Generic features of tertiary chromatin structure as detected in natural chromosomes - PubMed (original) (raw)
Generic features of tertiary chromatin structure as detected in natural chromosomes
Waltraud G Müller et al. Mol Cell Biol. 2004 Nov.
Abstract
Knowledge of tertiary chromatin structure in mammalian interphase chromosomes is largely derived from artificial tandem arrays. In these model systems, light microscope images reveal fibers or beaded fibers after high-density targeting of transactivators to insertional domains spanning several megabases. These images of fibers have lent support to chromonema fiber models of tertiary structure. To assess the relevance of these studies to natural mammalian chromatin, we identified two different approximately 400-kb regions on human chromosomes 6 and 22 and then examined light microscope images of interphase tertiary chromatin structure when the regions were transcriptionally active and inactive. When transcriptionally active, these natural chromosomal regions elongated, yielding images characterized by a series of adjacent puncta or "beads", referred to hereafter as beaded images. These elongated structures required transcription for their maintenance. Thus, despite marked differences in the density and the mode of transactivation, the natural and artificial systems showed similarities, suggesting that beaded images are generic features of transcriptionally active tertiary chromatin. We show here, however, that these images do not necessarily favor chromonema fiber models but can also be explained by a radial-loop model or even a simple nucleosome affinity, random-chain model. Thus, light microscope images of tertiary structure cannot distinguish among competing models, although they do impose key constraints: chromatin must be clustered to yield beaded images and then packaged within each cluster to enable decondensation into adjacent clusters.
Figures
FIG. 1.
Induction of MHCII by interferon and characterization of DNA probes for this region. (a) Interferon induction leads to expression of cell surface HLA antigens, as determined by immunofluorescence with antibodies against HLA-DP, HLA-DQ, and HLA-DR antigens. As determined with a t test, significant increases in intensity were observed after 2 and 5 h of interferon induction (P < 0.01). Standard errors are shown. (b) Two probes, PAC RP1-93N13 and BAC RP11-10A19, were used in DNA FISH to detect the MHCII region. These probes were end sequenced and found to align with the MHCII locus. Genes within the MHCII region are shown below the black line representing the locus; orange letters indicate genes induced by interferon. (c) Confirmation of probe specificity in DNA FISH by colocalization with known MHCII-specific probes.
FIG. 2.
Transcriptionally active domain on chromosome 22 and characterization of probes for this region. (a) Based on an analysis of microarray expression data (see Materials and Methods), a transcriptional score was calculated on the q arm of human chromosome 22 (chr 22q) for the expression of Jurkat cells relative to Raji cells. A peak was found in the domain from 4.3 to 6.1 Mb (encircled). (b) A finer-scale analysis of transcriptional levels within the 4.3- to 6.1-Mb domain revealed that transcriptional activity there peaked at between 5.3 and 5.8 Mb. (c) To probe tertiary chromatin structure in this domain, three probes were used. Their alignment with the 4.3- to 6.1-Mb domain was determined by end sequencing. Each probe was previously mapped by FISH (18), confirming its specificity. The red and blue BACs mark the end points of the 4.3- to 6.1-Mb domain and were used to measure the separation between these two locations. They span a 1,680-kb domain. The green and blue BACs were used to examine tertiary structure in the most transcriptionally active subdomain of the 4.3- to 6.1-Mb domain. These probes span a 440-kb domain.
FIG. 3.
Beaded images of tertiary structure arise in transcriptionally active regions in natural chromosomes. Projections of 3D deconvolved images of tertiary chromatin structure are shown. (a and c) Images of tertiary structure typically exhibit two to four beads in either the MHCII region on chromosome 6 after induction by interferon (a) or in a transcriptionally active domain on chromosome 22q in Jurkat cells (c). An assortment of decondensed states is shown, reflecting the natural variability in the cell population. Beads are always present; dimmer connecting strands sometimes are detected between them. In most images, the beads are nearly in contact, but wider separations sometimes occur. (b and d) In contrast, when these same natural domains are less transcriptionally active, images exhibit predominantly single beads. Single beads are detected when the MHCII region is examined before induction by interferon (b) or when chromosomal region 22q-11.21-22a is examined in Raji cells (d), which transcribe significantly less from this region than do Jurkat cells. (e and f) As a control for the preservation of structure by DNA FISH, we examined cells harboring the MMTV array after transcription was induced by hormone addition. (e) For live cells, images of the array containing up to ∼15 beads are observed with a GFP-tagged glucocorticoid receptor. (f) Similar images are seen for cells examined after DNA FISH with a probe for MMTV DNA. (The higher background in panel e is due to binding of the GFP-tagged glucocorticoid receptor throughout the nucleoplasm, while in panel f a DNA probe specific for the tandem array was used.)
FIG. 4.
Transcriptionally active states yield longer structures. Upper panels show examples of nondeconvolved images used for measurements of tertiary chromatin structures for the three different domains examined. Both the 440-kb domain on chromosome 22q and the 375-kb domain for MHCII were spanned by two nearly contiguous BAC or PAC probes. In contrast, the 1,680-kb domain on chromosome 22q was marked by a BAC at each end point, giving rise to images in which two separate structures could be discerned (b). Lower panels (bar charts) show the mean lengths of the structures measured in the different cell types in the presence (white bars) or absence (gray bars) of the transcriptional inhibitor DRB. The means are from 50 to 300 measurements, and each experiment was repeated at least three times with similar results; standard errors are shown. The brackets indicate that significant differences between means were found with a t test (P values are shown).
FIG. 5.
A NARC model can explain beaded images. (a and d) Examples of 400-kb simulated nucleosome chains. (b and e) High-magnification views of the simulated images reveal the presence of fluorescence puncta; most details of the nucleosome chain are lost. (c and f) Overlays show that less densely packed regions of the chain may extend beyond the puncta and may be invisible. (g) Sampling of simulated projected images following deconvolution for 400-kb chains with different degrees of decondensation. Note that these images exhibit beads resembling those in the actual images (compare to Fig. 3). (h) When the first third of the chain is colored blue, the middle third is colored green, and the last third is colored red, distinct largely nonoverlapping regions are obtained either for long chains of 256 kb (left panel) or for shorter chains of 64 kb (middle and right panels). These data indicate that, on a macroscopic scale, neighbor-neighbor relations are preserved, and so the chain could unravel without significant entanglement.
FIG. 6.
Chromonema models and radial-loop, protein scaffold models can be adapted to account for beaded images. (a) Chromonema models propose a hierarchically folded chromatin fiber that yields a series of different fiber thicknesses (4, 35, 40). Here, a thinner fiber is helically wound to yield a thicker fiber that would appear as a bead (green haze) when viewed in the light microscope. Yellow struts represent fiber-fiber interactions, for example, between chromatin cross-linking proteins, that could stabilize the folded structure (16). The diameter of the bead detected by the light microscope would depend on the thickness of the underlying fiber and the size of the hollow core. If the underlying fiber were 30 nm in diameter, then the hollow helical core would need to be ∼500 nm in diameter to match the measured bead diameter of 0.4 to 0.8 μm. However, if the underlying fiber were thicker, for example, 100 nm (4), then the hollow core would need to be only twice as large as the fiber diameter, namely, ∼200 nm. (b) Local disruption of fiber-fiber interactions would lead to unraveling of the fibers and would yield a light microscope image of two adjacent beads. (c) Radial-loop models (14, 27) propose that the chromatin fiber forms loops attached to a protein scaffold (yellow spheres). A cluster of loops would be imaged as a bead by light microscopy. If a typical loop were 80 kb, then the radius of the bead would correspond to 40 kb. Assuming that loops are formed from 30-nm fibers which are ∼40-fold compacted, the bead radius would be 40 kb × (1/40) × (0.34 nm/bp), or 0.34 μm, in reasonable agreement with the measured bead diameter of 0.4 to 0.8 μm. (d) The structure shown in panel c could decondense into two beads by local detachment of a loop from the scaffold, but in order to elongate into two beads, the scaffold would have to either fragment (as shown here) or stretch.
References
- Andersson, K., B. Bjorkroth, and B. Daneholt. 1980. The in situ structure of the active 75S RNA genes in Balbiani rings of Chironomus tentans. Exp. Cell Res. 130:313-326. - PubMed
- Beard, D. A., and T. Schlick. 2001. Computational modeling predicts the structure and dynamics of chromatin fiber. Structure (Cambridge) 9:105-114. - PubMed
- Belmont, A. S., S. Dietzel, A. C. Nye, Y. G. Strukov, and T. Tumbar. 1999. Large-scale chromatin structure and function. Curr. Opin. Cell Biol. 11:307-311. - PubMed
MeSH terms
Substances
LinkOut - more resources
Full Text Sources