Predicting the affinity of compstatin peptide with non-natural amino acids to human C3c protein by scoring molecular dynamics simulations

Abstract

ABSTRAC: Peptides are chemical entities composed of natural and non-natural amino acids that have been used successfully as drugs, vaccines, biomarkers, among others. However, these can be easily cleaved and degraded by proteases, where their breaking of a chemical bond in peptides gives smaller molecules or radicals, causing instability in some biological environments when we use peptides therapeutically or as medicines. One possible solution is the use of peptides with non-natural amino acids (NNAA). In the present study, we assessed the prediction of affinities in complexes between human Complement component 3 (C3c) protein bound to multiple compstatin peptide analogs with NNAAs. We used molecular dynamics simulations and six scoring functions to correlate the average score with the experimental binding data obtained from previous studies. Several correlation coefficients above 0.7 and one above 0.85 were detected, indicating an excellent correlation between these two variables. We found the highest Spearman correlation for the Nnscore and Cyscore scoring function, suggesting that these are the most adequate for ranking the binding of modified peptides to a protein target.