SAnDReS 2.0: Development of machine-learning models to explore the scoring function space.

Classical scoring functions may exhibit low accuracy in determining ligand binding affinity for proteins. The availability of both protein-ligand structures and affinity data make it possible to develop machine-learning models focused on specific protein systems with superior predictive performance. Here, we report a new methodology named SAnDReS that combines AutoDock Vina 1.2 with 54 regression methods available in Scikit-Learn to calculate binding affinity based on protein-ligand structures. This approach allows exploration of the scoring function space. SAnDReS generates machine-learning models based on crystal, docked, and AlphaFold-generated structures. As a proof of concept, we examine the performance of SAnDReS-generated models in three case studies. For all three cases, our models outperformed classical scoring functions. Also, SAnDReS-generated models showed predictive performance close to or better than other machine-learning models such as KDEEP, CSM-lig, and ΔVinaRF20. SAnDReS 2.0 is available to download at https://github.com/azevedolab/sandres.

PACSAB: Coarse-Grained Force Field for the Study of Protein-Protein Interactions and Conformational Sampling in Multiprotein Systems.

Molecular dynamics simulations of proteins are usually performed on a single molecule, and coarse-grained protein models are calibrated using single-molecule simulations, therefore ignoring intermolecular interactions. We present here a new coarse-grained force field for the study of many protein systems. The force field, which is implemented in the context of the discrete molecular dynamics algorithm, is able to reproduce the properties of folded and unfolded proteins, in both isolation, complexed forming well-defined quaternary structures, or aggregated, thanks to its proper evaluation of protein-protein interactions. The accuracy and computational efficiency of the method makes it a universal tool for the study of the structure, dynamics, and association/dissociation of proteins.

In silico and in vitro characterization of phospholipase A2 isoforms from soybean (Glycine max).

At the present, no secreted phospholipase A2 (sPLA2) from soybean (Glycine max) was investigated in detail. In this work we identified five sequences of putative secreted sPLA2 from soybean after a BLAST search in G. max database. Sequence analysis showed a conserved PA2c domain bearing the Ca²âº binding loop and the active site motif. All the five mature proteins contain 12 cysteine residues, which are commonly conserved in plant sPLA2s. We propose a phylogenetic tree based on sequence alignment of reported plant sPLA2s including the novel enzymes from G. max. According to PLA2 superfamily, two of G. max sPLA2s are grouped as XIA and the rest of sequences as XIB, on the basis of differences found in their molecular weights and deviating sequences especially in the N- and C-terminal regions of the isoenzymes. Furthermore, we report the cloning, expression and purification of one of the putative isoenzyme denoted as GmsPLA2-XIA-1. We demonstrate that this mature sPLA2 of 114 residues had PLA2 activity on Triton:phospholipid mixed micelles and determine the kinetic parameters for this system. We generate a model based on the known crystal structure of sPLA2 from rice (isoform II), giving first insights into the three-dimensional structure of folded GmsPLA2-XIA-1. Besides describing the spatial arrangement of highly conserved pair HIS-49/ASP-50 and the Ca⁺² loop domains, we propose the putative amino acids involved in the interfacial recognition surface. Additionally, molecular dynamics simulations indicate that calcium ion, besides its key function in the catalytic cycle, plays an important role in the overall stability of GmsPLA2-XIA-1 structure.

Kinetics of lipid-membrane binding and conformational change of L-BABP.

We designed an experimental approach to differentiate the kinetics of protein binding to a lipid membrane from the kinetics of the associated conformational change in the protein. We measured the fluorescence intensity of the single Trp6 in chicken liver bile acid-binding protein (L-BABP) as a function of time after mixing the protein with lipid membranes. We mixed the protein with pure lipid membranes, with lipid membranes in the presence of a soluble quencher, and with lipid membranes containing a fluorescence quencher attached to the lipid polar head group. We fitted simultaneously the experimental curves to a three-state kinetic model. We conclude that in a first step, the binding of L-BABP to the interfacial region of the anionic lipid polar head groups occurred simultaneously with a conformational change to the partly unfolded state. In a second slower step, Trp6 buried within the polar head group region, releasing contacts with the aqueous phase.

Binding and interactions of L-BABP to lipid membranes studied by molecular dynamic simulations.

Chicken liver bile acid-binding protein (L-BABP) is a member of the fatty acid-binding proteins super family. The common fold is a beta-barrel of ten strands capped with a short helix-loop-helix motif called portal region, which is involved in the uptake and release of non-polar ligands. Using multiple-run molecular dynamics simulations we studied the interactions of L-BABP with lipid membranes of anionic and zwitterionic phospholipids. The simulations were in agreement with our experimental observations regarding the electrostatic nature of the binding and the conformational changes of the protein in the membrane. We observed that L-BABP migrated from the initial position in the aqueous bulk phase to the interface of anionic lipid membranes and established contacts with the head groups of phospholipids through the side of the barrel that is opposite to the portal region. The conformational changes in the protein occurred simultaneously with the binding to the membrane. Remarkably, these conformational changes were observed in the portal region which is opposite to the zone where the protein binds directly to the lipids. The protein was oriented with its macrodipole aligned in the configuration of lowest energy within the electric field of the anionic membrane, which indicates the importance of the electrostatic interactions to determine the preferred orientation of the protein. We also identified this electric field as the driving force for the conformational change. For all the members of the fatty acid-binding protein family, the interactions with lipid membranes is a relevant process closely related to the uptake, release and transfer of the ligand. The observations presented here suggest that the ligand transfer might not necessarily occur through the domain that directly interacts with the lipid membrane. The interactions with the membrane electric field that determine orientation and conformational changes described here can also be relevant for other peripheral proteins.

On the Ewald artifacts in computer simulations. The test-case of the octaalanine peptide with charged termini.

The treatment of electrostatic interactions in molecular simulations is of fundamental importance. Ewald and related methods are being increasingly used to the detriment of cutoff schemes, which are known to produce several artifacts. A potential drawback of the Ewald method is the spatial periodicity that is imposed to the system, which could produce artifacts when applied in the simulation of liquids. In this work we analyze the octaalanine peptide with charged termini in explicit solvent, for which severe effects due to the use of Ewald sums were predicted using continuum electrostatics. Molecular Dynamics simulations for a total of 158 nanoseconds were performed in cells of different sizes. From the comparison of the results of different system sizes, no significant periodicity-induced artifacts were observed. It is argued that in current biomolecular simulations, the incomplete sampling is likely to affect the results to a larger extent than the artifacts induced by the use of Ewald sums.

Molecular dynamics simulation study of the interaction of trehalose with lipid membranes.

The interactions of the cryoprotective agent trehalose with a lipid membrane made of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine at 323 K were studied by means of molecular dynamics simulations. It was observed that trehalose binds to the phospholipid headgroups with its main axis parallel to the membrane normal. Trehalose establishes hydrogen bonds with the carbonyl and phosphate groups and replaces water molecules from the lipid headgroup. Notably, the number of hydrogen bonds (HBs) that the membrane made with its environment was conserved after trehalose binding. The HBs between lipid and trehalose have a longer lifetime than those established between lipid and water. The binding of the sugar does not produce changes either in the lipid area or in the lipid order parameter. The effect of trehalose on the dipole potential is in agreement with experimental results. The contribution of the different components to the membrane dipole potential was analyzed. It was observed that the binding of trehalose produces changes in the different components and the sugar itself contributes to the surface potential due to the polarization of its hydroxyl in the interface.