PROJECT SUMMARY

As part of the IMI antimicrobial resistance (AMR) programme New Drugs for Bad Bugs, TRANSLOCATION aims to increase the overall understanding of how to get antibiotics into multi-resistant Gram-negative bacteria such as Escherichia coli and Klebsiella pneumoniae and how to stop the bacteria from ejecting the drug. In sharing the knowledge and data discovered, TRANSLOCATION will develop guidelines for designing and developing new drugs to tackle antibiotic resistance and create an information centre for pre-existing and on-going antibacterial research data which will be used to establish best practices for future antibacterial drug discovery efforts.

On-line antibiotics database (Citation: A Database of Force-Field Parameters, Dynamics, and Properties of Antimicrobial Compounds).

WP2 - Understanding the impact of porin structure and intrinsic permeability

WP2 local team

• Prof. Matteo Ceccarelli (co-leader)
• Dr. Giuliano Malloci
• Dr. Carlo Guardiani
• Dr. Susruta Samanta
• Dr. Mariano Andrea Scorciapino
• Dr. Tommaso d'Agostino (PhD)
• Dr. Silvia Acosta Gutierrez (PhD)

WP2 description

The general goal of WP2 is to understand the permeability barrier imposed by the outer membrane of Gram-negative bacteria (GNB). In particular we will work to quantify the role of porins and siderophore receptors, in the translocation of antibiotics into the periplasmic space of GNB. Our objectives are the following:

• Identify the structural features conferring selectivity to porins and contributing to antibiotics translocation
• Identify the interaction of antibiotics with the siderophore binding proteins/receptors
• Elucidate the chemical/physical properties of molecules that modulate the permeability through porins
• Elucidate the permeability properties and mechanisms of those molecules that utilise siderophore receptors (so called "Trojan horse" antibiotics)

In order to achieve these objectives we will:

• Characterize the relevant porins in Enterobacteriaceae, P. aeruginosa and A. baumannii using the input from WP1. Identify a small set of antibiotics, fragments and other molecules for studying their interactions with above channels. Identify key siderophore receptors, their siderophore targets and "Trojan horse" antibiotics with input from WP3.
• Clone and purify targeted outer membrane proteins for transport assays and crystallization.
• Apply cell-free assays on a single molecular level to elucidate the permeation for a selected list of compounds. We aim to push the time-resolution limit of electrophysiology to quantify the transport substrate turnover number for a number of porins. Results will be combined with cell-based assays in WP1.
• Crystallize a number of porins and key siderophore receptors. We expect about 3 outer membrane proteins structures per year. Co-crystallise the receptors with natural and synthetic siderophore conjugates, as well as porins with antibiotics shown to have long residence time in the pores (this information will come from electrophysiology or predicted by MD simulations).
• Perform all-atom MD simulation of small molecules through porins in order to elucidate (i) transport properties of model compounds, (ii) help with interpretation ofelectrophysiology experiments and (iii) make suggestions for the binding modes of antibiotics in porins. Evaluation of the "Trojan horse" mechanism proposed for transport of siderophore conjugates through siderophore receptors.
• Develop novel simulation schemes to extract more rapidly key data on substrate transit properties.

MD modelling will be performed mainly at UNICA.

WP4 - Novel penetration/efflux targets from a genetic approach

WP4 local team

• Prof. Paolo Ruggerone (local leader)
• Dr. Attilio Vargiu
• Dr. Giuliano Malloci
• Dr. Pierpaolo Cacciotto (PhD)

WP4 description

Manipulating permeability in Gram-negative bacteria to our therapeutic advantage logically involves increasing influx or diminishing efflux. Efflux is the most well known component and decreasing its contribution through the use of pump inhibitors has been tried extensively in the past. Considerable hurdles exist, though. First because of lack of specificity of these agents leading to toxicity, and second because of problems of narrow spectrum of action. The alternative approach to increase influx has not been as well scrutinized and especially not beyond the idea to upregulate porins, the only well studied specific uptake systems for antimicrobials. Thus there are critical gaps to fill: can we address other targets beyond porins and specifically improve uptake, and is this approach viable? can we understand better efflux pumps so as to circumvent them? Consequently, this WP is focused on:

• Performing a genetic screen to identify genes specifically involved in the bacterial outer membrane integrity. Such genes and gene products could be targeted by an 'adjuvant' molecule that would improve the activity of otherwise poorly permeable antibiotics
• Enhancing our structural understanding of efflux to identify structural and chemical determinants that are involved in the recognition and transport of substrates by pumps in MDR strains.
• Determining the membrane permeability (associated to the integrity and fluidity) of OM and IM and the changes occurring in resistant clinical strains. Understanding/characterizing the effect of adjuvants that improve the antibiotic penetration on the membrane (OM and IM).

Genetic aspects of membrane permeability
Identification of genes from selection in clinical conditions: Three Enterobacter and three Pseudomonas isolates sequentially obtained from single patients and which showed increased MICs during antimicrobial therapy (strains already available) will be studied by: (i) whole genome sequencing, (ii) time-resolved whole transcriptome sequencing (RNA-seq), (iii) time-resolved whole-cell proteomics approach (as developed in WP1), and (iv) phenotypic microarrays.
Identification of genes by screening of transposon libraries: Libraries of ordered or random transposon mutants of all non-essential genes of E. coli and P. aeruginosa will be screened for altered membrane permeability using the most adapted assay from those set up above. A second assay will be used as a secondary screen to confirm primary screen.
Characterization of identified genes as targets for ‘adjuvants’ to antimicrobials: ‘Clean’ inactivated mutants of the genes identified above will be characterized by: (i) all membrane permeability assays developed below, especially sensitivity towards antimicrobials (ii) phenotypic microarray, especially how the mutants would become more sensitive to antimicrobials used for curing the infection. The genes to study will be prioritized according to their presence in the broadest range of species and clinical strains, their conservation, their in vivo expression and their drugability.

Characterization of Efflux
Structure determination: Though some X-ray structures of efflux systems exist, additional structures will be determined. Homology modelling and state-of-the art docking procedures will complement and integrate these structural data. Thus, known substrates and non-substrates will be considered to pinpoint structural and dynamical determinants of molecule-transporter interactions. Attention will be focused on efflux systems of P. aeruginosa and A. baumannii.
In silico experiments/Simulations: In particular, the influence of dynamic effects will be tested using in silico and experimental approaches. Microscopically accurate methods such as MD simulations will provide inventories of the relevant interactions determining affinity and transport properties of effluxed compounds. The key deliverable will be the identification of pharmacophoric groups with the potential to escape efflux pumps. Additionally, we will simulate entire efflux systems of the RND family starting from the available crystallographic data. Information acquired within this subtask will be crucial for the activity of WP5. Deliverables from WP1, WP2, and WP3 will provide an important input to these simulations.
In addition, we will exploit the characteristics of known efflux blockers (from EFPIA, literature, and partners of the consortium) to identify pharmacophoric groups with the potential to modulate the efflux activity in MDR isolates, aiming to increase intracellular concentrations of antibiotics. Strong connections with WP1 and WP2 will help assessing the activity and the permeability of the compounds.

Membrane permeability and adjuvants (chemosensitizers or membranotropic agents)
The membrane permeability is the key point for translocation and we will scrutinize the integrity of inner and outer membranes. It is known that membranotropic agents exhibit some toxicity on prokaryotic and eukaryotic cells (e.g. polymyxins) and it is important to discriminate between an effect targeting OM or IM and the alteration by using specific combined assays.
Membrane permeability and susceptibility. In clinical resistant isolates and in susceptible strains we will examine the integrity of and discriminate between IM and OM permeability by: (i) measuring the release of markers of cytoplasmic and periplasmic spaces (via enzymatic activities or immunodetection), (ii) determining the concentration of fluorometric probes (1-2'-DNA, Hoechst 33342, DISC, etc), (iii) establishing a baseline sensitivity to a panel of antimicrobials, building on data from WP1 regarding susceptibility to antimicrobials and effects of different conditions mimicking in vivo environment, (iv) measuring the kinetic effects on the two membranes and discriminating between the release of small and large molecules to precise the membrane state.
Adjuvants and membrane permeability. These methods will be used to study representative collections of small molecule gathered from the consortium. Biological (combination) and biochemical studies will define the adjuvant selectivity (best molecule, antibiotic, characterization of target, dose effect, etc.). The SAR of adjuvants improving the intracellular concentration of antibiotics and the molecular basis of activity restoring optimal bacterial susceptibility will be determined. Key properties for adjuvants will be determined by these studies combining structure/activity/combination selectivity. Strong connections with tasks A-B, WP1 and WP2 will help to assess the activity and uptake of the compounds.

Modeling, including molecular dynamics simulations, associated with WP4, section B, will be mainly performed by UNICA.

WP5 - Modelling and simulation

WP5 local team

• Prof. Paolo Ruggerone (co-leader)
• Dr. Attilio Vargiu
• Dr. Samuele Salis (PhD)

WP5 description

The goal of this WP is to cast information coming from the literature, from the other WPs, and from EFPIA in a model able to quantitatively understand interrelationships between key penetration and efflux mechanisms and to assess their impact on antibacterial efficacy. Through this model we aim to determine the antibiotic concentration in bacteria accounting for the whole-cell penetration and efflux. Our objectives are the following:

• To characterize the single-molecule kinetics of single systems (porins and efflux pumps);
• To validate and to extend the information extracted from MD simulations by using ligand-based approach;
• To describe whole-cell antibacterial penetration and efflux;
• To predict the effect of perturbations such as inactivation of a specific efflux pump or replacement of porins with differential permeability.

To achieve these objectives we will setup the following strategy:

• Use of kinetic Monte Carlo (KMC) simulations based on experiments and MD simulations already published or provided by the partners of the consortium. Quantitative insights into single-molecule kinetics will be acquired for influx and efflux.
• Use of ligand-based approaches to validate and extend insights extracted from MD simulations. After running one or more MD simulations for a given set of compound and systems, and identifying possible affinity sites, the ligand-based approach can be used to see if those sites are consistent with the translocation (or efflux) data for a set of related molecules.
• Derivation of quantitative ordinary differential equation (ODE) models that describe whole-cell antibacterial penetration and efflux by properly combining the outcomes from KMC runs and the quantitative protein abundance data as measured by proteomics (WP1).
• Use of sensitivity analysis to identify the most relevant parameters and relevant ranges to guide further experimental analysis.
• Prediction of the effect of perturbations such as inactivation of a specific efflux pump or replacement of porins with differential permeability on the concentration of antibiotics.

In summary, we will iteratively build a multilevel model to scout the space of the physical-chemical parameters of the compounds in the presence of realistic conditions in term of concentration and protein population, to guide further MD simulations and experiments.

UNICA is already strongly involved in the study of efflux and influx systems via computational techniques and has also developed KMC schemes.

WP6 - ND4BB Information Centre: governance structure and software development

WP6 local team

• Dr. Andrea Bosin (local leader)
• Dr. Giovanni Serra

WP6 description

We will provide a knowledge repository which will evolve into a major global resource for antibiotic resistance R&D. The ND4BB Information Centre (InfoCentre) will combine legacy data on successful and failed approaches to antibiotic drug discovery from EFPIA and public partners with the data and high level descriptors of the data generated within Topics 1, 2, and future related IMI Calls under ND4BB. Data stored within the repository will be broad even by the Pharmaceutical industry standards, encompassing: HTS and Hit to Lead "Discovery" studies; "Early Discovery" in-vitro, in-vivo and modelling data; and "Development" data including pre-clinical animal study and clinical trial findings.

The practical realisation of the InfoCentre, will begin by defining the governing framework for internal and external data aggregation, extraction, storage, access, and analytics. Once this governing framework is defined, WP6 will focus on building a detailed blue print of the InfoCentre architecture given a thorough review of user needs. The challenge will be identifying and implementing appropriate hardware and software solutions in a model which must serve a highly diverse set of users and be sustainable beyond the Call funding period. The selection of vendor(s)/partner(s) will be critical as the InfoCentre will be required to process many different data types and formats, including data stored in text documents. Automation of data aggregation, extraction, and triaging tasks will be essential due to the quantity of data that is anticipated. As the success of the Translocation programme is heavily dependent on the InfoCentre solution, we envision a three month vendor/partner evaluation process whereby needs versus capabilities are rigorously examined. Post funding sustainability will be among the criterion upon which the selection of vendor(s)/partner(s) will be based. In the interim period as the InfoCentre is under construction, a temporary data depository solution will be adopted to begin the upload of consortium data. Regardless of vendor/partner selection, successful implementation of the framework will be overseen via an ND4BB Project Data Co-ordinator working closely with a Key Account Manager within the vendor/partner to oversee delivery of ND4BB specific functionality.

In order to create a manageable process, we envision separate solutions for "Primary" (i.e., raw) and "Derived" (i.e. post-processing) data. The Primary generated within the project (e.g., unprocessed compound screening results, patient identifiable clinical trials data, x-ray diffraction patterns, uncurated gene sequences) would be stored externally to the InfoCentre by the partners involved in generating the data and in line with local regulations and clinical data protection requirements.

To support the management of this Primary data and promote appropriate standardisation within the consortia, the WP6 team will implement a web based electronic laboratory notebook for use by the public partners. Only "Derived" results would be transferred to the InfoCentre and made accessible to named consortia members. Within the preclinical space, Derived results will include, for example, compound structures and associated bioassay results such as IC50s, MICs, SAR tables, elucidated protein structures, in-vivo and clinical study reports. Within the clinical space, "Derived" results will include, for example, demography, efficacy, and safety findings. "Derived" clinical data will be informed by COMBACTE colleagues and will comply with all necessary ethical regulations (and the Translocation Ethical Governance Team (EGT) will help provide oversight to this aspect). "Derived" data content will be indexed and searchable, providing high levels of interrogation capabilities for consortium users. Data dictionaries and catalogues will provide a means to enforce specific ontologies for data and results.

Prior to the transfer, appropriate control procedures and rules governing the data format, structure, and quality would be established and approved by the Translocation EGT. This process would be fully decided in the early stages of the programme and would include such considerations as bioassay quality control specifications, application of ethical restrictions for clinical data, aggregation of profiling data into consistent SAR table formats. The governing framework covering ethical issues and the use and the sharing of data will be defined as a priority at the outset of the project. This framework will address user access (within and external to the consortia), security, standards for data curation and management, ethics, life-cycle management, and knowledge exploitation strategy. The Translocation EGT will ratify the governing framework and regularly audit implementation and compliance. A select group of member from the ND4BB COMBACTE team representing EFPIA and Clinical teams with specific responsibilities for COMBACTE Clinical data management will work with the InfoCentre operational team to help guide its activities. As further ND4BB Topics come on line, additional skill sets will be added by co-opting the new consortia representatives into the WP6 team.

When fully implemented the InfoCentre solution will provide a framework to govern data sharing with highly granular (individual level) security and access/privilege controls (such as read, create, update, delete) which would provide hierarchal levels of access to different parts of the systems content. Internal users will have full read access to the InfoCentre along with instructor lead training and supporting documentation. For external (non-consortia users) an InfoCentre-based web interface would give the scientific community access to high level overviews of data collected within individual discovery, pre-clinical and clinical programs. An easy to use system will allow external users to request the corresponding Derived data from the internal InfoCentre. All requests will be reviewed by the Translocation EGT and WP leadership and will be approved on the basis of scientific merit. Online help will be provided to both internal and external users and a help desk will be available to resolve. We have already laid the groundwork for vendor/partner selection by preliminarily investigating several strategies.

One possibility is the use of a TransMART-based platform through a partnership with eTRIKS (European Translational Information and Knowledge Management Services). Melding the eTRIKS solution into our ND4BB effort offers several advantages. First, the development of eTRIKS is a parallel IMI-funded project with a specific remit to develop knowledge management systems for ND4BB-like projects. The eTRIKS consortium will create an "Open Translational Knowledge Management system" based on an evolved version of TransMART. The TransMART software application was developed by the EFPIA ND4BB partner Johnson and Johnson and is widely used in academic and industrial translational research organisations including six current IMI projects. Second, eTRIKS and ND4BB are closely aligned and are fortunate to have a common core of EFPIA partners (AstraZeneca, GSK and Janssen) and corresponding individual personnel. Finally, eTRIKS works with IMI consortia partners in a federated service model. Some 400 FTE months of transMART development and 400 FTE months of Analytics development resource are available for use on behalf of eTRIKS' partners.

UNICA will provide hosting facilities and informatics support for the very large tranche of Primary data generated within the computational and molecular modeling efforts in WP2 and WP5. UNICA will work with ESP to implement consistent standards in handling of Primary data within the Public partners, through development and application of bespoke ND4BB templates and process workflows for the electronic notebook solution.