Module 2: Gluten, Gliadin, Glutenin

When flour is hydrated and kneaded, a single contiguous viscoelastic network emerges from individual storage protein molecules. Gluten is that network: a rheologically active, swollen, insoluble mass of gliadin and glutenin bound by disulfide bridges, hydrogen bonds, and hydrophobic interactions. It can be isolated by gentle water washing of a wetted dough ball, which strips away starch and water-soluble components while leaving the protein web intact.

Mechanically, gluten behaves as a critical gel: it is viscous enough to flow under slow stress (bubble expansion during proofing), yet elastic enough to spring back under fast stress (kneading, moulding, oven rise). This dual character is why wheat is the only cereal from which leavened bread can be reliably made. Rye lacks HMW-GS; barley has poor extensibility; oats have no gluten at all; maize lacks the sulfur-rich prolamins.

Gliadins are monomeric proteins of 30–40 kDa soluble in 60–70% ethanol and partitioned by electrophoretic mobility into four classes: α/β-gliadins, γ-gliadins, and ω-gliadins. They share a proline- and glutamine-rich primary sequence (P+Q > 50%) organised around repetitive motifs such as QPQPFPQ, PQQPY, and PQQPF. These repeats fold into characteristic poly-L-proline II (PPII) helical structures that resist proteolysis and are central to the immunological story below.

Sulfur-rich vs Sulfur-poor

• α/β-gliadins (~25–35 kDa): sulfur-rich, six cysteines forming three intramolecular disulfides; stabilised globular C-terminus + proline-rich N-terminus.
• γ-gliadins (~30–35 kDa): sulfur-rich, eight cysteines forming four intramolecular disulfides.
• ω-gliadins (~45–75 kDa): sulfur-poor, essentially no cysteines; almost pure repetitive prolamin domain; major contributors to loaf softness.

Because gliadins bear only intramolecular disulfides, they cannot covalently cross-link into polymers. They instead interact non-covalently with the glutenin matrix, serving as molecular plasticisers: their presence lowers dough elasticity, increases extensibility, and controls viscous flow during sheeting and moulding. Dough with too much gliadin is slack and unable to hold gas; with too little gliadin it is tight and prone to tearing.

Glutenins are disulfide-linked polymers, the largest protein molecules known in nature, reaching molecular weights of 1–10 MDa. They are built from high-molecular-weight (HMW-GS, 80–150 kDa) and low-molecular-weight (LMW-GS, 30–45 kDa) subunits. HMW-GS form the network backbone via head-to-tail intermolecular cystine bonds; LMW-GS branch off as side-chains connecting adjacent backbone strands.

HMW-GS Domain Architecture

Each HMW-GS polypeptide has three domains:

• N-terminal domain (~100 aa): conserved non-repetitive region carrying 3–5 cysteines, anchoring the subunit in the polymer.
• Central repetitive domain (~500 aa): hexapeptide (PGQGQQ) and nonapeptide (GYYPTSPQQ) repeats forming an extensible β-spiral (Tatham 2000).
• C-terminal domain (~40 aa): conserved non-repetitive region with 1 cysteine, another polymer anchor.

Solid-state NMR and molecular dynamics indicate the central repeats adopt a loose poly-β-turn spiral about 15 nm long in the fully extended state, which is the structural origin of the gluten’s reversible extensibility. Tatham & Shewry termed this the “loop and train” model: the central repeats are elastic loops between cystine-fixed “trains” of polymerised backbones.

Dx5+Dy10: The Bread-Making Allele

Wheat carries HMW-GS loci Glu-A1, Glu-B1, Glu-D1 encoding pairs of subunits (x and y types). The D-genome allele encoding Dx5 + Dy10 (the ubiquitous Chinese-Spring allele) confers superior bread-making quality relative to Dx2 + Dy12: Dx5 has an extra cysteine in the central repetitive domain (single-residue polymorphism) that seeds an additional backbone cross-link, strengthening the polymer. This single allele explains a large fraction of the variance in loaf volume between modern cultivars (Payne 1987).

Under small-amplitude oscillatory shear (SAOS) at strain <1%, dough obeys linear viscoelasticity: the shear stress response to a sinusoidal strain γ(t) = γ₀ sin(ωt) decomposes into an in-phase (elastic) component and a quadrature (viscous) component. The storage modulus G′ and loss modulus G″ are the moduli of these two components:

Wheat dough typically shows G′ > G″ over several decades of frequency (an elastic plateau), with tan(δ) ~ 0.25–0.35 at the technologically relevant frequency ω = 1 rad/s (Ng 2011). Frequency sweeps from 0.01 to 100 rad/s reveal the relaxation time spectrum H(τ), which for a well-developed dough shows two principal modes: a fast gliadin fluid mode at τ ~ 1 s and a slow glutenin network mode at τ ~ 100 s.

Generalised Maxwell Network

The generalised Maxwell model writes the relaxation modulus as a sum over modes of individual relaxation times:

Gliadin-dominant doughs show a higher ω-integrated G″ (more viscous dissipation); glutenin-dominant doughs show a higher low-ω plateau in G′ (more elastic energy storage). Dough mixing time (Farinograph), stability (Extensograph), and Alveograph W-value are industrial surrogates for these fundamental rheological quantities.

Flour and water alone, merely mixed, produce a crumbly mass. Dough development — the conversion of that mass into a cohesive, strong, extensible web — requires mechanical work input. A canonical bread dough receives ~15 kJ per kg of flour during mixing, corresponding to ~15–20 minutes in a spiral mixer or 3–6 minutes in a high-speed Chorleywood mixer.

Molecular Events During Mixing

• Hydration: water penetrates the amorphous protein matrix, plasticises the gluten proteins, reduces their effective T₋.
• Unfolding: shear work partially denatures compactly folded glutenin globules, exposing cysteine thiols.
• S-S interchange: free thiols exchange with intramolecular disulfides via thiol-disulfide exchange, rearranging the polymer topology.
• Network formation: the intermolecularly bonded network extends isotropically through the dough; G′ rises until it peaks (optimum development).
• Over-mixing breakdown: continued shear fragments the network, G′ falls, extensibility collapses.

Proof & Oven Rise

During proofing, yeast (Saccharomyces cerevisiae) ferments maltose from damaged starch and produces CO&sub2; at ~50 mL per gram of flour per hour at 30 °C. CO&sub2; bubbles nucleate on pre-existing air bubbles entrained during mixing and expand within the gluten network. Bubble growth must be fast enough to inflate the loaf, slow enough to avoid bursting the gluten walls. Oven rise (“oven spring”) is the final inflation as bubble CO&sub2; expands thermally and water vaporises, followed by network setting as starch gelatinises (Module 1) and proteins coagulate above ~75 °C.

Alveograph W-Value and Bread-Making Quality

Chopin’s Alveograph measures the work required to inflate a thin disc of dough to rupture; the area under the pressure–volume curve is the W-value. Good bread doughs have W = 200–350 × 10⁻⁴ J (“medium-strong”); pasta durum doughs target W = 150–250; cake flours W < 120. The tenacity-to-extensibility ratio P/L (from the same alveograph inflation) complements W.

Celiac disease is a T-cell-mediated enteropathy affecting roughly 1% of the worldwide population in which ingested gluten provokes chronic small-intestinal villous atrophy. The disease is genetically restricted: >95% of patients carry HLA-DQ2.5; the remainder essentially all carry HLA-DQ8. The HLA association is a necessary but not sufficient condition: roughly 30% of the European population carries DQ2/DQ8 but only 1% develops celiac disease (Sollid 2017).

The α-Gliadin 33-mer

Shan et al. (2002) identified a 33-residue fragment of α-gliadin, LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF, as the most immunodominant and proteolysis-resistant peptide in the human gluten response. The peptide’s proline density (~40%) protects it from gastrointestinal endoproteinases. It survives passage through the stomach, duodenum, and jejunum largely intact, accumulates in the lamina propria via transcytosis (possibly through compromised tight junctions in zonulin-dysregulated gut), and is presented to local tissue transglutaminase 2 (tTG2).

tTG2 Deamidation

tTG2 is a Ca²⁺-dependent transamidating enzyme that normally catalyses protein–protein cross-links (glutamine + lysine → isopeptide bond). In the presence of unrestricted Q-rich gliadin substrate and limited primary amine nucleophile, it short-circuits into deamidation, converting specific Q residues to E (glutamate). The introduction of a negative side-chain dramatically enhances binding of the peptide to the basic binding pocket of HLA-DQ2.5 (Kim 2004), typically by 20–50 fold.

T-cell Activation & Villous Atrophy

Deamidated 33-mer bound in HLA-DQ2.5 on intestinal antigen-presenting cells activates CD4+ gluten-reactive T cells in the lamina propria. The activated T cells release IFN-γ, IL-15, IL-21, and other cytokines that (i) drive plasma cells to produce IgA anti-tTG2 and anti-gliadin antibodies (the serological markers used in clinical diagnosis); (ii) activate intraepithelial cytotoxic lymphocytes via NKG2D-MICA interactions; and (iii) ultimately destroy the enterocyte lining, collapsing the villus–crypt architecture. The result is malabsorption of iron, folate, Ca, and fat-soluble vitamins.

Clinical Management

The only validated treatment is lifelong strict avoidance of wheat, barley, rye, and cross-contaminating oats (gluten-free diet, GFD). On a strict GFD, small-bowel biopsies typically heal within 6–12 months. Novel therapies under trial include orally administered proline-specific endopeptidases (ALV003/latiglutenase), blockers of HLA-DQ2.5 (small-molecule inhibitors), peptide tolerance vaccines (Nexvax2), and tTG2 inhibitors.

Beyond classical celiac disease (~1% prevalence) and true IgE wheat allergy (~0.2% prevalence), a much larger group of patients (5–10% of Western populations) report GI and extraintestinal symptoms on wheat consumption without anti-tTG2 antibodies or enteropathy on biopsy. This heterogeneous clinical entity is grouped under the label “non-celiac wheat sensitivity” (NCWS). Two leading non-mutually-exclusive mechanisms are invoked:

• FODMAPs: fructans in wheat are rapidly fermented by colonic microbiota, producing gas and osmotic water retention that drive IBS-like symptoms (Biesiekierski 2013). Double-blind, placebo-controlled cross-over trials repeatedly show that when gluten is reintroduced on a low-FODMAP background many self-reported “gluten-sensitive” patients fail to respond.
• ATIs: amylase–trypsin inhibitors (2–4% of wheat grain protein) activate TLR4 on intestinal monocytes (Junker 2012, Zevallos 2017), triggering innate pro-inflammatory cytokines independently of T-cell recognition. ATIs have been selectively enriched in modern bread-wheat cultivars bred for insect resistance.

Beyond these two mechanisms, increased intestinal permeability (“leaky gut”), altered microbiota, and psychogenic/nocebo factors are actively debated. Distinguishing celiac disease from NCWS remains a clinical challenge; serological testing (IgA anti-tTG2) plus endoscopic duodenal biopsy under a gluten-containing diet is the diagnostic standard.

Popular claims that modern bread wheat has “dangerously high” gluten relative to ancient wheats are only partially supported by the literature. Shewry (2018) and Kasarda (2013) compared the total protein and gluten content of einkorn, emmer, spelt, and modern bread wheat cultivars grown in common gardens and found no systematic increase in gluten content from ancient to modern.

Differences in Immunogenic Epitopes

What has changed is the distribution of specific T-cell epitopes. The canonical α-gliadin 33-mer occurs in some D-genome alleles but not others; einkorn (AA) lacks the DQ2.5 α-II immunodominant epitope that is present in all hexaploid modern wheats (Molberg 2005). This makes einkorn less immunogenic, though still not safe for celiac patients because it carries other γ-gliadin and ω-gliadin epitopes. True gluten-free alternatives remain rice, maize, sorghum, teff, and certified oats.

Rupture Test for Bread Volume Prediction

The rupture test (Dobraszczyk & Morgenstern 2003) measures the bubble-wall rupture strain of inflated dough discs. A key finding: final loaf volume correlates linearly with rupture strain, rather than with peak stress or alveograph W alone. This reflects the rheological requirement that the bubble wall must extend stably during proofing and oven rise without catastrophic failure, up to the moment when starch gelatinisation sets the crumb.

In developing wheat endosperm cells, gliadins and glutenins are synthesised on rough endoplasmic reticulum, translocated into the ER lumen as nascent polypeptides, and fold in the oxidising luminal environment with the help of protein disulfide isomerase (PDI), BiP (Hsp70-family chaperone), and calnexin. After folding they bifurcate along two trafficking routes (Tosi 2009):

• ER-derived protein bodies (Type I): gliadins and glutenins accumulate directly in the ER lumen, which swells into discrete Type I protein bodies. These are the predominant storage compartments.
• Vacuolar protein bodies (Type II): a fraction of storage protein is exported to the Golgi and then delivered to the protein-storage vacuole (PSV). Type II bodies contain mainly globulins and are more prominent in the aleurone.

During grain maturation the Type I protein bodies coalesce as water is withdrawn and the cell contents densify, yielding the amorphous protein matrix that one sees in mature endosperm by electron microscopy. The distinction between Type I and Type II bodies matters for genetic engineering: targeting a transgenic peptide to the ER vs the PSV determines its final storage site and folding environment.

ER Stress and Protein Quality

High rates of storage-protein synthesis stress the ER folding capacity. Wheat endosperm mounts a tissue-specific unfolded-protein response (UPR) mediated by bZIP transcription factors that upregulate BiP, PDI, and other ER chaperones. Disruption of the UPR (antisense bZIP28/60) impairs gluten polymer assembly and grain filling, illustrating the intimate coupling between storage-protein quality and the cellular machinery that builds it.

Bread-making quality is partly heritable and partly environmental (“G×E”). Heritable components are dominated by the HMW-GS alleles at the three Glu-1 loci, with the Payne quality score (Payne 1987) summing allele-specific contributions into an index predictive of loaf volume. Modern breeding programmes routinely genotype the Glu-1 and Glu-3 loci by PCR or KASP markers and select for favourable allele combinations.

• Glu-A1: allele “1” (or 2*) > null.
• Glu-B1: alleles 17+18 and 7+8 preferred over 7+9 or 20.
• Glu-D1: allele 5+10 (Dx5+Dy10) > 2+12 > 4+12.

Complementing HMW-GS selection, LMW-GS alleles at Glu-B3 and Glu-D3 fine-tune dough strength and extensibility; ω-gliadin alleles at Gli-B1 and Gli-D1influence extensibility. Indirect environmental levers (nitrogen nutrition, sulfur supply, post-anthesis temperature) strongly modulate final protein content and S:N ratio, which determines the gliadin:glutenin balance in the mature grain.

Industrial Quality Tests

• Farinograph (Brabender): records torque on mixer shaft during dough development; outputs water absorption, development time, stability.
• Extensograph: stretches a rested dough cylinder to rupture; records resistance and extensibility.
• Alveograph (Chopin): inflates a thin dough disc; outputs W (work), P/L (tenacity/extensibility).
• Mixograph: smaller-scale mixer curve for early-generation breeding screening.
• SDS sedimentation: rapid estimate of glutenin polymer size from suspension of flour in SDS+lactic acid.
• Baking trial: the gold standard; measures loaf volume, crumb grain, and sensory properties from a standard recipe.

Gluten is a soft-matter marvel: a self-assembled, hydration-triggered, work-developed viscoelastic polymer network whose properties arise from the interplay of covalent cystine cross-links (glutenin), non-covalent plasticisation (gliadin), and hydrogen-bonded β-turn secondary structure. Decades of cereal chemistry, polymer physics, and immunology have converged on a molecular description that now supports rational breeding for bread-making quality, celiac-safe “mini” wheats (Jouanin 2018), and industrial enzyme therapies.

But the gluten proteins are themselves the endpoint of a much larger biophysical chain that begins with light capture and CO&sub2; fixation. In Module 3 we turn to the source of the atoms: wheat as a C₃ photosynthesiser, its Rubisco biochemistry, and its canopy-level light-use efficiency — the solar-powered input that ultimately fills the endosperm with the starch and protein we have just dissected.

The α-gliadin 33-mer epitope cluster is encoded on the short arm of chromosome 6, with three homoeologues (α-Gli-A2, α-Gli-B2, α-Gli-D2). The main immunodominant DQ2.5-α-II epitope resides in the D-genome copy; A-genome α-gliadins lack the full-length immunogenic peptide. Because hexaploid buffering tolerates loss-of-function in individual homoeologues, it is in principle possible to knock out the D-genome copy while preserving bread-making quality from the A and B contributions.

CRISPR/Cas9 Multiplex Editing

Sánchez-León et al. (2018) used a single CRISPR/Cas9 construct with 12 guide RNAs to simultaneously edit 35 out of the ~45 α-gliadin genes in hexaploid wheat. The resulting lines showed ~85% reduction in gliadin content and a ~90% drop in immunogenic epitope load measured by R5 ELISA. Bread-making quality was partly preserved (loaf volumes ~70% of wild type), illustrating the trade-off between immunogenicity and functionality. Regulatory classification of CRISPR-edited hypoimmunogenic wheat remains a contentious issue in the EU and elsewhere (Jouanin 2018).

Oral Enzyme Therapy

Complementary to engineered wheat, oral enzyme therapy aims to digest the 33-mer in the small intestine before it reaches the lamina propria. Latiglutenase (ALV003) is a 1:1 mixture of a cysteine endoprotease from barley (EP-B2) and a bacterial proline-specific endopeptidase (PEP from Sphingomonas capsulata). The combination cleaves the proline-rich 33-mer at Q-residues (EP-B2) and P-residues (PEP), producing fragments too short to activate T cells. Phase II trials in celiac patients on a gluten-free diet showed modest protection from inadvertent gluten exposure; late-phase trials have not met primary endpoints.

Tolerance Induction

An alternative strategy seeks to induce antigen-specific tolerance by repeated administration of the immunodominant peptide cocktail (Nexvax2) to HLA-DQ2.5-positive patients. The rationale echoes allergen-specific immunotherapy. Phase II trials showed transient effects but no durable tolerance; later-phase development was halted in 2019. TG2 inhibitors (ZED1227) that block deamidation represent another tractable pharmacological target, with phase II data showing histologic improvement vs placebo in challenged celiac patients.