Defining the cross-reactivity between peanut allergens Ara h 2 and Ara h 6 using monoclonal antibodies

Abstract

In peanut allergy, Arachis hypogaea 2 (Ara h 2) and Arachis hypogaea 6 (Ara h 6) are two clinically relevant peanut allergens with known structural and sequence homology and demonstrated cross-reactivity. We have previously utilized X-ray crystallography and epitope binning to define the epitopes on Ara h 2. We aimed to quantitatively characterize the cross-reactivity between Ara h 2 and Ara h 6 on a molecular level using human monoclonal antibodies (mAbs) and structural characterization of allergenic epitopes. We utilized mAbs cloned from Ara h 2 positive single B cells isolated from peanut-allergic, oral immunotherapy–treated patients to quantitatively analyze cross-reactivity between recombinant Ara h 2 (rAra h 2) and Ara h 6 (rAra h 6) proteins using biolayer interferometry and indirect inhibitory ELISA. Molecular dynamics simulations assessed time-dependent motions and interactions in the antibody–antigen complexes. Three epitopes—conformational epitopes 1.1 and 3, and the sequential epitope KRELRNL/KRELMNL—are conserved between Ara h 2 and Ara h 6, while two more conformational and three sequential epitopes are not. Overall, mAb affinity was significantly lower to rAra h 6 than it was to rAra h 2. This difference in affinity was primarily due to increased dissociation of the antibodies from rAra h 6, a phenomenon explained by the higher conformational flexibility of the Ara h 6–antibody complexes in comparison to Ara h 2–antibody complexes. Our results further elucidate the cross-reactivity of peanut 2S albumins on a molecular level and support the clinical immunodominance of Ara h 2.

We utilized monoclonal antibodies cloned from peanut-allergic, oral immunotherapy–treated patients to probe the cross-reactivity between clinically relevant peanut allergens Ara h 2 and Ara h 6 on a molecular level.

Graphical Abstract

Graphical Abstract.

Introduction

Cross-reactivity, whereby allergen-specific IgE antibodies are able to bind to more than one allergen due to the structural similarities between them, occurs in many clinically relevant aspects of allergic disease, including oral allergy syndrome and multiple tree nut allergies. The cross-reactivity between allergens can result in a wide range of clinical manifestations, from isolated oropharyngeal pruritus to anaphylaxis [1]. Structural insights into the similarities and differences between the allergens inciting these reactions have allowed us to clinically prognosticate, manage, and treat a wide range of allergic symptoms. However, we continue to have an incomplete characterization of cross-reactive antibody–antigen interactions on a molecular level.

IgE-mediated peanut allergy is a severe, persistent condition affecting up to 10% of the population [2–5] with two immunodominant allergens Arachis hypogaea 2 (Ara h 2) and Arachis hypogaea 6 (Ara h 6). The immunodominant and most clinically relevant peanut allergen, Ara h 2, is a 18 kDa 2S albumin protein [6]. It is approximately 59% homologous with another clinically relevant 2S albumin seed storage peanut allergen with a molecular weight of 15 kDa, Ara h 6 [6, 7]. While Ara h 6 IgE levels are less predictive of clinical peanut allergy than Ara h 2 IgE levels, Ara h 6 is the second most clinically relevant allergen after Ara h 2 [8]. In fact, due to the structural and biochemical similarities between Ara h 2 and Ara h 6, it is remarkably difficult to fully separate the two natural proteins [9], as it is difficult to separate other components of peanut [10, 11]. Hence, many studies may have inadvertently utilized a mixture of both allergens.

While serum-specific IgE levels to Ara h 2 and Ara h 6 have a high correlation in multiple cohorts, in vitro assays with peanut-allergic serum have suggested varying degrees of cross-reactivity between Ara h 2 and Ara h 6 [9, 12–15]. More recently, a large panel of 69 monoclonal antibodies (mAbs) cloned from patients were analyzed for epitope specificity to Ara h 2. This work identified multiple conformational epitopes that occupy nonoverlapping regions of Ara h 2. Of these, the epitope that recognized alpha-helix 2 of Ara h 2 is known to bind to Ara h 2 and not Ara h 6. Two of the conformational epitopes, epitopes 1.2 and 3, have been characterized using neutralizing antibodies capable of disrupting multiple antibody–allergen interactions. These antibodies binding to conformational epitopes on Ara h 2 interact with amino acid residues also relevant for binding to sequential epitopes, including the KRELRNL sequential epitope [16].

This presents an opportunity to analyze the epitopes of Ara h 6, precisely characterize cross-reactive antibodies, and elucidate how the molecular allergen–antibody interactions impact the clinical relevance of these seed storage proteins. Given the previous ambiguity surrounding the extent of cross-reactivity, as well as the demonstrated difficulty in using natural proteins, we utilized recombinant Ara h 2 and Ara h 6 proteins to probe cross-reactivity using mAbs from peanut-allergic patients.

Materials and methods

Serum-specific IgE levels

To compare serum-specific IgE levels to Ara h 2 and Ara h 6 in peanut-sensitized patients seen within the healthcare system, we utilized electronic medical record data through the research patient data registry (RPDR) with Institutional Review Board approval (IRB 2021P003078). From the RPDR, a centralized data registry that stores clinical information from various hospitals in the Mass General Brigham (MGB) system, we identified 741 peanut-sensitized individuals with component testing entered in the system for analysis of their de-identified data in R using parseRPDR [17].

Clinical trials

Recombinant antibodies were cloned from allergen-specific B cells isolated from the peripheral blood of subjects undergoing oral immunotherapy from two previously published, IRB-approved clinical trials of peanut oral immunotherapy (NCT01324401, NCT01750879) [16, 18].

Recombinant Ara h 2-specific mAbs

Recombinant antibody cloning

Allergen-specific B cells were affinity selected using a a fluorescent natural Ara h 2-Alexa Flour 488 multimer, as previously described [16, 18], for subsequent sequencing and cloning of paired heavy and light chains into IgG1 heavy chain and kappa/lambda light chain vectors, as previously described [16].

Recombinant antibody production

Both HEK 293T adherent cells and Expi293 suspension cells were used for the expression of mAbs, according to manufacturer protocols, using transfection with jetPRIME versatile DNA/siRNA transfection reagent (Polyplus) and ExpiFectamine Transfection Reagent (ThermoFisher). Two days after transfection for the HEK 293T adherent cells, or 7 days after transfection for the Expi293 suspension cells, supernatants were harvested for incubation with immobilized protein G agarose beads (Thermo Scientific) and subsequent column-based elution.

Recombinant antibody quantification and validation

Purified mAbs were quantified on an Octet R2 Protein Analysis System (Sartorius) using anti-human Fab-CH1 biosensors (Sartorius) and PBS, as previously described [16].

Recombinant Ara h 2 and Ara h 6 allergens

Production of recombinant allergens

An avitag coding the amino acid sequence GLNDIFEAQKIEWHE (ggactaaatgatatatttgaagcgcagaagatcgaatggcatgaa) was engineered into pMALX(TEV)Arah2 vector containing rAra h 2.0101 (31–160) (sequence from PDB 8DB4) so that the construct expresses avitagMBP(TEV)Arah2. A similar vector was generated for Ara h 6 (9–124; sequence from PDB 1W2Q). These vectors were transformed into E.coli Origami B cells harboring a pACYCDuet-1 plasmid encoding thioredoxin, subsequently grown in Terrific broth, induced with Isopropyl-β-d-thiogalactopyranoside, grown overnight, then lysed by sonication [6]. Concentrated proteins were further purified by gel filtration with a Superdex 200 column that had been equilibrated in: PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 1.8 mM KH₂PO₄) for avitagMBP(TEV)Arah2, 25 mM Tris pH 7.5, 500 mM NaCl for pH6-MBP-TEV-BirA, and 25 mM Tris pH 7.5, 500 mM NaCl, 1 mM EDTA, and 5 mM maltose for avitagMBP(TEV)Arah6. Peak fractions were pooled, concentrated and flash-frozen in liquid nitrogen and stored at −80°C until used.

Biotinylation of proteins

To prepare biotin ligase, pH6-MBP-TEV-BirA (Addgene plasmid # 179694) was transformed and expressed into E. coli BL21 cells. The biotinylation of proteins was performed as described [19]. To remove the biotin ligase, 50% slurry of 100 μL Ni-NTA resin was removed by centrifugation and the protein was dialyzed against PBS to remove excess biotin overnight.

Biolayer interferometry assays

All biolayer interferometry (BLI) assays were run on an Octet R2 Protein Analysis System (Sartorius) at a plate temperature of 30°C with a shaking speed of 1000 rpm. Monoclonal antibodies, biotinylated peptides, and biotinylated Ara h 2 and Ara h 6 were all diluted in kinetics buffer (DPBS + 1% (w/v) BSA + 0.02% Tween 20 (Amresco) into 96-well black plates (Greiner Bio-One)). Sensors were regenerated no more than 15 times through a 30-second cycle alternating between glycine pH 1.5 (Bio-Rad) and kinetics buffer. Data were analyzed using Octet Analysis Studio Software version 12.2 (Sartorius).

Cross-reactivity assays

To identify which Ara h 2-specific antibodies were cross-reactive with Ara h 6, a streptavidin sensor was loaded for 80 seconds with 2 μg/mL recombinant avi-MBP-tev-Ara h 6. The sensor was then associated with each mAb at a concentration of 10 μg/mL for 150 seconds. The antibody was said to be cross-reactive, and thus specific to both Ara h 2 and Ara h 6, if the response rate was >0.1 nm to rAra h 6 after a baseline correction at the 150-second mark (4× the response rate of a negative control antibody).

Epitope binning assays

To perform epitope binning, biotinylated rAra h 6 diluted to a concentration of 2 μg/mL in kinetics buffer was loaded onto streptavidin sensors for 80 seconds. Primary mAbs were diluted to 10 μg/mL and secondary mAbs were diluted to 5 μg/mL in kinetics buffer. Primary antibodies were associated for 300 seconds. Then, after a 60-second baseline step, secondary antibodies were associated for 300 seconds. The sensors were regenerated in between each primary antibody–baseline–secondary antibody cycle. Antibody binding was measured by response rate (nm).

Kinetics assays

Affinities to recombinant Ara h 2 and recombinant Ara h 6 for each mAb were evaluated simultaneously on parallel sensors. For all cross-reactive antibodies, mAb affinities to recombinant Ara h 2 and Ara h 6 were determined using a protocol similar to the one previously published to assess mAb affinity to natural Ara h 2 [16]. Strepdavidin sensors (Sartorius) were loaded with either 2 μg/mL recombinant avi-MBP-tev-Ara h 2 or 2 μg/mL recombinant avi-MBP-tev-Ara h 6 for 80 seconds. To determine each affinity, 6 serial dilutions of antibodies, each by a factor of 3 usually beginning at a concentration of 10 μg/mL, were performed. A reference sample of kinetics buffer with no antibody added was used in each experiment to account for any buffer drift. After a 60-second baseline step, the streptavidin sensor was associated with a mAb (or the reference sample of kinetics buffer) for 150 seconds and then dissociated in kinetics buffer for 300 seconds before it was regenerated for 30 seconds. This sequence of steps was repeated for each dilution to create six affinity curves. A minimum of five curves were used to calculate the mAb affinity to recombinant Ara h 2 and Ara h 6 with a Chi-squared value <3 and a coefficient of determination R² value >0.95. If the Chi-squared value was >3 or the R² value was <0.95 for a particular mAb, the process was repeated at a different starting concentration. Affinities that did not meet criteria were not utilized in the analysis.

Peptide binding assays

In order to determine whether antibodies binding to the KRELRNL sequential epitope on Ara h 2 also bound to the homologous sequential epitope on Ara h 6, a custom-synthesized, lyophilized, biotinylated peptide with the amino acid sequence MVQHFKRELMNLPQQCNFGA (KRELMNL), labeled with a biotin and a hydrophilic linker (TTDS) on the N-terminus (BioTides, JPT Peptide Technologies, Berlin, Germany) was reconstituted in dimethyl sulfoxide (Sigma-Aldrich) for loading to the streptavidin sensor (Sartorius). Individual mAbs were associated with the sensor for 300 seconds each with a regeneration step in between. The negative controls were mAbs known to bind to conformational epitopes. Antibody binding to the peptide was measured by response rate (nm) after alignment with the preceding baseline.

Indirect competitive enzyme-linked immunoassay

In order to quantify the ability of mAbs to inhibit peanut-allergic serum IgE binding to Ara h 2 and Ara h 6, an indirect competitive enzyme-linked immunoassay (ELISA) was performed as described previously [16]. Briefly, microtiter platers were coated at 4°C overnight with 2.5 μg/mL of allergen, either rAra h 2 or rAra h 6, followed by a 2-hour blocking step (PBST, with 0.05% Tween 20 and 1% BSA) for 2 hours. After washing with PBST, mAbs (2.5–20 μg/mL) were added for 2 hours. After washing with PBST, wells were incubated with a 1:25 dilution of IgG-depleted peanut-allergic serum for 2 hours. After washing with PBST, polyclonal anti-IgE conjugated to HRP (1:5000, Bethyl Laboratories #A80-108B) was added to the plate for 1 hour. After washing with PBST, 3,3ʹ,5,5ʹ-tetramethylbenzidine substrate (BioLegend) was incubated for 1 hour, and the reaction was stopped with 1 M phosphoric acid (Thermo Fisher Scientific). Absorbance was measured at 450 nm on a SpectraMax iD5 spectrophotometer (Molecular Devises) with SoftMax Pro 7.1 (Molecular Devices). The background was subtracted from OD values before data analysis and the percentage of IgE inhibition was normalized to control wells without blocking mAbs as described previously [16].

The serum used in these assays was collected from peanut-allergic patients before they underwent OIT (oral immunotherapy) (n = 16) from two IRB-approved clinical trials conducted at Massachusetts General Hospital (NCT01324401 and NCT01750879).

Structure modeling and molecular dynamics simulations

To characterize the interactions of the 22S1 antibody in complex with Ara h 2 and Ara h 6, we performed molecular dynamics (MD) simulations of these antibody complexes. Homology modeling/structure modeling has been performed using Molecular Operating Environment (MOE) [20]. For the 22S1 antigen-binding fragment (Fab) in complex with Ara h 2, we used the available crystal structure (PDB 8DB4) as the starting structure for our simulations [16]. For the 22S1–Ara h 6 complex, we made a model using the available Ara h 6 NMR structure (PDB 1W2Q) [21] and the available complex X-ray structure as a template for the complex (PDB 8DB4). Furthermore, we generated a homology model using the WHO/IUIS Ara h 6 sequence (AAD56337) with the available NMR structure as the template (PDB 1W2Q). In addition, we used AlphaFold2 to generate a structure model of the WHO/IUIS Ara h 6 and modeled the complex based on the complex X-ray structure (PDB 8DB4) [22]. For all these antibody–antigen complexes, we performed each three replicas of classical MD simulations.

In the MD simulation protocol, the structure models were prepared in MOE using the Protonate3D tool [23]. With the help tool of the Amber Tools20 package [24], we explicitly bonded all existing disulfide bridges and placed the antibody–antigen complexes into cubic water boxes of TIP3P water molecules with a minimum wall distance to the protein of 12 Å [25–27]. Parameters for all antibody simulations were derived from the AMBER force field 14SB [28, 29]. To neutralize the charges, we used uniform background charges [30–32]. Each system was carefully equilibrated using a multistep equilibration protocol [33, 34]. MD simulations were performed using pmemd.cuda in an NpT ensemble to be as close to the experimental conditions as possible and to obtain the correct density distributions of both protein and water. Bonds involving hydrogen atoms were restrained by applying the SHAKE algorithm [35], allowing a timestep of 2.0 fs. Atmospheric pressure of the system was preserved by weak coupling to an external bath using the Berendsen algorithm [36]. The Langevin thermostat was used to maintain the temperature at 300K during simulations [37].

To calculate contacts of the antibody–antigen complexes, we used the GetContacts software [38]. This tool can compute interactions within one protein structure, but also between different protein interfaces and allows to monitor the evolution of contacts during the simulation. The contacts are defined based on the default geometrical criteria provided by GetContacts.

As a measure of the flexibility/conformational diversity of the antibody–antigen complex, we also performed a clustering analysis. The obtained trajectories were clustered in cpptraj [39] using the average linkage hierarchical clustering algorithm with a distance cut-off criterion of 12 Å and aligning on the respective antigen.

Statistical analysis

Statistical and graphical analyses were performed with GraphPad Prism, version 10.0.3. Multiple comparisons were performed with a one-way ANOVA, followed by Tukey’s multiple comparison test and adjusted P values were reported. Paired and unpaired comparisons were performed with a two-tailed Student’s t-test.

Results

Ara h 2 and Ara h 6 IgE levels are highly correlated

To dissect the role of cross-reactivity of serum IgE in peanut-sensitized individuals, we compared Ara h 2 and Ara h 6 IgE levels among all patients evaluated by peanut component testing at our institution (n = 741) (Fig. 1). We found a strong linear correlation between IgE levels to Ara h 2 and Ara h 6 (m = 0.75, R² = 0.82, P < 0.0001). This strong linear correlation has been previously identified in other studies [15]. The slope of the best-fit line is <1, indicating that overall, patient Ara h 2 sIgE levels are higher than Ara h 6 sIgE levels. Only 11 (1.5%) patients had IgE to Ara h 6 without IgE to Ara h 2 (median Ara h 6 IgE 0.13 kU/L, range 0.12–1.36 kU/L). In contrast, seven patients (0.94%) had elevated sIgE to Ara h 2 without elevated sIgE to Ara h 6; all of these patients had sIgE levels to Ara h 6 <1 kU/L and an elevated peanut sIgE test.

Figure 1.

Correlation of Ara h 2 and Ara h 6 IgE levels. Comparison of serum Ara h 2 and Ara h 6 IgE levels of 741 peanut-sensitized patients using a simple linear regression with Pearson’s coefficient, where m is the slope of the line of best fit.

Three discrete epitopes of Ara h 6 cross-react with Ara h 2

To define the epitope-based cross-reactivity underlying the positive correlation between patient sIgE levels to Ara h 2 and Ara h 6, we utilized a panel of human mAbs. We hypothesized that a significant number of Ara h 2-specific mAbs would be cross-reactive with Ara h 6 given the sequence homology between the two antigens and the correlated sIgE levels. BLI was used to evaluate the Ara h 6 specificity of a panel of Ara h 2-specific mAbs previously cloned from peanut-allergic, oral immunotherapy–treated patients [16].

Of the 69 mAbs, 57 bind to recombinant Ara h 2 (rArah2), and 34 bind to recombinant Ara h 6 (rArah6) (Fig. 2). Based on previous epitope mapping [16] of natural Ara h 2, cross-reactive mAbs that bind both to rAra h 2 and Ara h 6 bind to three epitopes: conformational epitopes 1.1 and 3 and the KRELRNL (Ara h 2)/KRELMNL (Ara h 6) sequential epitope. The cross-reactive mAbs included 28 of the 30 epitope 1.1 antibodies evaluated and all 3 of the epitope 3 antibodies evaluated, as well as all 3 antibodies binding to the KRELRNL/KRELMNL sequential epitope. Antibodies from 15 patients were characterized; antibodies from 11 of these patients were cross-reactive (8 patients for epitope 1.1, 3 patients for epitope 3, and 1 patient for KRELRNL/KRELMNL).

Figure 2.

Epitope specificity of cross-reactive antibodies. (A) Ribbon diagrams of Ara h 2 (gray) and Ara h 6 (purple). (B) Number of monoclonal antibodies (mAb) that cross-react with rAra h 6 (9–124, PDB 1W2Q) (purple) from those that bind to rAra h 2.0101 (31–160) (gray) on BLI. (C) Representative BLI sensorgrams of epitope-specific mAbs binding to biotinylated rAra h 6 bound to streptavidin. (D) Epitope specificity of individual rAra h 2-specific mAbs that cross-react with recombinant Ara h 6, shaded by epitope; unshaded mAbs bind to rAra h 2 but do not bind to rAra h 6.

None of the epitope 1.2 and epitope 2 antibodies were specific to Ara h 6, as well as none of the antibodies binding to the RRCQSQLER, QQEQQF, and DPYSP^OHS sequential epitopes (Fig. 2B and C). The Ara h 2 epitopes which are not cross-reactive with Ara h 6 are not conserved in the sequence of Ara h 6, suggesting that the difference in antibody binding to the two allergens may be dependent more on sequence rather than on differences in protein folding.

Of the 30 mAbs that are known to recognize epitope 1.1, the 2 mAbs that do not bind to rAra h 6 are also among the most dissimilar from the other 28 mAbs in the amino acid sequence of their heavy chain CDR3 regions (Supplementary Fig. S1).

The 12 (17.4%) mAbs that did not bind to rAra h 2 were previously found to bind to the DPYSP^OHS sequential epitope (Fig. 2D) [16], which lacks the post-translational hydroxylated proline (P^OH) critical for binding of these antibodies (Fig. 2D) [7, 40, 41].

Cross-reactive antibodies identify similar epitopes of Ara h 6 and Ara h 2

Next, we compared epitopes, using epitope binning, of the cross-reactive mAbs on rAra h 6. Previously, we used epitope binning by BLI to confirm that epitopes 1.1 and 3 are discrete on Ara h 2 by demonstrating that mAbs that bind these regions can bind successively [16]. Similarly, mAbs that bind these two regions on Ara h 2 also bind to nonoverlapping epitopes on Ara h 6, as the response rate on binding to rArah 6 increases after successive binding of an mAb specific to epitopes 1.1 and 3 as well as epitopes 1.1 and KRELRNL/KRELMNL (Fig. 3A). Antibodies previously identified to bind to the KRELRNL region of Ara h 2 [16] also bind to the KRELRMLP region of Ara h 6 (Fig. 3B). Based on sequence homology, the KRELRMLP of Ara h 6 overlaps with the region predicted to bind to epitope 3 (Fig. 3C), similar to the previous work in Ara h 2.

Figure 3.

Epitopes of cross-reactive mAbs. (A) Representative sensorgrams of epitope binning on rAra h 6 with 23P34 and 23P17 (Epitope 1.1), 06B1 and 27A4 (Epitope 3), 23P8 (KRELRNL/KRELRML). (B) Sensorgrams of binding curves of 23P8 and 23P16 to Ara h 6 peptide MVQHFKRELMNLPQQCNFGA. (C) Sequence alignments of a segment of Ara h 2 (130–172) and Ara h 6 (85–124), with amino acid residues involved in 22S1 binding to rAra h 2 colored in red.

Cross-reactive IgG antibodies inhibit allergen-IgE interactions more significantly on rAra h 6 than on rAra h 2

To evaluate epitope-specific inhibition of polyclonal IgE to rAra h 6 by mAb in comparison to epitope-specific inhibition to rAra h 2, we used an indirect inhibitory ELISA. We measured the percent inhibition of polyclonal peanut-allergic serum IgE (Supplementary Table S2) to immobilized rAra h 6 when saturated with individual mAbs representing each of the three cross-reactive epitopes. We identified that mAb to all three cross-reactive epitopes do inhibit variable IgE binding to rAra h 6. We then compared inhibition using the same three mAbs between rArah 2 and rAra h 6 and found that they inhibit a significantly higher percentage of IgE binding on rAra h 6 than they do on rAra h 2 (Fig. 4; Supplementary Fig. S2). These data support the fact that peanut-allergic patients may have a greater diversity of epitope recognition on Ara h 2 than Ara h 6, potentially contributing to why the Ara h 2 serum IgE level is a more accurate biomarker of clinical peanut allergy than the Ara h 6 serum IgE level.

Figure 4.

Serum IgE inhibition to rAra h 2 and rAra h 6 by individual cross-reactive mAb by ELISA. (A) Saturation curves of inhibitory ELISA using mAb to inhibit polyclonal serum IgE binding to rAra h 6 or rAra h 2. Lines colored by the mAb epitope, representative of 1 patient. (B) Percent inhibition of polyclonal serum IgE binding to rAra h 6 and rAra h 2 by cross-reactive mAbs at a single concentration of mAb. Each point is median of triplicates from each patient (n = 16). Paired Student’s t-test *P < 0.05.

Cross-reactive antibodies have different affinities to rAra h 2 and rAra h 6, driven by K_d

Next, we quantitatively compared the affinity of cross-reactive antibodies to rAra h 2 and rAra h 6 by simultaneously measuring affinity by BLI (Fig. 5A; Supplementary Fig. S3; Supplementary Table S1). The affinity of mAbs that recognize epitope 1.1 is higher than mAbs that recognize epitope 3 (Fig. 5B, adjusted P = 0.02), due to the difference in the K_dissociation (adjusted P < 0.0001). Overall, the affinity of all mAbs was significantly lower to rAra h 6 than to rAra h 2 (P = 0.03), due to the difference in the K_dissociation. Most mAb that recognize epitope 1.1 have similar or reduced affinities to rAra h 2 and rAra h 6, again due to the difference in K_dissociation (Fig. 5C); however, two mAbs had higher affinity to rAra h 6 (23P34 and 24D1, Supplementary Fig. S4).

Figure 5.

Affinities of cross-reactive mAb to rAra h 2 and rAra h 6. (A) Representative affinity curves for rAra h 2 and rAra h 6 for mAbs binding to each cross-reactive epitope. (B) Affinity (KD), association rate (K_a), and dissociation rate (K_d) to rAra h 6 by epitope. Ordinary one-way ANOVA with multiple comparisons. *adjusted P < 0.05, ****adjusted P < 0.0001. (C) Comparison of affinity (KD), association rate (K_a), and dissociation rate (K_d) for individual mAbs to rAra h 2 and rAra h 6. Paired student’s t-test, *P < 0.05.

Monoclonal antibody affinity is not correlated with OIT outcomes

Previously, we found that the differences in mAb affinity to natural Ara h 2 were not correlated with the clinical efficacy of OIT or epitope specificity [16]. Again, we did not find any significant difference in mAb affinity to rAra h 2 or rAra h 6, including KD, K_association, and K_dissociation, across clinical outcomes of OIT (Supplementary Fig. S5).

The increased flexibility of Ara h 6 correlates to lower affinity of cross-reactive mAbs

To better understand the lower affinity epitope–paratope interactions between cross-reactive antibodies and Ara h 6 at an atomic level, we performed MD simulations. MD simulations enable a description of conformational changes kinetically and thermodynamically in solution and provide insights into intramolecular and intermolecular interaction networks responsible for protein functions. Here, the MD stimulations were based on the previously published X-ray crystal structure of rAra h 2 bound to the Fab of mAbs. Of the mAbs previously used in X-ray crystallography, only epitope 3 is cross-reactive, so we used the rAra h 2 structure with the mAb 22S1. To identify differences between the antibody–allergen complexes with Ara h 2 and Ara h 6, we used AlphaFold2 to predict a structure for Ara h 6 using the sequence available in PDB: 1WQ2. In addition, we also used the available NMR structure of Ara h 6 (PDB:1W2Q) to model the complex. As a control, we compared the AlphaFold 2 structure of Ara h 2 with the X-ray crystal structure of rAra h 2 bound to maltose binding protein (PDB 3OB4) as well as the structure of rAra h 2 bound to the 2 Fab regions of 22S1 and 13T1 (PDB 8DB4), which were identical (data not shown). Modeling of Ara h 2 and Ara h 6 in complex with 22S1 revealed highly similar regions involved in epitope–paratope interactions (Fig. 6A), shown by highlighted residues within the ribbon cartoon representation. The disordered loop region of Ara h 2 is represented as a dotted line. We performed 3 × 1 µs of MD simulations of the complexes with Ara h 2 and Ara h 6 respectively and calculated hydrogen bond and salt bridge interactions in the binding interface. Figure 4A shows the top-view/open book representation of the epitope for both Ara h 2 and Ara h 6, showing similar residues involved in salt bridge interactions with the antibody, color-coded based on their occurrence. We found minor differences in the occurrence of salt–bridge interactions; there was a lower probability of salt bridges in Ara h 6, compared to Ara h 2 (Fig. 6B). On the other hand, the number and probability of relatively weaker hydrogen bonds are similar in both (14 in Ara h 2 and 13 in Ara h 6) (Supplementary Fig. S6).

Figure 6.

Modeling of antibody binding to Ara h 2 and Ara h 6. (A) Representative ribbon cartoon with a transparent surface model highlighting epitopes of Ara h 2 (gray) and Ara h 6 (purple). (B) Ribbon diagram of mAb 22S1 F’ab (heavy chain in dark blue and light chain in light blue) interacting with Ara h 2 (gray) or Ara h 6 (purple). Table of probabilities (Prob) of salt bridge interactions (SBI) between epitope–paratope residues by Kabat numbering. (C) Representative ribbon diagrams of allergen-22S1 F’ab from MD simulations by root mean square deviation.

Apart from characterizing interactions between the antibody and the allergens, we also wanted to compare the conformational diversity of both complexes. One way to quantify flexibility is to consider the number of clusters using the same cut-off distance criterion as a measure of variability. We aligned the complex on the respective allergen and applied a hierarchical average linkage clustering on the Fab residues using a Cα-RMSD distance cut-off criterion of 10 Å. We observed substantial differences in flexibility, reflected in a higher number of clusters for the Ara h 6–22S1 Fab complex (10 clusters) compared to the Ara h 2–22S1 Fab complex (2 clusters) (Fig. 6C). Another way to quantify global changes in flexibility is the B-factor, which is based on the root mean fluctuation. In line with the increase in flexibility demonstrated by clustering, Supplementary Fig. S6B shows an increase in the B-factor of Ara h 6 compared to Ara h 2. Altogether, the increase in flexibility of the Ara h 6–22S1 Fab complex emphasizes the importance of not only considering enthalpic contributions, such as stabilizing intermolecular interactions, but also conformational entropy as a key determinant for binding affinity.

Thus, the increased flexibility in interactions between Ara h 6 with antibodies in the MD simulations is likely a manifestation of increased entropy and explains the increased dissociation rate and therefore lower affinity of cross-reactive mAbs from Ara h 6.

Discussion

This work utilizes mAbs recombinantly cloned from peanut-allergic, oral immunotherapy–treated patients to characterize the cross-reactivity between 2S albumin seed storage allergens Ara h 2 and Ara h 6 on a molecular level. Previous studies utilizing peptide microarrays [42–45] have been inherently limited to the study of sequential epitopes; our study was the first to add characterization of the conformational epitopes conserved between Ara h 2 and Ara h 6.

This work identified 3 cross-reactive epitopes conserved across Ara h 2 and Ara h 6; however, the affinity of these antibodies is lower to Ara h 6 due to increased dissociation, particularly among antibodies that bind two of the three epitopes. Modeling suggests that there is increased flexibility in allergen–antibody interactions with Ara h 6 than with Ara h 2, which can result in increased dissociation. Furthermore, IgE inhibition assays suggest that the immunodominant epitope space on Ara h 6 may be more constrained than Ara h 2. Taken together, these cross-reactive antibodies drive the high degree of correlation in the Ara h 2 and Ara h 6 serum IgE level in peanut-sensitized individuals, and the greater epitope diversity of Ara h 2 may explain its superior function as a predictive biomarker of peanut allergy. Utilizing recombinant antigens enabled us to study cross-reactivity without the inherent potential for contamination of natural Ara h 2 and natural Ara h 6 with one another, which could create the appearance of cross-reactivity where there was none.

The frequency of cross-reactive antibodies to Ara h 2 and Ara h 6 is high, in part, because the most abundant mAbs bind to the conformational epitope 1.1 and are among those that are cross-reactive. In our previous work, we identified antibodies that recognize epitope 1.1 from B cells from over 50% of the profiled patients, and these antibodies emerged as convergent, or public, antibodies in multiple peanut-allergic patient cohorts [18, 46–48]. Together, this suggests that cross-reactive antibodies that recognize epitope 1.1 likely occur in a large number of individuals and may be dominant on a population level.

Consistent with previous structural characterization, we identified that the similarities between Ara h 2 and Ara h 6 extend beyond sequence homology, and include allergen size and the 4 alpha-helical structural structure. In our study, antibodies that bind to epitope 3 were also found to be cross-reactive, and these are known to interact with three helices on rAra h 2 by X-ray crystallography [16], suggesting the similarity is more than surface deep between the two allergens. Consistent with predictions, non-cross-reactive antibodies, such as those that bind to sequential epitopes such as the DPYSP^OHS loop region, RRCQSQLER, QQEQQF, or to the conformational epitope on alpha-helix 2, are located on Ara h 2 regions which are missing or substantially altered in the amino acid sequence of Ara h 6. As this work specifically identified B cells using a natural Ara h 2 multimer, less common or less dominant Ara h 6 epitopes may still exist, outside of these cross-reactive epitopes. Our current use of mAbs for epitope mapping is inherently limited, and future studies with larger cohorts would be useful.

In addition, while our use of recombinant allergens enables study of pure, non-contaminated allergenic proteins, recombinant allergens can have limitations. For instance, the recombinant proteins used here do not have the hydroxylated prolines, which are post-translational modifications present on the loop region of Ara h 2. The hydroxyproline within a repeated motif in this loop region has been long-recognized as an important sequential epitope of Ara h 2 but not Ara h 6. Since this Ara h 2-specific epitope is absent in Ara h 6, our comparison using a form of rAra h 2 without hydroxyproline to rAra h 6 allows for a direct comparison of those cross-reactive epitopes. Another limitation is the concern that recombinant allergens could be misfolded, thereby impacting conformational epitopes and mAb binding. We confirmed that conformational antibodies to Ara h 2 bind and epitope bin similarly with nAra h 2 and rAra h 2. Next, we utilized an identical methodology to create the rAra h 2 and rAra h 6 to create comparable recombinant allergens. Lastly, the specificity of these mAb is similar in nAra h 6 and rAra h 6. As we know that some of these mAb require three intact alpha-helices in a specific spatial configuration for binding [16], this data strongly suggest that the recombinant allergens are folded similarly to natural allergens.

The lower affinity of cross-reactive antibodies to Ara h 6 may contribute to the relative immunodominance of Ara h 2. We identified that the cross-reactive mAb affinity to rAra h 6 was significantly lower than it was to rAra h 2, especially for mAbs recognizing epitope 3 and the KRELRNL/KRELMNL sequential epitope, due to increased dissociation from Ara h 6. These findings align with previous research suggesting that antibodies bind to Ara h 6 with lower affinity than they do to Ara h 2, based on effector cell assays [15]. The dynamics of allergen–antibody interactions are particularly relevant in these assays, and the increased flexibility of the mAb-Ara h 6 complex suggests a mechanism for the lower affinity of these antibodies, which may result in less effector cell activation and a lower clinical significance as a result.

While a high degree of correlation in peanut-sensitized individuals between Ara h 2 sIgE and Ara h 6 sIgE has been observed in our and other cohorts, Ara h 2 sIgE is considered a more accurate biomarker of clinical peanut allergy, even in infancy [46]. Our work provides clues as to why Ara h 2 remains a more accurate biomarker of clinical peanut allergy than Ara h 6 despite their significant homology. The higher-affinity, more rigid interactions between Ara h 2 and mAbs in comparison to the interactions between mAbs and Ara h 6, in addition to the presumed greater number of epitopes on Ara h 2, help explain why sIgE levels to Ara h 2 have been demonstrated to be more clinically relevant. In conclusion, we identified three common, cross-reactive epitopes of the 2S albumin peanut allergens, Ara h 2 and Ara h 6, to better understand the clinical relevance of these allergens by dissecting the dynamics of their interactions with antibodies.

More broadly, these findings may speak fundamentally to the molecular dynamics (MD) that render Ara h 2 a more dominant and clinically relevant allergen than Ara h 6. Beyond providing insights into the clinical relevance of commonly utilized diagnostic biomarkers, this work starts to identify the molecular requirements for cross-reactivity by probing epitope–paratope interactions. Future studies of epitope-specific cross-reactivity across allergens will be essential to understand how allergen-specific antibodies develop and drive IgE-mediated food allergy.

Ethical Approval

This study was undertaken according to the principles in the Declaration of Helsinki. All patients in both clinical trials (NCT01324401 and NCT01750879) provided written and oral consent.

Conflict of Interests

S.U.P. filed patent PCT/US2022/022888 related to this work. S.U.P. has consultancy agreements with Mabylon and Buhlmann.

Funding

This work was funded by grants from the National Institute of Allergy and Infectious Diseases (NIAID), NIH (5R01AI155630, 1R21AI159732, to S.U.P.); the Division of Intramural Research of the National Institute of Environmental Health Sciences (NIEHS), NIH (1ZIAES102906, to GAM; 1ZICES102645, to LCP); Food Allergy Science Initiative award (SUP). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The clinical trial work was performed at the Harvard Clinical and Translational Science Center, supported by grant numbers 1UL1TR001102 and 8UL1TR000170 from the National Center for Advancing Translational Science, and 1UL1 RR025758 from the National Center for Research Resources. Cytometry was performed at the MGH Department of Pathology Flow and Image Cytometry Research Core through support from the NIH Shared Instrumentation program (grants 1S10OD012027-01A1, 1S10OD016372-01, 1S10RR020936-01, and 1S10RR023440-01A1). This work was supported by the Austrian Academy of Sciences APART-MINT postdoctoral fellowship to MFQ.

Data Availability

The data in this article will be shared on reasonable request to the corresponding author.

Author Contributions

O.M.R., M.F.Q., T.K., and S.U.P. designed experiments. O.M.R. and T.K. performed all experiments. M.F.Q. performed all molecular modeling. G.Z., L.C.P., and G.A.M. developed recombinant allergens and structural data. J.S. and S.U.P. compiled clinical data. O.M.R. and S.U.P. wrote the manuscript. All authors edited the manuscript.

Clinical Trial Registration

The monoclonal antibodies utilized in this study were cloned from the peripheral blood of patients enrolled in two clinical trials approved by the Mass General Brigham Institutional Review Board (NCT01324401 and NCT01750879).