Computational prediction of allosteric nucleic acids
Selected nucleic acid binding species (aptamers) have been shown to undergo conformational changes in the presence of ligands, and have been adapted to function as biosensors. We were interested in whether the secondary structures of aptamers could be rationally engineered to undergo ligand dependent conformational changes. To this end, we used rational and computational design methods to generate a number of aptamer biosensors. First, we built upon previous work that showed that antisense oligonucleotides bearing reporter moieties could be used to denature aptamers. Upon addition of ligands, the conformational equilibrium is shifted towards release of the antisense oligonucleotide and a concomitant increase in fluorescence. We attempted to adapt this format to the potential detection of ricin, but were unsuccessful. In order to better evaluate rational designs, we attempted to use computational modeling methods. Again, aptamer biosensors have previously been engineered based on ligand-induced reorganization of secondary structure (as opposed to oligonucleotide displacement), a so-called 'slip-structure' model. We developed an algorithm to evaluate different lisp structures, predicted both aptamers and aptazymes that should have undergone ligand-dependent changes in conformation, and experimentally evaluated the computationally predicted sequences. A number of robust biosensors that could respond to the cytokine VegF and the small molecule flavin were discovered. The computational model was further adapted to an aptamer biosensor that underwent a larger conformational change upon ligand-binding, an antiswitch. In this model, binding of the ligand stabilizes one hairpin structure at the expense of a competing structure (as opposed to merely changing the register of the hairpin as in the previously described slip structure model). Again, we were able to computationally identify a number of antiswitches that upon synthesis were responsive to the ligand theophylline. Finally we again attempted to use rational design methods to optimize not just the degree of signal but also the kinetic performance of aptamer biosensors. To this end, we developed biosensors that signaled within seconds the presence of the coagulation protein thrombin.