De novo formation of the double-membrane compartment autophagosome is seeded by small vesicles carrying membrane protein autophagy-related 9 (ATG9), whose function remains unknown. Here we find that ATG9A scrambles phospholipids of membranes in vitro. Cryo-EM structures of human ATG9A reveal a trimer with a solvated central pore, which is connected laterally to the cytosol through the cavity within each protomer. Similarities to ABC exporters suggest that ATG9A could be a transporter that uses the central pore to function. Moreover, molecular dynamics simulation suggests that the central pore opens laterally to accommodate lipid headgroups, thereby enabling lipids to flip. Mutations in the pore reduce scrambling activity and yield markedly small autophagosomes, indicating that lipid scrambling by ATG9A is essential for membrane expansion. We propose ATG9A acts as a membrane-embedded funnel to facilitate lipid flipping and to redistribute lipids added to the outer leaflet of ATG9 vesicles, thereby enabling growth into autophagosomes.

Adaptive quantum‐mechanics/molecular‐mechanics (QM/MM) methods feature on‐the‐fly reclassification of atoms as QM or MM during a molecular dynamics (MD) simulation, allowing the location and contents of the QM subsystem to be dynamically updated as needed. Such flexibility is a distinct advantage over conventional QM/MM, where a ‘static’ boundary is retained between the QM and MM subsystems. The ‘dynamic’ boundary in adaptive QM/MM allows a finite‐size QM to sustain simulations with an arbitrary length of time. To ensure smooth transitions between QM and MM, the energy or forces are interpolated. Special treatments are applied so that artifacts are eliminated or minimized. Recent developments have shed light on the relationship between the adaptive algorithms that describe Hamiltonian and non‐Hamiltonian systems. Originally developed to model an ion solvated in bulk solvent, adaptive QM/MM has been enhanced in many aspects, including the treatment of molecular fragments in macromolecules, monitoring molecules entering/leaving binding sites, and tracking proton transfer via the Grotthuss mechanism. Because the size of the QM region can be set as small as possible in adaptive QM/MM, the computational costs can be kept low. Small QM subsystems also facilitate the utilization of high‐level QM theory and long simulation time, which can potentially lead to new insights. WIREs Comput Mol Sci 2017, 7:e1310. doi: 10.1002/wcms.1310 This article is categorized under: Electronic Structure Theory > Combined QM/MM Methods Molecular and Statistical Mechanics > Molecular Dynamics and Monte-Carlo Methods Software > Molecular Modeling

Adaptive quantum-mechanics/molecular-mechanics (QM/ MM) dynamics simulations feature on-the-fly reclassification of atoms as QM or MM continuously and smoothly as trajectories are propagated. This allows one to use small, mobile QM subsystems, the contents of which are dynamically updated as needed. In this work, we report the first adaptive QM/MM simulations of H + transfer through a biological channel, in particular, the protein EcCLC, a chloride channel (CLC) Cl − /H + antiporter derived from E. coli. To this end, the H + indicator previously formulated for approximating the location of an excess H + in bulk water was extended to include Cl − ions and carboxyl groups as H + donors/acceptors. Furthermore, when setting up buffer groups, a new "sushi-roll" scheme was employed to group multiple water molecules, ions, and titratable residues along the onedimensional channel for adaptive partitions. Our simulations reveal that the H + relay path, which consists of water molecules in the pore, a bound Cl − ion at the central binding site (Cl − cen ) of the protein, and the external gating residue E148, exhibits certain mobility within the channel. A two-stage journey of H + migration was observed: the H + moves toward Cl − cen and is then shared between Cl − cen and nearby water molecules in the first stage and departs from Cl − cen via nearly concerted transfer to protonate E148 in the second stage. Most of the simulated trajectories show the bound Cl − ion in the channel to be transiently protonated, a possibility that was previously suggested by experiments and computations. Comparisons with conventional QM/MM simulations revealed that both adaptive and conventional treatments yield similar qualitative pictures. This work demonstrates the feasibility of adaptive QM/MM in the simulations of H + migration through biological channels.

No abstract