A model for chromatin networking via RNAPII-Ser5P. (A) A model for the formation of a loose spatial genome chromatin network via RNAPII-Ser5P, which can globally constrain chromatin dynamics. The P-TEFb complex (blue sphere in right panel) consisting of CYCT and CDK9 kinase, which interacts with RNAPII, forms a number of dynamic clusters/droplets in living cells (pink spheres in the center and right panels; Ghamari et al., 2013). Center: The P-TEFb cluster (pink sphere) can work as a hub to weakly connect multiple chromatin domains (green spheres) for a loose spatial genome network. Right: RNAPII-Ser5P (red) can function in the hub as glue for the weak interactions between P-TEFb (blue spheres) and transcribed DNA regions (green lines; Ghamari et al., 2013). Because after phosphorylation of RNAPII by P-TEFb, RNAPII-Ser2P seems to leave the hubs (P-TEFb clusters) for the elongation and processing process (Ghamari et al., 2013), it is unlikely to function as the glue for the connections (right). Note that this scheme is highly simplified. Besides the P-TEFb clusters, other clusters, including transcription factors, Mediator, and active RNAPII (Boehning et al., 2018; Boija et al., 2018; Cho et al., 2018; Chong et al., 2018; Sabari et al., 2018), might also work as hubs. (B) MSD plots (±SD among cells) of nucleosomes in CDK9-KD RPE-1 cells (siCDK9, pink) and control (siControl, black). For each condition, n = 20 cells. Note that the KD of CDK9 increased the chromatin dynamics. ***, P < 0.0001 (P = 1.3 × 10−6) by the Kolmogorov–Smirnov test. (C) CDK9 reduction in RPE-1 cells after RNA interference was verified by immunoblotting. H2B protein was used as a loading control.
