The genetic encoded toolbox for electron microscopy and connectomics

Targeting Functionally Characterized Synaptic Architecture Using Inherent Fiducials and 3D Correlative Microscopy

Brain circuits are highly interconnected three-dimensional structures fabricated from components ranging vastly in size; from cell bodies to individual synapses. While neuronal activity can be visualized with advanced light microscopy (LM) techniques, the resolution of electron microscopy (EM) is critical for identifying synaptic connections between neurons. Here, we combine these two techniques, affording the advantage of each and allowing for measurements to be made of the same neural features across imaging platforms. We established an EM-label-free workflow utilizing inherent structural features to correlate in vivo two-photon LM and volumetric scanning EM (SEM) in the ferret visual cortex. By optimizing the volume SEM sample preparation protocol, imaging with the OnPoint detector, and utilizing the focal charge compensation device during serial block-face imaging, we achieved sufficient resolution and signal-to-noise ratio to analyze synaptic ultrastructure for hundreds of synapses within sample volumes. Our novel workflow provides a reliable method for quantitatively characterizing synaptic ultrastructure in functionally imaged neurons, providing new insights into neuronal circuit organization.

Quintuple labeling in the electron microscope with genetically encoded enhanced horseradish peroxidase

Genetic encoded multilabeling is essential for modern cell biology. In fluorescence microscopy this need has been satisfied by the development of numerous color-variants of the green fluorescent protein. In electron microscopy, however, true genetic encoded multilabeling is currently not possible. Here, we introduce combinatorial cell organelle type-specific labeling as a strategy for multilabeling. First, we created a reliable and high sensitive label by evolving the catalytic activity of horseradish peroxidase (HRP). We then built fusion proteins that targeted our new enhanced HRP (eHRP) to three cell organelles whose labeling pattern did not overlap with each other. The labeling of the endoplasmic reticulum, synaptic vesicles and the plasma membrane consequently allowed for triple labeling in the EM. The combinatorial expression of the three organelle-specific constructs increased the number of clearly distinguishable labels to seven. This strategy of multilabeling for EM closes a significant gap in our tool set and has a broad application range in cell biology.