DNA is organized into TADs and TADs into chromosomes, but how this process occurs at the single-cell level was unknown until recently. One main reason was the lack of appropriate tools. For example, conventional microscopies lacked the resolution or specificity to visualize TAD structures in single cells. To solve these limitations, we developed two different methodologies. In the first, we adapted two-color stochastic optical reconstruction microscopy (STORM) to visualize active and repressed TADs in single cells, and used 3D-structured illumination microscopy (3D-SIM) to measure for the first time absolute interaction frequencies within a single TAD. In the second, we developed a unique combination of experimental and modeling to show that repressed TADs form a succession of discrete nanocompartments and are highly structurally heterogeneous.
Super-resolution-based methods allowed us to visualize the structures of TADs in single cells, however these technologies allow only for the visualization of a limited number of targets (1-2). Thus, while we could super-resolve the volume occupied by a TAD in a single cell, we could not restore the connectivity between genomic elements critical to understanding function (e.g. enhancers, promoters). Sequencing-based methods, in contrast, can access multiple interactions at once, but are difficult to perform in single cells and lack spatial resolution. To uplift these limitations of conventional microscopies and next generation sequencing-based techniques, we developed Hi-M. Hi-M provides the ability to simultaneously monitor transcription and chromosome organization in single cells while retaining spatial information within the cell and the organism. Critically, Hi-M can be performed in whole-mount embryos and in tissue sections. For these reasons, Hi-M has a tremendous potential to make transformative changes to fundamental and applied biomedical research, for example by providing new ways of understanding disease progression at the single-cell level, and by deriving new diagnostic tools. In the near future, we will use Hi-M to elucidate mechanisms of transcriptional activation, and repression, and characterize the intrinsic heterogeneities of DNA organization in single cells.
First, we leveraged the unique advantages of HiM to investigate how chromosomes folded within single cells during early Drosophila embryogenesis. We were particularly interested in testing if preferential contacts between enhancers and promoters within a model TAD interacted together and whether these interactions were conductive of activation. Surprisingly, we found that regulatory interactions are present in both transcriptionally active and inactive cells, supporting a model in which occupation of transcription factors is the main regulator of enhancer activity. We also applied Hi-M to shed light into the role of chromatin insulators in the folding of Drosophila TADs.
Next, we studied the functional importance of cis regulatory elements (PREs, enhancers) for the structure and function of Polycomb TADs. We showed that PcG dependent silencing and chromatin marks can be established in the absence of PRE looping, but 3D contacts between PREs might stabilize PcG-mediated gene repression during development. Also, we showed that PREs interact both in tissues where expression is active or repressed, and that insulators produce small changes to this organization, but have important transcriptional consequences. Finally, we investigated long-range Pc domain organization using HiM, and found that Pc domains colocalize in 3D sporadically and mostly in pairs. This data is in contrast to current models for human cells suggesting Pc domains form phase separated compartments.
