The growing challenge of antibiotic resistance calls for alternative therapeutic strategies beyond conventional antibiotics. Antibacterial Photodynamic Therapy (aPDT), based on the light-induced generation of Reactive Oxygen Species (ROS), offers a non-specific approach that is detrimental to bacterial cells at multiple levels. Here, we explore the photocatalytic properties of nitrogen-doped titanium dioxide (N-TiO2) and compare them with those of the N-TiO2/AuNRs composite, a colloidal hybrid system formed by combining N-TiO2 with gold nanorods (AuNRs) upon excitation with a blue visible light centered at 420 nm. The synthesized materials underwent thorough in-depth morphological characterization to investigate their structural and functional features. Their photocatalytic activity was first evaluated using methyl orange as a model molecule, and subsequently on DNA and model lipid membranes, key components of bacterial life. Spectroscopic techniques were employed to highlight details on the oxidative damages at the molecular level. Scavenging experiments were performed to identify the reactive species involved in the photocatalytic process. The results highlight the potential of the composite N-TiO2/AuNRs as promising agent for photodynamic therapy, providing insights into how oxidative processes alter molecular targets.

Proteins and peptides hold immense promises for treating fatal and rare diseases. However, their complex structures and intrinsic instability pose significant challenges in drug development. The spray drying process offers a continuous and rapid method to stabilize these biomacromolecules, converting them from liquid to solid formulations. However, the spray drying process remains trial-and-error based, requiring extensive resources. This study employs machine learning (ML) algorithms to predict key properties of spray-dried protein and peptide powders, including yield, particle size, residual solvent content, solid states properties, and aggregation, with the aim to help accelerate formulation development and optimize process parameters. In total, 321 yield data points, 288 particle size data points, 357 residual solvent content data points, 205 solid states of the dry powder data points, and 305 aggregation data points were collected and described using various molecular descriptors for model building. Seven ML algorithms were tested to identify the best model to predict the key properties. Light Gradient Boosting Machine (LightGBM) exhibited the best performance for regression tasks, particularly for residual solvent content (mean absolute error = 0.841), and logistic regression excelling in predicting solid state characteristics and aggregation. Feature importance analysis identified protein, excipient, processing parameters, and environmental conditions as critical factors influencing various properties of spray-dried proteins/peptides formulation. The generalizability of the models was experimentally validated using alpha-lactalbumin formulations, the mean absolute error (MAE) in the models of yield, particle size, and residual solvent content were 0.755, 1.591, and 14.492 respectively, and the accuracies in predicting solid states of the dry powder and aggregation were 78.0% and 100.0% respectively. This study demonstrates the potential of machine learning to streamline the development of spray-dried protein formulations, providing a material- and time-saving solution as a reference.

When evaluating the proclivity of a monoclonal antibody (mAb) for non-specific self-assembly, it is common to restrict charge distribution analyses to the variable region of mAbs, leaving out possible contributions from the constant region to the observed sticky behavior. Here, the aim was to study the relationship between charge asymmetry over the entire mAb surface and self-assembly propensity. To do so, we selected three mAbs with decreasing levels of charge asymmetry and evaluated their ability to engage in attractive self-interaction as a function of mAb concentration and ionic strength, using small-angle X-ray scattering, dynamic light scattering and micro-flow imaging. We show that the mAbs with oppositely charged Fab and Fc domains are characterized by overall attractive protein-protein interactions in solution amounting to diverse sub-visible morphologies, which vary non-linearly with mAb concentration and ionic strength. As a proof of concept, we also report the absence of any of such assemblies for the mAb with like-charged Fab and Fc domains, resulting in an overall repulsive behavior in solution. Altogether, we show how to utilize charge distribution analyses of full-length mAbs to rationally develop formulations that prevent unwanted self-assembly.

Protein-based nanoparticles hold great promise for bioactive molecule delivery, but conventional fabrication routes are often complex and rely on surfactants or organic solvents. Here, we reported a crosslinker-free strategy to engineer α-lactalbumin (ALA) nanoparticles via Ca²⁺ mediated ionic bridging and electrospraying, enabling efficient encapsulation of acidic fibroblast growth factor (aFGF). These aFGF@ALA nanoparticles were subsequently integrated into electrospun poly(vinyl alcohol) (PVA) nanofiber matrices to construct hybrid dressings (aFGF@ALA NPs/PVA ENMs). The resulting nanostructures combine the structural support of nanofibers with the controlled release capacity of protein nanoparticles, yielding a synergistic platform for therapeutic delivery. In vitro, the hybrid dressings promoted fibroblasts and keratinocytes proliferation and migration with excellent cytocompatibility. In vivo, they accelerated burn wound repair by enhancing re-epithelialization, collagen I/III remodeling, and angiogenesis. This work introduces a facile and generalizable strategy for designing functional protein-based nanoparticles and demonstrates their integration with nanofiber scaffolds as a versatile platform for growth factor delivery and tissue regeneration.

At high concentrations, therapeutic monoclonal antibodies (mAbs) often display increased self-association, which may result in liquid-liquid phase separation (LLPS) or high opalescence alone. Due to a suspected link with aggregation, formulation strategies aim at preventing their occurrence. However, the molecular underpinnings of LLPS and opalescence remain unclear, complicating their forecasting at early stages. By combining light and X-ray scattering with microscopy and microfluidics, we report the phase behavior of a model mAb (mAb1). This is characterized by clustering and LLPS in a narrow NaCl range, above which it transitions into an opalescent state devoid of micron-sized assemblies, yet retaining similar self-interaction fingerprints. Using Monte Carlo simulations, we show that the behavior of mAb1 is controlled by a positive patch in its Fab domain, whose degree of charge screening determines solution fate. Finally, we show that LLPS and opalescence are decoupled from stirring-induced aggregation.

Limited aqueous solubility, often associated with low drug bioavailability, represents one of the greatest challenges in the clinical translation of numerous drugs currently on the market or in development. New and safer strategies are needed to unlock the full potential of poorly water-soluble compounds. Here, we report the ability of potato protein isolate (PPI), a valorized side-stream product of the potato starch industry, to arrest precipitation of curcumin in aqueous-based solutions in a concentration-dependent manner, leading to a manifold improvement in the solubility of curcumin as compared to pure water. High concentrations of PPI and presolubilization of curcumin (in pure ethanol) before mixing with the protein isolate were necessary to achieve complete solubilization of 0.1 mg/mL curcumin in 90% (v/v) water. The effect of PPI as a solubility enhancer was found to be limited by its propensity to undergo gel transition when exposed to higher concentrations of ethanol. Finally, PPI was non-toxic to HeLa cells, as demonstrated by an in vitro cell viability study, whereas dispersions of curcumin induced a cytotoxic effect. Our findings highlight the potential of less conventional plant protein isolates, such as PPI, as a strategy to improve the aqueous solubility of poorly water-soluble compounds.

Wound healing is a complex, multi-phase biological process that continues to pose significant challenges in biomedical engineering. As such, the development of innovative therapeutic strategies and sustainable, multifunctional biomaterials capable of accelerating tissue repair remains a top priority. In this study, we present a green, multiprotein nanofiber-based system fabricated via waterborne electrospinning using water as the sole solvent. The nanofibers, composed of α-lactalbumin (ALA) and soy protein isolate (SPI) with up to 90% (w/w) protein content, were formulated with minimal amounts of polyethylene oxide (PEO). Comprehensive analysis of nanofiber morphology, water stability, and mechanical properties revealed that the combination of ALA and SPI provided enhanced structural tunability and performance compared to single-protein systems. In a rat model of third-degree burns, the ALA/SPI/PEO nanofibers significantly improved wound healing outcomes relative to controls, as evidenced by accelerated re-epithelialization, increased collagen deposition, and enhanced angiogenesis; all this being attributable to the synergistic effects of the two protein components. By integrating sustainability, material design and therapeutic efficacy into a single platform, our multiprotein nanofiber system offers a compelling blueprint for the next generation of eco-conscious and clinically translatable biomaterials.

This study reports a solid formulation consisting of electrospun nanofibers based on potato protein isolate (PPI), which facilitates the incorporation and release of soluble and active curcumin. The processing of the poorly water-soluble model compound curcumin into a solid dispersion of electrospun nanofibers was achieved from a water-based solution with only minute amounts (10% (v/v)) of ethanol. This was possible due to the presence of a high concentration of PPI that arrests the precipitation of curcumin. The cospinning polymer polyethylene oxide (PEO) was added to facilitate the electrospinning process. The electrospun curcumin-loaded PPI/PEO nanofibers displayed a uniform morphology with a size distribution centered around 479 +/- 129 nm and contained a very small amount of residual solvent (recorded weight loss of 4.7 +/- 0.3%, as analyzed by thermogravimetric analysis). After dissolution of the curcumin-loaded nanofibers in pure water, 52 +/- 3% of curcumin was solubilized corresponding to a similar to 40x improvement in its reported aqueous apparent solubility as quantified by a calibration curve of curcumin made in 90% (v/v) water and in the presence of PPI and PEO. Finally, an antibacterial assay confirmed that curcumin was not only solubilized but also remained active upon processing as demonstrated by inhibition of the growth of Escherichia coli.

Liquid-liquid phase separation (LLPS) and high opalescence are two self-association phenomena commonly encountered in monoclonal antibody (mAb) formulations. Because of their impact on colloidal stability, they are commonly avoided, due to a suspected link with aggregation and reduced product shelf-life. However, the molecular underpinnings and interrelation between these phenomena remain unclear, complicating predictions of their occurrence. By combining light and X-ray scattering techniques with microscopy and advanced microfluidic setups, we here report the delicate phase behavior of a model mAb, named mAb1. This is characterized by rapid clustering and LLPS in a narrow NaCl range, above which it transitions into an opalescent state devoid of micron-sized assemblies, yet retaining a similar interaction fingerprint. Using Monte Carlo simulations, we report that the macroscopic solution state of mAb1 is controlled by a positive patch, whose degree of charge screening determines whether LLPS or opalescence will take place. Specifically, neutralization of this patch via counterion interactions diminishes intermolecular repulsion and favors the concerted action of weaker dipole-dipole/hydrophobic interactions, amounting to the creation of a new solution phase, via LLPS. Further NaCl addition distributes ions more uniformly across the surface, attenuating these attractive interactions, leading to the dismantling of droplets while preserving solution opalescence. Finally, we show that LLPS and opalescence are decoupled from stirring-induced aggregation, challenging an unequivocal relationship between these phenomena.

J-domain proteins ( JDPs) act as major regulators of the proteostasis network by driving the specificity of the Hsp70 machine. Their important functions are mediated by a low-complexity glycine-/phenylalanine-rich region (GF-linker) that links the folded J-domain with the substrate binding domain. Recently, we and others have shown that in an autoinhibited JDP state, an α-helix formed within the GF-linker blocks the Hsp70 binding site on the J-domain. However, the role of the disordered GF-linker in autoinhibition and how the latter is released, are still not understood. Here, using autoinhibited DNAJB1 and DNAJB6 constructs, we show that in combination with the J-domain, the GF-linker creates a hydrophobic, partially collapsed cluster that shows a remarkable degree of long-range structural communication, disruption of which can lead to destabilisation of autoinhibition. Apart from this crucial intramolecular role, we reveal that the GF-linker can also be recognised by the substrate-binding domain of Hsp70 and dictate the lifetime of the entire JDP–Hsp70 complex. Strikingly, the GF-linkers of DNAJB1 and DNAJB6 display distinct structural properties that lead to different Hsp70 binding kinetics, showing that the behaviour of the GF-linker can vary dramatically even within the same class of JDPs.

Protein aggregation is associated with a spectrum of neurodegenerative diseases. Although many small ligands have been found to modulate or inhibit protein aggregation, their molecular mechanisms remain unclear. One reason for this is the inherent heterogeneity of protein aggregation pathways with different kinetics that result in the coexistence of multiple structures, for example, protein spherulites and fibrils, challenging the analysis of protein-ligand interactions. To address this issue, we evaluated the roles of betaine and proline in insulin aggregation. We employed our recently developed super-resolution microscopy real-time kinetics via binding and photobleaching localization microscopy (REPLOM) to directly observe the formation and morphological evolution of individual insulin aggregates in real time, with or without betaine/proline. Utilizing our machine learning approach, we monitor the effect of betaine and proline on the aggregation pathways and extract the growth kinetics of each individual aggregate type. Our results show that a high concentration of betaine or proline modulates the heterogeneity of the final aggregates, leading to the formation of smaller aggregates in a mixture with spherulites. The fraction of small aggregates increases with betaine/proline concentration, highlighting the heterogeneity of protein aggregation, and our toolbox can unravel the effects of small molecule ligands on individual protein aggregation pathways and the resulting aggregate types and abundances.

Liquid-liquid phase separation (LLPS) and high opalescence resulting from attractive interactions in the absence of large-scale self-association are commonly encountered in liquid monoclonal antibody (mAb) formulations and often associated with aggregation. Here, we evaluated the formulation dependent self-assembly behavior of a therapeutic mAb (termed mAb1) and provide a link to physical stability after applied stress. Using a combinatorial microdroplet platform to delineate the phase space of mAb1, we highlight its unusual solution properties. Specifically, mAb1 undergoes LLPS in a narrow window of low NaCl concentrations, in a formulation relevant design space, at room temperature (RT), without crowding agents and more than one pH unit lower than its isoelectric point. Using time-resolved small-angle X-ray scattering, we measured the attractive interactions underpinning LLPS to be on the sub second timescale and leading to the formation of small clusters, which altogether could be rapidly inhibited by L-arginine HCl. The elevated opalescence of mAb1 solutions extended beyond the LLPS regime, where a highly similar interaction fingerprint to the LLPS condition was observed but no microscopic association was detected. This suggests that LLPS and opalescence differ in the distribution of interactions within their networks, dictating their ability to form long-range assemblies (phase-separated droplets) or not. Finally, our physical stability assessment highlights that mAb1 LLPS and molecular assemblies giving rise to high opalescence both resist stirring stress and that the observed aggregation is uncorrelated with ionic strength at our selected timescale. This observation challenges an unequivocal relationship between LLPS/opalescence and reduced physical stability of mAb solutions. Significance Statement The development of formulations maximizing the stability of therapeutic monoclonal antibodies is a crucial part of their successful development. However, their proclivity for self-association often results in increased opalescence, which has been shown to indicate a propensity for liquid-liquid phase separation (LLPS), involving the de-mixing of solutions into two phases with different concentrations of solute. Both phenomena are commonly tied to a high propensity for aggregation. However, given the paucity of experimental studies, evidence for a general correlation between these solution properties and aggregation remains scant. In this work, we decouple stress-induced aggregation from LLPS and high opalescence without large-scale self-association through the fundamental study of these phenomena and solution integrity after applied stress. ### Competing Interest Statement Competing Interest Statement: N.L, M.N, J.S.M and M.G.J are employees of Novo Nordisk A/S.

Selenium nanoparticles (SeNPs) can be synthesised via physical, chemical, or biological methods. While physical and chemical approaches offer benefits, they often involve high costs, toxic reagents, and reduced biocompatibility due to stabilizers. Biological synthesis provides a greener, more biocompatible alternative, but faces challenges such as variability in biological sources, slow synthesis, low yields, and scalability issues due to pH sensitivity, toxic precursors, and costly equipment. To address these limitations, we propose a sustainable workflow for SeNPs production using the selenite (SeO32-)-reducing amino acid L-cysteine (Cys) and the eco- and bio-compatible surfactant sodium oleate (NaOl). We report that Cys not only reduces selenite but also induces phase transitions in NaOl, forming a confined environment of vesicles and lamellar structures that facilitate controlled SeNPs nucleation and growth. Through a detailed phase diagram analysis, we elucidate the critical Cys and NaOl concentration regimes and pH conditions thresholds that drive surfactant self-assembly, providing valuable insights into the optimized colloidal environment for SeNPs synthesis. We also report that these confined structures support SeNPs isotropic growth and long-term stability. This method provides a scalable and eco-friendly approach for producing biocompatible SeNPs, addressing some of the challenges of conventional methods.

Ion-protein interactions regulate biological processes and are the basis of key strategies of modulating protein phase diagrams and stability in drug development. Here, we report the mechanisms by which H-bonds and electrostatic interactions in ion-protein systems determine phase separation and amyloid formation. Using microscopy, small-angle X-ray scattering, circular dichroism and atomistic molecular dynamics (MD) simulations, we found that anions specifically interacting with insulin induced phase separation by neutralising the protein charge and forming H-bond bridges between insulin molecules. The same interaction was responsible for an enhanced insulin conformational stability and resistance to oligomerisation. Under aggregation conditions, the anion-protein interaction translated into the activation of a coalescence process, leading to amyloid-like microparticles. This reaction is alternative to conformationally-driven pathways, giving rise to elongated amyloid-like fibrils and occurs in the absence of preferential ion-protein binding. Our findings depict a unifying scenario in which common interactions dictated both phase separation at low temperatures and the occurrence of pronounced heterogeneity in the amyloid morphology at high temperatures, similar to what has previously been reported for protein crystal growth.

Our previous work shows that beta-lactoglobulin-stabilized amorphous solid dispersion (ASD) loaded with 70 % indomethacin remains stable for more than 12 months. The stability is probably due to hydrogen bond networks spread throughout the ASD, facilitated by the indomethacin which has both hydrogen donors and acceptors. To investigate the stabilization mechanisms further, here we tested five other drug molecules, including two without any hydrogen bond donors. A combination of experimental techniques (differential scanning calorimetry, X-ray power diffraction) and molecular dynamics simulations was used to find the maximum drug loadings for ASDs with furosemide, griseofulvin, ibuprofen, ketoconazole and rifaximin. This approach revealed the underlying stabilization factors and the capacity of computer simulations to predict ASD stability. We searched the ASD models for crystalline patterns, and analyzed diffusivity of the drug molecules and hydrogen bond formation. ASDs loaded with rifaximin and ketoconazole remained stable for at least 12 months, even at 90 % drug loading, whereas stable drug loadings for furosemide, griseofulvin and ibuprofen were at a maximum of 70, 50 and 40 %, respectively. Steric confinement and hydrogen bonding to the proteins were the most important stabilization mechanisms at low drug loadings (<= 40 %). Inter-drug hydrogen bond networks (including those with induced donors), ionic interactions, and a high Tg of the drug molecule were additional factors stabilizing the ASDs at drug loading greater than 40 %.

Protein self-assemblies in the form of ordered supramolecular structures such as particulates hold great potential as new biomaterials. However, research in this field is rarely conducted under physiologically relevant conditions but such studies are crucially needed to unravel the potential use of particulates and other amyloid structures in health sciences. In this study, particulates of α-lactalbumin (ALA) were prepared at different stages of maturation by thermal incubation. Disassembly of particulates in isotonic buffer, neutral pH and at 37 °C was investigated by simultaneously measuring Thioflavin T fluorescence intensity and light scattering. Freshly formed particulates quickly disassembled and displayed complete release of soluble ALA within 1 h. Mature particulates displayed slower disassembly kinetics with incomplete release of ALA within 1 h. The biocompatibility of particulates at different maturation stages to epithelial lung and fibroblast cells was assessed in vitro. Good cell compatibility was observed in the presence of the particulates and their released species. Our findings display protein particulates as biodegradable and highly tunable particles, promoting them as good candidates for drug delivery purposes.

Proteins acting as carriers in amorphous solid dispersions (ASDs) demonstrate a notable sensitivity to the spray drying process, potentially leading to changes in their conformation. The main aim of this study was to investigate the dissolution performance of ASDs based on proteins with different content of secondary structures, specifically β-sheet and α-helix structures. We prepared β-sheet-rich and α-helix-rich β-lactoglobulin (BLG), along with corresponding ASDs containing 10 wt% and 30 wt% drug loadings, through spray drying using celecoxib as the model drug. Circular dichroism and Fourier Transform Infrared Spectroscopy results revealed that even though changes in secondary structure were obtained in the spray-dried powders, the BLGs exhibited reversibility upon re-dissolving in phosphate buffer with varying pH levels. Both β-sheet-rich BLG and α-helix-rich BLG exhibited enhanced dissolution rates and higher solubility in the media with pH values far from the isoelectric point (pI) of BLG (pH 2, 7, 8, and 9) compared to the pH closer to the pI (pH 3, 4, 5, and 6). Notably, the release rate and solubility of the drug and BLG from both types of BLG-based ASDs at 10 wt% drug loading were largely dependent on the solubility of pure SD-BLGs. α-helix-rich BLG-ASDs consistently exhibited equivalent or superior performance to β-sheet-rich BLG-ASDs in terms of drug release rate and solubility, regardless of drug loading. Moreover, both types of BLG-based ASDs at 10 wt% drug loading exhibited faster release rates and higher solubility, for both the drug and BLG, compared to the ASDs at 30 wt% drug loading in pHs 2, 7, and 9 media.

A large part of the global population, including both immunocompromised and healthy individuals, suffers from fungal infections. The majority of current drug candidates for treating fungal infections exhibit poor water solubility, which hampers their permeability through biological barriers and limits their bioavailability. Here, we fabricated pea protein isolate (PPI)/poly(ethylene oxide) nanofibers (PPI/PEO nanofibers), with a high protein content [65% (w/w)] by waterborne electrospinning as an eco-friendly drug delivery system. X-ray diffraction results demonstrated that the solid-state properties of the individual components (PPI and PEO) were retained in the PPI/PEO nanofibers. We then encapsulated the poorly water-soluble drug, clotrimazole (CTZ), in the nanofibers (PPI/PEO/CTZ nanofibers), without heat treatment and/or use of an organic solvent or surfactant to presolubilize CTZ, generating an antifungal delivery system for topical administration. An in vitro study demonstrated that CTZ was successfully loaded in and released from the nanofibers. Additionally, the nanofibers were not toxic to HeLa cells. Finally, based on an antifungal disc agar diffusion study, CTZ-loaded nanofibers were shown to be effective against Candida albicans. The overall results demonstrate the potential of PPI-based nanofibers as a green platform for the generation of CTZ-loaded efficient drug delivery systems for antifungal treatment.

Hypothesis: Low-frequency Raman (LFR) spectroscopy is proposed as a novel non-destructive methodology to probe pH-related phase transitions in self-assembled lipid particles. In this case, dispersed lipid mesophases were composed of ionisable oleic acid (OA) or nicergoline (NG) in monoolein (MO). The sensitivity of LFR spectroscopy to low-energy intermolecular vibrations was hypothesised to be due to structural transformation in ionisable dispersed mesophases upon changes in pH. Method/experiment: Phase transitions of dispersed mesophases of MO mixed with OA or NG were induced by varying the pH of the aqueous buffer. The structural transformations were studied using LFR spectroscopy, recording the corresponding changes in the vibrational density of states (VDOS) upon changes in pH and analysed using principal component analysis (PCA). The results were correlated with structural transitions observed in simultaneous small-angle X-ray scattering (SAXS) measurements. Findings: The intensity of the VDOS signal of MO + OA mesophases scaled with phase-specific transformations, such as from the bi-continuous cubic Im3m phase (V2) or lamellar-based vesicles to the reversed hexagonal p6m phase (H2). For NG subtle changes in the lattice parameter of the V2 phase of NG + MO mesophases coincided with the apparent dissociation constant (pKaapp) of NG, however, slight variations between the pK a app of NG determined by equilibrated samples analysed using SAXS and non-equilibrated samples analysed using LFR suggest structural hysteresis upon changes in the protonation state of NG. This approach offers an efficient method for studying the phase behaviour of lipid systems under varying pH and potentially other conditions such as temperature.

Amorphous solid dispersions (ASDs) using proteins as carriers have emerged as a promising strategy for stabilizing amorphous drug molecules. Proteins possess diverse three-dimensional structures that significantly influence their own properties and may also impact the properties of ASDs. We prepared beta-lactoglobulin (BLG) with different contents of beta-sheet and alpha-helical secondary structures by initially dissolving BLG in different mixed solvents, containing different ratios of water, methanol/ethanol, and acetic acid, followed by spray drying of the solutions. Our findings revealed that an increase in alpha-helical content resulted in a decrease in the glass transition temperature (Tg) of the protein. Subsequently, we utilized the corresponding mixed solvents to dissolve both BLG and the model drug celecoxib (CEL), allowing the preparation of ASDs containing either beta-sheet-rich or alpha-helix/ random coil-rich BLG. Using spray drying, we successfully developed BLG-based ASDs with drug loadings ranging from 10 wt% to 90 wt%. At drug loadings below 40 wt%, samples prepared using both methods exhibited single-phase ASDs. However, heterogeneous systems formed when the drug loading exceeded 40 wt%. At higher drug loadings, physical stability assessments demonstrated that the alpha-helix/random coil-rich BLG structure exerted a more pronounced stabilizing effect on the drug-rich phase compared to the beta-sheet-rich BLG. Overall, our results highlight the importance of considering protein secondary structure in the design of ASDs.