Correlative fluorescence microscopy, transmission electron microscopy and secondary ion mass spectrometry (CLEM-SIMS) for cellular imaging

Abstract

Electron microscopy (EM) has been employed for decades to analyze cell structure. To also analyze the positions and functions of specific proteins, one typically relies on immuno-EM or on a correlation with fluorescence microscopy, in the form of correlated light and electron microscopy (CLEM). Nevertheless, neither of these procedures is able to also address the isotopic composition of cells. To solve this, a correlation with secondary ion mass spectrometry (SIMS) would be necessary. SIMS has been correlated in the past to EM or to fluorescence microscopy in biological samples, but not to CLEM. We achieved this here, using a protocol based on transmission EM, conventional epifluorescence microscopy and nanoSIMS. The protocol is easily applied, and enables the use of all three technologies at high performance parameters. We suggest that CLEM-SIMS will provide substantial information that is currently beyond the scope of conventional correlative approaches.

Fig 1. General overview of the workflow for CLEM-SIMS imaging.

The cells are cultured following standard protocols, and fluorescence labelling can be applied to address specific questions. The cells are then immobilized by high pressure freezing, followed by freeze substitution. After the polymerization of the resin, the blocks are sectioned and the slides are placed on electrically conductive and referenced grids. Once the thin-sections are placed on grids, the first imaging step is fluorescent microscopy, followed by TEM, and finally by nanoSIMS. When the same areas have been scanned/imaged by the three techniques, the images are processed for registration and are analyzed. Scale bars: 3 μm.

Fig 2. A large-view CLEM image.

A large field was imaged in both TEM and fluorescence, relying on the DAPI (blue) channel, which shows the uranyl acetate fluorescence, and on the mito-mCitrine channel (shown here in green). Scale bar: 10 μm. Boxed region 1 shows a higher zoom view on the light microscopy image as well as the correlated image with TEM of a cell marked by a white square in the overview image. Scale Bar: 2 μm. Boxed regions 2 and 3 show detailed images of CLEM and TEM in the regions of interest. Scale bar: 1 μm.

Fig 3. Detailed views of CLEM-SIMS in cell culture.

a) An individual cell was imaged in light microscopy, TEM, and nanoSIMS. The fluorescence channels show the uranyl acetate fluorescence (blue) and the MitoTracker fluorescence (magenta). The measured isotopes are indicated in the different nanoSIMS images. Nitrogen isotopes (¹⁴N, ¹⁵N) are measured as CN^- ions, while carbon isotopes (¹²C, ¹³C) are measured as C₂^- ions. ¹³C and ¹⁵N are rare isotopes, which explains the low intensity of the respective images. Scale bars: 2 μm. b) Higher zoom views of the areas marked by white squares in panel A. Correlations of fluorescence (light) microscopy (left panels), TEM (middle panels) and nanoSIMS (right panels, ¹⁴N) are shown. Scale bars: 500 nm.

Fig 4. An analysis of cultured cells with CLEM-SIMS shows different isotopic distribution in different cellular compartments.

a) CLEM-SIMS images of a cell labeled with MitoTracker (magenta). The overlay combines the TEM and fluorescence images. The arrowheads indicate several mitochondria not labeled by MitoTracker, but visible in TEM. Scale bars: 5 μm. The letters indicate mitochondria (m), granules (g), euchromatin (e), heterochromatin (h) and cytosol (c). b-d) NanoSIMS analysis of specific compartments in the cells, identified in the TEM images. The graphs show box plots from 34 to 89 cellular regions, from multiple analyzed cells. Ratios of different isotopes to the ubiquitous ¹²C (present in the cellular material and in the plastic resin) are shown. The middle line in the boxes indicates the median, while the boxes show the 25^th percentiles; the error bars show the 75^th percentiles, with outliers indicated by the symbols. Statistical differences were tested using Kruskal-Wallis tests, followed by Tukey post-hoc tests; * p<0.05; **p<0.01; ***p<0.001. b) ¹⁴N analysis. c) ³¹P analysis. d) ³²S analysis. e) A higher zoom view of mitochondria labeled with MitoTracker (magenta) or lacking the label (white arrowhead). Scale bar: 500nm. f) Analysis of the fluorescently-labeled and non-labeled mitochondria. The measurements compare 85 labeled mitochondria and 24 non-labeled ones. Statistical differences were tested using Mann-Whitney tests; ***p<0.001.