Seminars2012 – 2016

Transcription, the synthesis of RNA from a complementary DNA template, plays a crucial role in cellular regulation, including differentiation, development, and other fundamental processes. In this talk, I will discuss our results on modeling the RNA polymerase II (Pol II, a system with ∼400K atoms) Translocation and other functional conformational changes of this enzyme at sub-millisecond timescales. We have developed a novel algorithm, Hierarchical Nystrom Extension Graph method, to construct kinetic network models to extract long timescale dynamics from short simulations. For example, we reveal that RNA polymerase II translocation is driven purely by thermal energy and does not require the input of any additional chemical energy. Our model shows an important role for the bridge helix: Large thermal oscillations of this structural element facilitate the translocation by specific interactions that lower the free-energy barriers between four metastable states. The dynamic view of translocation presented in our study represents a substantial advance over the current understanding based on the static snapshots provided by X-ray structures of transcribing complexes. At the end of my talk, I will briefly discuss our recent progress on extending our kinetic network model to include sequence-dependent molecular dynamics of Pol II elongation to predict transcriptional accuracy in the genome-wide transcriptomic datasets. This model creates a critical link between the structural-mechanics understanding of Pol II fidelity and the genome-wide transcriptional accuracy.

RNA is implied in many fundamental biological processes, such as protein translation, gene regulation and catalysis, which are accomplished thanks to its intrinsic structural plasticity, derived from RNA motifs polymorphism. During my PhD I studied a particular class of these motifs, RNA tetraloops, formed by four nucleotides that cap helices inducing a backbone U-turn. One of the overlooked structural features of tetraloops is the stacking of backbone oxygen atoms with nucleobases, originating anion-π or lone pair-π stacking interactions. These two interactions can be used to define two folds for tetraloops and are significant for the local and global RNA plasticity, as well as helping still problematic RNA folding experiments. Tetraloops and their stacking interactions are modulated by water and ions interactions. Moreover, intracellular environments are crowded by macromolecules and metabolites. Our current knowledge on the crowding phenomena is limited to macroscopic effects, and much has to be discovered about molecular details. For that, I propose to study by full-atomistic MD a system embedded inside the tRNA T-box riboswitch (essential for gene expression in bacteria) that is composed by an intramolecular stacking interactions between two base pairs. The effects of molecular and macromolecular crowding during the thermally-induced “opening” of this system will be analyzed, in order to expand our knowledge about local modification on hydration structure and more generally on how crowding affects biomolecular recognition.

Assigning accurately chemical specie to solvent densities is one of the remaining challenges for crystallographers. While progress was made in refining macromolecules with workflows like Phenix or PDB_REDO, ion placement often results in errors. I will present examples of Mg²⁺ binding to purine N7 atoms assignment errors. As the affinity of Mg²⁺ for nitrogen is considerably smaller than for oxygen atoms, the former are not natural Mg²⁺ partners. Indeed, through a survey of small molecular assemblies from the Cambridge Structural Database (CSD) and the larger PDB macromolecular systems, we were able: (i) to define more precisely the binding patterns of Mg²⁺ ions towards purine N7, (ii) to assess that N7 is very rarely interacting with these ions and (iii) to establish that most of the Mg²⁺ ions placed in front of N7 atoms are monovalent ions, water molecules or transition metals. These results demonstrate that better ion placement methodologies need to be developed.
With that goal in mind, I started molecular dynamics (MD) simulations in crystallo. Here I present application of this method to refinement of a 0.6 å sarcin-ricin loop crystal structure. This ultra-high resolution allows to see not only detailed electron density for the RNA, but also well resolved first and second solvation shell. Although metal cations were present in the crystallization media, they were not identified among the solvent densities. Yet, ammonium sulfate was used in the crystallization media as well and it is likely that some densities considered as water conceal NH₄ ⁺ since the similarities of this ion to water make it an elusive specie. The only way to differentiate between H₂O and NH₄ ⁺ is by analyzing their hydrogen bonding modes. However, even at such remarkable resolution, some details are missing. By modelling dynamics in crystallo of the sarcin-ricin loop, we could infer positions of hydrogens that cannot be found in experimental densities including those of NH₄ ⁺ ions. We believe that such methodology can be further developed to help identifying solvent species in macromolecular structures.

I will briefly introduce past and present work that we have done in our group on RNA structure and dynamics with the help of Luigi D’Asenzo ad Filip Leonarski. I will especially show the importance of short hydrogen bonds that might arise in protein and RNA/protein or DNA/protein systems as a result of the protonation of Asp and Glu carboxylate groups. The existence of such short hydrogen bonds might challenge some of the current force fields used for molecular dynamics (MD) simulations. Further, I will emphasize the importance of correctly evaluating crystallographic structures that might sometimes lead to ambiguous or even wrong models that considerably complicate the creation of reliable structural databases that are essential for the validation of MD models.

The outer membrane of gram-negative bacteria is a unique asymmetric membrane bilayer that is composed of phospholipids in the inner leaflet and lipopolysaccharides (LPS) in the outer leaflet. Its function as a selective barrier is crucial for the survival of bacteria in many distinct environments, and it also renders gram-negative bacteria more resistant to antibiotics than their gram-positive counterparts. LPS comprises three regions: lipid A, core oligosaccharide, and O-antigen polysaccharide. In this talk, I will present our ongoing efforts on understanding various bacterial outer membranes and their interactions with outer membrane proteins, including (1) construction of a model of an E. coli R1 (core) O6 (antigen) LPS molecule using the CHARMM36 lipid and carbohydrate force fields and simulations of various E. coli R1.O6 LPS bilayers; (2) modeling of E. coli R2, R3, R4, and K12 cores and other O-antigens and their bilayer simulations; (3) development of LPS Modeler in CHARMM-GUI; (4) modeling and simulation of E. coli outer membranes with phospholipids in the inner leaflet and LPS in the outer leaflet as well as OmpLA in the outer membrane; (5) modeling and simulation of BamA in the E. coli outer membrane; (6) other ongoing outer membrane – protein simulations.

Coherent X-ray Diffraction Imaging (CXDI) is an imaging technique suitable for the whole structure analyses of non-crystalline and micrometer-size specimens without staining, sectioning or chemical labeling, due to the large penetration depth of short-wavelength X-rays. We applied the CXDI technique to the structure analysis of cells and cellular organelles at the XFEL facility SACLA. Here I would like to introduce the theory, experimental techniques and current application of the technique for biological specimens.

Internal dynamics of RNA is required for its proper biological function. For example, flexibility of ribosomal RNA and mRNA is essential for efficient translation of mRNA into polypeptides. In our laboratory we apply molecular dynamics simulations, absorbance and fluorescence spectroscopy to investigate the RNA dynamics. I will speak about the importance of RNA flexibility in two biologically-relevant RNA motifs. One is the mRNA decoding site in the ribosome, which is also the binding site of aminoglycoside antibiotics. I will present the dynamical properties of this site in the context of thermodynamics of aminoglycoside binding to bacterial and human ribosomes. The other one is an RNA hairpin that acts as a thermosensor responsible for translation initiation in bacteria upon heat shock. These short mRNA sequences respond to temperature changes and their local melting allows mRNA binding to the ribosome. I will present the mechanism of thermal unwinding of a fourU thermometer.

The ability of atomistic molecular dynamics simulations to model biological systems has increased dramatically in the past few years. This ability has come about through advances in hardware, software, and methodology. Recent examples include millisecond simulations of individual proteins, representations of membranes with intricate compositions, and even modeling of entire viruses at atomistic resolution. In this talk, I will describe the application of MD simulations to a specialized sub-cellular region, the bacterial periplasm. Situated between two membranes in Gram-negative bacteria, the periplasm is home to a number of unique component and systems. I will specifically address simulations of three outer-membrane proteins in their native environment: BtuB, a vitamin B12 transporter; BamA, responsible for outer-membrane-protein insertion; and LptD/E, a protein that inserts lipopolysaccharides into the outer leaflet of the outer membrane. I will also demonstrate that simulations can capture the mechanical properties of the bacterial cell wall, which is anchored to the outer membrane. Finally, I will describe our current efforts to unify all of these aspects into a model of the entire periplasmic space.

In general intermolecular interactions between pair of closed shell molecules can be represented by [-1, -6, +12] potential. Various intermolecular interaction varies significantly due to the differences in the weightage for each of the three terms. Spectroscopy and ab-initio calculations provide the reasonable understanding of many intermolecular interactions. However, each method has specific shortfalls. Physically meaningful models can only be constructed by adequately addressing these shortfalls while interpreting the data. The importance of each of the components of [-1, -6, +12] potential in understating hydrogen bonding, π–π stacking and in foldamers will be highlighted.

Because biomolecular processes are largely under thermodynamic control, dynamic extension of thermodynamics is necessary to uncover the mechanisms and driving factors of fluctuating processes. The fluctuating thermodynamics technology presented in this talk offers a practical means for the thermodynamic characterization of conformational dynamics in biomolecules. The use of fluctuating thermodynamics has the potential to provide a comprehensive picture of fluctuating phenomena in diverse biological processes. Through the application of fluctuating thermodynamics, we provide a thermodynamic perspective on the misfolding and aggregation of the various proteins associated with human diseases. In this talk, I will present the detailed concepts and applications of the fluctuating thermodynamics technology for elucidating biological processes.

Molecular graphics involves visualisation of molecules but rarely allows the user to engage their sense of touch to help learn about biomolecules.

We have developed three software tools that use a haptic (force-feedback) device: “Haptimol-ISAS”, “Haptimol-ENM” and “Haptimol-RD”(See: http://www.haptimol.com). Haptimol-ISAS allows the user to explore the solvent accessible surface of a biomolecule using a haptic device, Haptimol-ENM allows the user to apply forces to an elastic network model of a biomolecule and our latest software tool, Haptimol-RD allows the user to dock molecules rigidly. For Haptimol-RD methods will be described that enable the calculation interaction forces within the time constraint required for smooth perception of forces (1-2 milliseconds). These methods allow us to calculate interaction forces between very large biomolecules when implemented on the GPU.

Future developments will also be discussed, in particular tools that model protein flexibility for drug-protein and protein-protein interactions.

We are developing the special-purpose computer system for MD simulations, MDGRAPE-4. MDGRAPE-4 has a similar architecture as ANTON by D. E. Shaw research – it utilizes a system-on-chip architecture that integrates general-purpose processors, specialized pipelines, memories, and network interfaces.
The MDGRAPE-4 hardware has been completed on August 2014, and we are currently developing the software on it.
We will report the current status of the system and discuss future directions.

Amyloid fibrils are naturally occurring, self-assembled, supramolecular systems. Quantitative understanding of the kinetics of fibril formation and the molecular mechanism of transition from monomers to fibrils holds the key to describing the functions of amyloid fibrils. Significant advances using computations of protein aggregation in a number of systems have established generic and sequence specific aspects of the early steps in oligomer formation, as well as the ultimate formation of protofibrils and amyloid fibrils. Theoretical considerations, that view oligomer and fibril growth as diffusion in a complex energy landscape, and computational studies, involving minimal lattice and coarse-grained models, have revealed general principles governing the transition from monomeric protein to ordered fibrillar aggregates. Detailed atomistic calculations have explored the early stages of the protein aggregation pathway for a number of amyloidogenic proteins, most notably amyloid β (Aβ) protein and protein fragments. These computational studies have provided insights into the role of sequence, role of water, and specific interatomic interactions underlying the thermodynamics and dynamics of elementary kinetic steps in the aggregation pathway. More recently, studies have provided insight into the structural basis for the production of Aβ-peptides through interactions with secretases in the presence of membranes.
Recent results will be discussed, with an emphasis on theory and computation acting as a complement to experimental studies probing the principles governing protein aggregation.
(1) D. Thirumalai, G. Reddy, and J. E. Straub, Acc. Chem. Res. 45, 83-92 (2012).
(2) J.E. Straub and D. Thirumalai, Ann. Rev. Phys. Chem. 62, 437-463 (2011).
(3) J. E. Straub and D. Thirumalai, Curr. Opin. Struc. Bio. 20, 187-195 (2010).

The outer membrane of Gram-negative bacteria is a unique asymmetric lipid bilayer that is composed of phospholipids (PL) in the inner leaflet and lipopolysaccharides (LPS) in the outer leaflet. Its function as a selective barrier is crucial for the survival of bacteria in many distinct environments, and it also renders Gram-negative bacteria more resistant to antibiotics than their Gram-positive counterparts. LPS comprises three regions: lipid A, core oligosaccharide, and O-antigen polysaccharide. Utilizing the CHARMM36 lipid and carbohydrate force fields, we have constructed a model of an E. coli R1 (core) O6 (antigen) LPS molecule. Several all-atom bilayers are built and simulated with lipid A only (LIPA) and varying lengths of 0 (LPS0), 5 (LPS5), and 10 (LPS10) O6 antigen repeating units to investigate the impact of the molecular length on LPS bilayer structures. We also studied the structural properties of a model of the E. coli outer membrane and its interaction with various outer membrane proteins, including OmpLA, OmpF, and BamA, utilizing molecular dynamics simulations.

Docking (posing) calculations coupled with binding free energy estimates (scoring) are a mainstay of structure-based drug design. Docking and scoring methods have steadily improved over the years, but remain challenging because of the extensive sampling that is required, the need for accurate scoring functions and challenges encountered in accurately estimating entropy effects. This talk addresses the use of ensemble principles to directly address these issues and, thereby, accurately estimate protein-ligand binding free energies. In particular, we analytically demonstrate that sampling reduces computed binding free energy uncertainties and then highlight several methods that incorporate these concepts. For example, the moveable type method, employs an elegant approach to generate the necessary ensembles by using a “binned” pairwise knowledge-based potential combined with atom pair probabilities extracted from known protein-ligand complexes. This allows us to rapidly compute the ligand, protein and protein-ligand (inclusive of solvation effects) ensembles which then can be used to directly estimate protein-ligand binding free energies using basic statistical mechanical principles. This approach improves the quality of the potential (scoring) function by reducing computational uncertainty, sampling phase space in one shot and accurately incorporating entropy effects. This allows us to compute binding free energies rapidly, accurately and yields molecular poses at a minimal computational cost relative to currently available methods based on statistical mechanics.

This talk concerns the complex structures in the electronic states of π-conjugated molecules and the molecular dynamics using quantum chemical methods. Here, the word “complex” structure means that it cannot be reduced into effective one-body problems. At first, I’ll show several examples of such multireference (in other words, strongly-correlated) phenomena in organic π-conjugated systems. Then, I’ll describe how the state-of-the-art electronic structure theory, such as ab initio density matrix renormalization group, have been addressed to the following intriguing phenomena: Fluorescence spectra from the dark state of polyenes; instability of high spin states of polycarbenes; emergence of multi radical electrons on finite graphene nanoribbons. Finally, I’ll outline a new method for multireference quantum molecular dynamics. This method is designed for large-amplitude motions (and chemical reactions) where several degrees of freedom may strongly couple to each other. The key idea is to apply different methods to different coordinates: Strongly-correlated coordinates are treated by an expensive variational method; the remaining couplings are considered perturbatively. This scheme, which can be seen as an analogous vibrational wave function model for the MRMP method in electronic structure theory, allow us to treat a large syste ms with strongly coupled motions efficiently.

The ribosome is the universally conserved molecular machine responsible for protein synthesis. Over the past decade, we have focused on the mechanism by which the ribosome decodes genetic information (‘the decoding problem’, or ‘tRNA selection’). By performing large-scale molecular dynamics simulations of the ribosome, we are able to examine the inner workings of this molecular machine. A key rearrangement of the parts of this machine is called ‘accommodation’. Here, transfer RNAs (tRNAs) carrying protein building blocks (amino acids) move into the ribosome. We identified a new functional region of the ribosome (‘the accommodation corridor’) and predicted that certain parts of this corridor are important for ribosome function. Our predictions were recently validated in studies by three experimental groups. In an additional separate set of studies that combined our simulations with single molecule experiments, a new picture of ribosome function has emerged. Rather than the ribosome machine parts moving in lock-step, both simulations and single molecule experiments show the tRNAs making large-scale reversible excursions in a trial-and-error fashion. This picture is consistent with a dynamic energy landscape view of the ribosome. After studying the relatively tractable problem of ‘accommodation’, we are now investigating the mechanism of translocation, where a large conformational change involving the entire ribosome occurs. This motion allows the ribosome to move exactly 3 nucleotides along the messenger RNA to the next amino acid codon. Using microsecond sampling in explicit solvent for the full ribosome, in combination with experimentally measured rates, we are able to estimate barrier heights for various motions important for translocation. We have also used coarse-grained methods to simulate the various sub-steps of translocation. Our future goal is to use simulations of the ribosome to produce detailed energy landscapes of translocation.

The thiazole antibiotic thiostrepton inhibits bacterial protein synthesis by binding to a cleft formed by the ribosomal protein L11 and helix 43-44 of the 23S ribosomal RNA, a part of the GTPase associated region (GAR) on the 70S ribosome. It was proposed from ribosomal crystal structures that the ligand restricts the N-terminal movement of L11 and thus prevents proper binding of translation factors. An exact understanding of this mechanism at atomic resolution is, however, still missing. I will present our results from all-atom molecular dynamics simulations of the binary L11-23S complex and the ternary L11-23S-thiostrepton complex, which provides some new insights into this mechanism at atomic resolution. We demonstrate that thiostrepton has an impact on the protein and rRNA dynamics. Specifically, it restricts the conformational flexibility of the nearby N-terminal domain, and has a weak dynamic coupling to the distant C-terminal domain. We identified distinct conformations of the far more flexible ‘apo’ form of N-terminal domain that may reflect distinct interaction states with translation factors. If time permits, I will also introduce our new tool for the optimization of force field parameters.