Oleosin Cor a 15 is a novel allergen for Italian hazelnut allergic children.

BACKGROUND: Hazelnut allergy, which is characterized by symptoms that range from mild to severe, is one of the most common allergies in children throughout Europe, and an accurate diagnosis of this allergy is therefore essential. However, lipophilic allergens, such as oleosins, are generally underrepresented in diagnostic tests. We therefore sought to characterize the IgE reactivity of raw and roasted hazelnut oleosins, using the sera of hazelnut-allergic pediatric patients.
METHODS: Raw and roasted hazelnut oil body-associated proteins were analyzed by means of 1D and 2D electrophoresis and MS. Oleosin IgE reactivity was assessed by immunoblotting with the sera of 27 children who have confirmed hazelnut allergies and from 10 tolerant subjects. A molecular characterization of the oleosins was performed by interrogating the C. avellana cv. Jefferson and cv. TGL genomes, and through expression and purification of the recombinant new allergen.
RESULTS: A proteomic and genomic investigation allowed two new oleosins to be identified, in addition to Cor a 12 and Cor a 13, in hazelnut oil bodies. One of the new oleosins was registered as a new allergen, according to the WHO/IUIS Allergen Nomenclature Subcommittee criteria, and termed Cor a 15. Cor a 15 was the most frequently immunorecognized oleosin in our cohort. Oleosins resulted to be the only immunorecognized allergens in a subgroup of allergic patients who showed low ImmunoCAP assay IgE values and positive OFC and PbP. Hazelnut roasting resulted in an increase in oleosin immunoreactivity.
CONCLUSION: A novel hazelnut oleosin, named Cor a 15, has been discovered. Cor a 15 could play a role in eliciting an allergic reaction in a subgroup of pediatric patients that exclusively immunorecognize oleosins. The high prevalence of hazelnut oleosin sensitization here reported further confirms the need to include oleosins in routine diagnostic procedures.

The manuscript deals with the discovery of a new hazelnut allergen, Cor a 15. It also provides novel evidence concerning the role of oleosins in hazelnut allergy, the most common tree nut allergy in Europe. The research findings are relevant for clinical practice, since they confirm the need to include lipophilic allergens, such as oleosins, in the protein extracts used for routine diagnostic tests, as already found for peanuts.

INTRODUCTION

Hazelnut (Corylus avellana L.) allergy is the most common tree nut allergy in Europe , with a different prevalence being observed among European countries. Associated oil body (OB) proteins have been receiving increasing interest, since oleosins have been described as major allergens in peanuts , and in sesame. , Moreover, they have been associated with severe symptoms as a result of the consumption of hazelnuts. , In 2015, the EuroPrevall project established that hazelnut oleosin sensitization is quite common in Europe (which has from 10 to 25% of hazelnut‐allergic patients), with a higher prevalence in children than in adults (34% vs 11.4%). Akkerdaas et al. reported the two first allergenic oleosins in hazelnuts, Cor a 12 and Cor a 13.

Oleosins are usually underrepresented in skin prick test (SPT) extracts and in vitro reagents, thereby giving rise to false negatives in hazelnut allergy diagnosis. , Their absence in test reagents is due to their hydrophobic structure, which results in their eluding the standard protein purification procedures. Zuidmeer‐Jongejan et al. pointed out the absence of Cor a 12 and Cor a 13 in nine of the most commonly used SPT extracts.

The IgE binding potential of peanut oleosins has recently been found to be increased by in‐shell roasting. However, the effect of roasting on hazelnut oleosin immunorecognition remains unknown.

Our aim was to investigate the contribution of oleosins to hazelnut pediatric allergies and to assess the effect of roasting on hazelnut oleosin immunogenicity.

METHODS

Details about the methods used for most of the conducted experiments are available in the Online Repository.

Study population and clinical methods

Pediatric subjects with a convincing history of hazelnut allergy (HA patients, no. = 27) and pediatric subjects sensitized but tolerant to hazelnuts (HS patients, no. = 10) were retrospectively recruited from the Paediatric Allergy Unit, Regina Margherita Children’s Hospital (Città della Salute e della Scienza, Turin, Italy). Hazelnut allergy was defined as a convincing history of allergic symptoms (oral allergy syndrome (OAS), urticaria, angioedema, respiratory, cardiovascular, and/or gastrointestinal symptoms) appearing within 1 h after hazelnut ingestion. The allergy diagnosis was confirmed by means of an oral food challenge (OFC) (except for patients with severe anaphylaxis), by prick by prick (PbP), and by hazelnut sIgE dosed with ImmunoCAP (Thermo Fisher Scientific, Uppsala, Sweden) (Table 1).

TABLE 1

Clinical characterization of the patient cohort (N=37)

Panel A: first patient recruitment
ID	Patient classification	Sex	Age (y)	FHA	AD	FA	PA	NA	Hazelnut PbP (ø mm)		IgE ImmunoCAP (KU/l)						Roasted OFC Symptoms	OFC symptoms Severity grading	Reduced immunoblotting to RAW hazelnut OB		Reduced immunoblotting to ROASTED hazelnut OB
									Raw	Roasted	Hazelnut	rCor a 1	rCor a 8	rCor a 9	rCor a 14	rBet v 1			14‐kDa band	17‐kDa band	14‐kDa band	17‐kDa band
1	HA	M	15	Yes	Yes	Yes	No	No	6	6	75.80	58.80	1.02	19.20	67.80	55.80	OS, NS, V, ER, AP, LE, H	Severe	No	No	Yes	Yes
2	HA	M	5	Yes	Yes	No	Yes	No	13	5	13.70	0.10	0.25	9.80	4.50	<0.10	OAS, LAE	Mild	Yes	Yes	Yes	Yes
3	HA	M	11	Yes	No	Yes	No	No	12	6	51.70	72.30	<0.10	4.95	5.43	<0.10	OAS, PI, AP	Moderate	No	Yes	No	Yes
4	HA	F	7	Yes	Yes	Yes	Yes	Yes	10	10	28.90	<0.10	<0.10	27.40	39.30	<0.10	V, OAS, PI, ER	Moderate	No	No	Yes	Yes
5	HA	M	8	Yes	Yes	Yes	No	Yes	9	8	73.00	64.30	19.20	11.20	26.10	<0.10	V, AS, LE	Severe	No	No	Yes	Yes
6	HA	F	12	Yes	Yes	Yes	Yes	Yes	9	8	84.20	<0.10	<0.10	73.40	41.20	n.d.	AN	Severe	No	No	Yes	Yes
7	HA	M	7	Yes	No	Yes	Yes	Yes	6	4	55.90	<0.10	0.54	39.30	4.16	<0.10	V, AP, ER	Moderate	Yes	No	Yes	Yes
8	HA	F	7	Yes	Yes	Yes	No	Yes	5	5	9.55	<0.10	<0.10	19.90	3.94	<0.10	OAS, AP, NS, C, H	Severe	Yes	No	Yes	Yes
9	HA	M	8	Yes	Yes	Yes	Yes	Yes	8	5	16.60	11.50	<0.10	6.85	4.76	9.80	OAS, C, ER	Moderate	Yes	Yes	Yes	Yes
10	HA	M	12	Yes	No	No	Yes	Yes	10	14	>100	29.10	0.11	>100	>100	36.50	V, AP, U, LE, AS	Severe	No	Yes	Yes	Yes
11	HA	M	11	No	No	Yes	Yes	No	3	0	30.70	0.42	36.80	0.69	0.27	4.44	ER	Mild	No	Yes	No	Yes
12	HA	F	16	Yes	Yes	Yes	No	Yes	10	8	41.90	<0.10	<0.10	34.70	19.90	<0.10	OAS	Mild	Yes	Yes	Yes	Yes
13	HA	M	10	Yes	Yes	Yes	Yes	Yes	10	8	7.45	4.33	<0.10	1.92	1.49	4.40	OAS, PI, ER	Mild	No	Yes	Yes	Yes
14	HA	M	6	Yes	Yes	No	No	No	6	7.5	3.75	<0.10	<0.10	<0.10	4.67	<0.10	ER, U, AP, LAE	Moderate	No	Yes	Yes	Yes
15	HA	M	10	Yes	Yes	No	Yes	Yes	7	6.5	38.20	59.70	0.75	0.22	<0.10	>100	U, PI	Mild	No	Yes	No	Yes
16	HA	M	10	Yes	Yes	No	Yes	Yes	8	4.5	69.00	0.31	>100	29.10	81.50	0.52	U, PI	Mild	No	Yes	No	Yes
17	HA	M	14	Yes	Yes	Yes	Yes	Yes	7	8	>100	17.40	2.30	56.90	74.50	57.60	U, PI	Mild	No	Yes	Yes	Yes
18	HA	M	4	Yes	Yes	No	No	No	5	15	20.10	<0.10	0.73	8.57	26.30	<0.10	AS, C, R	Severe	No	Yes	Yes	Yes
19	HA	F	13	n.d.	Yes	Yes	Yes	Yes	11	7.5	>100	0.12	<0.10	>100	>100	<0.10	U	Mild	No	Yes	No	Yes
20	HA	M	14	Yes	Yes	Yes	Yes	Yes	10	7	3.62	<0.10	0.49	9.18	4.94	<0.10	LE, AP	Severe	No	Yes	No	Yes
21	HA	F	9	Yes	Yes	Yes	Yes	No	3	6	48.00	2.73	29.80	9.68	34.30	n.d.	OAS	Mild	Yes	Yes	Yes	Yes
22	HA	M	2	Yes	Yes	Yes	No	Yes	3	0	2.75	0.40	<0.10	n.d.	4.98	<0.10	LAE, U, LE	Severe	No	No	No	No
23	LIHA	M	11	Yes	No	Yes	No	Yes	5	5	0.76	<0.10	1.78	0.13	0.47	<0.10	OAS, PI, H	Severe	No	Yes	No	Yes
24	LIHA	M	12	Yes	Yes	Yes	Yes	Yes	12	7	1.29	0.55	0.13	<0.10	<0.10	<0.10	LAE, A, U	Mild	No	Yes	Yes	Yes
25	LIHA	M	16	Yes	Yes	Yes	Yes	No	7	5	0.71	<0.10	<0.10	0.36	0.76	<0.10	LAE	Mild	No	Yes	Yes	Yes
26	LIHA	M	5	n.d.	n.d.	Yes	No	No	5	6	1.59	1.88	<0.10	<0.10	0.70	4.98	LAE	Mild	No	Yes	No	Yes
27	LIHA	F	5	n.d.	n.d.	No	Yes	Yes	5	0	0.45	<0.10	<0.10	1.13	0.19	<0.10	OS, U, C	Moderate	No	Yes	No	Yes
28	HS	F	7	Yes	Yes	Yes	No	Yes	3	5	7.76	<0.10	6.50	0.32	<0.10	<0.10	None		No	Yes	Yes	Yes
29	HS	M	6	Yes	Yes	Yes	No	No	5	5	24.70	38.10	0.85	<0.10	<0.10	<0.10	None		Yes	Yes	Yes	Yes
30	HS	F	6	No	Yes	No	Yes	Yes	4.5	0	78.00	>100	1.37	1.74	0.69	>100	None		No	Yes	Yes	Yes
31	HS	M	17	Yes	Yes	Yes	Yes	No	10	10	7.42	7.53	<0.10	0.71	0.14	15.80	None		No	Yes	Yes	Yes
32	HS	F	12	Yes	No	Yes	Yes	No	4.5	0	47.40	58.70	<0.10	0.18	<0.10	67.30	None		No	Yes	Yes	Yes
33	HS	M	11	Yes	Yes	Yes	Yes	Yes	3	0	0.65	<0.10	<0.10	<0.10	0.30	<0.10	None		No	Yes	Yes	Yes
34	HS	M	11	Yes	No	No	Yes	Yes	5	0	15.20	<0.10	14.60	<0.10	<0.10	<0.10	None		No	Yes	Yes	Yes
35	HS	M	9	Yes	Yes	Yes	No	Yes	5	4	6.44	<0.10	0.18	1.00	0.91	<0.10	None		No	Yes	Yes	Yes
36	HS	M	7	Yes	No	Yes	No	No	12	9	6.14	<0.10	<0.10	0.65	1.37	<0.10	None		No	No	No	Yes
37	HS	M	7	Yes	Yes	No	Yes	Yes	5	0	54.90	<0.10	<0.10	79.00	6.40	n.d.	None		No	Yes	no	Yes

Abbreviations: A, angioedema; AD, atopic dermatitis; AN, anaphylaxis; AP, abdominal pains; AS, asthma; C, cough; ER, erythema; F, female; FA, other food allergy; FHA, family history of atopy; H, hoarseness; HA, hazelnut‐allergic patient; HS, hazelnut‐sensitized patient; HY, hypotension; LAE, labial edema; LE, laryngeal edema; LIHA, low ImmunoCAP hazelnut‐allergic patient; M, male; NA, other nut allergy; NS, nasal symptoms; OAS, oral allergy syndrome; OS, ocular symptoms; PA, pollen allergy; PI, pharyngeal itching; R, respiratory problems; U, urticaria; V, vomiting.

All the HA patients reported having had a reaction to foods containing roasted hazelnuts (cocoa, hazelnut spreads, chocolate, hazelnut ice cream, chocolate pralines with whole roasted hazelnuts, or hazelnut nougat). For this reason, the OFC was performed with finely ground roasted hazelnuts.

The OFC was initiated using one drop of physiological solution mixed with the hazelnut powder and placing the solution on the inner side of the lower lip. The challenge was continued with increasing doses of hazelnut powder (0.2 g, 0.4 g, 1 g, 2 g, 3 g, 6 g, 8 g, and 10 g) administered in 20‐min intervals, unless symptoms occurred or until a total cumulative dose of 30 g of hazelnut powder (equivalent to 15 hazelnuts). The OFC procedure was stopped after observation of the first objective symptom. In the case of mild objective or subjective symptoms, it was decided to wait longer for the subsequent dose and/or to repeat the same dose, and the OFC procedure was stopped in the case of persistent symptoms (lasting longer than 30 minutes) and/or worsening symptoms and in the case of repeated elicitation (over 3 doses). The scoring of the severity of the reaction to the OFC was evaluated independently by two allergy specialists (G.M. and R.C.), and the results were coded as mild, moderate, or severe, according to the Integrated Grading System for Scoring Reaction Severity of Oral Food Challenge Reactions of Zhu et al.

All the patients underwent PbP, using both raw and roasted hazelnuts. Histamine (10 mg/ml) was used as a positive control, and a saline solution was used as the negative control. PbP was considered positive if the wheal was higher than 3 mm, and there was no reaction to the negative control.

The IgEs sera were assayed against a hazelnut extract, rCor a 1, rCor a 8, rCor a 9, rCor a 14, and rBet v 1 using ImmunoCAP. The cutoff for positivity was set at 0.10 KU/l. rCor a 1, rCor a 8, rCor a 9, and rCor a 14 sIgE concentrations were standardized for the statistical analysis by subtracting the mean values of each variable and then dividing by the corresponding standard deviations. The sIgE values below the limit of detection (<0.10 KU/L) were assigned a value of 0.01; the sIgE values above the limit of detection (>100 KU/L) were assigned a value of 100. Missing data (codified in Table 1 with n.d.) were not considered in the statistical analysis. The standardized data of the patients classified as HA versus the patients classified as HS were analyzed by means of nonparametric Kruskal‐Wallis one‐way ANOVA, because of the non‐normal distribution. The correlation between the standardized sIgE concentrations of the different allergens was calculated by means of Kendall’s tau. All the analyses were performed using PAST software, version 2.17.

A subgroup of the HA patients (patient nos. 23, 24, 25, 26, and 27), who were characterized by low rCor a 1, rCor a 8, rCor a 9, and rCor a 14 ImmunoCAP IgE values, ranging from 0.10 kU/L to 1.88 kU/L, were defined as low ImmunoCAP hazelnut‐allergic (LIHA) patients.

The HS patients were characterized according to the presence of atopic dermatitis (AD) and/or other food allergy (FA) and/or other nut allergies (NA), with a positive PbP to hazelnut and/or positive sIgE to hazelnut, but negative to OFC.

The experimental controls were the sera of non‐nut‐allergic pediatric subjects (NNA) and of non‐food‐allergic pediatric subjects (NFA), used individually or pooled.

The study was approved by the local ethics committee of the Città della Salute e della Scienza (Turin, Italy) (approval no. 312, prot. no. 22050). The parents of all the patients gave written informed consent.

Hazelnut water‐soluble and OB‐associated protein extraction

Raw hazelnuts (Corylus avellana cv. Tonda Gentile delle Langhe, TGL) were obtained from experimental fields, roasted at a laboratory scale (120°C for 30 min), and used for the OFC, PbP, and immunoblotting experiments. The water‐soluble proteins were extracted according to Platteau et al. The hazelnut OBs were isolated according to Cao et al. and the OB‐associated proteins were extracted according to Lamberti et al., 2020.

Protein separation by means of LDS‐PAGE, 2‐dimensional electrophoresis (2‐DE), and immunoblotting analysis

The hazelnut water‐soluble proteins were separated by means of LDS‐PAGE. The OB‐associated proteins were separated by means of both LDS‐PAGE and 2‐DE. Protein bands and spots were electro‐transferred to PVDF membranes and incubated with the patient’s serum. Immunoblotting was developed using an Alkaline Phosphatase Substrate Kit (Bio‐Rad).

Use of sera

Individual serum was used for 1D immunoblotting of the OB protein extracts, ELISA, and an immunoblotting inhibition assay with a raw OB protein extract. Pooled sera were used for the remaining experiments: 2D immunoblotting of the OB protein extracts, an immunoblotting inhibition assay with rCor a 15, and immunoblotting of the total water‐soluble hazelnut protein extracts, due to the paucity of the available sera.

Protein identification by means of MALDI‐TOF/TOF and ESI‐Q‐TOF analyses

In‐gel digestion was performed to identify the allergens contained in the bands and spots from the gel electrophoresis, as in Nebbia et al., and protein identification was performed by coupling the MALDI‐TOF/TOF and ESI‐Q‐TOF‐MS techniques. The spectra were searched against the NCBI C. avellana database implemented with two new oleosin sequences derived from both the C. avellana cv Jefferson genome and C. avellana cv TGL.

Expression and purification of recombinant Cor a 15

In order to establish whether the newly discovered oleosin was an allergen, it was produced as a recombinant protein. Its complete coding sequence, obtained by means of RNA retrotranscription, was cloned in a pGEM‐T Easy vector (Promega, Madison, WI, The USA) and sent to PRIMM (Milan, Italy) for protein expression and purification. The full‐length allergen (rCor a 15) and the C‐terminal fragment (rCor a 15 C‐term, from aa 102 to aa 169) were used for the blot inhibition studies, and only the full‐length rCor a 15 was used for ELISA.

Quantification of Cor a 15‐specific IgEs by means of direct ELISA test

IgEs specific to rCor a 15 were quantified as described in Hilger et al., with some modifications. Briefly, patient sera (10 HA, 3 LIHA, 2 HS, 9 NNA, and 9 NFA patients) were transferred to 96‐well plates previously coated with 0.1 µg/ml of rCor a 15. In order to obtain a semi‐quantification of hazelnut‐allergic patient IgE against rCor a 15, the serum of a patient with a known titer of IgE specific to α‐lactalbumin (as determined by ImmunoCAP) was used as the reference standard. The reference serum was chosen since it contained IgE that was completely unrelated to the allergy that was the subject of this study. The reference serum was serially diluted in duplicate and added to α‐lactalbumin‐coated wells. The obtained standard curve allowed us to calculate the IgE titers specific to rCor a 15 for each patient’s serum.

RESULTS

A flowchart of the experiments and a summary of the main results are given in Figure 1.

FIGURE 1

Workflow of the experiments

Study population

The main characteristics of the pediatric subjects included in the study are summarized in Table 1. All the patients tested with PbP resulted positive to raw hazelnuts: 24/27 HA patients and 6/10 HS patients showed a wheal diameter ≥ 5 mm. Three HA patients out of 27 and 5/10 HS patients showed negative PbP results when tested for roasted hazelnuts.

The median scores of the sIgE of rCor a 1, rCor a 8, rCor a 9, and rCor a 14 for the HA patients were 0.31 (IQR 0.01–11.50), 0.11 (IQR 0.01–1.02), 9.18 (IQR 0.36–29.10), and 4.94 (IQR 0.76–39.30), respectively. The median scores of the sIgE of rCor a 1, rCor a 8, rCor a 9, and rCor a 14 for the HS patients were 0.01 (IQR 0.01–38.10), 0.10 (IQR 0.01–1.37), 0.49 (IQR 0.01–1.00), and 0.22 (IQR 0.01–0.91), respectively. The HA and HS patients differed significantly as far as the rCor a 9 and rCor a 14 values are concerned (pCora9 = 0.004, pCora14 = 0.002), while Cor a 1 and Cor a 8 did not discriminate the HA from the HS patients (pCora1 = 0.75, pCora8=0.98). Furthermore, the rCor a 9 and rCor a 14 values were directly correlated (r2 = 0.67; p < .001). Seventeen patients (12/27 HA patients and 5/10 HS patients) also suffered from a pollen allergy (PA). Finally, 17/27 HA patients and 6/10 HS patients were allergic to other tree nuts (NA). It was not possible to establish a correlation between the severity grading and oleosin immunorecognition because of the limited number of LIHA patients.

Discovery of two new hazelnut oleosins

The OB‐associated proteins were separated (Figures 2 and 3) and identified (Table 2, Table S1, and Table S2), and two new oleosin‐like proteins were found (Figure S1). Their sequences were submitted to GenBank under accession numbers MK737923 and MK737924. The MK737923 sequence showed 56% similarity with Cor a 12 and 41% with Cor a 13. The MK737924 sequence showed 63% similarity with Cor a 12 and 76% with Cor a 13 (Figure S2).

FIGURE 2

For each panel: first lane, stained LDS‐PAGE of raw (A) and roasted (B) OB‐associated protein extracts; lanes 1 to 37: immunoblotting with individual patient’s serum. C, a pool of five sera of non‐nut‐allergic (NNA) pediatric subjects was used as a control. CII, secondary antibody control. MW, molecular weight; HA, hazelnut‐allergic patients; LIHA, low ImmunoCAP hazelnut‐allergic patients; HS, hazelnut‐sensitized patients

FIGURE 3

2‐DE and immunoblotting of the OB‐associated protein extracts. 2‐DE of raw (A) and roasted (B) hazelnut OB protein extracts. Immunoblotting of raw (C) and roasted (D) hazelnut OB protein extracts with the pool of sera of the LIHA patients (nos. 23, 24, and 25). Immunoblotting of raw (E) and roasted (F) hazelnut OB protein extracts with a pool of five sera of non‐nut allergic (NNA) pediatric subjects. MW, molecular weight; pI, isoelectric point; NL, non‐linear

TABLE 2

1‐DE and 2‐DE protein identification

No band	Entry (NCBI)	Protein name	MWexp/MWtheor	Protein score	No of matching peptides	Protein coverage (%)	Molar fraction (%)
1‐DE protein identification
Raw hazelnut
1A	AAO65960	Cor a 13	14,000/14,723	776	7	36.6	78.4
	AAO67349	Cor a 12	14,000/16,745	455	4	26.9	21.6
2A	MK737923	Cor a 15	17,000/17,695	1045	7	45.9	100
Roasted hazelnut
1B	AAO65960	Cor a 13	14,000/14,723	1231	6	38.9	64.4
	AAO67349	Cor a 12	14,000/16,745	624	6	34.8	35.6
2B	MK737923	Cor a 15	17,000/17,695	1080	8	45.9	89.5
	AAO67349	Cor a 12	17,000/16,745	140	3	21.7	10.5
3B	AHA36627	Cor a 9	22,000/59,200	1853	14	31.2	69.5
	AAO65960	Cor a 13	22,000/14,723	203	4	26.9	30.6
4B	AHA36627	Cor a 9	23,000/59,200	2268	17	28.4	100
5B	AAO65960	Cor a 13	27,000/14,723	456	4	29.1	47
	AAO67349	Cor a 12	27,000/16,745	378	4	26.9	38.3
	AHA36627	Cor a 9	27,000/59,200	433	7	16.2	14.7
6B	MK737923	Cor a 15	29,000/17,695	293	5	36.6	40.6
	AAO65960	Cor a 13	29,000/14,723	401	4	29.1	27.2
	AAO67349	Cor a 12	29,000/16,745	410	4	26.9	22.2
	AHA36627	Cor a 9	29,000/59,200	539	8	18.6	10.1
7B	MK737923	Cor a 15	34,000/17,695	78	4	36.6	53.6
	AHA36627	Cor a 9	34,000/59,200	181	3	21.7	46.4
8B	AHA36627	Cor a 9	38,000/59,200	524	8	19.9	68.8
	AAO67349	Cor a 12	38,000/16,745	178	4	26.9	31.2

IgE reactivity to hazelnut OB‐associated proteins and discovery of Cor a 15

We hypothesized that other allergens, not included in the ImmunoCAP assay, could account for the positivity of the LIHA patients to OFC. We thus performed immunoblotting to evaluate the immunoreactivity of the single patients to the OB‐associated proteins. Two oleosin‐containing bands were found at 14 and 17 kDa (Figure 2).

One hundred percent of the allergic patients recognized at least one oleosin‐containing band (Table 1). Thirty‐three percent recognized the 14‐kDa band, containing both Cor a 12 and Cor a 13 (band 1A in Figure 2, Table 2). The 17‐kDa band, which exclusively contains the MK737923 oleosin (band 2A in Figure 2 and Table 2), was the most recognized one and was immunoreactive for 88% of the allergic patients. The MK737923 oleosin was therefore submitted as an allergen candidate to the WHO/IUIS Allergen Nomenclature Subcommittee, who approved it as a new hazelnut allergen and termed it Cor a 15. The three hazelnut oleosins were also immunorecognized by the HS patients, with a similar frequency to that of the allergic patients (Table 1 and Figure 2).

In order to verify that the LIHA patients did not recognize any other water‐soluble allergens than those found negative by ImmunoCAP, a total hazelnut water‐soluble protein extract was tested by immunoblotting against the pooled sera of the LIHA patients (nos. 23, 24, and 25) and against the pooled sera of the HA patients (nos. 1 to 14). The IgEs of the LIHA patients did not bind with any protein band, while several protein bands were recognized by the HA pool (Figure S3).

Effect of roasting on the hazelnut oleosin profile

We further assessed whether the hazelnut OB protein profile was affected by processing, since hazelnuts are usually consumed roasted. In addition to the 14‐kDa and 17‐kDa bands, higher MW bands appeared (Figure 2, panel B). The 14‐kDa band, containing a mixture of Cor a 12 and Cor a 13 (band 1B in Figure 2 and Table 2), was recognized by 62% of the allergic patients (Table 1). The 17‐kDa band, containing Cor a 15 with traces of Cor a 12 (band 2B in Figure 2 and Table 2), was again the most recognized band (96% of the allergic children) (Table 1). The bands around 20–40 kDa, generated by roasting, were recognized by almost all the patients. In addition to oleosins, these bands also contained Cor a 9 and/or Cor a 11 soluble hazelnut allergens (bands 3B to 8B in Figure 2, Table 2). To discriminate whether the LIHA patients’ sera immunoreacted to Cor a 9 and Cor a 11 or against oleosins, a 2DE separation of the OB‐associated proteins was performed, exploiting the different isoelectric points of these allergens (Figure 3, panels A and B). Most of the spots in the alkaline zone (right part) of the 2DE map were identified as oleosins, or as the basic subunit of Cor a 9, while those in the acidic zone were found to contain acidic subunits of Cor a 9 and Cor a 11 (Table 2).

In spite of the different 2DE protein profiles of the raw and roasted OB proteins (Figure 3, Panel A and B), the immunoreactivity profiles were very similar (Figure 3, Panels C and D) and exclusively localized in the alkaline region.

Hazelnut oleosins are major allergens for LIHA patients and Cor a 15 is the most frequently recognized oleosin

Most reactive spots on 2DE separation of raw and roasted extracts contained co‐migrating oleosins exclusively: Cor a 12, Cor a 13, and Cor a 15 (Figure 3, Table 2). Interestingly, the 1A2DE and 2A2DE spots, which showed a similar densitometric intensity in the 2DE gel, resulted in a highly different immunoreactivity. Spot 2A2DE, which contained 96.5% of Cor a 15, was much more immunoreactive than the 1A2DE spot, which exclusively contained Cor a 12 and 13 (Figure 3, Panels A and C, and Table 2).

In order to verify whether Cor a 15 was the most frequently recognized oleosin, full‐length Cor a 15 and its C‐term fragment (from aa 102 to aa 168) were expressed and purified (Figure 4, panels A and C). The C‐terminal fragment was expressed because it contains a similar peptide (11 aa out of 14 aa) to the most immunoreactive IgE epitope in Ara h 15 (Figure S4). The addition of full‐length rCor a 15 to both the LIHA and HA pooled patients’ sera completely inhibited the recognition of the 17‐kDa band exclusively containing Cor a 15 (Figure 4, panel B). The same amount of rCor a 15 C‐terminal did not result in the complete inhibition of the immunoreactivity of the sera (Figure 4, panel D). Moreover, the reverse was also true: the hazelnut oleosins present in the OB extract were able to inhibit the Cor a 15 immunorecognition by the patients’ sera (Figure 4, panel E).

FIGURE 4

Immunoblotting inhibition assay. Panels A and C: LDS‐PAGE of full‐length rCor a 15 (rCor a 15) and rCor a 15 C‐terminal (rC‐term), respectively. Immunoblotting with pooled sera of the LIHA patients (25, 26 and 27) and HA patients (2, 10, and 13). CII: secondary antibody control; MW: molecular weight. Panels B and D: LDS‐PAGE of raw hazelnut OB protein extracts (raw OB). Immunoblotting with the pooled sera of the LIHA patients (25, 26, and 27) and HA patients (2, 10, and 13), without (Tq) and with the addition of 0.2 and 2 µg of rCor a 15 and rCor a 15 C‐terminal to the patients’ sera, respectively. Panel E: Immunoblotting with the sera of 10 patients (nos. 4, 5, 9, 16, 17, 18, 19, 27, 36, and 37) without (Tq) and with the addition of 3 and 6 µg of raw OB protein extract to the patients’ sera. C1: pool of five sera of non‐nut‐allergic (NNA) pediatric subjects; C2: pool of five sera of non‐food‐allergic (NFA) pediatric subjects

All the patients showed specific IgEs toward Cor a 15

In order to obtain a semi‐quantification of the patients’ IgEs that were specific to Cor a 15 in the different subgroups of patients (HA, LIHA, and HS) and in the negative controls (NNA and NFA), the patients’ sera were tested by means of an in‐house direct ELISA test. The patients’ sera all showed specific IgEs for rCor a 15 (Figure 5); however, the ELISA test did not differentiate between the HA, LIHA, and HS patients on the basis of Cor a 15 immunorecognition, as already demonstrated by the immunoblotting experiments. An aspecific reaction was found in 3/18 of the control sera.

FIGURE 5

ELISA quantification of Cor a 15 IgEs in three subgroups of patients’ sera (10 HA, 3 LIHA, and 2 HS). A line marks the mean kUA/L for each subgroup of patients. NNA control: 10 sera of non‐nut allergic pediatric subjects; NFA control: 10 sera of non‐food‐allergic pediatric subjects

DISCUSSION

With the aim of conducting an immune‐allergological characterization of hazelnut oleosins in pediatric patients, we performed the isolation and identification of hazelnut OB‐associated proteins. We were able to separate oleosins from soluble allergens and to identify, for the first time, a new allergen belonging to the oleosin family, termed Cor a 15 by WHO‐IUIS. We then demonstrated the immunogenic power of the newly reported Cor a 15 oleosin for most of our hazelnut‐allergic pediatric patients. In addition, we succeeded in unambiguously ascribing the immunoreactivity of LIHA patients (characterized by low ImmunoCAP assay IgEs values and positive OFC and PbP) only to oleosins, thereby excluding the contribution of other allergens.

Regarding the IgE levels, the Cor a 9 and Cor a 14 IgE levels were clearly associated with the elicitation of the allergic reaction, according to Datema et al., and were useful to discriminate between the hazelnut‐allergic and hazelnut‐sensitized but tolerant patient groups. This was not found for the Cor a 1 and Cor a 8 IgE levels, for which a moderate diagnostic predictive value has been demonstrated. The IgE levels of the newly discovered Cor a 15 seemed to have a similar behavior, as they were unable to discriminate between allergic and sensitized patients.

The evaluation of oleosin immunoreactivity for hazelnut‐allergic patients was assessed by 1D and 2D immunoblotting, coupled with protein identification by means of mass spectrometry. The results indicated that all the HA patients immunorecognized at least one oleosin, although it was not possible to assess the role of the oleosins in eliciting the allergic reaction, as they were co‐sensitized to genuine hazelnut allergens. On the contrary, we found that a subset of LIHA patients exclusively immunorecognized oleosins and, among them, mainly Cor a 15. Cor a 15 appeared to be the best candidate for the allergic reaction in this subgroup of patients. This evidence highlights the relevance of dosing IgEs against Cor a 15 in hazelnut allergy diagnosis. We also observed IgE reactivity to hazelnut oleosins in the HS patients, which could be ascribed to a cross‐reactivity with other oleosins, at least for those patients with a confirmed allergy to other nuts.

Cor a 15, which we exclusively identified in the 17‐kDa band, was the most immunorecognized oleosin (88% of the allergic patients). This result was different from the results of Zuidmeer‐Jongejan, who detected both Cor a 12 and Cor a 13 in a band at the same MW. This is probably due to the lack of Cor a 15 gene sequence in the database at the time of the protein identification. We also identified Cor a 12 and 13 in a band at 14 KDa, with a recognition frequency of 33%, similar to what was reported in the EuroPrevall project for Cor a 12 in children (34%).

As already demonstrated by Schwager et al. for Ara h 15, we hypothesized that the epitope responsible the most for Cor a 15 immunorecognition is located at the C‐terminal, since both the full‐length rCor a 15 and rCor a 15 C‐terminal were able to inhibit IgEs against the 17‐kDa band containing only Cor a 15.

Our findings also show a resistance of oleosins to high temperatures, which resulted in an apparent enhancement of oleosin immunoreactivity after roasting, with a more evident effect on Cor a 12 and Cor a 13. The immunorecognition of roasted Cor a 12 and Cor a 13 exceeded 60% for the allergic patients, while Cor a 15 reached 96%. Similarly, Schwager et al. demonstrated an increase in oleosin immunoreactivity after roasting in‐shell peanuts. Glycation after roasting may play a role in increasing the affinity and/or the accessibility of oleosins to IgEs.

In conclusion, we have discovered a new hazelnut oleosin allergen, Cor a 15, which resulted to play a role in eliciting an allergic reaction in a subgroup of our pediatric patients that exclusively immunorecognized oleosins, thus confirming the relevance of including lipophilic allergens in the protein extracts used for diagnostic tests.

CONFLICT OF INTEREST

The authors have no conflict of interest.

AUTHOR CONTRIBUTION

Stefano Nebbia: Data curation (equal); Formal analysis (equal); Investigation (lead); Writing‐original draft (equal). Cristina Lamberti: Investigation (lead); Methodology (lead); Supervision (equal); Writing‐original draft (equal); Writing‐review & editing (equal). Simona Cirrincione: Data curation (equal); Investigation (equal); Methodology (equal). Alberto Acquadro: Investigation (equal); Writing‐review & editing (equal). Simona Abbà: Investigation (equal); Resources (equal). Marina Ciuffo: Investigation (equal); Writing‐review & editing (equal). Daniela Torello Marinoni: Resources (equal); Writing‐review & editing (equal). Marcello Manfredi: Investigation (equal); Writing‐review & editing (equal). Emilio Marengo: Investigation (equal); Writing‐review & editing (equal). Roberta Calzedda: Resources (equal); Writing‐review & editing (equal). Giovanna none Monti: Conceptualization (equal); Methodology (equal); Resources (lead); Writing‐review & editing (equal). laura cavallarin: Conceptualization (lead); Methodology (lead); Project administration (lead); Supervision (equal); Writing‐original draft (equal); Writing‐review & editing (lead). Maria Gabriella Giuffrida: Conceptualization (lead); Funding acquisition (equal); Methodology (lead); Supervision (lead); Writing‐original draft (equal); Writing‐review & editing (equal).

PEER REVIEW

The peer review history for this article is available at https://publons.com/publon/10.1111/pai.13579.

Fig S1

Click here for additional data file.

Fig S2

Click here for additional data file.

Fig S3

Click here for additional data file.

Fig S4

Click here for additional data file.

Table S1

Click here for additional data file.

Table S2

Click here for additional data file.