Understanding protein fouling by resorting to mesoscale modelling
R. Qiu,J. Xiao,X. D. Chen
2018-01-01
Abstract:In-depth molecular scale understanding of protein adsorption on heat exchanger surfaces will inevitably lead to promising anti-fouling strategies. It is however difficult to get sufficient experimental data on protein adsorption for mechanism analyses, especially for the in-process dynamic data. Starting from the simplest biofouling system, in this work, a meso-scale modeling method developed by us was extended to investigate multi-peptide adsorption. For a system with 12-Ala hydrophobic peptides, it was found that only a small proportion of peptide-chains would keep separated while most of them would tend to aggregate together. Compared with the aggregates, those individual peptide-chains have higher flexibilities and lower environmental sensibilities, which means that they could end up with an adsorbed state very quickly. This finding implies that the adsorbed individual peptide-chains might have changed the aggregated water layers adjacent to the solidliquid interface, which leads to barriers to the following aggregates’ adsorption. INTRODUCTION For the pasteurization process in dairy industry, undesired protein based deposits in heat exchangers would reduce the performance of heat transfer (Georgiadis et al., 1998; Jun et al., 2005; Bansal et al., 2006; Fickak et al., 2011; Pelegrine et al., 2012; Jimenez et al., 2013). And most importantly, this kind of undesired deposits have been proved to be pathogens’ excellent growing medium that would bring serious bio-pollution into dairy products (Punidadas et al., 1999; Michalski et al., 1999). A slight adjustment of surrounding environment such as ionic strengths (Christian et al., 2003; Blanpain-Avet et al., 2012; Jimenez et al., 2013) and pH (Pelegrine et al., 2012; Kröner et al., 2013) of the liquid raw material would significantly influence the adsorption rate. It can be attributed to the change of the degree of dissociation and configuration of the solute, i.e., protein in this case. When extended to new biomaterials, all operational parameters have to be resettled (Hanke et al., 2014). Although experimental methods could help us understand the influence of deposit on heat transfer at macroscale, poor understanding of molecular level adsorption mechanism could lead to long lab-working time and high production costs (Hanke et al., 2014). In order to open the ‘black-box’ of adsorption mechanism, microscale computational approach is a great choice. According to the interaction function and force fields, amino acid residues/atoms interact with surroundings via bonded and non-bonded interactions (Miller et al., 2010). Bonded interactions mainly work for the conformation change within one protein/peptide, and influence the interaction with surrounding molecules indirectly. While it is non-bonded potentials that are mainly in charge of reflecting the environment, leading to the adsorption of peptide chains to form fouling layers. Non-bonded potentials are Van der Waals potential and electrostatic potential that determine the residues’ water affinities based on their hydrophobicity (Penna et al., 2014; Pandey et al., 2009). At the same time, the steric structures of amino acid residues’ sidechains also play important roles in residue movements (Yu et al., 2012; Monticelli et al., 2008). Containing residues with tiny chemical groups, such as Alanine residue (Ala) and Glycine residue, a peptide would have higher backbone flexibility, which would increase its mobility in water (Yu et al., 2012; Sakiyama et al., 2006). Considering both interaction and steric influences of ordered amino acid residues, the results from all-atom simulations demonstrated that a single peptide-chain would move from bulk water towards an uncharged solid surface and end up with an adsorbed state via a biased diffusion phase followed by an anchoring phase and a binding phase (Yu et al., 2012; Penna et al., 2014; Qiu et al., 2017). When extended to protein simulation, structure flexibility is usually not taken into account because of the low computational efficiency of all-atom simulation and the insufficient experimental support (Hagiwara et al., 2009). Rigid structure of one single protein was used to study the influence of different charge distributions and specific types of residues’ distributions on its adsorption behavior (Hagiwara et al., 2009). In industrial practice however, fouling is mainly caused by the adsorption of multi-proteins (Bansal et al., 2006; Fickak et al., 2011; Jimenez et al., 2013). Their configurations would definitely change under the influence of the environment (Christian et al., 2003; Blanpain-Avet et al., 2012; Jimenez et al., 2013). Rigid structures of all atom simulation could not deal with this important steric effect. Meanwhile, the simulation case of one single protein/peptide could not represent the real fouling system (Qiu et al., 2017). Heat Exchanger Fouling and Cleaning – 2017 ISBN: 978-0-9984188-0-3; Published online www.heatexchanger-fouling.com 100 In the liquid phase, most single proteins would gather and form much larger aggregates that affect the moving trajectories of each other (Bansal et al., 2006; Blanpain-Avet et al., 2012; Jimenez et al., 2013; Bouvier et al., 2014). These experimental results show the limitation of all-atom method and the requirement in the development of a mesoscale coarse-grained method. By now, only a few papers have reported the simulations of multiple peptides. Pandey et al. (2009, 2010 and 2012) has contributed greatly in the simulation of multiple peptides with a unique all-residue Monte Carlo model. The adsorption density, binding energy and supramolecular assembly of multiple peptides were investigated. The adsorption preferences on different solid surfaces which have been reported in experimental studies were reproduced, which demonstrates the validity of the coarse-grained Monte Carlo model (Pandey et al., 2009; Eby et al., 2010). With further consideration of the root mean square displacement of the center of mass and the radius of gyration for single long peptide-chains, potential applications of this method to much more complex systems have been proposed (Foo et al., 1998, 1997; Pandey et al., 2012; Pandey et al., 2010; Pandey et al., 2009). The present study focuses on quantitative and mechanistic analyses of the adsorption behavior of 12-Ala hydrophobic peptide-chains. This module peptide was designed with the hydrophobic and denatured characters that most proteins have. The studies on peptide aggregation in bulk water and peptide adsorption on the surface can shed light on fundamental fouling mechanisms. MODELING METHODOLOGY In this study, a hybrid coarse-grained lattice Monte Carlo model was used to investigate peptides’ adsorption on Au (111) surface (Qiu et al., 2017). The movement of bead is partially accepted according to the Metropolis Algorithm. This model has been used in our previous research. It can reproduce the water distribution on the liquid-solid interface, and also multiple adsorption phases that were reported in other molecular dynamics simulation results (Qiu et al., 2017). The simulation box of 7.50 × 7.50 × 7.50 nm was constructed from identical cubes, whose sizes were 0.125 × 0.125 × 0.125 nm and could only host one coarse-grained bead to occupy its eight vertexes (Rouault et al., 1995; Carmesin et al., 1988; Xiao et al., 2009). Periodic boundary conditions were applied to the x and y directions. For the vertical direction (z) of the simulation box, within its bottom and top 1.2 nm space (equal to the cut-off radius) were filled with surface beads (providing the solid phase) and unmovable water beads (providing the bulk liquid phase), respectively. The Coarse-Grained (CG) beads, which were coarsegrained based on a four-to-one mapping scheme (Monticelli et al., 2008), can be divided into four types, i.e., polar, charged, uncharged and water beads according to specific distributions of charged sites (Monticelli et al., 2008; Yesylevskyy et al., 2010; Zhang et al., 2011; Jong et al., 2012; Qiu et al., 2017). Because most proteins are overall hydrophobic, a linear 12-Ala (AAAAAAAAAAAA) CG peptide-chain was used in this study. According to the CG map (Jong et al., 2012), the Ala residue is only represented by one P4 type CG bead that could remove the steric influence. Meanwhile, the CG bead type of Ala residues is the same as water beads, which means that the interaction parameters between water-surface and Ala-surface are the same. Without the influence from different potentials and specific sidechain steric hindrances, 12-Ala peptide-chain could allow us to focus more on the process of aggregation and adsorption. Gold surface is the most stable and well-studied metal surface (Heinz et al., 2008; Pandey et al., 2009; Yu et al., 2012; Penna et al., 2014). Using this surface could provide us with model validations and result comparisons (Yu et al., 2012; Penna et al., 2014). The adjusted Au surface structure was the one used in Heinz et al. (2008). The coordinates were (0.00, 0.00, 0.00), (0.125, 0.250, 0.00), (0.00, 0.125, 0.250), (0.125, 0.375, 0.250), (0.250, 0.125, 0.500) and (0.00, 0.375, 0.500) nm in an Au crystal cell with the lattice at (0.250, 0.500, 0.625) nm. The potential of a CG bead U(rij) summarized from all non-bonded interactions among its neighbors within the cutoff radius at 1.20 nm (Equation 1). The bonded interaction among one peptide-chain was replaced by Bond-Fluctuation Model (Carmesin et al., 1988; Xiao et al., 2009). The nonbonded interaction includes Lennard-Johns potential (switched at 0.9 nm) and the electrostatic potential (shifted from 0 nm to rcut-off) whose parameters are referenced from the MARTINI force field with the extension by the LorentzBerthelot rule (Monticelli et al., 2008; Yesylevskyy et al., 2010; Zhang et al., 2011; Jong et al., 2012; Xu et al., 2015) U(rij)=ULJ(rij)+UEle(rij) =4εij [( σij rij ) 12