Computer-mediated reduction of spectral properties to molecular structures : General Design and Structural Building Blocks

A basic design is outlined for the conversion of the aggregate spectral properties of an organic compound of unknown structure to a set of compatible alternative molecular structures which is small in number anal exhaustive in scope. The process is built around a precisely defined unit of substructure. A procedure is described for the exhaustive generation of the set of these fragments from any specified group of elements and allowed valences. A versatile, user-friendly data-management system (FOCISS) provides a convenient means of manipulating these fragments and correlating their structure and properties. Extensive carbon-13 n.m.r. data are in place and applications of FOCISS to fragment-oriented searches are illustrated. The creation of a subset of chemically-stable fragments is also described.     

Design of a Fragment Library that maximally represents available chemical space

Cheminformatics protocols have been developed and assessed that identify a small set of fragments which can represent the compounds in a chemical library for use in fragment-based ligand discovery. Six different methods have been implemented and tested on Input Libraries of compounds from three suppliers. The resulting Fragment Sets have been characterised on the basis of computed physico-chemical properties and their similarity to the Input Libraries. A method that iteratively identifies fragments with the maximum number of similar compounds in the Input Library (Nearest Neighbours) produces the most diverse library. This approach could increase the success of experimental ligand discovery projects, by providing fragments that can be progressed rapidly to larger compounds through access to available similar compounds (known as SAR by Catalog).     

GLYCAM06: a generalizable biomolecular force field. Carbohydrates

A new derivation of the GLYCAM06 force field, which removes its previous specificity for carbohydrates, and its dependency on the AMBER force field and parameters, is presented. All pertinent force field terms have been explicitly specified and so no default or generic parameters are employed. The new GLYCAM is no longer limited to any particular class of biomolecules, but is extendible to all molecular classes in the spirit of a small-molecule force field. The torsion terms in the present work were all derived from quantum mechanical data from a collection of minimal molecular fragments and related small molecules. For carbohydrates, there is now a single parameter set applicable to both alpha- and beta-anomers and to all monosaccharide ring sizes and conformations. We demonstrate that deriving dihedral parameters by fitting to QM data for internal rotational energy curves for representative small molecules generally leads to correct rotamer populations in molecular dynamics simulations, and that this approach removes the need for phase corrections in the dihedral terms. However, we note that there are cases where this approach is inadequate. Reported here are the basic components of the new force field as well as an illustration of its extension to carbohydrates. In addition to reproducing the gas-phase properties of an array of small test molecules, condensed-phase simulations employing GLYCAM06 are shown to reproduce rotamer populations for key small molecules and representative biopolymer building blocks in explicit water, as well as crystalline lattice properties, such as unit cell dimensions, and vibrational frequencies.     

General methodology to optimize damping functions to account for charge penetration effects in electrostatic calculations using multicentered multipolar expansions

We developed a methodology to optimize exponential damping functions to account for charge penetration effects when computing molecular electrostatic properties using the multicentered multipolar expansion method (MME). This methodology is based in the optimization of a damping parameter set using a two-step fast local fitting procedure and the ab initio (Hartree-Fock/6-31G** and 6-31G**+) electrostatic potential calculated in a set of concentric grid of points as reference. The principal aspect of the methodology is a first local fitting step which generates a focused initial guess to improve the performance of a simplex method avoiding the use of multiple runs and the choice of initial guesses. Three different strategies for the determination of optimized damping parameters were tested in the following studies: (1) investigation of the error in the calculation of the electrostatic interaction energy for five hydrogen-bonded dimers at standard and nonstandard hydrogen-bonded geometries and at nonequilibrium geometries; (2) calculation of the electrostatic molecular properties (potential and electric field) for eight small molecular systems (methanol, ammonia, water, formamide, dichloromethane, acetone, dimethyl sulfoxide, and acetonitrile) and for the 20 amino acids. Our results show that the methodology performs well not only for small molecules but also for relatively larger molecular systems. The analysis of the distinct parameter sets associated with different optimization strategies show that (i) a specific parameter set is more suitable and more general for electrostatic interaction energy calculations, with an average absolute error of 0.46 kcal/mol at hydrogen-bond geometries; (ii) a second parameter set is more suitable for electrostatic potential and electric field calculations at and outside the van der Waals (vdW) envelope, with an average error decrease >72% at the vdW surface. A more general amino acid damping parameter set was constructed from the original damping parameters derived for the small fragments and for the amino acids. This damping set is more insensitive to protein backbone and residue side-chain conformational changes and can be very useful for future docking and molecular dynamics protein simulations using ab initio based polarizable classical methods.     

Residue preference mapping of ligand fragments in the Protein Data Bank

The interaction between small molecules and proteins is one of the major concerns for structure-based drug design because the principles of protein-ligand interactions and molecular recognition are not thoroughly understood. Fortunately, the analysis of protein-ligand complexes in the Protein Data Bank (PDB) enables unprecedented possibilities for new insights. Herein, we applied molecule-fragmentation algorithms to split the ligands extracted from PDB crystal structures into small fragments. Subsequently, we have developed a ligand fragment and residue preference mapping (LigFrag-RPM) algorithm to map the profiles of the interactions between these fragments and the 20 proteinogenic amino acid residues. A total of 4032 fragments were generated from 71?798 PDB ligands by a ring cleavage (RC) algorithm. Among these ligand fragments, 315 unique fragments were characterized with the corresponding fragment-residue interaction profiles by counting residues close to these fragments. The interaction profiles revealed that these fragments have specific preferences for certain types of residues. The applications of these interaction profiles were also explored and evaluated in case studies, showing great potential for the study of protein-ligand interactions and drug design. Our studies demonstrated that the fragment-residue interaction profiles generated from the PDB ligand fragments can be used to detect whether these fragments are in their favorable or unfavorable environments. The algorithm for a ligand fragment and residue preference mapping (LigFrag-RPM) developed here also has the potential to guide lead chemistry modifications as well as binding residues predictions.     

Stalis: A Computational Method for Template-Based Ab Initio Ligand Design

Proteins interact with small molecules through specific molecular recognition, which is central to essential biological functions in living systems. Therefore, understanding such interactions is crucial for basic sciences and drug discovery. Here, we present S tructure t emplate-based a b initio li gand design s olution (Stalis), a knowledge-based approach that uses structure templates from the Protein Data Bank libraries of whole ligands and their fragments and generates a set of molecules (virtual ligands) whose structures represent the pocket shape and chemical features of a given target binding site. Our benchmark performance evaluation shows that ligand structure-based virtual screening using virtual ligands from Stalis outperforms a receptor structure-based virtual screening using AutoDock Vina, demonstrating reliable overall screening performance applicable to computational high-throughput screening. However, virtual ligands from Stalis are worse in recognizing active compounds at the small fraction of a rank-ordered list of screened library compounds than crystal ligands, due to the low resolution of the virtual ligand structures. In conclusion, Stalis can facilitate drug discovery research by designing virtual ligands that can be used for fast ligand structure-based virtual screening. Moreover, Stalis provides actual three-dimensional ligand structures that likely bind to a target protein, enabling to gain structural insight into potential ligands. Stalis can be an efficient computational platform for high-throughput ligand design for fundamental biological study and drug discovery research at the proteomic level. © 2019 Wiley Periodicals, Inc.     

A quantum-chemical approach to the analysis of intramolecular interactions using the fragment orbitals and MNDO method. Calculations for vinyl fluoride and vinyl iodide

A new quantum-chemical method for demarcation between the basic mechanisms of mutual influence of structural fragments in complex organic molecules (inductive effect, conjugation,etc.) and for assessment of their significance for particular physicochemical properties is proposed. The effects of different channels of intramolecular interaction on the molecular geometry and energy, charge distribution, and the molecular orbital structures and energies were considered taking vinyl halides as examples. In systems with an interfragment bond of high polarity, separation of the contribution of the inductive effect is to a great extent meaningless, while π-conjugation can be considered independently. The method allows a more valid interpretation of the results of quantum-chemical calculations in terms of theoretical organic chemistry.     

Ab initio molecular orbital calculations of the energetic,structural, vibrational and electronic properties of some hydrogen bonded complexes of water,ammonia and hydroxylamine

Recently published results of some ab initio molecular orbital calculations on 10 hydrogen bonded homodimers and heterodimers of water, ammonia and hydroxylamine have been analysed. The properties discussed include the interaction energies, the structural parameters of the hydrogen bonded fragments, RAH⋯B, the vibrational properties (wavenumbers of the bonded AH-stretching and RAH-bending vibrational modes, and their shifts, and the wavenumbers of the highest frequency intermolecular modes of the complexes), and the atomic polar tensors, their invariants, and the Mulliken charges of the bonded hydrogen atoms. It is concluded that such complexes are stabilized by hydrogen bonds of the OH⋯N, OH⋯O and NH⋯N type, listed in decreasing order of strength. The complexes tend to form preferentially open (more or less linear) structures, followed by large cyclic structures, in which ring strain is minimized, and finally small cyclic structures, in which ring strain is more important.     

Bayesian molecular design with a chemical language model

The aim of computational molecular design is the identification of promising hypothetical molecules with a predefined set of desired properties. We address the issue of accelerating the material discovery with state-of-the-art machine learning techniques. The method involves two different types of prediction; the forward and backward predictions. The objective of the forward prediction is to create a set of machine learning models on various properties of a given molecule. Inverting the trained forward models through Bayes’ law, we derive a posterior distribution for the backward prediction, which is conditioned by a desired property requirement. Exploring high-probability regions of the posterior with a sequential Monte Carlo technique, molecules that exhibit the desired properties can computationally be created. One major difficulty in the computational creation of molecules is the exclusion of the occurrence of chemically unfavorable structures. To circumvent this issue, we derive a chemical language model that acquires commonly occurring patterns of chemical fragments through natural language processing of ASCII strings of existing compounds, which follow the SMILES chemical language notation. In the backward prediction, the trained language model is used to refine chemical strings such that the properties of the resulting structures fall within the desired property region while chemically unfavorable structures are successfully removed. The present method is demonstrated through the design of small organic molecules with the property requirements on HOMO-LUMO gap and internal energy. The R package iqspr is available at the CRAN repository.     

First principles NMR calculations by fragmentation

The nuclear magnetic shielding tensor is a molecular property that can be computed from first principles. In this work we show that by utilizing the fragmentation approach, one is able to accurately compute this property for a large class of molecules. This is of great significance because the computational expense required in the evaluation of the shielding tensor for all nuclei in a large molecule is now subject to near linear scaling. On the basis of previous studies and this work, it is also very likely that all molecular properties that can be expressed as derivatives of the total energy of the system are also amenable to accurate evaluation via fragmentation. If only the chemical shifts for nuclei in a small part of a large molecule are of interest, then only those molecular fragments containing those nuclei need to have their shielding tensors evaluated. Further, the fragmentation approach allows one to construct a database of molecular fragments that could, in principle, be used in the NMR characterization of molecules and at the same time provide possible three-dimensional representations of these molecules.     

A theoretical study of molecular structure in the excited state and molecular luminescence : III. Types of molecular structure in the excited ππ singlet states

The molecular geometry (bond lengths) in the S₁, and S₂ excited states of a large number of conjugated organic molecules is calculated with the help of the SCF-CI-SC procedure, already described in previous communications. The analysis shows that all excited states studied can be classified in three basic types of structure: (1) with strong conjugation over the whole molecule or along its periphery; (2) with a long main conjugated framework and other shorter conjugated fragments moderately or slightly bonded with the framework and (3) structures consisting of two or more fragments slightly conjugated to each other. An estimation of the role of the possible intramolecular motions of low frequency is done, as well as some other factors contributing to the radiationless deactivation of a given excited state. The presence of structures with both lowest excited singlets of type 1 is shown.     

Coherent and incoherent orientation and alignment of ICN photoproducts

We report extended measurements of the rotational polarization and correlated angular distribution of CN photofragments from ICN photodissociation, with a particular emphasis on the creation and detection of molecular orientation with circularly-polarized light. Doppler profiles of the nascent photoproducts are measured by Frequency-Modulated (FM) transient absorption, and the resulting high signal-to-noise data are valuable for verifying the form of the angular correlations between the recoil velocity, the photofragment rotational angular momentum, and the space-fixed frame defined by the dissociation polarization. A space-fixed bipolar moment notation can be used for an unambiguous characterization of the maximal set of polarization properties that can be created with one-photon excitation and detected with one-photon Doppler-resolved absorption spectroscopy. Relating the observed polarization moments to the various coherent and incoherent, adiabatic and non-adiabatic mechanisms, that have been derived and verified extensively in the case of diatomic photodissociation to polarized atomic fragments, is not unambiguous in the case of diatomic fragments from triatomic precursors. Constraints among various polarization moments confirmed in the case of diatomic dissociation are not confirmed in this triatomic case, where the perpendicular transitions to non-degenerate A' and A' components of a linear Omega = 1 state are qualitatively different from excitation to degenerate Omega = +/-1 states in a diatomic molecule.     

Information about the biologically relevant properties of Clostridium pasteurianum rubredoxin obtained from modeling and dynamics simulations of molecular variants

Rubredoxins are small electron transfer proteins containing one iron atom at their active site. The rubredoxin from the anaerobic bacterium Clostridium pasteurianum has been subjected to molecular dynamics studies starting from the minimized solvated structure. The results of the simulations have been compared with identical ones carried out with selected mutated forms of the protein obtained by molecular modeling. Surface residues, which are highly conserved among rubredoxins and close to the cysteine ligands, can be replaced by glutamates, i.e. long chain carboxylates. The main structural consequence is a shift of the protein backbone bearing conserved aromatic residues. Reciprocally, substitution of the aromatic residue closest to the iron atom shifts the cysteine-containing peptide fragments. These observations have been related to the changes in electron transfer and redox properties previously measured for this set of rubredoxin molecular variants. Received: 16 May 1998 / Accepted: 4 August 1998 / Published online: 2 November 1998     

16-fold degeneracy of peptide plane orientations from residual dipolar couplings: analytical treatment and implications for protein structure determination

Residual dipolar couplings (RDCs) measured for internally rigid molecular fragments provide important information about the relative orientations of these fragments. Dependent on the symmetry of the alignment tensor and the symmetry of the molecular fragment, however, there generally exist more than one solution for the fragment orientation consistent with the measured RDCs. Analytical solutions are presented that describe the complete set of orientations of internally rigid fragments that are consistent with multiple dipolar couplings measured in a single alignment medium that is rhombic. For the first time, it is shown that, for a planar fragment such as the peptide plane, there generally exist 16 different solutions with their analytical expressions presented explicitly. The presence of these solutions is shown to be highly relevant for standard structure determination protocols using RDCs to refine molecular structures. In particular, when using standard protein structure refinement with RDCs that were measured in a single alignment medium as constraints, it is found that often more than one of the peptide plane solutions is physically viable; i.e., despite being consistent with measured RDCs, the local backbone structure can be incorrect. On the basis of experimental and simulated examples, it is rationalized why protein structures that are refined against RDCs measured in a single medium can have lower resolution (precision) than one would expect on the basis of the experimental accuracy of the RDCs. Conditions are discussed under which the correct solution can be identified.     

Exploring fragment spaces under multiple physicochemical constraints

We present a new algorithm for the enumeration of chemical fragment spaces under constraints. Fragment spaces consist of a set of molecular fragments and a set of rules that specifies how fragments can be combined. Although fragment spaces typically cover an infinite number of molecules, they can be enumerated in case that a physicochemical profile of the requested compounds is given. By using min-max ranges for a number of corresponding properties, our algorithm is able to enumerate all molecules which obey these properties. To speed up the calculation, the given ranges are used directly during the build-up process to guide the selection of fragments. Furthermore, a topology based fragment filter is used to skip most of the redundant fragment combinations. We applied the algorithm to 40 different target classes. For each of these, we generated tailored fragment spaces from sets of known inhibitors and additionally derived ranges for several physicochemical properties. We characterized the target-specific fragment spaces and were able to enumerate the complete chemical subspaces for most of the targets.     

Time-efficient flexible superposition of medium-sized molecules

We present an efficient algorithm for the structural alignment of medium-sized organic molecules. The algorithm has been developed for applications in 3D QSAR and in receptor modeling. The method assumes one of the molecules, the reference ligand, to be presented in the conformation that it adopts inside the receptor pocket. The second molecule, the test ligand, is considered to be flexible, and is assumed to be given in an arbitrary low-energy conformation. Ligand flexibility is modeled by decomposing the test ligand into molecular fragments, such that ring systems are completely contained in a single fragment. Conformations of fragments and torsional angles of single bonds are taken from a small finite set, which depends on the fragment and bond, respectively. The algorithm superimposes a distinguished base fragment of the test ligand onto a suitable region of the reference ligand and then attaches the remaining fragments of the test ligand in a step-by-step fashion. During this process, a scoring function is optimized that encompasses bonding terms and terms accounting for steric overlap as well as for similarity of chemical properties of both ligands. The algorithm has been implemented in the FLEXS system. To validate the quality of the produced results, we have selected a number of examples for which the mutual superposition of two ligands is experimentally given by the comparison of the binding geometries known from the crystal structures of their corresponding protein–ligand complexes. On more than two-thirds of the test examples the algorithm produces rms deviations of the predicted versus the observed conformation of the test ligand below 1.5 Å. The run time of the algorithm on a single problem instance is a few minutes on a common-day workstation. The overall goal of this research is to drastically reduce run times, while limiting the inaccuracies of the model and the computation to a tolerable level.