CHAPTER 1: Synthetic Modifications of Proteins
Published:18 Aug 2015
U. Glebe, B. Santos de Miranda, P. van Rijn, and A. Böker, in Bio-Synthetic Hybrid Materials and Bionanoparticles: A Biological Chemical Approach Towards Material Science, ed. A. Boker and P. van Rijn, The Royal Society of Chemistry, 2015, pp. 1-29.
Download citation file:
In the pursuit to develop new biohybrid materials based on protein structures or peptides, peptide chemistry is a convenient approach. In addition to conventional peptide chemistry in which reactive amino acids located on the protein surface are targeted, over the past few years many new approaches have been developed. Not only new reactions able to target the protein chemically but also other disciplines such as polymer chemistry and biocatalysis are now being added to the ever-broadening portfolio of protein/peptide alterations. Lysozyme, bovine serum albumin (BSA), ferritin and proteases such as chymotrypsin are examples that have been modified via the presented approaches. Hybrid conjugates with polymers or nanoparticles, for example, can be achieved not only by covalent linkages but also by non-covalent interactions. New tools introduced into the ‘chemical toolbox’ to modify protein structures, including controlled radical polymerization [especially atom transfer radical polymerization (ATRP) and reversible addition fragmentation chain transfer (RAFT)] and ‘click’ reactions, are now commonly used to create bionanoparticle-based conjugates with new functionalities and hence associated with it new materials and applications.
1.1 Introduction: General Approaches to Protein Modification
Proteins are fascinating building blocks with outstanding functions in Nature. Bioinspired material synthesis utilizes these unique properties of proteins when integrated into synthetic structures. The properties of both the protein and the synthetic compound, inorganic or organic, can be combined to create novel hybrid materials for a broad range of applications.1 Proteins provide a multitude of applications, e.g. as catalysts, recognition sites, transporters, nanochannels and much more. Proteins are defined by their three-dimensional structures with distinct shape and size dimensions. Under suitable conditions, these structures self-regulate and can be re-formed into their natural shape. Unfortunately, this nearly perfect system in Nature is a disadvantage for the synthesis of biohybrid materials. Protein self-assembly is a sensitive feature and protein stability can decrease strongly under unsuitable conditions. To preserve the protein shape, including the resulting functions mentioned above, satisfactory conditions need to be found that allow the integration of synthetic components into protein-based systems in order to form biohybrid structures. In many cases, conditions that mimic the natural environment provide a suitable foundation. Generally, proteins are stable in aqueous media. However, not all chemical reactions needed for protein modification can be realized in pure aqueous systems. In some cases, organic cosolvents can be used to dissolve a wider range of chemical reactants. Nevertheless, the specific ratio for each organic solvent that conserves functionality and the overall folded structures of the protein requires attention. Additionally, heat, drastic changes in pH and certain chemicals cause denaturation of proteins.
The size, folding and function of proteins are predetermined by the precise sequence of amino acids. Several approaches are feasible to modify proteins selectively without influencing their functionality. The use of biotechnological methods allows for the exchange of specific amino acids or the deletion of some parts of the peptide sequence while retaining the protein shape and stability. Such changes can be used to add functional groups and also binding sites, to stabilize a protein or to alter the polarity of certain parts.1 Proteins can be integrated into organic or inorganic structures. By forming covalent or non-covalent bonds, small molecules, specific functional groups or even much larger nanoparticles can be attached to proteins. Particularly important are polymers, which are frequently used to modify proteins. The residues of amino acids which are oriented to the outside of the protein cage can be used for direct and covalent chemical modification. If polarity, charge and steric demand are not altered significantly, the protein remains comparatively stable with respect to its natural occurrence.
Hybrid materials offer the possibility to combine the beneficial properties of inorganic, bio- and polymeric structures to generate materials with novel functions.2 Applications and objectives for protein-based hybrid structures are drug/gene delivery, biosensing and bioimaging, in addition to creating materials for electronic devices and functional membranes.1,3,4 Furthermore, the conjugation to polymers may improve protein stability, solubility and biocompatibility and also increase the interfacial activity, decrease the amount of denaturation of the protein on an interface or render the proteins more compatible with other polymer structures.1,5 Over the past few years, bioconjugation strategies have been investigated intensively.6
Protein–polymer systems are attractive for medical applications in the fields of diagnostics, therapeutics and theragnostics.7 Biocompatible or even biodegradable polymers could result in applicable hybrid systems. Although not yet ready for practical use, polymeric templates with enzymes and other proteins are envisaged to act as reaction spaces in nanoscale dimensions.7 Here, the insertion and encapsulation of biological compounds into polymer systems are possible, thereby protecting enzymatic reactions through the surrounding polymer structure.7 Such systems combine the stability and robustness of polymer structures with the specificity and selectivity that the biological compounds provide.7 Polymer conjugation can stabilize therapeutic proteins and protect them, especially from enzymatic degradation.8 It has recently been shown that a proline-specific endopeptidase (PEP) was stabilized to such an extent that it retained function even when presented with conditions as harsh as those found in the stomach. This was accomplished through the covalent attachment of a positively charged dendronized polymer.8,9
For the conjugation of proteins to synthetic structures, complementary functional groups are needed. Two types of reactions are commonly used: polymerization and ‘click’ reactions. Polymers can be attached in one step (grafting-to approach) or the polymerization can be performed directly from initiator sites on the protein (grafting-from approach). Section 1.3 addresses the conjugation of polymers to proteins.
‘Click’ chemistry is a still fairly recent field that comprises bond-forming reactions with the aim of linking molecular building blocks via heteroatoms. In general, ‘click’ reactions are highly selective, easy to perform with a high tolerance towards the presence of other functional groups and solvents and lead to products that can easily be isolated in high yield with little or no by-product formation.10,11 Many ‘click’ reactions are biocompatible and can be performed in water, which makes them ideal tools for protein functionalization.5,12–14 The applicability to bionanoparticle modification was demonstrated shortly after the initial development of ‘click’ chemistry.13 The reaction between azides and alkynes to form a 1,2,3-triazole ring is one reaction of the ‘click’ chemistry pool, known as copper-catalysed azide–alkyne cycloaddition (CuAAC). Biological structures normally do not have alkyne and azide groups; however, they can easily be incorporated into proteins.5,15 CuAAC is a very suitable reaction for a specific ligation strategy for proteins and the most commonly applied ‘click’ reaction for bioconjugation [Scheme 1.1(a)].16,17 However, the correct selection of the catalyst–ligand system, because of potential damage to biomolecular structures, and a comparably slow reaction rate at the low concentrations typically required for bioconjugation purposes must be emphasized.18 The copper-free variant of the reaction, strain-promoted azide–alkyne cycloaddition (SPAAC), provides a metal-free alternative for protein conjugation [Scheme 1.1(b)].15 The activation barrier of the reaction is lowered by taking advantage of the inherent ring strain of cyclooctynes.18
Copper(i)-catalysed 1,3-dipolar cycloaddition is the best known reaction, but other ‘click’ reactions are also commonly used and several of the reactions presented in this chapter for protein modifications indeed fulfil the requirements of ‘click’ reactions. Among them are the Diels–Alder reaction [Scheme 1.1(c)], Staudinger ligation [Scheme 1.1(d)], pyridyl disulfide, Michael addition and thiol–ene reactions [Scheme 1.1(e)], and also hydrazine and oxime formation.5,16,17,19–22 These copper-free ‘click’ reactions have the advantage that the toxicity and the possible denaturing character of CuI are avoided.5,15,19 Furthermore, Cu ions readily promote the generation of reactive oxygen species (ROS) that often harm biological structures.18 Sometimes proteins have to be functionalized first in order to apply a ‘click’ reaction. In other cases, native proteins can be used directly without modification, an example being the SH group of cysteine. The reaction of thiols with non-activated alkenes can proceed via a radical (thiol–ene chemistry) or anionic (thiol Michael addition) pathway.16,23 Additionally, ‘click’ reactions are well suited for the surface immobilization of small molecular components, including fluorescent labels, and also polymers on protein structures.16,22
1.2 Amino Acid Targeting for Synthetic Protein Modification
Since the functional groups of chemical compounds, including polymers, are highly diverse, the chemical strategy for the synthesis of protein hybrid conjugates is predetermined by the amino acid residues of the protein. First, 10 out of the 20 natural amino acids bear functional groups that are suitable for ligation chemistry.24 Second, it is important to consider whether or not these amino acids are accessible on the protein surface for chemical modifications. Lysine and cysteine are the two amino acids most commonly used for conjugation. In addition, tyrosine, glutamine, tryptophan, histidine, aspartic acid, glutamic acid, arginine and phenylalanine plus the N- and C-termini of the peptide chain are also possible reaction centres.24 For the other amino acids, almost no specific reaction that proceeds without damaging the native state of the protein or competing with functional groups of other amino acids is known.24 Furthermore, the modification of a specific amino acid residue should not influence the conformation and function of the protein.
Amino acids that are embedded in the globular structure of a protein are not accessible for chemical reactions. The average surface accessibility (ASA) indicates whether amino acids are addressable by chemical reactions. On average, aspartic acid, lysine, glycine, glutamic acid, glutamine and serine are mostly located on the surface of proteins.24,25 If one amino acid is accessible multiple times on the protein surface, this can lead to undefined and polydisperse products with varying degrees of modification. As a result, the monodisperse character that is a defining feature of proteins is lost.
The lysine residues and the N-terminus of proteins or peptides are the most popular sites for polymer conjugation.21,26 Lysines not only have a common location on the surface of proteins, but additionally can be used in a range of chemical reactions. The nucleophilicity of the amine group is higher than those of nucleophilic groups of other amino acids.24 Reactions with activated carboxylic acids are commonly used. For this purpose, N-hydroxysuccinimidyl (NHS) esters are widely utilized in addition to other carboxylic acid derivatives such as NHS carbonates, NHS carbamates, anhydrides and acid halides. In addition to such acylations, alkylation is also used for lysine modification.24 Here, a charge on the amine group can be retained, for example through amidation with imido esters.21,24 This may be advantageous if the charge on the residue is important for the stability and structure of the protein. Aldehydes react reversibly with lysine groups to form imines that can be reduced to amines under mild conditions.21,24 Scheme 1.2 gives an overview of common products of lysine modifications. In general, most proteins possess many accessible lysine residues. This, however, is accompanied with the loss of monodispersity due to differing degrees of amine modification, as mentioned earlier.26,27 For different proteins, the reactivity of lysine residues varies significantly owing to accessibility and pKa values.28 By adjusting the pH, one can influence whether the reaction takes place at the amine group of lysine (pKa ≈ 10.5) or at the N-terminus of the protein (pKa ≈ 7.8).21,24 However, in practice this is very difficult to achieve.5
Cysteine is also often used for conjugation and usually does not have the disadvantage of forming polydisperse bioconjugates.24 Cysteine is present in only small amounts on the protein surfaces and there are especially few cysteine residues that do not participate in disulfide bonds.5 In combination with the possibility of using reactions that are not feasible with other amino acids, the small amount of surface amino acids can be exploited for selective modification. Two reaction pathways are common for cysteine modifications: the formation of disulfide bonds and alkylation during a Michael addition or thiol–ene coupling (Scheme 1.2).24 Activated disulfide groups such as orthopyridyl disulfides or methanethiosulfonates react with cysteine residues with the formation of disulfide bonds.21,24 Reducing agents can cleave the disulfide linkages formed, which can be an advantage or disadvantage depending on the desired application.5,21 Activated alkenes such as maleimides and vinyl sulfones form stable thioether bonds with thiol groups in Michael additions.21,24,26 Recently, it was shown that also mono- and dibromomaleimides and electron-deficient alkynes can be used for conjugation to cysteine residues.24,26 Sometimes disulfide bonds have to be cleaved to make them accessible for ligation. It is possible to reduce selectively the surface-accessible disulfide bridges only.5 In many cases, the removal of a disulfide bond in a protein influences the overall three-dimensional structure. However, possibilities exist for preserving linkages in these positions by linking the two sulfur atoms via a three-carbon bridge. After cleavage of the disulfide bridge, a reaction with a bis(thiol)-specific reagent stabilizes the correct tertiary structure of the protein.5,21,24
In addition to the common protein modification approaches using lysine and cysteine residues, further chemical reactions addressing the residues of other amino acids should be mentioned that may be needed especially when the surface of the protein structures is devoid of any conventional addressable functionality. Tyrosine can be modified by palladium-catalysed alkylation of the hydroxyl group, by an electrophilic aromatic substitution such as a Mannich-type reaction or by a diazonium coupling.21,24 However, these approaches have rarely been used up to now.24 For glutamine, only enzymatic reactions mediated by transglutaminase are available for modification.21,24 The indole ring of tryptophan can be used for rhodium-catalysed reactions in which a rhodium carbene is formed in situ.21,24 Histidine residues form stable complexes with divalent transition metal ions. So-called His tags are composed of six histidine residues in a row and are used as a common binding motif for metal ions in peptide chemistry.24 In order to bind chemical compounds via this motif, these compounds need also to be equipped with a metal binding site. Carboxylic acid groups of aspartic acid and glutamic acid and the C-terminus of proteins are reacted with amines in reverse to lysine modification. This is achieved by enzymatic and biosynthetic approaches.24 The carboxylic acid group can be converted to very different chemical groups that may be suitable for further modification.
1.3 Synthetic Approaches for Polymer–Protein Hybrid Structures
The modification of proteins with polymers is one of the most widespread approaches to create hybrid structures and offers tremendous possibilities for functional systems.29,30 Both polymerization and coupling reactions are used to achieve polymer structures on protein surfaces.31 First, three major strategies have to be distinguished (Scheme 1.3). The grafting-to approach involves the synthesis of polymer chains with suitable end-groups that are then linked to proteins. Protein-reactive groups can be introduced to polymers in a post-polymerization step or protein-reactive initiators are used directly for the polymerization reaction. The latter guarantees that each polymer chain indeed has a protein-reactive end-group.17 The biomolecules are only involved in the last synthetic step as the polymers are synthesized separately. An excess of polymer is needed for the conjugation reaction, which hampers purification of the protein–polymer hybrid. Unreacted polymer and maybe even protein have to be removed from the conjugate.26 This disadvantage can be avoided with the grafting-from approach. Here, the polymer chains grow directly from a protein. First, small initiator molecules are linked to the protein surface able to initiate the polymerization from these so-called macroinitiators. Hence the purification of the product is easier because only small molecules, unreacted monomer and the catalyst, have to be removed.26,27 Furthermore, the location of the polymers is defined by the initiator positions and steric hindrance is negligible.26 Particularly in cases where it is intended to link a high density of polymer chains to a protein, grafting-from is superior to grafting-to because the steric hindrance of the growing, initially short, polymer chains is much smaller.27 Additionally, the synthesis of hydrophobic polymer–protein conjugates and the preparation of conjugates with high molecular weight homopolymers or block copolymers are easier via grafting-from.26,32 The drawback associated with the grafting-from approach is that polymerization conditions have to be chosen that do not affect the activity and folded structure of the protein. The third possibility is the grafting-through strategy.5,21,27 Here, monomers that contain proteins/peptides are polymerized so that every repeat unit of the comb-shaped polymer structure has a protein/peptide moiety. Grafting-through potentially allows the incorporation of a high amount of proteins/peptides into the conjugates. Alternatively, monomers with a protein/peptide binding site can be polymerized and the biocomponents attached to the polymer backbone in a post-polymerization step. More details and examples of the use of the different grafting strategies are given below.
Controlled radical polymerization (CRP) techniques allow the synthesis of a wide range of polymers with a narrow molecular weight distribution.27 The two most widely employed methods for synthesizing polymer–protein hybrid structures are atom transfer radical polymerization (ATRP) and reversible addition fragmentation chain transfer (RAFT) polymerization.26,27 During an ATRP reaction, the polymer chains grow by the addition of intermediate radicals to monomers.33,34 ATRP utilizes a CuI catalyst to establish an equilibrium state between an active and a dormant species [Scheme 1.4(b)]. The active species, the radical of the growing chain to which monomer is added, is formed by reaction of the dormant species, the halogenated polymer chain, with CuI ions. The oxidized transition metal atom rapidly deactivates the propagating polymer chain and reforms the dormant species.34 By keeping the concentration of the active species low, termination reactions are suppressed and highly monodisperse polymer chains can be achieved. Therefore, radical polymerizations behave in a nearly living or controlled manner.34,35 Mild reaction conditions, tolerance towards various chemical functionalities, experimental simplicity and the commercial availability of most initiators and catalysts make ATRP the most studied and applied CRP method for the synthesis of bioconjugates.6 RAFT polymerization does not require a metal catalyst. The RAFT group, typically a thiocarbonylthio group, reacts reversibly with propagating polymer radicals to form the dormant species [Scheme 1.4(c)]. Here, a degenerative chain-transfer process establishes reaction control.36 In view of the grafting-from approach, the RAFT chain-transfer agent (CTA) can be linked to amino acid residues via its R- or Z-group. With the R-group approach, polymer chains are grown between the protein surface and the CTA, which allows post-synthetic modification of the polymer end-group and also chain extension. The Z-group approach is not strictly grafting-from as the polymer chains do not grow from the protein, but to the CTA directly attached to the biomolecule. However, the advantages are that termination products remain in solution and the possible separate characterization of the polymers because of the labile bonding to the protein via the CTA.32 In addition to the ATRP and RAFT reactions, the third major method of controlled radical polymerizations, nitroxide-mediated radical polymerization (NMP),37 has also been used occasionally for the synthesis of protein hybrid structures. Here, nitroxides are persistent radicals that act as mediators for the polymerization. The nitroxides (N–O˙) react reversibly with the growing polymer with a radical chain-end (P˙) to give the dormant species P–ON [Scheme 1.4(a)].
Especially ATRP and RAFT have huge potential in the field of polymer–protein bioconjugates and have brought major improvements in this area.27,32 CRP methods made it possible to synthesize polymers with a high degree of control over architecture and insertion of functional groups. CRP allows for diverse chemical compositions since many monomers and initiators can be used in addition to organic and aqueous solvents. However, in vivo applications usually require the complete removal of the RAFT group and of the ATRP catalyst. Many proteins exhibit binding pockets for copper or iron ions and their biological integrities may be affected after performing a metal-catalysed polymerization reaction.38 Copper ions are known to be cytotoxic26,27,31 and the use of some ATRP catalysts is limited owing to their influence on protein structure and function. The method used for purification was shown to be critical in this context.26 Nevertheless, protein–polymer conjugates could have numerous medical applications. Many examples show that the biological activity of proteins in bioconjugates is preserved to a high degree.27
ATRP has been extensively studied in the grafting-from approach and especially biological-compatible reaction conditions can be used such as aqueous or buffered solutions, near-ambient temperature and catalyst systems that do not significantly harm protein structures.32 However, ATRP under aqueous conditions is challenging, mainly owing to high radical concentrations, and hence increased rate of termination and catalyst stability.39–42 Halide dissociation from CuII, competitive coordination of solvent, destabilization of the Cu–ligand complex, disproportionation or oxidation of CuI can occur in addition to hydrolysis of the alkyl halide initiators and chain ends.32 Proper reaction conditions need to be found that are successful in combination with protein structures. Some studies have systematically investigated ATRP-based polymerization reactions under biological-compatible conditions in the last few years.32,41,43 Several variants of CRP techniques have been investigated for polymerization in aqueous media in recent years and are promising for the modification of proteins. Some of these, namely ARGET ATRP and AGET ATRP, were successfully used for grafting polymers from proteins.41,43 For these ATRP variants, a reduced amount of copper salt is needed and they provide a more stable catalyst precursor, aimed at improving the polymerization from proteins. The substantially lower content of copper favours these types of polymerization for biomedical applications.
In the following, several examples of the synthesis of protein–polymer conjugates are presented. As only a few examples of particular proteins and recent developments are presented, the reader is referred to reviews for a more comprehensive overview.5,6,17,21,26,27,37,38,44,45
The grafting-to technique is the first and most widespread approach for the synthesis of protein–polymer bioconjugates.27 Well-defined polymers were synthesized containing functional groups—at either the α- or ω-terminus, or both21,27 —that were chosen for conjugation to proteins. Appropriate functional initiators are used in their native form or in a protected state for α-functionalization through radical, anionic and cationic polymerizations.21 For ω-functionalization, labile end-groups of polymers are modified into useful functionalities.21 Polymers with functionalities at the α- and ω-positions can be used to connect proteins and form dimers or multimers of higher orders.27 Some reviews provide an overview of protein-reactive initiators, terminators and mediating agents used for polymerization reactions prior to grafting-to.5,17,21,27 The preformed polymers are then linked to selected amino acid residues. PEGylation, the modification of proteins with poly(ethylene glycol) (PEG) polymers, is a typical example of grafting-to and was complemented with other PEG-containing polymers.17,27 The formation of PEGylated protein drugs was reviewed by Maynard and co-workers.46 Proteins such as lysozyme and bovine serum albumin (BSA) are typical examples that have been used in many grafting-to conjugation approaches. The amine groups provided by the lysine residues were often coupled with NHS esters or aldehyde groups of polymers whereas thiol groups of cysteine residues were reacted with maleimide or pyridyl disulfide α-functionalized polymers.27
CuAAC ‘click’ chemistry has also been used in combination with the grafting-to approach. Both combinations of azide-functionalized polymers and alkyne-bearing proteins or vice versa have been utilized.27 For polymer synthesis, ATRP initiators and RAFT agents with azide or alkyne functions can be used, in addition to post-synthetic substituents of halide atom chain ends.27 Especially the combination of CuAAC and ATRP has been used for the construction of hybrid bionanoparticles. The halogen end-groups of polymers prepared by ATRP can be easily functionalized and the catalyst–ligand system is very similar for both reactions.6
The grafting-from strategy is a straightforward, but not trivial, approach that has not been as widely investigated as grafting-to.6 Polymerization reactions on proteins cannot be initiated directly from amino acid residues. Hence amino acids are converted to polymerization initiators via functionalization with appropriate small molecular components. Proteins modified in this way are called macroinitiators. The strategies for the synthetic modification of amino acids presented in Section 1.2 are utilized for this purpose. For example, ATRP initiators such as 2-bromoisobutyric acid are coupled to amine groups via activated ester chemistry.24 Polymerization from protein macroinitiators was successful with ATRP, NMP and RAFT methods.21 CRP techniques are well suited for use in the presence of biological structures in view of their tolerance towards many functional groups. Among the first examples of grafting-from were the ATRP reaction of N-isopropylacrylamide (NIPAAm) on streptavidin and polymerization from BSA by Maynard and co-workers.47,48
An example of a protein to show various modification strategies is ferritin. Ferritin is an iron storage protein with a well-defined globular morphology that can be found in many animals, plants and prokaryotes. For ferritin, it was shown that several approaches for protein modification are successful, including grafting-to, grafting-from and ‘click’ reactions. Ferritin is stable when exposed to small amounts of organic solvents during synthesis. Wang and co-workers modified the lysine residues of apo-horse spleen ferritin (apo-HSF) with NHS ester chemistry.49 On one side, an alkyne moiety for subsequent CuAAC ‘click’ reaction was introduced [Scheme 1.5(a)]. A macroinitiator for ATRP was formed on the other side and PEG methacrylate was polymerized in this grafting-from approach, used for the first time with ferritin [Scheme 1.5(b)]. The lysine residues on the outside of the HSF nanocage were similarly used for modification by the groups of Russell and Emrick.50 A 2-bromoisobutyryl initiator unit was conjugated by reaction of the incorporated NHS ester bond with the amine groups of the lysine residues. Grafting-from polymerization of methacryloyloxyethyl phosphorylcholine (MPC) or PEG methacrylate (PEGMA) was performed by ATRP in order to control the polymer molecular weight through the monomer-to-initiator ratio. Alternatively, NHS-terminated poly(PEGMA) was synthesized in an ATRP reaction and conjugated to ferritin in a grafting-to approach. The polymer–ferritin particles displayed a different affinity towards polystyrene-b-polyethylene oxide (PS-b-PEO) block copolymer thin films and also a different interaction with antibodies compared with native ferritin.50 Another example of the synthesis of hybrid materials with ferritin is the construction of new membrane materials (see below) and complex networks, as developed by Böker and co-workers using the increased affinity of the hybrid particles towards polar/apolar interfaces.29,51–53
A few examples demonstrate the synthesis of protein–polymer conjugates containing block copolymers using a grafting-from approach. Sumerlin and co-workers showed that two consecutive RAFT polymerizations can be performed from the surfaces of lysozyme and BSA.54,55 A RAFT CTA was conjugated to both proteins via its R-group. This allowed the formation of block copolymers as the polymer chains grew between the protein surface and the RAFT group. The CTAs were linked to the lysine residues of lysozyme by reaction of the NHS-activated ester bond of the CTA or to a cysteine residue of BSA via Michael addition. After the first well-controlled polymerization of NIPAAm, the RAFT end-group was retained. Thus, the polymer chains could be extended by polymerization of N,N-dimethylacrylamide to yield the thermoresponsive conjugates containing block copolymer chains (Scheme 1.6). In addition to RAFT polymerization, block copolymers were also successfully grafted from a protein by ATRP. An ATRP initiator was linked to the lysine residues of chymotrypsin and the monomers sulfobetaine methacrylamide (SBAm) and NIPAAm were polymerized successively.56
The conjugates mentioned in the last paragraph contain responsive polymers that are occasionally used to create ‘smart’ hybrid conjugates. The polymer chains show responsiveness towards external stimuli such as pH, temperature, enzymes or ion/cofactor binding and undergo a change in size or structure.3 Poly(NIPAAm) is commonly used as a thermoresponsive polymer. As an example, Scheme 1.7 shows thermoresponsive conjugates that can be exploited for the separation of proteins from a complex mixture.57 Russell’s group recently discovered that the responsiveness of attached polymers could be used to tailor enzyme activity and stability.56,58,59 Chymotrypsin, a serine protease, was used as a model enzyme; 12 out of 14 surface lysine residues were modified by reaction with a water-soluble ATRP initiator [Scheme 1.8(a)].58 The subsequent grafting-from polymerization reaction conjugated poly[2-(dimethylamino)ethyl methacrylate] [poly(DMAEMA)] to chymotrypsin. With this temperature- and pH-responsive polymer, the polymer shell surrounding the protein could be changed in order to achieve higher stability and activity of the enzyme, even in a pH range where it is usually inactive or unstable [Scheme 1.8(b)].58 In addition, chymotrypsin conjugates with the thermoresponsive polymers poly(NIPAAm) and poly[N,N′-dimethyl(methacryloylethyl)ammonium propane sulfonate] [poly(DMAPS)] were synthesized and investigated in the context of this so-called ‘polymer-based protein engineering’.59 The conjugate of chymotrypsin with the polymer chains poly(SBAm-b-NIPAAm) introduced in the last paragraph responds to both low and high temperatures because of the different UCST or LCST values of the polymer blocks. The enzymes surrounded by a dual-zone shell showed higher stability even to harsh conditions of temperature, pH and protease degradation.56
Maynard and co-workers developed a strategy not to modify amino acids of formed proteins, but to design artificial amino acids with initiator units that will subsequently be incorporated into peptide strands.26,60 In this way, it is possible to run polymerization reactions precisely from predetermined amino acids even when the protein contains large amounts of the chosen amino acid (Scheme 1.9). The peptide macroinitiator was prepared by solid-phase peptide synthesis (SPPS). Matyjaszewski and co-workers recently reported a variation of this strategy for grafting from green fluorescent protein (GFP) that allows the direct analysis of the grafted polymer.61 A non-canonical amino acid with an ATRP initiator unit that has a base-labile ester bond was genetically incorporated into GFP. After polymerization, the polymer chain formed could be cleaved from the protein through hydrolysis of the ester bond and analysed by gel permeation chromatography (GPC).
The grafting-through technique leads to bioconjugates with increased local concentrations of peptides or proteins. Their observed biological activity can be significantly enhanced through possible multivalent interactions.5 Several systems have been developed with successful application of the grafting-through strategy.27 The polymerization of peptide/protein-based macromonomers has the advantage that all repeat units of the conjugate bear a peptide or protein unit. Peptides have been prepared that contain a broad range of functional groups used for polymerization reactions.21 However, synthesizing peptide-functional monomers is not trivial.5 For the alternative, the post-polymerization modification of a side-chain functional unit, it is difficult to achieve quantitative functionalization with peptide units.
Sometimes it is necessary initially to synthesize a protein. Possible reasons for this can be a low abundance of a natural protein, difficulties related to its isolation or purification, in addition to the aim of having a particular amino acid residue at a specific position in the peptide chain.21 There are three main methods for synthesizing peptide or protein structures, namely polymerization of amino acid N-carboxyanhydrides (NCAs), solid-phase peptide synthesis and protein biosynthesis. Non-canonical amino acids can also be incorporated into the structures.21 Particularly for protein–polymer conjugates, typically two different approaches are used to synthesize proteins by means of the above-mentioned methods. First, polymer and protein are synthesized separately and linked subsequently, or second, one of the compounds is synthesized on the other one which is prepared first.45 The combination of a peptide synthesis method with a CRP technique allows the preparation of well-defined hybrid conjugates. The polymerization of NCAs has been combined with either a CRP method or ‘click’ chemistry to construct polymer–peptide bioconjugates.27,45,62–64 The drawback of NCA ring-opening polymerization is the lack of precise control over the chain length and monomer sequence.45 Recent efforts have been aimed at the preparation of polypeptides with well-defined structures via NCAs.6 Solid-phase peptide synthesis is a routine technique with precise control over chain length and monomer sequence, but the maximum chain length is restricted to about 50 amino acids. The tailor-made peptide is built via stepwise addition of amino acids on a resin support.38 Here, the desired side-groups can be introduced at a specific step in the peptide synthesis. To synthesize conjugates, traditional peptide coupling conditions are frequently used as described before.21,45 Moreover, it is possible to synthesize the peptide domain from a soluble or solid-supported polymer and even the whole bioconjugate on a resin support.21,27,38 Longer peptides or proteins are conveniently prepared using biosynthetic methods that overcome the limitations of the NCA and SPPS methods. Although not a trivial method, through protein biosynthesis high molecular weight proteins with precisely controlled chain length and monomer sequence can be achieved. Again, polymers can be added by grafting-to in addition to ATRP or RAFT polymerization on protein macroinitiators.21,45
Several proteins are used as components in polymer-based membrane materials. Protein building blocks are of potential interest for membrane formation, because they form well-defined and uniform pores owing to their defined structure and also their variability in size and shape.2 In membranes, proteins can either act as templates for holes after denaturation or act as channels in the polymer structure on their own. The possibility of not only generating size-selective pores, but also selective ion or molecular transport is especially interesting.2 Two different approaches are used for the creation of polymer–protein membrane materials. In the first, the proteins are not covalently linked to the polymer material, but inserted into polymer membranes.65 In the second, bioconjugates are synthesized by covalently linking proteins to polymers to construct membranes with improved properties such as chemical and mechanical stability.
The first approach focuses on the construction of vesicle structures that can function as nanoreactors. These are usually composed of amphiphilic triblock copolymers that form spherical hollow vesicles, the so-called polymersomes, in aqueous solution that are advantageous compared with the traditionally used liposomes with respect to their higher stability.7,66,67 In principle, various supramolecular assemblies can be formed with amphiphilic block copolymers, ranging from dendrimers, spherical micelles, cylindrical micelles and capsules to polymersomes.7,68 The polymersome membrane is composed of the hydrophobic polymer part inside the two hydrophilic layers. The flexibility of the polymer systems permits the insertion of proteins into vesicle membranes if the hydrophobic part of the polymer membrane has dimensions that fit the hydrophobic part of the channel protein used. Such proteins are especially suitable because the polymer membrane mimics their natural environment and stabilizes the protein. If the hydrophobic parts of the protein and of the polymer membrane do not have exactly the same dimensions, this so-called hydrophobic mismatch renders it inefficient to incorporate a large amount of proteins into the membranes.2,67,69 Not only have suitable polymers been constructed, but also a protein has been engineered to match the polymer material.69 Polymersomes with incorporated proteins are nanocompartments that can be used for enzyme catalysis or as traps for compounds and hence are called synthosomes.70 The nanocontainer walls should be as impermeable as possible, so that the flux through the membrane is controlled only by the channel proteins. Examples include the insertion of the outer membrane protein F (OmpF),71–76 Aquaporin Z (AqpZ),77,78 Tsx,76 bacterial channel-forming protein (LamB),79 ferric hydroxamate uptake protein component A (FhuA),69,70,73,80 claudin-2,81 reduced nicotinamide adenine dinucleotide (NADH):ubiquinone reductase (complex I)82 and adenosine triphosphate (ATP) synthase together with bacteriorhodopsin83 into polymer vesicles. In nearly all examples, poly(2-methyloxazoline)-b-polydimethylsiloxane-b-poly(2-methyloxazoline) (PMOXA-b-PDMS-b-PMOXA) was used exclusively, which in part relates to the extreme flexibility of the hydrophobic block that adapts best to the requirements of the membrane proteins.84 Figure 1.1 illustrates how a reaction can be catalysed in a polymersome with incorporated protein channels enabling permeability. The examples given show that proteins can function as selective gates, chemical reactions take place exclusively inside the vesicles and compounds can be entrapped for specific release. Polymersomes are especially envisaged as confined reaction spaces and delivery vehicles for medical applications.7,85,86 Additionally, a biomimetic membrane with incorporated AqpZ channels was formed on a porous planar support by vesicle spreading.77
Other studies have focused on the direct ligation of polymers to proteins and subsequent formation of the membrane by cross-linking the covalently bound polymer chains. By this approach, the amount of incorporated proteins will be significantly higher than in the approach for synthosomes. As an example, ferritin was modified in a copolymerization with NIPAAm and cross-linkable 2-(dimethylmaleimido)-N-ethylacrylamide (DMIAAm) via 2-bromoisobutyryl-functionalized lysine residues on the surface of the protein and hybrid membranes were constructed from Pickering emulsions at liquid/liquid interfaces.29,51–53 After self-assembly and cross-linking of the conjugates at an air/water interface, a polymer membrane was formed in which ferritin acted as a template for uniform holes that were created through denaturation of the protein.87
A similar approach was utilized by Mann and co-workers, who ligated poly(NIPAAm) to BSA and linked the proteins after self-assembly of the conjugates at the oil/water interface.88 The amount of amine residues of BSA was increased by reaction of the carboxylic acid groups of aspartic and glutamic acid with 1,6-hexanediamine followed by grafting-to of functionalized poly(NIPAAm) [Figure 1.2(a)]. Proteins were cross-linked by ligation of unreacted amine groups with PEG-bis(N-succinimidyl succinate) after self-assembly at an oil/water interface [Figure 1.2(b)]. The assembly did not occur with the unmodified protein, indicating that the amphiphilicity of the protein–polymer construct was critical for stabilizing the water micro-droplets. These ‘proteinosomes’ exhibit protocellular properties.88
The examples described so far mostly consist of conjugates with one protein and one or multiple polymer chains. With either homo- or heterotelechelic polymers and also star-shaped structures, conjugates were formed with multiple biomolecules and even multiple proteins of different types.26 Higher binding affinities, efficient anchoring to surfaces and improved biosensors could be achieved by multiple anchoring.26
1.4 Non-Covalent Approaches for Polymer–Protein Conjugates
In addition to the described strategies to form covalently bound protein–polymer bioconjugates, apart from the polymersome approach, strategies with a non-covalent binding motif were developed. This is usually done by the use of cofactors. These are small organic molecules that fit perfectly into the active site of a certain protein usually having great importance for the activity of the protein. By covalently attaching the cofactor to a polymer, bioconjugates with non-covalent binding sites can be constructed.5 The best known example is the system of (strept)avidin and the cofactor biotin. Both avidin and streptavidin are constituted of four subunits, each of them having a biotin binding site.38 The interactions between avidin and biotin (Ka = 1015 M−1) and streptavidin and biotin (Ka = 1013 M−1) are actually seen as the strongest non-covalent biological interaction.27 The complexes are very stable even under harsh conditions and the kinetics of dissociation are extremely slow compared with the time scale of most experimental methods.38 The carboxylic acid group of biotin can easily be used for ligation to polymers and also for the incorporation of biotin into polymerization initiators.5,17 Further examples of the construction of polymer–protein bioconjugates based on non-covalent interactions use the haem cofactor protoporphyrin IX together with (apo-)myoglobin or (apo-)horseradish peroxidase (HRP) plus the barstar–barnase system (Ka = 1014 M−1).1,5,27
Glycopolymers are another example of the exploitation of non-covalent interactions to bind multiple proteins. Sugars are information-rich molecules owing to their structural diversity. Although the mechanisms are not completely clear, recognition processes are thought to proceed by specific, non-covalent, carbohydrate–protein interactions.38 Although protein–saccharide interactions are typically weak, a high affinity and a high specificity in the binding motif are reached through multivalent interactions.27,89 Glycopolymers, polymers bearing a saccharide moiety in every monomer, can act as a binding site for certain proteins. Lectins are such carbohydrate-binding proteins. Synthetic glycopolymers can be prepared by polymerization of either glycomonomers or monomers bearing reactive sites for subsequent functionalization with sugar moieties.27 The multivalency of glycopolymers provides the opportunity to bind lectins. Glycopolymers can be obtained with many incorporated functional groups, for example, biotin, maleimide, pyridyl disulfide or azide moieties.27 These functionalities were used for conjugation onto several proteins.90 Glycopolymers also proved applicable for protein stabilization.91,92
Electrostatic binding motifs are less specific non-covalent interactions used for the conjugation of polymers to proteins. Many biological structures exhibit charged surfaces that can be utilized for electrostatic interactions.6 Furthermore, histidine-tagged proteins can form a complex with immobilized nickel ions, for example.1
1.5 Protein–Nanoparticle Hybrids via Surface Conjugation
Nanoparticles are frequently used to construct hybrid conjugates with proteins. Biologically modified nanoparticles are an important goal in nanomedicine as the crossing of biological barriers, the delivery of therapeutic agents and the achievement of specific targeting could be improved.22,27,38,57,93 Hence nanoparticles biofunctionalized with (poly)peptides have been synthesized and often applied for drug delivery, biosensing, bioimaging and even bioelectronic devices.4,93 A wide variety of core materials, including metals, metal oxides and semiconductors, can be used. Protein–nanoparticle conjugates have been prepared by either covalent or non-covalent linkages. The interactions of nanoparticles with proteins may have several consequences for the biomaterial where the interface plays a crucial role, e.g. the probability of partial protein unfolding can be enhanced.94 This may lead to perturbance of the biological function due to the high local protein concentration on the nanoparticle surface. Especially upon non-specific covalent conjugation onto polymer nanoparticles, proteins tend to lose their biological activity.57 In other cases, the adsorption onto nanoparticles can stabilize the structure and activity of a protein.95 Covalent protein–nanoparticle conjugates formed by protein side-chain modifications or engineered ligands were used to master the challenge of preserving protein structure and activity.57,96 A biocompatible spacer, usually oligo(ethylene glycol) (OEG), is commonly used to minimize protein denaturation.96 The adsorption to nanoparticles through non-covalent interactions can similarly influence the conformation of the protein and thereby the forces that stabilize the structure. Non-covalent conjugates can be prepared using complementary electrostatic charges between the protein and the nanoparticle, in addition to metal-mediated and hydrophobic and bioaffinity interactions [biotin–strept(avidin), carbohydrate–lectin; see Section 1.3 for details].57,93,95,96 The reversible nature of non-covalently bound conjugates permits applications in sensing and protein delivery. In biological environments, proteins form a bionanointerface on the nanoparticle surface, the so-called ‘protein-corona’.94 This corona is composed of many proteins attached to the surface of the nanoparticle.
1.6 Biocatalytic Approaches for Biohybrid Structures
Enzymes are widely used as biocatalysts in order to produce covalent bonds, and the advantage of enzymatic reactions compared with chemical and physical methods is that they are biocompatible by nature.97,98 Enzymes promote site-specific modifications under mild reaction conditions and generally do not produce toxic by-products, which therefore allows applications in biomedicine such as tissue engineering, and also in drug and food processes. Certain enzymes are able to target specific amino acids and are able to modify them. This section explores the most frequently used enzymes together with some applications in protein modification for creating biohybrid materials.
Transferase is a class of enzymes able to transfer a functional group from one molecule to another and functions under conditions that ensure the preservation of the overall structure, folding and activity of most protein structures, which is a very interesting advantage. Wagner et al. presented a wide variety of substrates used by aminoacyl tRNA transferase to modify the N-terminus of a protein.99 By selectively targeting the N-terminus with respect to other available amine groups found in lysine residue containing peptide strands, one can synthesize very specific Janus-like structures with a high degree of control over position and amount of modification.
Oxidoreductase is a class of widely applied enzymes that have the same modification mechanism. These enzymes act as a catalyst in electron transfer (oxyreduction) reactions. Tyrosinase acts on tyrosine residues by transforming the phenolic group in an o-quinone. Furthermore, the modified protein can couple non-enzymatically with other molecules containing nucleophilic groups such as primary amines.97,100 Converting only one specific type of amino acid into a more reactive one without altering the overall three-dimensional structure is very convenient, otherwise it would mean redesigning the protein genetically in order to introduce more reactive amino acids. Another interesting enzyme is laccase, which is an oxidoreductase of the multinuclear copper-containing oxidoreductases.101 The enzyme functions under relatively mild conditions, consumes oxygen and produces water as its only by-product.102 It was shown that laccase can also modify small peptides in the tyrosine residues by transforming the tyrosine into a free radical, permitting a free radical polymerization.103 Peroxidases modify the tyrosine group, generating a radical with the abstraction of a proton, they are able to perform site-specific cross-linking and they use hydrogen peroxide as cofactor. It was shown that even genetically introduced tyrosine residues were modified specifically.104
Enzymes are used as biocatalysts of cross-linking reactions in food processing, leather and textile fabrication, tissue engineering and biochemical and biomedical research. A lot of effort has been put into developing a better understanding of these areas because of its potential on an industrial scale. Enzymes are already widely used on an industrial scale for the synthesis of various chemical intermediates ranging from amino acids to alcohols, acids and penicillin, and are also widely applied for protein cross-linking in food processes.97,105 This can add functionality, lower the immune-response and add nutritional value to the protein.
It is of great interest to improve an enzyme’s performance, stability, activity and also identify new substrates to be targeted.106 The immobilization of enzymes on a surfaces is one of the strategies to improve stability, allow re-use of the biomolecule and enhance the activity.107,108 During immobilization, the enzyme can lose part of its activity owing to unfavourable positioning during binding or induce denaturing effects. Using multipoint binding can reduce such effects. Improvements in enzyme immobilization and activity are very much related to synthetic modifications, as mentioned in previous sections. Also for protein immobilization, ‘click’ chemistry reactions have made a significant impact, and an overview of specific developments in this area was recently published by Palomo.109 Combining enzymatic chemical conversions with other synthetic methods broadens the overall approaches that can be adopted in order to construct new biohybrid materials. Although they are not yet as frequently used for this purpose as chemical modifications, in the near future more biocatalytic approaches will most likely emerge.
1.7 Discussion and Conclusion
While it has been recognized that biohybrid materials are of great importance, a high degree of control over particle synthesis is required to achieve control over the material properties. Various approaches ranging from non-covalent or covalent synthetic methods to biocatalytic modifications of protein/peptide structures are used for the synthesis of conjugates. Every synthetic modification has its specific conditions, which are not always compatible with delicate biological structures such as proteins. Therefore, it is pivotal that new and refined synthetic methods are developed to meet the requirements for the vast amount of protein structures available, each with their own stability, function and amino acid/chemical composition, and useful in different (bio)medical, sensing, food and cosmetic applications. So far the combination of proteins with polymers is regarded as the most versatile and hopeful approach for hybrid bioconjugates,30 since numerous polymers/monomers are available with all kinds of different chemical functionalities that can be used in different proportions with many possibilities also to change the polymer architecture, such as linear, block copolymer, branched, dendrimer, bottle-brush forms, etc. Even though many combinations have already been developed, the most recent examples of which have been highlighted here, new combinations and with them new functions and materials are still under intense investigation, stressing the huge potential of the presented approaches.