Understanding Beta-Lactam-Induced Lysis at the Single-Cell Level

Semi-mechanistic modeling of resistance development to β-lactam and β-lactamase-inhibitor combinations

The use of β-lactam (BL) and β-lactamase inhibitor (BLI) combinations, such as piperacillin-tazobactam (PIP-TAZ) is an effective strategy to combat infections by extended-spectrum β-lactamase-producing bacteria. However, in Gram-negative bacteria, resistance (both mutational and adaptive) to BL-BLI combination can still develop through multiple mechanisms. These mechanisms may include increased β-lactamase activity, reduced drug influx, and increased drug efflux. Understanding the relative contribution of these mechanisms during resistance development helps identify the most impactful mechanism to target in designing a treatment to counter BL-BLI resistance. This study used semi-mechanistic mathematical modeling in combination with antibiotic sensitivity assays to assess the potential impact of different resistance mechanisms during the development of PIP-TAZ resistance in a Klebsiella pneumoniae isolate expressing CTX-M-15 and SHV-1 β-lactamases. The mathematical models were used to evaluate the potential impact of several cellular changes as a sole mediator of PIP-TAZ resistance. Our semi-mechanistic model identified 2 out of the 13 inspected mechanisms as key resistance mechanisms that may independently support the observed magnitude of PIP-TAZ resistance, namely porin loss and efflux pump up-regulation. Simulation using the resulting models also suggested the possible adjustment of PIP-TAZ dose outside its commonly used 8:1 dosing ratio. The current study demonstrated how theory-based mechanistic models informed by experimental data can be used to support hypothesis generation regarding potential resistance mechanisms, which may guide subsequent experimental studies.

Discovery of antibiotics that selectively kill metabolically dormant bacteria

There is a need to discover and develop non-toxic antibiotics that are effective against metabolically dormant bacteria, which underlie chronic infections and promote antibiotic resistance. Traditional antibiotic discovery has historically favored compounds effective against actively metabolizing cells, a property that is not predictive of efficacy in metabolically inactive contexts. Here, we combine a stationary-phase screening method with deep learning-powered virtual screens and toxicity filtering to discover compounds with lethality against metabolically dormant bacteria and favorable toxicity profiles. The most potent and structurally distinct compound without any obvious mechanistic liability was semapimod, an anti-inflammatory drug effective against stationary-phase E. coli and A. baumannii. Integrating microbiological assays, biochemical measurements, and single-cell microscopy, we show that semapimod selectively disrupts and permeabilizes the bacterial outer membrane by binding lipopolysaccharide. This work illustrates the value of harnessing non-traditional screening methods and deep learning models to identify non-toxic antibacterial compounds that are effective in infection-relevant contexts.

Discovery of a structural class of antibiotics with explainable deep learning

The discovery of novel structural classes of antibiotics is urgently needed to address the ongoing antibiotic resistance crisis1-9. Deep learning approaches have aided in exploring chemical spaces1,10-15; these typically use black box models and do not provide chemical insights. Here we reasoned that the chemical substructures associated with antibiotic activity learned by neural network models can be identified and used to predict structural classes of antibiotics. We tested this hypothesis by developing an explainable, substructure-based approach for the efficient, deep learning-guided exploration of chemical spaces. We determined the antibiotic activities and human cell cytotoxicity profiles of 39,312 compounds and applied ensembles of graph neural networks to predict antibiotic activity and cytotoxicity for 12,076,365 compounds. Using explainable graph algorithms, we identified substructure-based rationales for compounds with high predicted antibiotic activity and low predicted cytotoxicity. We empirically tested 283 compounds and found that compounds exhibiting antibiotic activity against Staphylococcus aureus were enriched in putative structural classes arising from rationales. Of these structural classes of compounds, one is selective against methicillin-resistant S. aureus (MRSA) and vancomycin-resistant enterococci, evades substantial resistance, and reduces bacterial titres in mouse models of MRSA skin and systemic thigh infection. Our approach enables the deep learning-guided discovery of structural classes of antibiotics and demonstrates that machine learning models in drug discovery can be explainable, providing insights into the chemical substructures that underlie selective antibiotic activity.

Population dynamics of cross-protection against β-lactam antibiotics in droplet microreactors

Introduction

Bacterial strains that are resistant to antibiotics may protect not only themselves, but also sensitive bacteria nearby if resistance involves antibiotic degradation. Such cross-protection poses a challenge to effective antibiotic therapy by enhancing the long-term survival of bacterial infections, however, the current understanding is limited.

Methods

In this study, we utilize an automated nanoliter droplet analyzer to study the interactions between Escherichia coli strains expressing a β-lactamase (resistant) and those not expressing it (sensitive) when exposed to the β-lactam antibiotic cefotaxime (CTX), with the aim to define criteria contributing to cross-protection.

Results

We observed a cross-protection window of CTX concentrations for the sensitive strain, extending up to approximately 100 times its minimal inhibitory concentration (MIC). Through both microscopy and enzyme activity analyses, we demonstrate that bacterial filaments, triggered by antibiotic stress, contribute to cross-protection.

Discussion

The antibiotic concentration window for cross-protection depends on the difference in β-lactamase activity between co-cultured strains: larger differences shift the 'cross-protection window' toward higher CTX concentrations. Our findings highlight the dependence of opportunities for cross-protection on the relative resistance levels of the strains involved and suggest a possible specific role for filamentation.

Fluoroquinolone and beta-lactam antimicrobials induce different transcriptome profiles in Salmonella enterica persister cells

Here, we investigate the transcriptome profiles of two S. Enteritidis and one S. Schwarzengrund isolates that present different persister levels when exposed to ciprofloxacin or ceftazidime. It was possible to note a distinct transcript profile among isolates, time of exposure, and treatment. We could not find a commonly expressed transcript profile that plays a role in persister formation after S. enterica exposure to beta-lactam or fluoroquinolone, as only three DEGs presented the same behavior under the conditions and isolates tested. It appears that the formation of persisters in S. enterica after exposure to ciprofloxacin is linked to the overexpression of genes involved in the SOS response (recA), cell division inhibitor (sulA), iron-sulfur metabolism (hscA and iscS), and type I TA system (tisB). On the other hand, most genes differentially expressed in S. enterica after exposure to ceftazidime appeared to be downregulated and were part of the flagellar assembly apparatus, citrate cycle (TCA cycle), glycolysis/gluconeogenesis, carbon metabolism, bacterial secretion system, quorum sensing, pyruvate metabolism pathway, and biosynthesis of secondary metabolites. The different transcriptome profiles found in S. enterica persisters induced by ciprofloxacin and ceftazidime suggest that these cells modulate their response differently according to each stress.

Cationic Polymers Enable Internalization of Negatively Charged Chemical Probes into Bacteria

The bacterial cell envelope provides a protective barrier that is challenging for small molecules and biomolecules to cross. Given the anionic nature of both Gram-positive and Gram-negative bacterial cell envelopes, negatively charged molecules are particularly difficult to deliver into these organisms. Many strategies have been employed to penetrate bacteria, ranging from reagents such as cell-penetrating peptides, enzymes, and metal-chelating compounds to physical perturbations. While cationic polymers are known antimicrobial agents, polymers that promote the permeabilization of bacterial cells without causing high levels of toxicity and cell lysis have not yet been described. Here, we investigate four polymers that display a cationic poly(2-(dimethylamino)ethyl methacrylate (D) block for the internalization of an anionic adenosine triphosphate (ATP)-based chemical probe into Escherichia coli and Bacillus subtilis. We evaluated two polymer architectures, linear and micellar, to determine how shape and hydrophobicity affect internalization efficiency. We found that, in addition to these reagents successfully promoting probe internalization, the probe-labeled cells were able to continue to grow and divide. The micellar structures in particular were highly effective for the delivery of the negatively charged chemical probe. Finally, we demonstrated that these cationic polymers could act as general permeabilization reagents, promoting the entry of other molecules, such as antibiotics.

Whole-cell modeling of E. coli colonies enables quantification of single-cell heterogeneity in antibiotic responses

Antibiotic resistance poses mounting risks to human health, as current antibiotics are losing efficacy against increasingly resistant pathogenic bacteria. Of particular concern is the emergence of multidrug-resistant strains, which has been rapid among Gram-negative bacteria such as Escherichia coli. A large body of work has established that antibiotic resistance mechanisms depend on phenotypic heterogeneity, which may be mediated by stochastic expression of antibiotic resistance genes. The link between such molecular-level expression and the population levels that result is complex and multi-scale. Therefore, to better understand antibiotic resistance, what is needed are new mechanistic models that reflect single-cell phenotypic dynamics together with population-level heterogeneity, as an integrated whole. In this work, we sought to bridge single-cell and population-scale modeling by building upon our previous experience in "whole-cell" modeling, an approach which integrates mathematical and mechanistic descriptions of biological processes to recapitulate the experimentally observed behaviors of entire cells. To extend whole-cell modeling to the "whole-colony" scale, we embedded multiple instances of a whole-cell E. coli model within a model of a dynamic spatial environment, allowing us to run large, parallelized simulations on the cloud that contained all the molecular detail of the previous whole-cell model and many interactive effects of a colony growing in a shared environment. The resulting simulations were used to explore the response of E. coli to two antibiotics with different mechanisms of action, tetracycline and ampicillin, enabling us to identify sub-generationally-expressed genes, such as the beta-lactamase ampC, which contributed greatly to dramatic cellular differences in steady-state periplasmic ampicillin and was a significant factor in determining cell survival.

The Influence of Bacteriophages on the Metabolic Condition of Human Fibroblasts in Light of the Safety of Phage Therapy in Staphylococcal Skin Infections

Phage therapy has been successfully used as an experimental therapy in the treatment of multidrug-resistant strains of Staphylococcus aureus (MDRSA)-caused skin infections and is seen as the most promising alternative to antibiotics. However, in recent years a number of reports indicating that phages can interact with eukaryotic cells emerged. Therefore, there is a need to re-evaluate phage therapy in light of safety. It is important to analyze not only the cytotoxicity of phages alone but also the impact their lytic activity against bacteria may have on human cells. As progeny virions rupture the cell wall, lipoteichoic acids are released in high quantities. It has been shown that they act as inflammatory agents and their presence could lead to the worsening of the patient's condition and influence their recovery. In our work, we have tested if the treatment of normal human fibroblasts with staphylococcal phages will influence the metabolic state of the cell and the integrity of cell membranes. We have also analyzed the effectiveness of bacteriophages in reducing the number of MDRSA attached to human fibroblasts and the influence of the lytic activity of phages on cell viability. We observed that, out of three tested anti-Staphylococcal phages-vB_SauM-A, vB_SauM-C and vB_SauM-D-high concentrations (10⁹ PFU/mL) of two, vB_SauM-A and vB_SauM-D, showed a negative impact on the viability of human fibroblasts. However, a dose of 10⁷ PFU/mL had no effect on the metabolic activity or membrane integrity of the cells. We also observed that the addition of phages alleviated the negative effect of the MDRSA infection on fibroblasts' viability, as phages were able to effectively reduce the number of bacteria in the co-culture. We believe that these results will contribute to a better understanding of the influence of phage therapy on human cells and encourage even more studies on this topic.

Molecular Imaging of Isolated Escherichia coli DH5α Peptidoglycan Sacculi Identifies the Mechanism of Action of Cell Wall-Inhibiting Antibiotics

Antibiotic resistance of pathogenic bacteria needs to be urgently addressed by the development of new antibacterial entities. Although the prokaryotic cell wall comprises a valuable target for this purpose, development of novel cell wall-active antibiotics is mostly missing today. This is mainly caused by hindrances in the assessment of isolated enzymes of the co-dependent murein synthesis machineries, e.g., the elongasome and divisome. We therefore present imaging methodologies to evaluate inhibitors of bacterial cell wall synthesis by high-resolution atomic force microscopy on isolated Escherichia coli murein sacculi. With the ability to elucidate the peptidoglycan ultrastructure of E. coli cells, unprecedented molecular insights into the mechanisms of antibiotics were established. The nanoscopic impairments introduced by ampicillin, amoxicillin, and fosfomycin were not only identified by AFM but readily correlated with their known mechanism of action. These valuable in vitro capabilities will facilitate the identification and evaluation of new antibiotic leads in the future.

Single Cell Killing Kinetics Differentiate Phenotypic Bacterial Responses to Different Antibacterial Classes

With the spread of multidrug-resistant bacteria, there has been an increasing focus on molecular classes that have not yet yielded an antibiotic. A key capability for assessing and prescribing new antibacterial treatments is to compare the effects antibacterial agents have on bacterial growth at a phenotypic, single-cell level. Here, we combined time-lapse microscopy with microfluidics to investigate the concentration-dependent killing kinetics of stationary-phase Escherichia coli cells. We used antibacterial agents from three different molecular classes, β-lactams and fluoroquinolones, with the known antibiotics ampicillin and ciprofloxacin, respectively, and a new experimental class, protein Ψ-capsids. We found that bacterial cells elongated when treated with ampicillin and ciprofloxacin used at their minimum inhibitory concentration (MIC). This was in contrast to Ψ-capsids, which arrested bacterial elongation within the first two hours of treatment. At concentrations exceeding the MIC, all the antibacterial agents tested arrested bacterial growth within the first 2 h of treatment. Further, our single-cell experiments revealed differences in the modes of action of three different agents. At the MIC, ampicillin and ciprofloxacin caused the lysis of bacterial cells, whereas at higher concentrations, the mode of action shifted toward membrane disruption. The Ψ-capsids killed cells by disrupting their membranes at all concentrations tested. Finally, at increasing concentrations, ampicillin and Ψ-capsids reduced the fraction of the population that survived treatment in a viable but nonculturable state, whereas ciprofloxacin increased this fraction. This study introduces an effective capability to differentiate the killing kinetics of antibacterial agents from different molecular classes and offers a high content analysis of antibacterial mechanisms at the single-cell level. IMPORTANCE Antibiotics act against bacterial pathogens by inhibiting their growth or killing them directly. Different modes of action determine different antibacterial responses, whereas phenotypic differences in bacteria can challenge the efficacy of antibiotics. Therefore, it is important to be able to differentiate the concentration-dependent killing kinetics of antibacterial agents at a single-cell level, in particular for molecular classes which have not yielded an antibiotic before. Here, we measured single-cell responses using microfluidics-enabled imaging, revealing that a novel class of antibacterial agents, protein Ψ-capsids, arrests bacterial elongation at the onset of treatment, whereas elongation continues for cells treated with β-lactam and fluoroquinolone antibiotics. The study advances our current understanding of antibacterial function and offers an effective strategy for the comparative design of new antibacterial therapies, as well as clinical antibiotic susceptibility testing.

Reactive metabolic byproducts contribute to antibiotic lethality under anaerobic conditions

Understanding how bactericidal antibiotics kill bacteria remains an open question. Previous work has proposed that primary drug-target corruption leads to increased energetic demands, resulting in the generation of reactive metabolic byproducts (RMBs), particularly reactive oxygen species, that contribute to antibiotic-induced cell death. Studies have challenged this hypothesis by pointing to antibiotic lethality under anaerobic conditions. Here, we show that treatment of Escherichia coli with bactericidal antibiotics under anaerobic conditions leads to changes in the intracellular concentrations of central carbon metabolites, as well as the production of RMBs, particularly reactive electrophilic species (RES). We show that antibiotic treatment results in DNA double-strand breaks and membrane damage and demonstrate that antibiotic lethality under anaerobic conditions can be decreased by RMB scavengers, which reduce RES accumulation and mitigate associated macromolecular damage. This work indicates that RMBs, generated in response to antibiotic-induced energetic demands, contribute in part to antibiotic lethality under anaerobic conditions.

Benchmarking AlphaFold-enabled molecular docking predictions for antibiotic discovery

Efficient identification of drug mechanisms of action remains a challenge. Computational docking approaches have been widely used to predict drug binding targets; yet, such approaches depend on existing protein structures, and accurate structural predictions have only recently become available from AlphaFold2. Here, we combine AlphaFold2 with molecular docking simulations to predict protein‐ligand interactions between 296 proteins spanning Escherichia coli's essential proteome, and 218 active antibacterial compounds and 100 inactive compounds, respectively, pointing to widespread compound and protein promiscuity. We benchmark model performance by measuring enzymatic activity for 12 essential proteins treated with each antibacterial compound. We confirm extensive promiscuity, but find that the average area under the receiver operating characteristic curve (auROC) is 0.48, indicating weak model performance. We demonstrate that rescoring of docking poses using machine learning‐based approaches improves model performance, resulting in average auROCs as large as 0.63, and that ensembles of rescoring functions improve prediction accuracy and the ratio of true‐positive rate to false‐positive rate. This work indicates that advances in modeling protein‐ligand interactions, particularly using machine learning‐based approaches, are needed to better harness AlphaFold2 for drug discovery.

Single-Cell Analysis of the Antimicrobial and Bactericidal Activities of the Antimicrobial Peptide Magainin 2

Antimicrobial peptides (AMPs) inhibit the proliferation of or kill bacterial cells. To measure these activities, several methods have been used, which provide only the average value of many cells. Here, we report the development of a method to examine the antimicrobial and bactericidal activities of AMPs at the single-cell level (i.e., single-cell analysis) and apply this strategy to examine the interaction of an AMP, magainin 2 (Mag), with Escherichia coli cells. Using this method, we monitored the proliferation of single cells on agar in a microchamber and measured the distribution of the number of cells in each microcolony using optical microscopy. For method A, we incubated cells in the presence of various concentrations of AMPs for 3 h. The fraction of microcolonies containing only a single cell, Psingle, increased with the Mag concentration and reached 1 at a specific concentration, which corresponded to the MIC. For method B, after the interaction of a cell suspension with an AMP for a specific time, an aliquot was diluted to stop the interaction, and the proliferation of single cells then was monitored after a 3-h incubation; this method permits the definition of Psingle(t), the fraction of dead cells after the interaction. For the interaction of Mag with E. coli cells, Psingle(t) increased with the interaction time, reaching ~1 at 10 and 20 min for 25 and 13 μM Mag, respectively. Thus, these results indicate that a short interaction time between Mag and E. coli cells is sufficient to induce bacterial cell death. IMPORTANCE To elucidate the activity of antimicrobial peptides (AMPs) against bacterial cells, it is important to estimate the interaction time that is sufficient to induce cell death. We have developed a method to examine the antimicrobial and bactericidal activities of AMPs at the single-cell level (i.e., single-cell analysis). Using this method, we monitored the proliferation of single cells on agar in a microchamber and measured the distribution of the number of cells in each microcolony using optical microscopy. We found that during the interaction of magainin 2 (Mag) with E. coli cells, the fraction of dead cells, Psingle(t), increased with the interaction time, rapidly reaching 1 (e.g., 10 min for 25 μM Mag). This result indicates that Mag induces cell death after a short time of interaction.