Proteins can be considered the ultimate polymers for the synthetic chemist; they have high molecular weights and precisely defined primary structures and are perfectly monodisperse. Despite the impressive advances that have been made over the past decades, efficient chemical strategies for the preparation of high-molecular-weight polymers with uniform chain lengths and precisely controlled monomer sequences are still lacking. Yet, the chain length uniformity and monomer sequence (primary structure) are two important parameters that form the basis for the hierarchical folding and organization of polypeptide chains into functional proteins. The exquisite properties of many proteins, such as silk or enzymes, are directly related to their highly organized nanoscale structures. Proteins, however, also have limitations. Although they are constructed from natural building blocks, they can be toxic and elicit undesired immune responses. Furthermore, proteins in general are susceptible to enzymatic degradation and have limited solubility in nonaqueous solvents and only restricted temperature and pH stability.
The nanoscale structure of synthetic polymers is often described in terms of cooked spaghetti. This is not completely correct; semicrystalline polymers and block copolymers, for example, possess well-ordered crystalline or phase-separated microdomains, which are important for their (mechanical) properties. These nanoscale features, however, are by far not as complex as certain protein tertiary or quaternary structures found in nature. Also, unlike many proteins, which are able to recognize and bind certain guests, most synthetic polymers can be considered biologically inactive. However, whereas the toxicity and immunogenicity of de novo designed proteins are often difficult to predict, synthetic polymers can be prepared with a very specific set of biological and degradative properties.
This discussion on the scope and limitations of natural polymers (proteins) on the one hand and of synthetic polymers on the other hand is certainly not exhaustive and merely intends to show that each of these two classes of materials has its own specific characteristics. Moreover, it suggests that these two are largely complementary. Hybrid materials that contain both protein and synthetic polymer elements could synergetically combine the properties of the individual components and overcome several of their limitations.1 Although a relatively large number of reports have described the random conjugation of synthetic polymers to proteins,2 this article exclusively focuses on synthetic polymer–protein hybrids in which the polymer/protein ratio is precisely defined and the site of conjugation is exactly known. In most cases, the polymer/protein ratio will be 1, and the hybrids will have a block copolymer structure. This article does not discuss the relatively simple hybrid block copolymers that can be obtained by the homopolymerization of α-amino acid N-carboxyanhydrides with a synthetic macroinitiator. These materials have been extensively reviewed in other recent articles.3–5 The biological component of the hybrid materials that are discussed in this contribution is either a protein or a short peptide sequence derived from a protein. This article consists of two parts. In the first part, different synthetic methods that are used for the preparation of site-selectively conjugated protein or peptide hybrid block copolymers are discussed. The second part highlights several of the interesting properties and potential applications of this class of materials.
Most peptide/protein–synthetic polymer conjugates are prepared in homogeneous solutions. The synthesis of such hybrid block copolymers can be performed with both convergent and divergent strategies. Although the majority of the peptide/protein–synthetic polymer hybrids are obtained via convergent synthetic pathways, there are also a few examples known in which divergent approaches have been employed. In this article, several convergent and divergent routes are described.
Among the different convergent approaches that have been developed, chemoselective ligation strategies have found the most widespread use. Chemoselective ligation involves the regiospecific coupling of an unprotected peptide or protein with a synthetic polymer and requires a suitable combination of electrophilic and nucleophilic groups on the fragments that are to be coupled. Because of their high selectivity, these conjugation reactions allow precise control over the polymer/protein ratio and the site of conjugation, are compatible with a wide range of functional groups, can tolerate both organic and aqueous solvents, and do not require protective group strategies. If the conjugation site is properly chosen, chemoselective ligation does not affect the protein structure or function. For reasons that will become obvious in the next part of this article, poly(ethylene glycol) (PEG) has received particular attention for the preparation of synthetic polymer–protein hybrids. As a result, strategies to site-selectively modify proteins with PEG are relatively well established.6 Although they have been mainly used for protein PEG-ylation, this does not imply that they cannot be adapted to site-selectively conjugate other synthetic polymers.
The method most widely used for site-selective protein PEG-ylation involves the modification of the thiol group in the side chain of free cysteine residues. The success of this strategy lies in the relatively low natural abundance of cysteine and the ability of the thiol group to specifically react with a variety of reagents. Cysteine only amounts to approximately 1.7% of all amino acid residues in globular proteins. In proteins that lack cysteine, one or more residues can usually be easily introduced with genetic engineering techniques. Consequently, the number and position of the PEG chains that are to be introduced can be rather precisely controlled. The thiol group of cysteine residues can undergo Michael-type addition reactions with maleimide or vinyl sulfone residues and also reacts selectively with iodoacetamide or orthopyridyl disulfide groups.6 This is illustrated in Scheme 1, which shows the reaction of various monomethoxy poly(ethylene glycol) (mPEG) derivatives with thiol-functionalized peptides or proteins (HS-R). These conjugation reactions can be performed under mild conditions in aqueous solutions and do not require protective group chemistry. The unique reactivity of the cysteine sulfhydryl group has not just been used for protein PEG-ylation. Maleimide-functionalized7 and vinyl sulfone functionalized8–10 acrylamide-based copolymers as well as maleimide-end-functionalized polystyrene11 have also been successfully conjugated to various proteins with Michael-type addition chemistry.
An alternative method for site-selective PEG-ylation involves the reductive alkylation of a protein's N-terminal amine group with PEG-aldehyde derivatives (Scheme 2).6, 12 This reaction proceeds through a Schiff base intermediate, which is reduced in situ to a secondary amine linkage in the presence of sodium cyanoborohydride (NaCNBH3). Because of the difference between the pKa values of the N-terminal amine group (∼7.6–8.0) and the amine groups in the side chains of lysine residues (10.0–10.2), this reaction can be optimized to maximize the selectivity for N-terminal PEG-ylation by the reaction being performed under slightly acidic conditions (pH 5).
A final interesting example of a convergent chemoselective PEG-ylation strategy uses the enzyme transglutaminase (TGase).13 This enzyme catalyzes the acyl-transfer reaction between the γ-carboxyamide group of glutamine (Gln) residues in proteins and alkylamines (Scheme 3). The TGase-catalyzed PEG-ylation can be carried out under very mild conditions (pH 7.5) and leads to exclusive modification of the substrate's Gln residues. PEG-ylation of human interleukin 2 with this procedure was not found to lead to a decrease in bioactivity.
Although many proteins are exclusively composed of amino acids, there are a considerable number of proteins that consist of two components: an apoprotein (the polypeptide part) and a cofactor, which can be a metal ion such as copper or iron or various types of organic molecules. Cofactors are usually of great importance to protein function. Under certain conditions, the cofactors of a number of cofactor-bearing proteins such as hemoproteins, flavoenzymes, nicotinamide adenine dinucleotide (NADH), and pyrroloquinoline quinone (PQQ)-dependent enzymes can be extracted from the active pocket, chemically modified, and subsequently recombined with the enzyme (Fig. 1).14 This is a commonly used technique to study cofactor–apoprotein interactions at the molecular level. Cofactor reconstitution, however, is also a powerful method of functionalizing proteins with long alkyl chains and converting water-soluble proteins into membrane proteins or introducing photosensitizers, which can be used to photomodulate enzyme activity. Nolte et al.15 recently extended the concept of cofactor reconstitution to the preparation of synthetic–protein hybrid block copolymers. Their approach was based on the modification of the ferriprotoporphyrin IX cofactor of the horseradish peroxidase (HRP) enzyme with a single polystyrene chain with a molar mass of approximately 9500 g/mol. The reconstitution of the polystyrene-modified cofactor with apo-HRP successfully yielded an amphiphilic HRP–polystyrene hybrid block copolymer.
In addition to the aforementioned procedures, there are a number of other reactions that have not yet been applied to the site-selective conjugation of proteins and synthetic polymers but that offer great potential for the preparation of biological–synthetic hybrid materials. The Staudinger ligation16 (Scheme 4) and [3 + 2] Cu-catalyzed azide–alkyne cycloaddition17 (click chemistry; Scheme 5) are two examples of such reactions. Both the Staudinger ligation18, 19 and click chemistry20 have already been successfully employed for posttranslational protein engineering and the modification of cell surfaces. In both cases, an azide derivative is reacted with a phosphine or alkyne, respectively. Because methionine residues in proteins can be selectively replaced by azidohomoalanine with genetic engineering techniques, both the Staudinger ligation and click chemistry may be attractive alternative approaches for conjugating azide-modified proteins with phosphine- or alkyne-functionalized synthetic polymers. A third strategy, which is also based on recent developments in genetic engineering, relies on the incorporation of unnatural ketone-functionalized amino acids, such as m-acetyl-L-phenylalanine or p-acetyl-L-phenylalanine, into proteins.21, 22 The keto group is not found in the side chain of natural amino acids and reacts selectively with hydrazides and hydroxylamines in high yields and under mild conditions (both in vitro and in living cells). So far, this method has been successfully used to prepare glycoprotein mimics and fluorescent-dye-labeled proteins (for a specific example, see Scheme 6),21, 22 but it may as well be extended to the synthesis of polymer–protein hybrid block copolymers.
The strategies discussed so far represent convergent routes to site-selectively conjugate a protein or peptide and a synthetic polymer. In principle, however, protein/peptide–synthetic polymer conjugates can also be prepared with divergent approaches. Although they have gained significantly less attention, two divergent approaches that are worth mentioning are presented.
In the early 1970s, Bayer and Mutter24–26 extended the concept of solid-phase peptide synthesis (SPPS), which was introduced roughly 1 decade earlier by Merrifield,23 to what they called the liquid-phase method. In contrast to SPPS, in which the peptide chains are assembled on a swollen crosslinked polystyrene resin, liquid-phase synthesis is carried out in a homogeneous solution on a linear PEG support. After each coupling or deprotection step, the polymer-bound peptide is separated from excess reagents or byproducts via membrane dialysis or precipitation in diethyl ether. This is in contrast to the Merrifield solid-phase technique, which involves peptide synthesis under heterogeneous conditions, which may lead to anomalous reaction kinetics. The liquid-phase method also has the advantage that the progress of the reaction can be easily monitored with standard analytical techniques such as NMR spectroscopy. The liquid-phase method has been extensively explored by the groups of Bayer and Mutter for peptide synthesis and, in a few cases, has also been used for the preparation of PEG–peptide conjugates. In one example, the synthesis of a conjugate of an epitope region of the Hemophilus influenza P6 protein and PEG was described.27 In another example, Mutter et al.28 reported the synthesis and conformational properties of a series of PEG–oligoglycine block copolymers. Apart from these examples, the liquid-phase method has not found widespread application for the synthesis of synthetic polymer–peptide/protein hybrids. One possible reason for this could be that it is difficult to find suitable combinations of solvent and nonsolvent for polymers other than PEG. Another important drawback of the liquid-phase method is that it only works as long as the solubility properties of the conjugate are determined by the synthetic polymer support; this places restrictions on the nature and maximum length of the peptide fragments that can be attached.
A very interesting divergent approach to the synthesis of protein/peptide–synthetic polymer hybrids was published recently by Börner et al.29 In contrast to the convergent strategies discussed earlier, which involve the grafting of PEG onto a protein or peptide, or the divergent liquid-phase method, in which the peptide is assembled in a stepwise fashion from a soluble synthetic polymer support, the method introduced by Börner et al. is based on the controlled radical polymerization of vinyl monomers from a well-defined peptide fragment containing an appropriate initiator group. As shown in Scheme 7, the preparation of the hybrid block copolymers starts with solid-phase synthesis of the desired peptide fragment followed by on-bead acylation of the N-terminal glycine residue with 2-bromopropionic acid. Alkyl 2-bromopropionates are widely used as initiators for the atom transfer radical polymerization (ATRP) of acrylate monomers. In the next step, the oligopeptide initiator is carefully cleaved from the resin to avoid side-chain deprotection. In combination with the N,N,N′,N′,N″-pentamethyldiethylenetriamine (PMDETA)/CuBr/CuBr2 catalyst system, the oligopeptide-based initiator has been successfully used to initiate ATRP of n-butyl acrylate, yielding an oligopeptide–poly(n-butyl acrylate) block copolymer with a number-average molecular weight (Mn) of approximately 10.000 g/mol and a polydispersity (weight-average molecular weight/number-average molecular weight) of approximately 1.19.29 Although the published example uses only a relatively small oligopeptide initiator and the polymerization of n-butyl acrylate is carried out in dimethyl sulfoxide as the solvent, this approach seems very promising and may be further extended to the modification of proteins with the retention of structure and function, as well. The basic tools for this are available. Genetic engineering techniques can be used to site specifically modify proteins with cysteine residues or certain unnatural amino acids that allow the posttranslational introduction of functional groups, which could act as polymerization initiators. In addition, over the past years, a number of controlled ATRP and reversible addition–fragmentation chain transfer based polymerization processes have been developed that can be carried out in aqueous solutions.30, 31 It will be exciting to see whether in the near future these developments can be combined and allow the grafting of synthetic polymer segments from posttranslationally modified protein macroinitiators.
Solid-phase synthesis offers some distinct advantages over syntheses carried out in homogeneous solutions. Generally, reactions on solid supports can be driven to very high conversions because a large excess of reagents can be used and each reaction step is followed by a simple filtration-and-washing step to remove excess reagents and byproducts. Furthermore, because each reaction step is amenable to automation, solid-phase synthesis can be carried out on automated synthesizers. Solid-phase synthesis was developed in 1963 by Merrifield specifically for peptide synthesis.23 Over the past decades, solid-phase synthesis has matured and is now also routinely used for the synthesis of other biomolecules, such as nucleotides and saccharides, and it has found widespread application in (combinatorial) organic synthesis. Although the number of examples is still rather limited, solid-phase synthesis has also found first applications in the synthesis of synthetic polymer–peptide conjugates. Two conceptually different approaches for the solid-phase synthesis of synthetic polymer–peptide hybrids can be distinguished: (1) a grafting-onto or convergent strategy and (2) a grafting-from or divergent strategy. Although most of the examples that are discussed in this article focus on PEG-ylation, the conjugation chemistry that is involved may be equally applicable to other synthetic polymers.
The grafting-onto strategy was first introduced by Lu and Felix at Hoffmann-La Roche in the early 1990s.32 At that time, protein PEG-ylation had already received considerable attention, but attempts to prepare PEG-ylated peptides were rare. Peptides, nevertheless, are ideal substrates for PEG-ylation because solid-phase synthesis allows precise control over the extent and position of PEG-ylation through the choice of appropriate orthogonal protective groups. The original procedure involved 9-fluorenylmethoxycarbonyl (Fmoc) solid-phase synthesis of the peptide segment on a (2,4-dimethoxybenzyl)phenoxy acetic acid (DPA) modified methylbenzhydrylamine (MBHA) resin followed by benzotriazole-1-yloxytris(dimethylamino)phosphonium hexafluorophosphate (BOP) mediated acylation of the N-terminal amine group with an amide-linked mPEG norleucine conjugate (Mn = 2000 or 5000 g/mol) (Scheme 8).32 After resin cleavage and side-chain deprotection, the crude product was purified by means of reverse-phase high-performance liquid chromatography (HPLC) or dialysis to remove unreacted peptide. The success of N-terminal PEG-ylation depends both on the molecular weight of the PEG-ylating agent and on the N-terminal amino acid.33 PEG-ylation of unhindered N-terminal amino acids such as glycine could be carried out successfully with carboxylic acid functionalized mPEG derivatives with molar masses of 750, 2000, and 5000 g/mol. At a molar mass of 10,000 g/mol, however, the PEG-ylation reaction could not be driven to completion. PEG-ylation of a hindered N-terminal amino acid such as isoleucine only proceeded smoothly with carboxylic acid functionalized mPEG derivatives with molar masses of 750 and 2000 g/mol and was incomplete with PEG-ylating agents of high molar masses.33
The convergent solid-phase PEG-ylation strategy outlined in Scheme 8 was also extended to allow site-selective C-terminal and side-chain PEG-ylation.34 Site-selective C-terminal PEG-ylation was achieved by the attachment of a PEG-functionalized amino acid as a first residue to the solid support, followed by standard Fmoc solid-phase assembly of the residual peptide sequence. For the site-selective PEG-ylation of lysine and aspartic acid side chains, Lu and Felix34 developed two routes (Scheme 9). The first route (A) involves the incorporation of the appropriate PEG-ylated amino acid derivative during SPPS. The second route (B) is a solid-phase post-PEG-ylation procedure. In this case, orthogonally protected Nϵ-allyloxycarbonyl lysine or β-allyl aspartate residues are introduced at the desired PEG-ylation site in the peptide sequence during Fmoc solid-phase synthesis. When the assembly of the peptide sequence is completed, the alloc/allyl ester protective groups are selectively removed, and the amino acid side chains are modified with the appropriate mPEG amino acid conjugates.
Following the strategies outlined in Schemes 8 and 9, Lu and Felix33 were also able to prepare N-terminal/C-terminal and side-chain/C-terminal di-PEG-ylated peptides. This was found to work well for PEG-ylation agents with molecular weights up to 2000 g/mol. Attempts to prepare multiple-PEG-ylated peptides with higher molar mass PEG-ylating agents were unsuccessful.
Recently, we35 and others36 have prepared block copolymers of coiled-coil peptide sequences and PEG with procedures based on the strategies developed by Felix and Lu. In our own work, the Fmoc solid-phase synthesis of the desired peptide sequence on a Rink amide resin was followed by on-bead benzotriazole-1-yloxytris(pyrrolidino)phosphosium hexafluorophosphate mediated acylation of the N-terminal glycine residue with carboxylic acid functionalized mPEG derivatives with molecular weights of 750 and 2000 g/mol.35 Although PEG-ylation was found to proceed quantitatively for the lowest molar mass PEG-ylating agent, only approximately 40% of the peptide segments were modified when the PEG-ylation was carried out with the high-molar-mass PEG derivative. Kopeček et al.36 followed a slightly different approach. In this case, peptide synthesis was performed on a 2-chlorotrityl resin. First, Fmoc solid-phase fragment condensation was used to prepare coiled-coil peptide sequences that contained up to six [VSSLESK] heptad repeats. Subsequently, PEG-ylation was carried out by the treatment of the resin-bound peptide with an appropriate mPEG succinimidyl carbonate (Mn = 2000 g/mol) in the presence of N,N-diisopropylethylamine. Although virtually no unreacted peptide was detected for the shortest coiled-coil sequence, this amount increased to approximately 50% for the longest peptides, presumably because of the increased steric hindrance that hampered the reactivity of the two large species.
Although the convergent or grafting-onto solid-phase approaches outlined previously have found the most widespread use for the preparation of synthetic polymer–peptide conjugates, several divergent or grafting-from solid-phase strategies have also been reported. One divergent strategy involves the use of the so-called Tentagel PAP resin as a support for SPPS (Scheme 10). This resin consists of a lightly crosslinked polystyrene resin matrix onto which PEG is attached via a labile benzyl ether linkage that is sensitive to, for example, 100% CF3COOH. The synthesis of the PEG–peptide conjugates starts with the assembly of the desired peptide sequence on the amino-terminated PEG chain with standard Fmoc chemistry. After the completion of peptide synthesis, the treatment of the resin-bound product with CF3COOH simultaneously removes all side-chain protective groups and cleaves the linker between the polystyrene support and the PEG chains to afford the desired PEG–peptide block copolymer. In one example, this strategy was used for the synthesis of several PEG–lipopeptide conjugates.37 In addition, this strategy has been used to prepare various PEG-based synthetic hybrid block copolymers containing β-strand peptide sequences.38–40 The Tentagel PAP resin can also be used to prepare PEG–peptide–PEG triblock copolymers through the acylation of the N-terminal amino acid residue with an appropriate mPEG derivative before side-chain deprotection and resin cleavage.39
A second divergent strategy that has been recently introduced involves the solid-phase synthesis of the desired peptide sequence followed by controlled radical polymerization with the resin-bound peptide as the initiator.41, 42 A general overview of this strategy is presented in Scheme 11. Following this route, Wooley et al.41 first assembled an oligopeptide representing the protein transduction domain (PTD) of the HIV-1 TAT protein and extended this sequence with four additional glycine repeats with standard Fmoc solid-phase chemistry. Next, the N-terminal amine group of the resin-bound peptide was functionalized with a fluorinated alkoxyamine derivative, which could act as an initiator for nitroxide-mediated radical polymerizations (NMPs). The resin-supported peptide initiator was successfully used to generate a PTD–poly(tert-butyl acrylate) diblock copolymer and a PTD–poly(tert-butyl acrylate)–poly(methyl acrylate) triblock copolymer. Subsequent resin cleavage simultaneously removed the tert-butyl ester protective group and afforded the corresponding PTD–poly(acrylic acid) and PTD–poly(acrylic acid)–poly(methyl acrylate) block copolymers.41 Washburn et al.42 pursued a similar approach, but these authors explored ATRP instead of NMP to graft the synthetic polymer segment from a resin-supported peptide initiator. To this end, the N-terminus of a resin-bound GRGDS peptide was modified with 2-bromopropionylbromide after the completion of peptide assembly. The solid-supported secondary alkyl bromide was used as an initiator for the ATRP of 2-hydroxyethyl methacrylate. The polymerization was carried out in methyl ethyl ketone/1-propanol (70/30) and used the CuCl/bipyridyl catalyst system. Resin cleavage with a trifluoroacetic acid (TFA)/triisobutylsilane/water mixture (95/2.5/2.5) finally afforded the desired GRGDS–poly(2-hydroxyethyl methacrylate) block copolymer. The degree of polymerization of the poly(2-hydroxyethyl methacrylate) block was determined with 13C NMR experiments on the resin-bound block copolymer and was estimated to 34 ± 3.42
Although the preceding paragraphs have extensively discussed the utility and advantages of solid-phase synthesis for the preparation of peptide–synthetic polymer conjugates, it is also important to point out the drawbacks and restrictions of this strategy. The possibility of using a large excess of amino acids and reagents promotes very high yields in each of the reaction steps. However, even though the yields can be very high, they are generally not quantitative. Even when each of the coupling and deprotection reactions proceeds with 98% conversion, an oligopeptide composed of 20 amino acids, whose synthesis requires 20 coupling and 20 deprotection reactions, can only be obtained in a yield of at most (0.98)40 = 45%. This clearly illustrates that with increasing chain length the maximum achievable yields rapidly decrease, and as a result, solid-phase synthesis is practically most useful for the preparation of peptides containing at most 20–30 amino acid residues.
The aforementioned examples also show that there are two different possibilities for constructing the peptide segment of the hybrid block copolymers via solid-phase synthesis. The first is to prepare the desired peptide segment directly on the resin, followed by functionalization of the N-terminus with an appropriate activated synthetic polymer derivative or with a functional group that can initiate, for example, a controlled radical polymerization. The second option is to use a resin that is functionalized with the synthetic polymer of interest (e.g., a Tentagel PAP resin) and to assemble the peptide sequence at the end of the synthetic polymer chain. According to which of these routes is chosen, the purification of the crude hybrid block copolymer may prove to be relatively straightforward or more difficult. If SPPS includes a capping step after each amino acid coupling, on-bead N-terminal modification followed by resin cleavage and side-chain deprotection yields the desired hybrid block copolymer contaminated with defect peptide sequences. Because of the presence of the synthetic polymer segment, the difference in the molar masses of the desired product and the impurities is usually relatively large, and purification can be simply performed, for example, via dialysis. However, if the peptide is assembled on a Tentagel PAP type resin, the crude reaction product obtained after resin cleavage contains the desired block copolymer together with defect block copolymers that lack one or more amino acid residues. In this case, the difference in the molecular weight between the desired product and the impurities is not very pronounced. As a result, dialysis is not very likely to be successful, and purification may be more difficult and require, for example, HPLC.
PROPERTIES AND APPLICATIONS
Of all synthetic polymer–protein hybrid block copolymers, conjugates of PEG and pharmaceutically active peptides, proteins, or antibodies have attracted most attention. Peptide-, protein-, or antibody-based drugs can have several limitations, including a short plasma half-life, poor stability, and immunogenicity.43, 44 Already in the 1970s, it was discovered that these problems could be significantly reduced by the conjugation of PEG to the peptide, protein, or antibody of interest.45–47 PEG-ylation leads to increased solubility and stability (e.g., against enzymatic degradation), reduces immunogenicity, and results in increased plasma half-life. The last is due to the fact that PEG-ylation prevents rapid renal clearance of the proteins and also hampers their uptake by cells of the reticuloendothelial system. This is of course very attractive to patients because it requires less frequent dosing and promotes compliance. It has been proposed that the highly flexible nature of PEG and extensive hydration are important factors that contribute to these benefits. Interactions of PEG-ylated proteins with other proteins or cells would simultaneously involve the displacement of the hydrated water molecules, which is energetically unfavorable, and the compression of the PEG chains, which is entropically unfavorable. The enhanced plasma lifetime of PEG-ylated proteins also provides an opportunity to direct anticancer agents specifically to tumor tissue without the need for specific targeting ligands. This passive mode of targeting, which was first discovered by Maeda48, 49 in the early 1980s and is now called the enhanced permeability and retention (EPR) effect, is based on the tendency of macromolecules of sufficient high molecular weight to preferentially accumulate in tumor tissue. The EPR effect has been explained in terms of the disorganized pathology of tumor tissue with its discontinuous endothelium, which enhances permeation, in combination with the lack of effective lymphatic drainage, which enhances the retention of macromolecular substances. Table 1 provides a summary of several PEG–protein conjugates that are already on the market or are currently being studied in clinical trials.43, 44 With the exception of the PEG–granulocyte colony-stimulating factor (G-CSF) and PEG–anti-tumour necrosis factor (TNF) Fab conjugates, all other examples in Table 1 are prepared via non-site-selective PEG-ylation procedures. For some of these, reduced activities of the PEG-ylated proteins in comparison with the non-PEG-ylated native proteins have been reported. For PEG–anti-TNF Fab, however, PEG-ylation was found to have no negative effect on protein activity,55 and this illustrates the importance of site-selective PEG-ylation.
Table 1. Summary of several PEG–Protein Conjugate Therapeutics on the Market or Currently in Clinical Trials (Adapted from Ref. 44)
Enzyme Recovery, Affinity Separation, and Immunoassays
The conjugation of a synthetic polymer that possesses temperature-dependent solubility properties [i.e., which exhibits lower critical solution temperature (LCST) or upper critical solution temperature behavior] to a protein allows temperature-induced and reversible precipitation and solubilization of the polymer–protein conjugate. Among the different temperature-sensitive polymers that are known, acrylamide-based (co)polymers, which display LCST behavior, have been most widely studied.2 The possibility of reversibly precipitating and solubilizing polymer–protein conjugates has been used, for example, to separate enzymes from reaction solutions; in this way, the isolation of the reaction product and the recycling of the enzyme are facilitated [Fig. 2(a)]. Temperature-sensitive polymer–protein conjugates have also been used for affinity separation processes in which the protein of interest is first recognized and bound by the polymer–protein conjugate and is subsequently isolated by a temperature-induced precipitation of the affinity complex from solution [Fig. 2(b)]. This concept can be further extended to allow selective isolation and assaying of analytes from complex serum samples [Fig. 2(c)]. In this case, first a synthetic polymer–protein conjugate is added to the analyte solution, and it is followed by the introduction of a second, labeled antibody and a temperature-induced precipitation step. After the isolation of the precipitated complex of the polymer–protein conjugate, analyte, and antibody, excess antibody is removed in a washing step, and subsequently the analyte is assayed by redissolution of the complex in a (cold) solvent.
Most of the applications of synthetic polymer–protein conjugates for enzyme recovery, affinity separation, and immunoassays, which have been discussed previously, have involved the use of random polymer–protein conjugates, that is, conjugates in which the extent of conjugation, the site of conjugation, or both are not precisely defined. These nonselective conjugation strategies may lead to functionalization of the protein close to the active center or a binding site and influence biological activity. Some examples have been reported that indeed describe a loss of protein activity upon the random conjugation of a thermosensitive synthetic polymer.56 Although genetic engineering techniques in combination with site-selective conjugation strategies offer the potential to circumvent these problems, the number of examples in which this has been successfully demonstrated is still small. In one study, Hoffman and coworkers2, 7 site-selectively conjugated a genetically engineered mutant of cytochrome b5, which contained a unique cysteine residue, to a maleimide-terminated oligo(N-isopropylacrylamide). It was found that conjugation did not significantly affect the properties of the protein.
Modulation of Binding and Recognition Properties
Over the past decade or so, Hoffman et al. have impressively demonstrated the power of site-selective conjugation to modulate the binding and recognition properties of proteins.2 In one of their earliest examples, a genetically modified mutant of streptavidin, in which a cysteine residue was introduced near the outer edge of the biotin binding pocket, was conjugated with a vinyl sulfone terminated thermosensitive poly(N-isopropylacrylamide) (PNIPAAm) chain.8 Below the LCST of the PNIPAAm, the biotin binding properties of the polymer–protein conjugate were identical to those of the unmodified mutant or the wild-type streptavidin. At temperatures above the LCST of the PNIPAAm, however, an 84% decrease in the biotin binding capacity was observed, which was ascribed to the collapse of the PNIPAAm chain and the consequent blocking of the biotin binding site.8 A more recent article described the conjugation of a different thermosensitive polymer, poly(N,N-diethylacrylamide) (PDEAAm), to streptavidin at a position slightly further away from the binding pocket.57 In this case, the expansion of the PDEAAm chain below its LCST prevented the binding of biotinylated proteins. Above the LCST, however, the PDEAAm collapses, and this exposes the biotin binding site and allows the binding of biotinylated proteins.
In addition to temperature, light can also be used as a stimulus to modulate the properties of synthetic polymer–protein hybrids. The groups of Hoffman and Stayton9 have explored this concept by conjugating two different light-responsive vinyl sulfone-terminated polymers to a cysteine-modified streptavidin mutant. One of the polymers was a copolymer of N,N-dimethylacrylamide (DMA) and 4-phenylazophenylacrylate, which was soluble at 40 °C under UV irradiation but precipitated under visible-light irradiation. The second copolymer, which was composed of DMA and N-4-phenylazophenylacrylamide (AZAAm), was soluble at 40 °C upon irradiation with visible light, and precipitated upon exposure to UV irradiation. The biotin binding properties of these polymer–protein conjugates were found to correlate with the solubility characteristics of the appended polymers. Biotin binding was possible when the conjugated polymers were in their soluble state and was strongly hampered when the conjugated polymer collapsed and blocked the biotin binding site. The vinyl sulfone terminated DMA–AZAAm copolymer was also successfully conjugated to a cysteine-modified mutant of the enzyme endoglucanase EG 12A.10 Under irradiation with visible light, the soluble polymer chain did not alter substrate binding, and the catalytic activity of the polymer–enzyme conjugate was similar to that of the unmodified control enzyme. However, upon UV irradiation, the DMA–AZAAm copolymer collapsed and blocked substrate access; this was reflected in a strong decrease in the catalytic activity.
Self-Assembly and Supramolecular Architectures
A fourth reason that has driven part of the current interest in synthetic polymer–protein hybrids is related to the very specific self-assembly properties that can be programmed in the primary structure of the peptide segment. The self-assembly of common low-molecular-weight amphiphiles or high-molecular-weight block copolymers is generally driven by unspecific interactions. In the case of synthetic polymer–protein hybrids, however, this process is mediated by the very specific folding and organization properties of the peptide sequence, which may allow access to complex, hierarchically organized, and potentially functional materials that are difficult to generate from purely abiotic building blocks.58, 59 This concept has been pursued by a number of research groups, and the activities have essentially concentrated on two peptide motifs: β strands and coiled-coil-forming amino acid sequences. One of the first examples of a β-strand–synthetic polymer hybrid was reported by Meredith et al.,38 who investigated the dilute-solution aggregation properties of a diblock copolymer of the central domain of the β-amyloid peptide and PEG. In their studies, these authors found that the core of the fibrils formed by the block copolymer resembled the aggregates formed by β-amyloid. It was suggested that the solubility of the aggregates formed by the block copolymer could facilitate the characterization of the different steps in fibrillogenesis.60, 61 The self-assembly properties of hybrid diblock and triblock copolymers based on PEG and different β-strand peptide sequences have also been the subject of several more recent publications.39, 40 These studies confirmed the ability of the peptide segment to direct the self-assembly of the hybrid block copolymers into fibrils or hierarchically organized lamellar superstructures in solution and in the solid state, respectively.
In contrast to β strands, which intrinsically have the propensity to form infinite strands via lateral aggregation, coiled-coil peptide sequences fold and organize in discrete tertiary structures. The coiled coil, which is among the most frequently encountered folding motifs found in nature, is a superhelical structure that is composed of two to five α helices closely wrapped around each other.62, 63 Recently, we and others have explored the coiled-coil motif to direct the self-assembly of PEG-based hybrid block copolymers.35, 36 Circular dichroism and analytical ultracentrifugation experiments have indicated that the coiled-coil sequences mediate the formation of essentially uniform nanoobjects, in which a core composed of a discrete number of α-helical peptides is closely encapsulated in a PEG shell.64 A particularly interesting feature of these hybrid block copolymers is that the aggregation number and pH sensitivity of the nanoobjects can be controlled via directed single and double mutations of the peptide primary structure.65 This may be relevant in light of possible biomedical applications because preliminary experiments have shown that the biological properties of the PEG-ylated coiled coils correlate to their self-assembly behavior.
A final example that is worth mentioning is the work by Nolte et al.,11 who studied the self-assembly properties of a synthetic polymer–protein hybrid composed of a single polystyrene chain with a degree of polymerization of approximately 40 and the lipase B from Candida antarctica (CAL B), which has a molecular weight of approximately 40 g/mol. This hybrid block copolymer can be considered a giant amphiphile in which the polystyrene chain represents the hydrophobic tail and the enzyme acts as the hydrophilic head group. These giant amphiphiles were found to self-assemble in well-defined micrometer-long fibers that consisted of micellar rods (Fig. 3). Interestingly, the self-assembly properties of these giant amphiphiles seemed to be similar to those of their low-molar-mass counterparts.
The conjugation of biomacromolecules and synthetic polymers is an attractive strategy for preparing novel materials with interesting properties. The covalent combination of a biological polymer and a synthetic polymer can lead to materials that synergetically combine the properties of the individual components and overcome several of their limitations. This article has specifically focused on hybrid block copolymers composed of a peptide or protein segment and a synthetic polymer. Although the random conjugation of a peptide or protein and a synthetic polymer may have a negative effect on the material properties, the site-selective conjugation techniques that have been developed over the past years allow precise control over the protein/polymer ratio and the site of conjugation without the properties of the polymer being influenced. The conjugation of peptides and proteins with synthetic polymers has been demonstrated to be an effective strategy to control protein recognition and binding properties and also can be used to direct the self-assembly of synthetic polymers. Peptide/protein–synthetic hybrid polymers already find use as medicines and also show great potential for applications ranging from bioanalysis to bioseparations. The development of novel chemoselective reactions based on, for example, the Staudinger ligation or click chemistry, the availability of novel controlled radical polymerization techniques that can be carried out in aqueous solutions and tolerate a wide range of functional groups, and advances in genetic engineering that make it possible to introduce an increasing number of unnatural amino acids to proteins will further expand the range of biological–synthetic hybrid block copolymers that can be prepared. It will be exciting to see how these synthetic advances influence the development of novel biological–synthetic hybrid block copolymers and open the way to novel applications.
Harm-Anton Klok was born in 1971 and studied chemical technology at the University of Twente (Enschede, The Netherlands) from 1989 to 1993. He received his Ph.D. degree in 1997 from the University of Ulm (Germany) after working with M. Möller. After postdoctoral research with D. N. Reinhoudt (University of Twente) and S. I. Stupp (University of Illinois at Urbana–Champaign, Urbana, IL), he joined the Max Planck Institute for Polymer Research (Mainz, Germany) in early 1999 as a project leader in the group of K. Müllen. In November 2002, he was appointed to the faculty of the Swiss Federal Institute of Technology (Lausanne, Switzerland). There he is currently heading the Polymer Laboratory in the Institute of Materials.