Super-Resolution Fluorescence Microscopy of Synaptic Plasticity on Unmodified Brain Slices in Health and Tauopathy

Neuronal communication occurs at intercellular junctions called synapses, which can dynamically strengthen and weaken — termed synaptic plasticity. While synaptic plasticity underlies learning and memory, abnormal plasticity is associated with synapse loss and memory decline. Synaptic plasticity is expressed, in part, by

changes in the level of glutamate-gated AMPA receptors (AMPARs). The distance scale is ~10-25 nm, and thus, ~10 nm resolution is needed to ascertain structures like nanodomains and nanocolumns. Measuring such changes at the nanometer level in native brain slices is difficult due to the extraordinary high density of cells and

proteins, particularly in the hippocampus. While many super-resolution fluorescence microscopy (SRFM) tech- niques (most commonly dSTORM) exist to image AMPARs in dissociated neuronal culture, very few have been applied with high resolution to brain slices: the ones that do, look only near the edge (a few µm deep) or use

knock-in mice of epitope-tagged subunits. In addition, methods to decrowd proteins, such as expansion and clearing, involve extensive tissue manipulation. As a result, SRFM on unmodified brain slices is considered the ‘gold standard’. The goal of this technology proposal (PAR-22-127) is to develop new forms of SRFM that can

isolate native surface AMPARs in both synaptic and non-synaptic domains during synaptic plasticity on unmod- ified brain slices as a function of health and tauopathy that cause memory loss. To identify synaptic vs non- synaptic AMPARs on the cell membrane necessitates multi-color SRFM techniques to define the spatial resolu-

tion of synaptic proteins surrounding surface AMPARs. In our unpublished work, we selectively labeled native AMPAR subunits GluA2-4 on the cell surface of live 200 µm thick brain slices using a small chemical labeling agent called CAM2-Alexa647. After fixation and sectioning to ~30 µm, the slices were labeled on the postsynaptic

protein, Homer1, with a second SRFM dye (CF568). AMPAR and Homer1 were then imaged using two-color 3D dSTORM with an aberration corrected microscope and deformable mirrors, resulting in 20x20x90 nm 3D reso- lution on a native brain slice. This has not previously been accomplished. In Specific Aim 1, we will extend this

work to achieve 3-color SRFM with ~ 11 x 11 x 40 nm resolution to determine AMPAR distances from Bassoon or RIM1 in the presynaptic terminal, thereby identifying synaptic vs non-synaptic AMPARs. Hyperphosphorylated tau will be labeled with a 4th SRFM color. A new self-interferometric SRFM technique, called SELFI, along with

3D-dSTORM, will also be used with the goal of simplifying the optics. In Specific Aim 2, we aim for an improved

resolution (< 10 nm) by serially slicing the native brain slices into “thin” sections (0.7~4 µm) and digitally recom- bining their images to the original thickness. Equipped with new cameras, lasers, fluorescent dyes, and software, we expect a 30-100-fold speed improvement, possibly extending 3D dSTORM and SELFI to other parts of the

brain. We will validate these techniques in well-established systems for surface AMPAR change including hip- pocampal synaptic plasticity, fear-induced learning, and tauopathy.