A Deep Learning Approach to Antibiotic Discovery.

Due to the rapid emergence of antibiotic-resistant bacteria, there is a growing need to discover new antibiotics. To address this challenge, we trained a deep neural network capable of predicting molecules with antibacterial activity. We performed predictions on multiple chemical libraries and discovered a molecule from the Drug Repurposing Hub-halicin-that is structurally divergent from conventional antibiotics and displays bactericidal activity against a wide phylogenetic spectrum of pathogens including Mycobacterium tuberculosis and carbapenem-resistant Enterobacteriaceae. Halicin also effectively treated Clostridioides difficile and pan-resistant Acinetobacter baumannii infections in murine models. Additionally, from a discrete set of 23 empirically tested predictions from >107 million molecules curated from the ZINC15 database, our model identified eight antibacterial compounds that are structurally distant from known antibiotics. This work highlights the utility of deep learning approaches to expand our antibiotic arsenal through the discovery of structurally distinct antibacterial molecules.

A Comprehensive, CRISPR-based Functional Analysis of Essential Genes in Bacteria.

Essential gene functions underpin the core reactions required for cell viability, but their contributions and relationships are poorly studied in vivo. Using CRISPR interference, we created knockdowns of every essential gene in Bacillus subtilis and probed their phenotypes. Our high-confidence essential gene network, established using chemical genomics, showed extensive interconnections among distantly related processes and identified modes of action for uncharacterized antibiotics. Importantly, mild knockdown of essential gene functions significantly reduced stationary-phase survival without affecting maximal growth rate, suggesting that essential protein levels are set to maximize outgrowth from stationary phase. Finally, high-throughput microscopy indicated that cell morphology is relatively insensitive to mild knockdown but profoundly affected by depletion of gene function, revealing intimate connections between cell growth and shape. Our results provide a framework for systematic investigation of essential gene functions in vivo broadly applicable to diverse microorganisms and amenable to comparative analysis.

Identification of drugs including a dopamine receptor antagonist that selectively target cancer stem cells.

Selective targeting of cancer stem cells (CSCs) offers promise for a new generation of therapeutics. However, assays for both human CSCs and normal stem cells that are amenable to robust biological screens are limited. Using a discovery platform that reveals differences between neoplastic and normal human pluripotent stem cells (hPSC), we identify small molecules from libraries of known compounds that induce differentiation to overcome neoplastic self-renewal. Surprisingly, thioridazine, an antipsychotic drug, selectively targets the neoplastic cells, and impairs human somatic CSCs capable of in vivo leukemic disease initiation while having no effect on normal blood SCs. The drug antagonizes dopamine receptors that are expressed on CSCs and on breast cancer cells as well. These results suggest that dopamine receptors may serve as a biomarker for diverse malignancies, demonstrate the utility of using neoplastic hPSCs for identifying CSC-targeting drugs, and provide support for the use of differentiation as a therapeutic strategy.

Central metabolism is a key player in E. coli biofilm stimulation by sub-MIC antibiotics.

Exposure of Escherichia coli to sub-inhibitory antibiotics stimulates biofilm formation through poorly characterized mechanisms. Using a high-throughput Congo Red binding assay to report on biofilm matrix production, we screened ~4000 E. coli K12 deletion mutants for deficiencies in this biofilm stimulation response. We screened using three different antibiotics to identify core components of the biofilm stimulation response. Mutants lacking acnA, nuoE, or lpdA failed to respond to sub-MIC cefixime and novobiocin, implicating central metabolism and aerobic respiration in biofilm stimulation. These genes are members of the ArcA/B regulon-controlled by a respiration-sensitive two-component system. Mutants of arcA and arcB had a 'pre-activated' phenotype, where biofilm formation was already high relative to wild type in vehicle control conditions, and failed to increase further with the addition of sub-MIC cefixime. Using a tetrazolium dye and an in vivo NADH sensor, we showed spatial co-localization of increased metabolic activity with sub-lethal concentrations of the bactericidal antibiotics cefixime and novobiocin. Supporting a role for respiratory stress, the biofilm stimulation response to cefixime and novobiocin was inhibited when nitrate was provided as an alternative electron acceptor. Deletion of a gene encoding part of the machinery for respiring nitrate abolished its ameliorating effects, and nitrate respiration increased during growth with sub-MIC cefixime. Finally, in probing the generalizability of biofilm stimulation, we found that the stimulation response to translation inhibitors, unlike other antibiotic classes, was minimally affected by nitrate supplementation, suggesting that targeting the ribosome stimulates biofilm formation in distinct ways. By characterizing the biofilm stimulation response to sub-MIC antibiotics at a systems level, we identified multiple avenues for design of therapeutics that impair bacterial stress management.

CARD 2023: expanded curation, support for machine learning, and resistome prediction at the Comprehensive Antibiotic Resistance Database.

The Comprehensive Antibiotic Resistance Database (CARD; card.mcmaster.ca) combines the Antibiotic Resistance Ontology (ARO) with curated AMR gene (ARG) sequences and resistance-conferring mutations to provide an informatics framework for annotation and interpretation of resistomes. As of version 3.2.4, CARD encompasses 6627 ontology terms, 5010 reference sequences, 1933 mutations, 3004 publications, and 5057 AMR detection models that can be used by the accompanying Resistance Gene Identifier (RGI) software to annotate genomic or metagenomic sequences. Focused curation enhancements since 2020 include expanded ß-lactamase curation, incorporation of likelihood-based AMR mutations for Mycobacterium tuberculosis, addition of disinfectants and antiseptics plus their associated ARGs, and systematic curation of resistance-modifying agents. This expanded curation includes 180 new AMR gene families, 15 new drug classes, 1 new resistance mechanism, and two new ontological relationships: evolutionary_variant_of and is_small_molecule_inhibitor. In silico prediction of resistomes and prevalence statistics of ARGs has been expanded to 377 pathogens, 21,079 chromosomes, 2,662 genomic islands, 41,828 plasmids and 155,606 whole-genome shotgun assemblies, resulting in collation of 322,710 unique ARG allele sequences. New features include the CARD:Live collection of community submitted isolate resistome data and the introduction of standardized 15 character CARD Short Names for ARGs to support machine learning efforts.

A genetic platform to investigate the functions of bacterial drug efflux pumps.

Efflux pumps are a serious challenge for the development of antibacterial agents. Overcoming efflux requires an in-depth understanding of efflux pump functions, specificities and the development of inhibitors. However, the complexities of efflux networks have limited such studies. To address these challenges, we generated Efflux KnockOut-35 (EKO-35), a highly susceptible Escherichia coli strain lacking 35 efflux pumps. We demonstrate the use of this strain by constructing an efflux platform comprising EKO-35 strains individually producing efflux pumps forming tripartite complexes with TolC. This platform was profiled against a curated diverse compound collection, which enabled us to define physicochemical properties that contribute to transport. We also show the E. coli drug efflux network is conditionally essential for growth, and that the platform can be used to investigate efflux pump inhibitor specificities and efflux pump interplay. We believe EKO-35 and the efflux platform will have widespread application for the study of drug efflux.

An Omniphobic Spray Coating Created from Hierarchical Structures Prevents the Contamination of High-Touch Surfaces with Pathogens.

Engineered surfaces that repel pathogens are of great interest due to their role in mitigating the spread of infectious diseases. A robust, universal, and scalable omniphobic spray coating with excellent repellency against water, oil, and pathogens is presented. The coating is substrate-independent and relies on hierarchically structured polydimethylsiloxane (PDMS) microparticles, decorated with gold nanoparticles (AuNPs). Wettability studies reveal the relationship between surface texturing of micro- and/or nano-hierarchical structures and the omniphobicity of the coating. Studies of pathogen transfer with bacteria and viruses reveal that an uncoated contaminated glove transfers pathogens to >50 subsequent surfaces, while a coated glove picks up 104 (over 99.99%) less pathogens upon first contact and transfers zero pathogens after the second touch. The developed coating also provides excellent stability under harsh conditions. The remarkable anti-pathogen properties of this surface combined with its ease of implementation, substantiate its use for the prevention of surface-mediated transmission of pathogens.

Transparent and Highly Flexible Hierarchically Structured Polydimethylsiloxane Surfaces Suppress Bacterial Attachment and Thrombosis Under Static and Dynamic Conditions.

The surface fouling of biomedical devices has been an ongoing issue in healthcare. Bacterial and blood adhesion in particular, severely impede the performance of such tools, leading to poor patient outcomes. Various structural and chemical modifications have been shown to reduce fouling, but all existing strategies lack the combination of physical, chemical, and economic traits necessary for widespread use. Herein, a lubricant infused, hierarchically micro- and nanostructured polydimethylsiloxane surface is presented. The surface is easy to produce and exhibits the high flexibility and optical transparency necessary for incorporation into various biomedical tools. Tests involving two clinically relevant, priority pathogens show up to a 98.5% reduction in the biofilm formation of methicillin-resistant Staphylococcus aureus and Pseudomonas aeruginosa. With blood, the surface reduces staining by 95% and suppresses thrombin generation to background levels. Furthermore, the surface shows applicability within applications such as catheters, extracorporeal circuits, and microfluidic devices, through its effectiveness in dynamic conditions. The perfusion of bacterial media shows up to 96.5% reduction in bacterial adhesion. Similarly, a 95.8% reduction in fibrin networks is observed following whole blood perfusion. This substrate stands to hold high applicability within biomedical systems as a means to prevent fouling, thus improving performance.

A multiplexable assay for screening antibiotic lethality against drug-tolerant bacteria.

Antibiotic screens typically rely on growth inhibition to characterize compound bioactivity-an approach that cannot be used to assess the bactericidal activity of antibiotics against bacteria in drug-tolerant states. To address this limitation, we developed a multiplexed assay that uses metabolism-sensitive staining to report on the killing of antibiotic-tolerant bacteria. This method can be used with diverse bacterial species and applied to genome-scale investigations to identify therapeutic targets against tolerant pathogens.

Systems-Level Chemical Biology to Accelerate Antibiotic Drug Discovery.

Drug-resistant bacterial infections pose an imminent and growing threat to public health. The discovery and development of new antibiotics of novel chemical class and mode of action that are unsusceptible to existing resistance mechanisms is imperative for tackling this threat. Modern industrial drug discovery, however, has failed to provide new drugs of this description, as it is dependent largely on a reductionist genes-to-drugs research paradigm. We posit that the lack of success in new antibiotic drug discovery is due in part to a lack of understanding of the bacterial cell system as whole. A fundamental understanding of the architecture and function of bacterial systems has been elusive but is of critical importance to design strategies to tackle drug-resistant bacterial pathogens.Increasingly, systems-level approaches are rewriting our understanding of the cell, defining a dense network of redundant and interacting components that resist perturbations of all kinds, including by antibiotics. Understanding the network properties of bacterial cells requires integrative, systematic, and genome-scale approaches. These methods strive to understand how the phenotypic behavior of bacteria emerges from the many interactions of individual molecular components that constitute the system. With the ability to examine genomic, transcriptomic, proteomic, and metabolomic consequences of, for example, genetic or chemical perturbations, researchers are increasingly moving away from one-gene-at-a-time studies to consider the system-wide response of the cell. Such measurements are demonstrating promise as quantitative tools, powerful discovery engines, and robust hypothesis generators with great value to antibiotic drug discovery.In this Account, we describe our thinking and findings using systems-level studies aimed at understanding bacterial physiology broadly and in uncovering new antibacterial chemical matter of novel mechanism. We share our systems-level toolkit and detail recent technological developments that have enabled unprecedented acquisition of genome-wide interaction data. We focus on three types of interactions: gene-gene, chemical-gene, and chemical-chemical. We provide examples of their use in understanding cell networks and how these insights might be harnessed for new antibiotic discovery. By example, we show the application of these principles in mapping genetic networks that underpin phenotypes of interest, characterizing genes of unknown function, validating small-molecule screening platforms, uncovering novel chemical probes and antibacterial leads, and delineating the mode of action of antibacterial chemicals. We also discuss the importance of computation to these approaches and its probable dominance as a tool for systems approaches in the future. In all, we advocate for the use of systems-based approaches as discovery engines in antibacterial research, both as powerful tools and to stimulate innovation.

Discovery of an antivirulence compound that reverses ß-lactam resistance in MRSA.

Staphylococcus aureus is the leading cause of infections worldwide, and methicillin-resistant strains (MRSA) are emerging. New strategies are urgently needed to overcome this threat. Using a cell-based screen of ~45,000 diverse synthetic compounds, we discovered a potent bioactive, MAC-545496, that reverses ß-lactam resistance in the community-acquired MRSA USA300 strain. MAC-545496 could also serve as an antivirulence agent alone; it attenuates MRSA virulence in Galleria mellonella larvae. MAC-545496 inhibits biofilm formation and abrogates intracellular survival in macrophages. Mechanistic characterization revealed MAC-545496 to be a nanomolar inhibitor of GraR, a regulator that responds to cell-envelope stress and is an important virulence factor and determinant of antibiotic resistance. The small molecule discovered herein is an inhibitor of GraR function. MAC-545496 has value as a research tool to probe the GraXRS regulatory system and as an antibacterial lead series of a mechanism to combat drug-resistant Staphylococcal infections.

Antibacterial drug discovery in the resistance era.

The looming antibiotic-resistance crisis has penetrated the consciousness of clinicians, researchers, policymakers, politicians and the public at large. The evolution and widespread distribution of antibiotic-resistance elements in bacterial pathogens has made diseases that were once easily treatable deadly again. Unfortunately, accompanying the rise in global resistance is a failure in antibacterial drug discovery. Lessons from the history of antibiotic discovery and fresh understanding of antibiotic action and the cell biology of microorganisms have the potential to deliver twenty-first century medicines that are able to control infection in the resistance era.

Crystallographic analysis of TarI and TarJ, a cytidylyltransferase and reductase pair for CDP-ribitol synthesis in Staphylococcus aureus wall teichoic acid biogenesis.

The cell wall of many pathogenic Gram-positive bacteria contains ribitol-phosphate wall teichoic acid (WTA), a polymer that is linked to virulence and regulation of essential physiological processes including cell division. CDP-ribitol, the activated precursor for ribitol-phosphate polymerization, is synthesized by a cytidylyltransferase and reductase pair known as TarI and TarJ, respectively. In this study, we present crystal structures of Staphylococcus aureus TarI and TarJ in their apo forms and in complex with substrates and products. The TarI structures illustrate the mechanism of CDP-ribitol synthesis from CTP and ribitol-phosphate and reveal structural changes required for substrate binding and catalysis. Insights into the upstream step of ribulose-phosphate reduction to ribitol-phosphate is provided by the structures of TarJ. Furthermore, we propose a general topology of the enzymes in a heterotetrameric form built using restraints from crosslinking mass spectrometry analysis. Together, our data present molecular details of CDP-ribitol production that may aid in the design of inhibitors against WTA biosynthesis.

Crystallographic analysis of Staphylococcus aureus LcpA, the primary wall teichoic acid ligase.

Gram-positive bacteria, including major clinical pathogens such as Staphylococcus aureus, are becoming increasingly drug-resistant. Their cell walls are composed of a thick layer of peptidoglycan (PG) modified by the attachment of wall teichoic acid (WTA), an anionic glycopolymer that is linked to pathogenicity and regulation of cell division and PG synthesis. The transfer of WTA from lipid carriers to PG, catalyzed by the LytR-CpsA-Psr (LCP) enzyme family, offers a unique extracellular target for the development of new anti-infective agents. Inhibitors of LCP enzymes have the potential to manage a wide range of bacterial infections because the target enzymes are implicated in the assembly of many other bacterial cell wall polymers, including capsular polysaccharide of streptococcal species and arabinogalactan of mycobacterial species. In this study, we present the first crystal structure of S. aureus LcpA with bound substrate at 1.9 Å resolution and those of Bacillus subtilis LCP enzymes, TagT, TagU, and TagV, in the apo form at 1.6-2.8 Å resolution. The structures of these WTA transferases provide new insight into the binding of lipid-linked WTA and enable assignment of the catalytic roles of conserved active-site residues. Furthermore, we identified potential subsites for binding the saccharide core of PG using computational docking experiments, and multiangle light-scattering experiments disclosed novel oligomeric states of the LCP enzymes. The crystal structures and modeled substrate-bound complexes of the LCP enzymes reported here provide insights into key features linked to substrate binding and catalysis and may aid the structure-guided design of specific LCP inhibitors.

Physicochemical and Structural Parameters Contributing to the Antibacterial Activity and Efflux Susceptibility of Small-Molecule Inhibitors of Escherichia coli.

Discovering new Gram-negative antibiotics has been a challenge for decades. This has been largely attributed to a limited understanding of the molecular descriptors governing Gram-negative permeation and efflux evasion. Herein, we address the contribution of efflux using a novel approach that applies multivariate analysis, machine learning, and structure-based clustering to some 4,500 molecules (actives) from a small-molecule screen in efflux-compromised Escherichia coli We employed principal-component analysis and trained two decision tree-based machine learning models to investigate descriptors contributing to the antibacterial activity and efflux susceptibility of these actives. This approach revealed that the Gram-negative activity of hydrophobic and planar small molecules with low molecular stability is limited to efflux-compromised E. coli Furthermore, molecules with reduced branching and compactness showed increased susceptibility to efflux. Given these distinct properties that govern efflux, we developed the first efflux susceptibility machine learning model, called Susceptibility to Efflux Random Forest (SERF), as a tool to analyze the molecular descriptors of small molecules and predict those that could be susceptible to efflux pumps in silico Here, SERF demonstrated high accuracy in identifying such molecules. Furthermore, we clustered all 4,500 actives based on their core structures and identified distinct clusters highlighting side-chain moieties that cause marked changes in efflux susceptibility. In all, our work reveals a role for physicochemical and structural parameters in governing efflux, presents a machine learning tool for rapid in silico analysis of efflux susceptibility, and provides a proof of principle for the potential of exploiting side-chain modification to design novel antimicrobials evading efflux pumps.

Structure and mechanism of TagA, a novel membrane-associated glycosyltransferase that produces wall teichoic acids in pathogenic bacteria.

Staphylococcus aureus and other bacterial pathogens affix wall teichoic acids (WTAs) to their surface. These highly abundant anionic glycopolymers have critical functions in bacterial physiology and their susceptibility to ß-lactam antibiotics. The membrane-associated TagA glycosyltransferase (GT) catalyzes the first-committed step in WTA biosynthesis and is a founding member of the WecB/TagA/CpsF GT family, more than 6,000 enzymes that synthesize a range of extracellular polysaccharides through a poorly understood mechanism. Crystal structures of TagA from T. italicus in its apo- and UDP-bound states reveal a novel GT fold, and coupled with biochemical and cellular data define the mechanism of catalysis. We propose that enzyme activity is regulated by interactions with the bilayer, which trigger a structural change that facilitates proper active site formation and recognition of the enzyme's lipid-linked substrate. These findings inform upon the molecular basis of WecB/TagA/CpsF activity and could guide the development of new anti-microbial drugs.

A whole-cell, high-throughput hydrogenase assay to identify factors that modulate [NiFe]-hydrogenase activity.

[NiFe]-hydrogenases have attracted attention as potential therapeutic targets or components of a hydrogen-based economy. [NiFe]-hydrogenase production is a complicated process that requires many associated accessory proteins that supply the requisite cofactors and substrates. Current methods for measuring hydrogenase activity have low throughput and often require specialized conditions and reagents. In this work, we developed a whole-cell high-throughput hydrogenase assay based on the colorimetric reduction of benzyl viologen to explore the biological networks of these enzymes in Escherichia coli We utilized this assay to screen the Keio collection, a set of nonlethal single-gene knockouts in E. coli BW25113. The results of this screen highlighted the assay's specificity and revealed known components of the intricate network of systems that underwrite [NiFe]-hydrogenase activity, including nickel homeostasis and formate dehydrogenase activities as well as molybdopterin and selenocysteine biosynthetic pathways. The screen also helped identify several new genetic components that modulate hydrogenase activity. We examined one E. coli strain with undetectable hydrogenase activity in more detail (ΔeutK), finding that nickel delivery to the enzyme active site was completely abrogated, and tracked this effect to an ancillary and unannotated lack of the fumarate and nitrate reduction (FNR) anaerobic regulatory protein. Collectively, these results demonstrate that the whole-cell assay developed here can be used to uncover new information about bacterial [NiFe]-hydrogenase production and to probe the cellular components of microbial nickel homeostasis.

Meropenem potentiation of aminoglycoside activity against Pseudomonas aeruginosa: involvement of the MexXY-OprM multidrug efflux system.

Objectives: To assess the ability of meropenem to potentiate aminoglycoside (AG) activity against laboratory and AG-resistant cystic fibrosis (CF) isolates of Pseudomonas aeruginosa and to elucidate its mechanism of action. Methods: AG resistance gene deletions were engineered into P. aeruginosa laboratory and CF isolates using standard gene replacement technology. Susceptibility to AGs ± meropenem (at ½ MIC) was assessed using a serial 2-fold dilution assay. mexXY expression and MexXY-OprM efflux activity were quantified using quantitative PCR and an ethidium bromide accumulation assay, respectively. Results: A screen for agents that rendered WT P. aeruginosa susceptible to a sub-MIC concentration of the AG paromomycin identified the carbapenem meropenem, which potentiated several additional AGs. Meropenem potentiation of AG activity was largely lost in a mutant lacking the MexXY-OprM multidrug efflux system, an indication that it was targeting this efflux system in enhancing P. aeruginosa susceptibility to AGs. Meropenem failed to block AG induction of mexXY expression or MexXY-OprM efflux activity, suggesting that it may be interfering with some MexXY-dependent process linked to AG susceptibility. Meropenem potentiated AG activity versus AG-resistant CF isolates, enhancing susceptibility to at least one AG in all isolates and susceptibility to all tested AGs in 50% of the isolates. Notably, meropenem potentiation of AG activity was linked to MexXY in some but not all CF isolates in which this was examined. Conclusions: Meropenem potentiates AG activity against laboratory and CF strains of P. aeruginosa, both dependent on and independent of MexXY, highlighting the complexity of AG resistance in this organism.

Structure and Mechanism of Staphylococcus aureus TarS, the Wall Teichoic Acid ß-glycosyltransferase Involved in Methicillin Resistance.

In recent years, there has been a growing interest in teichoic acids as targets for antibiotic drug design against major clinical pathogens such as Staphylococcus aureus, reflecting the disquieting increase in antibiotic resistance and the historical success of bacterial cell wall components as drug targets. It is now becoming clear that ß-O-GlcNAcylation of S. aureus wall teichoic acids plays a major role in both pathogenicity and antibiotic resistance. Here we present the first structure of S. aureus TarS, the enzyme responsible for polyribitol phosphate ß-O-GlcNAcylation. Using a divide and conquer strategy, we obtained crystal structures of various TarS constructs, mapping high resolution overlapping N-terminal and C-terminal structures onto a lower resolution full-length structure that resulted in a high resolution view of the entire enzyme. Using the N-terminal structure that encapsulates the catalytic domain, we furthermore captured several snapshots of TarS, including the native structure, the UDP-GlcNAc donor complex, and the UDP product complex. These structures along with structure-guided mutants allowed us to elucidate various catalytic features and identify key active site residues and catalytic loop rearrangements that provide a valuable platform for anti-MRSA drug design. We furthermore observed for the first time the presence of a trimerization domain composed of stacked carbohydrate binding modules, commonly observed in starch active enzymes, but adapted here for a poly sugar-phosphate glycosyltransferase.