Charting the molecular landscape across the mammalian hippocampus using electron cryotomography
Authors
Jake Smith (1, 2), Calina Glynn (2), Matthew Case (2), Rebecca Csöndör (2, 3), Ana Katsini (2), Maria Sanita (1, 2), Thomas Glen (2), Avery Pennington (2), Michael Grange (1, 2)
Affiliations
- Division of Structural Biology, University of Oxford
- Rosalind Franklin Institute
- University College London
Summary
We outline a technical workflow to image brain tissue sections by electron cryo-tomography (cryoET). We demonstrate reproducible cryo-preservation of brain sections and high-throughput lamellae fabrication using plasma-sourced focused ion beam (PFIB) milling to interrogate the molecular organisation of mouse hippocampus at scale.
Introduction
Our understanding of neurological function and disease relies on knowledge of both the cellular architecture of brain tissue and the molecular organisation within neurons. CryoET is a powerful technique for investigating the structure and distribution of biological macromolecules in their native, cellular environment. Traditional sample preparation for cryoET of tissues involving chemical fixation and microtomy leads to artefacts that can obscure high-resolution detail, impeding biological interpretation on the molecular scale. Alternatively, tissues can be studied in their native, hydrated state by preparing vitreous samples with HPF, combined with nanofabrication approaches such as focused ion beam (FIB) milling.
The advent of PFIB instruments offers enhanced milling speeds, which have the potential to greatly improve the throughput of sample preparation for thick tissue samples [1]. Similarly, recent advancements in cryogenic lift-out of high-pressure frozen (HPF) samples on grids have shown the ability to excise large sections of material from HPF samples, and have enabled the analysis of more complex biological samples by cryoET [2,3]. To fabricate samples from thicker specimen (i.e. HPF brain tissue) routinely, a combination of PFIB and serial lift-out is almost obligatory but has yet to be demonstrated.
Methods
Mouse brain was sectioned sagittally on ice to 100 µm thickness and hippocampal biopsies were stained with nuclear dye before vitrification by HPF. After loading into the PFIB, hippocampal layers were identified using in-chamber cryo-fluorescence. A 350 µm region of CA1sr was targeted and trenches were milled using high-current (200 nA) Xe plasma. The isolated sample was removed with a cryogenically cooled manipulator, and serial sections were deposited on a copper grid followed by automated thinning to electron transparency. Tilt series were acquired, processed, and reconstructed for visualization.
Results
A novel milling approach was developed to isolate sections spanning multiple hippocampal layers in HPF mouse brain tissue by serial lift-out. A correlated light and electron microscopy (CLEM) approach identified the location of the pyramidal cell layer across the biopsy, which guided the lift-out of targets spanning the CA1 stratum oriens to the CA1 stratum lacunosum. The use of high current (60-200 nA) xenon plasma to shape the sample allowed deep trenches adjacent and above the region to be milled in less than 30 minutes. This vastly increased throughput with no visible damage artefacts being observed in the resulting lamellae. We could deposit up to 42 serial sections that sample across a 350-400 µm region of CA1, which were thinned to electron transparency with >90% success. In total, 246 tilt series were collected in the CA1 stratum radiatum (CA1sr).
Isolating a section of tissue within the plane of the sample by manipulation of the lift-out geometry provided access to a previously unobserved view of the CA1 stratum pyramidale (CA1sp) apical dendrites in 12% (n=30/246) of tomograms. This perspective allowed for the organisation of microtubule network and its interactions with other cellular components to be observed.
Additionally, 9% of tomograms in this dataset contained a confirmed synapse (n=22/246), with a further 11% (n=27/246) showing putative synapses. The quality of the tomograms reconstructed from this tissue sample was sufficient to identify canonical synaptic features with molecular resolution.
Conclusion
This work presents a routine approach for generating lamellae for cryo-ET of multiple layers of mammalian hippocampus from a single lift-out using PFIB. The developed pipeline incorporates a cryo-CLEM aspect to guide the lift-out of targets spanning 350-400 µm across the CA1 region, generating >40 deposited sections of HPF brain biopsy in a single 24-hour microscope session. By lifting out a section of tissue in the plane of the sample, previously inaccessible views of the CA1sp apical dendrite cytoskeletal network could be imaged in vitrified samples. The spatial relationship between sequential lamellae, which sample every 5-10 µm of the CA1sr, demonstrates the feasibility of acquiring cellular cryo-ET data with enhanced biological context using our approach. The workflow we present here has the potential to be adapted for different cryogenically preserved model systems and tissues, opening up possibilities for observing biomolecules within their most native environment.
References
[1] Burnett T.L, et.al., Ultramicroscopy, (2016). 161, 119-129. https://doi.org/10.1016/j.ultramic.2015.11.001
[2] Schiøtz O.H, et. al., Nature Methods, (2023). 20, 585-594. https://doi.org/10.1038/s41592-023-02113-
[3] Nguyen HTD, et.al., Nature Methods, (2024). 21, 1693-1701 https://doi.org/10.1038/s41592-024-02384-6
Keywords
Structural Neurobiology, Plasma-FIB Milling, CryoET, Serial Lift Out