Abstract
Ligand-binding aptamers obtained by SELEX (Systematic Evolution of Ligands by EXponential enrichment) often have low affinity or/and specificity, and post-SELEX optimization is usually needed. Due to experimental difficulty in determining three-dimensional (3D) structures of aptamer-ligand complexes, there are few structure-guided methods for rational post-SELEX optimization. Here, we employed a de novo optimization approach to engineer high-affinity variants for a G-quadruplex (GQ) aptamer (GO18-T-d) that specifically binds to marine toxin gonyautoxin 1/4 (GTX1/4). First, temperature-dependent modeling was carried out to predict the atomic structure of GO18-T-d. Then, to identify key bases for the optimization, spontaneous binding simulations were performed to reveal the complex structure of GO18-T-d with GTX1/4. Finally, binding energy analysis was conducted to evaluate the designed variants for high affinity. We predicted that GO18-T-d has the typical parallel GQ topology, consistent with circular dichroism (CD) measurements. Our simulations showed that the 5'-end of GO18-T-d hinders the GTX1/4 movement toward the binding pocket, leading to a designed variant that removes the first 5 nucleotides at the 5'-end. Microscale thermophoresis (MST) experiments verified that the binding affinity of the engineered aptamer increases by ~20 folds. Thus, this study not only provides a high-affinity variant of GO18-T-d, but also suggests that our computational approach is useful for the structure-guided optimization of GQ aptamers.