Cell envelope structural and functional contributions to antibiotic resistance in Burkholderia cenocepacia.

Antibiotic activity is limited by the physical construction of the Gram-negative cell envelope. Species of the Burkholderia cepacia complex (Bcc) are known as intrinsically multidrug-resistant opportunistic pathogens with low permeability cell envelopes. Here, we re-examined a previously performed chemical-genetic screen of barcoded transposon mutants in B. cenocepacia K56-2, focusing on cell envelope structural and functional processes. We identified structures mechanistically important for resistance to singular and multiple antibiotic classes. For example, susceptibility to novobiocin, avibactam, and the LpxC inhibitor, PF-04753299, was linked to the BpeAB-OprB efflux pump, suggesting these drugs are substrates for this pump in B. cenocepacia. Defects in peptidoglycan precursor synthesis specifically increased susceptibility to cycloserine and revealed a new putative amino acid racemase, while defects in divisome accessory proteins increased susceptibility to multiple ß-lactams. Additionally, disruption of the periplasmic disulfide bond formation system caused pleiotropic defects on outer membrane integrity and ß-lactamase activity. Our findings highlight the layering of resistance mechanisms in the structure and function of the cell envelope. Consequently, we point out processes that can be targeted for developing antibiotic potentiators.IMPORTANCEThe Gram-negative cell envelope is a double-layered physical barrier that protects cells from extracellular stressors, such as antibiotics. The Burkholderia cell envelope is known to contain additional modifications that reduce permeability. We investigated Burkholderia cell envelope factors contributing to antibiotic resistance from a genome-wide view by re-examining data from a transposon mutant library exposed to an antibiotic panel. We identified susceptible phenotypes for defects in structures and functions in the outer membrane, periplasm, and cytoplasm. Overall, we show that resistance linked to the cell envelope is multifaceted and provides new targets for the development of antibiotic potentiators.

A CRISPR-Cas-associated transposon system for genome editing in Burkholderia cepacia complex species.

Genome editing in non-model bacteria is important to understand gene-to-function links that may differ from those of model microorganisms. Although species of the Burkholderia cepacia complex (Bcc) have great biotechnological capacities, the limited genetic tools available to understand and mitigate their pathogenic potential hamper their utilization in industrial applications. To broaden the genetic tools available for Bcc species, we developed RhaCAST, a targeted DNA insertion platform based on a CRISPR-associated transposase driven by a rhamnose-inducible promoter. We demonstrated the utility of the system for targeted insertional mutagenesis in the Bcc strains B. cenocepacia K56-2 and Burkholderia multivorans ATCC17616. We showed that the RhaCAST system can be used for loss- and gain-of-function applications. Importantly, the selection marker could be excised and reused to allow iterative genetic manipulation. The RhaCAST system is faster, easier, and more adaptable than previous insertional mutagenesis tools available for Bcc species and may be used to disrupt pathogenicity elements and insert relevant genetic modules, enabling Bcc biotechnological applications. IMPORTANCE: Species of the Burkholderia cepacia complex (Bcc) have great biotechnological potential but are also opportunistic pathogens. Genetic manipulation of Bcc species is necessary to understand gene-to-function links. However, limited genetic tools are available to manipulate Bcc, hindering our understanding of their pathogenic traits and their potential in biotechnological applications. We developed a genetic tool based on CRISPR-associated transposase to increase the genetic tools available for Bcc species. The genetic tool we developed in this study can be used for loss and gain of function in Bcc species. The significance of our work is in expanding currently available tools to manipulate Bcc.

A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery.

Screening for novel antibacterial compounds in small molecule libraries has a low success rate. We applied machine learning (ML)-based virtual screening for antibacterial activity and evaluated its predictive power by experimental validation. We first binarized 29,537 compounds according to their growth inhibitory activity (hit rate 0.87%) against the antibiotic-resistant bacterium Burkholderia cenocepacia and described their molecular features with a directed-message passing neural network (D-MPNN). Then, we used the data to train an ML model that achieved a receiver operating characteristic (ROC) score of 0.823 on the test set. Finally, we predicted antibacterial activity in virtual libraries corresponding to 1,614 compounds from the Food and Drug Administration (FDA)-approved list and 224,205 natural products. Hit rates of 26% and 12%, respectively, were obtained when we tested the top-ranked predicted compounds for growth inhibitory activity against B. cenocepacia, which represents at least a 14-fold increase from the previous hit rate. In addition, more than 51% of the predicted antibacterial natural compounds inhibited ESKAPE pathogens showing that predictions expand beyond the organism-specific dataset to a broad range of bacteria. Overall, the developed ML approach can be used for compound prioritization before screening, increasing the typical hit rate of drug discovery.

Deep clustering of small molecules at large-scale via variational autoencoder embedding and K-means.

BACKGROUND: Converting molecules into computer-interpretable features with rich molecular information is a core problem of data-driven machine learning applications in chemical and drug-related tasks. Generally speaking, there are global and local features to represent a given molecule. As most algorithms have been developed based on one type of feature, a remaining bottleneck is to combine both feature sets for advanced molecule-based machine learning analysis. Here, we explored a novel analytical framework to make embeddings of the molecular features and apply them in the clustering of a large number of small molecules. RESULTS: In this novel framework, we first introduced a principal component analysis method encoding the molecule-specific atom and bond information. We then used a variational autoencoder (AE)-based method to make embeddings of the global chemical properties and the local atom and bond features. Next, using the embeddings from the encoded local and global features, we implemented and compared several unsupervised clustering algorithms to group the molecule-specific embeddings. The number of clusters was treated as a hyper-parameter and determined by the Silhouette method. Finally, we evaluated the corresponding results using three internal indices. Applying the analysis framework to a large chemical library of more than 47,000 molecules, we successfully identified 50 molecular clusters using the K-means method with 32 embeddings based on the AE method. We visualized the clustering result via t-SNE for the overall distribution of molecules and the similarity maps for the structural analysis of randomly selected cluster-specific molecules. CONCLUSIONS: This study developed a novel analytical framework that comprises a feature engineering scheme for molecule-specific atomic and bonding features and a deep learning-based embedding strategy for different molecular features. By applying the identified embeddings, we show their usefulness for clustering a large molecule dataset. Our novel analytic algorithms can be applied to any virtual library of chemical compounds with diverse molecular structures. Hence, these tools have the potential of optimizing drug discovery, as they can decrease the number of compounds to be screened in any drug screening campaign.

Improved Dynamic Range of a Rhamnose-Inducible Promoter for Gene Expression in Burkholderia spp.

A diverse genetic toolkit is critical for understanding bacterial physiology and genotype-phenotype relationships. Inducible promoter systems are an integral part of this toolkit. In Burkholderia and related species, the l-rhamnose-inducible promoter is among the first choices due to its tight control and the lack of viable alternatives. To improve upon its maximum activity and dynamic range, we explored the effect of promoter system modifications in Burkholderia cenocepacia with a LacZ-based reporter. By combining the bacteriophage T7 gene 10 stem-loop and engineered rhaI transcription factor-binding sites, we obtained a rhamnose-inducible system with a 6.5-fold and 3.0-fold increases in maximum activity and dynamic range, respectively, compared to the native promoter. We then added the modified promoter system to pSCrhaB2 and pSC201, common genetic tools used for plasmid-based and chromosome-based gene expression, respectively, in Burkholderia, creating pSCrhaB2plus and pSC201plus. We demonstrated the utility of pSCrhaB2plus for gene expression in B. thailandensis, B. multivorans, and B. vietnamiensis and used pSC201plus to control highly expressed essential genes from the chromosome of B. cenocepacia. The utility of the modified system was demonstrated as we recovered viable mutants to control ftsZ, rpoBC, and rpsF, whereas the unmodified promoter was unable to control rpsF. The modified expression system allowed control of an essential gene depletion phenotype at lower levels of l-rhamnose, the inducer. pSCRhaB2plus and pSC201plus are expected to be valuable additions to the genetic toolkit for Burkholderia and related species. IMPORTANCE Species of Burkholderia are dually recognized as being of attractive biotechnological potential but also opportunistic pathogens for immunocompromised individuals. Understanding the genotype-phenotype relationship is critical for synthetic biology approaches in Burkholderia to disentangle pathogenic from beneficial traits. A diverse genetic toolkit, including inducible promoters, is the foundation for these investigations. Thus, we sought to improve on the commonly used rhamnose-inducible promoter system. Our modifications resulted in both higher levels of heterologous protein expression and broader control over highly expressed essential genes in B. cenocepacia. The significance of our work is in expanding the genetic toolkit to enable more comprehensive studies into Burkholderia and related bacteria.

Phenylacetyl Coenzyme A, Not Phenylacetic Acid, Attenuates CepIR-Regulated Virulence in Burkholderia cenocepacia.

During phenylalanine catabolism, phenylacetic acid (PAA) is converted to phenylacetyl coenzyme A (PAA-CoA) by a ligase, PaaK, and then PAA-CoA is epoxidized by a multicomponent monooxygenase, PaaABCDE, before further degradation through the tricarboxylic acid (TCA) cycle. In the opportunistic pathogen Burkholderia cenocepacia, loss of paaABCDE attenuates virulence factor expression, which is under the control of the LuxIR-like quorum sensing (QS) system, CepIR. To further investigate the link between CepIR-regulated virulence and PAA catabolism, we created knockout mutants of the first step of the pathway (PAA-CoA synthesis by PaaK) and characterized them in comparison to a paaABCDE mutant using liquid chromatography-tandem mass spectrometry (LC-MS/MS) and virulence assays. We found that while loss of PaaABCDE decreased virulence, deletion of the paaK genes resulted in a more virulent phenotype than that of the wild-type strain. Deletion of either paaK or paaABCDE led to higher levels of released PAA but no differences in levels of internal accumulation compared to the wild-type level. While we found no evidence of direct cepIR downregulation by PAA-CoA or PAA, a low-virulence cepR mutant reverted to a virulent phenotype upon removal of the paaK genes. On the other hand, removal of paaABCDE in the cepR mutant did not impact its attenuated phenotype. Together, our results suggest an indirect role for PAA-CoA in suppressing B. cenocepacia CepIR-activated virulence.IMPORTANCE The opportunistic pathogen Burkholderia cenocepacia uses a chemical signal process called quorum sensing (QS) to produce virulence factors. In B. cenocepacia, QS relies on the presence of the transcriptional regulator CepR which, upon binding QS signal molecules, activates virulence. In this work, we found that even in the absence of CepR, B. cenocepacia can elicit a pathogenic response if phenylacetyl-CoA, an intermediate of the phenylacetic acid degradation pathway, is not produced. Instead, accumulation of phenylacetyl-CoA appears to attenuate pathogenicity. Therefore, we have discovered that it is possible to trigger virulence in the absence of CepR, challenging the classical view of activation of virulence by this QS mechanism. Our work provides new insight into the relationship between metabolism and virulence in opportunistic bacteria. We propose that in the event that QS signaling molecules cannot accumulate to trigger a pathogenic response, a metabolic signal can still activate virulence in B. cenocepacia.

Competitive Fitness of Essential Gene Knockdowns Reveals a Broad-Spectrum Antibacterial Inhibitor of the Cell Division Protein FtsZ.

To streamline the elucidation of antibacterial compounds' mechanism of action, comprehensive high-throughput assays interrogating multiple putative targets are necessary. However, current chemogenomic approaches for antibiotic target identification have not fully utilized the multiplexing potential of next-generation sequencing. Here, we used Illumina sequencing of transposon insertions to track the competitive fitness of a Burkholderia cenocepacia library containing essential gene knockdowns. Using this method, we characterized a novel benzothiadiazole derivative, 10126109 (C109), with antibacterial activity against B. cenocepacia, for which whole-genome sequencing of low-frequency spontaneous drug-resistant mutants had failed to identify the drug target. By combining the identification of hypersusceptible mutants and morphology screening, we show that C109 targets cell division. Furthermore, fluorescence microscopy of bacteria harboring green fluorescent protein (GFP) cell division protein fusions revealed that C109 prevents divisome formation by altering the localization of the essential cell division protein FtsZ. In agreement with this, C109 inhibited both the GTPase and polymerization activities of purified B. cenocepacia FtsZ. C109 displayed antibacterial activity against Gram-positive and Gram-negative cystic fibrosis pathogens, including Mycobacterium abscessus C109 effectively cleared B. cenocepacia infection in the Caenorhabditis elegans model and exhibited additive interactions with clinically relevant antibiotics. Hence, C109 is an enticing candidate for further drug development.

Essential Two-Component Systems Regulating Cell Envelope Functions: Opportunities for Novel Antibiotic Therapies.

Novel therapies are urgently needed to alleviate the current crisis of multiple drug-resistant infections. The bacterial signal transduction mechanisms, known as two-component systems (TCSs), are ideal targets of novel inhibitory molecules. Highly restricted to the bacterial world, TCSs control a diverse set of cellular functions, namely virulence, response to cell envelope stress, and drug efflux. Impaired regulation of any of these aspects could affect the susceptibility of bacterial pathogens to antibiotics, which highlights the potential of TCS as targets of antibiotic adjuvant therapies. Moreover, new high-density transposon mutagenesis methods have revealed the existence of TCSs required for growth and viability. Experimental validation of gene essentiality and phenotypic characterization of knockdown mutants indicate that essential TCSs regulate aspects of the cell envelope homeostasis in coordination with cell division. In this review, we describe essential TCSs, and their potentials as targets for antibacterial molecules. We also discuss methods for the identification of small molecules that inhibit TCSs and possible reasons why antibacterial molecules targeting essential TCSs have not yet reached clinical trials.

Competitive Growth Enhances Conditional Growth Mutant Sensitivity to Antibiotics and Exposes a Two-Component System as an Emerging Antibacterial Target in Burkholderia cenocepacia.

Chemogenetic approaches to profile an antibiotic mode of action are based on detecting differential sensitivities of engineered bacterial strains in which the antibacterial target (usually encoded by an essential gene) or an associated process is regulated. We previously developed an essential-gene knockdown mutant library in the multidrug-resistant Burkholderia cenocepacia by transposon delivery of a rhamnose-inducible promoter. In this work, we used Illumina sequencing of multiplex-PCR-amplified transposon junctions to track individual mutants during pooled growth in the presence of antibiotics. We found that competition from nontarget mutants magnified the hypersensitivity of a clone underexpressing gyrB to novobiocin by 8-fold compared with hypersensitivity measured during clonal growth. Additional profiling of various antibiotics against a pilot library representing most categories of essential genes revealed a two-component system with unknown function, which, upon depletion of the response regulator, sensitized B. cenocepacia to novobiocin, ciprofloxacin, tetracycline, chloramphenicol, kanamycin, meropenem, and carbonyl cyanide 3-chlorophenylhydrazone, but not to colistin, hydrogen peroxide, and dimethyl sulfoxide. We named the gene cluster esaSR for enhanced sensitivity to antibiotics sensor and response regulator. Mutational analysis and efflux activity assays revealed that while esaS is not essential and is involved in antibiotic-induced efflux, esaR is an essential gene and regulates efflux independently of antibiotic-mediated induction. Furthermore, microscopic analysis of cells stained with propidium iodide provided evidence that depletion of EsaR has a profound effect on the integrity of cell membranes. In summary, we unraveled a previously uncharacterized two-component system that can be targeted to reduce antibiotic resistance in B. cenocepacia.

Evaluation of the electron transfer flavoprotein as an antibacterial target in Burkholderia cenocepacia.

There are hundreds of essential genes in multidrug-resistant bacterial genomes, but only a few of their products are exploited as antibacterial targets. An example is the electron transfer flavoprotein (ETF), which is required for growth and viability in Burkholderia cenocepacia. Here, we evaluated ETF as an antibiotic target for Burkholderia cepacia complex (Bcc). Depletion of the bacterial ETF during infection of Caenorhabditis elegans significantly extended survival of the nematodes, proving that ETF is essential for survival of B. cenocepacia in this host model. In spite of the arrest in respiration in ETF mutants, the inhibition of etf expression did not increase the formation of persister cells, when treated with high doses of ciprofloxacin or meropenem. To test if etf translation could be inhibited by RNA interference, antisense oligonucleotides that target the etfBA operon were synthesized. One antisense oligonucleotide was effective in inhibiting etfB translation in vitro but not in vivo, highlighting the challenge of reduced membrane permeability for the design of drugs against B. cenocepacia. This work contributes to the validation of ETF of B. cenocepacia as a target for antibacterial therapy and demonstrates the utility of a C. elegans liquid killing assay to validate gene essentiality in an in vivo infection model.

Synthetic cystic fibrosis sputum medium diminishes Burkholderia cenocepacia antifungal activity against Aspergillus fumigatus independently of phenylacetic acid production.

Phenylacetic acid (PAA), an intermediate of phenylalanine degradation, is emerging as a signal molecule in microbial interactions with the host. In this work, we explore the presence of phenylalanine and PAA catabolism in 3 microbial pathogens of the cystic fibrosis (CF) lung microbiome: Pseudomonas aeruginosa, Burkholderia cenocepacia, and Aspergillus fumigatus. While in silico analysis of B. cenocepacia J2315 and A. fumigatus Af293 genome sequences showed complete pathways from phenylalanine to PAA, the P. aeruginosa PAO1 genome lacked several coding genes for phenylalanine and PAA catabolic enzymes. High-performance liquid chromatography analysis of supernatants from B. cenocepacia K56-2 detected PAA when grown in Luria-Bertani medium but not in synthetic cystic fibrosis sputum medium (SCFM). However, we were unable to identify PAA production by A. fumigatus or P. aeruginosa in any of the conditions tested. The inhibitory effect of B. cenocepacia on A. fumigatus growth was evaluated using agar plate interaction assays. Inhibition of fungal growth by B. cenocepacia was lessened in SCFM but this effect was not dependent on bacterial production of PAA. In summary, while we demonstrated PAA production by B. cenocepacia, we were not able to link this metabolite with the B. cenocepacia - A. fumigatus microbial interaction in CF nutritional conditions.

The attenuated virulence of a Burkholderia cenocepacia paaABCDE mutant is due to inhibition of quorum sensing by release of phenylacetic acid.

The phenylacetic acid degradation pathway of Burkholderia cenocepacia is active during cystic fibrosis-like conditions and is necessary for full pathogenicity of B. cenocepacia in nematode and rat infection models; however, the reasons for such requirements are unknown. Here, we show that the attenuated virulence of a phenylacetic acid catabolism mutant is due to quorum sensing inhibition. Unlike wild-type B. cenocepacia, a deletion mutant of the phenylacetyl-CoA monooxygenase complex (ΔpaaABCDE) released phenylacetic acid in the medium that favours infection in Caenorhabditis elegans. Addition of phenylacetic acid further decreased the pathogenicity of the ΔpaaABCDE, which cannot metabolize phenylacetic acid, but did not affect the wild-type, due to phenylacetic acid consumption. In line with reduced detection of acyl-homoserine lactones in spent medium, the ΔpaaABCDE exhibited transcriptional inhibition of the quorum sensing system cepIR. Phenotypes repressed in ΔpaaABCDE, protease activity and pathogenicity against C. elegans, increased with exogenous N-octanoyl-L-homoserine lactone. Thus, we demonstrate that the attenuated phenotype of B. cenocepacia ΔpaaABCDE is due to quorum sensing inhibition by release of phenylacetic acid, affecting N-octanoyl-L-homoserine lactone signalling. Further, we propose that active degradation of phenylacetic acid by B. cenocepacia during growth in cystic fibrosis-like conditions prevents accumulation of a quorum sensing inhibiting compound.

Chemical inhibition of bacterial ribosome biogenesis shows efficacy in a worm infection model.

The development of antibacterial compounds that perturb novel processes is an imperative in the challenge presented by widespread antibiotic resistance. While many antibiotics target the ribosome, molecules that inhibit ribosome assembly have yet to be used in this manner. Here we show that a novel inhibitor of ribosome biogenesis, lamotrigine, is capable of rescuing Caenorhabditis elegans from an established Salmonella infection, revealing that ribosome biogenesis is a promising target for the development of new antibiotics.

An electron transfer flavoprotein is essential for viability and its depletion causes a rod-to-sphere change in Burkholderia cenocepacia.

Essential gene studies often reveal novel essential functions for genes with dispensable homologues in other species. This is the case with the widespread family of electron transfer flavoproteins (ETFs), which are required for the metabolism of specific substrates or for symbiotic nitrogen fixation in some bacteria. Despite these non-essential functions high-throughput screens have identified ETFs as putatively essential in several species. In this study, we constructed a conditional expression mutant of one of the ETFs in Burkholderia cenocepacia, and demonstrated that its expression is essential for growth on both complex media and a variety of single-carbon sources. We further demonstrated that the two subunits EtfA and EtfB interact with each other, and that cells depleted of ETF are non-viable and lack redox potential. These cells also transition from the short rods characteristic of Burkholderia cenocepacia to small spheres independently of MreB. The putative membrane partner ETF dehydrogenase also induced the same rod-to-sphere change. We propose that the ETF of Burkholderia cenocepacia is a novel antibacterial target.

Genomic tools to profile antibiotic mode of action.

The increasing emergence of antimicrobial multiresistant bacteria is of great concern to public health. While these bacteria are becoming an ever more prominent cause of nosocomial and community-acquired infections worldwide, the antibiotic discovery pipeline has been stalled in the last few years with very few efforts in the research and development of novel antibacterial therapies. Some of the root causes that have hampered current antibiotic drug development are the lack of understanding of the mode of action (MOA) of novel antibiotic molecules and the poor characterization of the bacterial physiological response to antibiotics that ultimately causes resistance. Here, we review how bacterial genetic tools can be applied at the genomic level with the goal of profiling resistance to antibiotics and elucidating antibiotic MOAs. Specifically, we highlight how chemical genomic detection of the MOA of novel antibiotic molecules and antibiotic profiling by next-generation sequencing are leveraging basic antibiotic research to unprecedented levels with great opportunities for knowledge translation.

Rationally Designed Pooled CRISPRi-Seq Uncovers an Inhibitor of Bacterial Peptidyl-tRNA Hydrolase.

Pooled knockdown libraries of essential genes are useful tools for elucidating the mechanisms of action of antibacterial compounds, a pivotal step in antibiotic discovery. However, achieving genomic coverage of antibacterial targets poses a challenge due to the uneven proliferation of knockdown mutants during pooled growth, leading to the unintended loss of important targets. To overcome this issue, we describe the construction of CIMPLE ( C RISPR i - m ediated p ooled library of e ssential genes), a rationally designed pooled knockdown library built in a model antibiotic-resistant bacteria, Burkholderia cenocepacia. By analyzing growth parameters of clonal knockdown populations of an arrayed CRISPRi library, we predicted strain depletion levels during pooled growth and adjusted mutant relative abundance, approaching genomic coverage of antibacterial targets during antibiotic exposure. We first benchmarked CIMPLE by chemical-genetic profiling of known antibacterials, then applied it to an uncharacterized bacterial growth inhibitor from a new class. CRISPRi-Seq with CIMPLE, followed by biochemical validation, revealed that the novel compound targets the peptidyl-tRNA hydrolase (Pth). Overall, CIMPLE leverages the advantages of arrayed and pooled CRISPRi libraries to uncover unexplored targets for antibiotic action. Summary: Bacterial mutant libraries in which antibiotic targets are downregulated are useful tools to functionally characterize novel antimicrobials. These libraries are used for chemical-genetic profiling as target-compound interactions can be inferred by differential fitness of mutants during pooled growth. Mutants that are functionally related to the antimicrobial mode of action are usually depleted from the pool upon exposure to the drug. Although powerful, this method can fail when the unequal proliferation of mutant strains before exposure causes mutants to fall below the detection level in the library pool. To address this issue, we constructed an arrayed essential gene mutant library (EGML) in the antibiotic-resistant bacterium Burkholderia cenocepacia using CRISPR interference (CRISPRi) and analyzed the growth parameters of individual mutant strains. We then modelled depletion levels during pooled growth and used the model to rationally design an optimized CRISPR interference-mediated pooled library of essential genes (CIMPLE). By adjusting the initial inoculum of the knockdown mutants, we achieved coverage of the bacterial essential genome with mutant sensitization. We exposed CIMPLE to a recently discovered antimicrobial of a novel class and discovered it inhibits the peptidyl-tRNA hydrolase, an essential bacterial enzyme. In summary, we demonstrate the utility of CIMPLE and CRISPRi-Seq to uncover the mechanism of action of novel antimicrobial compounds.

The mechanism of action of auranofin analogs in B. cenocepacia revealed by chemogenomic profiling.

Drug repurposing efforts led to the discovery of bactericidal activity in auranofin, a gold-containing drug used to treat rheumatoid arthritis. Auranofin kills Gram-positive bacteria by inhibiting thioredoxin reductase, an enzyme that scavenges reactive oxygen species (ROS). Despite the presence of thioredoxin reductase in Gram-negative bacteria, auranofin is not always active against them. It is not clear whether the lack of activity in several Gram-negative bacteria is due to the cell envelope barrier or the presence of other ROS protective enzymes such as glutathione reductase (GOR). We previously demonstrated that chemical analogs of auranofin (MS-40 and MS-40S), but not auranofin, are bactericidal against the Gram-negative Burkholderia cepacia complex. Here, we explore the targets of auranofin, MS-40, and MS-40S in Burkholderia cenocepacia and elucidate the mechanism of action of the auranofin analogs by a genome-wide, randomly barcoded transposon screen (BarSeq). Auranofin and its analogs inhibited the B. cenocepacia thioredoxin reductase and induced ROS but did not inhibit the bacterial GOR. Genome-wide, BarSeq analysis of cells exposed to MS-40 and MS-40S compared to the ROS inducers arsenic trioxide, diamide, hydrogen peroxide, and paraquat revealed common and unique mediators of drug susceptibility. Furthermore, deletions of gshA and gshB that encode enzymes in the glutathione biosynthetic pathway led to increased susceptibility to MS-40 and MS-40S. Overall, our data suggest that the auranofin analogs kill B. cenocepacia by inducing ROS through inhibition of thioredoxin reductase and that the glutathione system has a role in protecting B. cenocepacia against these ROS-inducing compounds.IMPORTANCEThe Burkholderia cepacia complex is a group of multidrug-resistant bacteria that can cause infections in the lungs of people with the autosomal recessive disease, cystic fibrosis. Specifically, the bacterium Burkholderia cenocepacia can cause severe infections, reducing lung function and leading to a devastating type of sepsis, cepacia syndrome. This bacterium currently does not have an accepted antibiotic treatment plan because of the wide range of antibiotic resistance. Here, we further the research on auranofin analogs as antimicrobials by finding the mechanism of action of these potent bactericidal compounds, using a powerful technique called BarSeq, to find the global response of the cell when exposed to an antimicrobial.

ABT-MPNN: an atom-bond transformer-based message-passing neural network for molecular property prediction.

Graph convolutional neural networks (GCNs) have been repeatedly shown to have robust capacities for modeling graph data such as small molecules. Message-passing neural networks (MPNNs), a group of GCN variants that can learn and aggregate local information of molecules through iterative message-passing iterations, have exhibited advancements in molecular modeling and property prediction. Moreover, given the merits of Transformers in multiple artificial intelligence domains, it is desirable to combine the self-attention mechanism with MPNNs for better molecular representation. We propose an atom-bond transformer-based message-passing neural network (ABT-MPNN), to improve the molecular representation embedding process for molecular property predictions. By designing corresponding attention mechanisms in the message-passing and readout phases of the MPNN, our method provides a novel architecture that integrates molecular representations at the bond, atom and molecule levels in an end-to-end way. The experimental results across nine datasets show that the proposed ABT-MPNN outperforms or is comparable to the state-of-the-art baseline models in quantitative structure-property relationship tasks. We provide case examples of Mycobacterium tuberculosis growth inhibitors and demonstrate that our model's visualization modality of attention at the atomic level could be an insightful way to investigate molecular atoms or functional groups associated with desired biological properties. The new model provides an innovative way to investigate the effect of self-attention on chemical substructures and functional groups in molecular representation learning, which increases the interpretability of the traditional MPNN and can serve as a valuable way to investigate the mechanism of action of drugs.

Profiling cell envelope-antibiotic interactions reveals vulnerabilities to ß-lactams in a multidrug-resistant bacterium.

The cell envelope of Gram-negative bacteria belonging to the Burkholderia cepacia complex (Bcc) presents unique restrictions to antibiotic penetration. As a consequence, Bcc species are notorious for causing recalcitrant multidrug-resistant infections in immunocompromised individuals. Here, we present the results of a genome-wide screen for cell envelope-associated resistance and susceptibility determinants in a Burkholderia cenocepacia clinical isolate. For this purpose, we construct a high-density, randomly-barcoded transposon mutant library and expose it to 19 cell envelope-targeting antibiotics. By quantifying relative mutant fitness with BarSeq, followed by validation with CRISPR-interference, we profile over a hundred functional associations and identify mediators of antibiotic susceptibility in the Bcc cell envelope. We reveal connections between ß-lactam susceptibility, peptidoglycan synthesis, and blockages in undecaprenyl phosphate metabolism. The synergy of the ß-lactam/ß-lactamase inhibitor combination ceftazidime/avibactam is primarily mediated by inhibition of the PenB carbapenemase. In comparison with ceftazidime, avibactam more strongly potentiates the activity of aztreonam and meropenem in a panel of Bcc clinical isolates. Finally, we characterize in Bcc the iron and receptor-dependent activity of the siderophore-cephalosporin antibiotic, cefiderocol. Our work has implications for antibiotic target prioritization, and for using additional combinations of ß-lactam/ß-lactamase inhibitors that can extend the utility of current antibacterial therapies.

The Burkholderia contaminans prevalent phenotypes as possible markers of poor clinical outcomes in chronic lung infection of children with cystic fibrosis.

Burkholderia contaminans, a species of the Burkholderia cepacia complex-prevalent in certain Latin-American and European countries-can cause chronic pulmonary infection in persons with cystic fibrosis. Our aim was to gain insights into long-term lung infections with a focus on correlating how bacterial phenotypic traits in the chronic infection impact on patients' clinical outcome. Genotypic characteristics of 85 B. contaminans isolates recovered from 70 patients were investigated. For 16 of those patients, the clinical status and bacterial phenotypic characteristics, e.g. several virulence factors, phenotypic variants, and the antimicrobial susceptibility pattern, were evaluated. Two clones were found in the whole bacterial population: (i) the multiresistant ST 872 PCR-recA-RFLP-HaeIII-K-pattern clone, which carries a pathogenic island homologous to BcenGI11 of B. cenocepacia J2315, and (ii) the ST 102 PCR-recA-RFLP-HaeIII-AT-pattern clone. The emergence of certain bacterial phenotypes in the chronic infection such as the nonmucoid phenotype, small colony variants, brownish pigmented colonies, and hypermutators, proved to be, together with coinfection with Pseudomonas aeruginosa, the possible markers of more challenging infections and poor prognosis. The presence of cocolonizers and the bacterial phenotypes that are especially adapted to persist in long-term respiratory tract infections have a crucial role in patients' clinical outcomes.