Fragment-based Methods Research Articles

Recent Advances in Automated Structure-Based De Novo Drug Design.

As the number of determined and predicted protein structures and the size of druglike 'make-on-demand' libraries soar, the time-consuming nature of structure-based computer-aided drug design calls for innovative computational algorithms. De novo drug design introduces in silico heuristics to accelerate searching in the vast chemical space. This review focuses on recent advances in structure-based de novo drug design, ranging from conventional fragment-based methods, evolutionary algorithms, and Metropolis Monte Carlo methods to deep generative models. Due to the historical limitation of de novo drug design generating readily available drug-like molecules, we highlight the synthetic accessibility efforts in each category and the benchmarking strategies taken to validate the proposed framework.

Machine Learning-aided Computational Fragment-based Design of Small Molecules for Hypertension Treatment

With over 1 billion affected adults, hypertension is one of the most critical public health challenges worldwide. If left untreated over time, hypertension increases the likelihood of premature disability or death from cardiovascular diseases. Despite the range of medications available for the treatment of hypertension, many individuals do not respond positively to the treatment. Additionally, a significant percentage of the population does not take the medication as prescribed, which is sometimes attributed to intolerable side effects. Hence, there is still the need to develop new hypertension drugs that provide patients with favourable treatment outcomes. This paper explores the computational method of drug discovery to generate new lead drug molecules for hypertension by targeting the renin-angiotensin-aldosterone system (RAAS). Specifically, we proposed a framework that integrates computational fragment-based methods and an unsupervised machine learning technique to generate new lead Angiotensin-Converting Enzyme Inhibitor (ACEI) and Angiotensin-Receptor Blocker (ARB) molecules. The molecule generation process is initiated using all the approved agents acting on the RAAS that are available in the ChEMBL and DrugBank databases to create a fragment pool. The fragments are used to generate new molecules, which are categorised into ACEI and ARB clusters using unsupervised machine learning techniques. The generated molecules in each category are screened to determine their suitability as oral drug molecules, considering their physicochemical properties. Further screening is performed to determine the molecules’ suitability as ACEIs or ARBs, based on the presence of the appropriate functional groups and their similarities with existing drug molecules. The resultant molecules that passed screening are proposed as new lead antihypertensive agents. A synthesizability test is also performed on the final new lead molecules to determine the ease of making them compared to the original molecules.

Predicting 51V nuclear magnetic resonance observables in molecular crystals.

Solid-state nuclear magnetic resonance (NMR) spectroscopy and quantum chemical density functional theory (DFT) calculations are widely used to characterize vanadium centers in biological and pharmaceutically relevant compounds. Several techniques have been recently developed to improve the accuracy of predicted NMR parameters obtained from DFT. Fragment-based and planewave-corrected methods employing hybrid density functionals are particularly effective tools for solid-state applications. A recent benchmark study involving molecular crystal compounds found that fragment-based NMR calculations using hybrid density functionals improve the accuracy of predicted 51V chemical shieldings by 20% relative to traditional planewave methods. This work extends the previous study, including a careful analysis of 51V chemical shift anisotropy, electric field gradient calculations, and a more extensive test set. The accuracy of planewave-corrected techniques and recently developed fragment-based methods using electrostatic embedding based on the polarized continuum model (PCM) are found to be highly competitive with previous methods. Planewave-corrected methods achieve a 34% improvement in the errors of predicted 51V chemical shieldings relative to planewave. Additionally, planewave-corrected and fragment-based calculations were performed using PCM embedding, improving the accuracy of predicted 51V chemical shielding (CS) tensor principal values by 30% and values by 15% relative to traditional planewave methods. The performance of these methods is further examined using a redox-active oxovandium complex and a common 51V NMR reference compound.

Scalable generalized screening for high-order terms in the many-body expansion: Algorithm, open-source implementation, and demonstration.

The many-body expansion lies at the heart of numerous fragment-based methods that are intended to sidestep the nonlinear scaling of abinitio quantum chemistry, making electronic structure calculations feasible in large systems. In principle, inclusion of higher-order n-body terms ought to improve the accuracy in a controllable way, but unfavorable combinatorics often defeats this in practice and applications with n ≥ 4 are rare. Here, we outline an algorithm to overcome this combinatorial bottleneck, based on a bottom-up approach to energy-based screening. This is implemented within a new open-source software application ("Fragme∩t"), which is integrated with a lightweight semi-empirical method that is used to cull subsystems, attenuating the combinatorial growth of higher-order terms in the graph that is used to manage the calculations. This facilitates applications of unprecedented size, and we report four-body calculations in (H2O)64 clusters that afford relative energies within 0.1 kcal/mol/monomer of the supersystem result using less than 10% of the unique subsystems. We also report n-body calculations in (H2O)20 clusters up to n = 8, at which point the expansion terminates naturally due to screening. These are the largest n-body calculations reported to date using ab initio electronic structure theory, and they confirm that high-order n-body terms are mostly artifacts of basis-set superposition error.

IS-Seq: a bioinformatics pipeline for integration sites analysis with comprehensive abundance quantification methods

BackgroundIntegration site (IS) analysis is a fundamental analytical platform for evaluating the safety and efficacy of viral vector based preclinical and clinical Gene Therapy (GT). A handful of groups have developed standardized bioinformatics pipelines to process IS sequencing data, to generate reports, and/or to perform comparative studies across different GT trials. Keeping up with the technological advances in the field of IS analysis, different computational pipelines have been published over the past decade. These pipelines focus on identifying IS from single-read sequencing or paired-end sequencing data either using read-based or using sonication fragment-based methods, but there is a lack of a bioinformatics tool that automatically includes unique molecular identifiers (UMI) for IS abundance estimations and allows comparing multiple quantification methods in one integrated pipeline.ResultsHere we present IS-Seq a bioinformatics pipeline that can process data from paired-end sequencing of both old restriction sites-based IS collection methods and new sonication-based IS retrieval systems while allowing the selection of different abundance estimation methods, including read-based, Fragment-based and UMI-based systems.ConclusionsWe validated the performance of IS-Seq by testing it against the most popular analytical workflow available in the literature (INSPIIRED) and using different scenarios. Lastly, by performing extensive simulation studies and a comprehensive wet-lab assessment of our IS-Seq pipeline we could show that in clinically relevant scenarios, UMI quantification provides better accuracy than the currently most widely used sonication fragment counts as a method for IS abundance estimation.

Lipophilicity of BODIPY fluorophores and their distribution in 1-octanol–water system

The work covers synthesis and lipophilicity estimation of several BODIPY dyes. For these compounds, the distribution between 1-octanol and water layers is experimentally described and the corresponding partition coefficients LogP are calculated. The experimental LogP values are compared with popular fragment-based methods XLopP3, ALogPS, WLogP, SILICOS-IT and MLogP. Additionally, the hydrophobic and polar surface areas are found with quantum-mechanical calculations. That allowed to find a correlation between the LogP coefficient and the molecular surface topology, as well as to determine the corresponding incremental values of the methyl, acetyl, and phenyl substituents.

Accurate and efficient prediction of vibrational circular dichroism spectra of condensed-phase systems with the generalized energy-based fragmentation method

The vibrational circular dichroism (VCD) spectra could determine the molecular chirality of condensed-phase systems, but their quantum chemistry calculations are costly. The fragment-based methods have not been applied to the VCD spectra of periodic systems yet. In this work, we have extended the generalized energy-based fragmentation (GEBF) approach to accurately and efficiently compute the VCD spectra of chiral molecular crystals under periodic boundary conditions (PBCs) and macromolecules in solutions. In this approach, the Hessian matrices, atomic polar/axial tensors of a target system, are evaluated as a linear combination of the corresponding quantities from a series of small electrostatically embedded subsystems. Comparisons of GEBF and conventional VCD spectra of two large molecules show that this approach can satisfactorily reproduce the conventional computational results. Then, we applied the PBC-GEBF method to calculate the VCD spectra of a chiral molecular crystal and two macromolecules in solutions. Our results show that the experimental VCD spectrum can be roughly reproduced in terms of both band shape and vibrational peaks. The GEBF-VCD and PBC-GEBF-VCD approaches are expected to be practical tools to investigate the chirality of molecular crystals and macromolecules in solutions.

General, Rigorous Approach for the Treatment of Interfragment Covalent Bonds.

A generalized, projection-based transformation of the method-agnostic Fock operator in various ab initio fragment-based quantum chemistry methods has been developed for the treatment of interfragment covalent bonds. This transformation freezes the relevant localized molecular orbital associated with each interfragment bond, thereby restricting the variational subspace of the fragment wave functions, in order to maintain the proper physical characteristics of the involved covalent bonds. In addition, sets of orbitals that would lead to multiple occupancy of certain orbitals are explicitly removed from the variational space. The transformation is developed for the specific case of mutually orthonormal frozen and unfrozen orbitals within each fragment. The newly developed approach is then used to study model systems with two popular ab initio fragment-based methods, and the results of these calculations are compared to those obtained by existing methodologies. Analysis is focused on both quantitative and qualitative accuracy as well as computational scalability and stability. Other methods for which the developed formalisms are appropriate are outlined, and future extensions of the methods are discussed.

Using deep neural networks to explore chemical space

ABSTRACT Introduction The popularity and success of advanced AI methods like deep neural networks has led to novel ways for exploring chemical space. Their opaque nature poses challenges for model evaluation regarding novelty, uniqueness, and distribution of the chemical space covered. However, these methods also promise to be able to explore uncharted chemical space in novel ways that do not rely directly on structural similarity. Areas covered This review provides an overview of popular deep learning methods for chemical space exploration. Crucial aspects like choice of molecular representation, training for focused chemical space exploration, and criteria for assessing and validating chemical space coverage are discussed. Expert opinion Deep learning offers great potential for chemical space exploration beyond conventional fragment-based methods. Given the rarity of prospective applications and considering the difficulty in assessing representativeness and comprehensiveness of chemical space covered, developing criteria for assessing and validating generative models is of great significance. Latent space models like variational autoencoders are conceptually appealing for inverse QSAR/QSPR approaches as neighborhood relationships in latent space can be trained to reflect property similarities. Future research in understanding and interpreting generative models might lead to a better understanding of biologically relevant properties of molecules.

Fragment-based exploration of the 14-3-3/Amot-p130 interface.

The modulation of protein-protein interactions (PPIs) has developed into a well-established field of drug discovery. Despite the advances achieved in the field, many PPIs are still deemed as ‘undruggable’ targets and the design of PPIs stabilizers remains a significant challenge. The application of fragment-based methods for the identification of drug leads and to evaluate the ‘tractability’ of the desired protein target has seen a remarkable development in recent years. In this study, we explore the molecular characteristics of the 14-3-3/Amot-p130 PPI and the conceptual possibility of targeting this interface using X-ray crystallography fragment-based screening. We report the first structural elucidation of the 14-3-3 binding motif of Amot-p130 and the characterization of the binding mode and affinities involved. We made use of fragments to probe the ‘ligandability’ of the 14-3-3/Amot-p130 composite binding pocket. Here we disclose initial hits with promising stabilizing activity and an early-stage selectivity toward the Amot-p130 motifs over other representatives 14-3-3 partners. Our findings highlight the potential of using fragments to characterize and explore proteins' surfaces and might provide a starting point toward the development of small molecules capable of acting as molecular glues.

Scalable ab initio fragmentation methods based on a truncated expansion of the non-orthogonal molecular orbital model.

An alternative formulation of the non-orthogonal molecular orbital model of electronic structure theory is developed based on the expansion of the inverse molecular orbital overlap matrix. From this model, a hierarchy of ab initio fragment-based quantum chemistry methods, referred to as the nth-order expanded non-orthogonal molecular orbital methods, are developed using a minimal number of approximations, each of which is frequently employed in intermolecular interaction theory. These novel methods are compared to existing fragment-based quantum chemistry methods, and the implications of those significant differences, where they exist, between the methods developed herein and those already existing methods are examined in detail. Computational complexities and theoretical scaling are also analyzed and discussed. Future extensions for the hierarchy of methods, to account for additional intrafragment and interfragment interactions, are outlined.

Synthesis and Structural Characterization of Ricin Inhibitors Targeting Ribosome Binding Using Fragment-Based Methods and Structure-Based Design.

Ricin toxin A subunit (RTA) is the catalytic subunit of ricin, which depurinates an adenine from the sarcin/ricin loop in eukaryotic ribosomes. There are no approved inhibitors against ricin. We used a new strategy to disrupt RTA-ribosome interactions by fragment screening using surface plasmon resonance. Here, using a structure-guided approach, we improved the affinity and inhibitory activity of small-molecular-weight lead compounds and obtained improved compounds with over an order of magnitude higher efficiency. Four advanced compounds were characterized by X-ray crystallography. They bind at the RTA-ribosome binding site as the original compound but in a distinctive manner. These inhibitors bind remotely from the catalytic site and cause local conformational changes with no alteration of the catalytic site geometry. Yet they inhibit depurination by ricin holotoxin and inhibit the cytotoxicity of ricin in mammalian cells. They are the first agents that protect against ricin holotoxin by acting directly on RTA.

Comparisons of Molecular Structure Generation Methods Based on Fragment Assemblies and Genetic Graphs.

The use of quantitative structure-property relationships (QSPRs) helps in predicting molecular properties for several decades, while the automatic design of new molecular structures is still emerging. The choice of algorithms to generate molecules is not obvious and is related to several factors such as the desired chemical diversity (according to an initial dataset's content) and the level of construction (the use of atoms, fragments, pattern-based methods). In this paper, we address the problem of molecular structure generation by revisiting two approaches: fragment-based methods (FMs) and genetic-based methods (GMs). We define a set of indices to compare generation methods on a specific task. New indices inform about the explored data space (coverage), compare how the data space is explored (representativeness), and quantifies the ratio of molecules satisfying requirements (generation specificity) without the use of a database composed of real chemicals as a reference. These indices were employed to compare generations of molecules fulfilling the desired property criterion, evaluated by QSPR.

IOI: An Iterative Orbital Interaction Approach for Solving the Self-Consistent Field Problem.

An iterative orbital interaction (iOI) approach is proposed to solve, in a bottom-up fashion, the self-consistent field problem in quantum chemistry. While it belongs grossly to the family of fragment-based quantum chemical methods, iOI is distinctive in that (1) it divides and conquers not only the energy but also the wave function and that (2) the subsystem sizes are automatically determined by successively merging neighboring small subsystems until they are just enough for converging the wave function to a given accuracy. Orthonormal occupied and virtual localized molecular orbitals are obtained in a natural manner, which can be used for all post-SCF purposes.

Combining fragment docking with graph theory to improve ligand docking for homology model structures

Computational protein–ligand docking is well-known to be prone to inaccuracies in input receptor structures, and it is challenging to obtain good docking results with computationally predicted receptor structures (e.g. through homology modeling). Here we introduce a fragment-based docking method and test if it reduces requirements on the accuracy of an input receptor structures relative to non-fragment docking approaches. In this method, small rigid fragments are docked first using AutoDock Vina to generate a large number of favorably docked poses spanning the receptor binding pocket. Then a graph theory maximum clique algorithm is applied to find combined sets of docked poses of different fragment types onto which the complete ligand can be properly aligned. On the basis of these alignments, possible binding poses of complete ligand are determined. This docking method is first tested for bound docking on a series of Cytochrome P450 (CYP450) enzyme–substrate complexes, in which experimentally determined receptor structures are used. For all complexes tested, ligand poses of less than 1 Å root mean square deviations (RMSD) from the actual binding positions can be recovered. Then the method is tested for unbound docking with modeled receptor structures for a number of protein–ligand complexes from different families including the very recent severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) protease. For all complexes, poses with RMSD less than 3 Å from actual binding positions can be recovered. Our results suggest that for docking with approximately modeled receptor structures, fragment-based methods can be more effective than common complete ligand docking approaches.Electronic supplementary materialThe online version of this article (10.1007/s10822-020-00345-7) contains supplementary material, which is available to authorized users.

Visualizing drug binding interactions using microcrystal electron diffraction

Visualizing ligand binding interactions is important for structure-based drug design and fragment-based screening methods. Rapid and uniform soaking with potentially reduced lattice defects make small macromolecular crystals attractive targets for studying drug binding using microcrystal electron diffraction (MicroED). However, so far no drug binding interactions could unambiguously be resolved by electron diffraction alone. Here, we use MicroED to study the binding of a sulfonamide inhibitor to human carbonic anhydrase isoform II (HCA II). We show that MicroED data can efficiently be collected on a conventional transmission electron microscope from thin hydrated microcrystals soaked with the clinical drug acetazolamide (AZM). The data are of high enough quality to unequivocally fit and resolve the bound inhibitor. We anticipate MicroED can play an important role in facilitating in-house fragment screening for drug discovery, complementing existing methods in structural biology such as X-ray and neutron diffraction.

Discovery of Sulfonamide-Derived Agonists of SOS1-Mediated Nucleotide Exchange on RAS Using Fragment-Based Methods.

The nucleotide exchange factor Son of Sevenless (SOS) catalyzes the activation of RAS by converting it from its inactive GDP-bound state to its active GTP-bound state. Recently, we have reported the discovery of small-molecule allosteric activators of SOS1 that can increase the amount of RAS-GTP in cells. The compounds can inhibit ERK phosphorylation at higher concentrations by engaging a feedback mechanism. To further study this process, we sought different chemical matter from an NMR-based fragment screen using selective methyl labeling. To aid this process, several Ile methyl groups located in different binding sites of the protein were assigned and used to categorize the NMR hits into different classes. Hit to lead optimization using an iterative structure-based design paradigm resulted in compounds with improvements in binding affinity. These improved molecules of a different chemical class increase SOS1cat-mediated nucleotide exchange on RAS and display cellular action consistent with our prior results.

Efficient Construction of Excited-State Hessian Matrices with Machine Learning Accelerated Multilayer Energy-Based Fragment Method.

Recently, we have developed a multilayer energy-based fragment (MLEBF) method to describe excited states of large systems in which photochemically active and inert regions are separately treated with multiconfigurational and single-reference electronic structure method and their mutual polarization effects are naturally described within the many-body expansion framework. This MLEBF method has been demonstrated to provide highly accurate energies and gradients. In this work, we have further derived the MLEBF method with which highly accurate excited-state Hessian matrices of large systems are efficiently constructed. Moreover, in combination with recently proposed embedded atom neural network (EANN) model we have developed a machine learning (ML) accelerated MLEBF method (i.e., ML-MLEBF) in which photochemically inert region is entirely replaced with trained ML models. ML-MLEBF is found to improve computational efficiency of Hessian matrices in particular for large systems. Furthermore, both MLEBF and ML-MLEBF methods are highly parallel and exhibit low-scaling computational cost with multiple CPUs. The present developments could motivate combining various ML techniques with fragment-based electronic structure methods to explore Hessian-matrix-based excited-state properties of large systems.

Fragment Exchange Potential for Realizing Pauli Deformation of Interfragment Interactions.

In fragment-based methods, the lack of explicit short-range exchange interactions between monomers can result in unphysical deformation in charge density. In this study, we describe a fragment exchange potential (XFP) to explicitly account for interfragmental Pauli deformation. In our implementation, a Kohn-Sham exchange potential is adopted along with the Yukawa potential. The method has been validated by comparison of the computed exchange energies using the XFP potential with results obtained from antisymmetrized fragmental orbitals on the S66×8 data set containing 528 bimolecular interactions of equilibrium and arbitrary geometries. It was also found that it is only necessary to deploy numerical grids on atoms within their van der Waals contacts, significantly reducing the small, albeit extra, computational cost. We anticipate that the XFP presented here may be applied to molecular dynamics simulations of macromolecules using a fragment-based quantum mechanical potential with improved SCF convergence and computational accuracy.

Discovery of WD Repeat-Containing Protein 5 (WDR5)-MYC Inhibitors Using Fragment-Based Methods and Structure-Based Design.

The frequent deregulation of MYC and its elevated expression via multiple mechanisms drives cells to a tumorigenic state. Indeed, MYC is overexpressed in up to ∼50% of human cancers and is considered a highly validated anticancer target. Recently, we discovered that WD repeat-containing protein 5 (WDR5) binds to MYC and is a critical cofactor required for the recruitment of MYC to its target genes and reported the first small molecule inhibitors of the WDR5-MYC interaction using structure-based design. These compounds display high binding affinity, but have poor physicochemical properties and are hence not suitable for in vivo studies. Herein, we conducted an NMR-based fragment screening to identify additional chemical matter and, using a structure-based approach, we merged a fragment hit with the previously reported sulfonamide series. Compounds in this series can disrupt the WDR5-MYC interaction in cells, and as a consequence, we observed a reduction of MYC localization to chromatin.