Louise Hughes¹, Cheng Cheng², Stuart Searle², Iain Anderson¹
1 Oxford Instruments, High Wycombe, UK
2 Connectomx, Didcot, UK
AIM: Test the feasibility of using EDS to segment 3D SEM data.
INTRODUCTION:
3D SEM data collected using SBFSEM or array tomography provides ultrastructural information across large volumes of cells and tissues at high resolution. A significant bottleneck in data analysis is manual segmentation of 3D structures. Selective staining and maximum intensity thresholding of data allow rapid examination of cells and organelles but still require the selection of regions of interest. EDS provides information on sample chemistry and can be used as a tool to generate regions of interest without manual segmentation.
What is EDS?
Energy dispersive X-ray spectrometry measures and maps the chemistry of samples by detecting characteristic X-rays emitted from a sample exposed to an electron beam. EDS can be used on stained or unstained biological samples at low kV, avoiding sample damage and optimising image resolutions.
3D EDS data collected during SBFSEM
The Katana (Connectomx) SEM stage-mounted microtome combined with an Ultim Extreme EDS detector (Oxford Instruments) was used to collect BSE and EDS data at 3kV of Lamprey skin continuously for 190 slices (100nm thickness each slice). EDS was collected every slice without significant beam damage to the sample. 3D volumes of EDS and BSE data were assembled and combined to show regions containing significant osmium signal.
EDS data was successfully collected in combination with SBFSEM and Array tomography.
Stains and native elements can be used as a segmentation tool to select biologically relevant regions of interest.
3D EDS of stains and native elements
Liver tissue stained with osmium, sectioned at 100nm and mapped using EDS at 4kV using array tomography in combination with the Ultim Extreme EDS detector. Nitrogen, osmium, and phosphorus can be assembled into 3D data identifying cells, lipid droplets, and nuclei, respectively.
EDS data used to create regions of interest based on chemistry
Images and EDS maps of Venus’ flytrap leaf glands, stained with zinc-iodide-osmium stain (ZIO), were collected using array tomography at 5kV using an Ultim Extreme detector. 3 regions based on distinctive chemistry were identified using AZtec AutoPhaseMap (Oxford Instruments). These phase maps were used as 3D masks to segment biological tissue from surrounding resin and to also identify vacuoles in Venus’ fly trap glands cells containing higher levels of zinc, osmium and silicon. Maximum intensity projection of regions segmented by applying masks was used to image tissue ultrastructure.
