**5. Gluten peptides' fate in the pasta production chain**

**Simplified digestion**

**Adaptive immune response**

QLQPFPQPQLPY

QLQPFPQPQLPYPQPQPF

LQLQPFPQPQLPY

LQLQPFPQPQLPYPQPQPF

QLQPFPQPQLPYPQPQLPYPQPQPF

QLQPFPQPQLPYPQPHLPYPQPQPF

LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF

LPFPQQPQQPFPQPQ

**Innate immune response**

VRVPVPQLQPQNPSQQQPQEQVPLVQQQQF

QNPSQQQPQEQVPLVQQQ

VPVPQLQPQNPSQQQPQEQVPL

VRVPVPQLEPQNPSQQQPQEQVPL

VRVPVPQLQPQNPSQQQPQEQVPL

VRFPVPQLQPQNPSQQQPQEQVPL

PSSQVQWPQQQPVPQ

NMQVDPSGQVQWPQQQPF

SHIPGLEKPSQQQPLPL

Adapted with permission from Ref. [32]

**Table 2.**

α γ γ LMW

25.6

28.1

PQQPPFSQQQQPV

QQPPFSQQQPPPFS

QQQPLPL α, α‐gliadin; γ, γ‐gliadin; LMW, low‐molecular‐weight glutenin; Rt, retention time expressed in minutes. Epitopes and toxic sequences are underlined.

Pathogenic peptides identified in the digested samples obtained with the two different models, with the protein of origin.

23.0

QQQPPFS

29.8

QQQPL

α

28.8

SQQQQPV

α

26.6

α α α

28.0

26.8

RPQQPYPQPQPQ

LQPQNPSQQQPQEQVPL

LGQQQPFPPQQPYPQPQPFPS

28.2

LQPQNPSQQQPQ

**Prot**

**Rt**

**Innate immune response**

 α γ

29.6

α

33.3 26.5

α

34.1

**Prot**

α α α α

34.1

32.6

QAFPQQPQQPFPQ

TQQPQQPFPQQPQQPFPQ

PQTQQPQQPFPQFQQPQQPFPQPQQP

FPQQPQLPFPQQPQQPFPQPQQPQ

PFPQPQQPQQPFPQSQQPQQPFPQP

QPQLPFPQQPQQPFPQPQQPQQPSPQSQQPQQPFPQ

QQPQQPFPQPQQTFPQQPQLPFPQQPQQPFP

γ **Prot**

α α α α γ LMW LMW LMW LMW LMW

25.4

25.5

22.0

19.8

16.5

14.5

27.3

23.9

18.0

16.6

**Rt**

30.7

 γ

γ γ γ

29.3

29.8

29.3

26.8

32.7

SQQPQQPFPQPQ

30.5

TQQPQQPFPQ

**Rt**

**Physiological digestion**

**Adaptive immune response**

**Prot**

γ γ γ γ

24.9

24.4

320 Wheat Improvement, Management and Utilization

21.3

20.5

**Rt**

The quantification of gluten peptides reported in **Table 2** was carried out for six steps of pasta production (involving only durum wheat), to verify if some technological treatment have an influence on protein extractability/digestibility. Three different varieties were analyzed to exclude variations exclusively due to the genotype. Results are shown in **Figure 5** for whole wheat, flour, dough, extruded pasta, dried pasta and cooked pasta.

As observed in **Figure 5**, after the physiological digestion method, the amount of toxic pep‐ tides is approximately a half than with the simplified method, whereas immunogenic pep‐ tides are approximately twice, due to the different type of peptides generated. This should be taken into account when biological tests are done to assess immunological responses. It can be noted that the amount of toxic and immunogenic peptides remains largely constant along the pasta production chain, so none of the processing steps of pasta is at the moment able to decrease wheat immunogenic potential for celiac patients. In other words, if a vari‐ etal screening has to be performed, there is no need to use the end product; it is sufficient to test the basic wheat variety. The difference between the two digestion methods becomes more evident after pasta cooking; in fact, heat causes polymerization of gliadins through intermolecular disulphide bridge formation and to a lesser extent for dehydroalanine for‐ mation [34]. Thus, the heat treatment leads to the loss of gliadin extractability, which is the reason of the high underestimation of peptides generated after digestion of the ethanolic extract. It is interesting to note that independently from the method adopted, the differences between varieties maintain the same trend at all the steps of processing. This is an ulterior confirmation that traditional pasta processing leaves gluten immunogenic and toxic pep‐ tides unaffected.

**Figure 5.** Total amount of toxic and immunogenic peptides (upper and lower plots, respectively) quantified after the simplified and the physiological digestion model. Adapted with permission from Ref. [32].

#### **6. Conclusion**

Several digestion methods applied to gluten proteins are reported in literature. Generally, these models are very simple involving only the use of the main gastric and pancreatic pro‐ teases (pepsin, trypsin and chymotrypsin). A buffering agent is also used, to keep the correct pH value at each phase. However, a physiological digestion procedure was previously used in literature to assess the release of mycotoxins and heavy metals from food matrices. This method involves the use of digestive juices whose chemical composition strictly reflects the physiological one. These two methods were compared to assess gluten peptides generated. In both cases (simple and complex model), the peptides generated from the digestion were characterized using liquid chromatography coupled with mass spectrometry. In these *in vitro* experiments, the processes occurring in the human gastrointestinal tract during food diges‐ tion were simulated, and the outcome of the digestion was assessed by LC‐MS techniques. With the use of tandem mass spectrometry, the exact amino acid sequence of the peptides generated by the digestion was determined. Among all the peptides, the ones containing sequences known to be implied in celiac disease were identified. Strong differences were present between the two digestion models. First, with the simplified model almost all the peptides derive from α‐gliadin, whereas with the physiological method, they are equally dis‐ tributed among α‐gliadins and γ‐gliadins and LMW glutenins. This can be explained by the observation that α‐gliadin‐derived peptides of the simplified method are further proteolyzed into shorter peptides in the physiological model and often these shorter peptides did not contain immunotoxic sequences anymore. Moreover, in the physiological model are present enzymes other than proteases (like amylase and lipase) that, even if not directly implied in protein cleavage, can contribute (together with bile salts) to matrix degradation, thus improv‐ ing the extractability and digestibility of higher molecular weight proteins such as γ‐gliadins and glutenins.

be taken into account when biological tests are done to assess immunological responses. It can be noted that the amount of toxic and immunogenic peptides remains largely constant along the pasta production chain, so none of the processing steps of pasta is at the moment able to decrease wheat immunogenic potential for celiac patients. In other words, if a vari‐ etal screening has to be performed, there is no need to use the end product; it is sufficient to test the basic wheat variety. The difference between the two digestion methods becomes more evident after pasta cooking; in fact, heat causes polymerization of gliadins through intermolecular disulphide bridge formation and to a lesser extent for dehydroalanine for‐ mation [34]. Thus, the heat treatment leads to the loss of gliadin extractability, which is the reason of the high underestimation of peptides generated after digestion of the ethanolic extract. It is interesting to note that independently from the method adopted, the differences between varieties maintain the same trend at all the steps of processing. This is an ulterior confirmation that traditional pasta processing leaves gluten immunogenic and toxic pep‐

Several digestion methods applied to gluten proteins are reported in literature. Generally, these models are very simple involving only the use of the main gastric and pancreatic pro‐ teases (pepsin, trypsin and chymotrypsin). A buffering agent is also used, to keep the correct pH value at each phase. However, a physiological digestion procedure was previously used

**Figure 5.** Total amount of toxic and immunogenic peptides (upper and lower plots, respectively) quantified after the

simplified and the physiological digestion model. Adapted with permission from Ref. [32].

tides unaffected.

322 Wheat Improvement, Management and Utilization

**6. Conclusion**

Thus, in the case, a subsequent immunological experiments or biological trials have to be per‐ formed, the more physiological method is more suitable than the simplified one, because the peptides generated are really different and the complex method is more similar to what really happens in the human gastrointestinal tract.

The peptides containing immunotoxic sequences were quantified for both the *in vitro* digestion models along the pasta production chain, to evaluate also the suitability of the two methods for processed foods. The samples (kernels, semolina, dough, extruded pasta, dried pasta and cooked pasta) were obtained from three different durum wheat varieties (Svevo, Meridiano and Saragolla). The physiological digestion method produced lesser amount of toxic and a higher amount of immunogenic peptides compared to the simplified one, probably due to the different molecular weight of the peptides generated. A noticeable result is that the difference among the varieties tested remains unchanged, with Saragolla showing a lower content of peptides involved in celiac disease compared to Svevo and Meridiano. Another remarkable result is that the simplified method cannot be applied to thermally treated foods, because heating induces gluten polymerization leading to poor proteins extractability. The two dif‐ ferent models are very well correlated in terms of total amount of immunotoxic peptides generated. So, to perform a varietal screening, the simplified method is suitable when a large amount of samples has to be analyzed.

*In vitro* digestion of the prolamin extract was then applied to 45 durum wheat samples belonging to five different varieties and harvested in three different Italian regions (Argelato in the North of Italy, Falconara in the Centre and Lucera in the South). These findings showed no major differences due to the different cultivation place, consistently with the reserve role of that class of proteins (that thus is not affected by environmental factors).

For what concerns genotype influence, since the cultivar selection operated by breeders in the last years to achieve the desired rheological properties has led to a decrease in the genetic biodiversity of durum wheat varieties present nowadays on the market, 25 durum wheat accessions were selected from a durum wheat panel in order to maximize the genetic biodi‐ versity of the samples (and thus eventual differences in immunotoxic peptides production upon digestion). Results obtained from every single accession were mediated in five groups on the basis of phylogenetic affinity on dendrogram.

For toxic peptides, no significant differences were present while strong variability emerged for immunogenic peptides, with accessions of the second groups (International Center for Agricultural Research in the Dry Areas (ICARDA) accessions for temperate areas) showing a significantly lower content of peptides eliciting adaptive immune response.

The higher variability of immunogenic peptides compared to toxic peptides can be explained on the basis of gliadins sequence variability; in fact, toxic peptides usually derive from the N‐term region of the protein, which is the most conserved. On the contrary, immunogenic peptides derive from a region of the protein showing a much higher variability. So, differ‐ ent wheat genotypes can express different gliadins isoforms thus showing a different final content of immunogenic sequences.

Then, it is possible to select wheat varieties with good gluten content (and good rheological properties) but with a reduced amount of immunogenic sequences in order to reduce the exposure of people to a possible trigger for celiac disease.
