The specific roles of cohesin-SA1 and cohesin-SA2 in gene regulation and genome organization

The use of Chromosome Conformation Capture (3C)-derived techniques for probing 3D genome structure has shed new light on the principles of genome organization. DNA looping is recognized as the major mechanism responsible for this organization in which topologically associated domains (TADs) and compartments are acknowledged as hierarchical levels of genome folding at a megabase scale. The cohesin complex, working together with the insulator protein CTCF, has shown to be essential for partitioning the genome into chromatin loops and TADs. Cohesin is a ring-shaped protein complex best known for its role in sister chromatin cohesion that consists of four subunits: SMC1, SMC3, RAD21 and SA. In somatic vertebrate cells, the SA subunit can be either SA1 or SA2, thus giving rise to two different cohesin variants, cohesin-SA1 and cohesin-SA2. Studies in human and mouse cells have suggested that these variants have non-redundant functions, at least regarding cohesion. However, their differential contributions to gene regulation and genome organization have not been explored. In order to address this question, we first analysed the genome-wide distribution of the two cohesin variants in three human cell lines and found that cohesin-SA1 is almost exclusively present at CTCF-bound sites. In contrast, only a fraction of cohesin-SA2 colocalizes with cohesin-SA1 and CTCF, while another one occupies non-CTCF positions that are instead bound by transcriptional regulators. This population is enriched at active enhancers, in particular at super-enhancer elements responsible for establishment of cell identity. Importantly, cohesin-SA2 can reach CTCF positions in the absence of SA1, but cohesin-SA1 cannot occupy non-CTCF positions when SA2 is absent. Differential affinity of SA1 and SA2 for cohesin-releasing factor WAPL, CTCF and transcriptional regulators may contribute to the observed differences in the dynamics of their association to chromatin. Downregulation of cohesin-SA2 has a more pronounced effect on the transcriptome of MCF10A cells and leads to deregulation of core cell-identity genes. Using 4C and Hi-C analyses we also explored the consequences of cohesin variant depletion on local and global genome organization. Our results suggest that cohesin-SA1 works together with CTCF to define TADs while cohesin-SA2 facilitates more transient and local regulatory contacts. In summary, we propose that cohesin-SA2 contributes to cell-type specific gene regulation in a CTCFindependent fashion, a function that cannot be assumed by cohesin-SA1. Our work provides a new perspective on understanding the contribution of cohesin mutations to the pathology of human cancers.