Recent immunological data demonstrated that dendritic cells preferentially recognize advanced glycation end product (AGE)-modified proteins, upregulate expression of the receptor for AGE (RAGE), and consequently bias the immune response toward allergy.
Peanut extract was characterized by mass spectrometry (MS) to elucidate the specific residues and specific AGE modifications found in raw and roasted peanuts and on rAra h 1 that was artificially glycated by incubation with glucose or xylose. The binding of the RAGE-V1C1 domain to peanut allergens was assessed by PAGE and Western analysis with anti-Ara h 1, 2, and 3 antibodies. IgE binding to rAra h 1 was also assessed using the same methods.
AGE modifications were found on Ara h 1 and Ara h 3 in both raw and roasted peanut extract. No AGE modifications were found on Ara h 2. Mass spectrometry and Western blot analysis demonstrated that RAGE binds selectively to Ara h 1 and Ara h 3 derived from peanut extract, whereas the analysis failed to demonstrate Ara h 2 binding to RAGE. rAra h 1 with no AGE modifications did not bind RAGE; however, after AGE modification with xylose, rAra h 1 bound to RAGE.
AGE modifications to Ara h 1 and Ara h 3 can be found in both raw and roasted peanuts. Receptor for AGE was demonstrated to selectively interact with AGE-modified rAra h 1. If sensitization to peanut allergens occurs in dendritic cells via RAGE interactions, these cells are likely interacting with modified Ara h 1 and Ara h 3, but not Ara h 2.
In the United States, the estimated prevalence of food allergy is 2.7%, where 1.3% is specifically sensitive to peanuts . Adverse reactions to peanuts can be triggered by extremely small doses and are the most frequent type of fatal anaphylaxis to food allergens [2, 3]. However, the reasons for this hypersensitivity are poorly understood. It has been suggested that food processing may contribute modifications, such as advanced glycation end products (AGEs), that are inappropriately recognized by the immune system and contribute to allergic sensitization . Studies with antibodies raised against common AGEs and advanced lipoxidation end products (ALEs) concluded that the major peanut allergens Ara h 1 and Ara h 3 were commonly modified with carboxymethylysine, malondialdehyde, and hydroxynonenal, while such modifications were less common in Ara h 2 . In addition, IgE derived from allergic patients more strongly recognizes roasted peanut extracts than raw . It should be noted that patients only raise antibodies against antigens to which they are exposed, and in the United States, it is uncommon to encounter raw peanuts. So that while patient IgE recognizes roasted peanuts, it does not demonstrate that AGEs necessarily contribute to sensitization.
However, there are several new lines of evidence that AGE modifications do contribute to allergic sensitization. Buttari et al.  demonstrated that dendritic cells stimulated with AGE-modified protein activated more IL-4-producing T cells than IFN-γ-producing T cells, indicating a possible Th2 bias. Similar results were found by Hilmenyuk et al. , comparing dendritic cells pulsed with AGE-modified OVA and regular OVA. The CD4+ T cells co-cultured with the dendritic cells produced more of IL-5 and reduced production of IFN-γ in response to the AGE-modified OVA, again implying a Th2 bias. Another study also found enhanced CD4+ T-cell activation in response to AGE-modified OVA; however, the uptake by dendritic cells was mediated by scavenger receptor class A type I and II (SR-AI/II) and not RAGE (receptor for AGEs) or galectin-3 . These experiments provide good evidence that AGE modification of proteins can bias the immune response toward allergic sensitization.
The most logical receptors to recognize proteins with AGE modifications would be RAGE or other scavenger-type receptors with promiscuous affinity for a variety of structurally different ligands. Previous studies in Caco-2 cells, which are a model for intestinal epithelia, showed that RAGE activation by AGEs stimulated MAP-kinases , which were shown to influence cellular proliferation in a cancer model . Very recently, AGE-modified Ara h 1 was demonstrated to influence the proliferation of Caco-2 cells . The proliferation depended on the incubation time and temperature used to induce the formation of AGEs, indicating that specific chemical structures were important in influencing the pro-inflammatory network .
The previously mentioned study is among the few to characterize what chemical modifications occur to the peanut allergens . The antibody technology utilized was not capable of identifying which specific residues were modified. In this study, we attempted to characterize, at the atomic level, the modifications that occur in roasted peanuts and in controlled reactions with sugars using mass spectrometry. We tested the modified and unmodified proteins for binding to RAGE and allergic patient-derived IgE. The results are the first to directly demonstrate RAGE recognition of AGE-modified peanut allergens and should contribute to a better understanding of the innate immune system recognition of peanut allergens.
Materials and methods
Allergens and peanut extracts
Raw peanuts were oven-roasted at 177°C (350°F) for 15 min. Raw and roasted peanuts were crushed with a mortar and pestle. Two grams of ground peanut was extracted at 60°C with 30 ml of phosphate-buffered saline (PBS), 1M NaCl, pH 7.4 for 15 min, by manually shaking and vortexing the extraction mix. Peanut extracts were allowed to stand for 15 min in a 60°C bath, and the supernatant was centrifuged at 6000 g for 5 min at room temperature. The supernatants were further centrifuged at 16 000 g for 5 min and frozen at −20°C. Recombinant Ara h 1 was expressed and purified as previously described .
SDS-PAGE and Western blot analyses
Each of the samples was mixed with 3× sample loading buffer and incubated for 10 min at 65°C prior to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) on a 4–20% Novex Tris–HCl gel (Invitrogen, Grand Island, NY, USA). The gel was placed into Gel-Code Blue stain (Pierce, Rockford, IL, USA) solution for 1 h, then destained in distilled water for 1 h and photographed. Detailed procedures of the Western blotting are given in the supplemental material. Immunocap values for patients 1–5 are, respectively: >100, 23.1, 14.5, 21.5, and 37.1 kU/l.
Glycation of recombinant Ara h 1
Roasting-induced Maillard reaction or glycation was performed as described in a previously established simulated roasting model with slight modifications . Purified rAra h 1 (0.5 mg/ml) was solubilized in PBS and incubated at 55°C in the presence of 0.25 M glucose or xylose, for 0, 1, 2, 4, 7, or 10 days. Reactions were stopped by either addition of sample buffer for SDS-PAGE and Western blot analysis or frozen at −20°C for use in immunoprecipitation and mass spectrometry analysis.
RAGE pull-down experiments
Human RAGE V1C1 domain was expressed as a fusion with maltose-binding protein (MBP) and purified as described previously . Fresh amylose resin (50–200 ul of slurry) was washed 3–5× with PBS buffer by repeated centrifugation in small PCR tubes. The MBP-RAGE was incubated for 1 h while gently rocking with the amylose resin at 3 mg/ml of slurry, which was reported as the maximum capacity of the resin (New England Biolabs, Ispswich, MA, USA). Excess MBP-RAGE was removed with 3–5 washes of PBS. The resin and MBP-RAGE were incubated with the peanut extract, rAra h 1, or the negative control Bos d 6 for 1 h with gentle rocking. Again, excess protein was removed by repeated washing. The MBP-RAGE was eluted by adding PBS with 10 mM maltose. Xylose- and glucose-treated allergens rAra h 1 and Bos d 6 were obtained by incubating the allergens with 250 mM sugar for the times indicated, up to 10 days. The samples were dialyzed into PBS prior to the binding experiment to remove excess sugars that might interfere with the MBP-RAGE (maltose-binding protein) construct's binding to the amylose resin. Peanut extract concentrations were determined by Bradford assay to be 8–15 mg/ml. The amount of extract added to the resin was designed to not be more than 3 mg/ml.
Mass spectrometry sample preparation
Raw and roasted peanut extracts were analyzed by gel-enhanced liquid chromatography mass spectrometry (GeLC-MS) or off-line/on-line two-dimensional liquid chromatography mass spectrometry in efforts to enhance the detection of AGE modifications on the peanut allergens. For specific details, please see the supplementary material. Recombinant Ara h 1 was analyzed by standard nano LC-ESI-MS/MS.
RAGE binding to recombinant Ara h 1
RAGE is commonly known to promiscuously bind a variety of substrates. To insure that there were no endogenous AGE modifications or post-translational modifications on Ara h 1 that may interact with RAGE, the first binding experiments utilized rAra h 1 from bacteria. As described in the methods, MBP-tagged RAGE (V1C1 domain) was bound to an amylose resin and utilized to ‘pull-down’ potential ligands. Figure 1 shows a Western blot using an antibody selective for Ara h 1 to probe the samples obtained from the pull-down experiment. The results in panel A show the important negative control first, in which rAra h 1 that was not incubated with sugars is not retained by RAGE (lanes 4 and 5) and instead remains in the supernatant (lanes 2 and 3). Secondly, Fig. 1B shows that rAra h 1 incubated with glucose and xylose for 10 days forms multimers due to cross-linking of the subunits from AGE modifications. Clearly the xylose is more efficient than glucose in creating these multimeric modifications. Thirdly, the xylose-modified rAra h 1 is retained by RAGE in this pull-down experiment (panel B lanes 4 vs 5), while the glucose-modified rAra h 1 is not (panel B lanes 6 vs 7).
Mass spectrometry results (Table 1) confirmed that under the conditions examined, glucose only weakly allows for the formation of AGEs on recombinant Ara h 1. Interestingly, the Schiff base is the species detected. The Schiff base is the first step in the formation of AGEs and suggests that perhaps higher glucose concentration, longer incubation times, or higher temperatures may induce additional AGEs. In light of these mass spectrometric results, it is not surprising that glucose-modified rAra h 1 did not interact with RAGE and is likely related to the reduced efficiency of glucose vs xylose in forming AGE modifications in the same amount of time.
Table 1. Advanced glycation end product modifications on rAra h 1
MW obs (M + H)
MW theor (M + H)
Extensive y-ion series and precursor mass confirm this peptide with the addition of 116 Da, but the position of the 116 Da addition is ambiguous.
Table 1 compares the AGE modifications and the identity of the modified residues in the xylose- and glucose-treated samples. The most common modification identified was a carboxymethyl (CM) adduct, which occurred on both arginine and lysine residues. Some peptides, such as the IFLAG, were found modified with different AGEs. All of the peptides that were modified by glucose were also found to be modified in the sample incubated with xylose. Only the DLAF peptide was uniquely modified by xylose. Figure 2 and supplementary material display mass spectrometry results with AGEs present.
IgE binding to AGE-modified proteins
Figure 3 shows the binding of IgE from five peanut-allergic patients to the glucose- and xylose-treated rAra h 1. In panel A, the staining shows that the total protein content appears reduced at the longer incubation times with the sugars. A large number of free radicals that form during the progression of the Maillard reaction begin to attack and degrade the protein into smaller fragments that most likely run off of the bottom of the gel. What remains of the AGE-modified protein (and fragments thereof) is less visible and appears as smears, due to extensive levels of modifications. In each case, the intensity of the Western blot (Fig. 3B) closely matches the staining of the protein by Coomassie. This shows that the patient IgE recognizes both the monomeric and covalently linked multimeric forms of rAra h 1 that form due to AGE modifications. The total IgE binding appears to increase for some patients in the first time points and decrease for others. A slight reduction is seen in IgE binding to the xylose-modified samples at longer incubation times as the epitopes become extensively modified. Based on Fig. 3 and the MS analysis in Table 1, the patient IgE extensively recognizes both the AGE-modified and unmodified allergens. These results demonstrate that this in vitro system produces relevant modified allergens.
To further test whether the patient IgE recognized the AGE modifications exclusively, the milk allergen Bos d 6 was treated in the same in vitro system as was rAra h 1. Similar to the results in Fig. 1, the xylose-treated Bos d 6 interacted most strongly with RAGE (data not shown). Supplementary Fig. 1 demonstrates that the peanut-allergic patient sera did not recognize the AGE-modified Bos d 6. In summary, RAGE is specific for AGE modifications, while the patient IgE is allergen specific (See Supplementary Material).
RAGE interactions with allergens from roasted peanut extract
Figure 4 shows the results of a second RAGE pull-down experiment; however, this time raw and roasted peanut extracts were used. Panel A shows a stained SDS gel for proper comparison of the amount of protein used in the lanes containing raw vs roasted peanut. Western blots are shown in panels B, C, and D utilizing Abs specific for Ara h 1, Ara h 2, and Ara h 3, respectively. The results in panel B indicate that Ara h 1 is weakly pulled down by RAGE from peanut extract, from both the raw and roasted samples. Although the bands are in lanes 4 and 6 are weak, Ara h 1 was readily identified with 12 MS/MS spectra corresponding to 3 unique peptides (See supplemental material), with a Distinct Summed MS/MS Search Score of 42.2, and 5% sequence coverage. In panel B, Ara h 2 was not detected in the RAGE pull down at the lower molecular weight (9 kDa) usually associated with monomeric Ara h 2. There are faint bands at higher molecular weight that could indicate the presence of aggregates with some affinity for RAGE. MS analysis never detected any Ara h 2 fragments with AGE modifications (see Table 2 and discussion below).
Table 2. Advanced glycation end product modifications in peanut extract
MW obs (M + H)
MW theor (M + H)
extensive y-ion series and precursor mass confirm this peptide with the addition of 58 Da, but the exact position of the 58 Da addition can only be localized to the 5 C-terminal residues.
extensive y-ion series and precursor mass confirm this peptide with the addition of 58 Da localized on the extreme N-terminal residue.
Figure 4 panel D shows that Ara h 3 is most efficiently pulled down by RAGE according to the Western blot analysis. In addition, the faint band in lane 5 may indicate either nonspecific interactions of AGE-modified Ara h 3 with the amylose resin or that the RAGE binding was saturated, but the bands are significantly stronger when RAGE is present (Lanes 5 vs 6). The MS results confirmed the presence of Ara h 3 and/or ten other variants or iso-allergens of Ara h 3 (Table S1). In the NCBI database, there are 15 different proteins from Arachis hypogaea with >68% sequence identity to Ara h 3 (Table S2). According to the World Health Organization and International Union of Immunological Societies (WHO/IUIS) Allergen Nomenclature (www.allergen.org), an allergen with >67% sequence identity may be considered an iso-allergen. Because of the high sequence identity among the short peptides identified by MS, it is difficult to uniquely identify from which protein the peptide was derived. We are confident that the Western and MS data demonstrate that RAGE pulls down Ara h 3, or one or more of these iso-allergens.
In addition, Fig. 4 panel E shows IgE binding detected by Western blot using a pool of sera from patients 2 and 3 and the same samples used for the analyses in panels B–D. The patient's IgE appears to detect the roasted extract more strongly than the raw after co-elution with RAGE. This is due to less total peanut protein that co-elutes with RAGE from the raw extract which contains fewer AGE modifications. The bands in panel E most closely match those of Ara h 3, and Ara h 1 although there are some additional bands. These additional bands may be other peanut allergens with AGE modifications besides Ara h 1, 2, or 3.
Table 2 describes the AGE modifications found in Ara h 1 and Ara h 3 in peanut extracts. Only a few peptides could be confidently identified as having CM adducts in both the raw and roasted extracts. In the roasted extract, we were only able to identify CM adducts on four peptides from Ara h 1 and two peptides from Ara h 3. Two of the peptides were similarly identified with CM adducts in the artificial glycation with rAra h 1. Mass spectra with these AGE-modified peptides are shown in the supplemental material. Notably, no AGEs were found on Ara h 2 despite both Ara h 2.01 and Ara h 2.02 being readily detected in both raw and roasted peanut extracts. Ara h 2.01 was identified with a Distinct Summed MS/MS Score of 275.8, which corresponded to 334 spectra of 18 unique peptides and 78% sequence coverage. Ara h 2.02 was identified with a Distinct Summed MS/MS Score of 158.4, which corresponded to 50 spectra of 10 unique peptides and 48% sequence coverage. The fact that no glycation events could be detected was surprising, given the number of lysine and arginine residues on the surface of Ara h 2 in the recently described structure . Note that Tables 1 and 2 do not purport to be a complete catalogue of all possible AGE modifications. These are the modifications that were most confidently identified with mass spectrometry.
Figure 5 shows the location of the identified AGE modifications on the structure of Ara h 1 . The residues with modifications are distributed apparently randomly over the protein surface. Most of the lysine modifications occur on the ‘front face’, which is likely a consequence of the higher concentration of lysines on this surface.
The peanut has been identified as one of the most potent allergenic foods [15, 16]. Extracts from roasted peanuts were shown to bind higher IgE levels than extracts from raw peanuts , and other studies support a role for AGE modifications in this phenomenon . Recent studies have suggested that AGE–RAGE interactions may contribute to allergic sensitization [6, 7] so we assessed the recognition and binding of peanut allergens to RAGE.
This study describes evidence supporting RAGE binding to peanut allergens that have been specifically modified by AGEs. These studies establish that RAGE does not interact with unmodified rAra h 1, but does interact with AGE-modified rAra h 1 (Fig. 1). Secondly, we find that RAGE can bind AGE-modified Ara h 1 and Ara h 3 (Fig. 4). Mass spectrometry identified explicit AGE modifications associated with specific residues on Ara h 1 and Ara h 3 (Tables 1 and 2) and largely confirmed the analysis of the Western blots. We detected similar AGE modifications on the same peptides in Ara h 1 as were described by Hebling et al. . However, in contrast to our results, that study did not mention any AGE modifications in the raw peanut extract. The lack of modifications to Ara h 2 was similar to that described with mAb detection of AGEs by Chung et al. , and no modifications to Ara h 2 are mentioned in the MS analysis of Hebling et al. . The reported large number of modifications in the study by Chassaigne et al.  may seem to be in conflict with this result; however, the list of modifications to Ara h 2 in that study includes primarily N-terminal modifications and not AGEs. An important contribution of this study is that AGE modifications were found on both raw and roasted peanut allergens. These results suggest that future studies on the effect of AGE modifications from roasted peanut extract should utilize a recombinant protein instead of raw peanut extract as a control.
A comparison of the amount of Ara h 1 detected by Western blot analysis in raw vs roasted peanut extracts indicated a reduced signal intensity for the allergen from roasted peanut 1 (Fig. 4A, lanes 2 and 3). This result is apparently contradictory to previous work that demonstrated an increase in the amount of Ara h 1 in solution in peanut extracts that is dependent on the roasting time . This increase occurred in spite of a reduction, upon roasting, in the binding capacity of one of the two anti-Ara h 1 monoclonal antibodies from the ELISA used to measure the allergen. Similarly, the epitope recognized by the antibody used in the immunoblot analysis could have been affected by roasting, resulting in a reduced binding to Ara h 1 from roasted peanut. We emphasize that AGE-modified Ara h 1 is recognized by RAGE, as demonstrated in Fig. 1. The detection of Ara h 1 in the extract may also be weak because Ara h 3 in roasted extract is more effectively competing for RAGE binding.
It is interesting to discuss why the patient IgE does not appear to significantly differentiate between the AGE-modified and unmodified rAra h 1. There are several conclusions to draw. First, consumption of raw peanuts is rare; hence, the patients in question were likely primarily exposed to roasted peanuts. However, we have demonstrated here that even ‘raw’ peanuts may contain a substantial number of AGE modifications. Therefore, the IgE binding data here suggests that the AGE modifications do not substantially modify patient epitopes, because the IgE binds well to the control rAra h 1 that is untreated and contains no AGE modifications. The data on AGE-modified Bos d 6 further demonstrate that the patient IgE is not specific to the AGE modifications. In contrast, RAGE is specific to AGE modifications, and is not protein specific. To be clear on the separate issues of RAGE and IgE binding in the allergic response, RAGE is hypothesized to be important in the initial innate immune response that biases sensitization toward an allergic response. IgE binding is part of the later adaptive immune response from B cells.
In addition to AGE modifications, other protein modifications, including gylcosylation, phosphorylation, and cysteinylation, have been studied for their ability to modify the immunological response . In the case of the birch pollen allergen Bet v 1, nitrosylation of the protein was demonstrated to alter the presentation of Bet v 1 peptides on dendritic cells and increase proliferation, and IL-5 and IFN-γ production in splenocytes [22, 23]. While cataloging the AGE modifications to the peanut allergens, numerous other modifications were identified such as malondialdehyde, which is an ALE (data not shown). For this study, we have focused on reporting the AGE modifications and RAGE interactions. An important conclusion of this study is that RAGE does bind to AGE-modified Ara h 1 and Ara h 3. If allergic sensitization is enhanced by RAGE-dependent interactions, it is apparently feasible that AGE-modified Ara h 1 and Ara 3 are involved.
This research was supported in part by Research Project Number Z01-ES102885-01 to REL, Z01-ES50161 to KBT, and Z01-ES102488-05 to JGW in the Intramural Research Program of the National Institute of Environmental Health Sciences, National Institutes of Health. This work was supported in part by the Agricultural Research Service, U.S. Department of Agriculture.
GAM conceived of the study, performed RAGE binding experiments, and wrote the article. LLE and HP provided RAGE and contributed to the writing. KJ, AS, LJD, and JGW performed mass spectrometry experiments and analysis. REL and KBT analyzed data and contributed to the writing. SJM, BKH, HC, SR, and JBN performed Western analysis, provided allergens, and contributed to the writing. AP contributed peanut extracts, analysis and worked on the writing. GAM and SJM contributed equally.