Cofilin is best known for its ability to sever actin filaments and facilitate cytoskeletal recycling inside of cells, but at higher concentrations in vitro, cofilin stabilizes a more flexible, hyper-twisted state of actin known as “cofilactin”. While this filament state is well studied, a structural role for cofilactin in dynamic cellular processes has not been observed. With a combination of cryo-electron tomography and fluorescence imaging in neuronal growth cones, we observe that filopodial actin filaments switch between a fascin-linked and a cofilin-decorated state, and that cofilactin is associated with a variety of dynamic events within filopodia. The switch to cofilactin filaments occurs in a graded fashion and correlates with a decline in fascin cross-linking within the filopodia, which is associated with curvature in the bundle. Our tomographic data reveal that the hyper-twisting of actin from cofilin binding leads to a rearrangement of filament packing, which largely excludes fascin from the base of filopodia. Our results provide mechanistic insight into the fundamentals of cytoskeletal remodeling inside of confined cellular spaces, and how the interplay between fascin and cofilin regulates the dynamics of searching filopodia.

Cryo-electron tomography (cryo-ET) allows researchers to image cells in their native, hydrated state at the highest resolution currently possible. The technique has several limitations, however, that make analyzing the data it generates time-intensive and difficult. Hand segmenting a single tomogram can take from hours to days, but a microscope can easily generate 50 or more tomograms a day. Current deep learning segmentation programs for cryo-ET do exist, but are limited to segmenting one structure at a time. Here, multi-slice U-Net convolutional neural networks are trained and applied to automatically segment multiple structures simultaneously within cryotomograms. With proper preprocessing, these networks can be robustly inferred to many tomograms without the need for training individual networks for each tomogram. This workflow dramatically improves the speed with which cryo-electron tomograms can be analyzed by cutting segmentation time down to under 30 min in most cases.Further, segmentations can be used to improve the accuracy of filament tracing within a cellular context and to rapidly extract coordinates for subtomogram averaging.

Deep learning excels at cryo-tomographic image restoration and segmentation tasks but is hindered by a lack of training data. Here we introduce cryo-TomoSim (CTS), a MATLAB-based software package that builds coarse-grained models of macromolecular complexes embedded in vitreous ice and then simulates transmitted electron tilt series for tomographic reconstruction. We then demonstrate the effectiveness of these simulated datasets in training different deep learning models for use on real cryotomographic reconstructions. Computer-generated ground truth datasets provide the means for training models with voxel-level precision, allowing for unprecedented denoising and precise molecular segmentation of datasets. By modeling phenomena such as a three-dimensional contrast transfer function, probabilistic detection events and radiation-induced damage, the simulated cryo-electron tomograms can cover a large range of imaging content and conditions to optimize training sets. When paired with small amounts of training data from real tomograms, networks become incredibly accurate at segmenting in situ macromolecular assemblies across a wide range of biological contexts.

Cryo-electron tomography (cryo-ET) allows researchers to image cells in their native, hydrated state at the highest resolution currently possible. The technique has several limitations, however, that make analyzing the data it generates time-intensive and difficult. Hand segmenting a single tomogram can take from hours to days, but a microscope can easily generate 50 or more tomograms a day. Current deep learning segmentation programs for cryo-ET do exist, but are limited to segmenting one structure at a time. Here, multi-slice U-Net convolutional neural networks are trained and applied to automatically segment multiple structures simultaneously within cryotomograms. With proper preprocessing, these networks can be robustly inferred to many tomograms without the need for training individual networks for each tomogram. This workflow dramatically improves the speed with which cryo-electron tomograms can be analyzed by cutting segmentation time down to under 30 min in most cases.Further, segmentations can be used to improve the accuracy of filament tracing within a cellular context and to rapidly extract coordinates for subtomogram averaging.

No abstract