Emerging frontiers in virtual drug discovery: From quantum mechanical methods to deep learning approaches

Virtual screening-based approaches to discover initial hit and lead compounds have the potential to reduce both the cost and time of early drug discovery stages, as well as to find inhibitors for even challenging target sites such as protein–protein interfaces. Here in this review, we provide an overview of the progress that has been made in virtual screening methodology and technology on multiple fronts in recent years. The advent of ultra-large virtual screens, in which hundreds of millions to billions of compounds are screened, has proven to be a powerful approach to discover highly potent hit compounds. However, these developments are just the tip of the iceberg, with new technologies and methods emerging to propel the field forward. Examples include novel machine-learning approaches, which can reduce the computational costs of virtual screening dramatically, while progress in quantum-mechanical approaches can increase the accuracy of predictions of various small molecule properties.     

Restricted N-glycan Conformational Space in the PDB and Its Implication in Glycan Structure Modeling

Understanding glycan structure and dynamics is central to understanding protein-carbohydrate recognition and its role in protein-protein interactions. Given the difficulties in obtaining the glycan''s crystal structure in glycoconjugates due to its flexibility and heterogeneity, computational modeling could play an important role in providing glycosylated protein structure models. To address if glycan structures available in the PDB can be used as templates or fragments for glycan modeling, we present a survey of the N-glycan structures of 35 different sequences in the PDB. Our statistical analysis shows that the N-glycan structures found on homologous glycoproteins are significantly conserved compared to the random background, suggesting that N-glycan chains can be confidently modeled with template glycan structures whose parent glycoproteins share sequence similarity. On the other hand, N-glycan structures found on non-homologous glycoproteins do not show significant global structural similarity. Nonetheless, the internal substructures of these N-glycans, particularly, the substructures that are closer to the protein, show significantly similar structures, suggesting that such substructures can be used as fragments in glycan modeling. Increased interactions with protein might be responsible for the restricted conformational space of N-glycan chains. Our results suggest that structure prediction/modeling of N-glycans of glycoconjugates using structure database could be effective and different modeling approaches would be needed depending on the availability of template structures.     

Chemical compound navigator: a web-based chem-BLAST, chemical taxonomy-based search engine for browsing compounds

A novel technique to annotate, query, and analyze chemical compounds has been developed and is illustrated by using the inhibitor data on HIV protease-inhibitor complexes. In this method, all chemical compounds are annotated in terms of standard chemical structural fragments. These standard fragments are defined by using criteria, such as chemical classification; structural, chemical, or functional groups; and commercial, scientific or common names or synonyms. These fragments are then organized into a data tree based on their chemical substructures. Search engines have been developed to use this data tree to enable query on inhibitors of HIV protease (http://xpdb.nist.gov/hivsdb/hivsdb.html). These search engines use a new novel technique, Chemical Block Layered Alignment of Substructure Technique (Chem-BLAST) to search on the fragments of an inhibitor to look for its chemical structural neighbors. This novel technique to annotate and query compounds lays the foundation for the use of the Semantic Web concept on chemical compounds to allow end users to group, sort, and search structural neighbors accurately and efficiently. During annotation, it enables the attachment of "meaning" (i.e., semantics) to data in a manner that far exceeds the current practice of associating "metadata" with data by creating a knowledge base (or ontology) associated with compounds. Intended users of the technique are the research community and pharmaceutical industry, for which it will provide a new tool to better identify novel chemical structural neighbors to aid drug discovery.     

Computational design of peptide ligands

Peptides possess several attractive features when compared to small molecule and protein therapeutics, such as high structural compatibility with target proteins, the ability to disrupt protein-protein interfaces, and small size. Efficient design of high-affinity peptide ligands via rational methods has been a major obstacle to the development of this potential drug class. However, structural insights into the architecture of protein-peptide interfaces have recently culminated in several computational approaches for the rational design of peptides that target proteins. These methods provide a valuable alternative to experimental high-resolution structures of target protein-peptide complexes, bringing closer the dream of in silico designed peptides for therapeutic applications.     

MUSTA--a general, efficient, automated method for multiple structure alignment and detection of common motifs: application to proteins.

Here we present an algorithm designed to carry out multiple structure alignment and to detect recurring substructural motifs. So far we have implemented it for comparison of protein structures. However, this general method is applicable to comparisons of RNA structures and to detection of a pharmacophore in a series of drug molecules. Further, its sequence order independence permits its application to detection of motifs on protein surfaces, interfaces, and binding/active sites. While there are many methods designed to carry out pairwise structure comparisons, there are only a handful geared toward the multiple structure alignment task. Most of these tackle multiple structure comparison as a collection of pairwise structure comparison tasks. The multiple structural alignment algorithm presented here automatically finds the largest common substructure (core) of atoms that appears in all the molecules in the ensemble. The detection of the core and the structural alignment are done simultaneously. The algorithm begins by finding small substructures that are common to all the proteins in the ensemble. One of the molecules is considered the reference; the others are the source molecules. The small substructures are stored in special arrays termed combinatorial buckets, which define sets of multistructural alignments from the source molecules that coincide with the same small set of reference atoms (C(alpha)-atoms here). These substructures are initial small fragments that have congruent copies in each of the proteins. The substructures are extended, through the processing of the combinatorial buckets, by clustering the superpositions (transformations). The method is very efficient.     

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.     

Yeast "N"-hybrid systems for protein-protein and drug-protein interaction discovery

The majority of small molecule drugs act on protein targets to exert a therapeutic function. It has become apparent in recent years that many small molecule drugs act on more than one particular target and consequently, approaches which profile drugs to uncover their target binding spectrum have become increasingly important. Classical yeast two-hybrid systems have mainly been used to discover and characterize protein-protein interactions, but recent modifications and improvements have opened up new routes towards screening for small molecule-protein interactions. Such yeast "n"-hybrid systems hold great promise for the development of drugs which interfere with protein-protein interactions and for the discovery of drug-target interactions. In this review, we discuss several yeast two-hybrid based approaches with applications in drug discovery and describe a protocol for yeast three-hybrid screening of small molecules to identify their direct targets.     

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.     

From in silico target prediction to multi-target drug design: current databases, methods and applications

Given the tremendous growth of bioactivity databases, the use of computational tools to predict protein targets of small molecules has been gaining importance in recent years. Applications span a wide range, from the 'designed polypharmacology' of compounds to mode-of-action analysis. In this review, we firstly survey databases that can be used for ligand-based target prediction and which have grown tremendously in size in the past. We furthermore outline methods for target prediction that exist, both based on the knowledge of bioactivities from the ligand side and methods that can be applied in situations when a protein structure is known. Applications of successful in silico target identification attempts are discussed in detail, which were based partly or in whole on computational target predictions in the first instance. This includes the authors' own experience using target prediction tools, in this case considering phenotypic antibacterial screens and the analysis of high-throughput screening data. Finally, we will conclude with the prospective application of databases to not only predict, retrospectively, the protein targets of a small molecule, but also how to design ligands with desired polypharmacology in a prospective manner.     

Evidence for ligandable sites in structured RNA throughout the Protein Data Bank

RNA has attracted considerable attention as a target for small molecules. However, methods to identify, study, and characterize suitable RNA targets have lagged behind strategies for protein targets. One approach that has received considerable attention for protein targets has been to utilize computational analysis to investigate ligandable “pockets” on proteins that are amenable to small molecule binding. These studies have shown that selected physical properties of pockets are important parameters that govern the ability of a structure to bind to small molecules. This work describes a similar analysis to study pockets on all RNAs in the Protein Data Bank (PDB). Using parameters such as buriedness, hydrophobicity, volume, and other properties, the set of all RNAs is analyzed and compared to all proteins. Considerable overlap is observed between the properties of pockets on RNAs and proteins. Thus, many RNAs are capable of populating conformations with pockets that are likely suitable for small molecule binding. Further, principal moment of inertia (PMI) calculations reveal that liganded RNAs exist in diverse structural space, much of which overlaps with protein structural space. Taken together, these results suggest that complex folded RNAs adopt unique structures with pockets that may represent viable opportunities for small molecule targeting.     

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.     

Affinity Selection and Mass Spectrometry-Based Strategies to Identify Lead Compounds in Combinatorial Libraries

The screening of diverse libraries of small molecules created by combinatorial synthetic methods is a recent development which has the potential to accelerate the identification of lead compounds in drug discovery. We have developed a direct and rapid method to identify lead compounds in libraries involving affinity selection and mass spectrometry. In our strategy, the receptor or target molecule of interest is used to isolate the active components from the library physically, followed by direct structural identification of the active compounds bound to the target molecule by mass spectrometry. In a drug design strategy, structurally diverse libraries can be used for the initial identification of lead compounds. Once lead compounds have been identified, libraries containing compounds chemically similar to the lead compound can be generated and used to optimize the binding characteristics. These strategies have also been adopted for more detailed studies of protein–ligand interactions.     

Discovery of a novel small molecule binding site of human survivin

Survivin is one of the most tumor-specific genes in the human genome and is an attractive target for cancer therapy. However, small-molecule ligands for survivin have not yet been described. Thus, an interrogation of survivin which could potentially both validate a small-molecule therapy approach, and determine the biochemical nature of any of survivin's functions has not been possible. Here we describe the discovery and characterization of a small molecule binding site on the survivin surface distinct from the Smac peptide-binding site. The new site is located at the dimer interface and exhibits many of the features of highly druggable, biologically relevant protein binding sites. A variety of small hydrophobic compounds were found that bind with moderate affinity to this binding site, from which one lead was developed into a group of compounds with nanomolar affinity. Additionally, a subset of these compounds are adequately water-soluble and cell-permeable. Thus, the structural studies and small molecules described here provide tools that can be used to probe the biochemical role(s) of survivin, and may ultimately serve as a basis for the development of small molecule therapeutics acting via direct or allosteric disruption of binding events related to this poorly understood target.     

HYPROSP: a hybrid protein secondary structure prediction algorithm--a knowledge-based approach

We develop a knowledge-based approach (called PROSP) for protein secondary structure prediction. The knowledge base contains small peptide fragments together with their secondary structural information. A quantitative measure M, called match rate, is defined to measure the amount of structural information that a target protein can extract from the knowledge base. Our experimental results show that proteins with a higher match rate will likely be predicted more accurately based on PROSP. That is, there is roughly a monotone correlation between the prediction accuracy and the amount of structure matching with the knowledge base. To fully utilize the strength of our knowledge base, a hybrid prediction method is proposed as follows: if the match rate of a target protein is at least 80%, we use the extracted information to make the prediction; otherwise, we adopt a popular machine-learning approach. This comprises our hybrid protein structure prediction (HYPROSP) approach. We use the DSSP and EVA data as our datasets and PSIPRED as our underlying machine-learning algorithm. For target proteins with match rate at least 80%, the average Q₃ of PROSP is 3.96 and 7.2 better than that of PSIPRED on DSSP and EVA data, respectively.     

Developing hybrid molecule therapeutics for diverse enzyme inhibitory action: Active role of coumarin-based structural leads in drug discovery

Hybrid drugs featuring two or more potentially bioactive pharmacophores have been recognized as advanced and superior chemical entities to simultaneously modulate multiple drug targets of multifactorial diseases, thus overcoming the severe side effects associated with a single drug molecule. The selection of these chemical moieties to produce hybrid structures with druggable properties is generally facilitated by the observed and/or anticipated synergistic pharmacological activities of the individual molecules. In this perspective, coumarin template has extensively been studied in pursuit of structurally diverse leads for drug development due to high affinity and specificity to different molecular targets. This review highlights the most commonly exploited approaches conceptualizing the design and construction of hybrid molecules by coupling two or more individual fragments with or without an appropriate linker. In addition to the design strategies, this review also summarizes and reflects on the therapeutic potential of these hybrid molecules for diverse enzyme inhibitory action as well as their observed structure-activity relationship (SAR). Several key features of the synthesized hybrid structures that assert a profound impact on the inhibitory function have also been discussed alongside computational investigations, inhibitor molecular diversity and selectivity toward multiple drug targets. Finally, these drug discovery and development efforts should serve as a handy reference aiming to provide a useful platform for the exploration of new coumarin-based compounds with enhanced enzyme inhibitory profile.     

Inhibition of hepatitis C virus NS5A by fluoro-olefin based γ-turn mimetics

The HCV non-structural protein NS5A has been established as a viable target for the development of direct acting antiviral therapy. From computational modeling studies strong intra-molecular hydrogen bonds were found to be a common structural moiety within known NS5A inhibitors that have low pico-molar replicon potency. Efforts to reproduce these γ-turn-like substructures provided a novel NS5A inhibitor based on a fluoro-olefin isostere. This fluoro-olefin containing inhibitor exhibited picomolar activity (EC(50)=79 pM) against HCV genotype 1b replicon without measurable cytotoxicity. This level of activity is comparable to the natural peptide-based inhibitors currently under clinic evaluation, and demonstrates that a peptidomimetic approach can serve as a useful strategy to produce potent and structurally unique inhibitors of HCV NS5A.     

Intracellular antibody capture: A molecular biology approach to inhibitors of protein–protein interactions

Many proteins of interest in basic biology, translational research studies and for clinical targeting in diseases reside inside the cell and function by interacting with other macromolecules. Protein complexes control basic processes such as development and cell division but also abnormal cell growth when mutations occur such as found in cancer. Interfering with protein–protein interactions is an important aspiration in both basic and disease biology but small molecule inhibitors have been difficult and expensive to isolate. Recently, we have adapted molecular biology techniques to develop a simple set of protocols for isolation of high affinity antibody fragments (in the form of single VH domains) that function within the reducing environment of higher organism cells and can bind to their target molecules. The method called Intracellular Antibody Capture (IAC) has been used to develop inhibitory anti-RAS and anti-LMO2 single domains that have been used for target validation of these antigens in pre-clinical cancer models and illustrate the efficacy of the IAC approach to generation of drug surrogates. Future use of inhibitory VH antibody fragments as drugs in their own right (we term these macrodrugs to distinguish them from small molecule drugs) requires their delivery to target cells in vivo but they can also be templates for small molecule drug development that emulate the binding sites of the antibody fragments. This article is part of a Special Issue entitled: Recent advances in molecular engineering of antibody.     

Bioinformatics approaches for new drug discovery: a review

Prolonged antibiotic therapy for the bacterial infections has resulted in high levels of antibiotic resistance. Initially, bacteria are susceptible to the antibiotics, but can gradually develop resistance. Treating such drug-resistant bacteria remains difficult or even impossible. Hence, there is a need to develop effective drugs against bacterial pathogens. The drug discovery process is time-consuming, expensive and laborious. The traditionally available drug discovery process initiates with the identification of target as well as the most promising drug molecule, followed by the optimization of this, in-vitro, in-vivo and in pre-clinical studies to decide whether the compound has the potential to be developed as a drug molecule. Drug discovery, drug development and commercialization are complicated processes. To overcome some of these problems, there are many computational tools available for new drug discovery, which could be cost effective and less time-consuming. In-silico approaches can reduce the number of potential compounds from hundreds of thousands to the tens of thousands which could be studied for drug discovery and this results in savings of time, money and human resources. Our review is on the various computational methods employed in new drug discovery processes.     

Displacement of protein-bound aptamers with small molecules screened by fluorescence polarization

Small molecule inhibitors of proteins are invaluable tools in research and as starting points for drug development. However, their screening can be tedious, as most screening methods have to be tailored to the corresponding drug target. Here, we describe a detailed protocol for a modular and generally applicable assay for the identification of small organic compounds that displace an aptamer complexed to its target protein. The method relies on fluorescence-labeled aptamers and the increase of fluorescence polarization upon their binding to the target protein. The assay has high Z'-factors, making it compatible with high-throughput screening. It allows easy automation, making fluorescence readout the time-limiting step. As aptamers can be generated for virtually any protein target, the assay allows identification of small molecule inhibitors for targets or individual protein domains for which no functional screen is available. We provide the step-by-step protocol to screen for antagonists of the cytohesin class of small guanosine exchange factors.