Local nucleosome dynamics facilitate chromatin accessibility in living mammalian cells ヌクレオソームのゆらぎは、生細胞においてクロマチンへのアクセスビリティを容易にする

博士論文

In cell nuclei or mitotic chromosomes, long strings of genomic DNA are organized three-dimensionally to perform genome functions during cellular proliferation, differentiation, and development. The DNA is wrapped around histones, forming a nucleosome structure. The nucleosome had been assumed to be folded into a 30-nm chromatin fiber and other helical folding structures. However, recent studies, including our cryo-microscopy (cryo-EM) and synchrotron X-ray scattering analyses, have shown almost no visible 30-nm chromatin fibers or other regular structures in interphase nuclei and mitotic chromosomes. This suggests that chromosomes consist of irregularly folded nucleosome fibers comprising a polymer melt-like structure. Thus, nucleosome fibers may be constantly moving and rearranging at the local level. These local nucleosome dynamics could be crucial for various genome functions. I studied the dynamic aspects of nucleosomes in living cells. Because the dynamic chromatin environment in living cells is difficult to study using traditional fluorescence and electron microscopy, I used, with the help of my collaborators, a combination of fluorescence correlation spectroscopy (FCS), single molecule imaging, and Metropolis Monte Carlo computer simulations. First, to determine the chromatin environment in living cells, I employed FCS using free enhanced green fluorescent proteins (EGFPs). FCS detects the Brownian motion of free EGFPs in a small detection volume based on fluorescence intensity fluctuations and provides the diffusion coefficient (D) of the EGFPs, which indicates how far the molecules can move in a particular time. Thus, D gives useful information on their environment, with a smaller D indicating that the molecule exists in a more crowded environment, and vice versa. I measured Ds of EGFP-monomer, -trimer, or -pentamer molecules in interphase chromatin and mitotic chromosomes. Unexpectedly, D in mitotic chromosomes was quite comparable to that in interphase chromatin, thus suggesting that protein mobility in interphase chromatin and that the mitotic chromosome is comparable. Next, based on the physical parameters obtained via FCS, the chromatin environment in silico was reconstructed using the Metropolis Monte Carlo method to simulate EGFP behavior under various chromatin conditions. To simulate the diffusion of EGFP pentamers from the Stokes–Einstein relation, I represented EGFP-pentamer molecules as 13-nm-diameter green balls. Nucleosomes were represented as 10-nm-diameter red balls and fixed in a space at a concentration of 0.1 or 0.5 mM. The 0.5-mM condition corresponded to mitotic chromosomes and likely interphase heterochromatin. In the environment with 0.1-mM red balls under a fixed condition, the green balls moved around quite freely. However, with 0.5-mM fixed red balls, which represented a dense heterochromatin or chromosome environment, green balls could not move far from their starting position and were trapped in a confined space. Although this simulation suggested that EGFP pentamers in fixed chromatin environments cannot move around freely, it was inconsistent with FCS measurements in the living chromatin environment, in which the apparently free diffusion of EGFP pentamers was observed. To determine the conditions that could recapitulate the observations in vivo, next a simulation with fluctuating red balls was performed. Each red ball acted like “a dog on a leash,” being set in random motion within a certain distance range from the origin. In this dynamic chromatin environment, the green balls appeared to diffuse freely, even with 0.5-mM red balls. Strikingly, a 20-nm maximum displacement of red balls was sufficient for green balls to diffuse freely in the chromatin environment. This finding suggests that the dynamic fluctuation of nucleosomes facilitates the free diffusion of proteins in a compact chromatin environment, such as that in mitotic chromosomes, as well as dense heterochromatin. An obvious next question was whether the nucleosome fluctuations predicted by the simulation occur in living cells. Therefore, I performed single-particle imaging of nucleosomes in living cells. I fused photoactivatable (PA)-GFP with histone H4, which is a stable core histone component, and then expressed the fusion protein in cells at a very low level. For single nucleosome imaging, I used highly inclined and laminated optical sheet (HILO) microscopy. Unexpectedly, I found that a very small number of PA-GFP-H4s in the stable cells spontaneously activated without laser activation and could be observed as dots. With this imaging system, I recorded nucleosome signals in interphase chromatin and mitotic chromosomes at a video rate of ~30 ms/frame as a movie. The averaged displacements (movements) in 30 ms in interphase chromatin and mitotic chromosomes were 51 and 59 nm, respectively, and showed a similar fluctuation in both interphase and mitotic chromatin. Since the displacements of fluorescent beads on the glass surface or in cross-linked nucleosomes in glutaraldehyde-fixed cells were much smaller than those observed in living cells, I concluded that the majority of the displacement came from the movement of nucleosomes in living cells, and not from the drift of the microscopy system. Last, I examined whether local nucleosome dynamics drive chromatin accessibility or targeting in dense chromatin regions. To do so, I used immunostaining of condensin in mitotic chromosomes as a model system in dense chromatin regions. Immunostaining signals demonstrated that the antibodies (150 kDa, >15 nm) targeted the condensin complexes in the chromosome axes. Although I detected antibody signals in the chromosome axes of non-fixed cells, far fewer were observed in glutaraldehyde-fixed cells. This finding is consistent with the previous results and indicates that tight cross-linking of nucleosomes blocks antibody accessibility and targeting. In this study, I showed that interphase chromatin and dense mitotic chromosomes have comparable protein diffusibility. In both chromatins, I observed a novel local dynamics of individual nucleosomes (~50 nm movement/30 ms) caused by Brownian motion. The inhibition of this local dynamics by cross-linking impaired diffusibility and targeting efficiency in dense chromatin regions. I propose that this local movement of nucleosomes is the basis for scanning genome information.

総研大甲第1562号