Procedimiento Metodológico para la elaboración de un estudio de

As an initial attempt at isolating and identifying the undigested gluten peptides that are implicated as triggers of gluten related diseases, it was theorized that a normal human in-vivo enzymatic digestion of food could be simulated in an in-vitro fashion. This would produce a mixture of peptides from digested food similar to that found in the gut of patients with celiac disease. Store bought wheat gluten flour was used and was digested using the major human gastric and pancreatic digestive enzymes (pepsin, trypsin, chymotrypsin, elastase and carboxypeptidase A). Normal

in-vivo physiological conditions (i. e. ionic concentrations, pH, temperature, digestion

time) were simulated as closely as possible. In addition, recombinant prolyl endopeptidase (PEP) and rat brush border membrane enzymes (BBM) were employed in order to identify any proline-containing peptides that resulted from further proteolysis.

All extracts of proteolyzed gluten were analyzed using RP-HPLC with UV and ion trap-MSn detection. The resulting chromatograms were complicated and difficult to interpret because there were so many peaks present. A representative LC-MS total ion chromatogram (TIC) of wheat gluten flour proteolyzed with PTCECA is shown in Figure 2.1. Such complicated LC chromatograms and MS spectra were not unexpected due to the known complexity of the initial gluten matrix. However, over the 60 minute analytical time scale, it was difficult to monitor individual peaks

because most peaks represented mixtures of co-eluting peptides. Overall, it was noted from data from the pepsin proteolysis that no significant differences in the patterns of gluten peptides were observed over the entire course of treatment. The trypsin and chymotrypsin digestion experiments, however, did result in the

observation of a shift of LC peaks towards shorter retention times, indicating the peptides had been digested into smaller peptides. Some further shifts toward even shorter retention times were observed from the data from the treatment with elastase and carboxypeptidase A, but not nearly as much as was observed following the trypsin/chymotrypsin treatment. Upon addition of prolyl endopeptidase (PEP), even shorter retention time LC peaks appeared, along with the reduction and complete disappearance of some peptides. The profile of these peaks represents P-containing peptides. This is illustrated in Figures 2.2(A) and (B). The LC-MS total ion

chromatogram shows the degradation of one particular peak, (labeled with an arrow, in Figure 2.2(A)), from a large peak to a much smaller peak (Figure 2.2(B)) after the PEP was added. As expected, data from proteolyzed extracts that had been treated with PEP and the BBM enzymes showed that the only remaining peaks eluted quickly, presumably representing very small gluten peptides.

“PQ” sequences are present in every known immunogenic epitope (Table 1.1). An interesting means by which to identify the peptides that contained “PQ” sequences was to conduct a neutral loss scan. The neutral loss experiment involved performing CID and identifying the molecular weights of peptides that have lost PQ. It was theorized that this type of MS scan could simplify this enzymatic digestion data and make it easier to detect new potentially immunogenic PQ-containing target

peptides. Figures 2.3(A) and (B) show results from a representative neutral loss scan for “PQ” in proteolyzed wheat gluten flour. The peak labeled with an arrow in Figure 2.3(A) has totally disappeared after treatment with PEP, indicating it is a PQ- containing peptide that resisted complete proteolysis until the treatment with PEP. This indicated that this peptide may be physiologically relevant. This technique was able to identify some peptides that appeared to contain PQ sequences, but the sensitivity was quite low. In addition, it appeared to only identify those peptides where the PQ sequence was at the C-terminal end of the peptide.

Upon further evaluation of the data from these experiments the feasibility of this in-vitro enzymatic digestion approach towards the identification of immunogenic peptides from wheat gluten was proven in two ways. First, upon screening extracted ion chromatograms (EICs) from the data from proteolyzed wheat gluten for the presence of known immunogenic epitopes and peptides, the major immunodominant gluten peptide (α_{G-33) was discovered to be present. This is illustrated in Figure 2.4.}

The LC-MS total ion chromatogram in Figure 2.4(A) shows the various peptides that were released via the proteolysis procedure. The peak which is labeled at the

retention time of approximately 33 minutes corresponds to the 33mer peptide. Figure 2.4(B) shows the full scan MS spectrum of this peak, where the triply charged ion at m/z 1304.6 Da, the [M+4H]4+ ion at m/z 978.9 Da and the [M+5H]5+ ion at m/z 783.1 Da are all observed. Figure 2.4(C) shows the MS/MS spectrum of the most

abundant ion, the [M+3H]+3 ion at m/z 1304.8 Da. This MS/MS spectrum was sent to the MASCOT protein database which correctly identified the peptide as the 33mer, an α-gliadin peptide from wheat with the correct sequence.

Further evidence that this enzymatic digestion approach can successfully identify potentially immunostimulatory gluten peptides is shown in Figure 2.5. Figure 2.5(A) shows the LC-MS total ion current chromatogram of various digestion

products from wheat gluten that had undergone PTCECA digestion. One particular peak at a retention time of approximately 18 min. was shown to degrade upon treatment with PEP. Figure 2.5(B) shows full scan MS data for this peak at 18.3 minutes. Further evaluation of data in this full scan spectrum shows that several singly charged ions and one doubly charged ion can be seen with some intensity. The ion at m/z 980.5, corresponding to a [M+2H]+2 ion, was selected by auto MS/MS which produced a product ion spectrum, shown in Figure 2.5(C). This MS/MS data was submitted to the Mascot protein database for an identity search. The top hit from this search is shown in Figure 2.5(D) and the report indicates that the peptide is a wheat α−_{gliadin with the proposed sequence LQPQNPSQQQPQEQVPL, which}

does contain short known immunogenic epitopes.