ИСТИНА

Drosophila chromosomes are partitioned in large self-interacting domains (Topologically Associating Domains, TADs) separated by boundary regions or inter-TADs. How this spatial organization of chromatin is established and supported is not clear. We have performed Hi-C analysis on four Drosophila cell lines of different origins and annotated TADs using the Armatus software. Contrary to the previous studies, we did not observe a strong enrichment of TAD borders/inter-TADs with CTCF deposition sites, and instead another insulator protein Su(Hw) was preferentially present within the TADs. However, Drosophila inter-TADs harbor active chromatin and constitutively transcribed (house-keeping) genes. The tissue-specific genes reside preferentially in TADs and their transcription appears to correlate with a partial de-compaction of the TADs. Pairwise comparison of the cell lines demonstrated that, in some cases, activation of transcription within a TAD resulted in splitting of this TAD (i.e. in generation of a new boundary/inter-TAD). We proposed that integral features of active chromatin determine the TAD boundaries. Of particular importance may be the high level of histone acetylation that directly influences the ability of nucleosomes to interact with each other. Using computer simulations we demonstrated that polymers composed of long blocks of non-acetylated (capable to interact with each other) nucleosomes interspersed with shorter blocks of acetylated (incapable to interact with each other) nucleosomes adopts spatial configuration closely resembling organization of chromosomes in TADs. The important role of histone acetylation in determining TAD profiles was further confirmed by experiments with chemical inhibition of histone acetylases and histone deacetylases in S2 cells. An important question is whether TADs are present in individual genome or represent a population average. To address this question we have developed a methodology allowing construction of Hi-C maps for individual cells. Using this modified Hi-C protocol we have constructed HiC maps for 20 drosophila cells (line DmBG3c2). In the best cell we have captured ~10% of the theoretically available contacts. To analyze these sparse contact matrices, we have developed program tools allowing us to take into account the noise by comparing maps from individual cells with artificially generated random matrices. The results of our analysis demonstrate that contact chromatin domains do exist in individual chromosomes and are organized hierarchically. Importantly, using a number of statistical approaches we show that the observed profile of contact chromatin domains cannot be explained by random fluctuations. We also show that these domains do not coincide in randomly selected pairs of individual cells but tend to occupy some preferential positions in the genome as revealed by the comparison of several cells. Finally, we show that genomic regions that frequently harbor the contact domain borders possess specific epigenetic signatures. ACKNOWLEDGMENTS This work was supported by a Russian Science Foundation grant #14-24-00022.