Universal Glycosyltransferase Continuous Assay for Uniform Kinetics and Inhibition Database Development and Mechanistic Studies Illustrated on ST3GAL1, C1GALT1, and FUT1.

Chemical systems glycobiology requires experimental and computational tools to make possible big data analytics benefiting genomics and proteomics. The impediment to tool development is that the nature of glycan construction and mutation is not template driven but rests on cooperative glycosyltransferase (GT) catalytic synthesis. What is needed is the collation of kinetics and inhibition data in a standardized form to make possible analytics of glycan and glycoconjugate synthesis, mechanism extraction, and pattern recognition. Currently, kinetics assays in use for GTs are not universal in processing nucleoside phosphate UDP, GDP, and CMP donor-based glycosylation reactions due to limitations in accuracy and large substrate volume requirements. Here we present a universal glycosyltransferase continuous (UGC) assay able to measure the declining concentration of the NADH reporter molecule through fluorescence spectrophotometry and, therefore, determine reaction rate parameters. The development and parametrization of the assay is based on coupling the nucleotide released from GT reactions with pyruvate kinase, via nucleoside diphosphate kinase (NDK) in the case of NDP-based donor reactions. In the case of CMP-based reactions, the coupling is carried out via another kinase, cytidylate kinase in combination with NDK, which phosphorylates CMP to CDP, then CDP to CTP. Following this, we conduct kinetics and inhibition assay studies on the UDP, GDP, and CMP-based glycosylation reactions, specifically C1GAlT1, FUT1, and ST3GAL1, to represent each class of donor, respectively. The accuracy of calculating initial rates using the continuous assay compared to end point (noncontinuous) assays is demonstrated for the three classes of GTs. The previously identified natural product soyasaponin1 inhibitor was used as a model to demonstrate the application of the UGC assay as a standardized inhibition assay for GTs. We show that the dose response of ST3GAL1 to a serial dilution of Soyasaponin1 has time-dependent inhibition. This brings into question previous inhibition findings, arrived at using an end point assay, that have selected a seemingly random time point to measure inhibition. Consequently, using standardized Km values taken from the UGC assay study, ST3GAL1 was shown to be the most responsive enzyme to soyasaponin1 inhibition, followed by FUT1, then C1GALT1 with IC50 values of 37, 52, and 886 µM respectively.

CytoCopasi: a chemical systems biology target and drug discovery visual data analytics platform.

MOTIVATION: Target discovery and drug evaluation for diseases with complex mechanisms call for a streamlined chemical systems analysis platform. Currently available tools lack the emphasis on reaction kinetics, access to relevant databases, and algorithms to visualize perturbations on a chemical scale providing quantitative details as well streamlined visual data analytics functionality. RESULTS: CytoCopasi, a Maven-based application for Cytoscape that combines the chemical systems analysis features of COPASI with the visualization and database access tools of Cytoscape and its plugin applications has been developed. The diverse functionality of CytoCopasi through ab initio model construction, model construction via pathway and parameter databases KEGG and BRENDA is presented. The comparative systems biology visualization analysis toolset is illustrated through a drug competence study on the cancerous RAF/MEK/ERK pathway. AVAILABILITY AND IMPLEMENTATION: The COPASI files, simulation data, native libraries, and the manual are available on https://github.com/scientificomputing/CytoCopasi.

Molecular mechanisms from reaction coordinate graph enabled multidimensional free energies illustrated on water dimer hydrogen bonding.

Computing the free energies of molecular mechanisms in multidimensional space relies on combinations of geometrically complex reaction coordinates. We show how a graph theory implementation reduces complexity, and illustrate this on the arrangements of hydrogen bonding of a water dimer. The reaction coordinates and forces are computed using graphs that define the dependencies on the atoms in the Free Energy from Adaptive Reaction Coordinate Forces (FEARCF) library. The library can be interfaced with classical molecular dynamics as well as quantum molecular dynamics packages. Multidimensional interdependent reaction coordinates are constructed to produce complex free energy hypersurfaces. The reaction coordinates are graphed from atomic and molecular components to define points, distances, vectors, angles, planes and combinations thereof. The resultant free energy surfaces that are a function of the distance, angles, planes, and so on, can represent molecular mechanisms in reduced dimensions from the component atomic Cartesian coordinate degrees of freedom. The FEARCF library can be interfaced with any molecular package. Here, we demonstrate the link to NWChem to compute a hyperdimensional DFT (aug-cc-pVDZ basis set and X3LYP exchange correlation functionals) free energy space of a water dimer. Analysis of the water dimer free energy hypervolume reveals that while the chain and cyclic hydrogen bonding configurations are located in stable minimum energy wells, the bifurcated hydrogen bond configuration is a gateway to instability and dimer dissociation.

Multidimensional Free Energy and Accelerated Quantum Library Methods Provide a Gateway to Glycoenzyme Conformational, Electronic, and Reaction Mechanisms.

Enzyme reactions are complex to simulate accurately, and none more so than glycoenzymes (glycosyltransferase and glycosidases). A rigorous sampling of the protein frame and the conformationally plural carbohydrate substrate coupled with an unbiased treatment of the electron dynamics is needed to discover the true reaction landscapes. Here, we demonstrate the effectiveness of two computational methods ported in libraries that we have developed. The first is a flat histogram free energy method called FEARCF capable of multidimensional sampling and rapidly converging to a complete coverage of phase space. The second, the Quantum Supercharger Library (QSL), is a method that accelerates the computation of the ab initio electronic wave function as well as the integral derivatives on graphical processing units (GPUs). These QSL accelerated computations form the core components needed for direct quantum dynamics and QM/MM dynamics when coupled with legacy codes such as GAMESS and NWCHEM, making state of the art hyper-parallel electronic computations in chemistry and chemical biology possible. The combination of QSL (acceleration of ab initio QM computation) and FEARCF (multidimensional hyper-parallel reaction dynamics) makes the simulation of ab initio QM/MM reaction dynamics of enzyme catalysis feasible. Enzymes that process carbohydrates pose an added challenge as their pyranose ring substrates span multidimensional conformational space whose sampling is an intimate function of the catalytic mechanism. Here, we use the pairing of FEARCF and QSL to simulate the catalytic effect of the glycoenzyme ß-N-acetylglucosamine transferase (OGT). The reaction mechanism is discovered from a variable three bond reaction surface using SCCDFTB. The role of the OGT in distorting the pyranose ring of ß-N-acetylglucosamine (GlcNAc) away from the equilibrium 4C1 chair conformation toward the E3 envelope needed for the transition state is discovered from its pucker free energy hypersurfaces (or free energy volume, FEV). A complete GlcNAc ring pucker HF 6-31g FEV is constructed from ab initio QM dynamics in vacuum and ab initio QM/MM dynamics in the OGT catalytic domain. The OGT is shown to clearly lower the pathway toward the transition state E3 ring conformer as well as stabilize it by 1.63 kcal/mol. Illustrated here is the use of QSL accelerated ab initio QM/MM dynamics that thoroughly explores carbohydrate catalyzed reactions through a FEARCF multidimensional sampling of the interdependence between reaction and conformational space. This demonstrates how experimentally inaccessible molecular and electronic mechanisms that underpin enzyme catalysis can be discovered by directly modeling the dynamics of these complex reactions.

Anisotropic numerical potentials for coarse-grained modeling from high-speed multidimensional lookup table and interpolation algorithms.

A high-speed numerical potential delivering computational performance comparable with complex coarse-grained analytic potentials makes available models that have greater degrees of physical and chemical accuracy. This opens the possibility of increased accuracy in classical molecular dynamics simulations of anisotropic systems. In this work, we report the development of a high-speed lookup table (LUT) of four-dimensional gridded data, that uses cubic B-spline interpolations to derive off grid values and their associated partial derivatives that are located between the known grid data points. The accuracy of the coarse-grained numerical potential using a LUT from uniaxial Gay-Berne (GB) potential produced array of values is within a 3% and a 5% margin of error respectively for the interpolation of the uniaxial GB potential and its partial derivatives. The numerical potential model and partial derivatives speedup is made competitive with the analytical potential by exploiting graphics processing units on board functionality. The capability of the numerical potential is demonstrated by comparing minimizations of a box of 500 naphthalene molecules. The minimizations using a full atomistic (NAMD/CHARMM force field), a biaxial GB and a numerical potential from a LUT using data from the CHARMM pair potential was done. The numerical potential model is significantly more accurate in its approximation of the atomistic local minimum configuration than is the biaxial GB analytical potential function. This demonstrates that using a numerical potential founded on a direct lookup of the atomistic potential landscape significantly improves coarse grain (CG) modeling of complex molecules, possibly paving the way for accurate anisotropic system CG modeling.

Comparative ligand structural analytics illustrated on variably glycosylated MUC1 antigen-antibody binding.

When faced with the investigation of the preferential binding of a series of ligands against a known target, the solution is not always evident from single structure analysis. An ensemble of structures generated from computer simulations is valuable; however, visual analysis of the extensive structural data can be overwhelming. Rapid analysis of trajectory data, with tools available in the Galaxy platform, can be used to understand key features and compare differences that inform the preferential ligand structure that favors binding. We illustrate this informatics approach by investigating the in-silico binding of a peptide and glycopeptide epitope of the glycoprotein Mucin 1 (MUC1) binding with the antibody AR20.5. To study the binding, we performed molecular dynamics simulations using OpenMM and then used the Galaxy platform for data analysis. The same analysis tools are applied to each of the simulation trajectories and this process was streamlined by using Galaxy workflows. The conformations of the antigens were analyzed using root-mean-square deviation, end-to-end distance, Ramachandran plots, and hydrogen bonding analysis. Additionally, RMSF and clustering analysis were carried out. These analyses were used to rapidly assess key features of the system, interrogate the dynamic structure of the ligand, and determine the role of glycosylation on the conformational equilibrium. The glycopeptide conformations in solution change relative to the peptide; thus a partially pre-structuring is seen prior to binding. Although the bound conformation of peptide and glycopeptide is similar, the glycopeptide fluctuates less and resides in specific conformers for more extended periods. This structural analysis which gives a high-level view of the features in the system under observation, could be readily applied to other binding problems as part of a general strategy in drug design or mechanistic analysis.

BRIDGE: An Open Platform for Reproducible High-Throughput Free Energy Simulations.

Biomolecular Reaction and Interaction Dynamics Global Environment (BRIDGE) is an open-source web platform developed with the aim to provide an environment for the design of reliable methods to conduct reproducible biomolecular simulations. It is built on the well-known Galaxy bioinformatics platform. Through this, BRIDGE hosts computational chemistry tools on public web servers for internet use and provides machine- and operating-system-independent portability using the Docker container platform for local use. This construction improves the accessibility, shareability, and reproducibility of computational methods for molecular simulations. Here we present integrated free energy tools (or apps) to calculate absolute binding free energies (ABFEs) and relative binding free energies (RBFEs), as illustrated through use cases. We present free energy perturbation (FEP) methods contained in various software packages such as GROMACS and YANK that are independent of hardware configuration, software libraries, or operating systems when ported in the Galaxy-BRIDGE Docker container platform. By performing cyclin-dependent kinase 2 (CDK2) FEP calculations on geographically dispersed web servers (in Africa and Europe), we illustrate that large-scale computations can be performed using the exact same codes and methodology by collaborating groups through publicly shared protocols and workflows. The ease of public sharing and independent reproduction of simulations via BRIDGE makes possible an open review of methods and complete simulation protocols. This makes the discovery of inhibitors for drug targets accessible to nonexperts and the computer experiments that are used to arrive at leads verifiable by experts and reviewers. We illustrate this on ß-galactoside α-2,3-sialyltransferase I (ST3Gal-I), a breast cancer drug target, where a combination of RBFE and ABFE methods are used to compute the binding free energies of three inhibitors.

ProtoCaller: Robust Automation of Binding Free Energy Calculations.

ProtoCaller is a Python library distributed through Anaconda which automates relative protein-ligand binding free energy calculations in GROMACS. It links a number of popular specialized tools used to perform protein setup and parametrization, such as PDB2PQR, Modeller, and AmberTools. ProtoCaller supports commonly used AMBER force fields with additional cofactor parameters, and AM1-BCC is used to derive ligand charges. ProtoCaller also comes with an extensive PDB parser, an enhanced maximum common substructure algorithm providing improved ligand-ligand mapping, and a light GROMACS wrapper for running multiple molecular dynamics simulations. ProtoCaller is highly relevant to most researchers in the field of biomolecular simulation, allowing a customizable balance between automation and user intervention.

BRIDGE: An Open Platform for Reproducible Protein-Ligand Simulations and Free Energy of Binding Calculations.

Protein-ligand binding prediction is central to the drug-discovery process. This often follows an analysis of genomics data for protein targets and then protei n structure discovery. However, the complexity of performing reproducible protein conformational analysis and ligand binding calculations, using vetted methods and protocols can be a challenge. Here we show how Biomolecular Reaction and Interaction Dynamics Global Environment (BRIDGE), an open-source web-based compute and analytics platform for computational chemistry developed based on the Galaxy bioinformatics platform, makes protocol sharing seamless following genomics and proteomics. BRIDGE makes available tools and workflows to carry out protein molecular dynamics simulations and accurate free energy computations of protein-ligand binding. We illustrate the dynamics and simulation protocols for predicting protein-ligand binding affinities in silico on the T4 lysozyme system. This protocol is suitable for both novice and experienced practitioners. We show that with BRIDGE, protocols can be shared with collaborators or made publicly available, thus making simulation results and computations independently verifiable and reproducible.

Biomolecular Reaction and Interaction Dynamics Global Environment (BRIDGE).

MOTIVATION: The pathway from genomics through proteomics and onto a molecular description of biochemical processes makes the discovery of drugs and biomaterials possible. A research framework common to genomics and proteomics is needed to conduct biomolecular simulations that will connect biological data to the dynamic molecular mechanisms of enzymes and proteins. Novice biomolecular modelers are faced with the daunting task of complex setups and a myriad of possible choices preventing their use of molecular simulations and their ability to conduct reliable and reproducible computations that can be shared with collaborators and verified for procedural accuracy. RESULTS: We present the foundations of Biomolecular Reaction and Interaction Dynamics Global Environment (BRIDGE) developed on the Galaxy platform that makes possible fundamental molecular dynamics of proteins through workflows and pipelines via commonly used packages, such as NAMD, GROMACS and CHARMM. BRIDGE can be used to set up and simulate biological macromolecules, perform conformational analysis from trajectory data and conduct data analytics of large scale protein motions using statistical rigor. We illustrate the basic BRIDGE simulation and analytics capabilities on a previously reported CBH1 protein simulation. AVAILABILITY AND IMPLEMENTATION: Publicly available at https://github.com/scientificomputing/BRIDGE and https://usegalaxy.eu. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Denoising Autoencoder Self-Organizing Map (DASOM).

In this report, we address the question of combining nonlinearities of neurons into networks for modeling increasingly varying and progressively more complex functions. A fundamental approach is the use of higher-level representations devised by restricted Boltzmann machines and (denoising) autoencoders. We present the Denoising Autoencoder Self-Organizing Map (DASOM) that integrates the latter into a hierarchically organized hybrid model where the front-end component is a grid of topologically ordered neurons. The approach is to interpose a layer of hidden representations between the input space and the neural lattice of the self-organizing map. In so doing the parameters are adjusted by the proposed unsupervised learning algorithm. The model therefore maintains the clustering properties of its predecessor, whereas by extending and enhancing its visualization capacity enables an inclusion and an analysis of the intermediate representation space. A comprehensive series of experiments comprising optical recognition of text and images, and cancer type clustering and categorization is used to demonstrate DASOM's efficiency, performance and projection capabilities.

Producing DFT/MM enzyme reaction trajectories from SCC-DFTB/MM driving forces to probe the underlying electronics of a glycosyltransferase reaction.

The SCC-DFTB/MIO/CHARMM free energy surface for a glycosyltransferase, TcTS, is benchmarked against a DFT/MM reaction trajectory using the same CHARMM MM force field ported to the NWChem package. The popular B3LYP functional, against which the MIO parameter set was parameterized is used to optimize TS structures and run DFT reaction dynamics. A novel approach was used to generate reaction forces from a SCC-DFTB/MIO/CHARMM reaction surface to drive B3LYP/6-31G/MM and B3LYP/6-31G(d)/MM reaction trajectories. Although TS structures compare favorably, differences stemming primarily from a minimal basis set approximation prevented a successful 6-31G(d) FEARCF reaction dynamics trajectory. None the less, the dynamic evolution of the B3LYP/6-31G/MM-computed electron density provided an opportunity to perform NBO analysis along the reaction trajectory. Here, we illustrate that a successful ab initio reaction trajectory is computationally accessible when the underlying potential energy function of the semi-empirical method used to produce driving forces is sufficiently close to the ab initio potential. © 2017 Wiley Periodicals, Inc.

Conformational and electrostatic analysis of SN1 donor analogue glycomimetic inhibitors of ST3Gal-I mammalian sialyltransferase.

Mammalian sialyltransferases play a role in the metastasis of various cancers in humans. Inhibitors of these enzymes will in principle be able to directly inhibit aberrant sialylation in cancer. Inhibitors of ST3Gal-I resembling the donor component of SN1 Transition State structures were previously evaluated as part of a kinetics study. Here, using classical dynamics simulations and free energy perturbation calculations, we rationalize the performance of three of these donor analogue ST3Gal-I enzyme inhibitors. We find to inhibit the mammalian ST3Gal-I enzyme a donor analogue requires configurationally limited functionality. This is mediated by the binding of the inhibitor to the enzyme. The inhibitor's ability to interact with Y194 and T272 through a charged group such as a carboxylate is especially important. Furthermore, a conformational rigid form approximating the donor substrate is central. Here this is achieved by an intramolecular hydrogen bond formed between the carboxylate group and one of the ribose hydroxyl groups of the cytidine monophosphate (CMP) leaving group. This intramolecular interaction results in the donor substrate conformer complimenting the form of the catalytic binding site. Finally the carboxylate charge is essential for electrostatic pairing with the binding site. Substituting this group for an alcohol or amide results in severe weakening of the ligand binding. The carboxylate thus proves an to be an irreplaceable functional group and an essential pharmacophore.

The Glycome Analytics Platform: an integrative framework for glycobioinformatics.

MOTIVATION: Complex carbohydrates play a central role in cellular communication and in disease development. O- and N-glycans, which are post-translationally attached to proteins and lipids, are sugar chains that are rooted, tree structures. Independent efforts to develop computational tools for analyzing complex carbohydrate structures have been designed to exploit specific databases requiring unique formatting and limited transferability. Attempts have been made at integrating these resources, yet it remains difficult to communicate and share data across several online resources. A disadvantage of the lack of coordination between development efforts is the inability of the user community to create reproducible analyses (workflows). The latter results in the more serious unreliability of glycomics metadata. RESULTS: In this paper, we realize the significance of connecting multiple online glycan resources that can be used to design reproducible experiments for obtaining, generating and analyzing cell glycomes. To address this, a suite of tools and utilities, have been integrated into the analytic functionality of the Galaxy bioinformatics platform to provide a Glycome Analytics Platform (GAP).Using this platform, users can design in silico workflows to manipulate various formats of glycan sequences and analyze glycomes through access to web data and services. We illustrate the central functionality and features of the GAP by way of example; we analyze and compare the features of the N-glycan glycome of monocytic cells sourced from two separate data depositions.This paper highlights the use of reproducible research methods for glycomics analysis and the GAP presents an opportunity for integrating tools in glycobioinformatics. AVAILABILITY AND IMPLEMENTATION: This software is open-source and available online at https://bitbucket.org/scientificomputing/glycome-analytics-platform CONTACTS: chris.barnett@uct.ac.za or kevin.naidoo@uct.ac.za SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Glycosyltransferase Gene Expression Profiles Classify Cancer Types and Propose Prognostic Subtypes.

Aberrant glycosylation in tumours stem from altered glycosyltransferase (GT) gene expression but can the expression profiles of these signature genes be used to classify cancer types and lead to cancer subtype discovery? The differential structural changes to cellular glycan structures are predominantly regulated by the expression patterns of GT genes and are a hallmark of neoplastic cell metamorphoses. We found that the expression of 210 GT genes taken from 1893 cancer patient samples in The Cancer Genome Atlas (TCGA) microarray data are able to classify six cancers; breast, ovarian, glioblastoma, kidney, colon and lung. The GT gene expression profiles are used to develop cancer classifiers and propose subtypes. The subclassification of breast cancer solid tumour samples illustrates the discovery of subgroups from GT genes that match well against basal-like and HER2-enriched subtypes and correlates to clinical, mutation and survival data. This cancer type glycosyltransferase gene signature finding provides foundational evidence for the centrality of glycosylation in cancer.

Solution structures of chloroquine-ferriheme complexes modeled using MD simulation and investigated by EXAFS spectroscopy.

The interaction of chloroquine (CQ) and the µ-oxo dimer of iron(III) protoporphyrin IX (ferriheme) in aqueous solution was modeled using molecular dynamics (MD) simulations. Two models of the CQ-(µ-oxo ferriheme) complex were investigated, one involving CQ π-stacked with an unligated porphyrin face of µ-oxo ferriheme and the other in which CQ was docked between the two porphyrin rings. The feasibility of both models was tested by fitting computed structures to the experimental extended X-ray absorption fine structure (EXAFS) spectrum of the CQ-(µ-oxo ferriheme) complex in frozen aqueous solution. The docked model produced better agreement with experimental data, suggesting that this is the more likely structure in aqueous solution. The EXAFS fit indicated a longer than expected Fe-O bond of 1.87Å, accounting for the higher than expected magnetic moment of the complex. As a consequence, the asymmetric Fe-O-Fe stretch shifts much lower in frequency and was identified in the precipitated solid at 744cm(-1) with the aid of the O(18) isomer shift. Three important CQ-ferriheme interactions were identified in the docked structure. These were a hydrogen bond between the oxide bridge of µ-oxo ferriheme and the protonated quinolinium nitrogen atom of CQ; π-stacking between the quinoline ring of CQ and the porphyrin rings; and a close contact between the 7-chloro substituent of CQ and the porphyrin methyl hydrogen atoms. These interactions can be used to rationalize previously observed structure-activity relationships for quinoline-ferriheme association.

Quantum supercharger library: hyper-parallelism of the Hartree-Fock method.

We present here a set of algorithms that completely rewrites the Hartree-Fock (HF) computations common to many legacy electronic structure packages (such as GAMESS-US, GAMESS-UK, and NWChem) into a massively parallel compute scheme that takes advantage of hardware accelerators such as Graphical Processing Units (GPUs). The HF compute algorithm is core to a library of routines that we name the Quantum Supercharger Library (QSL). We briefly evaluate the QSL's performance and report that it accelerates a HF 6-31G Self-Consistent Field (SCF) computation by up to 20 times for medium sized molecules (such as a buckyball) when compared with mature Central Processing Unit algorithms available in the legacy codes in regular use by researchers. It achieves this acceleration by massive parallelization of the one- and two-electron integrals and optimization of the SCF and Direct Inversion in the Iterative Subspace routines through the use of GPU linear algebra libraries. © 2015 Wiley Periodicals, Inc.

Quantum supercharger library: hyper-parallel integral derivatives algorithms for ab initio QM/MM dynamics.

This article describes an extension of the quantum supercharger library (QSL) to perform quantum mechanical (QM) gradient and optimization calculations as well as hybrid QM and molecular mechanical (QM/MM) molecular dynamics simulations. The integral derivatives are, after the two-electron integrals, the most computationally expensive part of the aforementioned calculations/simulations. Algorithms are presented for accelerating the one- and two-electron integral derivatives on a graphical processing unit (GPU). It is shown that a Hartree-Fock ab initio gradient calculation is up to 9.3X faster on a single GPU compared with a single central processing unit running an optimized serial version of GAMESS-UK, which uses the efficient Schlegel method for s- and l-orbitals. Benchmark QM and QM/MM molecular dynamics simulations are performed on cellobiose in vacuo and in a 39 Å water sphere (45 QM atoms and 24843 point charges, respectively) using the 6-31G basis set. The QSL can perform 9.7 ps/day of ab initio QM dynamics and 6.4 ps/day of QM/MM dynamics on a single GPU in full double precision. © 2015 Wiley Periodicals, Inc.