I. Super-Resolution Optical Microscopy for Three Dimensions: iPALM & CLEM
iPALM: Majority of biological problems are 3-dimensional. However, most of the original super-resolution microscopy methods only had high resolution in lateral plane, and suffered substantially poorer resolution in axial (z-) direction. We developed an interferometric photoactivated localization microscopy (iPALM), the combination of photoactivated localization microscopy (PALM) with single-photon, simultaneous multiphase interferometry that provides sub-20-nm 3D protein localization with optimal molecular specificity. Watch a video of the personal story of how PALM was invented by Harald Hess and Eric Betzig.
CLEM: Correlative light and electron microscopy (CLEM) is attractive because it exploits two microscopy techniques that give very different and very complementary information. By combining light microscopy (LM) and electron microscopy (EM), one is able to achieve protein specific localization in the context of a global structure. We develop different correlative optical super-resolution and electron microscopy methods optimized for various types of biological problems and sample configurations.
II. High-Throughput Electron Microscopy for Three Dimensions: 3D FIB-SEM
3D FIB-SEM: Full neuronal circuit reconstruction demands high resolution and high throughput in all three dimensions. Focused Ion Beam Scanning Electron Microscopy (FIB-SEM) generates 3D images with isotropic voxels well below 10x10x10 nm3. However, the key barriers to wide adoption are the slower imaging speed and lack of long-term system stability, which in turn limit the maximum acquisition volume. Here we develop an innovative customized solution overcoming these barriers and paving the way for 3D FIB-SEM to become the mainstream imaging solution for connectomics. We have successfully sped up the image acquisition and improved the system reliability with numerous stabilizing control, monitoring, and automation techniques. These improvements have enabled the extension of the continuous imaged volume to more than 100x100x100 µm3 while maintaining 8x8x8 nm voxel resolution, with the system operating for several months seamlessly. Alternatively, image volume can be traded off for even finer resolution, for example a continuous volume of ~ 30x30x30 µm3 at 4x4x4 nm voxel resolution can be easily accommodated. We note that the ultimate volume size could be further extended with the aid of ultrathick partitioning. The greater ease of automated processing of such isotropic voxel data can improve the imbalance in connectomics studies pipeline among image acquisition, analysis, and the large human proof reading effort. Some exemplary data sets illustrating unique access to cell biology are also presented. The extended regime of fine resolution and total volume will arm researchers with a powerful new technique accelerating discoveries in connectomics and cell biology.