On Sunday-morning I left a warm and sunny Berlin to attend the Keystone conference “Imaging Across Scales: Leveraging the Revolution in Resolution” in Snowbird Utah. After a short stop-over in my home-country (and just enough time to buy some Gouda and stroopwafels!) a long flight took me directly to the white mountains surrounding Salt Lake City: what an impressive change of scenery! The goal of the conference was to bring together researchers from different imaging disciplines in order to stimulate new collaborations for imaging across multiple length scales and using different approaches. During the 3 days of the conference, we were treated with very diverse and outstanding talks on super-resolution fluorescence microscopy, electron tomography, atomic force microscopy, probe design, image analysis and much more. There were too many highlights to discuss them all in this blog-post, but I will try give a short summary of a number of them (whilst avoiding discussion of unpublished work):
The meeting was kicked off with an impressive keynote seminar by Xiaowei Zhuang (Harvard University). A few years ago, Zhuang’s lab has pioneered the development of MERFISH, a multiplexed single-molecule fluorescence in situ hybridization (smFISH) approach that can be used to image 100 to 1000 RNA species in a single cell (). Zhuang discussed her lab’s efforts to take this approach to the next level by creating spatial atlases of cells in complex tissues. As an example, they mapped ~70 neuronal populations in the preoptic region of the hypothalamus and were able to identify discrete neuronal cell-types involved in key social behaviours, such as parenting, aggression, and mating. Going beyond single-molecule RNA imaging, they have recently employed similar strategy for tracing the chromatin inside the nucleus (Oligopaint). This has enabled them to image the 3D organisation of the chromatin in single cells and to trace the positions of topologically associating domains (TADs) ( ), self-interacting genomic regions that are thought to control gene expression. Extension of this approach to super-resolution (STORM) imaging now allows them to image TADs and sub-TADs at 30kb resolution in single cells ( ). Zhang showed there is a striking resemblance between their data and previously published ensemble-averaged Hi-C contact matrices, and emphasizes the importance of being able to cross-validate Hi-C studies with Oligopaint data and vice versa.
Toshio Ando (Kanazawa University) showed an impressive range of examples wherein they have used high-speed AFM to image the conformation dynamics of cytoskeletal motor proteins and enzymes, among others. Most recently, they were able to directly visualise microtubule depolymerisation at nm-resolution and showed that microtubule inner proteins stabilize microtubules in Chlamydomonas axonemal microtubules (). They are now taking their high-speed AFM approach to the next stage by combining it with real-time manipulation of the sample by applying force.
Ru Gunawardane (Allen Institute for Cell Science) discussed her lab’s ongoing work to create a high-quality protein-localization map for different cellular states and behaviours. So far, they have created a collection of 38 fluorescent-protein labelled organelles and protein complexes using CRISPR. They do this in human induced pluripotent stem cells (hiPSCs) because they are diploid, non-transformed, pluripotent, proliferative. They perform extensive quality control on 100 clones for each cell line, which includes 17 assays to test to check for correct editing and exclude potential effects of the knock-in lines on cellular behaviour (growth rate, pluripotency markers, differentiation etc.). They are making all plasmids and validated cell lines available to the scientific community (!) and are continuously expanding their collection of cell lines (.
Clodagh O’Shea (The Salk Institute) talked about ChromEMT, her lab's recently developed approach for visualizing the genome at nucleosome resolution using electron microscopy tomography (EMT). The contrast of DNA in electron microscopy is normally very poor and in order to enhance the contrast, they take advantage of a DNA-binding fluorescent dye (DRAQ5). Upon excitation, DRAQ5 does not only emit a photon when returning to the ground-state, but also locally produces reactive oxygen species (ROS). They exploit this by enabling ROS-dependent catalysis and deposition of diaminobenzidine (DAB) polymers on the chromatin surface, which form a high-affinity binding site for the contrast agent OsO4. This allowed them to directly visualize the chromatin inside the nucleus at the individual nucleosome level (). To their surprise, they did however not observe clear higher order structures, which raises the question of how gene activity is controlled if not through different levels of chromatin packing? In order to explain their observations and to provide an alternative model for regulating gene accessibility, O’Shea made an interesting connection to polymer physics: a gelation step is known to occur when polymers exceed a critical concentration (typically 20-30%). She hypothesized that a similar mechanism might be in place inside the nucleus, where this critical concentration could be changed locally to promote or restrict access to genes. They determined that both mitotic chromosomes and heterochromatin have a DNA (polymer) density that would be above this critical concentration, whereas pluripotent ES cells appear to be largely below this critical concentration. In addition, they are currently designing contrast-labels to specifically visualise genomic loci of interest using EMT. This should allow them to study the organisation of individual genes at high resolution and to visualize how active genes looks compared to their inactive versions.
Gary Borisy (Forsyth) presented his lab’s work on mapping microbiomes at the micron scale by combining multiplexed FISH with spectral imaging. They have developed a set of FISH probes against bacterial 16S rRNA sequences of 15 laboratory-grown taxa. The highly repetitive nature of 16S rRNA provides them with a strong enough signal for fluorescence imaging, while the inter-species differences in the 16S rRNA sequences enables them to reliably identify different types of bacteria. They achieve high level multiplexing by spectral imaging (removing band-pass filters and detecting emission signatures by linear unmixing) and Borisy shows the most beautiful multi-colour images of the organisation of different types of bacteria within the microbial communities found in dental plaque (), gut ( ) and more. They can identify hierarchies and rules within these microbiome communities and are trying to quantitatively understand the higher-order formation and organisation of these communities.
I think it is clear from my notes that this an excellent conference, which covered a broad range of different but related topics and had a very strong focus on method development. The small scale of the meeting made this an excellent opportunity to interact with the other participants, not only to extend the scientific discussions but also to get to know the people behind all these creative ideas! I hope that the positive atmosphere during the meeting has encouraged the organisers to think about organising a second edition of this outstanding conference!