The antibiotic darobactin mimics a ß-strand to inhibit outer membrane insertase.

Antibiotics that target Gram-negative bacteria in new ways are needed to resolve the antimicrobial resistance crisis1-3. Gram-negative bacteria are protected by an additional outer membrane, rendering proteins on the cell surface attractive drug targets4,5. The natural compound darobactin targets the bacterial insertase BamA6-the central unit of the essential BAM complex, which facilitates the folding and insertion of outer membrane proteins7-13. BamA lacks a typical catalytic centre, and it is not obvious how a small molecule such as darobactin might inhibit its function. Here we resolve the mode of action of darobactin at the atomic level using a combination of cryo-electron microscopy, X-ray crystallography, native mass spectrometry, in vivo experiments and molecular dynamics simulations. Two cyclizations pre-organize the darobactin peptide in a rigid ß-strand conformation. This creates a mimic of the recognition signal of native substrates with a superior ability to bind to the lateral gate of BamA. Upon binding, darobactin replaces a lipid molecule from the lateral gate to use the membrane environment as an extended binding pocket. Because the interaction between darobactin and BamA is largely mediated by backbone contacts, it is particularly robust against potential resistance mutations. Our results identify the lateral gate as a functional hotspot in BamA and will allow the rational design of antibiotics that target this bacterial Achilles heel.

Polymyxin B1 in the Escherichia coli inner membrane: A complex story of protein and lipopolysaccharide-mediated insertion.

The rise in multi-drug resistant Gram-negative bacterial infections has led to an increased need for "last-resort" antibiotics such as polymyxins. However, the emergence of polymyxin-resistant strains threatens to bring about a post-antibiotic era. Thus, there is a need to develop new polymyxin-based antibiotics, but a lack of knowledge of the mechanism of action of polymyxins hinders such efforts. It has recently been suggested that polymyxins induce cell lysis of the Gram-negative bacterial inner membrane (IM) by targeting trace amounts of lipopolysaccharide (LPS) localized there. We use multiscale molecular dynamics (MD), including long-timescale coarse-grained (CG) and all-atom (AA) simulations, to investigate the interactions of polymyxin B1 (PMB1) with bacterial IM models containing phospholipids (PLs), small quantities of LPS, and IM proteins. LPS was observed to (transiently) phase separate from PLs at multiple LPS concentrations, and associate with proteins in the IM. PMB1 spontaneously inserted into the IM and localized at the LPS-PL interface, where it cross-linked lipid headgroups via hydrogen bonds, sampling a wide range of interfacial environments. In the presence of membrane proteins, a small number of PMB1 molecules formed interactions with them, in a manner that was modulated by local LPS molecules. Electroporation-driven translocation of PMB1 via water-filled pores was favored at the protein-PL interface, supporting the 'destabilizing' role proteins may have within the IM. Overall, this in-depth characterization of PMB1 modes of interaction reveals how small amounts of mislocalized LPS may play a role in pre-lytic targeting and provides insights that may facilitate rational improvement of polymyxin-based antibiotics.

High efficacy of the F-ATP synthase inhibitor TBAJ-5307 against nontuberculous mycobacteria in vitro and in vivo.

The F1FO-ATP synthase engine is essential for viability and growth of nontuberculous mycobacteria (NTM) by providing the biological energy ATP and keeping ATP homeostasis under hypoxic stress conditions. Here, we report the discovery of the diarylquinoline TBAJ-5307 as a broad spectrum anti-NTM inhibitor, targeting the FO domain of the engine and preventing rotation and proton translocation. TBAJ-5307 is active at low nanomolar concentrations against fast- and slow-growing NTM as well as clinical isolates by depleting intrabacterial ATP. As demonstrated for the fast grower Mycobacterium abscessus, the compound is potent in vitro and in vivo, without inducing toxicity. Combining TBAJ-5307 with anti-NTM antibiotics or the oral tebipenem-avibactam pair showed attractive potentiation. Furthermore, the TBAJ-5307-tebipenem-avibactam cocktail kills the pathogen, suggesting a novel oral combination for the treatment of NTM lung infections.

Variations of the Mycobacterium abscessus F-ATP synthase subunit a-c interface alter binding and potency of the anti-TB drug bedaquiline.

The anti-tuberculosis therapeutic bedaquiline (BDQ) is used against Mycobacterium abscessus. In M. abscessus BDQ is only bacteriostatic and less potent compared to M. tuberculosis or M. smegmatis. Here we demonstrate its reduced ATP synthesis inhibition against M. abscessus inside-out vesicles, including the F1FO-ATP synthase. Molecular dynamics simulations and binding free energy calculations highlight the differences in drug-binding of the M. abscessus and M. smegmatis FO-domain at the lagging site, where the drug deploys its mechanistic action, inhibiting ATP synthesis. These data pave the way for improved anti-M. abscessus BDQ analogs.

Antibody affinity versus dengue morphology influences neutralization.

Different strains within a dengue serotype (DENV1-4) can have smooth, or "bumpy" surface morphologies with different antigenic characteristics at average body temperature (37°C). We determined the neutralizing properties of a serotype cross-reactive human monoclonal antibody (HMAb) 1C19 for strains with differing morphologies within the DENV1 and DENV2 serotypes. We mapped the 1C19 epitope to E protein domain II by hydrogen deuterium exchange mass spectrometry, cryoEM and molecular dynamics simulations, revealing that this epitope is likely partially hidden on the virus surface. We showed the antibody has high affinity for binding to recombinant DENV1 E proteins compared to those of DENV2, consistent with its strong neutralizing activities for all DENV1 strains tested regardless of their morphologies. This finding suggests that the antibody could out-compete E-to-E interaction for binding to its epitope. In contrast, for DENV2, HMAb 1C19 can only neutralize when the epitope becomes exposed on the bumpy-surfaced particle. Although HMAb 1C19 is not a suitable therapeutic candidate, this study with HMAb 1C19 shows the importance of choosing a high-affinity antibody that could neutralize diverse dengue virus morphologies for therapeutic purposes.

Dengue Virus Capsid Protein Facilitates Genome Compaction and Packaging.

Dengue virus (DENV) is a single-stranded (+)-sense RNA virus that infects humans and mosquitoes, posing a significant health risk in tropical and subtropical regions. Mature virions are composed of an icosahedral shell of envelope (E) and membrane (M) proteins circumscribing a lipid bilayer, which in turn contains a complex of the approximately 11 kb genomic RNA with capsid (C) proteins. Whereas the structure of the envelope is clearly defined, the structure of the packaged genome in complex with C proteins remains elusive. Here, we investigated the interactions of C proteins with viral RNA, in solution and inside mature virions, via footprinting and cross-linking experiments. We demonstrated that C protein interaction with DENV genomes saturates at an RNA:C protein ratio below 1:250. Moreover, we also showed that the length of the RNA genome interaction sites varies, in a multimodal distribution, consistent with the C protein binding to each RNA site mostly in singlets or pairs (and, in some instances, higher numbers). We showed that interaction sites are preferentially sites with low base pairing, as previously measured by 2'-acetylation analyzed by primer extension (SHAPE) reactivity indicating structuredness. We found a clear association pattern emerged: RNA-C protein binding sites are strongly associated with long-range RNA-RNA interaction sites, particularly inside virions. This, in turn, explains the need for C protein in viral genome packaging: the protein has a chief role in coordinating these key interactions, promoting proper packaging of viral RNA. Such sites are, thus, highly consequential for viral assembly, and, as such, may be targeted in future drug development strategies against these and related viruses.

What have molecular simulations contributed to understanding of Gram-negative bacterial cell envelopes?

Bacterial cell envelopes are compositionally complex and crowded and while highly dynamic in some areas, their molecular motion is very limited, to the point of being almost static in others. Therefore, it is no real surprise that studying them at high resolution across a range of temporal and spatial scales requires a number of different techniques. Details at atomistic to molecular scales for up to tens of microseconds are now within range for molecular dynamics simulations. Here we review how such simulations have contributed to our current understanding of the cell envelopes of Gram-negative bacteria.

Computational modelling of flavivirus dynamics: The ins and outs.

Enveloped viruses such as the flaviviruses represent a significant burden to human health around the world, with hundreds of millions of people each year affected by dengue alone. In an effort to improve our understanding of the molecular basis for the infective mechanisms of these viruses, extensive computational modelling approaches have been applied to elucidate their conformational dynamics. Multiscale protocols have been developed to simulate flavivirus envelopes in close accordance with biophysical data, in particular derived from cryo-electron microscopy, enabling high-resolution refinement of their structures and elucidation of the conformational changes associated with adaptation both to host environments and to immunological factors such as antibodies. Likewise, integrative modelling efforts combining data from biophysical experiments and from genome sequencing with chemical modification are providing unparalleled insights into the architecture of the previously unresolved nucleocapsid complex. Collectively, this work provides the basis for the future rational design of new antiviral therapeutics and vaccine development strategies targeting enveloped viruses.

Site-Specific Steric Control of SARS-CoV-2 Spike Glycosylation.

A central tenet in the design of vaccines is the display of native-like antigens in the elicitation of protective immunity. The abundance of N-linked glycans across the SARS-CoV-2 spike protein is a potential source of heterogeneity among the many different vaccine candidates under investigation. Here, we investigate the glycosylation of recombinant SARS-CoV-2 spike proteins from five different laboratories and compare them against S protein from infectious virus, cultured in Vero cells. We find patterns that are conserved across all samples, and this can be associated with site-specific stalling of glycan maturation that acts as a highly sensitive reporter of protein structure. Molecular dynamics simulations of a fully glycosylated spike support a model of steric restrictions that shape enzymatic processing of the glycans. These results suggest that recombinant spike-based SARS-CoV-2 immunogen glycosylation reproducibly recapitulates signatures of viral glycosylation.

Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products.

Thrombin-derived C-terminal peptides (TCPs), including a major 11-kDa fragment (TCP96), are produced through cleavage by human neutrophil elastase and aggregate lipopolysaccharide (LPS) and the Gram-negative bacterium Escherichia coli However, the physiological roles of TCP96 in controlling bacterial infections and reducing LPS-induced inflammation are unclear. Here, using various biophysical methods, in silico molecular modeling, microbiological and cellular assays, and animal models, we examined the structural features and functional roles of recombinant TCP96 (rTCP96) in the aggregation of multiple bacteria and the Toll-like receptor (TLR) agonists they produce. We found that rTCP96 aggregates both Gram-negative and Gram-positive bacteria, including Staphylococcus aureus and Pseudomonas aeruginosa, and their cell-wall components LPS, lipid A, and lipoteichoic acid (LTA). The Gram-negative bacteria E. coli and P. aeruginosa were particularly sensitive to aggregation-induced bacterial permeabilization and killing. As a proof of concept, we show that rTCP96 reduces LPS-induced NF-κB activation in human monocytes, as well as in mouse models of LPS-induced subcutaneous inflammation. Moreover, in a mouse model of subcutaneous inoculation with P. aeruginosa, rTCP96 reduced bacterial levels. Together, these results link TCP-mediated aggregation of endotoxins and bacteria in vitro to attenuation of inflammation and bacterial levels in vivo.

Molecular basis of dengue virus serotype 2 morphological switch from 29°C to 37°C.

The ability of DENV2 to display different morphologies (hence different antigenic properties) complicates vaccine and therapeutics development. Previous studies showed most strains of laboratory adapted DENV2 particles changed from smooth to "bumpy" surfaced morphology when the temperature is switched from 29°C at 37°C. Here we identified five envelope (E) protein residues different between two alternative passage history DENV2 NGC strains exhibiting smooth or bumpy surface morphologies. Several mutations performed on the smooth DENV2 infectious clone destabilized the surface, as observed by cryoEM. Molecular dynamics simulations demonstrated how chemically subtle substitution at various positions destabilized dimeric interactions between E proteins. In contrast, three out of four DENV2 clinical isolates showed a smooth surface morphology at 37°C, and only at high fever temperature (40°C) did they become "bumpy". These results imply vaccines should contain particles representing both morphologies. For prophylactic and therapeutic treatments, this study also informs on which types of antibodies should be used at different stages of an infection, i.e., those that bind to monomeric E proteins on the bumpy surface or across multiple E proteins on the smooth surfaced virus.

The disordered plant dehydrin Lti30 protects the membrane during water-related stress by cross-linking lipids.

Dehydrins are intrinsically disordered proteins, generally expressed in plants as a response to embryogenesis and water-related stress. Their suggested functions are in membrane stabilization and cell protection. All dehydrins contain at least one copy of the highly conserved K-segment, proposed to be a membrane-binding motif. The dehydrin Lti30 (Arabidopsis thaliana) is up-regulated during cold and drought stress conditions and comprises six K-segments, each with two adjacent histidines. Lti30 interacts with the membrane electrostatically via pH-dependent protonation of the histidines. In this work, we seek a molecular understanding of the membrane interaction mechanism of Lti30 by determining the diffusion and molecular organization of Lti30 on model membrane systems by imaging total internal reflection- fluorescence correlation spectroscopy (ITIR-FCS) and molecular dynamics (MD) simulations. The dependence of the diffusion coefficient explored by ITIR-FCS together with MD simulations yields insights into Lti30 binding, domain partitioning, and aggregation. The effect of Lti30 on membrane lipid diffusion was studied on fluorescently labeled supported lipid bilayers of different lipid compositions at mechanistically important pH conditions. In parallel, we compared the mode of diffusion for short individual K-segment peptides. The results indicate that Lti30 binds the lipid bilayer via electrostatics, which restricts the mobility of lipids and bound protein molecules. At low pH, Lti30 binding induced lipid microdomain formation as well as protein aggregation, which could be correlated with one another. Moreover, at physiological pH, Lti30 forms nanoscale aggregates when proximal to the membrane suggesting that Lti30 may protect the cell by "cross-linking" the membrane lipids.

3D reconstruction and flexibility of the hybrid engine Acetobacterium woodii F-ATP synthase.

The Na+-translocating F1FO ATP synthase from Acetobacterium woodii (AwF-ATP synthase) with a subunit stoichiometry of α3:ß3:γ:δ:ε:a:b2:(c2/3)9:c1 represents an evolutionary path between ATP-synthases and vacuolar ATPases, by containing a heteromeric rotor c-ring, composed of subunits c1, c2 and c3, and an extra loop (γ195-211) within the rotary γ subunit. Here, the recombinant AwF-ATP synthase was subjected to negative stain electron microscopy and single particle analysis. The reference free 2D class averages revealed high flexibility of the enzyme, wherein the F1 and FO domains distinctively bended to adopt multiple conformations. Moreover, both the F1 and FO domains tilted relative to each other to a maximum extent of 28° and 30°, respectively. The first 3D reconstruction of the AwF-ATP synthase was determined which accommodates well the modelled structure of the AwF-ATP synthase as well as the γ195-211-loop. Molecular simulations of the enzyme underlined the bending features and flexibility observed in the electron micrographs, and enabled assessment of the dynamics of the extra γ195-211-loop.

The Molecular Basis for Purine Binding Selectivity in the Bacterial ATP Synthase ϵ Subunit.

The ϵ subunit of ATP synthases has been proposed to regulate ATP hydrolysis in bacteria. Prevailing evidence supports the notion that when the ATP concentration falls below a certain threshold, the ϵ subunit changes its conformation from a non-inhibitory down-state to an extended up-state that then inhibits enzymatic ATP hydrolysis by binding to the catalytic domain. It has been demonstrated that the ϵ subunit from Bacillus PS3 is selective for ATP over other nucleotides, including GTP. In this study, the purine triphosphate selectivity is rationalized by using results from MD simulations and free energy calculations for the R103A/R115A mutant of the ϵ subunit from Bacillus PS3, which binds ATP more strongly than the wild-type protein. Our results are in good agreement with experimental data, and the elucidated molecular basis for selectivity could help to guide the design of novel GTP sensors.

How Ligand Binding Affects the Dynamical Transition Temperature in Proteins.

The biochemical functions of proteins are activated at the protein glass transition temperature, which has been proposed to be dependent upon protein-water interactions. However, at the molecular level it is unclear how ligand binding to well-defined binding sites can influence this transition temperature. We thus report molecular dynamics (MD) simulations of the ϵ subunit from thermophilic Bacillus PS3 in the ATP-free and ligand-bound states over a range of temperatures from 20 to 300 K, to study the influence of ligand association upon the transition temperature. We also measure the protein mean square displacement (MSD) in each state, which is well established as a means to quantify this dynamical temperature dependence. We find that the transition temperature is largely unaffected by ligand association, but the MSD beyond the transition temperature increases more rapidly in the ATP-free state. Our data suggests that ligands can effectively "shield" a binding site from solvent, and hence stabilize protein domains with increasing temperature.

Extending the Martini Coarse-Grained Force Field to N-Glycans.

Glycans play a vital role in a large number of cellular processes. Their complex and flexible nature hampers structure-function studies using experimental techniques. Molecular dynamics (MD) simulations can help in understanding dynamic aspects of glycans if the force field parameters used can reproduce key experimentally observed properties. Here, we present optimized coarse-grained (CG) Martini force field parameters for N-glycans, calibrated against experimentally derived binding affinities for lectins. The CG bonded parameters were obtained from atomistic (ATM) simulations for different glycan topologies including high mannose and complex glycans with various branching patterns. In the CG model, additional elastic networks are shown to improve maintenance of the overall conformational distribution. Solvation free energies and octanol-water partition coefficients were also calculated for various N-glycan disaccharide combinations. When using standard Martini nonbonded parameters, we observed that glycans spontaneously aggregated in the solution and required down-scaling of their interactions for reproduction of ATM model radial distribution functions. We also optimized the nonbonded interactions for glycans interacting with seven lectin candidates and show that a relatively modest scaling down of the glycan-protein interactions can reproduce free energies obtained from experimental studies. These parameters should be of use in studying the role of glycans in various glycoproteins and carbohydrate binding proteins as well as their complexes, while benefiting from the efficiency of CG sampling.

Molecular dynamics simulations of bacterial outer membrane lipid extraction: Adequate sampling?

The outer membrane of Gram-negative bacteria is almost exclusively composed of lipopolysaccharide in its outer leaflet, whereas the inner leaflet contains a mixture of phospholipids. Lipopolysaccharide diffuses at least an order of magnitude slower than phospholipids, which can cause issues for molecular dynamics simulations in terms of adequate sampling. Here, we test a number of simulation protocols for their ability to achieve convergence with reasonable computational effort using the MARTINI coarse-grained force-field. This is tested in the context both of potential of mean force (PMF) calculations for lipid extraction from membranes and of lateral mixing within the membrane phase. We find that decoupling the cations that cross-link the lipopolysaccharide headgroups from the extracted lipid during PMF calculations is the best approach to achieve convergence comparable to that for phospholipid extraction. We also show that lateral lipopolysaccharide mixing/sorting is very slow and not readily addressable even with Hamiltonian replica exchange. We discuss why more sorting may be unrealistic for the short (microseconds) timescales we simulate and provide an outlook for future studies of lipopolysaccharide-containing membranes.

Aggregation of thrombin-derived C-terminal fragments as a previously undisclosed host defense mechanism.

Effective control of endotoxins and bacteria is crucial for normal wound healing. During injury, the key enzyme thrombin is formed, leading to generation of fibrin. Here, we show that human neutrophil elastase cleaves thrombin, generating 11-kDa thrombin-derived C-terminal peptides (TCPs), which bind to and form amorphous amyloid-like aggregates with both bacterial lipopolysaccharide (LPS) and gram-negative bacteria. In silico molecular modeling using atomic resolution and coarse-grained simulations corroborates our experimental observations, altogether indicating increased aggregation through LPS-mediated intermolecular contacts between clusters of TCP molecules. Upon bacterial aggregation, recombinantly produced TCPs induce permeabilization of Escherichia coli and phagocytic uptake. TCPs of about 11 kDa are present in acute wound fluids as well as in fibrin sloughs from patients with infected wounds. We noted aggregation and colocalization of LPS with TCPs in such fibrin material, which indicates the presence of TCP-LPS aggregates under physiological conditions. Apart from identifying a function of proteolyzed thrombin and its fragments, our findings provide an interesting link between the coagulation system, innate immunity, LPS scavenging, and protein aggregation/amyloid formation.

An Alternative HIV-1 Non-Nucleoside Reverse Transcriptase Inhibition Mechanism: Targeting the p51 Subunit.

The ongoing development of drug resistance in HIV continues to push for the need of alternative drug targets in inhibiting HIV. One such target is the Reverse transcriptase (RT) enzyme which is unique and critical in the viral life cycle-a rational target that is likely to have less off-target effects in humans. Serendipitously, we found two chemical scaffolds from the National Cancer Institute (NCI) Diversity Set V that inhibited HIV-1 RT catalytic activity. Computational structural analyses and subsequent experimental testing demonstrated that one of the two chemical scaffolds binds to a novel location in the HIV-1 RT p51 subunit, interacting with residue Y183, which has no known association with previously reported drug resistance. This finding supports the possibility of a novel druggable site on p51 for a new class of non-nucleoside RT inhibitors that may inhibit HIV-1 RT allosterically. Although inhibitory activity was shown experimentally to only be in the micromolar range, the scaffolds serve as a proof-of-concept of targeting the HIV RT p51 subunit, with the possibility of medical chemistry methods being applied to improve inhibitory activity towards more effective drugs.

Structure and subunit arrangement of Mycobacterial F1FO ATP synthase and novel features of the unique mycobacterial subunit δ.

In contrast to other prokaryotes, the Mycobacterial F1FO ATP synthase (α3:ß3:γ:δ:ε:a:b:b':c9) is essential for growth. The mycobacterial enzyme is also unique as a result of its 111 amino acids extended δ subunit, whose gene is fused to the peripheral stalk subunit b. Recently, the crystallographic structures of the mycobacterial α3:ß3:γ:ε-domain and c subunit ring were resolved. Here, we report the first purification protocol of the intact M. smegmatis F1FO ATP synthase including the F1-domain, the entire membrane-embedded FO sector, and the stator subunits b' and the fused b-δ. This enzyme purification enabled the determination of the first projected 2D- and 3D structure of the intact M. smegmatis F1FO ATP synthase by electron microscopy (EM) and single particle analysis. Expression and purification of the fused mycobacterial b-δ24-446 construct, excluding the membrane-embedded N-terminal amino acids, provided insight into its secondary structure. By combining these data with homology and ab-initio modeling techniques, a model of the mycobacterial peripheral stalk subunits b-δ and b' was generated. Superposition of the 3D M. smegmatis F-ATP synthase EM-structure, the α3:ß3:γ:ε and c-ring, and the derived structural models of the peripheral stalk enabled a clear assignment of all F-ATP synthase subunits, in particular with respect to the unique mycobacterial peripheral stalk subunit b' and the elongated δ fused with subunit b. The arrangement of δ relative to the N-termini of the catalytic α3ß3-headpiece and its potential as a drug target are discussed.