Protein pKa prediction with machine learning

Protein pKa prediction is essential for the investigation of pH-associated relationship between protein structure and function. In this work, we introduce a deep learning based protein pKa predictor DeepKa, which is trained and validated with the pKa values derived from continuous constant pH molecular dynamics (CpHMD) simulations of 279 soluble proteins. Here the CpHMD implemented in the Amber molecular dynamics package has been employed (Huang, Harris, and Shen J. Chem. Inf. Model. 2018, 58, 1372-1383). Notably, to avoid discontinuities at the boundary, grid charges are proposed to represent protein electrostatics. We show that the prediction accuracy by DeepKa is close to that by CpHMD benchmarking simulations, validating DeepKa as an efficient protein pKa predictor. In addition, the training and validation sets created in this study can be applied to the development of machine learning based protein pKa predictors in future. Finally, the grid charge representation is general and applicable to other topics, such as the protein-ligand binding affinity prediction.

Charge Interactions in a Highly Charge-depleted Protein

ABSTRACTInteractions between charged residues are difficult to study because of the complex network of interactions found in most proteins. We have designed a purposely simple system to investigate this problem by systematically introducing individual and pairs of charged and titratable residues in a protein otherwise free of such residues. We used constant pH molecular dynamics simulations, NMR spectroscopy, and thermodynamic double mutant cycles to probe the structure and energetics of the interaction between the charged residues. We found that the partial burial of surface charges contributes to a shift in pKa value, causing an aspartate to titrate in the neutral pH range. Additionally, the interaction between pairs of residues was found to be highly context dependent, with some pairs having no apparent preferential interaction, while other pairs would engage in coupled titration forming a highly stabilized salt bridge. We find good agreement between experiments and simulations, and use the simulations to rationalize our observations and to provide a detailed mechanistic understanding of the electrostatic interactions.SignificanceElectrostatic forces are important for protein folding and are favored targets of protein engineering. However, despite the many advances in the field of protein electrostatics, the prediction of changes in protein structure and function upon introduction or removal of titratable residues is still complicated. In order to provide a basic understanding of protein electrostatics we here characterize a highly charge-depleted protein and its titratable variants by a combination of NMR spectroscopy and constant pH molecular dynamics simulations. Our investigations reveal how strongly interacting residues engaged in salt bridging, can be characterized. Furthermore, our study may also enrich and facilitate the understanding of dehydration of salt-bridges and its potential effect on protein stability.

Spectral Tuning of Chlorophylls in Proteins – Electrostatics vs. Ring Deformation

In photosynthetic complexes, tuning of chlorophyll light-absorption spectra by the protein environment is crucial to their efficiency and robustness. Water Soluble Chlorophyll-binding Proteins from <i>Brassicaceae</i> (WSCPs) are useful for studying spectral tuning mechanisms due to their symmetric homotetramer structure, the ability to rigorously modify the chlorophyll’s protein surroundings, and the availability of crystal structures. Here, we present a rigorous analysis based on hybrid Quantum Mechanics and Molecular Mechanics simulations with conformational sampling to quantify the relative contributions of steric and electrostatic factors to the absorption spectra of WSCP-chlorophyll complexes. We show that when considering conformational dynamics, chlorophyll ring deformation accounts for about one-third of the spectral shift, whereas protein electrostatics accounts for the remaining two-thirds. From a practical perspective, protein electrostatics is easier to manipulate than chlorophyll conformations, thus, it may be more readily implemented in designing artificial protein-chlorophyll complexes with desired spectral shift.

Spectral Tuning of Chlorophylls in Proteins – Electrostatics vs. Ring Deformation

In photosynthetic complexes, tuning of chlorophyll light-absorption spectra by the protein environment is crucial to their efficiency and robustness. Water Soluble Chlorophyll-binding Proteins from <i>Brassicaceae</i> (WSCPs) are useful for studying spectral tuning mechanisms due to their symmetric homotetramer structure, the ability to rigorously modify the chlorophyll’s protein surroundings, and the availability of crystal structures. Here, we present a rigorous analysis based on hybrid Quantum Mechanics and Molecular Mechanics simulations with conformational sampling to quantify the relative contributions of steric and electrostatic factors to the absorption spectra of WSCP-chlorophyll complexes. We show that when considering conformational dynamics, chlorophyll ring deformation accounts for about one-third of the spectral shift, whereas protein electrostatics accounts for the remaining two-thirds. From a practical perspective, protein electrostatics is easier to manipulate than chlorophyll conformations, thus, it may be more readily implemented in designing artificial protein-chlorophyll complexes with desired spectral shift.

Formation of biomolecular condensates in bacteria by tuning protein electrostatics

Biomolecular condensates provide a strategy for cellular organization without a physical membrane barrier while allowing for dynamic, responsive organization of the cell. To date, very few biomolecular condensates have been identified in prokaryotes, presenting an obstacle to engineering these compartments in bacteria. As a novel strategy for bacterial compartmentalization, protein supercharging and complex coacervation were employed to engineer liquid-like condensates in E. coli. A simple model for the phase separation of supercharged proteins was developed and used to predict intracellular condensate formation. Herein, we demonstrate that GFP-dense condensates formed by expressing GFP variants of sufficient charge in cells are dynamic and enrich specific nucleic acid and protein components. This study provides a fundamental characterization of intracellular phase separation in E. coli driven by protein supercharging and highlights future utility in designing functional synthetic membraneless organelles.