Lipid Membrane State Change by Catalytic Protonation and the Implications for Synaptic Transmission

In cholinergic synapses, the neurotransmitter acetylcholine (ACh) is rapidly hydrolyzed by esterases to choline and acetic acid (AH). It is believed that this reaction serves the purpose of deactivating ACh once it has exerted its effect on a receptor protein (AChR). The protons liberated in this reaction, however, may by themselves excite the postsynaptic membrane. Herein, we investigated the response of cell membrane models made from phosphatidylcholine (PC), phosphatidylserine (PS) and phosphatidic acid (PA) to ACh in the presence and absence of acetylcholinesterase (AChE). Without a catalyst, there were no significant effects of ACh on the membrane state (lateral pressure change ≤0.5 mN/m). In contrast, strong responses were observed in membranes made from PS and PA when ACh was applied in presence of AChE (>5 mN/m). Control experiments demonstrated that this effect was due to the protonation of lipid headgroups, which is maximal at the pK (for PS: pKCOOH≈5.0; for PA: pKHPO4−≈8.5). These findings are physiologically relevant, because both of these lipids are present in postsynaptic membranes. Furthermore, we discussed evidence which suggests that AChR assembles a lipid-protein interface that is proton-sensitive in the vicinity of pH 7.5. Such a membrane could be excited by hydrolysis of micromolar amounts of ACh. Based on these results, we proposed that cholinergic transmission is due to postsynaptic membrane protonation. Our model will be falsified if cholinergic membranes do not respond to acidification.

DLA-Ranker: Evaluating protein docking conformations with many locally oriented cubes.

Proteins ensure their biological functions by interacting with each other, and with other molecules. Determining the relative position and orientation of protein partners in a complex remains challenging. Here, we address the problem of ranking candidate complex conformations toward identifying near-native conformations. We propose a deep learning approach relying on a local representation of the protein interface with an explicit account of its geometry. We show that the method is able to recognise certain pattern distributions in specific locations of the interface. We compare and combine it with a physics-based scoring function and a statistical pair potential.

Improved protein docking by predicted interface residues.

Scoring docking solutions is a difficult task, and many methods have been developed for this purpose. In docking, only a handful of the hundreds of thousands of models generated by docking algorithms are acceptable, causing difficulties when developing scoring functions. Today’s best scoring functions can significantly increase the number of top-ranked models but still fails for most targets. Here, we examine the possibility of utilising predicted residues on a protein-protein interface to score docking models generated during the scan stage of a docking algorithm. Many methods have been developed to infer the portions of a protein surface that interact with another protein, but most have not been benchmarked using docking algorithms. Different interface prediction methods are systematically tested for scoring >300.000 low-resolution rigid-body template free docking decoys. Overall we find that BIPSPI is the best method to identify interface amino acids and score docking solutions. Further, using BIPSPI provides better docking results than state of the art scoring functions, with >12% of first ranked docking models being acceptable. Additional experiments indicated precision as a high-importance metric when estimating interface prediction quality, focusing on docking constraints production. We also discussed several limitations for the adoption of interface predictions as constraints in a docking protocol.

PIPENN: Protein Interface Prediction with an Ensemble of Neural Nets

Motivation: Protein interactions play an essential role in many biological and cellular processes, such as protein—protein interaction (PPI) in signaling pathways, binding to DNA in transcription, and binding to small molecules in receptor activation or enzymatic activity. Experimental identification of protein binding interface residues is a time-consuming, costly, and challenging task. Several machine learning and other computational approaches exist which predict such interface residues. Here we explore if Deep Learning (DL) can be used effectively for this prediction task, and which learning strategies and architectures may be most efficient. We introduce seven DL architectures that are applied to eleven independent test sets, focused on the residues involved in PPI interfaces and in binding RNA/DNA and small molecule ligands. Results: We constructed a large data set dubbed BioDL, comprising protein-protein interaction data from the PDB and protein-ligand interactions (DNA, RNA and small molecules) from the BioLip database. Additionally, we reused our existing curated homo- and heteromeric PPI data sets. We performed several experiments to assess the impact of different data features, spatial forms, encoding schemes, network initializations, loss functions, regularization mechanisms, and activation functions on the performance of the predictors. Benchmarking the resulting DL models with an independent test set (ZK448) shows no single DL architecture performs best on all instances, but that an ensemble of DL architectures consistently achieves peak prediction performance. Our PIPENN's ensemble predictor outperforms current state-of-the-art sequence-based protein interface predictors on all interaction types, achieving AUCs of 0.718 (protein—protein), 0.823 (protein—nucleotide) and 0.842 (protein—small molecule) respectively. Availability: Source code and data sets at https://github.com/ibivu/

Improved protein docking by predicted interface residues

ABSTRACTScoring docking solutions is a difficult task, and many methods have been developed for this purpose. In docking, only a handful of the hundreds of thousands of models generated by docking algorithms are acceptable, causing difficulties when developing scoring functions. Today’s best scoring functions can significantly increase the number of top-ranked models but still fails for most targets. Here, we examine the possibility of utilising predicted residues on a protein-protein interface to score docking models generated during the scan stage of a docking algorithm. Many methods have been developed to infer the portions of a protein surface that interact with another protein, but most have not been benchmarked using docking algorithms. Different interface prediction methods are systematically tested for scoring >300.000 low-resolution rigid-body template free docking decoys. Overall we find that BIPSPI is the best method to identify interface amino acids and score docking solutions. Further, using BIPSPI provides better docking results than state of the art scoring functions, with >12% of first ranked docking models being acceptable. Additional experiments indicated precision as a high-importance metric when estimating interface prediction quality, focusing on docking constraints production. We also discussed several limitations for the adoption of interface predictions as constraints in a docking protocol.

Rational Design of Constrained Peptides as Protein Interface Inhibitors

The lack of progress in developing targeted therapeutics directed at protein–protein complexes has been due to the absence of well-defined ligand-binding pockets and the extensive intermolecular contacts at the protein–protein interface. Our laboratory has developed approaches to dissect protein–protein complexes focusing on the superfamilies of erbB and tumor necrosis factor (TNF) receptors by the combined use of structural biology and computational biology to facilitate small molecule development. We present a perspective on the development and application of peptide inhibitors as well as immunoadhesins to cell surface receptors performed in our laboratory.

Protein interface redesign facilitates the transformation of nanocage building blocks to 1D and 2D nanomaterials

Although various artificial protein nanoarchitectures have been constructed, controlling the transformation between different protein assemblies has largely been unexplored. Here, we describe an approach to realize the self-assembly transformation of dimeric building blocks by adjusting their geometric arrangement. Thermotoga maritima ferritin (TmFtn) naturally occurs as a dimer; twelve of these dimers interact with each other in a head-to-side manner to generate 24-meric hollow protein nanocage in the presence of Ca2+ or PEG. By tuning two contiguous dimeric proteins to interact in a fully or partially side-by-side fashion through protein interface redesign, we can render the self-assembly transformation of such dimeric building blocks from the protein nanocage to filament, nanorod and nanoribbon in response to multiple external stimuli. We show similar dimeric protein building blocks can generate three kinds of protein materials in a manner that highly resembles natural pentamer building blocks from viral capsids that form different protein assemblies.