Refinement of Docked Protein-Protein Complexes Using Repulsive Scaling Replica Exchange Simulations.

Accurate prediction and evaluation of protein-protein complex structures is of major importance to understand the cellular interactome. Predicted complex structures based on deep learning approaches or traditional docking methods require often structural refinement and rescoring for realistic evaluation. Standard molecular dynamics (MD) simulations are time-consuming and often do not structurally improve docking solutions. Better refinement can be achieved with our recently developed replica-exchange-based scheme employing different levels of repulsive biasing between proteins in each replica simulation (RS-REMD). The bias acts specifically on the intermolecular interactions based on an increase in effective pairwise van der Waals radii without changing interactions within each protein or with the solvent. It allows for an improvement of the predicted protein-protein complex structure and simultaneous realistic free energy scoring of protein-protein complexes. The setup of RS-REMD simulations is described in detail including the application on two examples (all necessary scripts and input files can be obtained from https://gitlab.com/TillCyrill/mmib ).

MISATO: machine learning dataset of protein-ligand complexes for structure-based drug discovery.

Large language models have greatly enhanced our ability to understand biology and chemistry, yet robust methods for structure-based drug discovery, quantum chemistry and structural biology are still sparse. Precise biomolecule-ligand interaction datasets are urgently needed for large language models. To address this, we present MISATO, a dataset that combines quantum mechanical properties of small molecules and associated molecular dynamics simulations of ~20,000 experimental protein-ligand complexes with extensive validation of experimental data. Starting from the existing experimental structures, semi-empirical quantum mechanics was used to systematically refine these structures. A large collection of molecular dynamics traces of protein-ligand complexes in explicit water is included, accumulating over 170 µs. We give examples of machine learning (ML) baseline models proving an improvement of accuracy by employing our data. An easy entry point for ML experts is provided to enable the next generation of drug discovery artificial intelligence models.

Despite the odds: formation of the SARS-CoV-2 methylation complex.

Coronaviruses modify their single-stranded RNA genome with a methylated cap during replication to mimic the eukaryotic mRNAs. The capping process is initiated by several nonstructural proteins (nsp) encoded in the viral genome. The methylation is performed by two methyltransferases, nsp14 and nsp16, while nsp10 acts as a co-factor to both. Additionally, nsp14 carries an exonuclease domain which operates in the proofreading system during RNA replication of the viral genome. Both nsp14 and nsp16 were reported to independently bind nsp10, but the available structural information suggests that the concomitant interaction between these three proteins would be impossible due to steric clashes. Here, we show that nsp14, nsp10, and nsp16 can form a heterotrimer complex upon significant allosteric change. This interaction is expected to encourage the formation of mature capped viral mRNA, modulating nsp14's exonuclease activity, and protecting the viral RNA. Our findings show that nsp14 is amenable to allosteric regulation and may serve as a novel target for therapeutic approaches.

Assembly-dependent Structure Formation Shapes Human Interleukin-23 versus Interleukin-12 Secretion.

Interleukin 12 (IL-12) family cytokines connect the innate and adaptive branches of the immune system and regulate immune responses. A unique characteristic of this family is that each member is anα:ßheterodimer. For human αsubunits it has been shown that they depend on theirßsubunit for structure formation and secretion from cells. Since subunits are shared within the family and IL-12 as well as IL-23 use the same ßsubunit, subunit competition may influence cytokine secretion and thus downstream immunological functions. Here, we rationally design a folding-competent human IL-23α subunit that does not depend on itsßsubunit for structure formation. This engineered variant still forms a functional heterodimeric cytokine but shows less chaperone dependency and stronger affinity in assembly with its ßsubunit. It forms IL-23 more efficiently than its natural counterpart, skewing the balance of IL-12 and IL-23 towards more IL-23 formation. Together, our study shows that folding-competent human IL-12 familyαsubunits are obtainable by only few mutations and compatible with assembly and function of the cytokine. These findings might suggest that human α subunits have evolved for assembly-dependent folding to maintain and regulate correct IL-12 family member ratios in the light of subunit competition.

Extrinsic stabilization of antiviral ACE2-Fc fusion proteins targeting SARS-CoV-2.

The angiotensin-converting enzyme 2 (ACE2) is a viral receptor used by sarbecoviruses to infect cells. Fusion proteins comprising extracellular ACE2 domains and the Fc part of immunoglobulins exhibit high virus neutralization efficiency, but the structure and stability of these molecules are poorly understood. We show that although the hinge between the ACE2 and the IgG4-Fc is highly flexible, the conformational dynamics of the two ACE2 domains is restricted by their association. Interestingly, the conformational stability of the ACE2 moiety is much lower than that of the Fc part. We found that chemical compounds binding to ACE2, such as DX600 and MLN4760, can be used to strongly increase the thermal stability of the ACE2 by different mechanisms. Together, our findings reveal a general concept for stabilizing the labile receptor segments of therapeutic antiviral fusion proteins by chemical compounds.

Small molecule mediated inhibition of protein cargo recognition by peroxisomal transport receptor PEX5 is toxic to Trypanosoma.

Trypanosomiases are life-threatening infections of humans and livestock, and novel effective therapeutic approaches are needed. Trypanosoma compartmentalize glycolysis into specialized organelles termed glycosomes. Most of the trypanosomal glycolytic enzymes harbor a peroxisomal targeting signal-1 (PTS1) which is recognized by the soluble receptor PEX5 to facilitate docking and translocation of the cargo into the glycosomal lumen. Given its pivotal role in the glycosomal protein import, the PEX5-PTS1 interaction represents a potential target to inhibit import of glycolytic enzymes and thus kill the parasite. We developed a fluorescence polarization (FP)-based assay for monitoring the PEX5-PTS1 interaction and performed a High Throughput Screening (HTS) campaign to identify small molecule inhibitors of the interaction. Six of the identified hits passed orthogonal selection criteria and were found to inhibit parasite growth in cell culture. Our results validate PEX5 as a target for small molecule inhibitors and provide scaffolds suitable for further pre-clinical development of novel trypanocidal compounds.

Explicit solvent repulsive scaling replica exchange molecular dynamics (RS-REMD) in molecular modeling of protein-glycosaminoglycan complexes.

Glycosaminoglcyans (GAGs), linear anionic periodic polysaccharides, are crucial for many biologically relevant functions in the extracellular matrix. By interacting with proteins GAGs mediate processes such as cancer development, cell proliferation and the onset of neurodegenerative diseases. Despite this eminent importance of GAGs, they still represent a limited focus for the computational community in comparison to other classes of biomolecules. Therefore, there is a lack of modeling tools designed specifically for docking GAGs. One has to rely on existing docking software developed mostly for small drug molecules substantially differing from GAGs in their basic physico-chemical properties. In this study, we present an updated protocol for docking GAGs based on the Repulsive Scaling Replica Exchange Molecular Dynamics (RS-REMD) that includes explicit solvent description. The use of this water model improved docking performance both in terms of its accuracy and speed. This method represents a significant computational progress in GAG-related research.

Evaluation of replica exchange with repulsive scaling approach for docking glycosaminoglycans.

Glycosaminoglycans (GAGs), long linear periodic anionic polysaccharides, are key molecules in the extracellular matrix (ECM). Therefore, deciphering their role in the biologically relevant context is important for fundamental understanding of the processes ongoing in ECM and for establishing new strategies in the regenerative medicine. Although GAGs represent a number of computational challenges, molecular docking is a powerful tool for analysis of their interactions. Despite the recent development of GAG-specific docking approaches, there is plenty of room for improvement. Here, replica exchange molecular dynamics with repulsive scaling (REMD-RS) recently proved to be a successful approach for protein-protein complexes, was applied to dock GAGs. In this method, effective pairwise radii are increased in different Hamiltonian replicas. REMD-RS is shown to be an attractive alternative to classical docking approaches for GAGs. This work contributes to setting up of GAG-specific computational protocols and provides new insights into the nature of these biological systems.

Efficient Refinement and Free Energy Scoring of Predicted Protein-Protein Complexes Using Replica Exchange with Repulsive Scaling.

Accurate prediction and evaluation of protein-protein complex structures are of major importance to understand the cellular interactome. Typically, putative complexes are predicted based on docking methods, and simple force field or knowledge-based scoring functions are applied to evaluate single complex structures. We have extended a replica-exchange-based scheme employing different levels of a repulsive biasing between partners in each replica simulation (RS-REMD) to simultaneously refine and score protein-protein complexes. The bias acts specifically on the intermolecular interactions based on an increase in effective pairwise van der Waals radii (repulsive scaling (RS)-REMD) without affecting interactions within each protein or with the solvent. The method provides a free energy score that correlates quite well with experimental binding free energies on a set of 36 complexes with correlation coefficients of 0.77 and 0.55 in explicit and implicit solvent simulations, respectively. For a large set of docked decoy complexes, significant improvement of docked complexes was found in many cases with the starting structure in the vicinity (within 20 Å) of the native complex. In the majority of cases (14 out of 20 in explicit solvent), near native docking solutions were identified as the best scoring complexes. The approach is computational demanding but may offer a route for refinement and realistic ranking of predicted protein-protein docking geometries.

Prediction of protein-protein complexes using replica exchange with repulsive scaling.

The realistic prediction of protein-protein complex structures is import to ultimately model the interaction of all proteins in a cell and for the design of new protein-protein interactions. In principle, molecular dynamics (MD) simulations allow one to follow the association process under realistic conditions including full partner flexibility and surrounding solvent. However, due to the many local binding energy minima at the surface of protein partners, MD simulations are frequently trapped for long times in transient association states. We have designed a replica-exchange based scheme employing different levels of a repulsive biasing between partners in each replica simulation. The bias acts only on intermolecular interactions based on an increase in effective pairwise van der Waals radii (repulsive scaling (RS)-REMD) without affecting interactions within each protein or with the solvent. For a set of five protein test cases (out of six) the RS-REMD technique allowed the sampling of near-native complex structures even when starting from the opposide site with respect to the native binding site for one partner. Using the same start structures and same computational demand regular MD simulations sampled near native complex structures only for one case. The method showed also improved results for the refinement of docked structures in the vicinity of the native binding geometry compared to regular MD refinement.

Evaluation of Predicted Protein-Protein Complexes by Binding Free Energy Simulations.

The accurate prediction of protein-protein complex geometries is of major importance to ultimately model the complete interactome of interacting proteins in a cell. A major bottleneck is the realistic free energy evaluation of predicted docked structures. Typically, simple scoring functions applied to single-complex structures are employed that neglect conformational entropy and often solvent effects completely. The binding free energy of a predicted protein-protein complex can, however, be calculated using umbrella sampling (US) along a predefined dissociation/association coordinate of a complex. We employed atomistic US-molecular dynamics simulations including appropriate conformational and axial restraints and an implicit generalized Born solvent model to calculate binding free energies of a large set of docked decoys for 20 different complexes. Free energies associated with the restraints were calculated separately. In principle, the approach includes all energetic and entropic contributions to the binding process. The evaluation of docked complexes based on binding free energy calculation was in better agreement with experiment compared to a simple scoring based on energy minimization or MD refinement using exactly the same force field description. Even calculated absolute binding free energies of structures close to the native binding geometry showed a reasonable correlation to experiment. However, still for a number of complexes docked decoys of lower free energy than near-native geometries were found indicating inaccuracies in the force field or the implicit solvent model. Although time consuming the approach may open up a new route for realistic ranking of predicted geometries based on calculated free energy of binding.

Origin of Ion Specificity of Telomeric DNA G-Quadruplexes Investigated by Free-Energy Simulations.

Telomeric DNA consists of tandem repeats of the sequence d(TTAGGG) that form G-quadruplex structures made of stacked guanines with monovalent cations bound at a central cavity. Although different ions can stabilize a G-quadruplex structure, the preferred bound ions are typically K+ or Na+. Several different strand-folding topologies have been reported for Q-quadruplexes formed from telomeric repeats depending on DNA length and ion solution condition. This suggests a possible dependence of the ion selectivity of the central pore on the folding topology of the quadruplex. Molecular dynamics free energy perturbation has been employed to systematically study the relative affinity of the central quadruplex pore for different cation types and the associated energetic and solvation contributions to ion selectivity. The calculations have been performed on two different common quadruplex folding topologies. For both topologies, the same ion selectivity was found with a preference for K+ followed by Rb+ and Na+, which agrees with the experimentally determined preference for most investigated quadruplexes. The selectivity is determined by a balance between attractive Coulomb interactions and loss of hydration but also modulated by van der Waals contributions. Specificity is mediated by the central guanines and no significant contribution of the nucleic acid backbone. The simulations indicate that different topologies might be stabilized by ions bound at the surface or alternative sites of the quadruplex because the ion specificity of the central pore does not depend on the strand folding topology.