Hydrogels are widely used as biomaterials for cell culture and tissue engineering. They consist of natural biopolymers (e.g., collagen, elastin, hyaluronic acid) or synthetic polymers (e.g., polyethylene glycol, polylactic acid, polyacrylamide). The mechanical properties of hydrogels, such as stiffness, toughness and stress relaxation, depend primarily on the molecular building blocks and the spatial organization of polymeric network chains and cross-links (network topology). Star-shaped, 4-arm polyethylene glycol (sPEG) is a widely used hydrogel building block. sPEG displays a very low polydispersity and carries exactly one terminal cross-linking site per arm. As a result, almost ideal covalently cross-linked networks have been synthesized from sPEG building blocks. In this thesis, my goal was to replace sPEG with a purely protein-based, 4-arm polymer building block, based on the tetrameric fluorescent protein DsRed and random coil polypeptides (RCPP). This strategy allowed for the recombinant production of fusion proteins consisting of DsRed, RCPP and a covalent or dynamic cross-linking unit thus providing purely protein-based hydrogels. More importantly, it gives access to sequence-controlled RCPPs with an exactly determined number of monomer units per chain (i.e., monodispersity). Employing a modular gene design, the length of the RCPPs can be easily varied and also polymer lengths beyond commercially available sPEGs can easily be obtained. In a first series of experiments, three different variants of DsRed were tested for their suitability to serve as the tetrameric core of the fusion proteins. Key criteria where their expression level in an E. coli expression system as well as their stability in a hydrogel network. The different variants were engineered so that they carry exactly one Cys residue per monomeric unit. The Cys residues were used to cross-link the DsRed proteins with maleimide-terminated sPEG to obtain a hydrogel network. After cross-linking, DsRed becomes the mechanically weakest link in the network and is expected to show force-induced dissociation when the hydrogel experiences strain. To probe possible pulling geometry effects, Cys was introduced either at the C-terminus or at an internal, surface-exposed position. Hydrogels were formed for all tested DsRed variants whereas no cross-linking took place for a Cys-free mutant. A simple stress-strain experiment was performed where the fluorescence signal of DsRed was monitored in parallel. It has been proposed that a shift from red to green fluorescence may occur upon tetramer dissociation; however, no such change was observed over a large range of applied strains. These experiments suggest that DsRed possesses a high stability and is well suited as a tetrameric core. Continuing with one DsRed variant, I genetically fused elastin-like polypeptides (ELPs) of controlled length to the C-terminus of the fluorescent protein. The fusion protein was further extended with either SpyCatcher or SpyTag to test the possibility of using this protein ligation system as a permanent covalent cross-link. I systematically varied the length of the ELP chain and tested the influence of this parameter on the resulting hydrogel properties, in combination with the total protein concentration. Rheology experiments show purely elastic behavior with a stiffness of around 10,000 Pa for the best hydrogel obtained (ELP length of 132 nm and protein concentration of 2 mM). This is a clear improvement over other protein-based hydrogels, which often show a stiffness less than 1000 Pa. The linear viscoelastic region scaled with ELP length, which is a clear indication that the hydrogels are indeed cross-linked via the terminal SpyCatcher/SpyTag fusion partners. Overall, these results show that purely protein-based hydrogels can be synthesized with a high level of molecular control, similar to what has already been obtained with sPEGs in other experiments. For most cell culture applications, dynamically cross-linked hydrogels are of advantage. It has been shown that cells respond to the viscoelastic properties of the hydrogel and the presence of dynamic bonds further allows cell migration and proliferation. In the next step, I thus replaced the SpyCatcher/SpyTag system with reversible coiled coil (CC) cross-links. To prevent hydrogel formation during recombinant protein expression, a heterodimeric CC (AB) was used. In contrast to the SpyCatcher/SpyTag system, expression and purification of these fusions proteins was more challenging. The DsRed-ELP-A fusion protein showed a low expression yield. DsRed-ELP-B had a strong tendency to aggregate, which most likely originated from the tendency of the B sequence to form homodimers. Preliminary experiments, where the B peptide was modified to reduce homodimer stability, did not show a significant improvement. Further optimization of the sequences is thus necessary before a sufficient amount of protein can be obtained for hydrogel synthesis. The above results have shown that ELPs are excellent RCPPs for hydrogel synthesis; however, ELPs also undergo a phase transition in response to salt concentration and temperature. As a possible replacement for ELPs, I tested another protein-based repeat sequence, consisting of serine, alanine and proline. Even though this PAS sequence has already been used as a replacement for PEG in protein therapeutics, its application as a hydrogel building block is novel. Initial expression and purification tests with the fusion proteins DsRed-PAS-A and DsRed-PAS-B also show a low yield, even lower than for the corresponding ELP fusion proteins. Overall, in this thesis I have tested several protein-based modules that can be used as tetrameric cores, random coil polymers and cross-linking modules for the synthesis of protein-based hydrogels. The results obtained from the above work show the potential of protein-based hydrogels but also highlight that it can be a challenge to obtain sufficient concentrations and overall quantities of the required fusions proteins. Despite the clear need for optimization of sequences and procedures, the building blocks introduced and tested will aid the future design of novel molecularly controlled hydrogels. Protein-based hydrogels are powerful candidates to serve as extracellular matrix mimicking, mechano-responsive networks with tunable mechanical properties.
