Investigation of noncovalent interactions between peptides with potential intrinsic sequence patterns by mass...

The conformation of a protein largely depends on the interactions between peptides. Specific and intrinsic sequence peptide patterns, such as DNA double helix backbones, may be present in proteins. A computational statistical deep learning method has supported this assumption, but it has not been experimentally proven. Mass spectrometry, as a fast and accurate experimental method, could be used to evaluate the interaction of biomolecules. The results would be of great value for further study of the mechanism of protein folding.

Several potential intrinsic peptides were chosen by the deep learning method, including seven groups of pentapeptides and five groups of nonapeptides. The noncovalent interactions between mixed polypeptides were investigated by electrospray ionization mass spectrometry (ESI-MS) in full-scan and collision-induced dissociation (CID) modes. Molecular dynamics and molecular mechanics Poisson-Boltzmann surface area (MD-MM/PBSA) analyses were also performed to support the results.

The ESI-MS spectra showed that 11 of the 12 groups of mixed polypeptides formed binary and ternary complexes with relatively high stability. The binding between nonapeptide groups was stronger than that between pentapeptide groups according to the relative intensity. The binding energies calculated by the MM/PBSA binding energy tool also provided strong evidence for the combination of the complexes. Electrostatic interactions, hydrophobic interactions, and van der Waals forces were thought to stabilize the complexes according to the binding models.

The results implied the formation of stable complexes between polypeptides and identified their noncovalent interactions, proving that specific sequences and combinations with relatively strong binding ability exist in potential intrinsic sequences of peptides in protein structures.