Molecular dynamics in multidimensional space explains how mutations affect the association path of neomycin to a riboswitch.

Riboswitches are messenger RNA (mRNA) fragments binding specific small molecules to regulate gene expression. A synthetic N1 riboswitch, inserted into yeast mRNA controls the translation of a reporter gene in response to neomycin. However, its regulatory activity is sensitive to single-point RNA mutations, even those distant from the neomycin binding site. While the association paths of neomycin to N1 and its variants remain unknown, recent fluorescence kinetic experiments indicate a two-step process driven by conformational selection. This raises the question of which step is affected by mutations. To address this, we performed all-atom two-dimensional replica-exchange molecular dynamics simulations for N1 and U14C, U14C[Formula: see text], U15A, and A17G mutants, ensuring extensive conformational sampling of both RNA and neomycin. The obtained neomycin association and binding paths, along with multidimensional free-energy profiles, revealed a two-step binding mechanism, consisting of conformational selection and induced fit. Neomycin binds to a preformed N1 conformation upon identifying a stable upper stem and U-turn motif in the riboswitch hairpin. However, the positioning of neomycin in the binding site occurs at different RNA-neomycin distances for each mutant, which may explain their different regulatory activities. The subsequent induced fit arises from the interactions of the neomycin's N3 amino group with RNA, causing the G9 backbone to rearrange. In the A17G mutant, the critical C6-A17/G17 stacking forms at a closer RNA-neomycin distance compared to N1. These findings together with estimated binding free energies coincide with experiments and elucidate why the A17G mutation decreases and U15A enhances N1 activity in response to neomycin.

Coarse-Grained Modeling of Multiple Pathways in Conformational Transitions of Multi-Domain Proteins.

Large-scale conformational transitions in multi-domain proteins are often essential for their functions. To investigate the transitions, it is necessary to explore multiple potential pathways, which involve different intermediate structures. Here, we present a multi-basin (MB) coarse-grained (CG) structure-based Go̅ model for describing transitions in proteins with more than two moving domains. This model is an extension of our dual-basin Go̅ model in which system-dependent parameters are determined systematically using the multistate Bennett acceptance ratio method. In the MB Go̅ model for multi-domain proteins, we assume that intermediate structures may have partial inter-domain native contacts. This approach allows us to search multiple transition pathways that involve distinct intermediate structures using the CG molecular dynamics (MD) simulations. We apply this scheme to an enzyme, adenylate kinase (AdK), which has three major domains and can move along two different pathways. Using the optimized mixing parameters for each pathway, AdK shows frequent transitions between the Open, Closed, and the intermediate basins and samples a wide variety of conformations within each basin. The explored multiple transition pathways could be compared with experimental data and examined in more detail by atomistic MD simulations.

evERdock BAI: Machine-learning-guided selection of protein-protein complex structure.

Computational techniques for accurate and efficient prediction of protein-protein complex structures are widely used for elucidating protein-protein interactions, which play important roles in biological systems. Recently, it has been reported that selecting a structure similar to the native structure among generated structure candidates (decoys) is possible by calculating binding free energies of the decoys based on all-atom molecular dynamics (MD) simulations with explicit solvent and the solution theory in the energy representation, which is called evERdock. A recent version of evERdock achieves a higher-accuracy decoy selection by introducing MD relaxation and multiple MD simulations/energy calculations; however, huge computational cost is required. In this paper, we propose an efficient decoy selection method using evERdock and the best arm identification (BAI) framework, which is one of the techniques of reinforcement learning. The BAI framework realizes an efficient selection by suppressing calculations for nonpromising decoys and preferentially calculating for the promising ones. We evaluate the performance of the proposed method for decoy selection problems of three protein-protein complex systems. Their results show that computational costs are successfully reduced by a factor of 4.05 (in the best case) compared to a standard decoy selection approach without sacrificing accuracy.

Refining evERdock: Improved selection of good protein-protein complex models achieved by MD optimization and use of multiple conformations.

A method for evaluating binding free energy differences of protein-protein complex structures generated by protein docking was recently developed by some of us. The method, termed evERdock, combined short (2 ns) molecular dynamics (MD) simulations in explicit water and solution theory in the energy representation (ER) and succeeded in selecting the near-native complex structures from a set of decoys. In the current work, we performed longer (up to 100 ns) MD simulations before employing ER analysis in order to further refine the structures of the decoy set with improved binding free energies. Moreover, we estimated the binding free energies for each complex structure based on an average value from five individual MD snapshots. After MD simulations, all decoys exhibit a decrease in binding free energy, suggesting that proper equilibration in explicit solvent resulted in more favourably bound complexes. During the MD simulations, non-native structures tend to become unstable and in some cases dissociate, while near-native structures maintain a stable interface. The energies after the MD simulations show an improved correlation between similarity criteria (such as interface root-mean-square distance) to the native (crystal) structure and the binding free energy. In addition, calculated binding free energies show sensitivity to the number of contacts, which was demonstrated to reflect the relative stability of structures at earlier stages of the MD simulation. We therefore conclude that the additional equilibration step along with the use of multiple conformations can make the evERdock scheme more versatile under low computational cost.

Practical Protocols for Efficient Sampling of Kinase-Inhibitor Binding Pathways Using Two-Dimensional Replica-Exchange Molecular Dynamics.

Molecular dynamics (MD) simulations are increasingly used to study various biological processes such as protein folding, conformational changes, and ligand binding. These processes generally involve slow dynamics that occur on the millisecond or longer timescale, which are difficult to simulate by conventional atomistic MD. Recently, we applied a two-dimensional (2D) replica-exchange MD (REMD) method, which combines the generalized replica exchange with solute tempering (gREST) with the replica-exchange umbrella sampling (REUS) in kinase-inhibitor binding simulations, and successfully observed multiple ligand binding/unbinding events. To efficiently apply the gREST/REUS method to other kinase-inhibitor systems, we establish modified, practical protocols with non-trivial simulation parameter tuning. The current gREST/REUS simulation protocols are tested for three kinase-inhibitor systems: c-Src kinase with PP1, c-Src kinase with Dasatinib, and c-Abl kinase with Imatinib. We optimized the definition of kinase-ligand distance as a collective variable (CV), the solute temperatures in gREST, and replica distributions and umbrella forces in the REUS simulations. Also, the initial structures of each replica in the 2D replica space were prepared carefully by pulling each ligand from and toward the protein binding sites for keeping stable kinase conformations. These optimizations were carried out individually in multiple short MD simulations. The current gREST/REUS simulation protocol ensures good random walks in 2D replica spaces, which are required for enhanced sampling of inhibitor dynamics around a target kinase.

Unraveling the Coupling between Conformational Changes and Ligand Binding in Ribose Binding Protein Using Multiscale Molecular Dynamics and Free-Energy Calculations.

Conformational changes of proteins upon ligand binding are usually explained in terms of several mechanisms including the induced fit, conformational selection, or their mixtures. Due to the slow time scales, conventional molecular dynamics (cMD) simulations based on the atomistic models cannot easily simulate the open-to-closed conformational transition in proteins. In our previous study, we have developed an enhanced sampling scheme (generalized replica exchange with solute tempering selected surface charged residues: gREST_SSCR) for multidomain proteins and applied it to ligand-mediated conformational changes in the G134R mutant of ribose-binding protein (RBPG134R) in solution. The free-energy landscape (FEL) of RBPG134R in the presence of a ribose at the binding site included the open and closed states and two intermediates, open-like and closed-like forms. Only the open and open-like forms existed in the FEL without a ribose. In the current study, the coupling between the conformational changes and ligand binding is further investigated using coarse-grained MD, multiple atomistic cMD, and free-energy calculations. The ribose is easily dissociated from the binding site of wild-type RBP and RBPG134R in the cMD simulations starting from the open and open-like forms. In contrast, it is stable at the binding site in the simulations from the closed and closed-like forms. The free-energy calculations provide the binding affinities of different structures, supporting the results of cMD simulations. Importantly, cMD simulations from the closed-like structures reveal transitions toward the closed one in the presence of a bound ribose. On the basis of the computational results, we propose a molecular mechanism in which conformational selection and induced fit happen in the first and second halves of the open-to-closed transition in RBP, respectively.

Visualizing proton antenna in a high-resolution green fluorescent protein structure.

"Proton-collecting antenna" are conjectured to consist of several carboxylates within hydrogen-bond (HB) networks on the surface of proteins, which funnel protons to the orifice of an internal proton wire leading to the protein's active site. Yet such constructions were never directly visualized. Here we report an X-ray structure of green fluorescent protein (GFP) of the highest resolution to date (0.9 A). It allows the identification of some pivotal hydrogen atoms pertinent to uncertainties concerning the protonation state of the chromophore. Applying a computer algorithm for mapping proton wires in proteins reveals the previously observed "active site wire" connecting Glu222 with the surface carboxylate Glu5. In addition, it is now possible to identify what appears to be a proton-collecting apparatus of GFP. It consists of a negative surface patch containing carboxylates, threonines, and water molecules, connected by a HB network to Glu5. Furthermore, we detect exit points via Asn146 and His148 to a hydrophobic surface region. The more extensive HB network of the present structure, as compared with earlier GFP structures, is not accidental. A systematic investigation of over 100 mutants shows a clear correlation between the observed water content of GFP X-ray structures and their resolution. With increasing water content, the proton wires become progressively larger. These findings corroborate the scenario in which the photodissociated proton from wild-type GFP can leak outside, whereafter another proton is recruited via the proton-collecting apparatus reported herein.

Mapping proton wires in proteins: carbonic anhydrase and GFP chromophore biosynthesis.

We have developed an algorithm for mapping proton wires in proteins and applied it to the X-ray structures of human carbonic anhydrase II (CA-II), the green fluorescent protein (GFP), and some of their mutants. For both proteins, we find more extensive proton wires than typically reported. In CA-II the active site wire exits to the protein surface, and leads to Glu69 and Asp72, located on an electronegative patch on the rim of the active site cavity. One possible interpretation of this observation is that positively charged, protonated buffer molecules dock in that area, from which a proton is delivered to the active site when the enzyme works in the dehydration direction. In GFP we find a new internal proton wire, in addition to the previously reported wire involved in excited state proton transfer. The new wire is located on the other face of the chromophore, and we conjecture that it plays a role in chromophore biosynthesis that occurs following protein folding. In the last step of this process, transient carbanion formation was suggested to occur on the bridge carbon [Pouwels et al. Biochemistry 2008, 47, 10111]. Residues on the new wire (Thr62, His181, Arg96) may participate in proton abstraction from this bridge carbon atom. A possible mechanism involves a rotation of the Thr62 side chain and completion of a short wire by which the proton is transported to His181, while the negative charge is transferred to the imidazolone carbonyl, producing a homoenolate intermediate that is stabilized by Arg96. Finally, comparison of the proton wires in the two proteins reveals common motifs, such as short internalized Ser/Thr-Glu hydrogen-bonded pairs for ultrafast proton abstraction, and threonine side chain rotation functioning as a proton wire switch.

Building a macro-mixing dual-basin Go model using the Multistate Bennett Acceptance Ratio.

The dual-basin Go-model is a structural-based coarsegrained model for simulating a conformational transition between two known structures of a protein. Two parameters are required to produce a dual-basin potential mixed using two single-basin potentials, although the determination of mixing parameters is usually not straightforward. Here, we have developed an efficient scheme to determine the mixing parameters using the Multistate Bennett Acceptance Ratio (MBAR) method after short simulations with a set of parameters. In the scheme, MBAR allows us to predict observables at various unsimulated conditions, which are useful to improve the mixing parameters in the next round of iterative simulations. The number of iterations that are necessary for obtaining the converged mixing parameters are significantly reduced in the scheme. We applied the scheme to two proteins, the glutamine binding protein and the ribose binding protein, for showing the effectiveness in the parameter determination. After obtaining the converged parameters, both proteins show frequent conformational transitions between open and closed states, providing the theoretical basis to investigate structure-dynamics-function relationships of the proteins.

Essential ion binding residues for Na+ flow in stator complex of the Vibrio flagellar motor.

The bacterial flagellar motor is a unique supramolecular complex which converts ion flow into rotational force. Many biological devices mainly use two types of ions, proton and sodium ion. This is probably because of the fact that life originated in seawater, which is rich in protons and sodium ions. The polar flagellar motor in Vibrio is coupled with sodium ion and the energy converting unit of the motor is composed of two membrane proteins, PomA and PomB. It has been shown that the ion binding residue essential for ion transduction is the conserved aspartic acid residue (PomB-D24) in the PomB transmembrane region. To reveal the mechanism of ion selectivity, we identified essential residues, PomA-T158 and PomA-T186, other than PomB-D24, in the Na+-driven flagellar motor. It has been shown that the side chain of threonine contacts Na+ in Na+-coupled transporters. We monitored the Na+-binding specific structural changes using ATR-FTIR spectroscopy. The signals were abolished in PomA-T158A and -T186A, as well as in PomB-D24N. Molecular dynamics simulations further confirmed the strong binding of Na+ to D24 and showed that T158A and T186A hindered the Na+ binding and transportation. The data indicate that two threonine residues (PomA-T158 and PomA-T186), together with PomB-D24, are important for Na+ conduction in the Vibrio flagellar motor. The results contribute to clarify the mechanism of ion recognition and conversion of ion flow into mechanical force.

Proton Wire Dynamics in the Green Fluorescent Protein.

Inside proteins, protons move on proton wires (PWs). Starting from the highest resolution X-ray structure available, we conduct a 306 ns molecular dynamics simulation of the (A-state) wild-type (wt) green fluorescent protein (GFP) to study how its PWs change with time. We find that the PW from the chromophore via Ser205 to Glu222, observed in all X-ray structures, undergoes rapid water molecule insertion between Ser205 and Glu222. Sometimes, an alternate Ser205-bypassing PW exists. Side chain rotations of Thr203 and Ser205 play an important role in shaping the PW network in the chromophore region. Thr203, with its bulkier side chain, exhibits slower transitions between its three rotameric states. Ser205 experiences more frequent rotations, slowing down when the Thr203 methyl group is close by. The combined states of both residues affect the PW probabilities. A random walk search for PWs from the chromophore reveals several exit points to the bulk, one being a direct water wire (WW) from the chromophore to the bulk. A longer WW connects the "bottom" of the GFP barrel with a "water pool" (WP1) situated below Glu222. These two WWs were not observed in X-ray structures of wt-GFP, but their analogues have been reported in related fluorescent proteins. Surprisingly, the high-resolution X-ray structure utilized herein shows that Glu222 is protonated at low temperatures. At higher temperatures, we suggest ion pairing between anionic Glu222 and a proton hosted in WP1. Upon photoexcitation, these two recombine, while a second proton dissociates from the chromophore and either exits the protein using the short WW or migrates along the GFP-barrel axis on the long WW. This mechanism reconciles the conflicting experimental and theoretical data on proton motion within GFP.

The hole in the barrel: water exchange at the GFP chromophore.

Internal water molecules in proteins are conceivably part of the protein structure, not exchanging easily with the bulk. We present a detailed molecular dynamics study of the water molecule bound to the green fluorescent protein (GFP) chromophore that conducts its proton following photoexcitation. It readily exchanges above 310 K through a hole that forms between strands 7 and 10, due to fluctuations in the 6-7 loop. As the hole widens, rapid succession of water exchange events occur. The exiting water molecule passes three layers of atoms, constituting the binding, internal, and surface sites. Along this pathway, hydrogen bonding protein residues are replaced with water molecules. The mean squared displacement along this pathway is initially subdiffusive, becomes superdiffusive as the water traverses the protein wall in a flip-flop motion, and reverts to normal diffusion in the bulk. The residence correlation function for the bound state decays biexponentially, supporting this three-site scenario. For a favorable orientation of the Thr203 side-chain, the hole often fills with a single file of water molecules that could indeed rapidly conduct the photodissociated proton outside the protein. The activation enthalpy for its formation, 26 kJ/mol, agrees with the experimental value for a protein conformation change suggested to gate proton escape.