I initiated the project by developing a technique to locally scramble LAD and iLAD sequences (see attached Figure 1, top panel). It consists of three main steps: 1) genomic integration of a cassette consisting of a SB transposon and two loxP sites, 2) SB hopping, to relocate one of the loxP sites within LAD sequences, and 3) Cre-mediated recombination between the loxP sites, creating either genomic inversions or deletions. After optimizing both SB hopping and Cre-mediated recombination, I generated long-range deletions and inversions ranging from 2.5 kb to 2 Mb and spanning both LAD and iLAD regions. Cre recombination was remarkably efficient (~10% at 2 Mb). This technique, performed in mouse embryonic stem cells, is now fully implemented in the lab.
I established a collection of 12 cell lines, dissecting two neighboring LADs in the mouse genome. I then assessed the LAD pattern in those recombined clones by pA-DamID, a technique already established in the lab. My data show that LADs are tethered to the NL by multiple elements. Interestingly, some elements are more potent than others. While the latter only has a modest affinity for the NL, the former autonomously interacts with the NL and even boosts the association of flanking sequences to the NL. Each of the two LADs inspected contained such potent subregions, suggesting that they could be a common tethering mechanism for LADs at the NL.
I then wondered whether entire LADs could cooperate to strengthen their association with the NL. I show that provided that they are close enough in the linear space, neighboring LADs can indeed cooperate to promote their association with the NL.
LADs are heterochromatic structures often associated with repressive histone marks such as H3K9me2 and H3K9me3. We therefore investigated whether the observed changes in NL interactions were mirrored by changes in H3K9me3. I performed pA-DamID to probe H3K9me3 deposition in several recombined and control cells. Interestingly, the changes in NL association were only partially mirrored by changes in H3K9me3.
Finally, I investigated the functional consequences of NL association and H3K9me3 changes by performing RNA-seq experiments in multiple recombined and control cells. As expected, gain in lamina association correlated with gene downregulation. However, it correlated slightly better with H3K9me3 changes. My findings suggest that changes in H3K9me3 are a slightly more important predictor of the changes in gene expression than changes in NL interactions. In conclusion, scrambling of LAD and iLAD sequences alters NL interactions, H3K9me3, and gene activity.
The project and associated findings were presented multiple times over the years at international conferences, through both posters and oral presentations. It included prestigious ones such as Cold Spring Harbor: Nuclear Organization & Function and Gordon Research Conferences. To better disseminate my results, I also wrote a scientific paper that is open to read by the scientific community on the public repository BioRxiv. I intend to publish those exciting results in a high-impact Open Access journal as soon as possible. To ensure transparency and favour data reproducibility, lab journals will be released along with the publication. Besides, to spread my findings as soon as possible, I used the lab website to update the community about my project, before its publication.
