Molecular Dynamics Simulations Predict the Pathways via Which Pristine Fullerenes Penetrate Bacterial Membranes

Abstract

Carbon fullerenes are emerging as effective devices for different biomedical applications, including the transportation of nanosized drugs and the extraction of harmful oxidants and radicals. It has been proposed that fullerenes could be used as novel antibacterial agents, given the realization that the nanoparticles can kill pathogenic Gram-negative bacteria. To explore this at the molecular level, we simulated C 60 fullerenes with bacterial membranes using the coarse-grain (CG) molecular dynamics (MD) Martini force field. We find that pristine C 60 has a limited tendency to penetrate (incomplete core) Re mutant lipopolysaccharide (LPS) leaflets, but the translocation of C 60 fullerenes into (complete core) Ra mutant LPS leaflets is not thermodynamically favored. Moreover, we show that the permeability of Re LPS bilayers depends sensitively on system temperature, the charge of ambient ions, and the prevalence of POPE defect domains. The different permeabilities are rationalized in terms of transitory head group pore formation, which underpins the translocation of C 60 into the lipid core. The Re LPS lipids readily form transient micropores when they are linked with monovalent cations, or when they are heated to a high temperature. POPE lipids are shown to be particularly adept at forming these transient surface cavities, and their inclusion into Re LPS membranes facilitates the formation of particularly large pores that are tunneled by C 60 aggregates of significant size (~ 5 nm wide). After inserting into the lipid core, the aggregates dissociate, and the disbanded nanoparticles migrate to the interface between separate POPE and LPS domains, where they weaken the boundaries between the coexisting lipid fractions and thereby promoting lipid mixing.

Through the Lipopolysaccharide Glass: A Potent Antimicrobial Peptide Induces Phase Changes in Membranes

“…[66] The Re LPS bilayers were formed in previous work with the GROMACS genconf utility; each bilayer contained 100 Re LPS molecules per leaflet. [31] The Re LPS membranes (being void of any PMB1 peptides) were equilibrated for 10 µs at different temperatures (300 K, 310 K, and 320 K) after initial energy minimization with the steepest decent algorithm. Simulations were run in the NPT ensemble, as the pressure was coupled to 1 bar using a semiisotropic barostat and the Parrinello-Rahman algorithm (time constant 1 ps).…”

“…Polygon-based tessellation of the LPS phosphate groups revealed that the PMB1 peptides follow the same translocation mechanism predicted for C 60 nanoparticles into lipid bilayers. [31][32][33] The hydrophobic groups enter the bilayer core as the phosphate groups transiently form small micropores in the membrane surface. As the hydrophobic moieties enter these transient micropores they expand the local defects, enabling their gradual penetration into the bilayer core.…”

The gram‐negative outer membrane modeler: Automated building of lipopolysaccharide‐rich bacterial outer membranes in four force fields

“…With the Protein‐Membrane option, users can upload a protein structure in PDB format and build a membrane around it, while the Membrane‐Only option allows them to simply construct a lipid bilayer. Finally, users can select the Pre‐Processed PDB and Topology option, to upload a processed structure and force field topology of their own; this last option can be useful in cases where the simulation system contains elements not available in standard force field parameters, for example, modifications to the protein structure, rationally designed compounds or even inorganic polymer structures like carbon nanotubes or fullerenes …”

“…These include variations of Lipid A from four different species ( E. coli , Pseudomonas aeruginosa , Helicobacter pylori and Neisseria meningitidis ) (Supporting Information Fig. S1), as well as three experimentally studied LPSs, containing Lipid A and the oligosaccharide core (Re and Ra LPS for E. coli and PAO1 LPS for P. aeruginosa ) . In addition, a number of standard lipids are also offered, representing the major categories most commonly found in bacteria .…”

“…Some glycolipid and glycan modeling tools can also be used to model LPS molecules, such as doGlycans for the AMBER/GLYCAM and OPLS force fields; however, these methods are not able to construct LPS‐containing bilayers, either homogeneous or mixed. As such, apart from CHARMM‐GUI, there are very limited options to automatically create LPS‐containing membranes with force fields or molecules outside the aforementioned systems (e.g., GROMOS, or MARTINI with LPSs from species other than E. coli ), even though valid Lipid A and LPS parameters exist . Instead, the literature shows that LPS‐containing simulation systems in such cases are, for the most part, built by hand, a process that can be time consuming and difficult, particularly for inexperienced users.…”