Three-dimensional architecture of actin filaments in Listeria monocytogenes comet tails

The intracellular bacterial pathogen Listeria monocytogenes is capable of remodelling the actin cytoskeleton of its host cells such that “comet tails” are assembled powering its movement within cells and enabling cell-to-cell spread. We used cryo-electron tomography to visualize the 3D structure of the comet tails in situ at the level of individual filaments. We have performed a quantitative analysis of their supramolecular architecture revealing the existence of bundles of nearly parallel hexagonally packed filaments with spacings of 12–13 nm. Similar configurations were observed in stress fibers and filopodia, suggesting that nanoscopic bundles are a generic feature of actin filament assemblies involved in motility; presumably, they provide the necessary stiffness. We propose a mechanism for the initiation of comet tail assembly and two scenarios that occur either independently or in concert for the ensuing actin-based motility, both emphasizing the role of filament bundling.Several pathogens, including Listeria monocytogenes, Shigella flexneri, and Rickettsiae, have developed means to hijack the actin cytoskeleton of their host cells to move inside the host’s cytosol and to spread from cell to cell (1, 2). The cytoplasmic comet tails assembled from actin and actin-interacting proteins propel the bacteria forward and form protrusions emanating from the cell surface, which then become engulfed by neighboring cells.Several studies using light microscopy and EM have attempted to visualize the supramolecular organization of Listeria cytoplasmic comet tails and protrusions (1–8). Despite advances in superresolution fluorescence microscopy, this technique has not yet resolved individual actin filaments in crowded environments. Furthermore, conventional EM applied to detergent-extracted and dehydrated samples suffers from artifacts or a complete collapse of the delicate cytoskeletal networks. Cytoplasmic comet tails were reported to consist of multiple short actin filaments forming a cross-linked and branched network (1, 2) with some degree of alignment at their periphery (7). Branching occurs at the bacterial surface through the interaction of the bacterial surface protein ActA with the Arp2/3 complex (9, 10), which nucleates daughter filaments at an angle of 70° from preexisting filaments (10–12). Isolated Listeria protrusions were described as containing bundles of long, axial filaments, interspersed by short, randomly oriented filaments (3). To date, the detailed molecular architecture of these networks, which is key to understanding actin-based motility, has remained elusive.We examined Listeria comet tails, stress fibers, and filopodia, in their native cellular environment using cryo-electron tomography (CET). CET combines the power of 3D imaging with a close-to-life preservation of cellular structures (13). We cultivated epithelial Potoroo kidney Ptk2 cells on EM grids and infected them with Listeria monocytogenes (SI Text). Uninfected, as well as infected, cells were subjected to plunge-freezing (13), and tomographic datasets of filopodia, stress fibers, and Listeria cytoplasmic comet tails near the periphery of cells or in protrusions were recorded (SI Text).For the interpretation of the tomograms, we applied an automated segmentation algorithm developed specifically for tracking actin filaments (14). Unlike manual segmentation, automated segmentation is fast and unbiased. Due to the low signal-to-noise ratio of CET data, we applied the algorithm conservatively, which tends to underestimate the frequency of branching and the length of filaments. Furthermore, short filaments (<100 nm in comet tails and <70 nm in stress fibers and filopodia) were excluded from the analysis to avoid false-positive results. Tomograms of five Listeria cytoplasmic comet tails (Fig. 1A and Figs. S1A and S2A), nine Listeria protrusions (Fig. 2A and Figs. S3B and S4 A and D), eight stress fibers (Fig. 3C and Fig. S5 A and D), and four filopodia (Fig. 3A and Fig. S6 A and F) were subjected to segmentation. The localization of individual filaments and the analysis of their local neighborhood were used to describe quantitatively the architecture of the networks.Open in a separate windowFig. 1.Hollow Listeria cytoplasmic comet tail contains closely packed parallel filaments. (A) Slice through the tomogram of a cytoplasmic comet tail (tail) in a PtK2 cell (cytoplasm is indicated by cyt) infected by Listeria (b). (Scale bar: 200 nm.) Distribution of XY-filaments (B) and Z-filaments (C) in the XZ plane, projected over the Y axis. The color scale ranges from high occurrence (red) to low occurrence (blue) (same color code in all relevant panels). (D) Two-dimensional histogram of interfilament distances, weighted by the distance, and relative orientations between the filaments. deg, degrees. (E) Two-dimensional histogram of the (ξ, ζ) coordinates of the neighboring filaments in an XY-bundle in the local plane perpendicular to the central filament (dark gray, drawn to scale). (F) XY-filaments projected into the XY plane. The color of the filaments corresponds to their angle with respect to the Y axis: 0–15° (blue), 15–30° (green), 30–45° (red). The cell wall of the bacterium is shown in gray. (G) XY-pairs of parallel filaments (black) among XY-filaments (orange).Open in a separate windowFig. 2.Listeria protrusion shows hexagonal bundles in the vicinity of the plasma membrane. (A) Slice through the tomogram of a protrusion formed by Listeria at the surface of a PtK2 cell. The electron micrograph is shown in Fig. S3A. (Scale bar: 200 nm.) Distribution of XY-filaments (B) and Z-filaments (C) in the XZ plane, projected over the Y axis. (D) Two-dimensional histogram of interfilament distances, weighted by the distance, and relative orientations between the filaments. (E) Two-dimensional histogram of the (ξ, ζ) coordinates of the neighboring filaments in an XY-bundle in the local plane perpendicular to the central filament (dark gray, drawn to scale). (F) XY-filaments projected into the XY plane (same color code as in Fig. 1F). The plasma membrane of the protrusion is shown in gray. (G) XY-pairs of parallel filaments (black) among XY-filaments (orange).Open in a separate windowFig. 3.Filopodia and stress fibers contain hexagonal bundles. Slices through the tomograms of a filopodium (A) and a stress fiber (C) of a PtK2 cell. (Scale bar: 200 nm.) (B and D) XY-pairs of parallel filaments (black) among XY-filaments (orange). The plasma membrane is shown in gray. (E and G) Two-dimensional histogram of interfilament distances, weighted by the distance, and relative orientations between the filaments. (F and H) Two-dimensional histogram of the (ξ, ζ) coordinates of the neighboring filaments in an XY-bundle in the local plane perpendicular to the central filament (dark gray, drawn to scale).We found that many filaments in the comet tails, stress fibers, and filopodia are organized into bundles with spacings between 12 and 13 nm. Moreover, we discovered an arrangement in which filaments are assembled into hexagonal close-packed arrays. Based on our results, we propose a mechanism for the initiation of comet tail assembly and two scenarios enabling Listeria motility, both illuminating the role of filament bundling.