograms of the reactive bands were analysed by scanning densitometry using the computer program Multigauge V3.0.

ELISA experiments

IgG quantitation was performed by ELISA in microtiter plates (Costar, Corning, NY, USA) coated with 100 μl/well of purified Sin a 2 (2 μg/ml) or yellow mustard seeds extract (20 μg/ml) [14]. Plates were incubated with increasing dilutions of the rabbit anti-Sin a 2 serum for titration. Then, the plates were incubated with horseradish peroxidase-labelled goat anti-mouse IgG as describe above and peroxidase reaction was developed using fresh enzyme substrate and measuring optical density (OD) at 492 nm. Each value was calculated as the mean of 2 determinations after blank subtraction.

For IgG-inhibitions ELISA, after being coated with 100 μl of Sin a 2 (2 μg/ml) or yellow mustard seeds extract (20 μg/ml), the plates were incubated with the rabbit anti-Sin a 2 serum (diluted 1/100000) previously preadsorbed with increasing amounts of Sin a 2 and mustard extracts as inhibitors for 2 h [38]. Then, the plates were incubated with horseradish peroxidase-labelled goat anti-mouse IgG and peroxidase reaction developed as described above. The percentage of inhibition was determined according to the formula: % Inhibition = (1- (OD_492nm with inhibitor/OD_492nm without inhibitor)) x 100. All the determinations were carried out as duplicates.

Sequence alignment and three-dimensional modelling

Multiple sequence alignments of Sin a 2, Pru du 6, Cor a 9, Pis v 2, Jug r 4 and Ara h 3 (accession numbers Q2TLW0, E3SH28, Q8W1C2, B7P073, Q2TPW5 and O82580, respectively) were performed with the Clustal Omega program [39]. Amino acid sequences of all 11S globulins are shown without signal peptide. Location of the signal peptide cleavage sites were predicted using SignalP 4.0 Server [40]. The 3D structure of Sin a 2 was modelled by using the services of the Swiss-Model Protein Modelling Server [41] and the structure of the soybean legumin Gly m 6 (PDB code 1od5A) [42] as template. The 3D structure of Ara h 3 correspond to PDB: 3c3v. Graphical processing of the three-dimensional structures was accomplished with PDB viewer PyMOL program.

Clinical features of mustard-allergic patients sensitized to Sin a 2 suggest that 11S globulins might be involved in cross-reactivity with tree nuts and/or peanut

By linking the clinical features of 34 well-characterized patients allergic to mustard with component-resolved approaches we previously showed that 2S albumin Sin a 1 is a marker of genuine sensitization to mustard, 11S globulin Sin a 2 is associated to severe symptoms after mustard ingestion and LTP Sin a 3 and profilin Sin a 4, two well-known panallergens, are associated with sensitization to other plant-derived foods from the Rosaceae family and pollens [16]. Although more than 50% of mustard allergic patients are also sensitized to tree nuts and/or peanut, significant clinical associations between sensitization to Sin a 2 and being allergic to other tree nuts or peanut were not found within our cohort of mustard allergic patients [16]. However, considering that recent studies suggested that 11S globulins are implicated in cross-reactivity among taxonomically unrelated tree nuts and peanut [30–33] and that different members of this protein family have been reported as relevant allergens [19–24], we investigated at the IgG and IgE level whether the 11S globulin Sin a 2 might be implicated in cross-reactivity among mustard, tree nuts and peanut within mustard allergic patients specifically sensitized to Sin a 2. We included in the study the 11 patients allergic to mustard with the general clinical characteristics summarized in Table 1. All patients presented allergic reactions within the first 30 minutes after mustard consumption and the inclusion criteria was based on positive specific IgE to yellow mustard seed extract and Sin a 2 as determined by SPT and ELISA (Tables 1 and 2). In line with our previous data, 10 of the 11 patients sensitized to Sin a 2 developed immediate systemic reactions after the ingestion of mustard and 7 attended the emergency room, where epinephrine was administrated [16]. All the patients included in the study had also positive SPT to tree nuts (almond or hazelnut or pistachio or walnut) or peanut and 6 of them (# 6, 7, 8, 9, 10 and 11) suffered from symptoms after ingestion of tree nuts (3 of them also with peanut, # 2, 4 and 5). Interestingly, only 3 patients showed positive specific IgE to the previously known cross-reactive mustard allergens Sin a 3 and 2 to Sin a 4 (Table 2), suggesting a potential role of Sin a 2, at least in the observed positive SPT to tree nuts and/or peanut. To further analyse the implication of Sin a 2 in cross-reactivity at the IgG and IgE level and to determine potential clinical relevance within the mustard allergic patients sensitized to Sin a 2, we obtained a rabbit anti-Sin a 2 serum and pooled patients’ sera into two different groups: i) Group 1, patients with positive SPT to tree nuts and/or peanut extracts without clinical symptoms to these allergenic sources; ii) Group 2, patients with positive SPT to tree nuts and/or peanut extracts with clinical symptoms to some of these allergenic sources (Table 1).

IgG epitopes of the 11S globulin Sin a 2 are present in almond, hazelnut, pistachio and walnut but not in peanut extracts

Compelling experimental evidence demonstrated that the use of well-defined purified allergens is very useful to improve diagnosis and treatment of allergic diseases [43, 44]. The availability of purified allergens allows the obtaining of specific polyclonal antibodies in mice or rabbit, which are key tools in the identification of potential cross-reactive homologous allergens in taxonomically related and non-related allergenic sources [14, 45]. We employed purified Sin a 2 to raise a rabbit anti-Sin a 2 serum as described in the methods section. The specific anti-Sin a 2 serum was titrated against purified Sin a 2 and yellow mustard seeds extract by ELISA and immunoblotting (Figure 1A and B). As shown in Figure 1A, the rabbit anti-Sin a 2 serum displayed equal and very high affinity to both purified Sin a 2 and mustard extract, demonstrating that purified Sin a 2 conserved the IgG epitopes and that the employed rabbit anti-Sin a 2 serum is specific to this allergen. The anti-Sin a 2 serum also recognized the protein band at around 51 kDa, corresponding to Sin a 2, in the mustard extract and the purified allergen in a dose-dependent manner in immunoblotting (Figure 1B). When the rabbit anti-Sin a 2 serum was employed at the highest concentration (dilution 1/50000), protein bands at around 30–32 kDa were slightly detected in both cases. These protein bands, which are also recognized by sera from mustard allergic patients, were previously identified as proteolytic fragments of Sin a 2 in mustard extract [13]. To further verify the specificity of the anti-Sin a 2 serum we performed IgG-inhibition experiments in ELISA and immunoblotting (Figure 1C and D). As shown in these figures, both purified Sin a 2 and mustard extract were able to block IgG binding to whole mustard extract or to Sin a 2-coated wells in a similar manner (Figure 1C). Complete inhibition of the binding of the anti-Sin a 2 serum to mustard extract or to purified Sin a 2 was reached when purified Sin a 2 or whole mustard extract were employed as inhibitors (Figure 1D). Collectively, all these data showed that the raised rabbit anti-Sin a 2 serum displays high affinity and specificity for the allergenic 11S globulin Sin a 2. Next, we tested the presence of IgG epitopes common to Sin a 2 in tree nut (almond, hazelnut, pistachio and walnut) and peanut extracts (Figure 2). These tree nuts and peanut extracts were selected because they were the most relevant allergenic sources to which the patients included in the study were allergic to or had positive SPT (Table 1). In Figure 2A, the Coomassie Blue staining after SDS-PAGE is displayed to visualize the protein content of the different extracts. The anti-Sin a 2 serum (lanes I) but not the pre-immune one (lanes P) reacted with protein bands at around 48–52 kDa in all the tested tree nut extracts. No reaction was detected when peanut extract was assayed. These results demonstrated that the previously described allergenic 11S globulins from almond (Pru du 6),[19] hazelnut (Cor a 9) [20], pistachio (Pis v 2) [24] and walnut (Jug r 4) [23] but not from peanut (Ara h 3) [22] share common IgG epitopes with Sin a 2. As shown in Figure 3, despite showing quite low percentages of identity (ranging between 28 and 39%) and similarity (between 46 and 56%), the compared 11S globulins present regions with conserved amino acid sequences such as for example those encompassing Gly⁸⁹-Asp⁹⁶, His¹⁵⁵-Asp¹⁶⁵ or Phe⁴²⁰-Ser⁴³⁶ (underlined in Figure 3). Ara h 3 is the 11S globulin showing the lowest identity (28%) and similarity (46%) with Sin a 2 and significant differences in the amino acid sequences of the above mentioned regions are also observed (Figure 3). Therefore, the presence of IgG binding sites at these regions, which do not completely overlap with the previously identified IgE hot spots (HS), might well justify the lack of reactivity of the anti-Sin a 2 serum against Ara h 3. To complement these data, we performed IgG-inhibition experiments in immunoblotting using BSA, purified Sin a 2 or mustard as controls, and almond, hazelnut, pistachio or walnut as potential inhibitors of the IgG binding to purified Sin a 2 (Figure 4A). The reactivity to Sin a 2 of the anti-Sin a 2 serum preadsorbed with purified Sin a 2 or mustard extract was nearly abolished (95 and 99% of inhibition respectively, quantified by scanning densitometry). When tree nut extracts were employed as inhibitors, an important reduction of the IgG-binding to Sin a 2 was also reached in all the cases. Almond and walnut were the extracts showing the highest inhibitory capacity (77% and 60% of inhibition, respectively) followed by hazelnut (53%) and pistachio (43%) as determined by scanning densitometry. Sin a 2 displays a region enriched in Gln and Gly (positions Gly¹¹⁸-Arg¹⁵², Figure 3), which is contained exclusively in Pru du 6. If this segment bears IgG binding sites, it might justify the highest percentage of inhibition reached by almond extract. As expected considering that the specific rabbit serum was raised against purified Sin a 2, the IgG binding to almond, hazelnut, pistachio and walnut was completely abolished when purified Sin a 2, mustard or the corresponding tree nut extract were used as inhibitors (Figure 4B). Collectively, all these results firmly demonstrated that Sin a 2 shows cross-reactivity with the allergenic 11S globulins from almond, hazelnut, pistachio and walnut at the IgG level, indicating that despite their low percentage of sequence identity they share IgG epitopes likely located in the most conserved regions of the protein family.

IgE cross-reactivity among allergenic 11S globulins from mustard, tree nuts and peanut

To gain insight into the potential role of Sin a 2 in cross-reactivity involving mustard, tree nuts and peanut at the IgE level, we performed IgE-inhibition experiments in immunoblotting using purified Sin a 2 and allergenic protein extracts from mustard, different tree nuts (almond, hazelnut, pistachio and walnut) and peanut. We pooled the patients’ sera into two different groups. In the group 1 we included the 5 patients without clinical symptoms to tree nuts and/or peanut at the moment of the study and in the group 2 the 6 patients that were allergic to tree nuts (3 of them also to peanut) (Table 1). As shown in Figure 5A, the reactivity to purified Sin a 2 of the pool of sera from group 1 preadsorbed with almond, hazelnut, walnut, pistachio or peanut was inhibited at different level depending on the employed extract. By scanning densitometry we quantified that almond (63% of inhibition) was the most potent inhibitor followed by walnut, pistachio, peanut and hazelnut (49%, 34%, 32%, and 15%, respectively). Interestingly, the pool of sera from group 1 recognized allergenic 11S globulins homologous to Sin a 2 in almond, hazelnut, pistachio, walnut and peanut. The IgE-binding to these proteins was nearly totally abolished when the pool of sera was preadsorbed with purified Sin a 2 except for hazelnut and peanut (Figure 5B), which might be due to co-sensitization to species-specific epitopes from 11S globulins contained in these allergenic extracts, the existence of different allergenic isoforms or the presence of additional allergens with the same molecular weight in hazelnut and peanut. These results show that for these patients Sin a 2 is the sensitizing allergen that shares common IgE epitopes with allergenic 11S globulins from the assayed extracts, which could explain why these patients despite of not suffering from clinical symptoms to tree nuts or peanut show in vivo positive SPT to such allergenic sources. To further investigate the potential clinical relevance of 11S globulin Sin a 2 in cross-reactivity, we performed the same type of experiments using the pool of sera from the mustard allergic patients sensitized to Sin a 2 with clinical symptoms to tree nuts and/or peanut (group 2). As shown in Figure 5C, the IgE-binding to purified Sin a 2 was significantly inhibited by almond, hazelnut, pistachio, walnut and peanut (83%, 62%, 60%, 67%, 69% and 67% of inhibition, respectively, as determined by scanning densitometry). This result firmly confirms that Sin a 2 share common IgE epitopes with allergenic 11S globulins from tree nuts and peanuts. This pool of sera (group 2) also recognized IgE-reactive protein bands at around 48–52 kDa that were not inhibited by Sin a 2 except in the case of almond (20% of inhibition by scanning densitometry) (Figure 5D). These results together with the fact that the inhibition of the IgE-binding to purified Sin a 2 by tree nut and peanut extracts was considerably higher for group 2 than group 1 indicated that mustard allergic patients sensitized to Sin a 2 and allergic to tree nuts and/or peanut might well be also primarily co-sensitized to 11S globulins from such allergenic sources. In addition, the pool of sera from group 2 recognized different low-molecular weight IgE-reactive protein bands (< 18 kDa) or epitopes from digested food that are not present in patient group 1 (Figure 5D and E), which could make the difference in the clinical symptoms associated with tree nuts and/or peanut ingestion. Interestingly, these IgE-reactive bands were not inhibited by 2S albumin Sin a 1 nor LTP Sin a 3 (Figure 5E), indicating that these families of allergenic proteins are not involved in the observed cross-reactivity. Collectively, our data suggest that Sin a 2 and the homologous allergenic 11S globulins from the studied tree nuts and peanut might contain both conserved cross-reactive and species-specific IgE epitopes. At this regard, several studies have previously shown the existence of five regions containing IgE-binding HS on 11S globulins from tree nuts (almond, hazelnut, cashew or walnut), peanut and soybean [19, 33]. Figures 3 and 6 show the amino acid sequence alignment of such regions comparing Sin a 2, Pru du 6, Cor a 9, Pis v 2, Jug r 4 and Ara h 3 as well as the location of the five IgE-binding HS on the 3D modelling of Sin a 2 and on the 3D structure of Ara h 3 . As shown in Figure 6A, the previously proposed IgE-binding HS-2 and HS-4 are quite conserved among the compared allergenic 11S globulins with percentages of identity with respect to Sin a 2 ranging from 40% (Pis v 2 and Ara h 3 in HS-4) to 66% (Pru du 6 and Cor a 9 in HS-4). The percentage of similarity with respect to Sin a 2 was even significantly higher and ranged from 66% (Jug r 4 in HS-2 and Pis v 2 in HS-4) to 86% (Pru du 6 in HS-2 and Pru du 6 and Cor a 9 in HS-4). As visualized in the 3D modelling of Sin a 2 and in the 3D structure of Ara h 3, both HS-2 and HS-4 are located at solvent-exposed areas of the monomer, thus representing ideal potential IgE cross-reactive epitopes in allergenic 11S globulins. When comparing HS-1, HS-3 and HS-5 lower percentages of identity were observed, being the HS-5 the most dissimilar. In addition, HS-1, HS-3 and HS-5 were not completely solvent-exposed on the 3D modelling of Sin a 2 or on the 3D structure of Ara h 3 suggesting that they might well constitute potential species-specific IgE epitopes. The identification of clinically relevant IgE binding sites to which patients are sensitized will contribute to improve diagnosis and accurate endotyping, which might well subsequently lead to develop more efficient and safer patient-tailored immunotherapy approaches [46–48]. Altogether, our study indicates that mustard allergic patients sensitized to Sin a 2 could present positive SPT to tree nuts and/or peanut without clinical manifestations against these allergenic sources due to IgE cross-reactivity involving 11S globulins. This finding is especially significant as it could help to avoid wrong diagnosis due to IgE cross-reactivity without clinical relevance. The fact that we also demonstrated IgE cross-reactivity involving Sin a 2 in mustard allergic patients sensitized to this allergen with clinical symptoms to tree nuts and/or peanut suggests the possibility that patients sensitized to Sin a 2 might eventually develop clinical episodes against these allergenic sources. Although the clinical relevance of the cross-reactive IgE-binding sites of the 11S globulin Sin a 2 needs to be studied in further detail, our results contribute to improve the diagnosis and management of the patients allergic to mustard sensitized to this allergen.