Mechanistic picture for conformational transition of a membrane transporter at atomic resolution

Abstract

During their transport cycle, ATP-binding cassette (ABC) transporters undergo large-scale conformational changes between inward-and outward-facing states. Using an approach based on designing system-specific reaction coordinates and using nonequilibrium work relations, we have performed extensive all-atom molecular dynamics simulations in the presence of explicit membrane/solvent to sample a large number of mechanistically distinct pathways for the conformational transition of MsbA, a bacterial ABC exporter whose structure has been solved in multiple functional states. The computational approach developed here is based on (i) extensive exploration of system-specific biasing protocols (e.g., using collective variables designed based on available lowresolution crystal structures) and (ii) using nonequilibrium work relations for comparing the relevance of the transition pathways. The most relevant transition pathway identified using this approach involves several distinct stages reflecting the complex nature of the structural changes associated with the function of the protein. The opening of the cytoplasmic gate during the outwardto inward-facing transition of apo MsbA is found to be disfavored when the periplasmic gate is open and facilitated by a twisting motion of the nucleotide-binding domains that involves a dramatic change in their relative orientation. These results highlight the cooperativity between the transmembrane and the nucleotidebinding domains in the conformational transition of ABC exporters. The approach introduced here provides a framework to study large-scale conformational changes of other membrane transporters whose computational investigation at an atomic resolution may not be currently feasible using conventional methods.orientation quaternion | bias-exchange umbrella sampling | conformational free energy A TP-binding cassette (ABC) transporters constitute one of the largest and most ancient protein superfamilies that are involved in diverse cellular processes (e.g., multidrug resistance) by actively transporting a wide variety of substrates (e.g., cytotoxic agents) across the cellular membrane (1). Their general architecture is characterized as a complex with two nucleotidebinding domains (NBDs) and two transmembrane domains (TMDs) (2). The NBDs dimerize when bound to ATP and dissociate in the absence of a bound nucleotide or in response to ATP hydrolysis (3, 4). The ATP-driven dimerization/dissociation of the NBDs is coupled to the conformational changes of the TMDs that facilitate the vectorial movement of the substrates across the membrane by alternating between "inward-facing" (IF) and "outward-facing" (OF) states, open toward the cytoand periplasm, respectively. Although evidence supports an IF ↔ OF transition during the transport cycle (4, 5) ("alternatingaccess" mechanism), the conformational changes involved in this transition are not fully understood in a mechanistic context.The resolved crystal structures of ABC exporters (5-10) reveal a great structural flexibility, particularly...

The Dynamic Behavior of the P2X4 Ion Channel in the Closed Conformation

“…Instead, a defined sequence of conformational changes, occurring at different parts of the protein, is required. Recently, an elegant and detailed computational study was able to reveal such a sequence of conformational changes for a membrane transporter (42). In this context, we wondered what interactions restrained the interchain movements in the absence of the intrachain deformations.…”

Investigating the dynamic nature of the ABC transporters: ABCB1 and MsbA as examples for the potential synergies of MD theory and EPR applications

“…A low-energy path for interconversion between the conformations could also be identified. The free energy of the transition from the inward-open to the inward-occluded conformation indicated that the conformation as observed in the crystal structure of nucleotide free MsbA is not the most stable conformation of the membrane-inserted transporter [70,72,73]. The energy of the NBD associated inward-closed state was only slightly higher than the free energy of the global minimum of slightly separated NBDs.…”

“…Results from targeted MD or steered DM in isolation are difficult to interpret at simulations length significantly below 100 ns, because the conversion is very fast, thereby not giving the transporter enough time to react, increasing the danger of producing artefacts. The targeted MD results could not be reproduced using steered MD simulations using a much slower rate constant [70]. Unbiased simulations [71] of 100 ns, starting from the outward facing conformation, showed that the transporter was conformationally dynamic at both the extracellular site and at the level of the NBDs in the apo state.…”

“…If carried out properly, MD simulations can lead to more detailed insights, because the positions of all atoms and the associated forces and energies are known. The path of the structural conversion from the outward-facing conformation to the inward-facing conformation has been investigated using a complex and system-specific biasing protocol [70,72] that used predefined biasing directions derived from available crystal structures. This approach identified a sequence of conformational changes consisting of periplasmic and cytoplasmic opening and closing of the TMD, NBD separation and the twisting motion of TMD vs. NBD.…”

“…Moore's law predicts that the computer power double every 17-18 months, indicating that it might be possible to carry out these unbiased simulations in 10 years from now. If the path of the conformational change can be defined correctly, then free energy path methods as the Potential of Mean Force calculations [79] (see above simulations of MsbA [70,72,73]) can derive the conformational ensemble from the energy profile along the reaction path. The information is equivalent with respect to a Boltzmann weighted distribution of protein conformations necessary to compare to EPR data, but requires much less computational effort.…”