Paul Walther (Ulm / DE), Nadia Wenske (Ulm / DE), Amanda Caroll (Ulm / DE), Deena D'Souza (Ulm / DE), Jens von Einem (Ulm / DE), Clarissa Read (Villinger) (Ulm / DE)
Abstract text (incl. figure legends and references)
Several approaches for volume scanning electron microscopy have been established during the last years. This article deals with array tomography by collecting SEM images of serial sections of high pressure frozen, freeze substituted and plastic embedded cultivated cells.
The objective of the study is to obtain TEM like resolution with an SEM on serial thin sections mounted on silicon chips.
Human cytomegalovirus (HCMV) infected human foreskin fibroblasts were grown on sapphire disks, high pressure frozen, freeze substituted with 0.2 % osmium tetroxide, 0.1 % uranyl acetate and 5 % of water in acetone and embedded in Epon. Serial thin sections (100 nm) were collected on silicon chips, partially post stained with lead citrate trihydrate (3% liquid) and imaged in a Hitachi S-5200 SEM at an accelerating voltage of 2 kV for recording the so-called composite signal, which consists of secondary electrons type 3 (that are produced by backscattered electrons) for intermediate magnifications (Fig. A), or with 5 kV accelerating voltage and recording the regular secondary electron signal for high magnifications (Figs. B and C).
When the sections were not poststained with lead, the secondary electron as well as the backscattered electron signal were very noisy. We first tried plasma cleaning to remove some of the Epon and to make the heavy metal better amenable, but in our hands, this did not significantly improve image quality. The breakthrough was post staining of the sections on the silicon chips with lead citrate. It was then possible to record 62 consecutive sections of cells infected with HCMV and make a three-dimensional reconstruction. Figure A shows a portion of one section at an intermediate magnification. Chromosomes, microtubules and the centriols are visible on this section.
Figures B and C show higher magnification images of an area of the cytoplasm with a viral particle in a vesicle. The images were recorded at 2 and at 5 kV. Best resolution was obtained at an accelerating voltage of 5 kV. The leaflets of the membrane bilayer are well visible and better resolved at 5 kV. These results fit well with our earlier studies of freeze substituted samples in FIB-SEM (Villinger et al., 2012, Histochem Cell Biol 138, 549-556)
With the presented approach it was possible to obtain very good resolution in X and Y of serial sections. All details visible in TEM on similar samples can also be seen in our SEM images. Best results were obtained after lead post staining at 5 kV accelerating voltage and using the secondary electron signal. The best accelerating voltage is a tradeoff of the smaller beam diameter and the higher penetration depth at 5 kV compared to 2 kV. The resolution in Z is obviously defined by the thickness of the sections and is 100 nm in the presented dataset.