Characterization of gliadin, secalin and hordein fractions using analytical techniques

Prolamins, alcohol soluble storage proteins of the Triticeae tribe of Gramineae family, are known as gliadin, secalin and hordein in wheat, rye and barley respectively. Prolamins were extracted from fifteen cultivars using DuPont protocol to study their physiochemical, morphological and structural characteristics. SDS-PAGE of prolamins showed well resolved low molecular weight proteins with significant amount of albumin and globulin as cross-contaminant. The β-sheet (32.72–37.41%) and β-turn (30.36–37.91%) were found higher in gliadins, while α-helix (20.32–28.95%) and random coil (9.05–10.28%) in hordeins. The high colloidal stability as depicted by zeta-potential was observed in gliadins (23.5–27.0 mV) followed secalins (11.2–16.6 mV) and hordeins (4.1–7.8 mV). Surface morphology by SEM illustrated the globular particle arrangement in gliadins, sheet like arrangement in secalins and stacked flaky particle arrangement in hordeins fraction. TEM studies showed that secalin and hordein fractions were globular in shape while gliadins in addition to globular structure also possessed rod-shaped particle arrangement. XRD pattern of prolamin fractions showed the ordered crystalline domain at 2θ values of 44.1°, 37.8° and 10.4°. The extracted prolamins fractions showed amorphous as well as crystalline structures as revealed by XRD and TEM analysis. Space saving hexagonal molecular symmetry was also observed in TEM molecular arrangement of prolamins which has profound application in development of plant-based polymers and fibres.


SDS-PAGE of prolamin fractions. Sodium Dodecyl Sulphate-Polyacrylamide Gel Electrophoresis (SDS-
PAGE) was used to analyse the extracted prolamin fractions by adopting the method of Siddiqi et al. 39 . 10 mg of prolamin fraction was thoroughly mixed with 1 ml of 2 × Laemmli sample buffer solution (pH 6.8 containing 62.5 mM Tris-HCl, 25% glycerol, 5% ß-mercaptoethanol, 2% SDS, 0.01% bromophenol blue) in 1.5 ml Eppendorf tubes. The tubes were vortexed to disperse the proteins and then shaken using an orbital shaker operated for 1 h at 151 rpm and 45 °C. Samples were heated at 100 °C for 5 min and then centrifuged (RC 4815S, Eltek, Mumbai, India) at 11000×g for 20 min. SDS-PAGE of prolamin fraction was performed by loading 10 μl of prepared sample supernatant in each well and resolved in 12% resolving gel at a constant current of 25 mA (Mini-Protean Tetra Cell, Bio-Rad Laboratories, Hercules, USA).
When the tracking dye reached the bottom of the gel, plates were disassembled to remove the gel which was stained overnight in the staining solution (0.1% Coomassie Brilliant Blue-R250 in 40% methanol and 10% acetic acid). The gels were destained using 30% methanol and 10% acetic acid in deionized water.
Classification of total prolamin was done according to Schalk et al. 13 . The relative proportion of prolamin was calculated using band percentage corresponding to the total extractable proteins. SDS-PAGE gels were performed in duplicates. Bio Rad Image Lab software 6.1 was used as working station and the maximum background subtraction was performed to all lane.
Amino acid analysis of prolamin fractions. Analysis of prolamin fraction was performed by following the procedure of Siddiqi et al. 39 with slight modification. Briefly, 5 mg of prolamin were hydrolysed in clean dried screw capped glass test tubes which were priorly dipped whole night in 2 N HCl to avoid any sort of contamination. Hydrolysis was carried out using 6 N HCl containing 0.1% ß-mercaptoethanol in autoclave at 110 °C ± 2 °C for 16 min screw tight capped closure test tubes to minimize the loss of digestion solvent during hydrolysis process. The filtrate was evaporated in a small round bottomed rotary flask under vacuum at 40 °C to dryness in a rotary evaporator (Buchi, Fawil, Switzerland). A suitable volume of 0.1 N HCl was added to each dried film of the hydrolyzed sample to dissolve all the soluble materials and then filtered through 0.22 µm filter paper (Millipore, Merck Life Science Private Limited, Mumbai, India). Amino acid analysis was performed using a Nexera Amino acid Analyzer (Shimazdu, Kyoto, Japan) equipped with a pre-column derivatization using three derivatizing reagents such as mercaptopropionic acid, o-phthaladehyde and 9-fluorenylmethyl chloroformate. A C-18 column (Waters-Spherisorb ODS2 Column; 5 µm, 80 Å, 4.6 mm × 250 mm) having pH stability 2-8 was used for chromatographic separation. Analysis was performed using 20 mmol/L phosphate (potassium) buffer (pH 6.5) as solvent A and 45/40/15 acetonitrile/methanol/water as solvent B. The separation was obtained at a flow rate of 1 ml/min using a gradient elution that allowed 2% B at 0.01st min, followed by linear raise of eluent B to 50% at 41st min and then decreased to 2% B at 43rd min. The temperature of the column oven was set at 40 °C and injection volume to 1 μL. Resolution of amino acid derivatives was performed with the help of fluorescence detector having excitation and emission set at 330 nm and 450 nm respectively. Labsolutions LC/ GC (Shimadzu, Kyoto, Japan) was used as a working station. The amino acid standard mixture was prepared by mixing eighteen amino acids (SRL, Mumbai, India) in 0.1 N HCl which included Aspartic acid (Asp), Glutamic acid (Glu), Serine (Ser), Glycine (Gly), threonine (Thr), Histidine (His), Alanine (Ala), Arginine (Arg), Tyrosine (Tyr), Valine (Val), Methionine (Met), Cystine (Cys), Phenylalanine (Phe), Tryptophan (Trp), Isoleucine (Ileu), Leucine (Leu), Lysine (Lys) and Proline (Pro). Each amino acid was identified by running it separately to determine its retention time. Then the mixture of 18 amino acids was named accordingly from the retention time of the individual amino acids as provided in Figure S2. The concentration of amino acids was estimated using single point calibration with no correction factor. The estimation of acid digested sample was carried out by comparing the area under the peak of standard mixture of each amino acid with that of the samples (Fig. 2 & Figure S3). Glutamine (Gln) and asparagine (Asn) were deaminated to glutamic acid (Glu) and aspartic acid (Asp) during acid hydrolysis 40 . Therefore, glutamic acid and aspartic acid were represented by combination of acid and its amide derivative as Glu + Gln and Asp and Asn.
Fourier-transform infrared spectroscopy (FTIR). FTIR spectral of prolamin fraction was recorded using FTIR/FIR Spectrometer (Perkin Elmer Inc, Waltham USA). The KBr pellet was made using finely grounded KBr salt (250 mg) with dried protein (concentration ≈1-5% Weight). The finely grounded mixture was then uniformly distributed in a suited cell and pressed with the force around 10 tons to yield fine transparent pellets (≈1-2 mm thickness) using hydraulic press. A pellet of pure KBr salt was used as a reference. Each sample was measured at least in triplicate, with total of 256 scans performed in the 400-4000 cm −1 region at 2 cm −1 resolutions. The spectral data acquisition and reference subtraction was performed using spectrum software, while Hue angle H o = tan −1 b * /a * Chroma C * = a * 2 + b * 2 0.5 Peak fitting. The non-linear fitting of spectral 1600-1700 amide I region data was performed by adopting the procedure of Sadat & Joye 41 using Peak analyzer-Levenberg-Marquardt algorithm to know the secondary structural components of the protein. Spectral 1600-1700 region plot was first subjected to baseline correction using second derivative (Zeros) method. Through peak analyzer, the hidden peaks were spotted on baseline corrected plot using second derivative method by setting the Savitsky-Golay function as smooth derivative and polynomial order as 2. After detecting the peak position based on negative peaks on secondary derivative, the multiple peak fitting was performed using the Voigt function (combination of Gaussian function and Lorentzian function). During initializing the parameters, fixed the y axis and perform simplex fit till the model (fitted cumulative curve) comes near to the observed (original curve) data. Then after iterations were performed till the model curve superimpose the observed curve and the best fit was achieved. Goodness-of-fit test was monitored to achieve best peak fitting i.e. chi square value (< 1E−6), R2 value (> 0.99), adjusted R2 value (> 0.99) and fit status. The relative area percentage of each structural components (α-helix, β-sheet, β-turn and random coil) was calculated by dividing the sum of area of assigned specific band position to that of total area. Dynamic light scattering (DLS) measurements. The hydrodynamic diameter (D h ) and zeta potential (ζ-potential) values of proteins were monitored on Zetasizer Nano ZS (Malvern Instruments Ltd., Worcestershire, UK) equipped with a helium-neon (He-Ne) laser (632.8 nm, 4mW), at back scattering angle of 173° to the incident beam and operated at 25 °C. The instrument monitors, time-dependent fluctuation in the light scattered by molecules present in solution to determine the diffusion rate due to the Brownian motion at a fixed scattering angle. The extracted prolamin(s) (1 mg ml −1 ) were prepared in acetonitrile: water: formic acid (50:50:0.1) solution. The refractive index values of the solvent system acetonitrile: water: formic acid (50:50:0.1) solution has been taken from the literature 42 . To avoid the formation of protein aggregates, the samples were vortexed gently and then filtered through 0.22 µm Millipore filters prior to measurements. An average of three measurements for each sample was considered as an experimental data.
Scanning electron microscopy (SEM) and energy dispersive X-ray analyser (EDX). The morphology of the isolated freeze-dried prolamin powder was studied with Carl Zeiss Supra-55 SEM (Carl Zeiss, Oberkochen, Germany) at an accelerating voltage of 10-15 kV. The samples were mounted on stubs using double-sided adhesive tape. Prior to SEM imaging, all samples were sputter-coated with gold for 2 min by Quorum Sputtered Coater (Q150R ES, Laughton, East Sussex, UK). Further, the elemental composition was analysed with Energy dispersive X-ray spectrometer (Oxford Instruments Nano Analysis & Asylum Research, High Wycombe, UK). X-ray diffraction (XRD). XRD studies, in the 2θ (Bragg's angle) range of 10-60° were performed using X-ray diffractometer (Shimadzu 7000, Shimadzu Corporation, Kyoto, Japan) in Bragg-Brentano geometry with Cu K α radiation (λ = 1.5405 A o ). Some of the prolamins were found to be nano-crystalline in nature i.e. these are composed of crystallites or grains that are the ordered building blocks units repeated periodically and separated by disordered grain boundaries. Debye-Scherrer equation was used to calculate the particle crystallite size on corresponding diffraction pattern based on the dimension of full-width at half-maximum (FWHM) values.
where, D is the mean size of the ordered (crystalline) domains, λ is the X-ray wavelength in nanometre (nm), β is the peak width of the diffraction peak profile at half of the maximum height (FWHM) in radians, θ is the Bragg angle, which can be in degrees or radians, cos θ corresponds to the same number and K is a constant related to crystallite shape, normally it can be taken as 0.89 or 0.9 for Full Width Half Maximum (FWHM) of spherical crystals with cubic unit cells 43 .
Further, inter-planar spacing (d), was calculated from the peak positions using Bragg's equation where d is inter-planar spacing, n is order of diffraction.
Transmission electron microscopy (TEM). The structures of different Prolamin fractions were investigated using Transmission Electron Microscope (JEM-2100, JEOL TEM, Tokyo, Japan). The samples (1 mg/ ml) were prepared in 50% acetonitrile containing 0.1% formic acid solvent and a drop of the diluted protein suspension solutions (15 μl) was poured on a polycarbon film supported on a copper grid and dried at ambient temperature. The samples were examined on TEM at an accelerating voltage of 200 kV. The sizes of the protein nano-structures were obtained using iTEM software (Olympus, Münster, Germany). The radius of the diffraction rings in SAED pattern is the inverse of the d-spacing of the associated lattice planes and radii of the rings were determined using iTEM software.
Statistical analysis. The results were expressed as a mean ± SD and compared statistically at p ≤ 0.05, using

Result and discussion
Wheat, rye and barley belonging to the Triticeae tribe, rich in gluten were extracted using DuPont et al. 11 protocol to obtain prolamin fractions known as gliadins in wheat, secalins in rye and hordeins in barley. We showed the properties of extracted freeze-dried prolamins fraction using different analytical techniques. The extracted prolamin fractions contained prolamins as a major component but with cross contamination of Alb + Glo which needs to be considered while interpreting the results. The color of freeze-dried powder of gliadins, hordeins and secalins fraction was light cream (Fig. 1).
Physiochemical characteristics. The protein content (PC) of the freeze-dried prolamin fractions extracted from wheat, rye and barley followed a decreasing order as gliadins (64.94-70.26%), secalins (47.50-60.70%) and hordeins (34.06-45.17%) respectively (Table S1). Among the cultivars, the gliadins fraction from HPW-349 contained the highest PC while hordein fraction from BH-959 contained the lowest. The PC of prolamin fractions was found to vary significantly (p ≤ 0.05). DuPont et al. 11 reported 40.1% protein content in extracted gliadins fraction which was slightly lowered than our findings. He et al. 12 reported higher protein content of dialyzed gliadin (85%) using alcohol extraction protocol and conversion factor of 6.25 instead of 5.7 used in this study. However, Schalk et al. 13 also reported the protein content of dialysed gliadin, secalin and www.nature.com/scientificreports/ hordein to be 93.5%, 89.4% and 87.3% respectively using 5.7 conversion factor and alcohol extraction protocol. The difference in the crude protein content may be due to difference in extraction protocol, dialysis performed and also to some extent genetic makeup of cultivars. The color characteristics of the prolamin fractions were explained via CIE color values (L*, a*, b*, Hue angle (H o ) and Chroma (C*). L* value which indicates the lightness was found to be in the range 53.13-68.12 (Table S1). In the investigated cereals, the L* value for hordein fraction from BH-902 was found to be significantly (p ≤ 0.05) higher while lowest for secalin fraction from MCTLG-1. The dark color (lower L* value) of the secalins fraction might be due to the presence of pigment and oxidative changes as compared to hordeins and gliadins fraction 44 . In case of coordinates a* and b*, the observed positive values for prolamin fraction fall in the range 0.53-2.06 and 6.90-13.93 implying that prolamin fractions had red and yellow tinge respectively. For the investigated cultivars, the a* value was found to be the highest for gliadins fraction (1.53-2.06) and the lowest for hordeins fraction (0.53-1.28). On the other hand, the higher b* values for gliadins fraction (9.68-13.93) reflected more yellow tint in comparison to secalins fraction (6.90-10.13) and hordeins fraction (7.41-9.61). On the whole, a significant (p ≤ 0.05) difference in a* and b* values were observed among the prolamin fractions at inter as well as at intra-cultivars levels.
Genetic proximity of various accessions in the present study was further evaluated through Jaccard Similarity Matrix by means of Un-weighted Pair Group Method with Arithmetic Averages (UPGMA) to construct a dendrogram ( Figure S4). Three distinct groups could be identified from the dendrogram of gliadins, secalins and hordeins fraction; moreover, gliadins and hordeins fractions were closer to each other than secalins fraction. In case of gliadins fraction, group-I had two accessions where HPW-42 was found to be dissimilar from rest of the other extracted gliadins with similarity coefficient of 0. The conventional methodology for determining amino acid composition, which included acid digestion followed by chromatographic analysis, was examined. Some amino acids undergo a various kind of chemical changes during acid hydrolysis. Tryptophan is destroyed; serine, cysteine, and threonine are partially destroyed; methionine is oxidized; tyrosine is halogenated or oxidized; valine and isoleucine require 72 h for complete hydrolysis and are only hydrolysed to about 50-70% at 110 °C in 24 h 40 . As no correction factors for these phenomena were determined in the present manuscript, the amino acid levels reported on are estimated values, rather than representing an exact quantitation.
The Tyr residue was found to be higher in hordein (3.85-4.93%) followed by gliadin (3.35-3.83%) and secalin (2.81-3.27%) fractions, whereas Arg content was found to be the highest in hordeins (3.46-4.36%), followed by secalins (2.56-3.96%) and gliadins (2.87-3.74%) fractions. This means that in Tyr-Tyr covalent interactions in the secalin was comparatively less active than hordeins and gliadins fraction. Hou et al. 50 suggested that the insolubilizing interactions has been facilitated by Asp, Glu, Arg and aromatic residues (Phe, Tyr, Trp, His). The role of Tyr-Tyr bond in stabilizing the toxic repetitive moiety of 33-mer gliadin was reported by Amundarain et al. 23 . The phenolic group of Tyr participates in formation of oxidative radical and thus stabilizes the intermediate proteins through the covalent bond of di-tyrosine.
Total non-essential amino acid (TNEAA) comprises of 66.45-73.91% of the total amino acid and being the highest in gliadins fraction from HPW-236 and the lowest from hordeins fraction from BH-959. The higher TNEAA of gliadins (69.90-73.91%) and secalins (68.52-73.34%) fractions are speculated to have more compact and stabilized protein-protein interaction as compared to hordeins (66.45-70.90%) fraction.
The NEAA profile obtained in present investigation is consistent with the literature. Similar results were reported by Field et al. 26 26 reported more serine and threonine in hordein than gliadin and secalin fractions which are similar to our finding, while ala was high in gliadin and secalin than hordein fraction. Our findings show non-significant (p ≥ 0.05) difference in the estimated amount of Ala in gliadin, secalin and hordein fraction. Gellrich et al. 27 analysed secalin for amino acid and reported similar results except lower value for Arg, Lys and higher content of pro (20.4%) in comparison to current findings.
The International ImMunoGeneTics (IMGT) system of information categorizes the amino acids into three 'hydropathy' groups (Table 4) namely hydrophilic (Lys, Arg, Gln + Glu and Asn + Asp), hydrophobic (Phe, Trp, Ileu, Leu, Met, Cys, Ala, and Val) and neutral (Tyr, Pro, Gly, Thr, Ser and His) 52 . Out of total amino acids, the hydrophilic amino acids group constituted about 47.08-52.22% in gliadins, 45.28-47.97% in secalins and 41.12-45.09% in hordeins fraction. The highest hydrophilic amino acids content was observed in gliadins fraction from cv. HPW-155 and the lowest in hordeins fraction from cv BH-959. The hydrophilic amino acids of prolamin fraction were further sub categorized into basic and acidic amino acids, which accounted for 5.33-8.70% and 32.60-46.86% of the total amino acids respectively. The acidic AA residues which also include their amines (Gln + Glu and Asn + Asp) were found to be high in gliadins fraction (41.02-46.86%) and low in hordeins fraction (32.60-36.39%) whereas basic AA residues were high in hordeins fraction (8.11-8.70%) and low in gliadins fraction (5.33-6.56%). Acidic, basic and total hydrophillic amino acids of extracted prolamin fraction showed a significant (p ≤ 0.05) difference at intra and inter cultivar level.
The hydrophobic AA constituted 24.46-29.46% of the total amino acids and comprised of aliphatic, S-containing and some aromatic amino acids which accounted for 18.02-20.88%, 0.69-1.68% and 5.19-7.79% of the total amino acids respectively. The hydrophobic aromatic amino acids found to be higher in hordeins (5.91-7.79%) and secalins (5.68-7.09%) while lower in gliadins (5.19-6.48%) fraction. Prolamin aliphatic, S-containing and www.nature.com/scientificreports/ total hydrophobic AA group showed a non-significant (p ≥ 0.05) difference while the hydrophobic aromatic amino acids varied significant (p ≤ 0.05) at intra and inter cultivar level. The total neutral amino acids of prolamin fraction varied from 22.21-30.48% and consisted of non-polar, polar and aromatic amino acids which accounted for 11.33-18.79%, 6.55-9.69%, and 2.81-4.93% respectively. The polar AAs were higher in hordeins (7.50-9.50%) and secalins (7.46-9.69%), while lower in gliadins (6.55-8.59%) fraction. Similarly, the non-polar AAs were higher in hordeins (14.07-18.79%) and secalins (12.53-16.92%), while lower in gliadins (11.33-15.96%) fraction. Hordeins fraction from BH-393 and gliadins fraction from HPW-236 accounted for the highest and the lowest proportion of polar amino acids respectively. Whereas the highest proportion of non-polar AA was observed in hordeins fraction cv BH-902 and the lowest in gliadins fraction cv HPW-155. The neutral aromatic AAs were found to be highest in hordeins fraction, BH-393 and the lowest in secalins fraction, MCTLG-3. This implies that Tyr-Tyr covalent interactions speculated to be more in hordein which attributed to greater aggregation tendency than gliadin and secalin fractions 23 .
The non-polar, polar, neutral aromatic and total neutral amino acids were observed to vary significantly (p ≤ 0.05) among prolamin fractions at inter and intra cultivar level except among wheat cultivars of neutral aromatic group where the values varied non-significantly (p ≥ 0.05). The variation in amino acid composition is considerably affected by genotype, irrigation practices, fertilizer application, and environmental conditions 53-58 . Secondary structure of prolamin fractions. The deconvoluted fitted curve and the relative amount of secondary structural components in amide I region (1600-1700) of extracted prolamin assigned to specific absorption frequencies are presented in Table S2, Fig. 3 and Figure S5. The structural component positioned between 1612-1624 cm −1 , 1630-1642 cm −1 , 1643-1648 cm −1 , 1649-1659 cm −1 , 1660-1685 cm −1 and 1686-1699 cm −1 has been assigned to inter-molecular β-sheet, β-sheet, random coil, α-helix, β-turn and β-turn + β-sheet secondary structure respectively 29,41,46,59 . The regions centred at 1699 cm −1 , 1695 cm −1 and 1688 cm −1 were assigned as β-turn + β-sheet of prolamin and found higher in gliadins (11.90-18.55%) followed by secalins (6.72-12.33%) and least in hordeins  www.nature.com/scientificreports/ in the range of 26.32-37.91%. Indeed, the higher proportion of β-turn in prolamin fraction was supposed to be attributed due to higher content of proline in repetitive domain of prolamin 45 . Wouters et al. 46 suggested that β-turn are rich in ω-gliadin and possess only lower amount of β-sheet and α-helix, which seems consistent with our findings. The ω-gliadin and C-hordein were also found also rich in β-turn as revealed by Purcell et al. 15 .
The previous report on the secondary structure of hydrated gluten via FTIR showed absence of random coil component 41 similar to our findings in the present work. Our results were also in consistent with that of Li et al. 60 , who reported more β-turn and relatively lesser amount of β-sheet structural configuration in soluble glutenin and gliadin protein fractions. Findings on the structural properties of gliadin containing β-sheet (39.46-40.18%), random coil (13.73-14.24%), α-helix (13.98-14.23%) and β-turn (31.62-32.46%) were in close agreement with our results 61 .
Particle characteristics. The prolamins exist as three-dimension structured particle and its microstructure was analysed using analytical instruments to elucidate its properties.
Dynamic light scattering (DLS). The prolamin fractions were analysed for particle size distribution (PSD) in terms of polydispersity index (PDI) and hydrodynamic diameter (D h ) in solution of acetonitrile:water:formic acid:: 50:50:0.1 under our experimental condition as presented in Table S3, Figure S6. PDI values for prolamins fall in the range 0.49-0.87. Lower PDI value i.e. less than 0.5 indicates more uniform PSD and existence of monodisperse system 62 . In the present study, PDI of the secalins (0.66-0.87), hordeins (0.58-0.83) and gliadins (0.49-0.64) were higher indicating disparity in particle size.
The gliadins and hordeins mainly existed in monomeric form with hydrodynamic diameter (D h ) in the ranges between 1.23-1.83 nm and 6.67-7.62 nm respectively. The hordeins from BH-946 (7.62 nm) had higher and gliadins from HPW-42 (1.23 nm) had lower hydrodynamic diameter. Gliadins showed monomeric units nonetheless hordeins shows the existence of some higher ordered aggregates. This fact is also supported by amino acid composition that the higher proportion of basic AA and S-containing AA residues in hordeins show their tendency in formation of aggregates through strong electrostatic and hydrostatic interaction 63 .
The secalins showed a bimodal PSD with two populations in the solution. The particle size of small and large sized population varied from 7.73-18.44 nm and 26.48-68.33 nm, whereas, their proportion ranged from 99.60-99.90% and 0.10-0.40% respectively. It indicated that the small sized particles were abundant in number while large sized particle were scanty. The large sized particles were observed in secalin fraction from MCTLG-3 (18.44 nm; 68.33 nm) and the small sized from MCTLG-4 (7.73 nm; 26.48 nm) in peak 1 and 2 respectively. The PDI and D h of the investigated prolamin have been found to vary significantly (p ≤ 0.05) among different extracted prolamin samples including intra-cultivar differences except wheat cultivars which shares non-significant differences (p ≥ 0.05).
The zeta potential (ZP) profile of the gliadins, secalins and hordeins fraction observed a positive value in the range of 23.53-27.00 mV, 11.23-16.60 mV and 4.10-7.98 mV respectively. The ZP values were found to be significantly (p ≤ 0.05) different at inter as well as intra-cultivar levels. The relatively higher zeta potential values of gliadin fraction indicated the high stability of their monomeric protein subunits. Gliadins and hordeins fractions had a monomodal PSD while secalins fraction had a bimodal PSD by number-based size distribution ( Figure S6a-d). The secalin fraction have relatively lower ZP than gliadins fraction and these lower ZP values are very well corroborated by the size measurement values of the secalins fraction where large size particles have been observed along with small particle that results in its bimodal PSD. The lower zeta potential values (< 30 mV) indicate the lower electrostatic charge on the particle hence lower repulsions among them that can leads to the aggregation of particles to form bigger aggregates hence the presence of large size particles in size distribution profile of secalins fraction can be correlated with the its lower ZP values.
Further the intensity-based size distribution ( Figure S6e-g) showed higher ordered aggregates in hordein followed by secalins and gliadins. On the other hand, the substantially lower ZP values for hordeins fraction reflected the greater tendency of protein motifs to form aggregates due to lower surface charge. The constituent protein units predominantly exist in higher ordered form in secalins and hordeins fraction compared to gliadins fraction, which was correlated by their corresponding PDI value. The low ZP in case of hordein fraction is probably due to its high basic amino acid contents which might be involved in π-interaction and increase aggregation propensities 50 .  64 has reported that zeta-potential of gliadin was about 30 mV and had remarkable colloidal stability. However, gliadin nano-particles prepared by controlled aggregation had D h , PDI and ZP of 190-220.6 nm, 0.067-0.232 and 14.8-18.3 mV respectively 46 . Peng et al. 65 have also demonstrated the high zeta potential which indicated greater electrostatic repulsions among the protein molecules and poses higher electrostatic barrier which prevents protein aggregation. The overall zeta-potential of gliadins, secalins and hordeins was observed to be in decreasing order. It further indicated the solution stability of these proteins was also in the same order. The higher colloidal stability of gliadins suspension is attributed to the high glutamine content which is engaged in hydrogen bonding with water 21 . The variation in zeta potential between inter and intra cultivar of present and earlier studies might be due to difference of dispersion media, amino acid side chains, pH, ionic strength of solution and temperature during extraction/analysis process 21,64 .
SEM and EDX. Scanning Electron Microscopy (SEM) and Energy Dispersive X-Ray Analyzer (EDX) were used to investigate the surface morphology and elemental composition of prolamin respectively as elucidated in Fig. 4, Figure S8 and Table S4. Furthermore, the cultivar HPW-42 gliadin, MCTLG-5 secalin and BH-393 hordein were selected for further morphological, internal structural and elemental analysis, based on relatively higher proportion of LMW group (α/β, γ-gliadin and LMW-GS) and lower proportion of HMW group at intra cultivar level. Morphology by SEM illustrated the globular particle arrangement of gliadin while sheet-like and stacked flaky structure was observed for secalin and hordein respectively. Similar globular particle arrangement has been reported in gliadin nanoparticles 5 20 also reported crystalline peaks in XRD analysis of gliadin, however these peaks were at different position. Similar crystalline diffraction was reported by Guan et al. 30 for electrospun fibre of hordein at 2θ values of 16.88°, 38.3° and 44.6°, which was in close proximity to our findings.
The inter-planar spacing, ' d' , was calculated from the peak positions using Bragg's equation (Table S6). The inter-planar spacing corresponding to the diffraction peak at 44. Transmission electron microscopy. Transmission Electron Microscopy was used to study the morphology and nanostructure of prolamin fractions. The micrograph was analysed for surface morphology (Fig. 4D-I), lattice planes and ring diffraction patterns of selected area (Fig. 5D-L). Morphology of gliadins, secalins and hordeins fraction showed a compact spherical structure (Fig. 4G-I) which might be due to active participation of α/βgliadins polypeptides 68 . A rod like structure (Fig. 4 D-F) was also observed in gliadins with an average diameter of 17.94 nm and length of 250 nm. Formation of such structures is attributed to involvement of ω and γ-gliadin polypeptides 68 . Ang et al. 17 also supported the existence of two structures of gliadins as compact globular and rod-shaped. The histogram shows the particle size distribution of compact globular structures in prolamin fractions and rod structure of gliadins fraction ( Figure S9). The average particle size was found to be smallest for gliadins (3.88 nm), followed by hordeins (4.32 nm) and largest in secalins (5.79 nm) fraction. The inset in panel (K) also displays hexagonal symmetry of the sample. This space saving molecular arrangement might be responsible for gliadin to act as plasticizer. Similar molecular arrangement of gliadin as hexagonal close packed was previously reported by Rasheed et al. 19 which is in line with our findings. www.nature.com/scientificreports/ It is to be noted that there is a lack of detailed XRD analysis and corresponding (hkl) values for different planes present in these sample. A few recent reports that appeared on same sample also lack the detailed XRD analysis and hence accurate Miller indices cannot be assigned to these planes 18-20 . Comparison among analytical techniques. Particle size of the prolamins (Table 5) observed for gliadins (1.12-6.06 nm), secalins (3.23-9.47 nm) and hordeins (2.13-6.83 nm) fractions using TEM whereas for gliadins (1.23-1.83 nm), secalins (6.45-11.96 nm) and hordeins (6.67-7.62 nm) fractions using DLS and for gliadins (44.1°, 12.29-13.67 nm; 37.8°, 11.23-14.33 nm), secalins (44.1°, 12.47-13.05 nm; 37.8°, 14.01 -17.31 nm) and hordeins (44.1°, 12.29 -14.13 nm; 37.8°, 13.40-16.63 nm) fractions using XRD. Among the prolamin fractions, the particle size of secalins was the largest, followed by hordeins and gliadins fraction. Among the analytical techniques, largest particle size was observed on XRD while TEM and DLS revealed particle size in a close range. However, the particle size range overlapped in all the three prolamin fractions.
The crystalline microstructure of prolamin was shown by both XRD and TEM studies. SEM images and low resolution TEM images confirmed the morphological differences between different samples. Further, HR-TEM images from samples indicated the presence of variable sized nanostructures.
The d-spacing was estimated from HR-TEM lattice plane pattern, Selected Area Electron Diffraction (SAED) pattern and XRD data. The d-spacing (Table S6) observed by HR-TEM, SAED and XRD for gliadins fraction was 0.21 nm, 0.29 nm and 0.21-0.85 nm whereas for hordeins fraction, it was 0.35 nm, 0.32 nm and 0.21-0.85 nm while for secalins fraction, it was 0.32 nm, 0.21-0.34 nm and 0.21-0.85 nm, respectively. For gliadins fraction the d-spacing obtained via XRD and HR-TEM imaging matches with each other. Further, as evident from SAED pattern and XRD data, all three samples have polycrystalline nature. These periodic regular arrangements probably attributed to the subsistence of β-sheet structure among the prolamin and disulphide bonding as a result of intra-molecular interaction which mainly involves cysteine residues in polypeptide chains 69 . Markgren et al. 63 highlighted the role of cysteine amino acid in intra-molecular disulphide bonding, and stated that the intramolecular interaction enhances the tendency of formation of more ordered β-sheet configuration. Similarly, Jung et al. 70 also reported the crystalline structure of proteins by XRD studies and attributed it to the β-sheet structure. Rasheed et al. 19 have also reported that the modified gliadin with nano-crystalline structure consists of high proportion of β-sheet structure along with irreversible linkages such as covalent disulphide bonds.

Conclusion
Freeze dried powder of gliadins, secalins and hordeins fraction had a light cream color. The crude protein content of extracted gliadins was found to be higher than secalins and hordeins fraction. The secondary structure of extracted prolamin revealed higher α-helix and random coil components in hordeins, while higher β-sheet and β-turn components in gliadins. Prolamin fractions were abundant in Gln + Glu acid and Pro content among the non-essential amino acids while Leu, Phe and Val among the essential amino acids. SDS-PAGE revealed that prolamin fractions contained minor amount of HMW-GS while cross-contamination with significant amount of albumin and globulin. This method seems to extract majority of ω-prolamin fractions and had well resolved LMW region, but still needs stepwise extraction with sodium chloride and water so that the prolamin fractions are free from these metabolic active proteins. Furthermore, the instrumental analysis reported here might be influenced by presence of albumin and globulin fractions. SEM images showed globular morphology of gliadin fraction while sheet like or stacked flaky morphology of hordein and secalin fractions. TEM analysis of gliadin fraction showed a compact spherical structure as well as a rod like structure. Particle size determined by TEM and DLS were in close proximity. XRD results indicated that prolamin had both amorphous as well as crystalline structure. Furthermore TEM-SAED pattern also revealed the polycrystalline nature and a hexagonal symmetry. The zeta potential of the gliadin fraction was higher followed by secalin and hordein fractions. The energy dispersive X-ray analyzer revealed the major elements present in the prolamin fractions as C, O and N, while minor elements as S, P, Na and I.