Biosensor for branched-chain amino acid metabolism in yeast and applications in isobutanol and isopentanol production

Branched-chain amino acid (BCAA) metabolism fulfills numerous physiological roles and can be harnessed to produce valuable chemicals. However, the lack of eukaryotic biosensors specific for BCAA-derived products has limited the ability to develop high-throughput screens for strain engineering and metabolic studies. Here, we harness the transcriptional regulator Leu3p from Saccharomyces cerevisiae to develop a genetically encoded biosensor for BCAA metabolism. In one configuration, we use the biosensor to monitor yeast production of isobutanol, an alcohol derived from valine degradation. Small modifications allow us to redeploy Leu3p in another biosensor configuration that monitors production of the leucine-derived alcohol, isopentanol. These biosensor configurations are effective at isolating high-producing strains and identifying enzymes with enhanced activity from screens for branched-chain higher alcohol (BCHA) biosynthesis in mitochondria as well as cytosol. Furthermore, this biosensor has the potential to assist in metabolic studies involving BCAA pathways, and offers a blueprint to develop biosensors for other products derived from BCAA metabolism.

OPEN leaf: an open-source cloud-based phenotyping system for tracking dynamic changes at leaf-specific resolution in Arabidopsis

The first draft of the Arabidopsis genome was released more than 20 years ago and despite intensive molecular research, more than 30% of Arabidopsis genes remained uncharacterized or without an assigned function. This is in part due to gene redundancy within gene families or the essential nature of genes, where their deletion results in lethality (i.e., the dark genome). High-throughput plant phenotyping (HTPP) offers an automated and unbiased approach to characterize subtle or transient phenotypes resulting from gene redundancy or inducible gene silencing; however, commercial HTPP platforms remain unaffordable. Here we describe the design and implementation of OPEN leaf, an open-source HTPP system with cloud connectivity and remote bilateral communication to facilitate data collection, sharing and processing. OPEN leaf, coupled with the SMART imaging processing package was able to consistently document and quantify dynamic morphological changes over time at the whole rosette level and also at leaf-specific resolution when plants experienced changes in nutrient availability. The modular design of OPEN leaf allows for additional sensor integration. Notably, our data demonstrate that VIS sensors remain underutilized and can be used in high-throughput screens to identify characterize previously unidentified phenotypes in a leaf-specific manner.

Small Molecule Drugs Targeting Non-Coding RNAs as Treatments for Alzheimer’s Disease and Related Dementias

Despite the enormous burden of Alzheimer’s disease and related dementias (ADRD) on patients, caregivers, and society, only a few treatments with limited efficacy are currently available. While drug development conventionally focuses on disease-associated proteins, RNA has recently been shown to be druggable for therapeutic purposes as well. Approximately 70% of the human genome is transcribed into non-protein-coding RNAs (ncRNAs) such as microRNAs, long ncRNAs, and circular RNAs, which can adopt diverse structures and cellular functions. Many ncRNAs are specifically enriched in the central nervous system, and their dysregulation is implicated in ADRD pathogenesis, making them attractive therapeutic targets. In this review, we first detail why targeting ncRNAs with small molecules is a promising therapeutic strategy for ADRD. We then outline the process from discovery to validation of small molecules targeting ncRNAs in preclinical studies, with special emphasis on primary high-throughput screens for identifying lead compounds. Screening strategies for specific ncRNAs will also be included as examples. Key challenges—including selecting appropriate ncRNA targets, lack of specificity of small molecules, and general low success rate of neurological drugs and how they may be overcome—will be discussed throughout the review.

Co-evolution based machine-learning for predicting functional interactions between human genes

Over the next decade, more than a million eukaryotic species are expected to be fully sequenced. This has the potential to improve our understanding of genotype and phenotype crosstalk, gene function and interactions, and answer evolutionary questions. Here, we develop a machine-learning approach for utilizing phylogenetic profiles across 1154 eukaryotic species. This method integrates co-evolution across eukaryotic clades to predict functional interactions between human genes and the context for these interactions. We benchmark our approach showing a 14% performance increase (auROC) compared to previous methods. Using this approach, we predict functional annotations for less studied genes. We focus on DNA repair and verify that 9 of the top 50 predicted genes have been identified elsewhere, with others previously prioritized by high-throughput screens. Overall, our approach enables better annotation of function and functional interactions and facilitates the understanding of evolutionary processes underlying co-evolution. The manuscript is accompanied by a webserver available at: https://mlpp.cs.huji.ac.il.

Stable Cell Clones Harboring Self-Replicating SARS-CoV-2 RNAs for Drug Screen

The development of antivirals against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been hampered by the lack of efficient cell-based replication systems that are amenable to high-throughput screens in biosafety level 2 laboratories. Here we report that stable cell clones harboring autonomously replicating SARS-CoV-2 RNAs without S, M, E genes can be efficiently derived from the baby hamster kidney (BHK-21) cell line when a pair of mutations were introduced into the non-structural protein 1 (Nsp1) of SARS-CoV-2 to ameliorate cellular toxicity associated with virus replication. In a proof-of-concept experiment we screened a 273-compound library using replicon cells and identified three compounds as novel inhibitors of SARS-CoV-2 replication. Altogether, this work establishes a robust, cell-based system for genetic and functional analyses of SARS-CoV-2 replication and for the development of antiviral drugs. IMPORTANCE: SARS-CoV-2 replicon systems that have been reported up to date were unsuccessful in deriving stable cell lines harboring non-cytopathic replicons. The transient expression of viral sgmRNA or a reporter gene makes it impractical for industry-scale screening of large compound libraries using these systems. Here, for the first time, we derived stable cell clones harboring the SARS-CoV-2 replicon. These clones may now be conveniently cultured in a standard BSL-2 laboratory for high throughput screen of compound libraries. This achievement represents a ground-breaking discovery that will greatly accelerate the pace of developing treatments for COVID-19.

Union is strength: target-based and whole-cell high throughput screens in antibacterial discovery

Antimicrobial resistance is one of the greatest global health challenges today. For over three decades antibacterial discovery research and development has been focused on cell-based and target-based high throughput assays. Target-based screens use diagnostic enzymatic reactions to look for molecules that can bind directly and inhibit the target. Target-based screens are only applied to proteins that can be successfully expressed, purified and the activity of which can be effectively measured using a biochemical assay. Often times the molecules found in these in vitro screens are not active in cells due to poor permeability or efflux. On the other hand, cell-based screens use whole cells and look for growth inhibition. These screens give higher number of hits than target-based assays and can simultaneously test many targets of one process or pathway in their physiological context. Both strategies have pros and cons when used separately. In the past decade and a half our increasing knowledge of bacterial physiology has led to the development of innovative and sophisticated technologies to perform high throughput screening combining these two strategies and thus minimizing their disadvantages. In this review we discuss recent examples of high throughput approaches that used both target-based and whole-cell screening to find new antibacterials, the new insights they have provided and how this knowledge can be applied to other in vivo validated targets to develop new antimicrobials.

AtomNet PoseRanker: Enriching Ligand Pose Quality for Dynamic Proteins in Virtual High Throughput Screens

Structure-based, virtual High Throughput Screening (vHTS) methods for predicting ligand activity in drug discovery are important when there are no or relatively few known compounds that interact with a therapeutic target of interest. State-of-the-art computational vHTS necessarily relies on effective methods for pose sampling and docking to generate an accurate affinity score from the docked poses. However, proteins are dynamic; in vivo, ligands bind to a conformational ensemble. In silico docking to the single conformation represented by a crystal structure can adversely affect the pose quality. Here we introduce AtomNet PoseRanker, a graph convolutional network trained to identify, and re-rank crystal-like ligand poses from a sampled ensemble of protein conformations and ligand poses. In contrast to conventional vHTS methods that incorporate receptor flexibility, a deep learning approach can internalize valid cognate and non-cognate binding modes corresponding to distinct receptor conformations. AtomNet PoseRanker significantly enriched pose quality in docking to cognate and non-cognate receptors of the PDBbind v2019 dataset. Improved pose rankings that better represent experimentally observed ligand binding modes improve hit rates in vHTS campaigns, and thereby advance computational drug discovery, especially for novel therapeutic targets or novel binding sites.

Machine learning methods, databases and tools for drug combination prediction

Combination therapy has shown an obvious efficacy on complex diseases and can greatly reduce the development of drug resistance. However, even with high-throughput screens, experimental methods are insufficient to explore novel drug combinations. In order to reduce the search space of drug combinations, there is an urgent need to develop more efficient computational methods to predict novel drug combinations. In recent decades, more and more machine learning (ML) algorithms have been applied to improve the predictive performance. The object of this study is to introduce and discuss the recent applications of ML methods and the widely used databases in drug combination prediction. In this study, we first describe the concept and controversy of synergism between drug combinations. Then, we investigate various publicly available data resources and tools for prediction tasks. Next, ML methods including classic ML and deep learning methods applied in drug combination prediction are introduced. Finally, we summarize the challenges to ML methods in prediction tasks and provide a discussion on future work.

CFTR Lifecycle Map—A Systems Medicine Model of CFTR Maturation to Predict Possible Active Compound Combinations

Different causative therapeutics for CF patients have been developed. There are still no mutation-specific therapeutics for some patients, especially those with rare CFTR mutations. For this purpose, high-throughput screens have been performed which result in various candidate compounds, with mostly unclear modes of action. In order to elucidate the mechanism of action for promising candidate substances and to be able to predict possible synergistic effects of substance combinations, we used a systems biology approach to create a model of the CFTR maturation pathway in cells in a standardized, human- and machine-readable format. It is composed of a core map, manually curated from small-scale experiments in human cells, and a coarse map including interactors identified in large-scale efforts. The manually curated core map includes 170 different molecular entities and 156 reactions from 221 publications. The coarse map encompasses 1384 unique proteins from four publications. The overlap between the two data sources amounts to 46 proteins. The CFTR Lifecycle Map can be used to support the identification of potential targets inside the cell and elucidate the mode of action for candidate substances. It thereby provides a backbone to structure available data as well as a tool to develop hypotheses regarding novel therapeutics.