The core challenge is architectural. Many plant proteins fold into compact globular shapes stabilized by dense networks of hydrogen bonds, hydrophobic interactions, and disulfide linkages. These bonds provide stability but also bury the functional groups that proteins need to dissolve in water, stabilize emulsions, form foams, and create gel networks. The result is poor solubility, weak interfacial behavior, limited digestibility, and sensory performance that often falls short of animal proteins. The review emphasizes that these constraints operate across a hierarchy of scales, from the molecular conformation of secondary and tertiary structures to the microstructural organization of protein aggregates and the macroscopic properties of gel networks, and it is this multiscale organization that ultimately governs how plant proteins behave in complex food matrices.

Ultrasound processing has emerged as one of the most versatile tools for breaking through these structural barriers. Operating at frequencies between roughly 20 and 100 kilohertz, ultrasound triggers acoustic cavitation, a process in which microscopic bubbles form, grow, and collapse violently within liquid protein suspensions. Each bubble collapse generates localized shear forces, microjets, and shock waves that can disrupt protein aggregates and promote partial unfolding without the extensive thermal damage associated with conventional heating. The review highlights that these mechanical forces primarily target the non-covalent interactions stabilizing protein structure, weakening hydrogen bonds and hydrophobic associations that hold aggregates together.

The structural consequences are measurable and significant. Jambrak and colleagues demonstrated that ultrasound treatment altered the secondary structure of soy proteins, reducing the proportion of alpha-helical structures while increasing random coil configurations, which enhanced protein flexibility and improved solubility and emulsifying capacity. Li and co-workers reported that ultrasound reduced the particle size of pea protein aggregates from 831.6 to 346.7 nanometers under controlled conditions, with the resulting smaller particles exhibiting improved interfacial activity and producing more stable emulsions over time. Hu and colleagues observed that ultrasound treatment increased the availability of free sulfhydryl groups in soy protein isolates, promoting stronger gel formation and improved water-holding capacity. These molecular-level changes cascade upward: disrupted hydrogen bonds drive partial unfolding, which exposes previously buried hydrophobic residues and reactive groups, which then drive controlled aggregation or interfacial adsorption, ultimately determining the macroscopic solubility, emulsification, and gelation properties of the final food product.

Pulsed electric field processing operates through a fundamentally different mechanism. Short bursts of high-voltage electricity, typically between 5 and 50 kilovolts per centimeter with microsecond pulse durations, temporarily disrupt cell membranes through electroporation, creating nanoscale pores that facilitate the release of intracellular proteins without harsh solvents or severe thermal treatments. Barba and colleagues reported that PEF pretreatment improved protein extraction yields from plant materials by approximately 30 percent. Beyond its effects on cellular structures, PEF directly influences protein molecules by interacting with molecular dipoles and charged amino acid residues, causing polarization and reorientation that can disrupt hydrogen bond networks and modify the hydration layer surrounding proteins. The review offers a particularly elegant physical explanation for why alpha-helical segments are preferentially affected: the alpha-helix carries a substantial net dipole moment along its axis because the peptide-bond dipoles are aligned in the same direction, making these segments more strongly coupled to an externally applied electric field than beta-sheet segments, which leave a comparatively small net dipole.

Cold plasma treatment takes yet another approach, acting primarily at the protein surface rather than inducing extensive unfolding throughout the molecular structure. Plasma is a partially ionized gas containing energetic electrons, ions, ultraviolet photons, and reactive oxygen and nitrogen species. When these reactive species interact with exposed amino acid residues, they initiate oxidation-reduction reactions that modify protein side chains. Sulphur-containing amino acids such as cysteine and methionine are especially susceptible, forming sulfoxides, disulfide cross-links, or carbonyl derivatives. Misra and colleagues demonstrated that cold plasma treatment increased the emulsifying activity of soy protein isolates because the introduction of oxygen-containing groups enhanced protein surface polarity and interfacial adsorption. Segat and colleagues reported improved foaming capacity in plasma-treated wheat proteins, while Ji and colleagues showed enhanced solubility and dispersibility of plasma-treated peanut proteins. Because cold plasma modifies surfaces selectively, it allows fine control over interfacial behavior while minimizing structural degradation.

Ohmic heating, which the review carefully distinguishes as an electro-thermal rather than strictly non-thermal technology, passes alternating electric current directly through the food material, generating heat internally through electrical resistance. Unlike conventional heating, where thermal energy transfers gradually from external surfaces inward, ohmic heating produces rapid and uniform volumetric heating throughout the entire food matrix. This minimizes temperature gradients and prevents localized overheating, allowing greater control over protein denaturation and aggregation behavior. Sakr and Liu observed that ohmic heating promoted more homogeneous protein unfolding and network formation compared with conventional thermal methods, with the resulting protein gels exhibiting improved water-holding capacity, elasticity, and structural stability. The review notes that this controlled heating approach is particularly valuable for developing plant-based meat analogues and high-moisture protein systems where uniform network formation is essential for replicating the fibrous texture of animal muscle tissue.

One of the most striking findings of the review is that no single set of processing parameters is optimal across all plant protein sources. Soy and pea proteins are dominated by 7S vicilin and 11S legumin globulins, with the more loosely associated 7S fraction unfolding and solubilising more readily than the disulfide-rich 11S fraction under identical ultrasound treatment. Oat protein, dominated by a compact, highly disulfide-stabilised 12S globulin with unusually high thermal and structural stability, requires substantially higher ultrasound intensities to achieve comparable particle-size reduction and solubility gains. Wheat gluten proteins, being prolamins rather than globulins, respond to ultrasound primarily through changes in disulfide crosslink density and aggregate structure rather than through classical alpha-helix-to-beta-sheet transitions, since their native structure already contains little ordered secondary structure. Faba bean globulins occupy an intermediate position, sharing the 7S/11S architecture of legumes but with a subunit composition that produces a somewhat lower unfolding threshold and a correspondingly higher risk of re-aggregation if treatment is not carefully controlled.

The review also sounds a cautionary note about overprocessing. Beyond a defined favourable range, the same mechanisms that improve functionality at moderate intensity begin to cause functional deterioration. At high acoustic intensities, water sonolysis generates hydroxyl and other free radicals that oxidise newly exposed free sulfhydryl groups back into disulfide bonds, promoting irreversible protein re-aggregation rather than continued unfolding. Excessive cold plasma exposure can over-oxidise surface amino acid residues, producing carbonylation and cross-linking reactions that stiffen the protein surface and reduce interfacial activity. Once excessive aggregation has occurred, the resulting dense aggregates can become less accessible to digestive proteases than either the native protein or the moderately processed form, producing an inverted U-shaped relationship between processing intensity and digestibility. The point at which structural modification becomes detrimental is technology- and protein-source-specific, and is generally reached well before complete protein unfolding or maximal cavitation intensity.

Looking forward, the review identifies computational modelling and machine learning as critical tools for optimizing non-thermal protein processing. Molecular dynamics simulations allow researchers to observe atomic-level structural changes in proteins subjected to mechanical stress, electric fields, or pressure variations, providing insights into how protein folding patterns respond to different processing environments. Machine learning algorithms such as artificial neural networks and random forest models can identify relationships between processing parameters and functional outcomes that may not be evident through conventional statistical analysis, predicting how variations in ultrasound intensity, treatment time, or PEF strength influence solubility, emulsifying capacity, and gel strength. The integration of real-time sensors with machine learning algorithms enables adaptive process control, where processing parameters are adjusted dynamically based on incoming data to ensure consistent product quality despite variations in raw materials. As the review concludes, the most energy-efficient operating point for a given technology is generally the same moderate-intensity window that avoids overprocessing, meaning that energy efficiency and functional quality are not competing objectives but are, in practice, optimised together, bringing the food industry closer to producing plant-based products that are nutritionally balanced, environmentally sustainable, and appealing to consumers.

Subject of Research: Food Science and Soft Matter

Article Title: Multiscale structural modulation of plant proteins via ultrasound and electro-physical processing: from extraction to functional food matrices

Article References: Majeed, N., Younis, K., & Yousuf, O. (2026). Multiscale structural modulation of plant proteins via ultrasound and electro-physical processing: from extraction to functional food matrices. Frontiers in Soft Matter, 6, Article 1881300. https://doi.org/10.3389/frsfm.2026.1881300

Image Credits: AI Generated

DOI: 10.3389/frsfm.2026.1881300

Keywords: plant proteins, ultrasound, pulsed electric fields, cold plasma, ohmic heating, non-thermal processing, food structure, sustainable food systems, protein functionality, multiscale, structural, modulation