Revisiting the Role of VraTSR in Staphylococcus aureus Response to Cell Wall-Targeting Antibiotics.

Exposure of Staphylococcus aureus to cell wall inhibitors leads to the activation of the VraTSR three-component sensory regulatory system. This system is composed of VraS, a membrane histidine kinase; VraR, its cognate response regulator, and VraT, a protein required for the full activity of VraTSR. The exact function of VraT remains mostly uncharacterized, although it has been proposed to detect the unknown stimulus sensed by the VraTSR system. Here, we elucidate the topology of VraT, showing that its C-terminal domain is extracellular. We also demonstrate that the signal sensed by VraTSR is not an intermediate in the peptidoglycan synthesis pathway, as previously suggested. Instead, the specific inhibition of the penicillin-binding protein (PBP)2 leads to strong activation of the system. IMPORTANCE The Gram-positive bacterial pathogen Staphylococcus aureus is currently the second most frequent cause of global deaths associated with antibiotic resistance. Its response to cell wall-targeting antibiotics requires the VraTSR three-component system, which senses cell wall damage. Here, we show that the signal sensed by VraTSR is not an intermediate in the peptidoglycan synthesis pathway, as previously suggested. Instead, the specific inhibition of the penicillin-binding protein (PBP)2, the major peptidoglycan synthase in S. aureus, leads to strong activation of the system. Identifying the exact cell wall damage signal is key to fully understanding the response of S. aureus to cell wall-targeting antibiotics.

Extracting Relational Explanations From Deep Neural Networks: A Survey From a Neural-Symbolic Perspective.

The term "explainable AI" refers to the goal of producing artificially intelligent agents that are capable of providing explanations for their decisions. Some models (e.g., rule-based systems) are designed to be explainable, while others are less explicit "black boxes" for which their reasoning remains a mystery. One example of the latter is the neural network, and over the past few decades, researchers in the field of neural-symbolic integration (NSI) have sought to extract relational knowledge from such networks. Extraction from deep neural networks, however, has remained a challenge until recent years in which many methods of extracting distinct, salient features from input or hidden feature spaces of deep neural networks have been proposed. Furthermore, methods of identifying relationships between these features have also emerged. This article presents examples of old and new developments in extracting relational explanations in order to argue that the latter have analogies in the former and, as such, can be described in terms of long-established taxonomies and frameworks presented in early neural-symbolic literature. We also outline potential future research directions that come to light from this refreshed perspective.

Brain-behaviour modes of covariation in healthy and clinically depressed young people.

Understanding how variations in dimensions of psychometrics, IQ and demographics relate to changes in brain connectivity during the critical developmental period of adolescence and early adulthood is a major challenge. This has particular relevance for mental health disorders where a failure to understand these links might hinder the development of better diagnostic approaches and therapeutics. Here, we investigated this question in 306 adolescents and young adults (14-24 y, 25 clinically depressed) using a multivariate statistical framework, based on canonical correlation analysis (CCA). By linking individual functional brain connectivity profiles to self-report questionnaires, IQ and demographic data we identified two distinct modes of covariation. The first mode mapped onto an externalization/internalization axis and showed a strong association with sex. The second mode mapped onto a well-being/distress axis independent of sex. Interestingly, both modes showed an association with age. Crucially, the changes in functional brain connectivity associated with changes in these phenotypes showed marked developmental effects. The findings point to a role for the default mode, frontoparietal and limbic networks in psychopathology and depression.

A comparative genomics approach for identifying host-range determinants in Streptococcus thermophilus bacteriophages.

Comparative genomics has proven useful in exploring the biodiversity of phages and understanding phage-host interactions. This knowledge is particularly useful for phages infecting Streptococcus thermophilus, as they constitute a constant threat during dairy fermentations. Here, we explore the genetic diversity of S. thermophilus phages to identify genetic determinants with a signature for host specificity, which could be linked to the bacterial receptor genotype. A comparative genomic analysis was performed on 142 S. thermophilus phage genomes, 55 of which were sequenced in this study. Effectively, 94 phages were assigned to the group cos (DT1), 36 to the group pac (O1205), six to the group 5093, and six to the group 987. The core genome-based phylogeny of phages from the two dominating groups and their receptor binding protein (RBP) phylogeny corresponded to the phage host-range. A role of RBP in host recognition was confirmed by constructing a fluorescent derivative of the RBP of phage CHPC951, followed by studying the binding of the protein to the host strain. Furthermore, the RBP phylogeny of the cos group was found to correlate with the host genotype of the exocellular polysaccharide-encoding operon. These findings provide novel insights towards developing strategies to combat phage infections in dairies.

SEDS-bPBP pairs direct lateral and septal peptidoglycan synthesis in Staphylococcus aureus.

Peptidoglycan (PGN) is the major component of the bacterial cell wall, a structure that is essential for the physical integrity and shape of the cell. Bacteria maintain cell shape by directing PGN incorporation to distinct regions of the cell, namely, through the localization of late-stage PGN synthesis proteins. These include two key protein families, SEDS transglycosylases and bPBP transpeptidases, proposed to function in cognate pairs. Rod-shaped bacteria have two SEDS-bPBP pairs, involved in elongation and division. Here, we elucidate why coccoid bacteria, such as Staphylococcus aureus, also possess two SEDS-bPBP pairs. We determined that S. aureus RodA-PBP3 and FtsW-PBP1 probably constitute cognate pairs of interacting proteins. A lack of RodA-PBP3 resulted in more spherical cells due to deficient sidewall PGN synthesis, whereas depletion of FtsW-PBP1 arrested normal septal PGN incorporation. Although PBP1 is an essential protein, a mutant lacking PBP1 transpeptidase activity is viable, showing that this protein has a second function. We propose that the FtsW-PBP1 pair has a role in stabilizing the divisome at midcell. In the absence of these proteins, the divisome appears as multiple rings or arcs that drive lateral PGN incorporation, leading to cell elongation. We conclude that RodA-PBP3 and FtsW-PBP1 mediate sidewall and septal PGN incorporation, respectively, and that their activity must be balanced to maintain coccoid morphology.

Combining heterogeneous data sources for neuroimaging based diagnosis: re-weighting and selecting what is important.

Combining neuroimaging and clinical information for diagnosis, as for example behavioral tasks and genetics characteristics, is potentially beneficial but presents challenges in terms of finding the best data representation for the different sources of information. Their simple combination usually does not provide an improvement if compared with using the best source alone. In this paper, we proposed a framework based on a recent multiple kernel learning algorithm called EasyMKL and we investigated the benefits of this approach for diagnosing two different mental health diseases. The well known Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset tackling the Alzheimer Disease (AD) patients versus healthy controls classification task, and a second dataset tackling the task of classifying an heterogeneous group of depressed patients versus healthy controls. We used EasyMKL to combine a huge amount of basic kernels alongside a feature selection methodology, pursuing an optimal and sparse solution to facilitate interpretability. Our results show that the proposed approach, called EasyMKLFS, outperforms baselines (e.g. SVM and SimpleMKL), state-of-the-art random forests (RF) and feature selection (FS) methods.

The pentaglycine bridges of Staphylococcus aureus peptidoglycan are essential for cell integrity.

Bacterial cells are surrounded by cell wall, whose main component is peptidoglycan (PG), a macromolecule that withstands the internal turgor of the cell. PG composition can vary considerably between species. The Gram-positive pathogen Staphylococcus aureus possesses highly crosslinked PG due to the presence of cross bridges containing five glycines, which are synthesised by the FemXAB protein family. FemX adds the first glycine of the cross bridge, while FemA and FemB add the second and the third, and the fourth and the fifth glycines, respectively. Of these, FemX was reported to be essential. To investigate the essentiality of FemAB, we constructed a conditional S. aureus mutant of the femAB operon. Depletion of femAB was lethal, with cells appearing as pseudomulticellular forms that eventually lyse due to extensive membrane rupture. This deleterious effect was mitigated by drastically increasing the osmolarity of the medium, indicating that pentaglycine crosslinks are required for S. aureus cells to withstand internal turgor. Despite the absence of canonical membrane targeting domains, FemA has been shown to localise at the membrane. To study its mechanism of localisation, we constructed mutants in key residues present in the putative transferase pocket and the α6 helix of FemA, possibly involved in tRNA binding. Mutations in the α6 helix led to a sharp decrease in protein activity in vivo and in vitro but did not impair correct membrane localisation, indicating that FemA activity is not required for localisation. Our data indicates that, contrarily to what was previously thought, S. aureus cells do not survive in the absence of a pentaglycine cross bridge.

Peptidoglycan synthesis drives an FtsZ-treadmilling-independent step of cytokinesis.

Peptidoglycan is the main component of the bacterial wall and protects cells from the mechanical stress that results from high intracellular turgor. Peptidoglycan biosynthesis is very similar in all bacteria; bacterial shapes are therefore mainly determined by the spatial and temporal regulation of peptidoglycan synthesis rather than by the chemical composition of peptidoglycan. The form of rod-shaped bacteria, such as Bacillus subtilis or Escherichia coli, is generated by the action of two peptidoglycan synthesis machineries that act at the septum and at the lateral wall in processes coordinated by the cytoskeletal proteins FtsZ and MreB, respectively. The tubulin homologue FtsZ is the first protein recruited to the division site, where it assembles in filaments-forming the Z ring-that undergo treadmilling and recruit later divisome proteins. The rate of treadmilling in B. subtilis controls the rates of both peptidoglycan synthesis and cell division. The actin homologue MreB forms discrete patches that move circumferentially around the cell in tracks perpendicular to the long axis of the cell, and organize the insertion of new cell wall during elongation. Cocci such as Staphylococcus aureus possess only one type of peptidoglycan synthesis machinery, which is diverted from the cell periphery to the septum in preparation for division. The molecular cue that coordinates this transition has remained elusive. Here we investigate the localization of S. aureus peptidoglycan biosynthesis proteins and show that the recruitment of the putative lipid II flippase MurJ to the septum, by the DivIB-DivIC-FtsL complex, drives peptidoglycan incorporation to the midcell. MurJ recruitment corresponds to a turning point in cytokinesis, which is slow and dependent on FtsZ treadmilling before MurJ arrival but becomes faster and independent of FtsZ treadmilling after peptidoglycan synthesis activity is directed to the septum, where it provides additional force for cell envelope constriction.

A quinolinol-based small molecule with anti-MRSA activity that targets bacterial membrane and promotes fermentative metabolism.

In a loss-of-viability screen of small molecules against methicillin-resistant Staphylococcus aureus (MRSA) USA300, we found a small molecule, designated DNAC-2, which has an MIC of 8 µg ml-1. DNAC-2 is a quinolinol derivative that is bactericidal at 2X MIC. Macromolecular synthesis assays at 2 × MIC of DNAC-2 revealed inhibition of DNA, cell wall, RNA and protein synthesis within fifteen to thirty minutes of treatment when compared to the untreated control. Transmission electron microscopy of DNAC-2-treated cells revealed a significantly thicker cell wall and impaired daughter cell separation. Exposure of USA300 cells to 1 × MIC of DNAC-2 resulted in mislocalization of PBP2 away from the septum in an FtsZ-independent manner. In addition, membrane localization with FM4-64, as well as depolarization study with DiOC2 and lipophilic cation TPP+ displayed membrane irregularities and rapid membrane depolarization, respectively, in DNAC-2-treated cells vs -untreated control. However, DNAC-2 exhibited almost no toxicity toward eukaryotic membranes. Notably, DNAC-2 drives energy generation toward substrate level phosphorylation and the bacteria become more sensitive to DNAC-2 under anaerobic conditions. We propose that DNAC-2 affects USA300 by targeting the membrane, leading to partial membrane depolarization and subsequently affecting aerobic respiration and energy-dependent functional organization of macromolecular biosynthetic pathways. The multiple effects may have the desirable consequence of limiting the emergence of resistance to DNAC-2.

Predictive modelling using neuroimaging data in the presence of confounds.

When training predictive models from neuroimaging data, we typically have available non-imaging variables such as age and gender that affect the imaging data but which we may be uninterested in from a clinical perspective. Such variables are commonly referred to as 'confounds'. In this work, we firstly give a working definition for confound in the context of training predictive models from samples of neuroimaging data. We define a confound as a variable which affects the imaging data and has an association with the target variable in the sample that differs from that in the population-of-interest, i.e., the population over which we intend to apply the estimated predictive model. The focus of this paper is the scenario in which the confound and target variable are independent in the population-of-interest, but the training sample is biased due to a sample association between the target and confound. We then discuss standard approaches for dealing with confounds in predictive modelling such as image adjustment and including the confound as a predictor, before deriving and motivating an Instance Weighting scheme that attempts to account for confounds by focusing model training so that it is optimal for the population-of-interest. We evaluate the standard approaches and Instance Weighting in two regression problems with neuroimaging data in which we train models in the presence of confounding, and predict samples that are representative of the population-of-interest. For comparison, these models are also evaluated when there is no confounding present. In the first experiment we predict the MMSE score using structural MRI from the ADNI database with gender as the confound, while in the second we predict age using structural MRI from the IXI database with acquisition site as the confound. Considered over both datasets we find that none of the methods for dealing with confounding gives more accurate predictions than a baseline model which ignores confounding, although including the confound as a predictor gives models that are less accurate than the baseline model. We do find, however, that different methods appear to focus their predictions on specific subsets of the population-of-interest, and that predictive accuracy is greater when there is no confounding present. We conclude with a discussion comparing the advantages and disadvantages of each approach, and the implications of our evaluation for building predictive models that can be used in clinical practice.

A multiple hold-out framework for Sparse Partial Least Squares.

BACKGROUND: Supervised classification machine learning algorithms may have limitations when studying brain diseases with heterogeneous populations, as the labels might be unreliable. More exploratory approaches, such as Sparse Partial Least Squares (SPLS), may provide insights into the brain's mechanisms by finding relationships between neuroimaging and clinical/demographic data. The identification of these relationships has the potential to improve the current understanding of disease mechanisms, refine clinical assessment tools, and stratify patients. SPLS finds multivariate associative effects in the data by computing pairs of sparse weight vectors, where each pair is used to remove its corresponding associative effect from the data by matrix deflation, before computing additional pairs. NEW METHOD: We propose a novel SPLS framework which selects the adequate number of voxels and clinical variables to describe each associative effect, and tests their reliability by fitting the model to different splits of the data. As a proof of concept, the approach was applied to find associations between grey matter probability maps and individual items of the Mini-Mental State Examination (MMSE) in a clinical sample with various degrees of dementia. RESULTS: The framework found two statistically significant associative effects between subsets of brain voxels and subsets of the questions/tasks. COMPARISON WITH EXISTING METHODS: SPLS was compared with its non-sparse version (PLS). The use of projection deflation versus a classical PLS deflation was also tested in both PLS and SPLS. CONCLUSIONS: SPLS outperformed PLS, finding statistically significant effects and providing higher correlation values in hold-out data. Moreover, projection deflation provided better results.

The SpoIIQ-SpoIIIAH complex of Clostridium difficile controls forespore engulfment and late stages of gene expression and spore morphogenesis.

Engulfment of the forespore by the mother cell is a universal feature of endosporulation. In Bacillus subtilis, the forespore protein SpoIIQ and the mother cell protein SpoIIIAH form a channel, essential for endosporulation, through which the developing spore is nurtured. The two proteins also form a backup system for engulfment. Unlike in B. subtilis, SpoIIQ of Clostridium difficile has intact LytM zinc-binding motifs. We show that spoIIQ or spoIIIAH deletion mutants of C. difficile result in anomalous engulfment, and that disruption of the SpoIIQ LytM domain via a single amino acid substitution (H120S) impairs engulfment differently. SpoIIQ and SpoIIQ(H120S) interact with SpoIIIAH throughout engulfment. SpoIIQ, but not SpoIIQ(H120S) , binds Zn(2+) , and metal absence alters the SpoIIQ-SpoIIIAH complex in vitro. Possibly, SpoIIQ(H120S) supports normal engulfment in some cells but not a second function of the complex, required following engulfment completion. We show that cells of the spoIIQ or spoIIIAH mutants that complete engulfment are impaired in post-engulfment, forespore and mother cell-specific gene expression, suggesting a channel-like function. Both engulfment and a channel-like function may be ancestral functions of SpoIIQ-SpoIIIAH while the requirement for engulfment was alleviated through the emergence of redundant mechanisms in B. subtilis and related organisms.

Cell shape dynamics during the staphylococcal cell cycle.

Staphylococcus aureus is an aggressive pathogen and a model organism to study cell division in sequential orthogonal planes in spherical bacteria. However, the small size of staphylococcal cells has impaired analysis of changes in morphology during the cell cycle. Here we use super-resolution microscopy and determine that S. aureus cells are not spherical throughout the cell cycle, but elongate during specific time windows, through peptidoglycan synthesis and remodelling. Both peptidoglycan hydrolysis and turgor pressure are required during division for reshaping the flat division septum into a curved surface. In this process, the septum generates less than one hemisphere of each daughter cell, a trait we show is common to other cocci. Therefore, cell surface scars of previous divisions do not divide the cells in quadrants, generating asymmetry in the daughter cells. Our results introduce a need to reassess the models for division plane selection in cocci.

Characterization of a novel small molecule that potentiates ß-lactam activity against gram-positive and gram-negative pathogens.

In a loss-of-viability screen using small molecules against methicillin-resistant Staphylococcus aureus (MRSA) strain USA300 with a sub-MIC of a ß-lactam, we found a small molecule, designated DNAC-1, which potentiated the effect of oxacillin (i.e., the MIC of oxacillin decreased from 64 to 0.25 µg/ml). Fluorescence microscopy indicated a disruption in the membrane structures within 15 min of exposure to DNAC-1 at 2× MIC. This permeabilization was accompanied by a rapid loss of membrane potential, as monitored by use of the DiOC2 (3,3'-diethyloxacarbocyanine iodide) dye. Macromolecular analysis showed the inhibition of staphylococcal cell wall synthesis by DNAC-1. Transmission electron microscopy of treated MRSA USA300 cells revealed a slightly thicker cell wall, together with mesosome-like projections into the cytosol. The exposure of USA300 cells to DNAC-1 was associated with the mislocalization of FtsZ accompanied by the localization of penicillin-binding protein 2 (PBP2) and PBP4 away from the septum, as well as mild activation of the vraRS-mediated cell wall stress response. However, DNAC-1 does not have any generalized toxicity toward mammalian host cells. DNAC-1 in combination with ceftriaxone is also effective against an assortment of Gram-negative pathogens. Using a murine subcutaneous coinjection model with 10(8) CFU of USA300 as a challenge inoculum, DNAC-1 alone or DNAC-1 with a sub-MIC of oxacillin resulted in a 6-log reduction in bacterial load and decreased abscess formation compared to the untreated control. We propose that DNAC-1, by exerting a bimodal effect on the cell membrane and cell wall, is a viable candidate in the development of combination therapy against many common bacterial pathogens.

Differential localization of LTA synthesis proteins and their interaction with the cell division machinery in Staphylococcus aureus.

Lipoteichoic acid (LTA) is an important cell wall component of Gram-positive bacteria. In Staphylococcus aureus it consists of a polyglycerolphosphate-chain that is retained within the membrane via a glycolipid. Using an immunofluorescence approach, we show here that the LTA polymer is not surface exposed in S. aureus, as it can only be detected after digestion of the peptidoglycan layer. S. aureus mutants lacking LTA are enlarged and show aberrant positioning of septa, suggesting a link between LTA synthesis and the cell division process. Using a bacterial two-hybrid approach, we show that the three key LTA synthesis proteins, YpfP and LtaA, involved in glycolipid production, and LtaS, required for LTA backbone synthesis, interact with one another. All three proteins also interacted with numerous cell division and peptidoglycan synthesis proteins, suggesting the formation of a multi-enzyme complex and providing further evidence for the co-ordination of these processes. When assessed by fluorescence microscopy, YpfP and LtaA fluorescent protein fusions localized to the membrane while the LtaS enzyme accumulated at the cell division site. These data support a model whereby LTA backbone synthesis proceeds in S. aureus at the division site in co-ordination with cell division, while glycolipid synthesis takes place throughout the membrane.

Restoring methicillin-resistant Staphylococcus aureus susceptibility to ß-lactam antibiotics.

Despite the need for new antibiotics to treat drug-resistant bacteria, current clinical combinations are largely restricted to ß-lactam antibiotics paired with ß-lactamase inhibitors. We have adapted a Staphylococcus aureus antisense knockdown strategy to genetically identify the cell division Z ring components-FtsA, FtsZ, and FtsW-as ß-lactam susceptibility determinants of methicillin-resistant S. aureus (MRSA). We demonstrate that the FtsZ-specific inhibitor PC190723 acts synergistically with ß-lactam antibiotics in vitro and in vivo and that this combination is efficacious in a murine model of MRSA infection. Fluorescence microscopy localization studies reveal that synergy between these agents is likely to be elicited by the concomitant delocalization of their cognate drug targets (FtsZ and PBP2) in MRSA treated with PC190723. A 2.0 Å crystal structure of S. aureus FtsZ in complex with PC190723 identifies the compound binding site, which corresponds to the predominant location of mutations conferring resistance to PC190723 (PC190723(R)). Although structural studies suggested that these drug resistance mutations may be difficult to combat through chemical modification of PC190723, combining PC190723 with the ß-lactam antibiotic imipenem markedly reduced the spontaneous frequency of PC190723(R) mutants. Multiple MRSA PC190723(R) FtsZ mutants also displayed attenuated virulence and restored susceptibility to ß-lactam antibiotics in vitro and in a mouse model of imipenem efficacy. Collectively, these data support a target-based approach to rationally develop synergistic combination agents that mitigate drug resistance and effectively treat MRSA infections.