Protein-Protein Docking in Drug Design and Discovery.

Protein-protein docking is an in silico method to predict the formation of protein complexes. Due to the limited computational resources, the protein docking approach has been originated and developed under the assumption of rigid docking, in which one of the two protein partners remains rigid during the protein associations. The rigid docking has successfully predicted structures of various protein complexes, but often failed if the proteins acquire conformational changes or are driven by influences of other factors (e.g. solvent) while interacting. Formulation of the initial rigid docking contains two main stages: (1) searching for all possible surface matches on a rotational and translational sampling space and (2) ranking those possible solutions to distinguish the correct predictions by locating them on the top high ranks. To obtain better results of the protein rigid docking, one can improve it by optimizing the solutions in the search space or by developing more effective ranking methodology to discriminate the correct predictions from the incorrect or false positive ones. However, while development and improvement are reported in the searching stage, it seems to-date that most initial docking algorithms find it difficult or even fail to locate successfully the correct predictions apart from the others in the ranking stage, especially for Antigen/Antibody complexes. To tackle this issue, a new energy-based scoring function is proposed in this research, namely IFACEwat, to re-rank the results of an initial rigid docking algorithm and therefore further improve the discrimination of the near-native structures from the other false positives. Unlike other re-ranking techniques, in which the solvent effect is ignored or implicitly presented, the IFACEwat implements interfacial water into the protein interfaces and explicitly reflects the water influences on the protein interactions. The IFACEwat was implemented based on the interface Atomic Contact Energy (IFACE) of the initial rigid docking algorithm ZDOCK3.0.2 and the derived energies of water-involved interactions at the protein interfaces. The IFACEwat therefore not only takes advantages of shape complementarity from the initial rigid docking algorithm for protein recognitions but also accounts for the water-mediated contacts during the protein associations. Evaluated for various types of protein complexes, the IFACEwat both increases the numbers of the near-native structures and improves their ranks (i.e. most of them ranked at the top-1) as compared to the original rigid docking. In fact, it achieves a success rate of 83.8% for Antigen/Antibody complexes, which is 10% better than ZDOCK3.0.2. As compared to another re-ranking technique ZRANK, the IFACEwat obtains success rates of 92.3% (8% better) and 90% (5% better) respectively for medium and difficult cases. When comparing with the latest published re-ranking method in F2Dock, it performs equivalently well or better for several Antigen/Antibody complexes. In conclusion, with the inclusion of interfacial water, the IFACEwat improves significantly the results of the initial rigid docking, especially for Antigen/Antibody complexes. The improvement is achieved by explicitly accounting the contribution of water during the protein interactions, which was ignored or not fully presented by the initial rigid docking and other re-ranking techniques. In addition, the IFACEwat maintains the sufficient computational efficiency of the initial docking algorithm, yet improves the ranks as well as the number of the near-native structures found.

Domain-Based Protein Docking with Extremely Large Conformational Changes

Native top-down mass spectrometry (nTDMS) has emerged as a powerful structural biology tool that can localize post-translational modifications (PTMs), explore ligand-binding interactions, and elucidate the three-dimensional structure of proteins and protein complexes in the gas-phase. Fourier-transform ion cyclotron resonance (FTICR) MS offers distinct capabilities for nTDMS, owing to its ultrahigh resolving power, mass accuracy, and robust fragmentation techniques. Previous nTDMS studies using FTICR have mainly been applied to overexpressed recombinant proteins and protein complexes. Here, we report the first nTDMS study that directly analyzes human heart tissue lysate by direct infusion FTICR MS without prior chromatographic separation strategies. We have achieved comprehensive nTDMS characterization of cardiac contractile proteins that play critical roles in heart contraction and relaxation. Specifically, our results reveal structural insights into ventricular myosin light chain 2 (MLC-2v), ventricular myosin light chain 1 (MLC-1v), and alpha-tropomyosin (α-Tpm) in the sarcomere, the basic contractile unit of cardiac muscle. Furthermore, we verified the calcium (Ca2+) binding domain in MLC-2v. In summary, our nTDMS platform extends the application of FTICR MS to directly characterize the structure, PTMs, and metal-binding of endogenous proteins from heart tissue lysate without prior separation methods.

Cooperative interactions in protein-protein interfaces demonstrate the interdependency or the linked network-like behavior of interface interactions and their effect on the coupling of proteins. Cooperative interactions also could cause ripple or allosteric effects at a distance in protein-protein interfaces. Although they are critically important in protein-protein interfaces it is challenging to determine which amino acid pair interactions are cooperative. In this work we have used Bayesian network modeling, an interpretable machine learning method, combined with molecular dynamics trajectories to identify the residue pairs that show high cooperativity and their allosteric effect in the interface of G protein-coupled receptor (GPCR) complexes with G proteins. Our results reveal a strong co-dependency in the formation of interface GPCR:G protein contacts. This observation indicates that cooperativity of GPCR:G protein interactions is necessary for the coupling and selectivity of G proteins and is thus critical for receptor function. We have identified subnetworks containing polar and hydrophobic interactions that are common among multiple GPCRs coupling to different G protein subtypes (Gs, Gi and Gq). These common subnetworks along with G protein-specific subnetworks together confer selectivity to the G protein coupling. This work underscores the potential of data-driven Bayesian network modeling in elucidating the intricate dependencies and selectivity determinants in GPCR:G protein complexes, offering valuable insights into the dynamic nature of these essential cellular signaling components.

G protein-coupled receptors (GPCRs) are a key drug target class. They account for over one-third of current pharmaceuticals, and both drugs that inhibit and promote receptor function are important therapeutically; in some cases, the same GPCR can be targeted with agonists and inhibitors, depending upon disease context. There have been major breakthroughs in understanding GPCR structure and drug binding through advances in X-ray crystallography, and membrane protein stabilization. Nonetheless, these structures have predominately been of inactive receptors bound to inhibitors. Efforts to capture structures of fully active GPCRs, in particular those in complex with the canonical, physiological transducer G protein, have been limited via this approach. Very recently, advances in cryo-electron microscopy have provided access to agonist:GPCR:G protein complex structures. These promise to revolutionize our understanding of GPCR:G protein engagement and provide insight into mechanisms of efficacy and coupling selectivity and how these might be controlled by biased agonists. Here we review what we have currently learned from the new GPCR:Gs and GPCR:Gi/o complex structures.

The simultaneous activation of many distinct G protein-coupled receptors (GPCRs) and heterotrimeric G proteins play a major role in various pathological conditions. Pan-inhibition of GPCR signaling by small molecules thus represents a novel strategy to treat various diseases. To better understand such therapeutic approach, we have characterized the biomolecular target of BIM-46187, a small molecule pan-inhibitor of GPCR signaling. Combining bioluminescence and fluorescence resonance energy transfer techniques in living cells as well as in reconstituted receptor-G protein complexes, we observed that, by direct binding to the Galpha subunit, BIM-46187 prevents the conformational changes of the receptor-G protein complex associated with GPCR activation. Such a binding prevents the proper interaction of receptors with the G protein heterotrimer and inhibits the agonist-promoted GDP/GTP exchange. These observations bring further evidence that inhibiting G protein activation through direct binding to the Galpha subunit is feasible and should constitute a new strategy for therapeutic intervention.

The current model for cell-to-cell movement of plant viruses holds that transport requires virus-encoded movement proteins that intimately associate with endoplasmic reticulum membranes. We have examined the early stages of the integration into endoplasmic reticulum membranes of a double-spanning viral movement protein using photocross-linking. We have discovered that this process is cotranslational and proceeds in a signal recognition particle-dependent manner. In addition, nascent chain photocross-linking to Sec61alpha and translocating chain-associated membrane protein reveal that viral membrane protein insertion takes place via the translocon, as with most eukaryotic membrane proteins, but that the two transmembrane segments of the viral protein leave the translocon and enter the lipid bilayer together.

Of 12 naturally occurring, activating mutations in the seven-transmembrane (7TM) domain of the human Ca2+ receptor (CaR) identified previously in subjects with autosomal dominant hypocalcemia (ADH), five appear at the junction of TM helices 6 and 7 between residue Ile819 and Glu837. After identifying a sixth activating mutation in this region, V836L, in an ADH patient, we studied the remaining residues in this region to determine whether they are potential sites for activating mutations. Alanine-scanning mutagenesis revealed five additional residues in this region that when substituted by alanine led to CaR activation. We also found that, whereas E837A did not activate the receptor, E837D and E837K mutations did. Thus, region Ile819-Glu837 of the 7TM domain represents a "hot spot" for naturally occurring, activating mutations of the receptor, and most of the residues in this region apparently maintain the 7TM domain in its inactive configuration. Unique among the residues in this region, Pro823, which is highly conserved in family 3 of the G protein-coupled receptors, when mutated to either alanine or glycine, despite good expression severely impaired CaR activation by Ca2+. Both the P823A mutation and NPS 2143, a negative allosteric modulator that acts on the 7TM through a critical interaction with Glu837, blocked activation of the CaR by various ADH mutations. These results suggest that the 7TM domain region Ile819-Glu837 plays a key role in CaR activation by Ca2+. The implications of our finding that NPS 2143 corrects the molecular defect of ADH mutations for treatment of this disease are also discussed.

Advances in template-based protein docking by utilizing interfaces towards completing structural interactome

Concerted interactions between all the cell components form the basis of biological processes. Protein-protein interactions (PPIs) constitute a tremendous part of this interaction network. Deeper insight into PPIs can help us better understand numerous diseases and lead to the development of new diagnostic and therapeutic strategies. PPI interfaces, until recently, were considered undruggable. However, it is now believed that the interfaces contain "hot spots," which could be targeted by small molecules. Such a strategy would require high-quality structural data of PPIs, which are difficult to obtain experimentally. Therefore, in silico modeling can complement or be an alternative to in vitro approaches. There are several computational methods for analyzing the structural data of the binding partners and modeling of the protein-protein dimer/oligomer structure. The major problem with in silico structure prediction of protein assemblies is obtaining sufficient sampling of protein dynamics. One of the methods that can take protein flexibility and the effects of the environment into account is Molecular Dynamics (MD). While sampling of the whole protein-protein association process with plain MD would be computationally expensive, there are several strategies to harness the method to PPI studies while maintaining reasonable use of resources. This chapter reviews known applications of MD in the PPI investigation workflows.

Protein-Protein Docking: From Interaction to Interactome

Proteins perform many important functions in our body. Influencing the behaviour of proteins helps to control the spread of disease. Since a protein's function is directed by the molecules with which it binds, the ultimate goal of drug design is to design molecules that selectively inhibit or activate particular proteins and hence modulate their function. Identifying small molecules that modulate protein-protein interactions is a major challenge of drug discovery, to the extent that such targets have been labelled undrugable. Protein-protein interactions involve large flat surfaces usually burying greater than 1100Ǻ2 [95]. This differs significantly from traditional drug targets that contain small molecule binding cavities. However, a minor fraction of the interface residues within protein-protein interactions account for the majority of free energy of binding [28]. Such 'hot spots' appear to be common to protein-protein interactions, and tend to be clustered together at the center of the interface [19]. If the hot spots consist of short continuous binding domains then the discovery of mimetics is feasible, as exemplified by the small molecule integrin inhibitors [16]. However, most protein-protein interfaces consist of noncontinuous binding epitopes, and little is known about the common structure (if any) of protein hot-spots and protein recognition surfaces in general. Molecular recognition is a surface phenomena and this surface is dictated by both the chemistry and topology of exposed functional groups. Thus, protein recognition is dictated by a combination of the 20 naturally occurring amino acids found on the binding surface and how these side chains are topologicals arranged. Drug design techniques are focussed on the discovery of suitable scaffolds and functional group attachments to present the required electrostatic and steric surface for binding to a target and inducing a functional response. To focus molecular design towards the discovery of molecules that modulate protein-protein interactions, we have developed algorithms and tools to cluster protein contact surfaces and have identified common side chain positions of proteins involved in molecular recognition events. These populated shapes are found at the centre of the interface suggesting that they may describe the geometries of protein hot-spots. Consequently using these common motifs in drug design will result in the discovery of scaffolds that mimic the shape of functional regions of protein surfaces, and appending suitable functionality to these scaffolds should result in the discovery of molecules to mimic the function of proteins. To achieve this, we have developed a Population Based Incremental Learning algorithm [13] for efficient conformational searching of potential scaffolds. This algorithm performs strongly in a number of respects, including its ability to calculate the global minimum energy conformation of large flexible molecules and, more importantly, to determine in a short period a large number of low energy conformations of rigid drug-like molecules. This then allows the creation of virtual libraries of molecules which are then filtered through the common motifs to identify appropriate scaffolds. Finally, this process is illustrated by presentation of scaffolds that match common protein surfaces.

Orexin G protein-coupled receptors (OxRs) and their cognate agonists have been implicated in a number of disorders since their recent discovery, ranging from narcolepsy to formation of addictive behavior. Bioluminescence resonance energy transfer assays of agonist-occupied OxRs provided evidence for a strong dose-dependent interaction with both trafficking proteins β-arrestin 1 and 2 that required unusually high agonist concentrations compared with inositol phosphate signaling. This appears to be reflected in functional differences in potency with respect to orexin A (OxA) and OxR2-dependent ERK1/2 phosphorylation after 90 min compared with 2 min, potentially consistent with β-arrestin-mediated versus G protein-mediated signaling, respectively. Furthermore, extended bioluminescence resonance energy transfer kinetic data monitoring OxA-dependent receptor-β-arrestin and β-arrestin-ubiquitin proximity suggested subtype-specific differences in receptor trafficking, with OxR2 activation resulting in more sustained receptor-β-arrestin-ubiquitin complex formation than elicited by OxR1 activation. Enzyme-linked immunosorbent assay (ELISA) data also revealed that OxR1 underwent significantly more rapid recycling compared with OxR2. Finally, we have observed sustained OxA-dependent ERK1/2 phosphorylation in the presence of OxR2 compared with OxR1. Although both OxR subtypes could be classified as class B receptors for β-arrestin usage based on the initial strength of interaction with both β-arrestins, our temporal profiling revealed tangible differences between OxR subtypes. Consequently, OxR1 appears to fit uneasily into the commonly used β-arrestin classification scheme. More importantly, it is hoped that this improved profiling capability, enabling the subtleties of protein complex formation, stability, and duration to be assessed in live cells, will help unlock the therapeutic potential of targeting these receptors.

Co-evolution has an important function in the evolution of species and it is clearly manifested in certain scenarios such as host–parasite and predator–prey interactions, symbiosis and mutualism. The extrapolation of the concepts and methodologies developed for the study of species co-evolution at the molecular level has prompted the development of a variety of computational methods able to predict protein interactions through the characteristics of co-evolution. Particularly successful have been those methods that predict interactions at the genomic level based on the detection of pairs of protein families with similar evolutionary histories (similarity of phylogenetic trees: mirrortree). Future advances in this field will require a better understanding of the molecular basis of the co-evolution of protein families. Thus, it will be important to decipher the molecular mechanisms underlying the similarity observed in phylogenetic trees of interacting proteins, distinguishing direct specific molecular interactions from other general functional constraints. In particular, it will be important to separate the effects of physical interactions within protein complexes (‘co-adaptation') from other forces that, in a less specific way, can also create general patterns of co-evolution.

Understanding protein-protein interactions (PPIs) is quite important to elucidate crucial biological processes and even design compounds that interfere with PPIs with pharmaceutical significance. Protein-protein docking can afford the atomic structural details of protein-protein complexes, but the accurate prediction of the three-dimensional structures for protein-protein systems is still notoriously difficult due in part to the lack of an ideal scoring function for protein-protein docking. Compared with most scoring functions used in protein-protein docking, the Molecular Mechanics/Generalized Born Surface Area (MM/GBSA) and Molecular Mechanics/Poisson Boltzmann Surface Area (MM/PBSA) methodologies are more theoretically rigorous, but their overall performance for the predictions of binding affinities and binding poses for protein-protein systems has not been systematically evaluated. In this study, we first evaluated the performance of MM/PBSA and MM/GBSA to predict the binding affinities for 46 protein-protein complexes. On the whole, different force fields, solvation models, and interior dielectric constants have obvious impacts on the prediction accuracy of MM/GBSA and MM/PBSA. The MM/GBSA calculations based on the ff02 force field, the GB model developed by Onufriev et al. and a low interior dielectric constant (εin = 1) yield the best correlation between the predicted binding affinities and the experimental data (rp = -0.647), which is better than MM/PBSA (rp = -0.523) and a number of empirical scoring functions used in protein-protein docking (rp = -0.141 to -0.529). Then, we examined the capability of MM/GBSA to identify the possible near-native binding structures from the decoys generated by ZDOCK for 43 protein-protein systems. The results illustrate that the MM/GBSA rescoring has better capability to distinguish the correct binding structures from the decoys than the ZDOCK scoring. Besides, the optimal interior dielectric constant of MM/GBSA for re-ranking docking poses may be determined by analyzing the characteristics of protein-protein binding interfaces. Considering the relatively high prediction accuracy and low computational cost, MM/GBSA may be a good choice for predicting the binding affinities and identifying correct binding structures for protein-protein systems.