Advances in high-resolution imaging--techniques for three-dimensional imaging of cellular structures

Abstract

A fundamental goal in biology is to determine how cellular organization is coupled to function. To achieve this goal, a better understanding of organelle composition and structure is needed. Although visualization of cellular organelles using fluorescence or electron microscopy (EM) has become a common tool for the cell biologist, recent advances are providing a clearer picture of the cell than ever before. In particular, advanced light-microscopy techniques are achieving resolutions below the diffraction limit and EM tomography provides high-resolution three-dimensional (3D) images of cellular structures. The ability to perform both fluorescence and electron microscopy on the same sample (correlative light and electron microscopy, CLEM) makes it possible to identify where a fluorescently labeled protein is located with respect to organelle structures visualized by EM. Here, we review the current state of the art in 3D biological imaging techniques with a focus on recent advances in electron microscopy and fluorescence super-resolution techniques.

Fig. 1.

3D EM tomography. To examine ER structure at high resolution, West et al. generated tomograms from 200-nm thick sections of yeast cells (West et al., 2011). (A) A two-dimensional EMT image shows the nuclear envelope (N, orange), Golgi (pink) and ER structures. Scale bar: 200 nm. (B) The reconstructed 3D model reveals the cisternae (yellow, CecER), tubular (green, TubER) ER and plasma-membrane-associated ER (blue, PmaER). Scale bar: 100 nm. Image reproduced from West et al., 2011 (West et al., 2011).

Fig. 2.

Demonstration of several techniques that produce images with resolution better than the confocal microscope. (A) 4pi microscopy uses two opposing lenses focused at the same plane to effectively increase the lens aperture, resulting in improved axial resolution (left) From Hell, 2007. Reprinted with permission from AAAS. The ~100-nm axial resolution of 4pi microscopy (right) reveals mitochondrial fragmentation that was not apparent in confocal microscopy images (center) (Dlasková et al., 2010). Note the lower resolution in the axial dimension (blue axis) as compared to the lateral dimensions. Images reprinted from Dlasková et al., 2010 with permission from Elsevier. (B) In SIM, a series of images are collected while scanning the illumination pattern across the sample (left) and are used to reconstruct an image with a lateral resolution improved by about a factor of two compared to confocal microscopy. The three images to the right show 3D SIM images of nuclear pore complex epitopes (red) with respect to the nuclear lamina (green). A nuclear stain (blue) is used as a reference. The resolution achieved was 100 nm and 300 nm in the lateral and axial directions, respectively. Single planes of 3D data sets are compared. The confocal image is on the left, the confocal plus mathematical reconstruction image is in the center and the 3D SIM on the right. From Schermelleh et al., 2008. Reprinted with permission from AAAS. (C) A thin ‘Bessel’ beam is swept through the sample while it is maintained in the focal plane of a second imaging objective (left). The resulting 300-nm isotropic resolution is combined with faster imaging speeds (50 two-dimensional frames per second) and dramatically reduced photo-bleaching as compared with that achieved with confocal imaging. Comparison of confocal microscopy (top center) and Bessel beam imaging (bottom center) that used two-photon excitation. The right panel shows one 3D frame of a 73-frame image series. Scale bar: 10 μm. (Planchon et al., 2011). Images reprinted by permission from Macmillan Publishers Ltd: Nat. Methods (Planchon et al., 2011), copyright 2011. (D) 3D SML-SR of microtubules (green) and mitochondria membrane (magenta). Localization precision was 30 nm laterally and 70 nm axially. Single planes of 3D data sets are shown. The diagram on the left demonstrates how the cylindrical lens encodes the z-position in the single molecule image (left). The three images on the right show comparisons between conventional wide field image (left), super-resolution x-y horizontal section (center), and super-resolution x-y vertical section (right) are shown. Scale bars are 500 nm. From Huang et al, 2008a. Reprinted with permission from AAAS. (E) The STED beam (orange) inhibits fluorescence outside of the center of the focal spot (left). From Hell, 2007. Reprinted with permission from AAAS. The right two images show 3D STED with 50-nm lateral and 100-nm axial resolution images of neurofilaments. The confocal image is in the center and 3D STED image on the right. Images reprinted with permission from Wildanger et al., 2009. Scale bar: 1 μm.

Fig. 3.

CLEM allows for localization of proteins to subcellular structures. (A–D) Localization of GFP-tagged Rvs167 to endocytic invagination. (A) Wide-field image of Rsv167–EGFP and Abp1–mCherry expressed in S. cerevisiae. Inset: GFP signal used for localization of the molecule (cross marks). (B) Image of the fiducial beads used for image correlation. (C) Corresponding fiducial markers in the EM tomograph. (D) High-resolution EM image taken from the tomographic reconstruction that reveals the presence of an endocytic invagination near the location of the Rvs167–GFP. Localization probability of 50% is designated by the inner dashed circle (33 nm) and outer circle of 80% (89 nm). (E–H) The improved resolution of PALM reveals the localization of α-liprin–tdEOS to the presynaptic dense projection. (E) Sum TIRF image of α-liprin–tdEOS, representing the typical resolution of the light microscope. (F) PALM image. (G) EM image. (H) Overlay of PALM and EM image. Scale bars: 1 μm (A–C), 50 nm (D), 500 nm (E). Images A–D reproduced from Kukulski et al. (Kukulski et al., 2011). Images E–H reprinted by permission from Macmillan Publishers Ltd: Nat. Methods (Watanabe et al., 2011), copyright 2011.