(540c) Computational Discovery and Characterization of Peptide Chains That Recognize the tRNALys3 Anticodon Loop | AIChE

(540c) Computational Discovery and Characterization of Peptide Chains That Recognize the tRNALys3 Anticodon Loop

HIV (human immunodeficiency virus) and AIDS (acquired immunodeficiency syndrome) are responsible for the deaths of over 25 million people. Our efforts are aimed at discovering potential therapies capable of breaking the reverse transcription of HIV. Since HIV replication depends strongly on the tRNALys3 primer, Agris and coworkers have been developing peptide sequences to bind to primer tRNALys3 with a higher affinity and specificity than viral RNA. A short promising peptide sequence, 15 amino acids in length — RVTHHAFLGAHRTVG, was synthesized and tested in their lab. Our goal in this work is to use simulation to: (1) understand the mechanism by which this sequence recognizes and binds to tRNALys3 and (2) design additional sequences with superior affinity and specificity. The AMBER package was used to simulate the binding behavior of tRNALys3 and the 15-amino acid peptide at the atomic level. Our results revealed that the arginine, which is positively charged, on the peptide chain, dominates the binding of the peptide sequence due to its interaction with the negatively-charged phosphate backbone on the tRNA. However this binding is general and does not enhance the specificity that is necessary for tRNALys3 recognition. In contrast the hydrophilic histidine binds selectively to the loop of tRNALys3, conferring specific recognition. The phenylalanine contributes positively to both binding formation and peptide recognition. The loop on  tRNALys3 plays an important energetic role in both the binding affinity and specificity. By calculating and analyzing the binding free energy, we find that the van der Waals interaction is the key to recognizing specific molecules, while the electrostatic interaction essentially determines the strength of the binding affinity. We are developing an algorithm that combines self-consistent mean field (SCMF) with a Monte Carlo (MC) search algorithm in sequence space to determine optimal peptide sequences which can then be tested for binding affinity and specificity  with tRNALys3 in the Agris lab.