Structural basis of internal peptide recognition by PDZ domains using molecular dynamics simulations

Abstract

PDZ domains are important peptide recognition modules which usually recognize short C-terminal stretches of their interaction partners, but certain PDZ domains can also recognize internal peptides in the interacting proteins. Due to the scarcity of data on internal peptide recognition and lack of understanding of the mechanistic details of internal peptide recognition, identification of PDZ domains capable of recognizing internal peptides has been a difficult task.Since Par-6 PDZ domain can recognize both C-terminal and internal peptides, we have carried out multiple explicit solvent MD simulations of 1 μs duration on free and peptide bound Par-6 PDZ to decipher mechanistic details of internal peptide recognition. These simulations have been analyzed to identify residues which play a crucial role in internal peptide recognition by PDZ domains. Based on the conservation profile of the identified residues, we have predicted 47 human PDZ domains to be capable of recognizing internal peptides in human. We have also investigated how binding of CDC42 to the CRIB domain adjacent to the Par6 PDZ allosterically modulate the peptide recognition by Par6 PDZ. Our MD simulations on CRIB-Par6_PDZ didomain in isolation as well as in complex with CDC42, indicate that in absence of CDC42 the adjacent CRIB domain induces open loop conformation of PDZ facilitating internal peptide recognition. On the other hand, upon binding of CDC42 to the CRIB domain, Par6 PDZ adopts closed loop conformation required for recognition of C-terminus peptides. These results provide atomistic details of how binding of interaction partners onto adjacent domains can allosterically regulate substrate binding to PDZ domains. In summary, MD simulations provide novel insights into the modulation of substrate recognition preference of PDZ by specific peptides, adjacent domains and binding of interaction partners at allosteric sites.

Interconnected Protein Networks: Insights Towards CRIB-Par6 Protein Allostery Through a Graph-Theoretic Analysis

“…PDZ domains are often associated with receptors (e.g., NMDA and AMPA/kainite) as well as membrane channels (e.g., Shaker K1 and TRP Ca21) and are involved in clustering proteins into functional complexes at the plasma membrane (Kong & Karplus, 2009), in addition to sorting, assembling, and anchoring multidomain interaction complexes (van den Berk et al, 2004). Disruption of PDZ mediated interactions is therefore implicated in the onset of a variety of human diseases including neurological disorders and cancer (Sain & Mohanty, 2019). Usually, the PDZ-containing protein has several PDZ domains (up to 10 domains), which can concurrently interact with different protein, and thereby form functional protein complexes (Kong & Karplus, 2009).…”