A CLEM study in biology aims at providing knowledge on the identity and localisation of cellular constituents of interest within a cell’s ultrastructure. The key limitation of such experiments, however, is the registration precision between light and electron microscopic data. Precise registration is usually achieved using fiducial markers, visible in both imaging modalities [1].

Integrated microscopes, i.e. light and electron microscopy are performed in one machine [2], in contrast, offer an inherent high-precision correlation. In such systems, photon emission from the sample, is either triggered by a light source (e.g. laser) or the electron beam. A glass objective lens, mounted below the sample collects the emitte light, that is ultimately detected by a photomultiplier tube or a camera [3].

This geometry is particularly advantageous for detecting a sample’s cathodoluminescence (CL) signal, since roughly 80% of the emitted photons are emitted into the forward direction [3], and it allows for simultaneous, unobstructed secondary electron (SE) imaging.

We evaluate a CL detector similar to the one described by Narváez et al. [3], with respect to its applicability to CL imaging of small synthetic fluorophores. As these fluorophores are readily damaged by the electron beam [4], we study the CL signal of these molecules dependent on primary energy and beam current. Based on the results of these experiments we image the CL signal of fluorescently labelled, resin embedded biological specimen. Our data indicate the feasibility of CL imaging for CLEM of biological specimen.

The CL signal of 200 nm sized fluorescently labelled polystyrene beads, scales with primary energy (figure 1). Increasing the beam current increases the CL intensity (figure 2A) of individual images. Comparing image series taken at different beam currents (240 pA (240 frames), 480 pA (120 frames) and 1440 pA (40 frames)), a dose rate effect on the cumulative retrievable intensity (figure 2B) is observed. The signal-to-noise ratio can be maximized by either repeated scanning of the same area on the sample at low beam currents or by short pixel dwell times at higher beam currents. In both cases the final image is obtained by averaging the acquired frames.

Analyses of the influence of beam current and primary energy on the CL signal were performed at a pixel dwell time of 6.4 µs. Cumulative intensities in figure 2 are presented as mean ± standard deviation of 4 different areas on the sample per condition.

Having established that CL imaging of organic fluorophores is feasible, we imaged DNA stained with 1 µM Sytox® Green in mammalian cells. Figure 3 shows CL and SE images of 100 nm sections, of LR White embedded Hela cells before and after staining, deposited on ITO (Indium Tin Oxide) coated cover slips. CL signal of stained cell nuclei was detected at different magnifications. Following the imaging guidelines established for fluorescently labelled beads, CL of stained DNA was recorded by several fast scans (100 ns pixel dwell time) of the same position and averaging of the individual frames. Sections were imaged at 1 kV (unstained cells) and 2 kV (stained cells) primary energy, respectively, and a beam current of 480 pA.

Currently we investigate, whether low temperatures (120 K) increase the beam stability of the CL signal. Another factor we currently address is the background signal in images of resin embedded samples (Figure 3), which results from photon emission upon electron beam excitation of the resin or the glass substrate.

References:

[1] Kukulski et al., J. Cell Biol. 192 (2011), p. 111

[2] Peddie et al., Ultramicroscopy 143 (2014), p. 3

[3] Narváez et al., Opt. Express 21 (2013), p. 29968

[4] Niitsuma et al., J Electron Microsc 54 (2005), p. 325

[5] Our work is funded by the German Federal Ministry of Education and Research (13GW0044).

[6] We thank Marina van Ark for technical assistance.

Figures:

Figure 1: Increasing the acceleration voltage increases the recorded CL intensity of 200 nm sized fluorescently labelled polystyrene beads. CL intensity plateaus at acceleration voltages larger than 2 kV. Beads were image at a beam current of 480 pA and a pixel dwell time of 6.4 µs.

Figure 2: High beam currents produce high CL intensities, but lead to rapid bleaching of the signal. A) CL intensity is increased by increasing the beam current. B) Repeated scanning of the same area at low beam currents yields the highest cumulative intensities, suggesting that the bleaching is dose-rate dependent.

Figure 3: CL and SE images of ultra-thin sectioned, resin embedded mammalian cells before and after staining with Sytox® Green. Stained cells show strong CL signal in the nucleus. 50 fast scanned frames (pixel dwell time 100 ns) of stained cells were averaged to produce the final image.

« Back to The 16th European Microscopy Congress 2016

EMC Abstracts - https://emc-proceedings.com/abstract/cathodoluminescence-microscopy-of-biological-samples-for-correlative-light-and-electron-microscopy-clem-using-organic-fluorophores/