In addition, the bands corresponding to the acidic and basic subunits of edestin are still visible under non-reducing conditions, although with lower intensity. At both reducing and non-reducing conditions, minor bands at ≤10 kDa were identified as the water-soluble 2S albumin storage protein, which consists of 2 subunits of 7 and 3 kDa linked by disulphide bonding . At non-reducing conditions, a low molecular weight band between 10 and 14 kDa was observed, suggesting the presence of the albumin complex. As all gels were loaded with the same protein content, slight differences between the intensity of some of these bands could be ascribed to different globulin ratio, protein conformation, as well as other interactions between non-protein minor components. Hemp 3 and Hemp 1 presented albumin bands that were visibly more intense than those for the other hemp samples. Moreover, at the top of the gels, these samples did not show any diffused band corresponding to high molecular weight protein aggregates. Remarkably, Hemp 3 and Hemp 1 were the only samples deprived from any wet processing step, suggesting that any of the protein enrichment processes used in Hemp 2, Hemp 4, and Hemp 5 leads to albumin loss or formation of protein aggregates. Tanger, Engel, and Kulozik reported the same phenomenon for AE-IP-extracted pea protein. Precisely, the albumin fraction was not precipitated during the precipitation step and remained soluble in the supernatant due to the higher isoelectric point of the albumin than the globulin fraction . Therefore, albumin fractions are mostly discarded during AE-IP process of plant proteins . In addition, Hemp 5 showed the lowest band intensity at non-reducing conditions suggesting the highest amount of disulfide bonds in this protein . Wang and Xiong reported that shifting the pH from neutral to alkaline conditions can significantly increase the disulfide bonds in hemp protein. In general, cannabis dry rack aqueous protein extraction is performed first in an alkaline solution to solubilize the protein and then in an acidic solution to precipitate them.
The former process might have increased disulfide bonding of Hemp 5, as it underwent two aqueous extraction steps.SDS-PAGE of pea protein showed bands ranging from 100 to 12 kDa, which represent mainly 11S basic and 11S acidic subunits and 7S subunits.Vicilin was identified as the main fraction based on a previously reported study . Native-PAGE of pea and hemp samples generally displayed an intense unresolved/diffused band at high molecular weight . Native-PAGE electrophoresis provides non-denaturing conditions for proteins to migrate into the gel without any change in their intra, intermolecular interactions. Hemp 1 and Hemp 3 displayed the most intense and unresolved band at high molecular weight, suggesting that these proteins retained their native solubility and molecular weight. However, Hemp 1, with an additional pelletizing step compared to Hemp 3, showed that some proteins resolve throughout the gel , suggesting that pelletizing could dissociate some bonds that link the different edestin subunits. This effect was even more noticeable for Hemp 2 and Hemp 4, indicating that AE-IP further breaks down the oligomeric state of hemp proteins. Moreover, this occurrence would suggest that pea protein concentrate was also subjected to AE-IP. Interestingly, Hemp 5 did not show any visible bands , which indicates that the double AE-IP performed in Hemp 5 led to an extensive denaturation and aggregation of proteins that were then removed during centrifugation . In fact, AE-IP has been previously reported to increase the aggregation of hemp protein isolate from hemp seed cake/meal , which would explain the lack of unresolved bands in the Native-PAGE electropherogram . A faint band appeared around 10 kDa in Hemp 2 and Hemp 4 samples which might be attributed to either low molecular weight peptides formed upon protein breakage or to the albumin fraction that was only visible in those Native-PAGE electropherograms that were subjected to a single AE-IP step . Among AE-IP samples, Hemp 5 showed the least intense albumin band, which could have resulted from an extensive protein denaturation and aggregation , or more albumin loss during the double AE-IP processing steps. This occurrence agrees with SDS-PAGE results.
The extent of thermal denaturation of proteins during heating was investigated by DSC from 25 to 120 ◦C . Protein concentrate samples exhibited a single endotherm, which has been previously associated with the rupture of hydrogen bonds, or with the unfolding of the most predominant protein fraction, edestin . This single endotherm of hemp samples presented enthalpy values ranging from 7.5 to 23.7 J/g protein , which agrees with previous work on hemp protein isolates . Hemp 3 exhibited the highest enthalpy , indicating that this protein enrichment technology, involving only the removal of the hull fraction, resulted in a concentrate with proteins of higher structural order, i.e., larger proportion of undenatured edestin. The rest of hemp samples, characterized by additional protein enrichment steps, had lower enthalpies , suggesting that some protein damage could have occurred during any of these processes . This was particularly evident in Hemp 5 , which together with Native-PAGE results, suggested that the double AE-IP leads to greater hemp protein denaturation, as reported before by comparing AE-IP and micellisation . Hemp 2 and Hemp 4 presented a visibly higher Td50 compared to the other samples tested in this work, which exhibited Td50 ranging from 91.4 to 95.2 ◦C . Hemp 2 and Hemp 4 also showed distinctively higher viscosity peaks during a heating/cooling cycle than the rest of the pea and hemp samples . They also seemed to exhibit higher propensity to form gels after RVA . We note that RVA data of hemp samples was very noisy, which has been reported to be due to the formation of stiff chunks during the RVA analysis . Interestingly, Hemp 2 and Hemp 4 also possessed a significantly higher potassium content than the rest of protein samples , which is a well-known kosmotrope according to the Hofmeister series . The kosmotropes are strongly hydrated and have stabilizing and salting-out effects on proteins . In other words, the enhanced stabilizing interactions in folded conformations of edestin chains could have been due to the high concentration of kosmotropes. The higher Td50 could also be due to the diverging denaturation temperatures of edestin and other protein fractions. In this regard, Raman spectroscopy was employed to study the vibrational modes of the amide-I region and revealed that AE-IP samples possessed a significantly higher relative proportion of ordered secondary structures , i.e., α-helices and β-sheets, that could contribute to the higher thermal stability or Td50 as reported by Motta, Fambri, and Migliaresi .
In this regard, the significant protein denaturation and aggregation of Hemp 5 , together with its lower content of kosmotropes and higher protein content, could explain its lower Td50 than Hemp 2 and Hemp 4. Further studies are needed to understand the contribution of these two explanations to the thermal stability of hemp proteins, which should be taken into account during food structuring processes. Hemp 1–3 showed roughly similar secondary structures, whereas Hemp 4 and Hemp 5 displayed higher α-helix content and lower β-sheet contents. This could be attributed to a different hemp genotype, environmental and soil growth conditions, and/or to the different processing enrichment processing parameters used to produce Hemp 4 and Hemp 5. Surface hydrophobicity of proteins is an important feature dictating interfacial and functional properties.Although the amount of hydrophobic amino acids in pea and hemp protein is roughly the same, hemp protein concentrates possessed higher surface hydrophobicity than pea protein , indicating a higher number of non-polar patches on the hemp protein surface to which the ANS fluoroprobe can attach. The different H0 values among hemp samples could be explained by the differences in protein denaturation upon protein enrichment processing , which was extensively discussed in the previous section. In fact, protein unfolding can cause exposure of both polar and non-polar amino acid residues and, subsequently, an increase in H0. However, if protein remains in denaturing conditions for longer times, exposed hydrophobic groups can interact and form protein aggregates that results in a decrease in H0 ; Therefore, surface hydrophobicity cannot solely provide complete information on the extent of protein aggregation and denaturation. Hemp 2 displayed the highest H0 value among all the samples, which can be attributed to intermediate processing conditions leading to unfolded proteins with a high exposure of hydrophobic amino acids. Meanwhile, Hemp 4, with virtually the same processing steps as Hemp 2, plant racks showed different surface hydrophobicity, which could likely be explained by its higher lipid content. More specifically, lipid molecules could mask hydrophobic patches on the surface of the protein and inhibit ANS molecules from attaching to the non-polar residues. Hemp 5 and pea protein showed the lowest H0 values . Especially in the case of hemp, this might be due to the extensive unfolding and subsequent refolding upon the dual AE-IP.
The visual appearance of protein concentrates is shown in Fig. 3. Hemp 3, the only sample extracted from dehulled seeds, possessed the lightest colour among the hemp protein concentrates. The hempseed hulls are darker than the inner part of the seed, resulting in Hemp 3 protein concentrate having a lighter colour, in agreement with CIE La*b* colour measurements . Likewise, Hemp 1, the sample with the highest hull ratio displayed visibly darker spots and a heterogeneous visual appearance. Hemp 5, followed by Hemp 4 and Hemp 2, exhibited a homogeneous darker colour than pea and Hemp 3 samples. As discussed before, all protein concentrates exhibited a high content of phenolic compounds, presumably phenolic acids, which might be greatly oxidised to quinones during the alkaline extraction of proteins. This is of paramount importance, since phenolic acid quinones can react with the amine groups of proteins, resulting in dark coloured adducts . Therefore, the richer content of phenolic acids of hemp seeds, as compared to other plant grains, such as peas, is particularly important to consider during alkaline extraction of proteins, which explains the darker colour of alkaline extracted samples. The solubility of pea and hemp samples increased at alkaline conditions as shown in Fig. 4 and as typically reported for pea and hemp proteins . Although hemp proteins exhibited higher surface hydrophobicity than pea , their solubility was generally higher at neutral and alkaline pH . The amount of proteins present in the supernatant is influenced by their processing-related aggregation, which results in sedimentation during centrifugation . Hence, a high pea protein denaturation and aggregation could partially explain these phenomena.In fact, chaotropic salts are able to bind to hydrophobic patches on the protein surface and disrupt intra- or inter-molecular hydrophobic interaction that increases their solubility . The solubility of Hemp 3 was noticeably higher than other hemp proteins and this can be attributed to a higher proportion of the albumin fraction, as observed by SDS-PAGE . Hemp albumins are characterized by a higher solubility compared to the globulins due to the reduced level of aromatic and hydrophobic amino acids and less rigid conformational structure . Nevertheless, Hemp 1, with similar proportions of albumins and H0 as Hemp 3 , exhibited lower solubility, which could be explained by differences in their secondary or tertiary structure and denaturation levels . Pea protein showed a significantly greater WHC than hemp samples, which can be explained by its higher number of exposed hydrophilic groups that can be easily hydrated. This feature also resulted in native pea protein to uniquely display a constant high initial viscosity around 500 cP in the RVA curve . In fact, hydration of proteins can improve protein elongation and increase its hydration volume . Upon heating, the WHC, HWHC and viscosity development of pea protein decreased, whereas those of hemp samples increased . Hemp proteins are tightly bound in a globular matrix with disulfide bonds , in which proteins unfold upon heating and cause the polar groups in the structure to potentially bind more water, increasing their HWHC and making it more similar to that of pea protein.