No abstract

Bacterial membranes are complex organelles composed of a variety of lipid types. The differences in their composition are a key factor in determining their relative permeabilities. The success of antibacterial agents depends upon their interaction with bacterial membranes, yet little is known about the molecular-level interactions within membranes of different bacterial species. To address this, we have performed molecular dynamics simulations of two bacterial membranes: the outer membrane of E. coli and the cell membrane of S. aureus . We have retained the chemical complexity of the membranes by considering the details of their lipidic components. We identify the extended network of lipid-lipid interactions that stabilize the membranes. Our simulations of electroporation show that the S. aureus cell membrane is less resistant to poration than the E. coli outer membrane. The mechanisms of poration for the two membranes have subtle differences; for the E. coli outer membrane, relative differences in mobilities of the lipids of both leaflets are key in the process of poration.

Bacterial outer membrane porins have a robust ␤-barrel structure and therefore show potential for use as stochastic sensors based on single-molecule detection. The monomeric porin OmpG is especially attractive compared with multisubunit proteins because appropriate modifications of the pore can be easily achieved by mutagenesis. However, the gating of OmpG causes transient current blockades in single-channel recordings that would interfere with analyte detection. To eliminate this spontaneous gating activity, we used molecular dynamics simulations to identify regions of OmpG implicated in the gating. Based on our findings, two approaches were used to enhance the stability of the open conformation by site-directed mutagenesis. First, the mobility of loop 6 was reduced by introducing a disulfide bond between the extracellular ends of strands ␤12 and ␤13. Second, the interstrand hydrogen bonding between strands ␤11 and ␤12 was optimized by deletion of residue D215. The OmpG porin with both stabilizing mutations exhibited a 95% reduction in gating activity. We used this mutant for the detection of adenosine diphosphate at the single-molecule level, after equipping the porin with a cyclodextrin molecular adapter, thereby demonstrating its potential for use in stochastic sensing applications.gating ͉ MD simulation ͉ OmpG ͉ stochastic sensor E ngineered protein pores can be used as stochastic sensors for single-molecule detection (1). The ionic current flowing through a pore under an applied potential is altered when an analyte binds within the lumen (1). Measurement of the frequency of occurrence of the binding events allows the determination of the concentration of an analyte, while the nature of the events (e.g., their amplitude or duration) aids in analyte identification. To date, studies of proteinaceous stochastic sensing elements have mostly focused on staphylococcal ␣-hemolysin (␣HL), a ␤-barrel pore-forming toxin (1, 2). However, the stoichiometry and symmetry of the heptameric ␣HL pore generate a large number of combinations and permutations when more than one type of subunit is used (3), which makes it difficult to fine-tune the properties of the pore. Therefore, it would be highly desirable to design a stochastic sensor based on a monomeric ␤-barrel.OmpG is a monomeric porin from the outer membrane of Escherichia coli that presents an attractive alternative to ␣HL for stochastic sensing (4-7). Previous studies have shown that OmpG undergoes pH-dependent, voltage-dependent, and spontaneous gating (4, 5). Specifically, the pore tends to close when the pH decreases to below 7 or when the voltage is higher than Ϯ100 mV (5). At neutral pH and an applied potential lower than Ϯ100 mV, OmpG exhibits a spontaneous gating behavior-i.e., it rapidly switches between open and closed states (Fig. 1A). Such spontaneous gating interferes with the application of a pore as a biosensor. Therefore, eliminating or significantly reducing the intrinsic gating activity of OmpG would be an important step in the development of an al...

Complete determination of a membrane protein structure requires knowledge of the protein position within the lipid bilayer. As the number of determined structures of membrane proteins increases so does the need for computational methods which predict their position in the lipid bilayer. Here we present a coarse-grained molecular dynamics approach to lipid bilayer self-assembly around membrane proteins. We demonstrate that this method can be used to predict accurately the protein position in the bilayer for membrane proteins with a range of different sizes and architectures.

Molecular dynamics simulations provide a route to studying the dynamics of lipid bilayers at atomistic or near atomistic resolution. Over the past 10 years or so, molecular dynamics simulations have become an established part of the biophysicist's tool kit for the study of model biological membranes. As simulation time scales move from tens to hundreds of nanoseconds and beyond, it is timely to re-evaluate the accuracy of simulation models. We describe a comparative analysis of five freely available force fields that are commonly used to model lipid bilayers. We focus our analysis on 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) bilayers. We show that some bilayer properties have a pronounced force field dependence, while others are less sensitive. In general, we find strengths and weaknesses, with respect to experimental data, in all of the force fields we have studied. We do, however, find some combinations of simulation and force field parameters that should be avoided when simulating DPPC and POPC membranes. We anticipate that the results presented for some of the membrane properties will guide future improvements for several force fields studied in this work.