Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Evaluation of the discriminatory potential of antibodies created from synthetic peptides derived from wheat, barley, rye and oat gluten

  • David Poirier,

    Roles Conceptualization, Data curation, Formal analysis, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing

    Affiliations Department of Food Science and Nutrition, Pavillon Paul-Comtois, Université Laval, Québec, Québec, Canada, Institute of Nutrition and Functional Foods, Université Laval, Québec, Québec, Canada

  • Jérémie Théolier,

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review & editing

    Affiliations Department of Food Science and Nutrition, Pavillon Paul-Comtois, Université Laval, Québec, Québec, Canada, Institute of Nutrition and Functional Foods, Université Laval, Québec, Québec, Canada

  • Riccardo Marega,

    Roles Formal analysis, Methodology, Writing – review & editing

    Affiliation Analytical Laboratory, CER Groupe, Marloie, Belgium

  • Philippe Delahaut,

    Roles Writing – review & editing

    Affiliation Analytical Laboratory, CER Groupe, Marloie, Belgium

  • Nathalie Gillard,

    Roles Funding acquisition, Writing – review & editing

    Affiliation Analytical Laboratory, CER Groupe, Marloie, Belgium

  • Samuel Benrejeb Godefroy

    Roles Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – review & editing

    samuel.godefroy@fsaa.ulaval.ca

    Affiliations Department of Food Science and Nutrition, Pavillon Paul-Comtois, Université Laval, Québec, Québec, Canada, Institute of Nutrition and Functional Foods, Université Laval, Québec, Québec, Canada

Abstract

Celiac disease (CD) is triggered by ingestion of gluten-containing cereals such as wheat, barley, rye and in some cases oat. The only way for affected individuals to avoid symptoms of this condition is to adopt a gluten-free diet. Thus, gluten-free foodstuffs need to be monitored in order to ensure their innocuity. For this purpose, commercial immunoassays based on recognition of defined linear gluten sequences are currently used. These immunoassays are designed to detect or quantify total gluten regardless of the cereal, and often result in over or underestimation of the exact gluten content. In addition, Canadian regulations require a declaration of the source of gluten on the label of prepackaged foods, which cannot be done due to the limitations of existing methods. In this study, the development of new antibodies targeting discrimination of gluten sources was conducted using synthetic peptides as immunization strategy. Fourteen synthetic peptides selected from unique linear amino acid sequences of gluten were bioconjugated to Concholepas concholepas hemocyanin (CCH) as protein carrier, to elicit antibodies in rabbit. The resulting polyclonal antibodies (pAbs) successfully discriminated wheat, barley and oat prolamins during indirect ELISA assessments. pAbs raised against rye synthetic peptides cross-reacted evenly with wheat and rye prolamins but could still be useful to successfully discriminate gluten sources in combination with the other pAbs. Discrimination of gluten sources can be further refined and enhanced by raising monoclonal antibodies using a similar immunization strategy. A methodology capable of discriminating gluten sources, such as the one proposed in this study, could facilitate compliance with Canadian regulations on this matter. This type of discrimination could also complement current immunoassays by settling the issue of over and underestimation of gluten content, thus improving the safety of food intended to CD and wheat-allergic patients.

Introduction

Celiac disease is estimated to affect approximately 1% of the world’s population [1]. Symptoms of this condition are triggered by ingestion of cereals containing gluten, such as wheat, barley, rye, and in rare cases oats [2]. To date, adhering to a gluten-free diet is the preferred strategy for the prevention of celiac disease symptoms. A reliable method for gluten quantification is therefore necessary to determine the level of this component in foods. In 1979 the Codex Alimentarius adopted the Standard for Foods for Special Dietary Use for Persons Intolerant to Gluten. Revised in 2008, it states that a food not exceeding a gluten content of 20 mg/kg can be declared as gluten-free [3]. This same standard specifies that the preferred method for the determination of gluten is the R5 Méndez enzyme-linked immunosorbent assay (ELISA). This method has been endorsed by the AOAC as an official method and as a type I method by the Codex Committee of Methods of Analysis and Sampling since 2006 [4, 5]. A type I method “determines a value that can only be arrived at in terms of the method per se and serves by definition as the only method for establishing the accepted value of the item measured” [6].

Since then, several authors have highlighted the harmonization and performance issues of gluten quantification by ELISAs. More specifically for the type I method, it has been shown that the test response varied depending on the type of gluten detected (e.g. barley, rye, wheat) thus leading to over and underestimation [710]. Besides, the standard used to calibrate the immunoassay may also impact the results. Most authors agree on the need for better test calibration and better standards [1119]. However, the use of the most appropriate reference material is still extensively debated [1119]. The use of a reference material for each grain, namely wheat, barley, rye and oats, as a calibration standard makes it possible to correct or at least reduce this variability [9, 11, 20]. Appropriate calibration with prolamin or glutelin as the analyte has been shown to reduce the discrepancy between measured and actual amounts of gluten [8, 12]. However, this calibration is often impossible in food analysis, since the source of gluten, if present due to cross-contamination, is usually unknown. Therefore, being able to identify the source of gluten in a food sample would help in selecting the appropriate calibration standards. In addition, Canadian legislation requires the declaration of the source of gluten specifying the name of the original grain on the labels of prepackaged foods, which makes the availability of a method to differentiate gluten sources in food samples all the more essential [21]. The ability to distinguish sources of gluten would also increase the food supply of the population suffering from wheat allergy who is deprived of consuming rye, barley and oats because of current analytical limitations. The aim of this study was to create new antibodies capable of distinguishing between different sources of gluten. These new antibodies must therefore target unique linear or conformational epitopes belonging to wheat, barley, rye and oats. The immunization strategy for obtaining the current commercially available antibodies uses native wheat (Skerritt) or rye (R5) prolamins as immunogen. [22, 23]. However, the use of prolamin in its native form confined the reactivity of antibodies to only a few external antigenic sites [24]. G12, in contrast, developed against a 33-mer peptide of α-gliadin, has been identified as a primary initiator of the inflammatory response to gluten in celiac patients [25, 26]. The sensitivity of this monoclonal antibody (mAbs) to wheat, rye, barley and oats does not allow for its use to distinguish different sources of gluten [25]. However, the strategy of immunization using synthetic peptides is promising to obtain antibodies able to differentiate gluten sources. In this study, immunogens were developed by bioconjugating synthetic peptides from respectively unique amino acid sequences of wheat, barley, rye and oats to a carrier protein. Polyclonal antibodies (pAbs) were thus elicited in rabbits, and the evaluation of their relative sensitivity and specificity by indirect ELISA showed their ability to discriminate between different sources of gluten from wheat, barley, rye and oats.

Materials and methods

Materials

All chemicals and solvents were HPLC grade as a minimum. Ultrapure water was used for FPLC and buffers (Merck KGaA, Darmstadt, Germany). Grains of wheat cultivar (cv.) AAC harlaka, barley cv. AAC synergy and rye cv. danko were provided by Semican International Inc. (Plessisvile, QC, CA). Oat cv. Ruffian was provided by Avena Foods (Regina, AL, CA). Flours from rice, split pea, chickpea, millet, and soy were purchased from local stores.

Research and selection of sequences

Gluten protein sequences listed in Table 1 were searched and extracted from the National Center for Biotechnology Information (NCBI) database [27]. For each protein, a multiple alignment of all available sequences was performed using Multiple Sequence Alignment (MSA) tool from Clustal Omega [28]. Then, consensus sequences (CS) were obtained for each gluten protein types (GPT) using the sequence editor Jalview(2.11.0) [29, 30]. Pairwise sequence alignment was then performed on such CS with every combination of two proteins from Table 1 using Jalview(2.11.0) to identify shared and unshared amino acid (a.a.) strands. All unshared sequences from the CS of each grain equal or containing more than 6 a.a. were compiled and kept in a database. Prolamins from maize (zein) [31] and soy (glycinin) [32] were also compared with the CS to discard shared sequences between GPT and these prolamins. The Grand Average of Hydropathy (GRAVY) of the potential candidates as haptens was calculated according to Kyte and Doolittle (1982) [33]. Based on several parameters, such as the consecutive number of the same a.a., the occurrence of the sequence in the GPT and the number of a.a., 14 sequences were selected to produce immunogens.

thumbnail
Table 1. Prolamins and glutelins from gluten-containing grains.

https://doi.org/10.1371/journal.pone.0257466.t001

Hapten synthesis

To the 14 selected peptides, minor modifications were made to their sequences during their synthesis (Bio Basic Inc., Markam Ontario, Canada). The modifications regard the N-terminal addition of a cysteine (when none was present in the original sequence), in order to introduce a thiol moiety for the subsequent “click-coupling” with protein carriers surfacing maleimido moieties, and N-terminal acetylation / C-terminal amidation to better mimic the polypeptide structure.

Bioconjugation of immunogen

Each of the 14 synthetic peptides was coupled to hemocyanin from Concholepas concholepas (CCH) [34] (Blue Carrier, Biosonda S.A., Santiago, Chile) using the protocol described by Hermanson et al. (2008) [35] with minor modifications. Briefly, 20 mg of CCH were dissolved in 1 mL of coupling buffer (100 mM sodium phosphate, 0.3M NaCl, pH 7.4) to which were added 2 mg of the heterobifunctional cross-linker, sulfosuccinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate (sulfo-SMCC) (Thermo Fisher Scientific, MA, USA), previously dissolved in 200 μl of ultrapure water. The mixture was incubated for 30 minutes at room temperature on an orbital shaker (Fisher Scientific, MA, USA). Then, 20 μl of 100 mM glycine in coupling buffer were added to neutralize excess of sulfo-SMCC, followed by purification of the CCH-sulfo-SMCC complexes by FPLC (ÄKTA avant; GE Healthcare, IL, USA) on desalting columns (HiTrap; GE Healthcare, IL, USA). Ellman’s assays were conducted to ensure a good level of maleimide activation of CCH [35, 36], by using L-cysteine as external calibration curve. Maleimide-activated CCH was conjugated by 1.2-fold molar excess of synthetic peptides and allowed to react 1 hour at room temperature, and the peptide loading (nmol peptide/mg CCH) was indirectly obtained upon quantitation of the unreacted thiolated peptide (Ellman assay). The conjugates were snap frozen in liquid nitrogen and stored at -20°C until further use.

BSA bioconjugate

Each of the 14 synthetic peptides was coupled to UltraPure Bovine Serum Albumin (BSA) (Thermo Fisher Scientific, MA, USA) using the same protocol described above for bioconjugation of CCH, but with 4 mg of Mal-PEG4-NHS (Carbosynth, Compton, UK) as cross-linker.

Immunization of rabbits

All of the experimentation involving animals was done under the frame of the ethical protocol CE/Sante/E/001 (immunization and production of sera/polyclonal antibodies) approved by the ethical committee of CER Groupe (agreement nb. LA1800104). The agreement LA1800104 was bestowed by the Federal Public Service of the Walloon Region (Belgium). The experimentation respected the legislation in force at the moment of the studies, thus following the guidelines established at the European level (Directive 2010/63/EU revising Directive 86/609/EEC on the protection of animals used for scientific purposes), Belgian level (Arrêté royal relatif à la protection des animaux d’expérience, AR 2013/05/29), and Regional level (Code Wallon du Bien-être animal 03/10/2018). Each CCH-peptide conjugate was injected to three different rabbits for a total of 42 rabbits. Polyclonal antibodies were raised in rabbits by subcutaneous injection of 200 micrograms of CCH-peptide conjugates emulsified with Freund’s complete adjuvant for the first injection, or Freund’s incomplete adjuvant for all following injections (Becton Dickinson Benelux, Erembodegem, Belgium). Injections were administered on a fortnightly basis and then, from the third injection onward, at the rhythm of one injection every 28 days. Test bleeds were collected 10 days after each immunization (from the third immunization onward). The blood was centrifuged, and the collected serum was stored at -20°C until used. An aliquot of such serum was diluted 1/10 in a solution of 50% assay buffer (phosphate buffer 65 mM,NaCl 150 mM, 0.2% gelatin, 0.05% Tween 20, 0.01%, 8-anilino-1-naphthalenesulfonic acid ammonium salt, and ascorbic acid 28 mM) and 50% ethyleneglycol, yielding diluted pAbs solutions that were kept at -20°C prior to their use for titer and specificity assessment.

Preparation of gluten protein fractions

Wheat, barley, rye and oat grains were turned into flours with a Grindomix GM 200 (Retsch, Haan, NRW, DE). Flours purity was confirmed by PCR using primers for wheat, barley, rye and oat designed by Sandberg et al. (2003) [37]. Defatting of the flours and extraction of the different prolamins was performed according to Schalk et al. (2017) [38]. The solvent used for defatting was replaced by hexane [39].

Titer determination by indirect ELISA

Microtiter plates (F8 Maxisorp Nunc-Immuno Module, 96-well plates, Thermo Scientific) were coated overnight at room temperature with 1 μg/mL BSA-peptide bioconjugate diluted in 0.05 M carbonate-bicarbonate buffer, pH 9.6 (Fisher Scientific, Pgh, USA). After being washed three times with washing buffer (0.15 M NaCl and 0.05% Tween 20), 250 μl of blocking buffer (PBS with 1% gelatin) were added to each well and the plates were incubated 2 hours at 37°C. After washing, 100 μl per well of a serial dilution (1:2 000 to 1:256 000 overall) of the previously yielded pAbs solutions in assay buffer were added, along with non-specific binding, and incubated for 1 hour at 37°C. Plates were washed again and 100 μl per well of anti-Rabbit IgG (whole molecule)–peroxidase antibody produced in goat (Sigma, Saint-Louis, USA, reference# A6154-1ML) diluted 1:10,000 in assay buffer was added. After another three washings step, the chromogenic substrate 3,3′,5,5′-tetramethylbenzidine (TMB, D-Tek, Belgium) was added, and the wells were incubated at room temperature. After 30 minutes in the dark, the reaction was stopped with 1.8 N H2SO4 and optic density (O.D.) was measured at 450 nm in a microplate reader (Multiskan FC; Thermo Fisher Scientific, MA, USA). Titer value was defined as the highest dilution which triggered an O.D. value of 1.000 at 450 nm.

Relative sensitivity and specificity tests

The same protocol used for titer determination was used for relative sensitivity and specificity tests against gluten prolamins in native and denatured state, and against rice, split pea, chickpea, millet, and soy flours with modifications regarding the coating of the microtiter plate. Coating with native prolamins was performed by firstly dissolving previously prepared prolamins into 60% (v/v) ethanol at a concentration of 25 mg/mL, and then coating the microplate with a 50 μg/mL native prolamin solution diluted in carbonate-bicarbonate buffer, pH 9.6. Denaturation of prolamins was performed at 25 mg/mL using cocktail solution according to García et al. (2005) [40]. The microplates were coated at 50 μg/mL with a denatured prolamin solution in carbonate-bicarbonate buffer, pH 9.6. The same dilutions of pAbs used for the relative sensitivity tests were used for specificity tests. Results were expressed in intervals of the strongest signal obtained for each serum. The cut-off threshold for discriminating background noise was determined based on the average of several blanks and was statistically fixed at 99.9% confidence level using Frey et al.’s (1998) endpoint titer determination method for immunoassays [41].

Results and discussion

Selection of the sequences

The bioinformatic strategy for the selection of sequences to be used as immunogens was based on an MSA combined to a CS (S1 Table). This strategy is the foundation of the results presented in this study and therefore needs to be discussed. Among each GPT, there are often small differences due to substitution, deletion and/or insertion of nucleotides, which contribute to the heterogeneity within each type [38]. Thus, the NCBI database contains a wide range of sequences even within a same GPT. It was therefore important for the present work to be able to separate non-well-preserved regions from well-preserved ones as the former are more representative of the GPT range. When aligning sequences with shared origin such as GPT, the use of MSA allows to effectively identify conserved residues in the dataset [42, 43]. Compiling these aligned patterns results in a CS, defined as a compiled sequence of the most common a.a. at each position [29, 44]. The execution of the MSA before the creation of the CS makes the CS model more accurate [29]. It has also been shown that the use of CS to create an immunogen is an effective strategy that minimizes the degree of variable elements within a created immunogen [45]. There are, however, limitations to the use of this bioinformatics strategy. Indeed, a CS is generated independently of the conserved or unpreserved character of the GPT. The quality of the CS directly depends on the quality of the MSA performed upstream. This therefore creates segments of CS showing notable dissimilarity with GPT, and others that are conservative. However, for large data sets, it is difficult to determine which segment of the sequence is from variable or conserved regions, which can result in the selection of less representative synthetic peptides from the GPT. Shared patterns recognized through the pairwise alignment performed on each CS, led to identification of unique sequences for each GPT. Every unshared sequence containing more than 6 a.a., the length that has been showed to consistently elicit antibodies that bind to the original protein, were conserved (158 unique sequences) [46]. The workflow of the subsequent steps is presented in Fig 1. After elimination of cross-reacting sequences with prolamins of corn, soy and rice, the remaining sequences were divided based on their hydropathy index. Hydrophilic sequences were conserved because of (1) their bioconjugation capacity in biological buffers such as PBS, and (2) their hydrophilic surface-oriented epitopes would be accessible to antibodies [4650]. However, well-conserved sequences are most often found in the internal structure of the protein molecule and are therefore more hydrophobic [51]. Nevertheless, the authors estimated that the selection of hydrophilic peptides instead of hydrophobic peptides was an adequate strategy for facilitating the creation of the immunogen, and for maximizing the final recognition rate of the antibodies generated, by presenting an antigen to the host with a loose and more linear structure [4648]. These two constraints make the selection of synthetic peptides challenging. The selection strategy used in this study makes the well-conserved segments of the GPT more reliable. However, the majority of these segments are hydrophobic and therefore less suitable for antibodies production. Conversely, the use of hydrophilic segments as synthetic peptides is less reliable, since they are mostly in less well-preserved segments of the GPT, but is structurally more suitable to raise antibodies. Furthermore, it is known that specificity to an antigen can be highly modified by a change as small as a single a.a. substitution [52]. Thus, the choice of more hydrophilic sequences as immunogen may have caused a decrease in the specificity of the antibodies generated towards the GPT of the respective grains.

The hydropathy index-based selection led to 116 unique sequences. Among them, 52 sequences with too many consecutive repetitions of a same a.a. were rejected for specificity reasons. The remaining 64 sequences were ranked according to their occurrence in their respective GPT’s consensus sequences, their number of a.a., and the proportion of their respective GPT in gluten of wheat, barley, rye and oat based on Schalk et al. (2017) data [38]. If the site of the selected peptide is poorly exposed or even criptic, the raised antibody is more likely to be unable to recognize the native protein [53]. Thus, N- or C-termini external, charged and polar regions are often good choices for the selected peptides [49]. Nevertheless, in the present study, the selected peptides firstly aimed to discriminate GPT homologically close, regardless of their conformations, gluten extraction is not always performed under denaturing conditions [40, 54, 55]. Peptides with more than 8 a.a were prioritized to improve recognition of the original protein [48]. On the other hand, peptides longer than 20 a.a were discarded as they may adopt conformations that will no longer resemble the origin protein, leading to poorer specificity [48]. These considerations restrained the choice to 14 sequences (six for wheat, four for barley, three for rye and one for oat) that were selected to produce immunogens to raise pAbs.

Titer determination

Except for rabbits 24 and 41, which died before the first bleed, titer determinations were performed on the pre immune sera and on test bleed number 4, in order to maximize the occurrence of stable titers levels. It should be pointed out that the peptide-BSA conjugates used for these ELISA assessments further differ from the corrresponding CCH conjugates in light of the different spacer used in the heterobifunctional crosslinker (tetraethyleneglycol instead of cyclohexyl), which ensures that reactive antibodies are specific to the peptides.

Results of the indirect ELISA (Fig 2) against synthetic peptides used for the titer determination reveal titer determinations higher than 1:10 000 for all the sera with the exception of those obtained from immunogens produced with peptides 10 (barley) and 13 (Rye). Very high dilutions (above 1:96 000) were obtained for the immunogens produced with peptides 2 and 5 (wheat), and peptide 9 (barley). Equivalent values could not be obtained for any immunogen containing rye or oat peptides.

thumbnail
Fig 2. Titer determination by indirect ELISA against 1μg/mL of BSA-(Mal-PEG4-NHS)-peptide (1–14).

Highest sera dilution triggering an O.D. value of ≈1.000 at 450 nm. Wheat peptide (01–06); Rabbits (01–18). Barley peptides (07–10); Rabbits (19–30). Rye peptides (11–13); Rabbits (31–39). Oat peptide (14); Rabbits (40–42). Rabbits 24 and 41 died before test bleed 1.

https://doi.org/10.1371/journal.pone.0257466.g002

Variations in antibody production or affinity among the same immunogen can be explained by biological variation between different hosts and initial immunogen processing by B-cells, which result in differences in raised pAbs [56, 57]. Titers below or equal to 1:4 000 were discarded since pAbs production was not considered strong enough for the purpose of this study: rabbit 9 (wheat), rabbit 28 (barley) and rabbit 31 (rye).

Sensibility and specificity tests

Recognition of the GPT from which the synthetic peptides are sourced is not guaranteed as generation of antibodies is still not completely understood and is still based on empirical processes [46, 47, 53]. Screening results by indirect ELISA (Table 2) reveal low sensitivity for most of the pAbs against native and/or denatured GPT. These low recognition rates for certain prolamins contrast with the titers obtained against the BSA conjugates even if an approximately 10-fold molar excess was used for the prolamin coating over the peptides of the BSA conjugates, based on the MW of the different GPT.

For wheat, rabbits 1 to 6 and rabbit 18 present results above 1:8 000 against at least one of the two forms of GPT. On the other hand, sensitivity against native or denatured proteins collapsed for rabbit 7 to rabbit 17. For example, sera obtained with peptide 5 presented good results against the BSA bioconjugate, but negative or lower than 1:2 000 for the corresponding GPT. For barley, only the immunogen made with peptide 9 produced pAbs sensitive for the native form of the GPT. These results confirm the high titers obtained against the BSA bioconjugates (1:128 000 to 1:256 000). The sensitivities of all barley pAbs were not strong enough against the denatured form of the GPT with results equal to or lower than 1:2 000. For rye, the raised pAbs were not sensitive enough for either form of the GPT, which reflect the lower titer obtained with the related synthetic peptides. For oat, peptide 14 resulted in equally sensitive pAbs against either form of the GPT. Rabbit 42 gave pAbs even more reactive with the GPT (1:64 000) than the synthetic peptide (1:48 000).

These results were expected since the epitope occurrence is relatively higher in the BSA bioconjugate, covered with synthetic peptides, than in the corresponding GPT. Several other factors can also explain these results. For native GPT, the targeted sequence is not necessarily present at the protein surface, limiting or preventing the fixation of the pAbs. For denatured GPT, even if the protein structure is disturbed, the targeted sequence could present a conformation that makes it unrecognizable by the pAbs [58]. It has been discussed earlier that hydrophilic peptides were selected since ideal antigenic haptens are hydrophilic and surface-oriented, making the epitopes exposed to the solvent [4650]. However, hydropathy index of a peptide does not necessarily reflect the conformation of the corresponding stretch in the original protein. Indeed, secondary structure such as α-helices and β-sheets can hinder recognition by antibodies of the original protein, since linear peptides are able to assume a more random structure [58, 59]. These structures are inherently present in the native form of the prolamins, but also in the denatured state. It has been demonstrated that even in strong denaturant, some structures remain folded at some extent [60]. In milder conditions, such as in the buffer of the sensibility tests, the state of most denatured proteins would present considerable secondary structures [60]. Further analysis such as X-ray biocrystallography, NMR Spectroscopy or cryo-electron microscopy should be performed to better evaluate the repartition of these macromolecular structures in the native and denatured state of the GPT [6163].

Only pAbs giving an O.D. of 1.000 at a dilution equal or higher than 1:4 000 for either native or denatured form of the prolamins were conserved. Then, only 11 sera were tested for potential cross-reactivity from the 37 tested (i.e. Rabbits: 01–06, 18, 25, 27, 40 and 42). Specificity tests results are shown in Table 3. Preliminary tests were conducted against wheat, barley, rye and oat prolamins. All the sera presented high specificity for their original prolamins, with the notable exception of rabbits 02 and 03, which were produced with a peptide originating from wheat but cross-reacted with all the prolamins independently of the cereal. For this reason, these two sera were discarded from the second phase of the cross-reaction tests. Regarding the other sources of proteins, results of the cross-reaction testing show that some cross-reaction within multiple matrices occurred with rabbits 06 and 40. The pAbs from rabbit 06 cross-reacted with chickpea, while those from rabbit 40 did so with rice, split-pea and chickpea. Highly similar a.a. profiles between the different GPT and the ability of antibodies to recognize an extensive number of related epitopes, especially in pAbs, may explain these cross-reactions [64, 65]. Epitope mapping of the cross-reacting pAbs could point out the shared reactive sites [64, 66].

Several sera presented interesting results. One of these pAbs was specific to native and denatured wheat prolamins (rabbit 4: peptide 2 from α/β-gliadins). Another was specific to native barley prolamins (rabbit 25: peptide 9 from γ3-hordeins) with a residual reaction for oats, which could be discarded by immunoaffinity columns combined with epitope mapping [46, 66]. Finally, rabbits 40 and 42 (peptide 14 from avenins) presented specificity to both forms of oat prolamins. No specific pAbs were obtained against rye, but a pAbs with similar reactivity to native and denatured prolamins of wheat and rye was obtained (rabbit 18: peptide 6 from HMW-GS of wheat).

As no purification of the pAbs was done beforehand, rabbits serum proteins could have contributed to the residual signal, since the secondary antibody used in all the indirect ELISAs was a labelled goat anti-rabbit antibody. Moreover, the blanks used to determine the cut-off values were based on assay buffer, which does not contain rabbit serum to avoid nonspecific binding by the secondary antibody. Thus, this can slightly underestimate the threshold values when using Frey et al.’s (1998) statistical model, which might explain the presence of some negligible signals (+/-) [41]. In order to produce ELISA tests, a purification of the antibodies by protein A affinity would be necessary in order to overcome those signals [67].

Conclusions

The immunization strategy based on synthetic peptides used in this study led to the identification of pAbs specific to native and denatured forms of wheat and oat prolamins, as well as pAbs specific to native barley prolamins. This strategy can therefore be regarded as a promising tool for the development of specific antibodies, thus facilitating compliance with Canadian regulations on the declaration of gluten sources in foods. In addition, this would directly benefit the wheat-allergic and coeliac populations by increasing their dietary options. Indeed, the R5, the method currently used for the detection of gluten and of traces of wheat in food, has a significant limitation, as it also measures barley and rye. This means that patients allergic to wheat have no option but to ban rye and barley from their diet. In addition, since R5 does not allow for detection of oats, the results of this study are also promising for European and Australian jurisdictions, which require declaration of oats on prepackaged food labels [68, 69]. Detection and quantification of the source of gluten can also complement current immunoassays by setting the issue of over and underestimation of gluten content. Further work should be conducted to raise an antibody specific to rye prolamins, since the one obtained in this study was equally specific to wheat and rye prolamins. Nonetheless, the latter could see applications in combination with the other specific pAbs to differentially determine the source of gluten [70]. Further developments for the creation of monoclonal antibodies based on the immunization strategy presented here are planned, as well as further experimentations against processed food products. This ongoing work is expected to overcome the residual cross-reactivity between the various GPT, since pAbs are characterized by a lower specificity, as compared to monoclonal antibodies (i.e. epitope-specific) [46] and could pave the way for the production of new ELISA assays.

Supporting information

S1 Table. Consensus sequences obtained from multiple alignments sequences (MSA) and compilation.

https://doi.org/10.1371/journal.pone.0257466.s001

(DOCX)

Acknowledgments

Many thanks to Virginie Barrere who helped with the PCR protocols, Anne-Catherine Huet who provided technical assistance for the immunoassays, Sylvia Dominguez who helped with the language correction and Semican International Inc. and Avena Foods who generously supplied the grains used in this study.

References

  1. 1. Singh P, Arora A, Strand TA, Leffler DA, Catassi C, Green PH, et al. Global Prevalence of Celiac Disease: Systematic Review and Meta-analysis. Clin Gastroenterol Hepatol. 2018;16: 823–836.e2. pmid:29551598
  2. 2. La Vieille S, Pulido OM, Abbott M, Koerner TB, Godefroy S. Celiac Disease and Gluten-Free Oats: A Canadian Position Based on a Literature Review. In: Canadian Journal of Gastroenterology and Hepatology [Internet]. 2016 [cited 26 Mar 2018]. pmid:27446825
  3. 3. CODEX Alimentarius. STANDARD FOR FOODS FOR SPECIAL DIETARY USE FOR PERSONS INTOLERANT TO GLUTEN. FAO-WHO; 2015 p. 3. Report No.: CXS 118–1979. http://www.fao.org/fao-who-codexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FStandards%252FCXS%2B118-1979%252FCXS_118e_2015.pdf
  4. 4. AOACI. 2012.01: Gliadin as a Measure of Gluten in Foods Containing Wheat, Rye, and Barley. Enzyme Immunoassay Method. 2012. http://www.aoacofficialmethod.org/index.php?main_page=product_info&cPath=1&products_id=2965
  5. 5. CCMAS. Distribution of the Report of the 27th Session of the Codex Committee on Methods of Analysis and Sampling. Budapest, Hungary: CODEX ALIMENTARIUS COMMISSION; 2006 May p. 52. Report No.: ALINORM 06/29/23. http://www.fao.org/fao-who-codexalimentarius/meetings/detail?meeting=CCMAS&session=27
  6. 6. CCMAS. Comprehensive guidance for the process of submission, consideration and endorsement of methods for inclusion in CXS234. Budapest, Hungary; 2019 p. 8. Report No.: MAS/40 CRD/27. http://www.fao.org/fao-who-codexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FMeetings%252FCX-715-40%252FCRD%252Fmas40_CRD27x.pdf
  7. 7. Diaz-Amigo C, Popping B. Accuracy of ELISA detection methods for gluten and reference materials: a realistic assessment. J Agric Food Chem. 2013; 5681–5688. pmid:23713744
  8. 8. Kanerva PM, Sontag-Strohm TS, Ryoppy PH, Alho-Lehto P, Salovaara HO. Analysis of barley contamination in oats using R5 and omega-gliadin antibodies. J Cereal Sci. 2006; 347–352.
  9. 9. Lexhaller B, Tompos C, Scherf KA. Comparative analysis of prolamin and glutelin fractions from wheat, rye, and barley with five sandwich ELISA test kits. Anal Bioanal Chem. 2016;408: 6093–6104. pmid:27342795
  10. 10. Wehling P, Scherf KA. Preparation of Validation Materials for Estimating Gluten Recovery by ELISA According to SMPR 2017.021. J AOAC Int. 2020;103: 210–215. pmid:31277725
  11. 11. Martinez-Esteso MJ, Brohee M, Norgaard J, O’Connor G. Label-free proteomic analysis of wheat gluten proteins and their immunoreactivity to ELISA antibodies. Cereal Chem. 2017; 820–826.
  12. 12. Lexhaller B, Tompos C, Scherf KA. Fundamental study on reactivities of gluten protein types from wheat, rye and barley with five sandwich ELISA test kits. Food Chem. 2017;237: 320–330. pmid:28764003
  13. 13. Bugyi Z, Török K, Hajas L, Adonyi Z, Poms RE, Popping B, et al. Development of Incurred Reference Material for Improving Conditions of Gluten Quantification. J AOAC Int. 2012;95: 382–387. pmid:22649923
  14. 14. Diaz-Amigo C, Panda R, Koerner T, Downs M, Poms RE. Analysis of gluten in foods: where are we and where do we need to go? J Food Prot. 2016.
  15. 15. Panda R, Zoerb HF, Cho CY, Jackson LS, Garber EAE. Detection and quantification of gluten during the brewing and fermentation of beer using antibody-based technologies. J Food Prot. 2015; 1167–1177. pmid:26038908
  16. 16. Rzychon M, Brohée M, Cordeiro F, Haraszi R, Ulberth F, O’Connor G. The feasibility of harmonizing gluten ELISA measurements. Food Chem. 2017;234: 144–154. pmid:28551218
  17. 17. Sharma GM. Immunoreactivity and Detection of Wheat Proteins by Commercial ELISA Kits. J AOAC Int. 2012;95: 364–371. pmid:22649920
  18. 18. Slot IDB, van der Fels-Klerx HJ, Bremer MGEG, Hamer RJ. Immunochemical Detection Methods for Gluten in Food Products: Where Do We Go from Here? Crit Rev Food Sci Nutr. 2016;56: 2455–2466. pmid:25779856
  19. 19. Torok K, Hajas L, Bugyi Z, Balazs G, Tomoskozi S. Investigation of the effects of food processing and matrix components on the analytical results of ELISA using an incurred gliadin reference material candidate. Acta Aliment. 2015; 390–399.
  20. 20. Haraszi R, Ikeda TM, Peña RJ, Branlard G. Gluten Analysis. In: Igrejas G, Ikeda TM, Guzmán C, editors. Wheat Quality For Improving Processing And Human Health. Cham: Springer International Publishing; 2020. pp. 109–143.
  21. 21. Goverment of Canada. Consolidated federal laws of canada, Food and Drug Regulations. 16 Dec 2019 [cited 23 Apr 2020]. https://laws-lois.justice.gc.ca/eng/regulations/c.r.c.%2C_c._870/index.html
  22. 22. Skerritt JH, Smith RA, Wrigley CW, Underwood PA. Monoclonal antibodies to gliadin proteins used to examine cereal grain protein homologies. J Cereal Sci. 1984;2: 215–224.
  23. 23. Sorell L, López JA, Valdés I, Alfonso P, Camafeita E, Acevedo B, et al. An innovative sandwich ELISA system based on an antibody cocktail for gluten analysis. FEBS Lett. 1998;439: 46–50. pmid:9849874
  24. 24. Lerner RA. Tapping the immunological repertoire to produce antibodies of predetermined specificity. Nature. 1982;299: 592–596. pmid:6181415
  25. 25. Moron B, Cebolla A, Manyani H, Alvarez-Maqueda M, Megias M, del Carmen-Thomas M, et al. Sensitive detection of cereal fractions that are toxic to celiac disease patients by using monoclonal antibodies to a main immunogenic wheat peptide. J Clin Nutr. 2008; 405–414. pmid:18258632
  26. 26. Shan L, Molberg Ø, Parrot I, Hausch F, Filiz F, Gray GM, et al. Structural Basis for Gluten Intolerance in Celiac Sprue. Science. 2002;297: 2275–2279. pmid:12351792
  27. 27. Wieser H. Chemistry of gluten proteins. Food Microbiol. 2007;24: 115–119. pmid:17008153
  28. 28. Madeira F, Park Y mi, Lee J, Buso N, Gur T, Madhusoodanan N, et al. The EMBL-EBI search and sequence analysis tools APIs in 2019. Nucleic Acids Res. 2019;47: W636–W641. pmid:30976793
  29. 29. Gribskov M. Identification of Sequence Patterns, Motifs and Domains. In: Ranganathan S, Gribskov M, Nakai K, Schönbach C, editors. Encyclopedia of Bioinformatics and Computational Biology. Oxford: Academic Press; 2019. pp. 332–340.
  30. 30. Waterhouse AM, Procter JB, Martin DMA, Clamp M, Barton GJ. Jalview Version 2—a multiple sequence alignment editor and analysis workbench. Bioinformatics. 2009;25: 1189–1191. pmid:19151095
  31. 31. Coleman CE, Larkins BA. The Prolamins of Maize. In: Shewry PR, Casey R, editors. Seed Proteins. Dordrecht: Springer Netherlands; 1999. pp. 109–139.
  32. 32. Katsube T, Kurisaka N, Ogawa M, Maruyama N, Ohtsuka R, Utsumi S, et al. Accumulation of Soybean Glycinin and Its Assembly with the Glutelins in Rice. Plant Physiol. 1999;120: 1063–1074. pmid:10444090
  33. 33. Kyte J, Doolittle RF. A simple method for displaying the hydropathic character of a protein. J Mol Biol. 1982;157: 105–132. pmid:7108955
  34. 34. Oliva H, Moltedo B, De Ioannes P, Faunes F, De Ioannes AE, Becker MI. Monoclonal Antibodies to Molluskan Hemocyanin from Concholepas concholepas Demonstrate Common and Specific Epitopes among Subunits. Hybrid Hybridomics. 2002;21: 365–374. pmid:12470479
  35. 35. Hermanson GT. Bioconjugate techniques. 2nd ed. Amsterdam: Elsevier Academic Press; 2008. http://ariane.ulaval.ca/cgi-bin/recherche.cgi?qu=i9780123705013
  36. 36. Riddles PW, Blakeley RL, Zerner B. Ellman’s reagent: 5,5′-dithiobis(2-nitrobenzoic acid)—a reexamination. Anal Biochem. 1979;94: 75–81. pmid:37780
  37. 37. Sandberg M, Lundberg L, Ferm M, Yman IM. Real time PCR for the detection and discrimination of cereal contamination in gluten free foods. Eur Food Res Technol. 2003; 344–349.
  38. 38. Schalk K, Lexhaller B, Koehler P, Scherf KA. Isolation and characterization of gluten protein types from wheat, rye, barley and oats for use as reference materials. PLOS ONE. 2017;12: e0172819. pmid:28234993
  39. 39. Megahad OA, Kinawy OSE. Studies on the extraction of wheat germ oil by commercial hexane. Grasas Aceites. 2002;53: 414–418.
  40. 40. García E, Llorente M, Hernando A, Kieffer R, Wieser H, Méndez E. Development of a general procedure for complete extraction of gliadins for heat processed and unheated foods. Eur J Gastroenterol Hepatol. 2005;17: 529–539. pmid:15827444
  41. 41. Frey A, Di Canzio J, Zurakowski D. A statistically defined endpoint titer determination method for immunoassays. J Immunol Methods. 1998;221: 35–41. pmid:9894896
  42. 42. Thompson JD, Linard B, Lecompte O, Poch O. A Comprehensive Benchmark Study of Multiple Sequence Alignment Methods: Current Challenges and Future Perspectives. PLOS ONE. 2011;6: e18093. pmid:21483869
  43. 43. Edgar RC, Batzoglou S. Multiple sequence alignment. Curr Opin Struct Biol. 2006;16: 368–373. pmid:16679011
  44. 44. Liljas L. Consensus Sequences. In: Maloy S, Hughes K, editors. Brenner’s Encyclopedia of Genetics (Second Edition). San Diego: Academic Press; 2013. pp. 163–164.
  45. 45. Ramanathan MP, Kutzler MA, Kuo Y-C, Yan J, Liu H, Shah V, et al. Coimmunization with an optimized IL15 plasmid adjuvant enhances humoral immunity via stimulating B cells induced by genetically engineered DNA vaccines expressing consensus JEV and WNV E DIII. Vaccine. 2009;27: 4370–4380. pmid:19497647
  46. 46. Greenfield EA. Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory Press; 2014. https://books.google.ca/books?id=URpxnAEACAAJ
  47. 47. Lee B-S, Huang J-S, Jayathilaka LP, Lee J, Gupta S. Antibody Production with Synthetic Peptides. Methods Mol Biol Clifton NJ. 2016;1474: 25–47. pmid:27515072
  48. 48. Hancock DC, OReilly NJ. Synthetic Peptides as Antigens for Antibody Production. Immunochemical Protocols. Humana Press; 2005. pp. 13–25.
  49. 49. Rubinstein ND, Mayrose I, Halperin D, Yekutieli D, Gershoni JM, Pupko T. Computational characterization of B-cell epitopes. Mol Immunol. 2008;45: 3477–3489. pmid:18023478
  50. 50. Kahlenberg F, Sanchez D, Lachmann I, Tuckova L, Tlaskalova H, Méndez E, et al. Monoclonal antibody R5 for detection of putatively coeliac-toxic gliadin peptides. Eur Food Res Technol. 2006;222: 78–82.
  51. 51. Lodish H, Berk A, Zipursky SL, Matsudaira P, Baltimore D, Darnell J. Hierarchical Structure of Proteins. Mol Cell Biol 4th Ed. 2000 [cited 12 Oct 2020]. https://www.ncbi.nlm.nih.gov/books/NBK21581/
  52. 52. Skovbjerg H, Koch C, Anthonsen D, Sjöström H. Deamidation and cross-linking of gliadin peptides by transglutaminases and the relation to celiac disease. Biochim Biophys Acta BBA—Mol Basis Dis. 2004;1690: 220–230. pmid:15511629
  53. 53. Trier NH, Hansen PR, Houen G. Production and characterization of peptide antibodies. Methods. 2012;56: 136–144. pmid:22178691
  54. 54. Fallahbaghery A, Zou W, Byrne K, Howitt CA, Colgrave ML. Comparison of Gluten Extraction Protocols Assessed by LC-MS/MS Analysis. J Agric Food Chem. 2017;65: 2857–2866. pmid:28285530
  55. 55. Gessendorfer B, Wieser H, Koehler P. Optimisation of a solvent for the complete extraction of prolamins from heated foods. J Cereal Sci. 2010;52: 331–332.
  56. 56. Jensen PE. Recent advances in antigen processing and presentation. Nat Immunol. 2007;8: 1041–1048. pmid:17878914
  57. 57. Vyas JM, Van der Veen AG, Ploegh HL. The known unknowns of antigen processing and presentation. Nat Rev Immunol. 2008;8: 607–618. pmid:18641646
  58. 58. Grant GA. Synthetic Peptides for Production of Antibodies that Recognize Intact Proteins. Curr Protoc Mol Biol. 2002;59. pmid:18265297
  59. 59. Gromiha MM. Chapter 5—Protein Structure Prediction. In: Gromiha MM, editor. Protein Bioinformatics. Singapore: Academic Press; 2010. pp. 143–207.
  60. 60. Dill KA, Shortle D. Denatured states of proteins. Annual Review of Biochemistry. 1991. pp. 795–825. pmid:1883209
  61. 61. Wüthrich K. Protein structure determination in solution by NMR spectroscopy. J Biol Chem. 1990;265: 22059–22062. pmid:2266107
  62. 62. Ilari A, Savino C. Protein Structure Determination by X-Ray Crystallography. In: Keith JM, editor. Bioinformatics: Data, Sequence Analysis and Evolution. Totowa, NJ: Humana Press; 2008. pp. 63–87.
  63. 63. Fernandez-Leiro R, Scheres SHW. Unravelling biological macromolecules with cryo-electron microscopy. Nature. 2016;537: 339–346. pmid:27629640
  64. 64. Van Regenmortel MHV. What Is a B-Cell Epitope? In: Schutkowski M, Reineke U, editors. Epitope Mapping Protocols: Second Edition. Totowa, NJ: Humana Press; 2009. pp. 3–20.
  65. 65. Wild D. The Immunoassay Handbook: Theory and Applications of Ligand Binding, ELISA and Related Techniques. Elsivier. Newnes; 2013.
  66. 66. Tye-Din JA, Stewart JA, Dromey JA, Beissbarth T, van Heel DA, Tatham A, et al. Comprehensive, quantitative mapping of T cell epitopes in gluten in celiac disease. Sci Transl Med. 2010;2. pmid:20650871
  67. 67. Hober S, Nord K, Linhult M. Protein A chromatography for antibody purification. J Chromatogr B. 2007;848: 40–47. pmid:17030158
  68. 68. Regulation (EU) No 1169/2011 of the European Parliament and of the Council on the provision of food information to consumers. Sect. Annex II—M2 (1), No 1169/2011 Oct 25, 2011. http://data.europa.eu/eli/reg/2011/1169/2018-01-01
  69. 69. Australia New Zealand Food Standards Code—Standard 1.2.3—Information requirements—warning statements, advisory statements and declarations. Sect. 1.2.3–4 Mandatory declaration of certain foods or substances in food, F2017C00418. https://www.legislation.gov.au/Details/F2017C0041870.
  70. 70. Marega R, Desroche N, Huet A-C, Paulus M, Suarez Pantaleon C, Larose D, et al. A general strategy to control antibody specificity against targets showing molecular and biological similarity: Salmonella case study. Sci Rep. 2020; 10, Article number: 18439. pmid:33116156