Dendritic and self-assembling linear RGD peptides - from integrin binding to responsive hydrogels

Several studies exist in the literature which utilise either dendritic (covalent) or self-assembling (non-covalent) strategies to achieve multivalent binding to a biological target, but rarely are the two explored together. Herein, we compare and contrast dendritic and self-assembling approaches to organise a multivalent array of ligands to bind the protein, integrin αvβ3. In the first instance, linear RGD (Arg-Gly-Asp) peptides were covalently attached to first and second generation dendritic frameworks, and a positive (monovalent) and negative control were synthesised. A fluorescence polarisation (FP) competition assay was used to quantify the binding. The first generation dendron (2.16) was the most effective binder, with an EC50 of 125 μM (375 μM per peptide unit); significantly better than the monovalent ligand (2.19), the binding of which could not be quantified in our assay, even at 1 mM concentration; whilst the second generation dendron (2.17) was somewhat less effective, indicating that there is an optimum number of ligands that can be displayed before the multiple ligand array becomes disadvantageous to the binding. To explore the non-covalent approach, the linear RGD peptide was conjugated to either a single hydrophobic C12 aliphatic chain (3.2), an aromatic pyrene (3.4), a C22 aliphatic chain (3.6), or two C12 aliphatic chains (3.18), which gave rise to amphiphilic peptides which were able to self-assemble to differing degrees and into markedly different morphologies, as shown by a Nile Red encapsulation study and TEM imaging. Spherical micelles were formed by amphiphiles 3.2 and 3.4, whereas 3.6 and 3.18 produced cylindrical, rod-like micelles. Compounds 3.2 and 3.4 were the most effective integrin binders at concentrations of 200 μM and 110 μM, respectively, whilst 3.6 and 3.18 failed to produce a quantifiable binding concentration. The results therefore show that not only does the micellar self-assembly approach yield multivalent ligand displays with improved efficiency of binding compared with the dendritic method, but the morphology of the self-assembled system can also be detrimental to the recognition of the protein, at least in our FP assay and using purified integrin in solution. Finally, we report on a family of linear RGD peptide conjugate hydrogelators. Of particular interest was the novel bolaamphiphile C12-[urea-RGD]2 (4.4), comprised of linear RGD peptide head groups connected to either end of a hydrophobic C12 aliphatic chain via urea linkages. The molecule undergoes thermoreversible chiral self-assembly in water and generates a sample-spanning nanofibrous gel network, as determined by circular dichroism spectroscopy and TEM/SEM imaging, respectively. Gels were formed at a minimum gel concentration (MGC) of 0.06 wt % (0.6 mg/ml, 0.5 mM), one of the lowest MGCs reported and represents a “super” hydrogel. We report on its responsiveness (breakdown) towards charge dense basic anions such as phosphate and acetate, but its stability in the presence of more charge diffuse halide and nitrate anions. Furthermore, we demonstrate the passive diffusion of encapsulated low MW additives out of the gel phase, whereas high MW, protein-sized molecules remain trapped within the fibrous network.