Introduction

Plastics and polymers are indispensable; they provide protection, convenience, comfort, and play critical roles in enabling sustainable energy technologies. More than 400 million tons of plastics is produced yearly, nearly all of which comes from fossil fuels.[1] Most plastics do not degrade at reasonable time scales except for a relatively small fraction of industrially compostable polyesters and biopolymers.[2] Concerns over the environmental effects of fossil fuel extraction and the accumulation of plastics in the environment have motivated the search for alternative carbon sources for materials manufacture, with significant efforts in developing biomass-derived monomers or platform chemicals. Another approach is to utilize biopolymers as the polymeric material, and natural proteins are promising in this regard as they are renewable and biodegradable.

Proteins are polyamides with amino acids linked in sequences that are encoded in the genome. With 20 naturally occurring proteinogenic amino acids in eukaryotes and the ability of cells to add post-translational modifications, proteins differ from many other biopolymers in their monomer diversity. In addition, proteins are characterized by their ability to form secondary structures through hydrogen bonds between the carbonyl oxygen and amino hydrogen in the backbone. Hydrophobic and electrostatic interactions also play important roles in driving protein folding and stabilization of complex structures, enabling proteins to perform crucial biological roles ranging from catalysis in enzymes to mechanical reinforcement in structural proteins. A few examples of non-ribosomally synthesized proteins can also be found in nature, including cyanophycin in cyanobacteria, ε-polysine in Streptomyces albulus, and poly-γ-glutamate in Bacillus strains and other species.[3]

Proteins are attractive as feedstock for materials manufacture due to their renewability, biodegradability, and capability to form cohesive matrices in a range of glassy to rubbery materials. Active research on protein-based films, fibers, and adhesives date back to the early twentieth century,[4] and a few food protein-based fibers (e.g., Fibrolane, Lannital, Vicara) were even commercialized for a time, until they were ultimately replaced by petrochemical plastics.[5] Today, protein-based materials are gaining renewed interest as alternatives to fossil fuel-based materials and as value-added byproducts from agriculture and integrated biorefineries. The biodegradability of proteins makes them well suited as single-use packaging materials, and the good oxygen barrier properties of proteins can be advantageous in food packaging materials or coatings. Edible proteins may also be leveraged in specific applications such as edible films and water-soluble pouches for delivering pre-measured quantities. Proteins can have high mechanical stiffness due to extensive intermolecular interactions, which enable them to function as reinforcements in composite materials. Besides films and fibers, a few examples of protein foams and aerogels have also been demonstrated.[6] Despite broad potential applications, issues in material performance, processing issues, and cost hinder commercialization of protein-based materials.

Unmodified proteins lack the mechanical robustness, water resistance, and the processability that many synthetic polymers exhibit.[7] Additionally, protein feedstocks are often heterogeneous, containing mixtures of proteins and non-protein biomolecules that are expensive to separate.[8] These challenges continue to inspire research on novel material design and fabrication methods, which include various physical blending and chemical modification strategies.[7] As research fields at the interface of protein engineering and polymer science grow, tools developed for synthetic protein design and bioconjugation chemistries can also advance efforts in utilizing nature’s building blocks in sustainable materials.

In this prospective, we will review non-food and non-therapeutic uses of proteins as commodity and engineering plastics feedstock, with a particular focus on material design strategies and structure–mechanical property relationships. We will first introduce common protein physicochemical characteristics that are relevant to fabrication of protein-based materials. Current strategies for improving processability and mechanical properties through physical blending and chemical modifications will be summarized, along with an examination of structure–property relationships. Finally, we will summarize properties and production potential for a select number of ‘industrial’ protein sources from agriculture, animal farming, and aquaculture.

Protein physicochemical characteristics

While proteins have a broad spectrum of compositions and structures, certain solution and solid behaviors are broadly observed due to the general composition of proteins and the presence of hydrogen bonding, electrostatic, and hydrophobic interactions. Protein solubilities are affected by primary sequences and 3D conformations. Native globular proteins are generally soluble in aqueous solutions due to the burial of non-polar hydrophobic groups and the presence of charged polar groups on the surface, and their solubilities are higher at pH’s far from the isoelectric point as the electrostatic repulsion prevents aggregation. At low ionic strengths, solubilities increase with salt concentrations due to screening of attractive electrostatic charges. At high ionic strengths, salting out is observed, where ions compete for associations with water and decrease protein solubilities. Acid, base, heat, chaotropic agents, and surfactants can disrupt intramolecular associations and unfold proteins to different degrees. The exposure of hydrophobic groups leads to solubility losses after denaturation, which may be compounded by the formation of intermolecular disulfide bonds. Protein solubilities can therefore be significantly affected by processing or extraction conditions; for example, spray drying greatly reduces protein solubilities when compared to gentler recovery methods.[9] Native fibrous proteins such as keratin, which are characterized by their extended, parallel structures, are generally insoluble due to high levels of covalent disulfide or physical crosslinking. Chemical treatment or protein hydrolysis may be employed to generate soluble proteins and polypeptides.[10,11]

Globular proteins are known to aggregate and form a variety of structures when denatured in solution, subsequently gelling when percolating networks are formed. Depending on the balance of electrostatic repulsion and attraction upon unfolding, proteins may form random aggregates, branched structures, flexible strands, or fibrils.[12] The impact of processing conditions on protein gel formation, structure, and kinetics has been extensively reported due to their importance to the food industry.[13] More recently, the formation of nanofibrils from globular food proteins has attracted much attention.[14] They resemble amyloid fibrils responsible for many neurodegenerative diseases and are postulated to be the most thermodynamically stable state of proteins.[15] The most common fibrillation condition involves heat treatment at low pH and low ionic strength, but several other conditions that utilize agitation, reducing conditions, or chaotropic agents have also been demonstrated.[16] Diverse fibril morphologies, including ribbons, nanotubes, and various crystal polymorphs, arise from the assembly of β-sheet building blocks into protofilaments with cross-β structures, which then twist into multistrand fibrils with diameters of 1–10 nm and lengths of 1–10 μm. In solution, the protein fibrils can form gels at lower concentrations due to their high aspect ratios when compared to non-specific aggregates. When dried, the densely hydrogen bonded fibrils are reported to have some of the highest Young’s moduli in proteinaceous materials.[14]

Dehydrated proteins are glassy and generally very brittle at room temperature. Extensive inter and intramolecular secondary forces greatly hinder protein chain mobility, resulting in high glass transition temperatures (Tg). Protein Tg may not always be detected due to the weak transition, but studies on different proteins report values at around 150–220°C.[17] The temperatures at which proteins soften are also high and often close to their degradation temperatures, which start at around 220°C.[18] Glassy, dried globular proteins have very low mechanical robustness, necessitating modifications to transform them into useful materials, most typically by plasticization, covalent attachment to other polymers, or covalent crosslinking. There are notable exceptions where natural fibrous proteins such as silk and keratin are highly robust and provide structural support in biological components, but regenerated fibers and films prepared from these proteins are also often brittle and require other modifications or hydration.[19] Besides tuning physical properties of proteinaceous materials, plasticization and polymer conjugation are also important strategies for enabling material processing.

Proteins have large numbers of reactive functional groups that provide handles for modification, crosslinking, and polymer attachment. Commonly targeted native reactive groups include amines from lysines, carboxylic acids from aspartic and glutamic acids, and thiols from cysteines. Glutamine and asparagine may be converted into glutamic acid or aspartic acid by deamidation, which is relevant for proteins extracted or processed in acidic or alkaline conditions.[20] Extensive lists of protein reactions can be found in a multitude of reviews, and some examples are shown in Fig. 1.[21,22,23] Reactivities of the protein functional groups are influenced by solvent accessibility, protein environment effects, and reaction conditions. For reactions involving abundant, surface-exposed residues that have similar reactivities, broad distributions in the number and location of modification sites are typically observed. Advances in protein bioconjugation have expanded strategies that enable more site-selective modifications, which include chemistries that target amino acids that are naturally less abundant, or reactions that leverage unique reactivities of functional groups at the termini.

Figure 1
figure 1

Chemical modification of native amino acids.

Full size image

Transforming proteins into materials

Proteins are transformed into films, fibers, or other plastic forms through several solution and melt processing strategies. Solution casting is one of the most common methods for converting protein solutions into films at the laboratory scale. Small molecule or polymeric additives are typically co-dissolved in protein solutions, followed by solvent evaporation [Fig. 2(a)]. Fiber wet spinning also involves first solubilizing proteins into dope solutions, followed by extrusion into poor solvents [Fig. 2(b)]. The protein filaments that precipitate in the coagulation bath are then typically drawn and elongated. During solution processing, good solvents swell proteins and enable reconfiguration of native protein structures. This increases chain entanglements, which are crucial for the formation of cohesive materials.[24] Compared to solvent casting, thermoplastic processing methods such as molding and extrusion are generally more scalable and industrially relevant. Here, proteins are plasticized and heated above their softening or glass transition temperatures (Tg), where they become malleable and can be fused into cohesive materials as chain interdiffusion occurs [Fig. 2(c)]. Plasticizers, which may be solvents or non-volatile small molecules, play important roles in enabling stable extrusion by reducing protein Tg and tuning melt viscosities. More recently, additive manufacturing has been applied to prepare protein-based materials, utilizing techniques such as vat polymerization and material extrusion which involve similar processing considerations as other solution and melt-based strategies.[25,26]

Figure 2
figure 2

Copyright 1998 John Wiley and Sons.

Processing strategies for protein-based materials. (a) Solution processing of films by solvent casting. (b) Wet spinning of fibers by extrusion into a coagulation bath. (c) Schematic representation of thermoplastic processing of proteins, where proteins are plasticized and heated above their softening temperatures for shaping. Adapted with permission from Ref. 27.

Full size image

Plasticized proteins as polymer matrices

A vast majority of efforts that transform proteins into continuous matrices in films and fibers rely on using external plasticizers or additives that disrupt the intermolecular hydrogen bonding of proteins, allowing them to be more flexible and malleable. Plasticizers modulate thermal and mechanical properties through a few possible mechanisms: they may increase lubricity and allow the polymer molecules to slide over each other by reducing intermolecular friction, they may disrupt the three-dimensional physical network of polymers by breaking polymer–polymer interactions, or they may increase the free volume and thereby the motion of polymers.[28] Plasticizer incorporation into protein-based materials is typically achieved by directly compounding proteins and plasticizer additives above the mixture Tg,[29] or by co-dispersing both components in a suitable solvent.[30] Generally, plasticizers decrease Tg’s and viscosities to allow thermal processability. At constant temperatures, increasing plasticizer content decreases stiffness and increases elongation-at-break and toughness. Material performance requirements and processing needs are typically the main factors for consideration when selecting or designing an efficient plasticizer.

Plasticizer molecular structure and material properties

Plasticizer efficiency is broadly defined, and usually relates to the relative amount of plasticizer required to achieve a certain desirable property. There are no universally effective plasticizers for all protein-based materials, and a plasticizer suited for achieving one material property may not be equally as effective for achieving others. Nevertheless, secondary bonding interactions and compatibility between plasticizer and polymer are known to be crucial for plasticization. Polar low molecular weight plasticizers are commonly selected to alleviate the challenges stemming from the extensive intermolecular hydrogen bonding in proteins, including the excessive brittleness and limited thermal processability. By mass, water is likely the most powerful plasticizer due to its low Tg, small size, and ability to hydrogen bond to and solvate protein molecules. Protein Tg is therefore highly sensitive to moisture, and mechanical properties can vary greatly with humidity changes.[31] The most frequently used non-volatile plasticizer for proteins is glycerol, which has three hydroxyl groups capable of participating in hydrogen bonding and high compatibility with many proteins.[30] It is also hygroscopic, and the large amounts of water absorbed at ambient conditions lead to further plasticization. Generally, a minimum glycerol loading of ~ 20 wt% is required for proteins to be thermally processable via extrusion or compression molding.[32] As more glycerol is incorporated, proteins transition from glassy to rubbery. Higher concentration of glycerol leads to decreased mechanical strength and increased elongation-at-break until a saturation value (Fig. 3). On the other hand, less polar molecules such as propylene glycol and waxes have been shown to be relatively ineffective at increasing flexibility (Fig. 3).[33] Other polar plasticizers explored and reviewed in several works include polyols, amines, mono-, di-, and polysaccharides, and fatty acids.[7,29,34,35]

Figure 3
figure 3

Ultimate tensile strength and elongation-at-break for whey protein plasticized with glycerol or beeswax.[33,36,37,38] Mass percentages of plasticizer incorporated is noted in the legend. For whey protein, glycerol addition generally increases material flexibility, and its effects are much larger than that of beeswax. Mechanical property trends influenced by heat denaturation and covalent crosslinking are also noted. Each data point is numbered according to the corresponding reference and details are provided in Table S1. All materials tested were equilibrated at 23–25°C and 50% RH.

Full size image

In addition to hydrogen bonding, hydrophobic interactions also contribute significantly to protein intra- and intermolecular associations. Synergistic plasticizing effects may be observed when amphiphilic and polar molecules are combined in plasticizer blends, potentially due to these molecules binding to different parts of proteins.[35,39] For example, solution-cast zein films were observed to have the lowest Tg and highest elongation-at-break when an optimal combination of oleic acid and glycerol was used.[40] When used alone, amphiphilic molecules generally have low plasticizing efficiencies compared to hydrophilic molecules, but some have been shown to improve properties of hydrophobic proteins when incorporated at concentrations below the macrophase separation limit.[35] For the relatively hydrophobic corn proteins, addition of octanoic acid leads to cohesive matrices, while water and glycerol at the same mass ratios lead to crumbly materials.[29] The inverse is observed for the more hydrophilic wheat gluten, where octanoic acid has the smallest effect on Tg when compared to water, glycerol, butanediol, and lactic acid at the same mole fractions.[41]

Several studies have suggested that plasticizer properties such as molecular structure, number of theoretical hydrogen bonds, and molar mass have an impact on plasticizer efficiency. Bulky molecules, such as sucrose, are hypothesized to have reduced plasticizer efficiency due to their lower mobility and limited ability to interact with proteins.[30] On the other hand, small plasticizers like glycerol have been hypothesized to be efficient due to the ease of insertion into three-dimensional networks. However, comparisons done on a mole plasticizer per mole protein basis can show greater plasticizing effects for large molecules when each molecule has more functional groups that can interact with proteins.[30] There are also studies that show the contrary; for example, ethylene glycol was shown to be less efficient than di- and tri-ethylene glycol at increasing extensibility of wheat gluten films.[42] Plasticizer with different hygroscopicity will also lead to varying water contents in the materials at equilibrium. Properly accounting for moisture effects may therefore be necessary, given the powerful plasticizing capability of water.

Plasticizer selection often involves weighing different material property tradeoffs to reach a compromise. Since proteins are hydrophilic and poor barriers to water vapor, hydrophobic plasticizers such as fatty acids and waxes have been introduced to increase their performance in packaging applications where low water vapor and oxygen permeabilities are desired. However, these additives provide limited improvements in material flexibility (Fig. 3), and can macrophase separate and exude at high loadings.[33,35] Another desirable plasticizer characteristic is permanence, which limits the loss of plasticizer and drifts in material properties. Larger plasticizer molecules have a greater permanence and stability but may not be as efficient at increasing chain mobility. Sustainability and regulatory concerns can also influence plasticizer selection and are particularly important in food-related applications. Mechanistic understanding of plasticizer–protein interactions (Fig. 4) can therefore help tailor plasticizers or additives and streamline the material design process, as broad ranges of property targets and variabilities in plasticizer effectiveness across proteins are typically encountered.

Figure 4
figure 4

Factors that influence plasticizer effectiveness and strategies for tuning material properties.

Full size image

Protein secondary structures and material properties

Material mechanical properties are correlated with protein conformations and β-sheet hydrogen bonding strength, which can be influenced by protein–plasticizer interactions. Plasticizers increase chain mobility and enable protein restructuring. At very low concentrations, glycerol is observed to increase ordered α-helix and β-sheet structures in many proteins. The trend reverses at high plasticizer concentrations as protein–protein interactions are disrupted.[43] With increasing plasticization, changes in both protein conformations and moduli are more pronounced for polar proteins with low amounts of cysteines. This is attributed to the increase in plasticizer binding sites provided by polar side groups, and the lack of conformation stability by having fewer disulfide bonds.[43] Some studies have also showed that plasticizers can vary intermolecular hydrogen bonding strength in proteins, and the agreement between trends in peptide hydrogen bonding strength and moduli suggests strong connections between plasticizers, protein conformation, and macroscopic properties.[44]

Processing aids in the form of reducing agents and denaturants may also be added to modulate protein unfolding during thermal processing. While not efficient plasticizers on their own, denaturants like sodium dodecyl sulfate (SDS) and urea have been utilized to improve extrudability by slowing down protein aggregation and enabling viscosity control.[45,46] At excessively high concentrations, these additives may function as solid fillers. Reducing agents are used to reduce melt viscosities of proteins with cysteines. During extrusion of keratin, a melt viscosity minimum is achieved by an optimal amount of sodium sulfite. Smaller amounts of the additive lead to insufficient reduction of intermolecular disulfide bonds, while excessive sodium sulfite may cleave too many intramolecular disulfide bonds, allowing proteins to align and crystallize and viscosities to increase.[47]

Tuning properties of plasticized films

Material properties may be tuned by inducing the formation of intermolecular β-sheets. In solution-cast films, drying induces formation of these structures, which serve as physical crosslinking points and stabilize networks.[44,48] They may also increase as a result of denaturation, where native secondary structures are disrupted, and previously buried hydrophobic and reactive functional groups are exposed. Elevated temperatures, mechanical energy input, and strong acids and alkali are therefore handles for increasing intermolecular association and chain entanglements. When plasticized proteins are heat treated, significant increases in both tensile strengths and elongation-at-break may be observed (Fig. 3).[38]

Introducing chemical crosslinks into plasticized protein films is a common approach for increasing their mechanical and chemical stabilities. One of the crosslinking methods includes heat treatment, which not only induces physical crosslinking through protein conformation changes, but also formation of new covalent bonds from native reactive functional groups. Disulfide bond formation or exchanges are commonly observed, and they contribute to the thermosetting and gelation behavior of heat treated or extruded plasticized proteins that contain cysteine residues. In cysteine-poor proteins or insufficiently plasticized materials, the contribution of disulfide bonds on mechanical properties may be small relative to the extensive non-covalent forces.[49] Other non-reducible covalent crosslinks may also arise from heating. Possible reactions include ester, dityrosine, isopeptide, and lysinoalanine formation, but these reaction products are generally difficult to quantify.[50] To a lesser extent, protein-based materials can also be crosslinked via ionizing and UV radiation.[34]

Small molecule or polymeric crosslinkers may also be added for controlling crosslink densities, and amine-reactive reagents are popular due to the natural abundance of lysine. Formaldehyde, as an example, has been employed as a curing agent since the early days of leather tanning and protein resin and fiber manufacturing.[5] Other aldehyde crosslinkers of varying sizes and sources include glyoxal, glutaraldehyde, and polysaccharides with di- and multifunctional aldehydes.[27,34] Crosslinkers sourced from biomass, such as tannic acid, genipin, and enzymes such as transglutaminase are also utilized. Crosslinking glycerol plasticized films by heat treatment, small molecule crosslinker, and irradiation have been observed to increase modulus and tensile strength to varying extents (Fig. 3), from negligible changes to up to ~ 5-fold increases. Large stiffening effects, however, come at the expense of significant decreases in extensibility.[34]

Self-assembly of proteins is emerging as a powerful tool that enables material properties to be tailored through nanostructure formation. Recent work has demonstrated that high aspect ratio nanofibrils can be fabricated from a wide range of protein sources.[16] When incorporated into plasticized films, stiff nanocrystalline protein fibrils enhance mechanical performance by serving as reinforcements.[51,52] They can also adopt nematic order and align parallel to the film plane during solution casting.[53] Figure 5 shows a strategy leveraging the self-assembly of proteins to overcome limitations of poorly soluble proteins while improving mechanical performance of soy protein isolate (SPI) films.[51] Fibril formation was performed at solution conditions that unfold and partially hydrolyze proteins, which led to solubilization and eventually formation of fine-stranded aggregates. Owing to the β-sheet nanocrystals, self-assembled films exhibited significantly higher Young’s moduli and tensile strengths relative to unstructured films.

Figure 5
figure 5

Copyright 2021 Springer Nature.

(a) Schematic representation of the self-assembly of soy protein isolate (SPI). Heat, sonication, and an acetic acid-based solution is used to solubilize SPI, followed by controlled cooling of the solution to form β-sheet rich fibrillar aggregates. Self-assembled films containing glycerol as the plasticizer were obtained by drying. (b) TEM image of SPI fibrillar aggregates, scale bar is 500 nm. (c) TEM image of dried films showing β-sheet nanocrystals, scale bars are 5 nm for the main image and 2 nm for the inset. (d) Representative stress–strain curves for SPI-glycerol films. The self-assembled films with fibrillar proteins exhibit superior mechanical performance when compared to the non-structured films prepared with unmodified SPI. Reproduced with permission from Ref. 51.

Full size image

Protein-polymer blends

Blending polymers or fillers with proteins provides opportunities to combine advantages of multiple components in composite materials. In plasticized films where proteins function as the flexible continuous matrix, changing plasticizer content allows flexibility to be tuned at the cost of stiffness. Incorporation of a second stiffer component is therefore attractive, as reinforcing polymers, fibers, or inorganic fillers can compensate for the loss of mechanical strength. Fillers may also be selected to impart properties such as stability, conductivity, and fire retardancy. Fully bio-based composites have been demonstrated with lignin, chitin, and other polysaccharides as fillers in molded protein plastic parts, fibers, and 3D printing filaments.[26,54] Alternatively, instead of serving as the matrix, proteins may be utilized as reinforcing fillers since they are stiff in the absence of plasticizers. Naturally occurring proteinaceous fibers such as feathers are attractive for these applications due to their mechanical strength and low densities,[55] and various other proteins have been incorporated into synthetic polymer matrices such as polyisoprene, polyethylene, and polyvinyl alcohol. High aspect ratio protein nanofibrils are also attractive in the design of high-performance composites, as they have been reported to demonstrate higher reinforcing capability than non-fibrillated protein aggregates.[56]

In these protein-based blends, compatibility between components is necessary for mechanical performance. Incompatible polymers will macrophase separate, and a lack of adhesion between phases or filler/matrix leads to inefficient stress transfer and low mechanical robustness. Proteins are generally more compatible with other biopolymers with polar functional groups than non-polar synthetic polymers. Charged biopolymers, such as chitosan, are also able to electrostatically complex with proteins.[57] On the other hand, many commercially relevant synthetic polymers are hydrophobic and chemically incompatible with proteins. This poses additional mixing challenges, as solution blending is restricted due to the lack of common good solvents, and melt blending results in poor dispersions due to aggregation. Compatibilization may therefore be necessary, which may be achieved by adding a compatibilizer that resides on the protein/polymer interface, or covalent modifying protein surfaces to tune their hydrophobicity. Addition of coupling agents and reactive extrusion are also well-established strategies for compatibilizing blends, and some approaches include melt mixing proteins and polymers with reactive functional groups and introducing radical initiators to promote crosslinking between phases.[54,58]

Protein copolymers

Covalently conjugating or grafting polymers onto proteins expands the design space of hybrid materials and enables access to unique microstructures. The presence of chemically distinct polymer segments can drive self-assembly in hybrid materials, and unlike in physical blends, covalent linkages inhibit macrophase separation. Due to the large monomer possibilities, protein copolymers have greatly expanded design space and engineered properties. Judicious selection of the protein and conjugated polymer has enabled the design of advanced functional materials in a multitude of applications, such as protein therapeutics with improved pharmacokinetics, highly active immobilized enzymes, and stimuli-responsive materials.[23]

Drawing inspiration from exceptionally strong natural materials such as silk proteins, copolymers with good mechanical properties have been designed with polypeptides or proteins functioning as physically crosslinkers. Polypeptides with various conformations such as the nature-inspired β-sheets-forming silk-based peptides, amyloid-β fragments, and coiled-coil domains have been demonstrated to impart mechanical strength to gels and films.[59,60] Similar strengthening capabilities may be observed in proteins with a high propensity to aggregate, and rational design of the non-protein polymeric component allows further tunability of mechanical behavior and processability. As an illustration, hybrid copolymers with proteins and soft polyacrylate were shown to exhibit high strength and toughness (Fig. 6). The copolymer motif is similar to that of silk proteins and synthetic high-performance thermoplastic elastomers, where the microphase separated protein and polyacrylate contribute to the hard and soft domains, respectively.[61]

Figure 6
figure 6

Copyright 2017 American Chemical Society.

Copolymers consisting of stiff proteins and rubbery polymer segments. (a) Methacrylated whey protein is incorporated into the copolymer via grafting-through polymerization. A rubbery polyacrylate, polyhydroxypropyl acrylate (PHPA), is selected to provide extensibility to the elastomer. (b) Representative stress–strain curves for a polyacrylate crosslinked with a small molecule (black), a whey protein/PHPA blend (red), and a methacrylated whey protein-PHPA copolymer (blue). (c) AFM phase images of the copolymer showing non-uniform microphase separated domains, scale bar is 200 nm. Reproduced with permission from Ref. 61.

Full size image

Bioconjugation strategies

Approaches to prepare protein copolymers include grafting-to, grafting-through, and grafting-from strategies (Fig. 7). Grafting-to involves coupling reactive polymers to proteins and requires functional groups with high reactivities to overcome steric barriers and achieve high coupling efficiencies. Since the polymers are pre-fabricated, they may be synthesized in conditions incompatible with proteins. In grafting-from, polymer grafts are grown from protein macroinitiators or macro-chain transfer agents. Conjugated proteins can be prepared with high efficiencies using excess small molecule reactants and can be easily purified, but the polymerization conditions must be compatible with proteins. Grafting-through enables preparation of branched polymers via copolymerization with protein macromonomers. It has similar advantages as grafting-from strategies, but the distribution of proteins along the polymer chain may not be well controlled.

Figure 7
figure 7

Protein copolymer architectures obtained from grafting-to (a), grafting-through (b), and grafting-from (c) polymerization. Linear, branched, and crosslinked networks can be achieved, depending on the number of reactive functional groups per protein.

Full size image

By controlling the number of attachment sites per protein, linear, branched, and crosslinked copolymers may be obtained. Non-specific reactions involving abundant reactive groups such as amines and carboxylic acids typically produce heterogeneous products resulting from broad distributions in the number and location of modification. When multifunctional proteins participate in polyaddition or grafting-through copolymerization, the formation of crosslinked networks is likely.[61,62] On the other hand, grafting-from copolymerization with multifunctional proteins results in branched or comb polymers (Fig. 7).[63] If a particular copolymer architecture or a non-crosslinked copolymer is desired, proteins must be modified site-selectively. Some developed strategies include targeting naturally less abundant amino acids or unique positions such as the protein termini, and biosynthetic incorporation of noncanonical amino acids or substrates recognized by enzymes.[23] Targeting amino acid residues such as cysteine, tyrosine, and tryptophan affords some site selectivity as they are present at low frequencies at protein surfaces, but the effectiveness of such strategies will be highly dependent on the protein feedstock. Selectively targeting a unique site such as the N-terminus is immensely attractive, particularly as some strategies such as the pyridoxal-5’-phosphate (PLP)-mediated transamination are broadly applicable toward multiple N-terminal residues.[64] Although this site-selective bioconjugation has yet to be widely adopted for bioplastics synthesis, their potential at enabling structural control opens opportunities for targeting high performance and processable materials.

Reactive groups present on native proteins, notably ε-amino groups of lysine residues and thiols of cysteine residues, are valuable targets for direct attachment to polymers. Polymers end-functionalized with activated carboxylic acids, isocyanates/isothiocyanates, and epoxies can be coupled to proteins via reactions with amines, while polymers with Michael acceptors can react with both amines and thiols. Multiblock copolymers have been prepared by polyaddition of telechelic polypeptides with cysteine residues at chain ends and maleimide terminated poly(ethylene oxide).[65] Crosslinked networks may be obtained when proteins with more than two reactive functional groups are used, as demonstrated when poly(ethylene oxide) diglycidyl ether is grafted onto proteins.[62] Proteins and polypeptides have also been used in polyurethane synthesis, where primary amines react with diisocyanates or bis(cyclic carbonate)s to form polyurethane/ureas or polyhydroxyurethane.[66] In these polyaddition reactions, stoichiometric balance is required to achieve high molar mass copolymers. This may pose a challenge for industrial proteins where the number of reactive groups may not be known precisely, and the addition of other small molecule chain extenders or crosslinkers may be necessary.

Several native protein functional groups can function as initiating groups in grafting-from polymerizations. During free radical polymerization of vinyl monomers in the presence of proteins, mixtures of homopolymers and small fractions of graft protein copolymers are obtained.[67] Thiols and disulfide bonds in cysteines play a significant role in the formation of these protein-polymer grafts, as they can act as initiators, transfer agents, and terminators in radical polymerizations.[68,69] Percent monomer incorporated into protein copolymers can therefore be increased by unfolding proteins to expose cysteines. A few examples utilizing protein amine and hydroxyl groups as initiators in ring opening polymerization of cyclic esters have also been demonstrated.[70] While currently underexplored, these materials may be interesting as polyesters are degradable and can be bio-based. To ensure high grafting densities, water-free conditions are required to prevent excessive homopolymerization from adventitious initiation.

To increase protein reactivity in radical polymerizations and to enable preparation of well-defined protein-polymer conjugates, proteins may be functionalized with polymerizable, initiating, or chain transfer groups. A versatile strategy that converts proteins into polymerizable macromonomers involves functionalization with (meth)acryloyl groups, which enables incorporation of proteins into copolymers via grafting-through copolymerization.[71] Proteins may be copolymerized with various vinyl monomers including (meth)acrylates, (meth)acrylamides, and styrenes, and the strategy has enabled fabrication of hydrogels, films, and selectively photocured 3D printed parts.[25,61,72] Proteins may also be functionalized with alkyl halide atom transfer radical polymerization (ATRP) initiators or reversible addition-fragmentation chain transfer (RAFT) agents to enable grafting-from (Fig. 8). Aqueous ATRP may be performed on protein macroinitiators modified with 2-isobromobutyrate groups, using optimized combinations of ligand, halide, and reaction conditions.[21,73] Similarly, RAFT polymerization may be performed on proteins functionalized with molecules containing thiocarbonylthio moieties (Z-C(= S)S-R). Typically, proteins are attached to the reinitiating R groups, which enable the propagating polymer to remain covalently bound to the protein where the labile thiocarbonylthio is situated at the free polymer end [Fig. 8(b)]. If the protein is instead attached to the stabilizing Z-group, propagation occurs on free polymer chains and the resulting copolymer has a thiocarbonylthio linkage between the protein and polymer.[74] To ensure good polymerization control, care should be taken to avoid hydrolysis of the RAFT agents particularly in aqueous conditions and in the presence of nucleophilic protein reactive groups.[75] These controlled radical polymerization techniques not only enable control of polymer dispersities, but also expands the polymer design space by enabling block copolymer synthesis and post-polymerization modifications.

Figure 8
figure 8

Grafting-from protein copolymer synthesis via controlled radical polymerization. (a) ATRP with proteins bearing initiating group. (b) RAFT polymerization with proteins attached to the R-group (top) and Z-group (bottom).

Full size image

The feasibility of processing protein copolymers into films, molded or printed parts, fibers, and coatings dictates their relevance in materials manufacture. Strategies for increasing the processability of protein copolymers are therefore crucially needed, especially since crosslinking is often observed in multifunctional proteins. Crosslinked copolymers possess high mechanical, thermal, and chemical stability, but have limited processability post-polymerization. To shape these materials into their final forms, copolymer synthesis and crosslinking can be performed in molds or by selective photocuring in laser-based stereolithography 3D printers.[25] However, when solvents are used, post-polymerization drying is required, which can lead to dimensional instability as the crosslinked materials shrink during solvent evaporation. To mitigate this, the Olsen group has developed a solvent-free melt polymerization approach utilizing surfactants as non-volatile compatibilizers and plasticizers, thereby eliminating solvent drying steps.[72] Other strategies such as bioconjugation chemistries that enable non-crosslinked architectures, and the use of dynamic covalent chemistries may also be considered for imparting processability and recyclability in next-generation protein-based copolymers.

Structure–property relationships of protein copolymers

Microphase separation of protein copolymers provides another route to control material properties through microstructure. In well-defined monodisperse globular protein copolymer diblocks, ordered nanostructures such as lamellar, hexagonal, and disordered phases have been observed, and the ordering quality was hypothesized to be dependent on protein molar masses and secondary structure.[76] However, many commercial proteins are complex mixtures with varying reactivities, and thus far, copolymers prepared with them are microphase separated but lack long-range order. Despite the lack of uniform domains, the segregated protein-rich domains contribute to mechanical stiffness through physical crosslinking (Fig. 6). When combined with a rubbery polymer, hybrid materials with mechanical properties comparable to that of bio-based polyurethanes have been demonstrated.[61]

Like in plasticized protein-based materials, protein secondary structure in the copolymers influences macroscopic properties. Increasing intermolecular β-sheet in protein-polyacrylate copolymers by thermal treatment leads to increases in modulus, mechanical strength, and toughness through ordered domains with intermolecular associations.[77] Different secondary structures may also influence protein–polymer interactions and consequently long-range order and crystallinity, which provides a handle for tuning mechanical properties in poly(ethylene glycol)-based polyurea-polypeptide hybrids.[78] Developments in mechanistic understanding of the structure–property relationships along with strategies to control the effects of feedstock diversity are expected to advance the design of protein copolymers and potentially enable broader ranges of material properties to be accessed.

Protein sources and characteristics

Currently, proteins can be sourced from the food industry and obtained as co-products of oils, starch, and biofuels. The production of food is associated with generation of wastes, sometimes unavoidable in the process of transforming raw materials to products or in the form of inedible parts.[8] Utilization of protein byproducts has the added advantage of reducing pollution, particularly as protein-containing effluents have high biological oxygen demand (BOD) and need to be properly treated. Aside from food and feed, some crops are also cultivated for biofuels, which generate protein byproducts. In this section, we will briefly discuss the sources and unique characteristics of several types of proteins. The primary and secondary structural characteristics of these proteins are noted in Table I, including the presence of cysteine residues that can be important for the formation of intermolecular covalent bonds.

Table I Characteristics of a number of industrial proteins and their characteristics.
Full size table

Dairy

Bovine milk protein is made up of 80% casein and 20% whey protein. Cheese and casein manufacturing produces liquid whey as the byproduct after casein is coagulated with acid or rennet. Roughly 240 million tons of whey is produced globally every year, of which 12 wt% are proteins.[8] Whey proteins can be concentrated and purified to remove lactose, salts, and other small compounds, and are commercially available in multiple grades of protein purity, namely whey protein isolate (> 90%) and whey protein concentrate (45–80%). The primary proteins are β-lactoglobulin (57%) and α-lactalbumin (19%).[79,80] β-lactoglobulin and α-lactalbumin are compact globular proteins with structures and functions that are very well characterized. β-lactoglobulin consists of an eight-stranded β-barrel, a three-turn α-helix, and a β-strand, while α-lactalbumin has a large α-helical domain connected to a small β-sheet domain via a calcium binding loop.[89,90] Since β-lactoglobulin represents a large fraction of whey protein, it heavily influences the solution and gelation behavior of the protein mixture.[91] β-lactoglobulin is highly water soluble in the native state but is prone to gelation when heated or treated in alkaline conditions, driven by hydrophobic interactions and thiol-disulfide bond exchanges.[92] α-lactalbumin on the other hand has a lower tendency to gel when isolated, but aggregates at a faster rate in the presence of other whey proteins due to favorable interactions with β-lactoglobulin.

Casein proteins possess high nutritional value and are utilized as biomaterials but may be less relevanαt in large scale materials production. They are classified based on their electrophoretic mobility, and are known as the αs1, αs2, β, and κ fractions. Many of the casein proteins are block-like, amphiphilic, and post translationally phosphorylated on serine and occasionally on threonine.[81,93] In solution, the αs1, αs2, β caseins aggregate with calcium phosphate to form disperse micelles with diameters of roughly 200 nm, which are stabilized by hairy κ-casein on the surface.[82] Caseins can be precipitated by disrupting κ-casein through enzymatic hydrolysis or collapsing κ-casein at its isoelectric point and dissolving the calcium phosphate. A water-soluble form of casein can be recovered by readjusting the pH to ~ 7 using NaOH, KOH, NH4OH, or Ca(OH)2.

Cereals

Cereals, including corn, rice, and wheat, are staple food crops with global annual production volumes of thousands of millions of tons.[94] Two plant storage proteins, wheat gluten and corn zein, will be briefly discussed here.

Wheat proteins are obtained as co-products in starch production, and can be extracted by washing hydrated doughs with water to separate the protein fraction from starch and other soluble components.[83] The dried commodity protein contains a minimum protein content of 80%, ~ 85% of which is gluten. Wheat gluten is elastic when hydrated, and currently has primary uses in baked goods and as functional ingredients in the food industry. It is highly disperse, with roughly half of the protein comprising monomeric gliadins (30–75 kDa) and the rest polymeric glutenins (up to tens of millions Da).[95] Glutenin can be separated into high molecular weight (HMW, 60–100 kDa) and low molecular weight (LMW, 30–50 kDa) subunits in reducing conditions. Large polymeric glutenin is thought to be HMW subunits linked together in a head-to-tail fashion via disulfide bonds, while the LMW subunits attach to the HMW “backbone” like side chains. Smaller polymers or oligomers of only LMW subunits are also found. Gliadins associate with glutenin through secondary forces, primarily hydrogen bonding. Gluten is likely disordered when dry, but in the presence of water, loose β-spirals in the HMW subunits orient and form β-sheet-like structures, facilitated by the repetitive glutamine and proline-rich sequences. Despite the presence of hydrophilic residues, unaltered gluten is insoluble in water due to disulfide bonds and non-covalent interactions, but may be solubilized in the presence of reducing or chaotropic agents.[95]

Corn proteins constitute 9–12% of the kernel mass and are attractive as a material source by virtue of the sheer scale of commodity corn production. Of these proteins, zein makes up 44–79% of corn endosperm proteins, and may be extracted from whole ground corn or from corn byproducts. Commercial zein is mainly produced from corn gluten meal, which is obtained after separation of corn components by wet milling. Corn gluten meal contains at least 60% protein, and zein with higher purities are produced by applying multiple alcohol extraction/precipitation cycles.[96] Zein may also be extracted from distiller’s dried grains with solubles (DDGS), a co-product of dry milled corn for ethanol production. DDGS is cheaper, produced at larger volumes, and may experience growth along with the bioethanol industry.[97] However, DDGS contains less proteins and more polysaccharides than corn gluten meal, and the harsh processing conditions can alter zein extractability and characteristics. Using a nomenclature proposed by Esen, zein can be divided into α-, β-, and γ-, and δ-types.[98] α-zein is the primary zein protein and represents most of the protein extracted in commercial wet milled zein. It is insoluble in water but is generally soluble in solvent mixtures with water and small alcohols, ketones, or glycols[5]. β-, and γ-, and δ zein contain higher amounts of cysteines, and when co-extracted with α-zein in the presence of reducing agents, contribute to gelation.[84] α-Zein has sharply defined hydrophobic and hydrophilic surfaces and can undergo evaporation-induced self-assembly in aqueous ethanol, where decreasing solvent quality drives formation of spheres, lamellar, or bicontinuous phases.[99]

Soy

Soybean is the largest oilseed crop by volume and is mainly processed for oil and soybean meal for animal feed in the United States. The coproduced soy protein also has an established market, valued at ~ $170 million in 2022.[100] Conventional soy proteins are available as flours or grits (40–50% protein), concentrates (70% protein), and isolates (90%), where higher protein purities are obtained through a combination of dilute alkali extraction or alcohol washes followed by acid precipitation.[101] Due to the harsh extraction and spray drying processes, the obtained proteins may have poor solubilities in water. Components in soy protein are typically distinguished based on their sedimentation coefficients obtained from ultracentrifugation. They are namely the 2S, 7S, 11S, and 15S (S, Svedberg unit) fractions. The main storage proteins, β-conglycinin and glycinin, make up more than half of the 7S fraction and most of the 11S fractions, respectively. β-conglycinin is a trimer that is formed from various combinations of the three glycosylated subunits α, α′, and β, and are held together by hydrophobic and electrostatic interaction. Both the α and α′ subunits contain one cysteine residue, while the β subunit does not. Glycinin is a hexamer with two trimers stacked on top of one another, and like β-conglycinin, is also molecularly heterogeneous as it can be made up of combinations of subunits. Each of these subunits are further composed of an acidic and a basic polypeptide linked by disulfide bonds. The globulins have different aggregation behavior and mechanical properties, with glycinin reported to gel faster and have higher tensile strength in cast films.[86,102]

Meat byproducts

Keratin, found in the outer protective layer in vertebrates, is one of the most important structural proteins in nature. It is present in many inedible animal parts such as skin, hair, nails, feathers, and horns. Keratins are fibrous proteins and are categorized into α-keratin or β-keratin based on their helical or β-sheet structures.[88] α-keratin is made up of two right-handed α-helices that form a coiled-coil dimer held together with disulfide crosslinks. The dimers aggregate end-to-end to form protofilaments, which then laterally associate in pairs to form protofibrils. Protofibrils combine into fibrils, which wind together to form supercoiled structures that are embedded in a protein matrix. On the other hand, β-keratins are single proteins that make up both the filament and the matrix. The central region of the protein folds into lateral β-strands stabilized by hydrogen bonds, which twist to form distorted β-sheets. Two pleated sheets then assemble into filaments embedded within the matrix that is made up of terminal parts of the protein. α-keratin is found in mammals and can be sourced from low quality wool byproducts in sheep farmed for dairy and meat. β-keratin is found in birds and reptiles and is less studied. However, its utilization is very attractive as poultry feathers, which have annual production volumes of millions of tons, are ~ 90% keratin. Keratins have high mechanical and chemical resistance due to the large number of cysteine residues participating in intermolecular and intramolecular disulfide bonds and are consequently insoluble in water and organic solvents. They also have high thermal transitions, with crystalline melting temperatures around 230–240°C.[39] Keratin may be ground up into small particulates for use in material applications or solubilized with the aid of acid or alkali treatment, reducing agents, oxidizing agents, and hydrolysis.[10]

Collagen is an abundant structural protein found in connective tissues. It is traditionally obtained from the skins, tendons, and cartilages of animals such as pigs and cows, and is valued in the food, biomedical, and cosmetics sectors. Mammalian collagen utilization in materials manufacture may be limited by consumer concerns regarding religious and safety issues. On the other hand, valorization of collagen and myofibrillar proteins from aquatic sources are increasing in importance as part of efforts to recover and upgrade wastes from fishery by-catch and processing.[8] Some of the most common types of collagens found in meat and fish byproducts are type I collagen from bones, skin, tendons, and ligaments, and type II collagen from cartilages.[103] Collagen is made up of three ~ 100 kDa helical polypeptides with repeating sequences of Gly-X-Y, where X and Y are often proline and hydroxyproline, respectively. These α-chains pack into triple helices stabilized by intra and intermolecular hydrogen bonding.[103] Lysine, hydroxylysine, and their aldehyde derivatives located on the helical ends undergo crosslinking reactions to stabilize the fibrils. Mild hydrolysis and denaturation produce gelatin as a mixture of water-soluble polypeptides, which can regain some residual triple helical structures from native collagen. Commercial gelatins are available at different gel (Bloom) strengths, which are related to the content of triple helices that serve as physical crosslinking points. The sol–gel transition of gelatin at around ~ 41°C in aqueous media corresponds to the coil–helix transition.[104]

Algae

Algae proteins are diverse, with compositions and protein levels dependent on the species, culture conditions, and growth phase. Explorations of algae-derived materials are still in the nascent stage, and studies on protein-based materials are even more limited when compared to other algal biopolymers such as polysaccharides and polyhydroxybutyrate.[105] However, material production from algae is attractive as they have high growth rates, can be cultivated at high densities, and do not compete for resources needed for food production.[106] Microalgae are single cell microorganisms with high productivities, with protein levels up to 40–65% of the dry weight in green and blue-green algae.[107] They have attracted attention as a source of biofuels, which can be obtained in the form of ethanol and biodiesel from carbohydrate fermentation and lipid extraction, respectively. The remaining biomass residue after fermentation/extraction is rich in proteins and can be potentially transformed into materials. Most algae proteins have low contents of cysteine and methionine, and in some microalgae species, supramolecular protein–pigment complexes or phycobiliproteins may represent a large fraction of total cell protein.[107] The proteins can be extracted by disrupting the cell followed by dissolution in alkali under reducing conditions followed by precipitation, or by hydrolysis to solubilize the polypeptides.[106]

Outlook and conclusion

Proteins are attractive sources of materials and can play an important role in advancing materials sustainability. They are renewable and biodegradable, and their utilization is advantageous for the circular bioeconomy. Proteins can be sourced from a range of biomass, including from byproducts in the food and biofuels industries. Valorization of these byproducts mitigates environmental impacts of waste disposal and can contribute positively to market competitiveness of biorefineries. In addition, proteins may become available in large volumes; replacement of 10% of the global fuel demand with biofuels is estimated to supply an additional 100 million tons of protein per year, which exceeds the human consumption needs.[108] Many plant and animal proteins have been characterized to different degrees, and rising interests in their utilization can inspire further efforts to understand their properties and improve protein extraction processes.

Several material design challenges will have to be addressed for proteins to be established as alternatives to fossil-fuel-based plastics. One of the reasons for the ubiquity of plastics is their mechanical robustness, but proteins are brittle when dry. To increase material flexibility and improve processability of proteins, they are typically blended with small molecule plasticizers that disrupt intermolecular interactions and lower the glass transition temperatures. The resulting flexible films are commonly investigated as alternatives to single-use food packaging or edible films. They have tensile strengths on the order of ~ 10 MPa, comparable to that of LDPE, but much lower elongation-at-break.[34] Although proteins are good barriers to oxygen and oil, they have high water vapor permeabilities and therefore do not provide sufficient barrier properties to maximize food shelf lives. The poor barrier properties are often compounded by hydrophilic small molecule plasticizers. Replacing hydrophilic plasticizers with hydrophobic molecules and increasing crosslink densities were reported to have small to moderate effectiveness in reducing water vapor permeabilities.[33] Major improvements, or the design of multilayer composite materials will likely be required for these protein films to be competitive.

Advances in protein bioconjugation chemistries and knowledge of self-assembly behaviors provide exciting opportunities to further improve material properties and enable proteins to be harnessed in higher value applications. Nanostructured proteins, for instance, have been explored as novel functional materials in applications ranging from tissue engineering to water purification.[14] Since proteins have high reinforcing capability due to the extensive inter- and intramolecular interactions, they can also be incorporated into composites and copolymers in the design of high-performance engineering plastics. The design space for copolymers and composites is enormous and can enable a greater range of material properties to be targeted. For example, hybrid microphase separated materials combining the advantages of stiff proteins with rubbery polymer domains have been demonstrated to possess both high strength and toughness. A few challenges may be encountered in the design of protein-based composites and copolymers. The low protein solubility in non-aqueous solutions limits reactions and solution processing. Since proteins are hygroscopic, strategies that mitigate excessive plasticization from water absorption must be developed to achieve stable mechanical performance. Additionally, the effects of monomer and feedstock diversity should be accounted for when designing modification chemistries and self-assembled materials.

Life cycle assessments and biodegradability studies will be crucial in the design of next-generation protein-based materials. Although proteins are renewable and inherently biodegradable, feedstock farming, protein extractions, and materials fabrication processes may have negative environmental impacts, and the benefits of bio-based feedstock must be weighed against the use of fertilizers and pesticides which contribute to acidification potential and eutrophication.[2] Data on the degradation behavior of protein-based materials will be needed to ascertain the stability of these materials during use, and for determining any potential adverse effects of degradation products. Encouragingly, life cycle analyses performed on plasticized protein films and protein-synthetic copolymer hybrids show that the protein-based materials have lower environmental impact when compared to bioplastics like poly(lactic acid).[8,109] As new materials are developed, life cycle analyses with appropriate boundary conditions that reflect the manufacturing practices, material properties, potential uses, and end-of-life options will provide more accurate assessments on the sustainability of protein-based materials, while encouraging various aspects of the materials’ life cycle to be considered during material design processes.

In summary, proteins are promising biopolymers that can be transformed into plastics through a range of plasticization, blending, physical, and covalent modification strategies. Their reactive functional groups provide handles for modification, and the non-covalent interactions that drive folding and self-assembly can potentially introduce novel structures and properties into materials. Innovative strategies will be needed to improve their mechanical properties and processability, and better understanding of structure–property relationships will be important for addressing gaps in performance. The pressing need for sustainable materials, along with the broad applications of proteins in the biomedical, nanotechnology, energy, and purification sectors, is expected to catalyze these efforts.