34 doi:10.1017/S1551929521001073 www.microscopy-today.com • 2021 September Cryo-Confocal Imaging for CLEM Mapping in Brain Tissues Connon I. Thomas, Nicolai T. Urban, Ye Sun, Lesley A. Colgan, Xun Tu, Ryohei Yasuda, and Naomi Kamasawa* Max Planck Florida Institute for Neuroscience, Jupiter, FL 33458 *naomi.kamasawa@mpf.org Abstract: In correlative light and electron microscopy (CLEM) workfows, identifying the same sub-cellular features in tissue by both light (LM) and electron microscopy (EM) remains a challenge. Furthermore, use of cryo-fxation for EM is desirable to capture rapid biological phenomena. Here, we describe a workfow that incorporates cryo-confocal laser scanning microscopy into the CLEM process, mapping cells in brain slices to re-image them with serial section scanning electron microscopy (ssSEM) array tomography. The addition of Airyscan detection increased the signal-to-noise ratio (SNR), allowing individual spines in thick frozen tissue to be visualized at a suf fcient spatial resolution, providing a new tool for a CLEM approach to capture biological dynamics. Keywords: CLEM, cryo-confocal, 2-photon LM, high-pressure freez- ing, SEM array tomography Introduction One method for studying live cellular events in the brain is to section fresh brain tissue and culture it in a solution of artifcial cerebrospinal fuid. Using an organotypic slice culture preparation [1], neural activity can be optically driven and observed using light microscopy (LM), and the ultrastructure of neurons can be observed with electron microscopy (EM). At the intersection of these two techniques is correlative light and electron microscopy (CLEM), which allows for a comprehensive investigation of the mechanisms behind neural plasticity. One subcellular target involved in neural plasticity is the dendritic spine, where about 90% of the excitatory synapses in the brain are located [2]. Repetitive stimulation of glutamate receptors at the synapse results in rapid and sustained functional and structural changes of the dendritic spines. Tis structural plasticity of dendritic spines is thought to be the basis of learning and memory and has been extensively characterized [3–5]. Following excitation, the tissue can be fxed and observed using CLEM to characterize ultrastructural changes induced in the spine [6–7]. Since initial structural changes in the dendritic spine start on a short time scale ( ∼several seconds), capturing the process of structural plasticity requires rapid fxation of the tissue. Chemical fxation requires minutes to hours for complete fxation, but high- pressure freezing can be applied to halt structural changes, even in samples too thick for traditional freezing methods [8], with tight temporal control (that is, 2–3 minutes afer stimulation). From this point, tissue can be processed for EM via freeze-substitution. We integrated immunogold labeling in the process to support the CLEM workfow. However, even with such labeling, it can be difcult to fnd the cell of interest without a correlative map of the slice, especially when using frozen tissue. Cryo-confocal laser scanning microscopy (cryo-CLSM) provides a way to overcome these problems and to image vitri- fed samples with confocal microscopy (CLSM) at liquid nitrogen (LN 2 )-temperatures. Under cryogenic conditions, fuorescence is preserved and bleaching reduced, allowing for imaging of fro- zen samples [9]. Cryo-CLSM-guided CLEM has been used suc- cessfully for vitrifed samples of microorganisms [10] and for lamellar preparations from FIB-SEM lif-out experiments with cultured cells on grids [11]. It has not yet been applied to brain tissue slices [12]. We integrated this technology into the freeze- substitution workfow to image a 2-photon glutamate-uncaged and cryo-fxed organotypic slice culture from the hippocampus before preparation for serial section array tomography scanning electron microscopy (ssSEM). Cryo-confocal images provided us an overview of the tissue and location of fuorescent neurons. Using the enhanced sensitivity of Airyscan detection technology (Carl Zeiss Microscopy, LLC), we were able to resolve the location of the dendritic spine that had undergone structural plasticity. Tis information was used during cell and dendrite correlation to identify the target spine in our volumetric EM data. Materials and Methods Instrumentation. Two-photon LM was performed on a custom-built microscope equipped with a Ti:sapphire laser (Figure 1A, for details see [6]). High-pressure freezing of the cultured slice on a grid was done using a Leica HPM 100 with 4.6 mm carriers (Figures 1B, 1C). Cryo-CLSM was performed using an upright ZEISS LSM 980 equipped with a Linkam Cryo-Correlative Microscopy Stage (Figure 1D). A custom- made carrier adapter (Figure 1D, insert, Linkam Scientifc Instruments) was used to image the cultured tissue slice on the carrier. Freeze substitution was performed using a Leica EM AFS-2 (Figure 1E). Te embedded tissue was serially sectioned using an ATUMtome (RMC Boeckeler, Figure 1F). EM images were captured on a ZEISS Gemini 300 SEM equipped with a Gatan OnPoint ™ BSD detector (Figure 1G). Two-photon LM glutamate uncaging and high-pressure freezing. A 350 μm thick organotypic slice of mouse hippo- campus was cultured for 16 days on the top of an index-gold TEM grid (G200F1-Au, EMS) placed on the culture membrane insert. Green fuorescent protein (GFP) was expressed via a biolistic transfection. A pyramidal neuron expressing GFP fuorescence was imaged with 2-photon light microscopy using a 60× water immersion lens (LUMPlan FLN 60× 1.00, OLYM- PUS) with 1×–30× digital zoom to capture the target cell body, dendrite, and spine. Glutamate uncaging was performed on one single spine as previously described [7] to induce struc- tural plasticity. Subsequent spine growth was monitored for approximately 1 minute (Figure 1A). Te grid holding the slice was then transferred to the fat side of a 4.6 mm diameter gold specimen carrier (16770130, Leica Microsystems) lined with a ring of approximately 200 μm thick double-sided tape (Fig- ure 1B), covered in a thin layer of artifcial cerebrospinal fuid, Downloaded from https://www.cambridge.org/core. IP address: 54.82.56.168, on 26 Sep 2021 at 12:28:05, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S1551929521001073