Coupling of Proton Transfer and Multidrug Expulsion in the Inner Membrane Translocase AcrB

Abstract

With the introduction of the first antibiotics in the 1940s, lethal bacterial infections like tuberculosis or pulmonary inflammation were widely believed to be defeated. However, bacteria inherently hold a variety of effective resistance mechanisms enabling them to rapidly adapt to new environmental conditions. As a consequence, the wide spread use of of antibiotics simultaneously has created selective pressure for antibiotic resistances in bacterial strains. In multidrug resistance (MDR), whereby a host displays immunity to ≥ 3 antibacterial drug classes, the active efflux of compounds by MDR efflux transporters plays a dominant role. In Escherichia coli, the AcrAB-TolC complex is one of the best studied MDR efflux transporters causing resistances against a broad range of structurally and chemically unrelated compounds. Spanning the entire periplasmic space, drugs are captured directly from periplasm and subsequently expelled to the outer medium before they can reach lethal concentrations or pass the inner membrane. The tripartite complex consists of the inner membrane translocase AcrB, the outer membrane factor TolC, and up to six AcrA adaptor proteins. The AcrA molecules not only stabilize the complex formation but significantly enhance AcrB pump activity. The homotrimer AcrB constitutes the engine of the complex using the proton gradient over the inner membrane to drive the conformational changes necessary for substrate translocation. In this work we applied atomistic molecular dynamics (MD) simulations of membrane-embedded asymmetric AcrB focusing on two questions: 1) How are protons transported through AcrB's transmembrane domain? 2) How are substrates recruited and transported by AcrB's porter domain?<br /> To date, only five residues of the transmembrane domain (TMD) are known to be essential for AcrB proton transfer rendering the pump non-functional when mutated to alanine. Located centrally in the proton-transferring transmembrane domain they are believed to be directly involved in proton translocation. As internal waters are essential for proton transfer through proteins, we calculated TMD-internal water distributions. Based on the resulting average hydration pattern we detected three possible routes of proton transfer through AcrB's TMD. We find that the water accessibility of the transduction routes is regulated by four groups of gating residues in a combination of residue side-chain re-orientations and shifts of transmembrane helices. Furthermore, we identified new key residues candidates by the quantification of each residue's frequency of water hydrogen contact. The existence of alternative proton pathways could explain the robustness of the transporter against substitutions of other TMD residues beside the known key residues. <br /> In the second part of this work we analyzed the dynamics of the substrate binding and transporting porter domain addressing the question why all available X-ray structures display nearly identical porter domain conformations within the same reaction cycle intermediate. The porter domain can be divided into the substrate channel entry, a central phenylalanine-rich hydrophobic binding pocket which is open only in the B-intermediate state, and the porter domain exit. We find that the porter domain is more flexible than previously assumed displaying clear opening and closing motions of its entrance and exit regions. Concurrently we observe a predominantly closed hydrophobic binding pocket conformation in all monomers and all simulations. Exploring currently unreported conformations in our simulations, we propose that the structural similarity between the crystal structures is caused by bound but structurally unresolved buffer or detergent molecules. Additionally, if our simulations are correct and the opening and closing motions of the entry region inhibits drug binding in vivo, the structural changes provide a reasonable explanation for the AcrB activity enhancing effect of AcrA by stabilizing the open and thus drug-accessible conformation of the porter domain. With no evidence for a transition from an ABC to an AAA "resting state"-like conformation during 100 ns simulation time, our findings further support the existence of an asymmetric AcrB conformation also in the absence of substrates for which our simulations provide the first insight. <br /> Addressing the question of porter domain dynamics in presence of substrate we finally added 25 hexane molecules to the simulation system. We observed binding events to three of the four proposed substrate binding sites supporting their function in protein/drug interaction. Furthermore, hexane passage through one of the suggested alternative transport tunnels connecting the transmembrane domain with the porter domain, confirms the proposed substrate capture from the membrane bilayer. The fact that hexane bindings to the hydrophobic binding pocket in the B-intermediate states restored its open, crystal structure-like conformation, strengthens our assumptions that the high structural homogeneity displayed by the crystal structures is caused by bound but unresolved substrates and the existence of an asymmetric conformation in vivo.