Recruitment of chromatin remodelers by XIST B-repeat region is variably dependent on HNRNPK.

Abstract

X-chromosome inactivation is triggered by the long non-coding RNA XIST, whose structure is characterized by tandem repeats that modularly recruit different proteins and chromatin remodelers. Previously, we reported that the addition of the mouse PID region to a transgene with human repeat regions A, F and E (miniXIST; 5.1 kb) enabled binding of HNRNPK and also enabled the induction of silencing and recruitment of H3K27me3, UbH2A and H4K20me1, but only partially. As the 680 bp PID region enabled so many features of inactivation, we hypothesized that augmenting the PID with more mouse or human sequences rich in CCC motifs would allow us to design a short transgene which was as effective as Full XIST. Three new transgenes using the A, F and E human domains as a backbone were tested for ability to induce silencing and heterochromatic mark recruitment. The all human-derived BhB-BhB transgene (4.9 kb) was as good as our previous miniXIST, suggesting that these domains are the human equivalent of the mouse PID region. A PID-PID transgene (5.8 kb) was not statistically different from Full XIST and could be potentially used for chromosome therapy. Adding BhB to PID (BhB-PID, 5.4 kb) had an intermediate efficacy compared to the other two transgenes, suggesting that the most important component for silencing and heterochromatic mark recruitment is the number of CCC motifs, not the species of origin. Finally, we created a heterozygous HNRNPK deletion and observed a disproportionate impact on HNRNPK and UbH2A recruitment to XIST, reflecting complex roles for the PID and HNRNPK in X-chromosome inactivation.