Energy Dispersive X-ray Spectrometry in Combination with 3D SEM Facilitates the Identification and Segmentation of Cells and Organelles

Abstract

An ongoing challenge in the field of three-dimensional (3D) scanning electron microscopy (SEM) of biological materials is the identification and subsequent segmentation of datasets rich in complex information. Techniques such as serial block face SEM (SBFSEM) and array tomography produce increasingly large amounts of data and a major resource bottleneck is the amount of research time used for processing and analysing this information (Peddie and Collinson, 2014). Selective staining techniques can aid in rapid analysis by enabling researchers to threshold data based on stain intensity (Kittelmann et al., 2016), but there are a limited range of stains that can specifically target regions or organelles of interest and the technique is hampered by the inclusion of artefacts at similar greyscale thresholds. The aim of this investigation was to test if elemental analysis using energy dispersive X-ray spectrometry (EDS) could be used to identify endogenous elements and stains in biological samples and separate them from common artefacts. Array tomography and SBFSEM was used to collect 3D data of several model systems, including animal and plant tissue, prepared using standard fixation, staining and embedding techniques for electron microscopy. Array tomography and SBFSEM was used to collect BSE images and EDS maps of tissues, cells and organelles. Serial sections were cut at a thickness of 100nm and collected onto carbon coated Kapton tape for array tomography. Blocks of resin embedded samples were serially sectioned using SBFSEM at a thickness of 100-150nm. EDS was used to generate element-based maps, which were used as masks and applied to 3D datasets. EDS maps enabled fast identification of regions of interest, helped to remove imaging artefacts and enabled rapid volumetric and quantitative analysis compared to the equivalent manual segmentation.