Skip to Main Content
From Computational Biophysics to Systems Biology (CBSB2017)
May 18-20, 2017  Cincinnati, Ohio
OU link

Children's Hospital link

OU link

NSF link

  • Nabeel Ahmed (Pennsylvania State University), Poster
  • Erik Alred (University of Oklahoma), Poster
  • Mohammad Sadegh Avestan (University of Cincinnati), Poster
  • Ivet Bahar (University of Pittsburgh), Keynote Talk
  • Nathan Bernhardt (University of Oklahoma), Best Poster Award
  • Jacek Biesiada (University of Cincinnati)
  • Nicholas Bodmer (Washington University), Poster
  • Julia Borbas (University of Konstanz), Poster
  • Samuel Bowerman (Illinois Institute of Technology), Best Poster Award
  • Bernard Brooks (NIH), Keynote Talk
  • Kathleen Carter (Georgia State University), Poster
  • Arghya Chakravorty (Clemson University), Outstanding Young Researcher Award Talk
  • Catherine Chaton (University of Cincinnati)
  • Weizhong Dai (Louisiana Tech University)
  • Manish Datt (Ahmedabad University)
  • Jiajie Diao (University of Cincinnati),Poster
  • Ruxandra Dima (University of Cincinnati)
  • Thomas Dodd (Georgia State University),Poster
  • Andrew Eisenhart (University of Cincinnati)
  • Eshel Faraggi (Indiana University),
  • K. Anton Feenstra (Vrije Universiteit, Amsterdam), Invited Talk
  • Yasan Fonseka (University of Cincinnati), Poster
  • Yiqin Gao (Peking University), Invited Talk
  • Angel Garcia (Los Alamos National Laboratory), Invited Talk
  • Debajyoti Ghosh (University of Cincinnati)
  • Michael Gribskov (Purdue University), Contributed Talk
  • Ulrich H.E. Hansmann (University of Oklahoma)
  • Jason Harris (University of North Carolina Chapel Hill), Contributed Talk
  • Teresa Head-Gordon (UC Berkeley), Keynote Talk
  • Andrew Herr (CincinnatiChildren's Hospital Medical Center)
  • Toshiko Ichiye (Georgetown University), Invited Talk
  • Abdolreza Javidialesaadi (University of Cincinnati), Poster
  • Zhe Jia (Clemson University), Contributed Talk
  • Rui Jiang (University of Cincinnati), Poster
  • Khatuna Kachlishvili (Cornell University), Poster
  • Raphael Kopan (Cincinnati Children's Hospital), Invited Talk
  • Konda Reddy Karnati (Albany State University), Poster
  • Daisuke Kihara (Purdue University), Contributed Talk
  • In-Kwon Kim (University of Cincinnati)
  • Andrzej Kloczkowski (Nationwide Children Hospital), Invited Talk
  • Michael Kotliar (Cincinnati Childrens Hospital Medical Center)
  • Sarah Leininger (Penn State University), Best Poster Award
  • Chen Li (Illinois Institute of Technology), Outstanding Young Researcher Award Talk
  • Oliver Lichtarge (Baylor College of Medicine), Invited Talk
  • James Liman (Rice University)
  • Xiaoming Lu (Cincinnati Childrens Hospital Medical Center)
  • Avi Ma'ayan (Mount Sinai), Invited Talk
  • Gia Maisuradze (Cornell University), Contributed Talk
  • Girik Malik (Nationwide Children Hospital), Poster
  • Abdullah Al Mamun (University of Louisville),Poster
  • Parthiban Marimuthu (Albany State University), Poster
  • Jarek Meller (University of Cincinnati)
  • Dale Merz (University of Cincinnati), Poster
  • David Minh (Illinois Institute of Technology), Contributed Talk
  • Trung Hai Nguyen (Illinois Institute of Technology), Poster
  • Daniel Nissley (Penn State), Best Poster Award
  • Edward O'Brien (Penn State University), Invited Talk
  • Yunhui Peng (Clemson University), Poster
  • Tina Perica (UC, San Francisco), Outstanding Young Researcher Award Talk
  • Joseph Persichetti (Pennsylvania State University), Poster
  • Marcin Pilarczyk (University of Cincinnati), Poster
  • Adolfo Poma (Polish Academy of Sciences), Outstanding Young Researcher Award Talk
  • Yasin Pourfarjam (University of Cincinnati), Poster
  • Mario Pujato (Cincinnati Children's Hospital Medical Center)
  • Karissa Sanbonmatsu (Los Alamos National Laboratory), Invited Talk
  • Stephan C. Schürer (University of Miami), Invited Talk
  • Ajeet Sharma (Pennsylvania State University), Poster
  • Harinder Singh (Cincinnati Children's Hospital Medical Center), Invited Talk
  • George Stan (University of Cincinnati)
  • Yu-Hsuan Shih (University of Cincinnati), Poster
  • Allison Talley (University of Cincinnati), Poster
  • Marcus Thomas (Carnegie Mellon University), Contributed Talk
  • Alexander Thorman (University of Cincinnati), Poster
  • Rohith Anand Varikoti (University of Cincinnati), Poster
  • Juozas Vasiliauskas (University of Cincinnati)
  • Wenhua Wang (University of Oklahoma), Poster
  • Qi Wang (University of Cincinnati), Poster
  • David Wolfe (Pennsylvania State University)
  • Wenhui Xi (University of Oklahoma), Poster
  • Bing Xie (Illinois Institute of Technology), Poster
  • VIVEK YADAV (Temple University), Poster
  • Chunli Yan (Georgia State University), Poster
  • J. Alexander Yates (University of Oklahoma),Poster
  • Alexander Yarawsky (University of Cincinnati)i
  • Jun Yi (Nanjing University of Science and Technology), Contributed Talk
  • Hexi Zhang (Illinois Institute of Technology)
  • Xiaochuan Zhao (University of Vermont), Outstanding Young Researcher Award Talk



  • Exploring reaction mechanism and folding pathways using the Action-CSA method
    Bernhard Brooks
    National Heart, Lung, and Blood Institute

    Conformational Space Annealing (CSA) is the most robust conformational search algorithm available to us. We have extended the CSA machinery to search, not for structures, but for entire for reaction pathways, using action integrals for pathway evaluation. We continue to develop and refine the algorithms, targeting small molecule reactions at the QM level, enzyme mechanism involving QM/MM, and larger all-atom peptide and protein folding problems. Initial results involving the optimization of Onsager-Machlup action integrals with CSA will be presented.
    Juyong Lee, In-Ho Lee, InSuk Joung, Jooyoung Lee & Bernard R. Brooks „Finding multiple reaction pathways via global optimization of action‰, in press, Nature Communications (2017).
    Back to the top of the page

  • Computational Methods and Models for Intrinsically Disordered Peptides
    Teresa Head-Gordon
    Departments of Chemistry, Bioengineering, and Chemical and Biomolecular Engineering
    University of California, Berkeley, CA 94720

    Intrinsically disordered proteins (IDPs) represent a new frontier in structural biology since the primary characteristic of IDPs is that structures need to be characterized as diverse ensembles of conformational sub-states. This talk pertains to determining the diverse conformational substates of IDPs using the interplay of solution-based NMR experiments that are analyzed with state-of-the art molecular simulation methods and Bayesian probabilistic models for generating, refining, and validating IDP structural ensembles
    Back to the top of the page

  • Dynamics of Dopamine Transport: Molecular-to-Cellular Simulations
    Ivet Bahar
    University of Pittsburgh

    Efficient clearance of dopamine (DA) molecules from the synapse is key to regulating dopaminergic signaling. This role is fulfilled by dopamine transporters (DATs). Recent advances in the structural characterization of DAT from Drosophila (dDAT) at the molecular level, and imaging of dopamine neurons and the distribution and trafficking of DATs at the cellular level, now permit us to gain a mechanistic understanding of DA reuptake events at multiple scales in silico. Molecular simulations have now elucidated the conformational changes undergone by DAT during DA transport cycle (Cheng & Bahar, Structure 2015). However, the efficiency of transport at the chemical synapses are affected by many other features beyond the structural dynamics of DATs, such as the diffusion of DAs in the synapse or extra-synaptic region, the structural heterogeneity of the synaptic regions, or the surface distribution of DATs on the plasma membrane. Using electron microscopy images and immunofluorescence labeling of transgenic knock-in mice brain, we reconstructed a realistic simulation environment for Monte Carlo simulations of DA transmission using MCell, a package for simulating subcellular, microphysiological events. The heterogeneous distribution of DATs as well as complex morphology of axon terminals near active zones give rise to large fluctuations in the extracellular DA density. The probability of attaining local DA concentrations sufficiently high to activate high affinity, and even low affinity, DA receptors, exhibit a strong dependency on the pattern of DA release, phasic/periodic vs tonic/random, the former resulting in a higher probability of receptor activation. The study highlights the importance of taking account of the irregular extra-synaptic geometry, DAT surface density heterogeneities, DAT equilibrium thermodynamics between outward- and inward-facing states, for a realistic description of the DA transport dynamics.
    Back to the top of the page

Invited Talks

  • Molecular Simulations of Cosolvent effects on Protein and RNA Stability
    Angel Garcia
    Molecular Simulations of Cosolvent effects on Protein and RNA Stability

    Proteins are marginally stable, and the folding/unfolding equilibrium of proteins in aqueous solution can easily be altered by the addition of small organic molecules known as cosolvents. Cosolvents that shift the equilibrium toward the unfolded ensemble are termed denaturants, whereas those that favor the folded ensemble are known as protecting osmolytes. Urea is a widely used denaturant in protein folding studies, and the molecular mechanism of its action has been vigorously debated in the literature. We we review recent experimental as well as computational studies that show an emerging consensus in this problem. Urea has been shown to denature proteins through a direct mechanism, by interacting favorably with the peptide backbone as well as the amino acid side chains. In contrast, the molecular mechanism by which the naturally occurring protecting osmolyte trimethylamine N-oxide (TMAO) stabilizes proteins is not clear. Detailed molecular simulations, when used with force fields that incorporate these interactions, can provide insight into this problem.
    Back to the top of the page

  • Computational Identification and Screening for Deleterious Mutants
    Andrzej Kloczkowski
    Nationwide Children Hospital

    The practice of precision medicine relies on individual patient genome sequences, among other data. The challenge for the interpretation of genome sequences is to distinguish between the normal and the disease-related mutations. The human proteome contains over 20,000 structures, and the consideration of possible mutations increases the number of potential protein structures enormously. The experimental determination of all these structures and assessment of phenotypic effects of mutations (beneficial, deleterious or neutral) is currently far beyond experimental reach, and we need to rely on efficient computational methods. The structural and dynamical information greatly advances our ability to correctly predict phenotypic effects of protein mutations. To achieve these goals, we significantly improved protein structure evaluations by considering the effects of amino acid variants on protein stability, and we have shown that the outliers in stability are typically aberrant proteins. In addition, protein dynamics is usually strongly impacted by deleterious mutants. Recently, we have made significant progress in understanding protein stability and dynamics by computing protein free energies extracted from structures to account for the high packing densities in proteins, including important novel evaluations of protein entropies. Preliminary results show large improvements over previous potentials for assessing protein stabilities. Likewise, when this is applied to a large set of mutants ˆ the deleterious mutants show extreme changes in stability, with either higher or lower stability. We have shown for some mutants of a kinase that the cooperativity of the functional dynamics is very substantially changed for a non-functional mutant. Our results demonstrate that there are substantial gains in specificity from combining the sequence with structural and protein dynamic data. These developments may significantly impact the advancement of precision/personalized medicine
    Back to the top of the page

  • Extreme Biophysics: Proteins and the Intracellular Environment
    Toshiko Ichiye
    Georgetown University

    Understanding how extreme conditions of pressure (P), temperature (T), and chemical composition (X) affect biological macromolecules is important in understanding how life exists in extreme conditions and how extreme conditions can be used to kill pathogens, as well as for bioengineering new functions for macromolecules or whole cells. However, the P-T-X conditions of these macromolecules in the cell are often not considered in experiments. In addition, while changes in protein sequences or chemical composition of membranes may protect against extremes, cells may use also regulation of the chemical composition “X” of the intracellular environment. These issues will be discussed as well as progress and challenges in simulations.
    Back to the top of the page

  • Towards Clinical Applications of AI and Evolutionary Theory
    Olivier Lichtarge
    Baylor College of Medicine

    A persistent challenge is the integration of biological data into knowledge, and better therapies. The complexity, heterogeneity, and sheer mass of information are daunting, however. Here, we will describe a stepwise computational approach. The first step integrates information over networks. An example is a prediction of gene function from a network spanning hundreds of species, leading to a possible new target for the best current antimalarial drug. The second step is text mining. We show how automated reasoning over the all PubMed abstracts suggests new protein interactions, including new p53 kinases. Perhaps most surprising is the third step˜an analytic model of evolution that adds information on patient specific genome variations. This model reduces to an equation that describes the genotype-phenotype relationship in terms of perturbations in the fitness landscape. Mutational, clinical, and population genetic data show that the equation agrees fairly well with the effect of point mutations in diverse proteins; correlates significantly with morbidity and mortality in disease-causing gene mutations; and may explain human coding polymorphism frequencies. Taken together, these different approaches point to an integrative framework that may soon reflect structured and unstructured data that is personalized to the relevant mutational fluctuations of individual patients. Hopefully, diverse applications across biology and precision medicine should benefit.
    Back to the top of the page

  • Non-equilibrium physics in nascent protein structure and function
    Edward O'Brien
    Penn State University

    An emerging paradigm is that translation kinetics can influence nascent protein behavior in vivo over long time scales. Introducing synonymous codon mutations into an mRNA molecule, which changes the rate at which codon positions are translated by the ribosome but not the amino acid encoded, has been shown to influence whether a nascent protein will fold and function, misfold and malfunction, aggregate, or efficiently translocate to a different cellular compartment. Synonymous mutations that can change translation rates have now been linked to a variety of diseases, including subtypes of hemophilia and cancer. These findings are a shift away from the thermodynamic-control paradigm of protein structure and function, to one in which kinetic control can play an important role in determining nascent protein behavior. My lab is developing theoretical and computational tools understand, model and predict the coupling of in vivo protein behavior and translation-elongation kinetics. I will present results from recent studies in which we demonstrate that chemical kinetic modeling can accurately predict co-translational folding curves in vivo, and that by using coarse-grained molecular-dynamics simulations we can explain the physical origin of the experimental observation of critical-codon positions. Critical codons are codon positions within a messenger RNA that, upon a synonymous substitution, cause the nascent protein to behave in a radically different manner. Finally, I will present results indicating that evolution has encoded translation-rate information into mRNA sequences to temporally-coordinate chaperone binding.
    Back to the top of the page

  • DNA methylation landscape and modelling of 3D chromatin structure
    Yiqin Gao
    Peking University

    Biological functions are determined by both genetics and epigenetics. 3D chromatin structure is expected to play an important role in determining the genomic features. Recent Hi-C experimental studies have been dramatically deepening our understanding of the 3D chromatin. In this talk, we describe the modelling of the chromatin structure utilizing experimental Hi-C data, which provides the spatial organization of the human chromatin. By mapping a plethora of genome features onto the chromatin model, we quantitatively and systematically reinforce the importance of chromatin architecture for genome function. We find that the co-localization of genome features on a linear map coincides their 3D segregation, and thus the latter provides a mechanism for the regulation of various genome properties. Especially, we will discuss the effects of DNA methylation on the structure of DNA and that the overall DNA methylation patter reflects the 3-D organization of the human chromatin.
    Back to the top of the page

  • How can a hypomorphic and a hypermorphic mutation in Notch cause the same cancer?`
    Raphael Kopan
    Cincinnati Children's Hospital Medial Center)

    Notch signaling pathway plays a central role in kidney development as well as acute and chronic kidney injury. Sending extracellular signals through Notch involves hetero-oligomerization of the membrane receptors, cleavage and translocation to the nucleus of the intracellular domain, and formation of transcription activating complexes with Mastermind and other proteins. Biochemical and structural studies, coupled with systems biology approaches, have elucidated many critical protein-protein interactions and the effects of sequence variants modulating Notch signaling. Here, we ask the question how hypomorphic (ie. leading to reduced activity) and hypermorhic mutations can lead to the same phenotype. We show that the answer lies in the effect loss of dimerization has on the rate of degradation.
    Back to the top of the page

  • Global Analysis and Visualization of Thousands of Expression Signatures for Drug and Target Discovery
    Avi Ma'ayan
    Maunt Sinai

    In the past several decades, drug discovery was obsessed with first identifying a worthy target and then designing small molecules that would specifically bind to the target to modulate its activity. This approach is still prominent today. However, there is increase appreciation that most successful drugs bind to many targets, and the mechanisms of action of many are not fully understood. High throughput gene expression and phenotypic drug screening strategies are a powerful alternative approach toward drug discovery. Coupled with machine learning methods, such approaches can produce target agnostic novel therapeutics quickly. The Library of Network-based Integrated Cellular Signatures (LINCS) project is a central resource for providing data and tools that can enable this type of drug discovery approach. As the data coordination and integration center (DCIC) for LINCS, we have developed methods to: prioritize small molecules to be tested for various complex diseases, identify mimickers for highly successful drugs, predict small molecules to assist in cellular reprogramming and directed differentiation, and better predict adverse events for small molecules before these are tested in human.
    Back to the top of the page

  • Towards automated fitting of high resolution cryo-EM data
    Karisaa Sanbonmatsu
    Los Alamos National Laboratory

    An explosion of new data from high-resolution cryo-electron microscopy (cryo-EM) studies has produced a large number of data sets for many species of ribosomes in various functional states over the past few years. While many methods exist to produce structural models for lower resolution cryo-EM reconstructions, high-resolution reconstructions are often modeled using crystallographic techniques and extensive manual intervention. Here, we present an automated fitting technique for high-resolution cryo-EM data sets that produces all-atom models highly consistent with the EM density.[1] Using a molecular dynamics approach, atomic positions are optimized with a potential that includes the cross-correlation coefficient between the structural model and the cryo-EM electron density, as well as a biasing potential preserving the stereochemistry and secondary structure of the biomolecule. Specifically, we use a hybrid structure-based/ab initio molecular dynamics potential to extend molecular dynamics fitting. We obtain atomistic models of the human ribosome consistent with high-resolution cryo-EM reconstructions of the human ribosome. Automated methods such as these have the potential to produce atomistic models for a large number of ribosome complexes simultaneously that can be subsequently refined manually.
    [1] Using Molecular Simulation to Model High-Resolution Cryo-EM Reconstructions. Serdal Kirmizialtin, Justus Loerke, Elmar Behrmann, Christian MT Spahn, Karissa Y Sanbonmatsu. 2015. Methods in enzymology Volume 558, Pages 497-514.
    Back to the top of the page

  • Novel therapeutic combinations and targeted poly-pharmacology in Glioblastoma
    Stephan C. Schürer
    University of Miami

    Glioblastoma Multiforme (GBM) is the most common and aggressive adult brain tumor. The standard of care, surgical recection followed by radiotherapy and temozolamide chemotherapy yields a 5-year survival rate of less than 10%.
    Using transcriptional response data from the Library of Integrated Network-based Cellular Signatures (LINCS) we identified novel compound combinations that show synergy in GBM PDX models with the potential for overcome resistance and improved clinical efficacy. Drugs with targeted polypharmacology represent a potentially attractive alternative to combination therapy. To demonstrate this approach, we illustrate the computational design of multi-target EGFR kinase and bromodomain-containing protein 4.
    Back to the top of the page

  • Binary cell fate choices as violations of symmetry in biological systems
    Harinder Singh
    Division of Immunobiology and the Center for Systems Immunology
    Cincinnati Children's Hosptial Medical Center

    Binary cell fate choices are a universal feature of developmental processes in biology. Recurring cellular bifurcations enable the elaboration of diverse differentiated states that are associated with unique functional features. The ground state of a progenitor can be maintained through symmetric cell division. When such a cell division generates two distinct differentiation states this process can be viewed as a violation of cellular symmetry. Our work has focused on various binary cell fate choices in the hematopoietic and immune systems. I will elaborate on two such exemplars, one controls the generation of monocytes and neutrophils of the blood and the second the bifurcating dynamics of antibody producing B cells of the immune system. In the former case, single-cell genomic profiling has uncovered a metastable monocyte-neutrophil intermediate that appears to manifest dynamic instability. In the latter instance, a regulatory circuit comprising of sequential double negative feedback loops generates exceptional bifurcating cellular trajectories that tune the quality of an antibody response to infection or vaccination
    Back to the top of the page

  • Seeing the Trees through the Forest: Sequence-based Homo- and Heteromeric Protein-protein Interaction sites prediction using Random Forest
    K. Anton Feenstra
    Vrije Universiteit, Amsterdam

    Protein-protein interactions play a central role in virtually most cellular processes. Identification of interaction sites between two interaction proteins is essential to understand complex formation and investigate their functions. Genome sequencing is now producing ever-increasing amounts of protein sequences. Few of these have experimentally validated annotations, however, making the prediction of protein interaction sites from sequence information increasingly attractive. \r\nWe present our recently developed random-forest-based prediction method for protein-protein interface using sequence information only. Our final general interface predictor was trained on the combined homomeric and heteromeric dataset, and we show on independent test-sets that in addition to predicting homomeric interfaces (AUC-ROC 0.724; second-best method 0.601), it is also able to pinpoint interface residues in heterodimers (AUC-ROC 0.636; second-best method 0.614). The success of our random forest model suggests that homodimer and heterodimer interfaces have much more in common than is often assumed.\r\nAs the method is sequence-based and not sensitive to the type of interaction, we expect our research to be of interest to many biological and biomedical researchers in academia and industry, especially when protein structure data is unavailable. For non-expert users we implemented the SeRenDIP web-server ( It simply takes the sequence of a protein of interest, and outputs a table with interaction position predictions and scores indicating likelihood and confidence of the predictions
    Back to the top of the page

Outstanding Young Researcher Award Talks

  • Multiscale Modeling of Membrane Proteins in a Heterogeneous Environment
    Xiaochuan Zhao
    University of Vermont

    G protein-coupled receptors (GPCRs) represent a superfamily of proteins that play indispensible roles in cellular signal transduction, and thus have been extensively studied as therapeutic targets. Embedded in the biological membrane, however, GPCR oligomers are challenging to simulate at attainable computational cost using all-atom models. Toward a practical solution to this challenge, we have investigated a prototypial GPCR heterodimer, which contains two distinct receptor proteins ˜ the adenosine A2A receptor (A2AR) and the dopamine D2 receptor (D2R). It has been long known that, activation of A2AR deactivate D2R and vice versa, a phenomenon that requires molecular-level knowledge to understand. Therefore, using the conventional all-atom (AA), our newly developed mixed-resolution (AACG), and the MARTINI coarse-grained (CG) force fields, we built five complex models with distinct interfaces of the heterodimers and compared them in AA, AACG, and CG simulations. For the same model at different resolutions, we observed consistent structural stability but various levels of speed-ups in the protein dynamic. The AACG and CG models show two and four times faster protein diffusion than the all-atom models, in addition to a 4- and 400-fold speedup in the simulation performance. Therefore, we obtained both the high-resolution details from AA models and dynamic behaviors on large time scale from CG models. This multiscale modeling strategy can be instructive in illuminating other complex systems. To further improve the ability of modeling even larger complex protein system, we are developing a new CG modeling method, in which several amino acids are represented by one CG site and the CG force field is being parameterized with fluctuation matching. We expect that the new method can dramatically reduces the simulation cost, allowing us to model large protein assemblies.
    Back to the top of the page

  • Reweighting the apo to the holo ensemble
    Chen Li
    llinois Institute of Technology

    Molecular dynamics (MD) simulations have become a powerful and popular method for the study of protein allostery, the widespread phenomenon in which a stimulus at one site on a protein influences the properties of another site. Simulations can enable the discovery of allosteric binding sites and elucidate the mechanism of allostery, providing a foundation for applications including rational drug design and protein engineering. However, performing a separate simulation for every ligand can be computationally expensive. Here we demonstrate a proof of principle that conformations sampled from the apo state can be reweighted according to the binding potential of mean force (BPMF) to predict ensemble averages in a holo state (bound to a ligand). Using BPMFs for the binding of AMP and ATP to adenylate kinase (AdK), an enzyme that plays an important role in the process of energy homeostasis, we construct a hinge angle distribution consistent with alchemical binding free energy calculations.
    Back to the top of the page

  • Modeling Electrostatic Potential around Proteins: Role of Bound Ions and Implications for Zeta-Potential
    Arghya Chakravorty
    Clemson University

    A new feature of the popular software DelPhi is developed and reported, allowing for computing the surface averaged electrostatic potential (SAEP) of macromolecules. The user is given the option to specify the distance from the van der Waals surface where the electrostatic potential will be\r\noutputted. In conjunction with DelPhiPKa and the BION server, the user can adjust the charges of titratable groups according to specific pH values, and add explicit ions bound to the macromolecular surface. This approach is applied to a set of four proteins with „experimentally‰ delivered zeta (ζ)-\r\npotentials at different pH values and salt concentrations. It has been demonstrated that the protocol is capable of predicting ζ-potentials in the case of proteins with relatively large net charges. This protocol has been less successful for proteins with low net charges. The work demonstrates that in the case of proteins with large net charges, the electrostatic potential should be collected at distances about 4 Å away from the vdW surface and explicit ions should be added at a binding energy cutoff larger than 1−2kT, in order to accurately predict ζ-potentials. The low salt conditions substantiate this effect of ions on SAEP.
    Back to the top of the page

  • GoMARTINI: Studying Conformational transition in proteins with the Martini force-field
    Adolfo Poma
    Polish Academy of Sciences

    The application of Coarse-Grained (CG) models in biology is essential to access large length and time scales needed for the description of several biological processes (e.g. self-assembly of beta-peptides into disease/functional amyloid fibrils, dissociation of protein-protein complexes, large movement of proteins under high mechanical stress, etc). The ELNEDIN protein model is based on the well-known MARTINI CG force-field and incorporates additionally „harmonic bonds‰ of a certain spring constant within a defined cutoff distance between pairs of amino acid residues, in order to retain the native structure of the protein. In this case, the use of unbreakable harmonic bonds hinders the study of unfolding and folding processes. To overcome this barrier we have replaced the harmonic bonds with LennardˆJones interactions based on the contact map of the native protein structure as is done in Go-like models. This model exhibits very good agreement with all-atom and the ELNEDIN simulations. Moreover, it can capture the structural motion linked to particular catalytic activity in the Man5B protein, in\r\nagreement with all-atom simulations. Furthermore, our model is based on the van der Waals radii, instead of a cutoff distance, which results in a smaller number of interactions compared to the ELNEDIN model. In conclusion, we anticipate that our model will provide further possibilities for studying biological systems beyond the scope of the ELNEDIN protein model.
    Back to the top of the page

  • Mutational perturbation of a fundamental biological switch
    Tina Perica
    UC, San Francisco

    Point mutations in primary protein sequence differentially affect structural and biochemical properties of a gene/protein and are thus expected to have quantitatively different effects on the phenotype. As a case study for how small changes in primary sequence affect the phenotype we chose Gsp1/Ran, a Ras superfamily member from S. cerevisiae. Gsp1/Ran is an appealing model protein because (i) it has a well-understood molecular function and works as a biological GTPase switch), (ii) is a cellular protein-protein interaction (PPI) hub, and (iii) is central to essential biological processes which include nuclear transport and cell cycle regulation.\r\n\r\nWe designed 56 Gsp1 point mutants spanning its protein-protein interfaces, and then measured their effects on the yeast phenotype by quantifying their genetic interactions with thousands of yeast genes. Our results show that the phenotypic effects of mutations do not cluster by position on the protein structure, or by the protein-protein interfaces they affect. Instead, they cluster by the main GTPase cycle regulators, such as the RanGAP and RanGEF. These data reveal that the importance of intrinsic allostery of a small GTPase for its function and regulation is higher than initially anticipated.\r\n\r\nFinally, we are building a comprehensive model of a molecular switch as a biological system by combining the genetic interaction data with biochemical data on protein-protein interactions and enzymatic kinetics parameters for nucleotide hydrolysis and exchange. We argue this approach is essential for understanding the biology of multifunctional, protein network hub proteins, as well as systems diseases such as cancer.
    Back to the top of the page

Contributed Talks

  • A simulation study of integrin alpha-2 I domain activation
    Zhe Jia
    Clemson University

    Integrin alpha-2 I domain, the ligand binding domain of the signaling protein - integrin alpha-2, has two known conformations. They are specified as open and closed conformation that correspond to high and low binding affinity to ligands, respectively. The closed conformation is more stable in the ligand free state, and the open conformation is more populated when bound with ligand. The conformation change from closed to open is known to be triggered by a directional force on the C-terminal. To explain the mechanisms of the conformation change, we built a model of the I domain bound state in the closed conformation (C-terminal in closed conformation, MIDAS binding site in partially open conformation) using steered molecular dynamics simulation. Combining correlated motion analysis with gene sequence analysis, we found a clear pathway that transduces the pulling force between the C-terminal residues and ligand binding site. Residues in the pathway are well conserved among the evolution of integrin alpha-2. Further, in MM/PBSA energy calculation, the I domain is more stable in the unbound closed conformation than in unbound open conformation, which is consistent with experiments. However, the electrostatic force analysis indicates that the binding site in the closed conformation has stronger attractive electrostatic force to the ligand when they are separated by 5-20Å. In our model of the closed conformation in ligand bound state, the stability of the partially closed conformation falls in between of the closed and the open conformation. It indicates the existence of an I domain ligand bound closed conformation as an intermediate state between the open bound state and the closed unbound state.
    Back to the top of the page

  • Protein misfolding and neurodegenerative diseases
    Gia Maisuradze
    Cornell University

    Neurodegenerative disorders, such as Alzheimer‚s, Parkinson‚s, Huntington‚s, and Creutzfeldt-Jakob‚s diseases, are caused by incorrect protein folding followed by aggregation and accumulation of protein deposits in neuronal cells. Moreover, protein folding intermediates are associated with formation of amyloid fibrils, which are responsible for neurodegenerative diseases. Therefore, elucidation of important sites and mechanisms for amyloid fibril formation, and the origins of formation of intermediates and finding ways to prevent them are very important. All these problems will be addressed in this presentation by investigating (i) folding trajectories of the 37-residue triple beta-strand WW domain from the Formin binding protein 28 (FBP28) and its six mutants (L26D, L26E, L26W, E27Y, T29D, and T29Y); (ii) the computationally-modeled entire process of Ab fibril elongation; (iii) molecular dynamics of two forms, monomeric and tetrameric, of alpha-synuclein. Experimental validation of theoretical findings in these studies will also be presented.
    Back to the top of the page

  • Protein-Ligand Binding Free Energy Calculations between T4 Lysozyme at 141 Small Molecules based on Multiple Rigid Receptor Conformations
    David Minh
    Illinois Institute of Technology

    Binding free energies between a flexible ligand and rigid receptor, exponentially averaged over the apo ensemble of the receptor, are equal to the standard binding free energy with the flexible receptor. My group is developing ways to accelerate protein-ligand binding free energy calculations based on this concept. I will present out recent work in receptor sampling, free energy calculations based on the fast Fourier transform, and the prediction of holo energy landscapes.
    Back to the top of the page

  • ScrubChem: Cleaning of PubChem Bioassay Data to Create Diverse and Massive Bioactivity Datasets for Use in Modeling Applications
    Jason Harris
    University of North Carolina Chapel Hill

    The PubChem Bioassay database is a non-curated public repository with bioactivity data from many sources, including: ChEMBL, BindingDb, DrugBank, Tox21, NIH Molecular Libraries Screening Program, and various academic, government, and industrial contributors. However, this data is difficult to use in data-driven research, mainly due to lack of interoperability and standardization among its 1.2 million assay records. Methods for extracting this public data into high-quality, computable datasets, useable for predictive and analytical research, presents several big-data challenges for which ScrubChem is being developed as a manageable solution. Our approach was to use logic-based text and language processing rules in order to digitally curate and correct the many issues related to the flexible deposition structure of PubChem (e.g., variable placement of biological target information, variable endpoint terminologies, result readouts not distinguished from non-result readouts, improper use of the null vs zero concept, incorrect spellings). Currently, ScrubChem contains approximately 694 million bioactivity values and related meta-data within PubChem and maps this data to 9 thousand biological targets and 2.3 million chemical structures. This work presents case issues identified and resolved through ScrubChem and provides an example dataset for the human androgen receptor with over 85 thousand reference bioactivities to further illustrate the results of the cleaning process. Data and updates are available at
    Back to the top of the page

  • Computational modeling of flexible protein-protein and protein-drug and interactions
    Daisuke Kihara
    Purdue University

    Many important cellular processes in a cell are carried out by protein-protein and protein-ligand interactions. We have developed a series of computational methods for predicting the tertiary structures for protein complexes and protein-ligand interactions. Among them, first we will present our recent method for predicting structures of docking conformation between a disordered protein and a structured protein. Protein-protein interaction with an intrinsically disordered protein (IDP) are prevalent in the cell, including important signaling and regulatory pathways. However, experimental methods have difficulty in solving disordered PPIs and existing protein-protein and protein-peptide docking methods are not able to model them. Our method, IDP-LZerD, was able to model structures with IDPs of up to 69 amino acid residue long by docking and connecting local fragments from IDPs. In the latter part of the talk, we present our molecular-surface-based binding ligand screening method, PL-Patch-Surfer2. PL-Patch-Surfer2 uses a novel mathematical surface representation of molecular surface, which performs well even when apo form of binding sites or computationally modelled binding sites are used for a target. Prediction by the computational method was combined with experimental ligand-binding screening using mass spectrometry to identify novel NAD binding proteins in the E. coli proteome. Both methods push the limitation of existing molecular modeling tools that have difficulty in dealing with flexibility of biomolecules.
    Back to the top of the page

  • Using Bayesian Optimization to Learn Parameters of Molecular Self-Assembly
    Marcus Thomas
    Carnegie Mellon University

    The ability of collections of molecules to spontaneously organize and assemble into large functional complexes is central in biology. Virus capsid assembly has long been an important model system for understanding general self-assembly, yet despite decades of experimental and theoretical investigation, important aspects of capsid assembly such as binding rate constants remain unknown. Using local rule based, discrete event simulation methods in conjunction with bulk indirect experimental data, we can meaningfully constrain the space of possible assembly pathways and allow inference of experimentally unobservable features of the real system. We assess optimality of binding rate parameters with an objective function quantifying their ability to produce simulations that agree with experimental scattering data. Here, we advance this strategy for data-driven model inference in two directions. First, we extend our prior work to encompass small-angle X-ray/neutron scattering (SAXS/SANS) as possibly richer experimental data sources than the previously used static light scattering (SLS). Second, we explore the problem of quantifying uncertainty in parameter estimation by constructing a probabilistic model of the objective using Gaussian process (GP) models. These models specify a prior on the space of all possible objective functions. As simulations at successive parameter values are completed, the prior is updated, forming the posterior which is used in prediction. New data points for sampling are selected based on the current properties of the GP and trade-offs between exploration and exploitation of the input space. This iterative Bayesian optimization approach is better able to quantify uncertainty in parameter assignments than our current models which are based on local surrogate functions. Additionally, the formalism allows for predictions at test points using information about the smoothness and self-similarity of the objective both locally and globally. We apply this strategy to mo! del fitting on synthetic SAXS data generated from a known ground truth.
    Back to the top of the page

  • Graph theoretic approaches to RNA topology comparison
    Michael Gribskov
    Purdue University

    RNA, along with DNA and proteins, is one of the fundamental macromolecules found in living systems. While it was once thought to be a primarily structural molecule, for instance, acting as a scaffold in the ribosome, it is now known to play a dynamic role in both catalysis and regulation within the cell. One of the most powerful approaches in molecular biology has been to identify evolutionarily conserved regions of macromolecules ˆ such conserved regions are the result of constraints on the allowable mutations, placed on the molecule by its structure and function. Approaches based on string comparison have been highly successful in understanding protein structure and function, and in identifying important DNA regulatory regions (such as transcription factor binding sites). However, in the case of RNA, the nucleotide sequence itself is often not conserved ˆ it is the pattern of internal base-pairs inside the molecule that is most important. This type of conservation cannot be detected using traditional string comparison methods.\r\n\r\nWe have developed a graph theoretic approach that describes the topology of RNAs (XIOS graph) in terms of the internally base-paired regions (stems) and the relationships between the stems. RNAs can be compared by identifying the greatest common subgraph in a group of RNAs. This is a difficult (NP-hard) problem for which I will present two solutions: the first relies on enumerating a dictionary of all biologically possible RNA topologies, and using this dictionary to characterize individual RNA graphs in terms of their subgraph spectrum. The subgraph spectrum, which we call an RNA fingerprint, can be used to identify common topologies in the absence of sequence similarity. The second approach couples a maximal frequent subgraph mining algorithm (MARGIN) with MCMC sampling to efficiently discover the longest common graphs in a topological graphset. Both of these approaches identify conserved frequent subgraphs in large complex graphs. In our a! pplication, these conserved subgraphs have a biological interpretation, but similar approaches should be applicable to many other kinds of complex graphs.\r\n\r\n Huang J, Li K, Gribskov M. Accurate Classification of RNA Structures Using Topological Fingerprints. PLoS One. 11:e0164726, 2016.\r\n Liu, M., Gribskov M. MMC-Margin: Identification of Maximum Frequent Subgraphs by Metropolis Monte Carlo Sampling. IEEE International Conference on Big Data, 2015. \r\n
    Back to the top of the page

  • Crystallographic Characterization of the Nitric Oxide and the C-Nitroso Derivatives of R-State Human Hemoglobin
    Jun Yi
    Nanjing University of Science and Technology

    Nitric oxide (NO), a critical signaling molecule to human, is biosynthesized and initiates its signal transduction cascade via several heme-containing proteins such as NO synthase and soluble guanylate cyclase. It is known that human hemoglobin could reduce inorganic nitrite to NO under hypoxic conditions and also metabolize organic nitro compounds to their C-nitroso (R-N=O; R = alkyl, aryl) derivatives in vivo. The latter interaction results in the formation of ferrous Fe(RNO) derivatives that have been characterized by electronic absorption spectroscopy.
    We have successfully solved the single crystal structure of R-state human HbNO derivative obtained by soaking the precursor R-state Hb(ONO) crystal in the presence of dithionite for 5 minute. The differences in the FeNO bond parameters and H-bonding patterns between the alpha and beta subunits may contribute to understanding of the observed enhanced stability of the alpha(FeNO) moieties relative to the beta(FeNO) moieties in the human R-state HbNO.
    We have reinvestigated the interactions of alkyl C-nitroso compounds (MeNO and EtNO) with human hemoglobin, and have characterized the products by single crystal X-ray crystallography. Both the Hb(MeNO) and Hb(EtNO) products display N-binding of the C-nitroso compound with heme Fe. The Hb(MeNO) crystal structure shows the MeNO ligand in both the alpha and beta subunits. In contrast, the Hb(EtNO) crystal structures shows the EtNO ligand bound to Fe in the alpha subunit but not the beta subunit. The complexity of the beta subunit structure is evident by the observed 4.9 Å heme slippage towards the exterior of the protein. In the future work, molecular dynamic simulations will be carried out to investigate the transition processes of the heme loss of human Hb induced by C-nitroso compound
    Back to the top of the page


  • Mechanistic insights into HIV-1 protease inhibition by liphophilic adamantyl P1 ligands through molecular dynamics simulations
    Konda Reddy Karnati
    Albany State University

    The HIV-1 protease is an important drug target for HIV/AIDS therapy. HIV-1 protease performs an essential role in viral maturation by processing specific cleavage sites in the Gag and Gap-Pol precursor polyproteins to release the mature proteins. To date, the Food and Drug Administration (FDA) has approved nine inhibitors against protease and several other inhibitors are in development. Meanwhile, the high mutation rate of the virus causes a very high degree of adaptability. In the present study, molecular dynamics (MD) simulations of HIV-1 protease with three adamantyl P1 ligands 1) GRL-097-13A (bis-THF in P2 and isobutylamine in P1‚), 2) (GRL-007-14A (bis-THF in P2 and benzylamine in P1‚), and 3) GRL-031-14A ( tetrahydropyrano-tetrahydrofuran in P2 and isobutylamine in P1‚) were performed extensively (lasting 500 ns each) in order to explore their selectivity. The size and steric bulk of the adamantane group is quite different from the phenyl ring inherent to many potent protease inhibitors, including darunavir and TMC-126. The adamantane group form stable interaction with hydrophobic groups of HIV-1 protease active site. All three ligands also formed stable van der Waals interactions with 80s loop residues, flap residues and Gly27 from the catalytic triad. The key hydrogen interactions of the bis-THF moiety of P1 ligands with the Asp29 and Asp30 residues were quite stable throughout simulations. The binding free energies that are calculated by molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) method agrees well with the experimental data. The present study sheds light on the dynamics of HIV-1 protease inhibition by a liphophilic adamantyl P1 ligands, providing useful information to the design of more potent and effective HIV-1 protease inhibitors.
    Back to the top of the page

  • Unfolding and Translocation Mechanism of Green Fluorescent Protein Mediated by the ClpB Biological Nanomachine
    Rui Jiang
    University of Cincinnati

    Proteins form toxic aggregates under environmental pressure, causing neurologic diseases such as Parkinson‚s disease. Protein disaggregation and degradation by the ClpB nanomachine, a member of the HSP100 ATPase family, are thus crucial to prevent lethal protein fibrillization and maintain cellular homeostasis. We study the molecular mechanisms of ClpB-mediated disaggregation by using computer simulations of threading Green Fluorescent Protein (GFP) substrates through the central pore of ClpB. We observe a stable Δβ11 GFP intermediate during the unfolding process
    Back to the top of the page

  • Computer Simulations of the Dynamics of IgG Antibody Binding-Region
    Yasan Fonseka
    University of Cincinnati

    IgG Antibody is the most common antibody in human circulation. This Y-shaped biomolecule contains two heavy chains and two light chains. Each heavy (light) chain has four (two) Ig-like domains. In the V-shaped Fab region of IgG, each light chain binds to a heavy chain to form the arms of the structure. The top region of each arm participates in binding to Antigens. Disulfide bonds in the flexible hinge region which connect two heavy chains affect the dynamics the Fab region and IgG binding affinity. The molecular details of the role of disulfide bonds on the flexibility of Fab region is not fully understood yet. Here we use an implicit solvent model of IgG1 and IgG4 antibodies and perform Langevin dynamics simulations to decipher the effect of hinge-region disulfide bonds on the flexibility of Fab region. We find that the presence of disulfide bonds in the hinge region of IgG1 and IgG4 stabilizes the Fab region by limiting the motions of the arms which results in decreasing the distance between the binding regions.
    Back to the top of the page

  • Coarse-grained Simulations of Green Fluorescent Protein Unfolding Mediated by 26S Proteasome
    Mohammad Sadegh Avestan
    University of Cincinnati

    The Ubiquitin-Proteasome System (UPS) participates in maintaining cell viability by unfolding and degradation of ubiquitinated substrates. 26S proteasome is a major eukaryotic ATP-dependent molecular motor which contains more than 30 different subunits arranged in two functional groups, core particle (CP) and regulatory particle (RP). The barrel shaped 20S-CP comprises 4 heptameric rings that are stacked axially. Each end of 20S-CP is capped by a 19S-RP containing a heterohexameric AAA+ ATPase ring that unfolds Ubiquitinated substrates by threading them through its central pore and delivers them to 20S-CP. The molecular details of protein unfolding by 19S-RP are not fully understood. Here, we use a coarse-grained model of 19S-RP and perform Langevin dynamics simulation of mechanical threading of the Green Fluorescent Protein (GFP) through the central pore of 19S-RP. We compare and contrast 19S-mediated unfolding of GFP with mechanical pulling of the substrate protein along the N-C direction to probe direction-dependent unfolding of GPF.
    Back to the top of the page

  • Photochemistry of Coenzyme B12-Dependent Ethanoline Ammonia Lyase
    Abdullah Al Mamun
    University of Louisville

    Adenosylcobalamin is a highly complex organometallic molecule that acts as a cofactor in numerous enzymatic reactions. The initial step involves the homolytic cleavage of the cofactor‚s Co-C bond leading to the formation of reactive radical intermediates. This Co-C bond can be photolyzed with visible or near UV light through a complex photochemical reaction. The specific mechanism of photolysis depends upon whether the cofactor‚s environment is solvent or enzyme. The mechanism is not yet clearly understood in the enzymatic environment. A detailed understanding of this photolytic mechanism requires the combination of both theoretical/computational and experimental study. In this project, methods of computational chemistry were applied to gather detailed knowledge about the potential energy surface (PES) and the electronically excited states. DFT calculations combined with molecular mechanics (QM/MM) were performed to construct the ground state potential energy surface (PES). TD-DFT calculations were also performed to obtain the low-lying singlet and triplet excited states. The key step in photolysis involves singlet radical pair generation from the first electronically singlet excited state (S1). The S1 PES constructed by constraining Co-C and Co-NIm axial coordinates of adenosylcobalamin, is characterized by the crossing of two states S1 and S2 that are separated by a seam between two minima. These minima indicate metal-to-ligand charge transfer (MLCT) at the partially elongated axial bonds and the ligand field (LF) at the elongated axial bonds. These results indicate that the possible pathway of photolysis is likely the base-off form in the enzymatic environment. This is initiated by the elongation of the Co-NIm bond followed by the elongation of both axial bonds Co-C and Co-NIm
    Back to the top of the page

  • PU.1 and ETS-1 sequence specificity divergence through differential ETS-DNA complex hydration
    Kathleen Carter
    Georgia State University

    The DNA-binding domains of ETS family transcription factors are highly structurally homologous despite strongly divergent primary sequences. At opposite ends of the ETS sequence homology spectrum lie the PU.1 and Ets-1 ETS family members (~30% sequence homology). This relatively low sequence homology results in high functional divergence, with PU.1 and Ets-1 exhibiting strongly differing preferences for DNA binding site sequence, despite their very high structural homology (~1.4Å RMSD). While experiments have shown that PU.1 and Ets-1 sequence preference differences are driven at least partially by hydration and salt interactions, the dynamical and conformational differences underlying these effects have remained elusive. We have applied Grid Inhomogeneous Solvation Theory (GIST) to molecular dynamics trajectories to monitor hydration energetics of in a set of PU.1-DNA and Ets-1-DNA complexes differing by PU.1 binding affinity, finding that increases in hydration in the GIST analysis correlate well with increases in PU.1 binding affinity. Dynamical network analysis reveals that these differences in hydration are coupled to allosteric community formation in the DNA. Higher PU.1-affinity sequences exhibit more coherent dynamical DNA communities in PU.1.
    Back to the top of the page

  • Understanding the Allosteric Mechanism Behind the E3 Ligase Parkin
    Thomas Dodd
    Georgia State University

    The ubiquitin-proteasome pathway, conserved from yeast to mammals, is an important process required for the targeted degradation of most shortlived proteins in the eukaryotic cell. The main targets include regulatory proteins, as well as proteins unable to fold properly within the endoplasmic reticulum. E3 ligases facilitate the transfer of ubiquitin from an ubiquitin conjugating enzyme (E2) to a target protein, initiating the degradation process. Parkin is a RING-between-RING (RBR) E3 ligase that functions in the covalent attachment of ubiquitin to specific substrates. Mutations in Parkin are linked to Parkinson‚s disease, cancer and mycobacterial infection. We have modeled Parkin in complex with an E2-ubiquitin conjugate using existing crystallographic data of HOIP, an E3 ligase within the same RBR subfamily and structurally similar to that of Parkin. Initial molecular dynamics simulations of Parkin indicate that the activated ubiquitin binds to the RING2 domain of Parkin via a hydrogen-bonding network similar to that of HOIP. These data suggest that all RBRs have a catalytic domain capable of accommodating the di-Arginine motif found on ubiquitin. Through dynamic network analysis, we hope to elucidate key details behind the allosteric mechanism that allows Parkin to bind E2-ubiquitin conjugates.
    Back to the top of the page

  • Exploring the Molecular Mechanisms of Microtubule Severing
    Rohith Anand Varikoti
    University of Cincinnati

    Microtubules (MTs) are the filamentous intracellular biopolymers of a globular protein, called tubulin which is a dimer comprising of two polypeptides, alpha-tubulin and beta-tubulin (tubulin dimers). MTs along with ~300 associated proteins (called microtubule-associated proteins, MAPs), regulate various vital cellular processes. MAPs are involved in nucleation, depolymerizing or severing of protofilaments. The MT severing enzymes such as katanin and spastin, are part of the ATP-dependent homo-hexamerases of ATPases associated with various cellular activities (AAA+) family of enzymes. Spastin is a homo-hexameric protein with a center pore which binds to MTs in different orientations. Its interactions with the acidic residue rich carboxy-terminal tails (CTTs) of tubulins, are important for both binding and severing. The binding mechanism and transfer of CTTs through the center pore of spastin is unknown. A combination of docking studies with Molecular Dynamics simulations and Normal Mode Analysis are used to investigate the initial binding and translocation of CTTs to spastin.
    Back to the top of the page

  • Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in the DNA Damage Response
    Chunli Yan
    Georgia State University

    Proliferating cell nuclear antigen (PCNA) is a pivotal replication protein, which also controls cellular responses to DNA damage. Posttranslational modification of PCNA by SUMO and ubiquitin modulate these responses. How the modifiers alter PCNA-dependent DNA repair and damage tolerance pathways is largely unknown. We used hybrid methods to identify atomic models of PCNA(K107)-Ub and PCNA(K164)-SUMO consistent with small-angle X-ray scattering data of these complexes in solution. We show that SUMO and ubiquitin have distinct modes of association to PCNA. Ubiquitin adopts discrete docked binding positions. By contrast, SUMO associates by simple tethering and adopts extended flexible conformations. These structural differences are the result of the opposite electrostatic potentials of SUMO and Ub. The unexpected contrast in conformational behavior of Ub-PCNA and SUMO-PCNA has implications for interactions with partner proteins, interacting surfaces accessibility, and access points for pathway regulation.
    Back to the top of the page

  • Normal Mode Analysis of Conformational Changes in the ClpP Peptidase
    Qi Wang
    University of Cincinnati

    Maintaining homeostasis at the protein level is prerequisite for cellular viability and functionality. Caseinolytic peptidases (ClpP) are molecular machines responsible for cell homeostasis in bacteria. The ClpP tetradecamer comprises two stacked heptameric rings that encloses a roughly spherical hollow cavity with 14 protease active sites. Upon formation of complexes between ClpP and the hexameric ClpA or ClpX ATPases, substrate proteins are unfolded and translocated through the ClpP axial pores into the proteolytic chamber for degradation. In the absence of ATPase components, ClpP can only degrade small peptides. When associated with Clp ATPases, ClpP peptidase is activated and undergoes significant conformational changes between distinct functional states, resulting in the release of substrate segments in the degradation chamber. Recent studies hypothesize that the degraded proteins are released from the chamber through equatorial pores that are formed due to conformational switching of ClpP. In this work, a normal mode analysis (NMA) method is used to study the conformational changes of the monomer and allosteric motions of the tetradecamer of ClpP from open state to a zinc-binding state. We observed a large deformation of the axial loop and handle domains in the monomer normal mode, while the allosteric transitions of the tetradecamer are mainly contributed by conformational changes of the handle region. We also found moderate inter-ring cooperativity and weak intra-ring cooperativity of the double-ring structure. Overall, our results support a model that ClpP releases degraded peptides through the equatorial regions.
    Back to the top of the page

  • Simulating Protein Fold Switching by Replica Exchange with Tunneling
    Nathan Bernhardt
    University of Oklahoma

    Past experiments have demonstrated the importance of amino acid sequence in selecting the native fold as well as non-native forms a protein exhibits. These various folds may be important to the function of the protein or in its ability to fold. Such forms populate a multi-funnel energy landscape where mutations or changes to the environment can alter how each fold is populated. A classic example of this was demonstrated in experiments by Bryan Orban in which 2 subdomains of protein- G, GA and GB, underwent multiple mutations to increase sequence identity. GA and GB are known to take very different folds with GA having a 3 helix bundle and GB 4 beta sheets and a single alpha helix interacting along one of the beta sheet faces. After each mutation variants were assessed for functional loss as well as structural alteration. Orban was able to pinpoint a single mutational difference at residue 45 shown to be critical in selecting one fold or the other (Leucine for the GA fold and Tyrosine for the GB fold). These mutational variants were termed GA98 and GB98. To determine how a single mutation selects such different folds we use Hamiltonian Replica Exchange Molecular Dynamics to study the free energy landscape of both peptide sequences. We find that in accordance with experiment the GA fold is favored in the GA98 sequence over the GB fold. However, we find that the GA and GB folds of GB98 are similarly populated with a large energy barrier separating them. Such findings suggest that the GA fold of GB98 may be kinetically difficult to observe in experiment.
    Back to the top of the page

  • Exploring the mechanics of a muscle complex involved in mechanotransduction
    Nicholas Bodmer
    Washington University

    The main anchoring point of the megadalton protein titin in the Z-disc of the sarcomere is a complex composed of two titin chains and telethonin, an immunoglobulin domain with an unusual fold sandwiched in between the titin chains. Experimental and computational investigations established that the titin-telethonin complex resists to unusually high forces along the direction of the C-termini of the titin chains. In contrast, the N-termini direction is weak. The sarcomere is one of the most crowded cellular environments, resulting in a large number of interactions between the titin-telethonin complex and other proteins. Interestingly, most interacting partners target telethonin, whose exact role in the complex is still unclear. To shed light on the origin of the directional response of the titin-telethonin complex, as well as the role of telethonin, we followed the mechanical response of the complex using simulations based on a minimalist protein model. Previously, this model, combined with results from single molecule experiments, proved successful in elucidating the molecular factors leading to the C-termini mechanical behavior of the complex. Employing a similar approach, our present simulations show that titin-telethonin displays considerable kinetic partitioning when pulled along the N-termini orientation, as several distinct pathways emerge across a single order in pulling speed, which is in stark contrast with the pathway conservation exhibited by the complex along the C-termini direction. We found that the complex can be stabilized mechanically by interactions with other proteins in the Z-disc that form interaction surfaces such as the ones found in the dimer structure of the complex. Importantly, our simulations indicate that telethonin resists to high forces when torque is applied to the complex, supporting its proposed role in torque mechanotransduction.
    Back to the top of the page

  • Computer Simulations of Protein Remodeling by Ring-Shaped Biological Nanomachines
    Abdolreza Javidialesaadi
    University of Cincinnati

    Ring-shaped AAA+ (ATPases Associated with diverse cellular Activities) motors, such as Bacterial Caseinolytic proteases (Clp), support cell viability by mediating protein degradation mechanisms. Cycles of ATP-driven allosteric motions of ring subunits result in alternating between „open‰ and „closed‰ configurations of the central pore of the AAA+ machine. These conformational changes are coupled with mechanical forces applied to the substrate protein (SP) to promote SP threading through the central channel. Here, we perform multiscale molecular dynamics simulations of the ClpYΔI ATPase and Titin I27 to compare and contrast ClpY-mediated SP unfolding and translocation as probed in laser optical tweezer (LOT) experiments, in which the SP N-terminus is restrained, with in vivo mechanisms, in which N-terminus is not restrained. The LOT setup restricts ClpY-mediated pulling along the N-C direction of the SP, which results in unfolding via a shearing mechanism. By contrast, in vivo-like ClpY-action results in pulling along softer mechanical directions of I27 and unfolding occurs via an unzipping mechanism. We find that factors that affect these distinct mechanisms are SP-ClpY surface interactions, the size of the SP relative to the ClpY pore size, the SP mechanical resistance, and the presence of other substrate domains.
    Back to the top of the page

  • Absolute binding free energies between T4 lysozyme and 141 small molecules: calculations based on multiple rigid receptor structures
    Bing Xie
    Illinois Institute of Technology

    Free energy calculations predict binding affinities of noncovalent binding partners, which can play an important role in drug design. However, the calculations are difficult due to insufficient sampling, solvation effects, and limited time. Our group is developing a new free energy prediction program, AlGDock , which is based on implicit ligand theory . This approach combines advantages of molecular docking and alchemical free energy calculations, obtaining results more accurate than the former and more quickly than the latter. Here, we apply this new method to estimate the affinity between T4 lysozyme and 141 different small molecules. Calculated binding free energies are highly correlated to results from YANK , which uses a flexible receptor. Primarily due to a more advanced force field, the method is better than DOCK6 in distinguishing active and inactive molecules.
    Back to the top of the page

  • Probing the Mechanism of Protein Unfolding and Translocation by the ClpY ATPase during Protein Degradation
    Yu-Hsuan Shih
    University of Cincinnati

    Clp ATPases play indispensable roles in the protein quality control system by dispatching stress-encountered aggregates, which are often associate with neurodegenerative diseases, such as Alzheimer‚s, Parkinson‚s diseases. To guard against such deleterious pathway, Clp ATPases utilize the energy from ATP hydrolysis to induce large conformational changes and extrude the substrate proteins (SPs) through their scaffold. The detailed mechanisms underlying these powerful ATP-driven nanomachines still remain unclear. In this study, we developed a novel coarse-grained model coupled with Langevin dynamics simulations and normal mode analysis (NMA) to pinpoint the mechanical response of the green fluorescent protein (GFP) to the action of the ClpY ATPase during protein degradation. Our simulations using NMA cooperate the dynamical open and closed motions of ClpY with SPs during degradation. In addition, we employ the mechanical pulling to simulate such large ClpY nanomachine at biologically relevant timescales. We observed that SP unfolding and translocation mechanism catalyzed by ClpY reveal stepped transitions due to the local mechanical resistance. Unraveling GFP from C-terminus transits into two major meta-stable intermediateswhich are consistent with single-molecule experiments. On the other hand, unraveling from N-terminus of GFP results in decomposing the central helix which facilitate to destabilize the rest of strands. We also compared GFP wild type to the engineered circular GFP permutants, providing information for the mechanical stiffness along the corresponding opening. This finding indicates that mechanical resistance of partially unfolded intermediates is correlated with the topology of local domains, rather than the global SP fold. This study serves to help bridge the gap between experiments and simulations, providing vivid microscopic details on the degradation capability of Clp proteases.
    Back to the top of the page

  • Conformational Switching and Amyloidogenesis Mechanism Study of N-terminal of Serum Amyloid A
    Wenhua Wang
    University of Oklahoma

    Amyloidosis amyloid occurs as a reaction to chronic inflammatory, infectious, and neoplastic diseases, such as rheumatoid arthritis, inflammatory bowel disease, tuberculosis, and renal cell carcinoma. Serum amyloid A protein (SAA), a major acute-phase protein, is the main constituent of secondary amyloidosis in amyloidosis protein family. Previous experimental results reveal the SAA1.1, one of the typically protein in SAA family, assemble into hexamer complex with mostly α-helical conformation which is conflict from other amyloid fibrils characterizing as highly ordered β-sheet structures. Here we study the N-terminal region of SAA, which is regarded as the aggregation section, with replica exchange molecular dynamics(REMD) simulation and multi-exchange methods. The first thirteen residues show less stability than to the other residues in first 27 residues and β-sheet structures have been found in this region. Our results reveal the highly hydrophobic N-terminal helix is particularly crucial to amyloidogenesis and the mechanism of aggregation related to conformational transition from α-helix to β-strand in this region.
    Back to the top of the page

  • The fast Fourier transform accelerates binding free energy calculations between T4 lysozyme and 141 small molecules
    Trung Hai Nguyen
    Illinois Institute of Technology

    We apply the fast Fourier transform (FFT) method to calculate the binding free energies for a set of 141 T4 lysozyme - ligand complexes. In this approach, the ligand is sampled from the apo state using MD simulations and random rotations. Then the FFT correlation method is used to compute interaction energies for a set of discrete translational positions of the ligand relative to the receptor. Then an exponential average of the interaction energy over the ligand translations, rotations and conformations and receptor conformations gives the standard binding free energy. This approach is computationally efficient and the results are in good agreement with other methods such as YANK and AlGDock. To our knowledge, this is the first time FFT method has been used to estimate binding free energies of protein-ligand complexes.
    Back to the top of the page

  • Building Atomistic Models from SAXS Experiments Using Bayesian Refinements of Accelerated Molecular Dynamics Trajectories
    Samuel Bowerman
    Illinois Institute of Technology

    Small angle X-ray scattering (SAXS) experiments capture the full solution ensemble of flexible biomolecules, but SAXS data is typically low-dimensional and difficult to interpret without additional structural knowledge from complementary techniques. In principle, this information may be provided by molecular dynamics (MD) simulations, but conventional MD (cMD) trajectories are often restricted to local minima. One method for enhancing MD sampling is accelerated MD (aMD), which reduces the height of potential energy barriers by introducing a bias to the underlying energy landscape, and unlike many similar techniques does not require the definition of a reaction coordinate. However, the bias potential disturbs the Boltzmann distribution, and determining the appropriate populations of states directly from an aMD simulation can be difficult, especially for large complexes. Therefore, it is beneficial to use these two techniques together, SAXS and aMD, to produce atomistic-level models of biomolecular solution ensembles. Here, we present a method for fitting aMD simulations of poly-ubiquitin trimers with SAXS data using an iterative Bayesian procedure. Candidate scattering states are first identified from MD trajectories, and then their populations are re-weighted against empirical data using a Bayesian Monte Carlo approach. Resistance to ensemble over-fitting is achieved by iteratively considering increasing subsets of scattering states and by reducing experimental data to the Shannon sampling limit. Using this protocol, we find that aMD simulation can be used to produce higher quality models in shorter timescales than standard cMD simulations. We analyze several different trimer linkages and find that the location of each linkage is strongly coupled to the populations of the resulting models. This site-specific alteration of solution flexibility may play a key role in the disparate roles of each linkage in cellular signaling and processes.
    Back to the top of the page

  • Comparing NMR and X-ray protein structure: Lindemann-like parameters and NMR disorder
    Eshel Faraggi
    Indiana University

    Disordered protein chains and segments are fast becoming a major pathway for our understanding of biological function, especially in more evolved species. However, the standard definition of disordered residues: the inability to constrain them in X-ray derived structures, is not easily applied to NMR derived structures. We carry out a statistical comparison between proteins whose structure was resolved using NMR and using X-ray protocols. We start by establishing a connection between these two protocols for obtaining protein structure. We find a close statistical correspondence between NMR and X-ray structures if fluctuations inherent to the NMR protocol are taken into account. Intuitively this tends to lend support to the validity of both NMR and X-ray protocols in deriving biomolecular models that correspond to in-vivo conditions. We then establish Lindemann-like parameters for NMR derived structures and examine what order/disorder cutoffs for these parameters are most consistent with X-ray data and how consistent are they. Finally, we find critical value of $L=4$ for the best correspondence between X-ray and NMR derived order/disorder assignment, judged by maximizing the Matthews correlation, and a critical value $L=1.5$ if a balance between false positive and false negative prediction is sought. We examine a few non-conforming cases, and examine the origin of the structure derived in X-ray. This study could help in assigning meaningful disorder from NMR experiments.
    Back to the top of the page

  • RNA Mediated Conversion of Prions
    Erik Alred
    University of Oklahoma

    Diseases associated with prion proteins involve the induction of fold of mis-folded proteins onto the native form. This transition is rate-limited by the initial conversion of native prion protein to the mis-folded state in the absence of mis-folded templates. Recent experiments indicate that polyadenosine RNA could be an agent in the initial generation of seed structures. In order to explore the RNA-induced transition, we dock polyadnesione RNA to two different prion sequences, mouse and human. By comparing the change in the hydrogen bonding network, as well as secondary structure, between docked and undocked systems we are able to suggest mechanisms that lead to the initial generation of mis-folded prion structures.
    Back to the top of the page

  • Eliminating a Protein Folding Intermediate by Tuning a Local Hydrophobic Contact
    Khatuna Kachlishvili
    Cornell University

    Intermediate states in protein folding may slow folding, and sometimes can provide a starting point for aggregation. Recently, the origins of formation of an intermediate state involved in amyloid formation and ways to prevent it were illustrated with the example of the Formin binding protein 28 (FBP28) WW domain, which folds with biphasic kinetics [1]. Here, we combine new simulations over a broad temperature range with experimental temperature-jump data to study this site in more detail. We replace Leu26 by Asp26 or Trp26 to alter the folding scenario from three-state folding towards two-state or downhill folding at temperatures below the melting point, whereas the wild type shows two-state behavior only near its melting temperature. We offer an explanation of this behavior mainly in terms of principles of hydrophobic interactions.
    1. G.G. Maisuradze, J. Medina, K. Kachlishvili, P. Krupa, M.A. Mozolewska, P. Martin-Malpartida, L. Maisuradze, M.J. Macias and H.A. Scheraga, Proc. Natl. Acad. Sci. USA, 112, 13549-13554, 2015.
    Back to the top of the page

  • Fluctuations across subdomains reveal hinge mechanics of the Hsp70 substrate binding domain
    Dale Merz
    University of Cincinnati

    The regulation of protein function through ligand-induced conformational changes is crucial for many signal transduction processes. Heat shock protein (70 kDa) is one example of a protein undergoing conformational changes upon ligand hydrolysis. It comprises a nucleotide binding domain (NBD) covalently linked to a substrate binding domain (SBD) via an interdomain linker. Using a combination of coarse-grained pulling molecular dynamics simulations [1] along with corresponding single molecule experiments, we uncovered the mechanics of the SBD during ligand binding, hydrolysis, and exchange of ATP. Fluctuations in both the alpha and beta subdomains of the SBD and flipping kinetics between the two during the pulling experiments were determined to be a mechanical communication across the connecting loop. Stabilization in the latter β-sheet induces folding in Helix A and B and is driven allosterically by hydrolysis of ATP in the nucleotide binding domain[2].
    [1] Zhmurov, A., Dima, R. I., Kholodov, Y., & Barsegov, V. (2010). Sop-GPU: accelerating biomolecular simulations in the centisecond timescale using graphics processors. Proteins, 78, 2984–99.
    [2] Nanomechanics of the substrate binding domain of Hsp70 determine its allosteric ATP-induced conformational change. Soumit S. Mandal, Dale R. Merz Jr., Maximilian Buchsteiner, Ruxandra I. Dima, Matthias Rief, and Gabriel Žoldák. Submitted PNAS.
    Back to the top of the page

  • Deciphering General Characteristics of Residues Constituting Allosteric Communication Paths
    Girik Malik
    Nationwide Children Hospital

    There has been great interest in studying proteins‚ dynamics and structural changes involved in function to gain insights into the machinery of proteins. Due to difficulty in retaining atomic details in mode decomposition of large system dynamics, there have been significant computational challenges, which make the study of large system dynamics very complex.\r\n\r\nConsidering all the PDB annotated proteins from the Allosteric Database (ASD) [1,2] belonging to four different classes (kinases, nuclear receptors, peptidases, transcription factors), this work has attempted to decipher certain consistent patterns present in the residues constituting the allosteric communication sub-system (ACSS). While the thermal fluctuations of hydrophobic residues in ACSSs were found to be significantly higher than those present in the non-ACSS part of the same proteins, thermal fluctuations recorded for the polar residues showed the opposite trend.\r\n\r\nWhile the basic residues and hydroxyl residues were found to be slightly more predominant than the acidic residues and amide residues in ACSSs, hydrophobic residues were found extremely frequently in kinase ACSSs. Despite having different sequences and different lengths of ACSS, they were found to be structurally quite similar to each other, suggesting a preferred structural template for communication.\r\n\r\nACSS structures recorded low RMSD and high Akaike Information Criterion (AIC) scores among themselves. While the ACSS networks for all the groups of allosteric proteins recorded low degree centrality and closeness centrality, the betweenness centrality magnitudes revealed non-uniform behavior. Though cliques and communities could be identified within the ACSS, maximal-common-subgraph considering all the ACSS could not be generated, primarily due to the diversity in the dataset. Barring one particular case, the entire ACSS for any class of allosteric proteins did not demonstrate „small world‰ behavior, except for a few sub-graphs.\r\n\r\nSuc! h a report can be expected to benefit the protein engineering community, those who attempt to decipher the general mechanism of allostery, and in general, long-distance communication within protein structures, both from knowing the topological invariants of communication paths and from knowing the biophysical-biochemical-and-structural patterns therein.\r\n\r\n[1] Huang Z, Zhu L, Cao Y, et al. ASD: a comprehensive database of allosteric proteins and modulators. Acids Res. 2011; 39 (Database issue):D663-D669.\r\n\r\n[2] Huang Z, Mou L, Shen Q, et al. ASD v2.0: updated content and novel features focusing on allosteric regulation. Nucleic Acids Res. 2014; 42 (Database issue):D510-D516.
    Back to the top of the page

  • Small Molecule Effects on Surfactant Microstructure and Dynamics: A Case Where MD and NMR Drive Understanding in How Partitioning Influences Physical Properties
    Allison Talley
    Univeristy of Cincinnati

    The physical properties of surfactant and amphiphile aggregate structures, for example biological membrane bilayers or consumer cleaning products, are strongly modulated by the presence\r\nof guest molecules such as cholesterol, drugs, or flavors/fragrances. The ability to study structure and dynamics in these systems, and to assess the impact of guest molecules, can aid\r\nunderstanding the origin of many properties ranging from the mechanical strength and curvature of membranes to the viscosity, stability, and consumer appeal of soaps and detergents. NMR\r\nrelaxometry provides standard experimental approaches to studying dynamics and structure in systems where the overall global structure is well defined (as in folded proteins and nucleic\r\nacids). However, in conformationally and configurationally dynamic systems like surfactant and amphiphile aggregates, and in related structured fluids, interpretation of NMR relaxation\r\ntechniques becomes more difficult because of the complex time- and structure-dependence of interatomic interactions. Here we present an approach where atomistic molecular dynamic\r\nsimulations of amphiphile aggregates are used in conjunction with NMR relaxometry to understand the structure and dynamics of amphiphile aggregates and their guests.
    Back to the top of the page

  • Molecular Dynamics Study of Ring-like N-fold Models of Abeta42 fibrils
    Wenhui Xi
    University of Oklahoma

    The brains of patients with Alzheimer‚s disease are characterized by the presence of amyloid fibrils that are made of amyloid-β (Aβ) peptides, most commonly Aβ1ˆ40 peptides but also the more toxic Aβ1ˆ42 species. The structures of Aβ peptides are polymorphism. At least five different structures of Aβ1-40 amyloid fibrils have been resolved by solid state NMR (ss-NMR), all sharing as a common U-shape like motif that the individual chains form two β-strands connected by a loop region. In 2015 a high-resolution model of Aβ1-42 amyloid fibrils are developed with experimental methods (PDB id: 2MXU) and this assembling motif arrange as S-shaped three-stranded chains. Our previous study proved that the S-shape motif of Aβ peptide are incompatible with Aβ1-40 which can only assume U-shaped conformations. This allows Aβ42 peptides to assemble pore-like structures that may explain their higher toxicity. Basing on this hypothesize, we propose a scalable model of ring-like assemblies of S-shaped Aβ1ˆ42 chains based on the packing of inter-chain salt-bridges between residue K16 and residues E22/D23. Using extensive molecular dynamics(MD) simulations, we investigate the stability and structural properties of these N-fold assemblies with symmetric of N=4,5,6,12. All of them are stable in our simulations and three-fold Aβ1-42 fibril model is more stable than the patient-derived Aβ1-40 fibril model (PDB-Id: 2M4J) that has the same three-fold symmetry. The outer diameters that are comparable with experimentally observed fatty-acid-catalysed oligomers and with pore-forming Aβ assemblies seen in membranes. As our assemblies are characterized by a unique pattern of inter-chain sidechain electrostatic interactions between residues K16 and either E22 or D23, they could be experimentally identified by the resulting specific NMR signal. We conjecture that the ability to form these arrangements with high N-fold symmetry may expla! in the high toxicity of Aβ1ˆ42 amyloids.
    Back to the top of the page

  • Back to the top of the page

  • Comparative analysis of the structural determinants of endogenous cannabinoids and activity of illicit drugs on Cannabinoid receptors
    Temple University

    Cannabinoid (CB) receptors belong to G protein-coupled receptor (GPCR) family and are activated by endogenous, phytogenic and synthetic modulators. There are two known subtypes of CB receptors, CB1 and CB2. Till date, there is no available crystal structure for CB receptors, and therefore we used multiple template comparative homology modeling algorithm to construct 3D models for CB1 and CB2. Cannabinoid receptor type-2 (CB2) plays vital roles in various physiological events like pain transmission, immune and neurodegenerative conditions. We built the 3D structure of CB2 using Prime. CB2 ligands were prepared and docked into the generated receptor grid. Induced fit protocol (IFD) were used for flexible protein docking of four ligands. Molecular dynamics with DESMOND were performed to monitor the protein-ligand interactions and to define the commonly involved amino acid residues in ligand binding. Molecular properties of the ligands including molecular, polar and solvent accessible surface areas, and intra-molecular hydrogen bonds were evaluated throughout the course of molecular dynamics simulations.
    Back to the top of the page

  • Predicting binding free energy change caused by missense mutations in protein-DNA interactions using modified MM/PBSA method
    Yunhui Pen
    Clemson University

    ProteinˆDNA interactions are essential for the regulations of many important cellular processes, such as transcription, replication, recombination and translation. Missense mutations occurring in DNA-binding protein have profound effect on protein-DNA interactions and are related with many diseases. In particular, the effect of mutations on binding affinity is one of the most important component of the overall disease effects. Hence, accurate prediction of missense mutations\\\' effect on proteinˆDNA binding affinity is essential to understanding disease-causing mechanism and guiding protein engineering. A new methodology is developed to accurately predict the changes of the binding free energy caused by missense mutations on protein-DNA interactions. This new method utilizes modified molecular mechanics Poisson-Boltzmann Surface Area (MM/PBSA) approach along with an additional set of knowledge based terms delivered from investigation of the physico-chemical properties of protein-DNA complexes. An experimental dataset (combined ProNIT database with recent references) of experimentally determined binding free energy changes caused by 105 mutations in 13 proteins is used for training and benchmarking, resulting in a good agreement between experimental data and prediction.(correlation coefficient of 0.72)
    Back to the top of the page

  • Binding of Lisinopril and Indicators to Barrel Aβ Structures
    J. Alexander Yates
    University of Oklahoma

    The proper folding of complex proteins is vital to life in all species. When proteins fold improperly, their biochemical function is significantly altered, leading to a variety of illnesses, the worst of which can be life-threatening. An example of such a protein which, when folded improperly, can have severe consequences is the Beta Amyloid (Aβ). Denatured Aβ structures have been linked to the onset and progression of Alzheimer‚s Disease. As such, understanding their folding mechanisms, what causes them to fold improperly, and how to prevent such is vital to the treatment of that disease. Multiple studies have now found a link between ACE Inhibitors, used to treat hypertension, and lessened chances and symptoms of Alzheimer‚s Disease. Thus, it has been proposed that the interaction between ACE Inhibitors and Aβ structures prevents the aggregation of improperly folded versions of the latter. This study builds off of another in the exploration of this idea: the previous showed that for a patient-derived Aβ structure, ACE Inhibitors were no more likely to bind and stay bound to the structure than Congo Red (an indicator which turns red in the presence of Aβ structures, but has no effect on their folding). In this study, we use the GROMOS 43A1 forcefield to explore the binding sites of Congo Red, Lisinopril, and Thioflavin T (another indicator), to see which has the most stable binding to a barrel Aβ structure.
    Back to the top of the page

  • Integrative LINCS (iLincs): Connecting diseases, drugs and mechanisms of actions
    Marcin Pilarczyk
    University of Cincinnati

    iLINCS (Integrative LINCS) is an integrative web platform for analysis of omics data and signatures of cellular perturbations. The portal consists of biologists-friendly user interfaces for finding and analyzing datasets and signatures, backend databases with a large collection of datasets (>3,000), pre-computed signatures (>200,000) and their connections (>2*109). The portal integrates R analytical engine via several R tools for web-computing (rserve, opencpu, shiny, rgl) and other public domain web tools and open-source applications (e.g., FTreeView, Enrichr, L1000CDS2) into a coherent web platform for omics data analysis. Analytical tools are organized into three interconnected analytical workflows. The “Dataset” workflow facilitates comprehensive analysis of primary omics datasets. In a typical use case, the user starts with an omics dataset of interest (e.g., GEO dataset corresponding to a disease of interest), performs differential gene expression analysis to construct the signature of the disease. Performs enrichment, pathway and network analysis of differentially expressed genes. Identifies “connected” drug signatures that can implicate a potential therapeutic agent for the disease. The “Signatures” workflow facilitates “connectivity analysis” with a large collection of pre-computed signatures that include LINCS drug perturbation signatures, ENCODE transcription factor binding signatures and a library of “disease related signatures” extracted from public domain omics datasets. User can either select one or more pre-computed signatures, or upload their own signatures to use in the analysis. One of the use-cases involves uploading a custom disease signature, identifying the connected LINCS chemical perturbagen signatures which can then provide putative agents for treating the disease. The “Genes” workflow starts with a user supplied list of genes which are then used to query and analyze primary data and pre-computed signatures. The portal can be accessed freely and does not require user registration (
    Back to the top of the page

  • An insilico approaches for designing novel organophosphate derivatives towards Acetylcholinesterase inhibition
    Parthiban Marimuthu
    Albany State University

    Organophosphate compounds (OPC) — have become the primary choice of insecticides and is widely used across the world. Additionally, OPCs were also commonly used as a chemical warfare agent that triggers a great challenge to public safety. The exposure of OPCs to human causes immediate excitation of cholinergic neurotransmission through transient elevation of synaptic acetylcholine (Ach) levels and accumulations. Furthermore, the OPC exposure can affect the processes in immune response, carbohydrate metabolism, cardiovascular toxicity and several others. Studies revealed that inhibition of acetylcholinesterase (AchE) is the major mechanism of toxicological responses to OPCs. Therefore, in order to assess precise and comprehensive effects of various OPCs we performed “Pharmacophore based 3D-QSAR model” based on known experimentally derived compounds along with its LD50 values. The five-featured pharmacophore model — AAHHR.305 — which has two hydrogen bond acceptor, two hydrophobic features and one aromatic ring feature, which exhibits the highest correlation coefficient (R2=0.9166), cross validated correlation coefficient (Q2=0.8221), F=63.2, Pearson-R=0.9615 and low RMS error=0.2621 values at five component PLS factor respectively. The model was further validated by enrichment studies using external decoy sets. The well-validated model was then used as a 3D query to search novel OPCs with different chemical scaffold using high-throughput virtual screening technique. The resulting outputs were further sorted by applying ADMET properties, Lipinski’s rule and molecular docking studies to refine the retrieved hits. Furthermore molecular dynamics simulation was employed to study the stability of docked conformation and to investigate the binding interactions in detail. Several important interactions were revealed at the binding site of AchE. Overall, this study suggests that the proposed hits may be more effective inhibitors for AchE inhibition.
    Back to the top of the page

  • Increasing protein production rates can decrease the rate at which functional protein is produced and their steady-state levels
    Ajeet Shama
    Pennsylvania State University

    The rate at which soluble, functional protein is produced by the ribosome has recently been found to vary in complex and unexplained ways as various translation-associated rates are altered through synonymous codon substitutions. To understand this phenomenon, here, we combine a wellestablished ribosome-traffic model with a master-equation model of co-translational domain folding to explore the scenarios that are possible for the protein production rate, 𝑱, and the functional-nascent protein production rate, 𝑭, as the rates of various translation processes are altered for five different E. coli proteins. We find that while 𝑱 monotonically increases as the rates of translation-initiation, -elongation and -termination increase, 𝑭 can either increase or decrease. 𝑭’s non-monotonic behavior arises from two opposing trends: the tendency for increased translation rates to produce more total protein but less co-translationally folded protein. We further demonstrate that under certain conditions these non-monotonic changes in 𝑭 can result in nonmonotonic variations in post-translational, steady-state levels of functional protein. These results provide a potential explanation for recent experimental observations that the specific activity of enzymatic proteins can decrease with increased synthesis rates and our model can in principle be used to rationally-design transcripts to maximize the production of functional nascent protein by simultaneously optimizing translation initiation, elongation and termination rates.
    Back to the top of the page

  • Accurate Prediction of Co-translational Folding with Molecular Dynamics Simulations of Förster Resonance Energy Transfer
    Daniel Nissley
    Pennsylvania State University

    Protein folding, the assembly of a protein molecule or domain into a tertiary structure, can occur as a protein is being synthesized by the ribosome in a process referred to as co-translational folding. Förster resonance energy transfer has recently emerged as a technique for monitoring co-translational folding in vitro and probing its pathways. Holtkamp et al. 2015 found using FRET assays that the N-terminal domain of the E. coli protein Hemk folds co-translationally via a compact state. Our computational study addresses key questions about the co-translational folding of the Hemk N-terminal domain. We will determine if coarse-grained molecular dynamics simulations with explicit representations of FRET dyes are able to accurately reproduce experimental FRET curves. We will also ascertain by what pathways the co-translational folding of HemkNTD occurs. This project will also experimentally verify our approach to modeling co-translational protein folding.
    Back to the top of the page

  • The factors governing tensile force generation by co-translational protein folding
    Sarah Leininger
    Penn State University

    Tensile forces generated from cotranslational protein folding can affect the translation elongation rate, which will in turn modify the protein’s ability to fold and function properly. Here, we use computational methods to determine the magnitude and scaling of the tensile forces generated by the cotranslational folding of several protein domains. We also use these domains to determine how the magnitude of the tensile force depends on physical factors. These results show that cotranslational folding can exert a force which can influence the translation elongation rate, including rescuing a stalled sequence. This force is biologically significant because it can change the orientation of the PTC, which can change translation elongation rates, which affects cotranslational folding and other downstream processes.
    Back to the top of the page

  • Understanding the Dependence of the Ensemble of Pathways and Temperature in the Observed Rate Constants of Enzyme Catalysis
    Joseph Persichetti
    Pennsylvania State University

    Experimental characterization of transitions states (TSs) formed along enzyme catalysis pathways remains challenging. Theoretical and computational models, such as chain-of-states (COS) methods, can elucidate TS structures by constructing a chain that connects two metastable states on an energy surface from a minimum free energy path The finite-temperature string COS method defines a transition tube which connects reactant and product states. The string is discretized into images of the system that can each sample a region of conformational space. Each image is representative of an ensemble average at its location along the reaction coordinate. Quantum mechanical/molecular mechanical (QM/MM) simulations are commonly used to gain insight into enzyme mechanisms. Such methods treat the active site quantum mechanically to accurately model bond breakage and subsequent rearrangement of electrons. The enzyme complex and surrounding environment, however, are modeled classically. We aim to predict the relevant TS structures and observed rate constants of enzymatic reactions more accurately by considering an ensemble of QM/MM reaction pathways.
    Back to the top of the page

  • Conformational ensembles of short polyQ peptides studied by atomistic REMD simulations
    Julia Borbas
    University of Konstanz

    Because of the atomic level description and temporal resolution in the picosecond range, atomistic MD simulations coupled with enhanced sampling methods are well suited for the detailed study of the conformational ensembles of intrinsically disordered peptides like those containing extended glutamine stretches (polyQ). In the present work, five 14-residue-long polyQ-containing peptides in explicit water are investigated with atomistic REMD simulations. The sketch-map dimensionality reduction algorithm is applied to visualize and compare the obtained conformational spaces, and also to highlight structures containing specific interactions. These peptides were considered as potential experimental model systems for kinetic study of the initial steps of polyQ aggregation. Into three of the sequences known β-hairpin-promoting motifs were incorporated: pG-insertion as turn-inducing residues and Coulomb-interaction between oppositely charged ends for the stabilization of the short hairpin in K2Q4pGQ4D2, electrostatic interaction between ends in the case of K2Q10D2, and a potential hydrophobic pocket formed between tryptophan residues in K2QWQWQQWQWQK2. The repelling, charged Lys residues at the ends are used to somewhat hinder the fast aggregation of polyQ peptides like in K2Q10K2, and the pure polyQ of the same length, Q14 is simulated, too. The available conformational spaces at this short timescale are very similar for all the studied peptides, though the K2Q4pGQ4D2 exhibits somewhat constrained sampling limited to areas corresponding to bended, horsshoe-shaped structures. Its ensemble contains significant proportion of β-strand elements, yet, these could be mostly considered misfolded due to being not well-registered. The other peptides are present mainly as helical and unstructured conformers. While helical and β-hairpin-like regions occupy distinct areas of the energy landscape, these regions are shared with unstructured conformers. The results imply that these peptides do have ordered structures, but because of the impermanent nature of - especially the mainchain-mainchain hydrogen bonds as one of their defining features, these structures are so transient and ill-defined that they would not be registered in an experiment because of the rapid interconversions between them. Such conformations may nevertheless be important for subsequent ordering and structure-formation processes like aggregation.
    Back to the top of the page

  • Membrane packing defects recruit the C-terminal domain of complexin-1 to curved membranes
    Jiajie Dao
    University of Cincinnati

    Complexin-1 regulates spontaneous neurotransmitter release and enhances Ca2+-evoked release. The C-terminal domain localizes to synaptic vesicles and contributes to Ca2+-independent fusion regulation. Recent studies demonstrate that this domain preferentially associates with highly curved membranes, like synaptic vesicles, to correctly position complexin-1 to clamp/inhibit spontaneous fusion. In this prospective, we summarized recent experimental results about the C-terminus of complexin-1. To further investigate this mechanism, we also performed in silico simulations probing interactions between the C-terminal domain and membranes of varying curvature. The results demonstrate that membrane defects are likely the driving force behind the curvature sensitivity of this domain.
    Back to the top of the page

  • Identifying the ribosome’s A- and P-site locations on ribosome-protected mRNA fragments using Linear Programming
    Nabeel Ahmed
    Pennsylvania State University

    The ability to accurately identify the A- and P-site locations of ribosomes on mRNA fragments generated during Ribosome Profiling experiments is central and fundamental to quantitatively analyzing transcriptome-wide translation properties of organisms. Many methods to identify these sites utilize the mRNA fragments mapped near the start codon of transcripts. In this study, we propose a method using Linear Programming to identify the A-site location as a function of the fragment size and frame. Applying our method to very high-coverage pooled dataset; we get a clear unambiguous A-site location within a ribosome-protected fragment for several fragment length and frame combinations. However, we find that there are certain length-frame combinations where accurate assignment of the A-site is not possible mostly likely due to the stochastic nature of RNase cleavage of mRNA. Our method performs better or similar than the contemporary methods in detecting higher ribosome density at known the ribosome pause sites. By increasing the accuracy of A- and P-site identification, the underlying features of translation can become more apparent.
    Back to the top of the page

  • Towards Discovery of Novel Small Molecule Inhibitors of BCL2A1
    Alexander Thorman
    University of Cincinnati

    exacerbated by the resistance of activated immune cells to apoptosis due to protection via pro-survival proteins. Bcl-2 family proteins are critical in the balance between cell survival and cell death, driving a survival phenotype by sequestration of BH3-only peptides. Within the immune system, BCL2A1 (A1/Bfl1) maintains activated states of neutrophils and CD4+ T-cells at sites of inflammation, promoting survival upon NF-êB stimulation. We hypothesize that by disrupting the interaction between A1/Bfl1 and BH3-only peptides, apoptosis can be induced specifically in the activated immune cells at sites of inflammation, thus reducing severity of a multitude of autoimmune diseases. Using a combination of in silico and in vitro approaches, potential small molecule inhibitors targeting two major pockets of this protein system have been identified. Briefly, a subset of the NCI library was selected and in silico docking simulations were performed using both the mouse ortholog, A1, and the human ortholog, Bfl1, enriching for compounds likely to interact with both A1 and Bfl1. Predicted binding compounds were acquired and differential scanning fluorimetry (DSF) was utilized to identify compounds that were binding to purified protein. Compounds demonstrated to bind via DSF were observed in a fluorescence polarization competition assay for the ability to bind specifically within the binding groove and displace BH3-only peptide. Utilizing these approaches, a number of putative lead compounds have been identified.
    Back to the top of the page

  • Mechanism and regulation of human Poly(ADP-ribose) Glycohydrolase (PARG)
    Yasin Pourfarjam and In-Kwon Kim
    University of Cincinnati

    The reversible modification of proteins by poly(ADP-ribose) (PAR) regulates chromatin structure, repair of DNA damage, and cell-fate. The PARG-catalyzed removal of PAR chains from post-translationally modified proteins restores protein function and releases oligo-(ADP-ribose) chains that bind to PAR acceptor proteins and signal damage. However, the detailed mechanism and regulation of PAR turnover remain enigmatic. We previously reported the first mammalian PARG structure showing unique features of mammalian PARG and identified a unique substrate-binding element, which we named as “Tyr-clasp”. We also reported the first fluorescence-based high-throughput PARG activity assay that can conveniently monitor PAR binding and turnover activities in real-time. To investigate the mechanism that regulates PARG activity, we decided to identify new proteins that interact with the C-terminal catalytic core of PARG (PARGCAT) through protein-microarray technique using a biotinylated PARGCAT. As a first step, we successfully purified biotinylated PARGCAT and confirmed the biotinylation.
    Back to the top of the page