Heterogeneous fluid-like movements of chromatin and their implications to transcription

Abstract

Eukaryotic chromatin is a complex of genome DNA and associated proteins, and its structure and dynamics play a crucial role in regulating DNA functions. Chromatin takes rather irregular structures in the nucleus and exhibits heterogeneous sub-diffusive movements as polymers fluctuating in a fluid state. Using genome-wide single-nucleosome tracking data, heterogeneity of movements was statistically analyzed, which categorized chromatin into two types: slow chromatin that moves under structurally constrained environments and fast chromatin that moves with less constraints. Interactions of chromatin to various protein factors determine the motional constraints. For example, loss of the cohesin complex that bundles the chromatin chains reduces the motional constraints and increases the population of fast chromatin. Another example is the transcriptional machinery. While it was previously thought that the transcriptional activity is associated with more open and dynamic chromatin structure, recent studies suggested a more nuanced role of transcription in chromatin dynamics: dynamic association/dissociation of active RNA polymerase II (RNAPII) and other transcription factors and Mediators (TF-Meds) transiently bridges transcriptionally active DNA regions, which forms a loose network of chromatin and constrains chromatin movement, enhancing the slow chromatin population. This new view on the dynamical effects of transcription urges a reflection on the traditional model of transcription factories and invites the more recent models of condensates/phase-separated liquid droplets of RNAPII, transcription factors, and Mediators. The combined procedure of genome-wide single-nucleosome tracking and its statistical analysis would unveil heterogeneity in the chromatin movement, which should provide a key to understanding the relations among chromatin dynamics, structure, and function.

Keywords: Cohesin; Liquid droplets; Live-cell imaging; Nucleosome; RNA polymerase II; Statistical analyses; Transcription factory.

Fig. 1

Fast and slow nucleosomes. a The van Hove self-correlation (vHc) function, $2\mathit{\pi r}{G}_s^0\left(r,t\right)$, directly calculated from the single-nucleosome trajectories observed in living HeLa cells by Nozaki et al. (2017) (red) is compared with the one obtained from the RL method, 2πrG_s(r, t) (black). Here, vHc is shown by multiplying a factor 2πr for the area under the curve to be normalized. bP(M, t) obtained with the RL method for an example cell. cP(M, 0.5 s) vs M/M^∗ calculated for 10 cells. M^∗ is the separation point of bimodal peaks indicated. d MSD calculated for the fast (black) and slow (red) nucleosomes based on the nucleosome identification discussed in the main text. Dashed lines are MSD of individual 10 cells and real lines are the average over them. Reprinted from Ashwin et al. (2019)

Fig. 2

Effects of perturbations on HeLa cells and focusing on heterochromatin. Features of the effects on the distribution of MSD, P(M, t) at t = 0.5 s, of single nucleosomes: a the ratio of the number of fast nucleosomes to the number of slow nucleosomes, b the mean MSD of fast nucleosomes, and c the mean MSD of slow nucleosomes. Box plots of the data from 10 cells. Reprinted from Ashwin et al. (2019)

Fig. 3

Increased chromatin dynamics with RNAPII inhibitors. a MSD plots (±SD among n = 20 cells) of nucleosomes in the living human retinal pigment epithelium RPE-1 cells treated with RNAPII inhibitors, α-AM (pink) and ActD (brown). As untreated control, dimethyl sulfoxide (DMSO, gray) was added. b A model for the formation of a loose spatial genome chromatin network via active RNAPII-Ser5P, which can globally constrain chromatin dynamics. Chromatin domains, each formed by cohesin binding, are bound to a condensate/cluster/droplet of transcription factors and Mediators, which works as a “Hub” of loose network of chromatin. c–e Computational modeling of chromatin domain network via active RNAPII. Brownian dynamics of four chains of chromatin domains (green spheres) connected by springs (invisible) and four hubs, clusters of transcription factors/Mediators (pink spheres), were simulated. Chromatin chains bind N_glue RNAPII-Ser5P (red spheres), and transient attractive interactions were assumed between RNAPII-Ser5P-bound chromatin and hubs. c Snapshots of the Brownian dynamics with no glues (left) and with 64 glues of RNAPII-Ser5P (red spheres) (right). d The average MSD of total chromatin domains in the system calculated with various glue numbers bound on chromatin domains from N_glue = 0 to 64. e The MSD (0.5 s) distribution plots of total chromatin domains in the system with (gray) and without RNAPII-Ser5P glues (black). An arrow indicates the contribution from chromatin sites neighboring RNAPII-bound loci. Reprinted from Nagashima et al. (2019)