Minimalist Model for the Dynamics of Helical Polypeptides: A Statistic-Based Parametrization

Abstract

Low-resolution models are often used to address macroscopic time and size scales in molecular dynamics simulations of biomolecular systems. Coarse graining is often coupled to knowledge-based parametrization to obtain empirical potentials able to reproduce the system thermodynamic behavior. Here, a minimalist coarse grained (GC) model for the helical structures of proteins is reported. A knowledge-based parametrization strategy is coupled to the explicit inclusion of hydrogen-bonding-related terms, resulting in an accurate reproduction of the structure and dynamics of each single helical type, as well as the internal conformational variables correlation. The proposed strategy of basing the force field terms on real physicochemical interactions is transferable to different secondary structures. Thus, this work, though conclusive for helices, is to be considered the first of a series devoted to the application of the knowledge-based, physicochemical model to extended secondary structures and unstructured proteins.

Protein secondary-structure description with a coarse-grained model

“…ScrewFrame may also be used a starting point for the development of minimalistic models for protein structure and dynamics, similar to the worm-like chain model (Doi & Edwards, 1986), which has been successfully applied to DNA (Marko & Siggia, 1995). Our method may also be used to analyze dynamic processes such as the folding and unfolding of peptides (Spampinato & Maccari, 2014) and it can describe the fold of intrinsically disordered proteins.…”

Low-Resolution Models for the Interaction Dynamics of Coated Gold Nanoparticles with β2-microglobulin

“…The protein is an un-branched chain of beads, representing as single amino-acids and placed on the Cαs (Figure 6A). In general, the force field (FF) consists of bonded and non-bonded terms similar to those present in standard atomistic FFs, though using less simple functional forms, parameterized based on a combination of structure-based and statistics-based parameterization [27]. The latest version includes a set of numerical pair potentials (including both the hydrophobic/steric component and electrostatics, and with implicit solvent) for the non-bonded part, capable of reproducing the different amino acids aggregation tendencies [17], even in the presence of other elements, such as cytoplasm “crowders” molecules [28].…”

Evolutionary Switches Structural Transitions via Coarse-Grained Models

“…L being a set of ''local nonbonded contacts,'' typically including the hydrogen bond network, disulfide, and salt bridges or other persistent and strong contacts. u local is the term usually retaining the bias toward the reference structure (Tozzini et al, 2007;Trovato et al, 2013;Spampinato et al, 2014), bringing structural accuracy to the model. The set of local contacts can be defined either geometrically based on a cutoff distance (Di Fenza et al, 2009) or on chemical-physical considerations (Trovato et al, 2013), for example, the known network of H-bonds (Tozzini et al, 2006) and different functional forms were previously used, either generic single well potentials (Di Fenza et al, 2009) or more sophisticated amino acid-dependent potentials including sequence information, electrostatics, hydrophobicity, and desolvation effects (Trovato et al, 2013).…”

“…But at variance with that, we consider a fully reductionist approach keeping only the essential elements and parameters. First, we consider the Ca-only-based representation of the polypeptide, also called ''minimalist'' coarse grained (MCG) (Trovato and Tozzini, 2012), which was demonstrated to be the simplest possible still capable of accounting uniquely of the secondary structure (Spampinato et al, 2014), and allowing the reconstruction of the atomistic backbone structure (Tozzini et al, 2006). This representation displays a great flexibility, being capable of easily passing from structurally biased [Go or network models (Tirion, 1996)] to unbiased (Yap et al, 2008) or partially biased parameterizations (Di Fenza et al, 2009).…”