Project Summary/Abstract Tremendous technological progress in the use of serial electron microscopy (EM) for brain circuit mapping has been made over the last decade, and it is possible to directly visualize synaptic connectivity in larger volumes of brain tissue than ever before. Meanwhile, advances in molecular biology have shed new light on the diversity among neurons, particularly with respect to their patterns of gene expression. Currently, there are no methods available to efficiently integrate molecular labels into serial EM reconstructions. The ability to distinguish many molecular cell type markers in serial EM volumes would greatly enhance our ability to study circuit function, neuronal diversity, and neuroplasticity, and to determine how these are affected in disease states. The major barrier to visualizing molecules in serial EM is that the methods used to preserve morphology and generate contrast for serial EM are incompatible with most labels. Several methods have been developed to accommodate this limitation, mainly by using genetic tools to introduce labels before tissue samples are prepared for EM. These approaches are restrictive in that only a few labels can be used in a single sample, genetic manipulation is required, and endogenous molecules cannot be localized. This project will develop approaches that allow many different molecular labels to be differentiated within a single serial EM tissue volume. An innovative strategy will be used: instead of working around the standard serial EM protocol by designing labels that are compatible with it, the focus will be on replacing the incompatible elements of the standard protocol. Circuit reconstruction by serial EM requires a high degree of morphological preservation, which is typically accomplished with harsh chemicals that damage and denature molecules. However, all that is fundamentally required to preserve morphology is to retain as many molecules in the tissue as possible, which is also necessary for molecular labeling. Therefore, tissue preservation protocols will be developed to minimize extraction of molecules without damage or irreversible denaturation. A combination of strategies will be employed, including novel combinations of chemical crosslinkers and embedding resins. These new approaches will offer a means of revealing valuable information about circuit organization and neuronal diversity that is presently inaccessible.