Leveraging structure determination with fragment screening for infectious disease drug targets: MECP synthase from Burkholderia pseudomallei

As part of the Seattle Structural Genomics Center for Infectious Disease, we seek to enhance structural genomics with ligand-bound structure data which can serve as a blueprint for structure-based drug design. We have adapted fragment-based screening methods to our structural genomics pipeline to generate multiple ligand-bound structures of high priority drug targets from pathogenic organisms. In this study, we report fragment screening methods and structure determination results for 2C-methyl-D-erythritol-2,4-cyclo-diphosphate (MECP) synthase from Burkholderia pseudomallei, the gram-negative bacterium which causes melioidosis. Screening by nuclear magnetic resonance spectroscopy as well as crystal soaking followed by X-ray diffraction led to the identification of several small molecules which bind this enzyme in a critical metabolic pathway. A series of complex structures obtained with screening hits reveal distinct binding pockets and a range of small molecules which form complexes with the target. Additional soaks with these compounds further demonstrate a subset of fragments to only bind the protein when present in specific combinations. This ensemble of fragment-bound complexes illuminates several characteristics of MECP synthase, including a previously unknown binding surface external to the catalytic active site. These ligand-bound structures now serve to guide medicinal chemists and structural biologists in rational design of novel inhibitors for this enzyme.     

NMR screening in drug discovery

NMR methods in drug discovery have traditionally been used to obtain structural information for drug targets or target-ligand complexes. Recently, it has been shown that NMR may be used as an alternative approach for identification of ligands that bind to protein drug targets, shifting the emphasis of many NMR laboratories towards screening and design of potential drug molecules, rather than structural characterization.     

Identification of B. anthracis N5-carboxyaminoimidazole ribonucleotide mutase (PurE) active site binding compounds via fragment library screening

The de novo purine biosynthesis pathway is an attractive target for antibacterial drug design, and PurE from this pathway has been identified to be crucial for Bacillus anthracis survival in serum. In this study we adopted a fragment-based hit discovery approach, using three screening methods—saturation transfer difference nucleus magnetic resonance (STD-NMR), water-ligand observed via gradient spectroscopy (WaterLOGSY) NMR, and surface plasmon resonance (SPR), against B. anthracis PurE (BaPurE) to identify active site binding fragments by initially testing 352 compounds in a Zenobia fragment library. Competition STD NMR with the BaPurE product effectively eliminated non-active site binding hits from the primary hits, selecting active site binders only. Binding affinities (dissociation constant, K_D) of these compounds varied between 234 and 301 μM. Based on test results from the Zenobia compounds, we subsequently developed and applied a streamlined fragment screening strategy to screen a much larger library consisting of 3000 computationally pre-selected fragments. Thirteen final fragment hits were confirmed to exhibit binding affinities varying from 14 μM to 700 μM, which were categorized into five different basic scaffolds. All thirteen fragment hits have ligand efficiencies higher than 0.30. We demonstrated that at least two fragments from two different scaffolds exhibit inhibitory activity against the BaPurE enzyme.     

Discovery of Fragment Molecules That Bind the Human Peroxiredoxin 5 Active Site

The search for protein ligands is a crucial step in the inhibitor design process. Fragment screening represents an interesting method to rapidly find lead molecules, as it enables the exploration of a larger portion of the chemical space with a smaller number of compounds as compared to screening based on drug-sized molecules. Moreover, fragment screening usually leads to hit molecules that form few but optimal interactions with the target, thus displaying high ligand efficiencies. Here we report the screening of a homemade library composed of 200 highly diverse fragments against the human Peroxiredoxin 5 protein. Peroxiredoxins compose a family of peroxidases that share the ability to reduce peroxides through a conserved cysteine. The three-dimensional structures of these enzymes ubiquitously found throughout evolution have been extensively studied, however, their biological functions are still not well understood and to date few inhibitors have been discovered against these enzymes. Six fragments from the library were shown to bind to the Peroxiredoxin 5 active site and ligand-induced chemical shift changes were used to drive the docking of these small molecules into the protein structure. The orientation of the fragments in the binding pocket was confirmed by the study of fragment homologues, highlighting the role of hydroxyl functions that hang the ligands to the Peroxiredoxin 5 protein. Among the hit fragments, the small catechol molecule was shown to significantly inhibit Peroxiredoxin 5 activity in a thioredoxin peroxidase assay. This study reports novel data about the ligand-Peroxiredoxin interactions that will help considerably the development of potential Peroxiredoxin inhibitors.     

Intracellular antibodies and cancer: New technologies offer therapeutic opportunities

Since the realisation that the antigen‐binding regions of antibodies, the variable (V) regions, can be uncoupled from the rest of the molecule to create fragments that recognise and abrogate particular protein functions in cells, the use of antibody fragments inside cells has become an important tool in bioscience. Diverse libraries of antibody fragments plus in vivo screening can be used to isolate single chain variable fragments comprising VH and VL segments or single V‐region domains. Some of these are interfering antibody fragments that compete with protein‐protein interactions, providing lead molecules for drug interactions that until now have been considered difficult or undruggable. It may be possible to deliver or express antibody fragments in target cells as macrodrugs per se. In future incarnations of intracellular antibodies, however, the structural information of the interaction interface of target and antibody fragment should facilitate development of binding site mimics as small drug‐like molecules. This is a new dawn for intracellular antibody fragments both as macrodrugs and as precursors of drugs to treat human diseases and should finally lead to the removal of the epithet of the ‘undruggable’ protein‐protein interactions.     

Computer-Aided Design of Fragment Mixtures for NMR-Based Screening

Fragment-based drug discovery is widely applied both in industrial and in academic screening programs. Several screening techniques rely on NMR to detect binding of a fragment to a target. NMR-based methods are among the most sensitive techniques and have the further advantage of yielding a low rate of false positives and negatives. However, NMR is intrinsically slower than other screening techniques; thus, to increase throughput in NMR-based screening, researchers often assay mixtures of fragments, rather than single fragments. Herein we present a fast and straightforward computer-aided method to design mixtures of fragments taken from a library that have minimized NMR signal overlap. This approach enables direct identification of one or several active fragments without the need for deconvolution. Our approach entails encoding of NMR spectra into a computer-readable format that we call a fingerprint, and minimizing the global signal overlap through a Monte Carlo algorithm. The scoring function used favors a homogenous distribution of the global signal overlap. The method does not require additional experimental work: the only data required are NMR spectra, which are generally recorded for each compound as a quality control measure before its insertion into the library.     

Binding-Site Assessment by Virtual Fragment Screening

The accurate prediction of protein druggability (propensity to bind high-affinity drug-like small molecules) would greatly benefit the fields of chemical genomics and drug discovery. We have developed a novel approach to quantitatively assess protein druggability by computationally screening a fragment-like compound library. In analogy to NMR-based fragment screening, we dock ∼11000 fragments against a given binding site and compute a computational hit rate based on the fraction of molecules that exceed an empirically chosen score cutoff. We perform a large-scale evaluation of the approach on four datasets, totaling 152 binding sites. We demonstrate that computed hit rates correlate with hit rates measured experimentally in a previously published NMR-based screening method. Secondly, we show that the in silico fragment screening method can be used to distinguish known druggable and non-druggable targets, including both enzymes and protein-protein interaction sites. Finally, we explore the sensitivity of the results to different receptor conformations, including flexible protein-protein interaction sites. Besides its original aim to assess druggability of different protein targets, this method could be used to identifying druggable conformations of flexible binding site for lead discovery, and suggesting strategies for growing or joining initial fragment hits to obtain more potent inhibitors.     

Identification of novel fragment compounds targeted against the pY pocket of v-Src SH2 by computational and NMR screening and thermodynamic evaluation

Discovery of small molecule inhibitors of protein-protein interactions is a major challenge to pharmaceutical development. Fragment-based approaches have begun to be widely adopted as an effective way of exploring chemical space on a protein surface with reduced library size. On completion of a fragment screen, the subsequent selection of appropriate "hit" molecules for development is a key decision point. Thermodynamic parameters can be used in this decision process. In this work, a fragment identification protocol based on a virtual fragment analysis and selection followed by 19F NMR screening was directed at the phosphotyrosine binding site of the Src SH2 domain. Three new ligands were identified. Isothermal titration calorimetry was used to provide thermodynamic parameters for the physiologically relevant ligand and the selected fragments. One of these fragments possesses a highly favorable enthalpic contribution to complex formation compared to other fragments and to the physiologically relevant ligand suggesting that it would make a good candidate for compound development.     

Thermodynamic analysis of water molecules at the surface of proteins and applications to binding site prediction and characterization

Water plays an essential role in determining the structure and function of all biological systems. Recent methodological advances allow for an accurate and efficient estimation of the thermodynamic properties of water molecules at the surface of proteins. In this work, we characterize these thermodynamic properties and relate them to various structural and functional characteristics of the protein. We find that high-energy hydration sites often exist near protein motifs typically characterized as hydrophilic, such as backbone amide groups. We also find that waters around alpha helices and beta sheets tend to be less stable than waters around loops. Furthermore, we find no significant correlation between the hydration site-free energy and the solvent accessible surface area of the site. In addition, we find that the distribution of high-energy hydration sites on the protein surface can be used to identify the location of binding sites and that binding sites of druggable targets tend to have a greater density of thermodynamically unstable hydration sites. Using this information, we characterize the FKBP12 protein and show good agreement between fragment screening hit rates from NMR spectroscopy and hydration site energetics. Finally, we show that water molecules observed in crystal structures are less stable on average than bulk water as a consequence of the high degree of spatial localization, thereby resulting in a significant loss in entropy. These findings should help to better understand the characteristics of waters at the surface of proteins and are expected to lead to insights that can guide structure-based drug design efforts.     

Comparing Binding Modes of Analogous Fragments Using NMR in Fragment-Based Drug Design: Application to PRDX5

Fragment-based drug design is one of the most promising approaches for discovering novel and potent inhibitors against therapeutic targets. The first step of the process consists of identifying fragments that bind the protein target. The determination of the fragment binding mode plays a major role in the selection of the fragment hits that will be processed into drug-like compounds. Comparing the binding modes of analogous fragments is a critical task, not only to identify specific interactions between the protein target and the fragment, but also to verify whether the binding mode is conserved or differs according to the fragment modification. While X-ray crystallography is the technique of choice, NMR methods are helpful when this fails. We show here how the ligand-observed saturation transfer difference (STD) experiment and the protein-observed ¹⁵N-HSQC experiment, two popular NMR screening experiments, can be used to compare the binding modes of analogous fragments. We discuss the application and limitations of these approaches based on STD-epitope mapping, chemical shift perturbation (CSP) calculation and comparative CSP sign analysis, using the human peroxiredoxin 5 as a protein model.     

A combination of 19F NMR and surface plasmon resonance for site-specific hit selection and validation of fragment molecules that bind to the ATP-binding site of a kinase

¹⁹F NMR has recently emerged as an efficient, sensitive tool for analyzing protein binding to small molecules, and surface plasmon resonance (SPR) is also a popular tool for this purpose. Herein a combination of ¹⁹F NMR and SPR was used to find novel binders to the ATP-binding pocket of MAP kinase extracellular regulated kinase 2 (ERK2) by fragment screening with an original fluorinated-fragment library. The ¹⁹F NMR screening yielded a high primary hit rate of binders to the ERK2 ATP-binding pocket compared with the rate for the SPR screening. Hit compounds were evaluated and categorized according to their ability to bind to different binding sites in the ATP-binding pocket. The binding manner was characterized by using isothermal titration calorimetry and docking simulation. Combining ¹⁹F NMR with other biophysical methods allows the identification of multiple types of hit compounds, thereby increasing opportunities for drug design using preferred fragments.     

Knowledge-based Fragment Binding Prediction

Target-based drug discovery must assess many drug-like compounds for potential activity. Focusing on low-molecular-weight compounds (fragments) can dramatically reduce the chemical search space. However, approaches for determining protein-fragment interactions have limitations. Experimental assays are time-consuming, expensive, and not always applicable. At the same time, computational approaches using physics-based methods have limited accuracy. With increasing high-resolution structural data for protein-ligand complexes, there is now an opportunity for data-driven approaches to fragment binding prediction. We present FragFEATURE, a machine learning approach to predict small molecule fragments preferred by a target protein structure. We first create a knowledge base of protein structural environments annotated with the small molecule substructures they bind. These substructures have low-molecular weight and serve as a proxy for fragments. FragFEATURE then compares the structural environments within a target protein to those in the knowledge base to retrieve statistically preferred fragments. It merges information across diverse ligands with shared substructures to generate predictions. Our results demonstrate FragFEATURE''s ability to rediscover fragments corresponding to the ligand bound with 74% precision and 82% recall on average. For many protein targets, it identifies high scoring fragments that are substructures of known inhibitors. FragFEATURE thus predicts fragments that can serve as inputs to fragment-based drug design or serve as refinement criteria for creating target-specific compound libraries for experimental or computational screening.     

Fragment-based virtual screening approach and molecular dynamics simulation studies for identification of BACE1 inhibitor leads

Traditional structure-based virtual screening method to identify drug-like small molecules for BACE1 is so far unsuccessful. Location of BACE1, poor Blood Brain Barrier permeability and P-glycoprotein (Pgp) susceptibility of the inhibitors make it even more difficult. Fragment-based drug design method is suitable for efficient optimization of initial hit molecules for target like BACE1. We have developed a fragment-based virtual screening approach to identify/optimize the fragment molecules as a starting point. This method combines the shape, electrostatic, and pharmacophoric features of known fragment molecules, bound to protein conjugate crystal structure, and aims to identify both chemically and energetically feasible small fragment ligands that bind to BACE1 active site. The two top-ranked fragment hits were subjected for a 53 ns MD simulation. Principle component analysis and free energy landscape analysis reveal that the new ligands show the characteristic features of established BACE1 inhibitors. The potent method employed in this study may serve for the development of potential lead molecules for BACE1-directed Alzheimer’s disease therapeutics.     

Hit clustering can improve virtual fragment screening: CDK2 and PARP1 case studies

Virtual fragment screening could be a promising alternative to existing experimental screening techniques. However, reliable methods of in silico fragment screening are yet to be established and validated. In order to develop such an approach we first checked how successful the existing molecular docking methods can be in predicting fragment binding affinities and poses. Using our Lead Finder docking software the RMSD of the binding energy prediction was observed to be 1.35 kcal/mol(-1) on a set of 26 experimentally characterized fragment inhibitors, and the RMSD of the predicted binding pose from the experimental one was <1.5 ?. Then, we explored docking of 68 fragments obtained from 39 drug molecules for which co-crystal structures were available from the PDB. It appeared that fragments that participate in oriented non-covalent interactions, such as hydrogen bonds and metal coordination, could be correctly docked in 70-80% of cases suggesting the potential success of rediscovering of corresponding drugs by in silico fragment approach. Based on these findings we've developed a virtual fragment screening technique which involved structural filtration of protein-ligand complexes for specific interactions and subsequent clustering in order to minimize the number of preferable starting fragment candidates. Application of this method led to 2 millimolar-scale fragment PARP1 inhibitors with a new scaffold.     

Microtubule-associated protein (MAP) 4 interacts with microtubules in an intrinsically disordered manner

We previously used nuclear magnetic resonance (NMR) to analyze the structure of a synthetic tricosapeptide corresponding to an active site of microtubule-associated protein 4 (MAP4). To further the structural analysis, we have constructed a minimal active domain fragment of MAP4, encompassing the entire active site, and obtained its NMR spectra. The secondary structure prediction using partially assigned NMR data suggested that the fragment is largely unfolded. Two other independent techniques also demonstrated its unfolded nature, indicating that MAP4 belongs to the class of intrinsically disordered proteins (IDPs). The NMR spectra of the fragment-microtubule mixture revealed that the fragment binds to the microtubule using multiple binding sites, apparently contradicting our previous quantitative studies. Given that MAP4 is intrinsically disordered, we propose a mechanism in which any one of the binding sites is active at a time, which is one of the typical interaction mechanisms proposed for IDPs.     

In silico studies for the rational discovery of anticonvulsant compounds

Theoretical models to virtual screening and rational design of anticonvulsant compounds based on a topological sub-structural molecular design (TOSS-MODE) approach are developed. These models, developed on the basis of data sets of great structural variability, permit the classification of compounds as active/inactive anticonvulsants and predict the quantitative anticonvulsant potency of such compounds. The classification model is applied to a virtual screening of anticonvulsant compounds by analyzing a data set of molecules reported in the literature. More than 88% of them were well classified by the current model. Active and inactive fragments are identified by using the present approach. Some of the active fragments are identified in anticonvulsant molecules as potential pharmacophores and one of them is analyzed in detail. The three-dimensional (3-D) features of this fragment are investigated in a series of five anticonvulsant compounds. Some structure–anticonvulsant activity relationships are derived on the basis of the 3-D structure of this fragment and some findings reported in the literature that indicate that it is an important pharamacophore are outlined.     

Structure based virtual screening to identify inhibitors against MurE Enzyme of Mycobacterium tuberculosis using AutoDock Vina

The Mur E enzyme of Mur pathway of Mycobacterium tuberculosis is an attractive drug target as it is unique to bacteria and is absent in mammalian cells. The virtual screening of large libraries of drug like molecules against a protein target is a common strategy used to identify novel inhibitors. However, the method has a large number of pitfalls, with large variations in accuracy caused in part by inaccurate protocols, use of improper standards and libraries, and system dependencies such as the potential for nonspecific docking from large active-site cavities. The screening of drug-like small molecules from diversity sets can, however, be used to short-list potential fragments as building blocks to generate leads with improved specificity. We describe a protocol to implement this strategy, which involves an analysis of the active site and known inhibitors to identify orthospecific determinants, virtual screening of a drug-like diversity library to identify potential drug primitives, and inspection of the potential docked fragments for both binding potential and toxicity. The protocol is implemented on the M.tb Mur E protein which has a large active site with poor enrichment of known positives and a set of drug-like molecules that meets this criteria is presented for further analysis.

Abbreviations

MTB - Mycobacterium tuberculosis, NCI - National Cancer Institute, PDB - Protein Databank.     

A novel approach to local similarity of protein binding sites substantially improves computational drug design results

We present a novel notion of binding site local similarity based on the analysis of complete protein environments of ligand fragments. Comparison of a query protein binding site (target) against the 3D structure of another protein (analog) in complex with a ligand enables ligand fragments from the analog complex to be transferred to positions in the target site, so that the complete protein environments of the fragment and its image are similar. The revealed environments are similarity regions and the fragments transferred to the target site are considered as binding patterns. The set of such binding patterns derived from a database of analog complexes forms a cloud-like structure (fragment cloud), which is a powerful tool for computational drug design. It has been shown on independent test sets that the combined use of a traditional energy-based score together with the cloud-based score responsible for the quality of embedding of a ligand into the fragment cloud improves the self-docking and screening results dramatically. The usage of a fragment cloud as a source of positioned molecular fragments fitting the binding protein environment has been validated by reproduction of experimental ligand optimization results.     

Dynameomics: Data‐driven methods and models for utilizing large‐scale protein structure repositories for improving fragment‐based loop prediction

Protein function is intimately linked to protein structure and dynamics yet experimentally determined structures frequently omit regions within a protein due to indeterminate data, which is often due protein dynamics. We propose that atomistic molecular dynamics simulations provide a diverse sampling of biologically relevant structures for these missing segments (and beyond) to improve structural modeling and structure prediction. Here we make use of the Dynameomics data warehouse, which contains simulations of representatives of essentially all known protein folds. We developed novel computational methods to efficiently identify, rank and retrieve small peptide structures, or fragments, from this database. We also created a novel data model to analyze and compare large repositories of structural data, such as contained within the Protein Data Bank and the Dynameomics data warehouse. Our evaluation compares these structural repositories for improving loop predictions and analyzes the utility of our methods and models. Using a standard set of loop structures, containing 510 loops, 30 for each loop length from 4 to 20 residues, we find that the inclusion of Dynameomics structures in fragment‐based methods improves the quality of the loop predictions without being dependent on sequence homology. Depending on loop length, ～25–75% of the best predictions came from the Dynameomics set, resulting in lower main chain root‐mean‐square deviations for all fragment lengths using the combined fragment library. We also provide specific cases where Dynameomics fragments provide better predictions for NMR loop structures than fragments from crystal structures. Online access to these fragment libraries is available at http://www.dynameomics.org/fragments .     

A new strategy for improved secondary screening and lead optimization using high-resolution SPR characterization of compound-target interactions

Biophysical label-free assays such as those based on SPR are essential tools in generating high-quality data on affinity, kinetic, mechanistic and thermodynamic aspects of interactions between target proteins and potential drug candidates. Here we show examples of the integration of SPR with bioinformatic approaches and mutation studies in the early drug discovery process. We call this combination 'structure-based biophysical analysis'. Binding sites are identified on target proteins using information that is either extracted from three-dimensional structural analysis (X-ray crystallography or NMR), or derived from a pharmacore model based on known binders. The binding site information is used for in silico screening of a large substance library (e.g. available chemical directory), providing virtual hits. The three-dimensional structure is also used for the design of mutants where the binding site has been impaired. The wild-type target and the impaired mutant are then immobilized on different spots of the sensor chip and the interactions of compounds with the wild-type and mutant are compared in order to identify selective binders for the binding site of the target protein. This method can be used as a cost-effective alternative to high-throughput screening methods in cases when detailed binding site information is available. Here, we present three examples of how this technique can be applied to provide invaluable data during different phases of the drug discovery process.