Introduction

Despite the rapid progress in sequencing eukaryotic genomes, our current abilities to interpret sequence information are still limited. Progress is expected from knowledge about the functional links between nuclear architecture and gene expression patterns on one hand and the establishment and maintenance of expression patterns during development on the other. Along these lines, the principles are explored that account for the compartmentalization of replication and transcription machineries within the nuclear compartments. These compartments assemble factors to an extent enabling protein-protein and protein-DNA interactions, and they serve the integration of regulatory signals into specific signal transduction pathways.

During recent years, epigenetic, chromatin-activating principles have entered the stage (review: Hake et al. 2004) together with the elements that delimit differentially regulated domains, so-called genomic insulators and/or boundary elements (Goetze et al. 2005). New evidence has emerged to help understand the role of chromosome territories (CTs), i.e. the structural equivalent of metaphase chromosomes at interphase, together with the interchromatin domain compartment (ICD), originally interpreted as a chromatin-free channel system in between the CTs. Ultimately, such a simplistic model was not consistent with the high frequency of complex chromosomal aberrations, which indicated the presence of inter-chromosome contacts within this space (Bode et al. 2000a; Branco and Pombo 2006).

When Kanda et al. (1998) stained the entire chromatin compartment in living cells, using histone H2b-GFP fusions they could localize putative factor storage sites, such as speckles, Cajal bodies and PML bodies, to extended portions of the interchromatin space, and they demonstrated that chromatin loops can in fact expand into this compartment. This led to the view that active genes interact with the transcriptional machinery only if they are positioned at the surface of CTs or on its looped extensions. Transcriptionally "potentiated" (otherwise called "poised") genes such as the quiescent, but inducible type-I interferon genes were found in a related position (Winkelmann, 2007). Quiescent genes on the other hand were thought to reside within the CTs. This model had to be refined once more when transcription and splicing could not only be observed at the periphery, but also appeared to extend into the territories. This was later ascribed to a highly folded CT structure that still permits access to certain genes that line this interior ICD-channel system (Albiez et al. 2006, Branco and Pombo 2006).

The question if a polymer meshwork, a so-called "nuclear matrix" or "nuclear scaffold", is an essential component of the in vivo nuclear architecture is still a matter of debate (Martelli et al. 2002). While there are arguments that the relative position of CTs may be maintained due to steric hindrance or electrostatic repulsion forces between the apparently highly structured CT surfaces, such an idea has to be reconciled with the following pilot observations:

• When Maniotis and colleagues (1997) "harpooned" nuclei, they could pull out all of the nucleoplasm on a string in interphase and all the chromosomes on a string in metaphase. Depending on the presence of Mg++, they observed unwinding and rewinding of these structures. This effect was lost upon mild DNAse treatment, indicating that the structure of DNA and its associated scaffolds are responsible for this phenomenon;

• Ma et al. (1999) treated cells in situ with the classic extraction procedures that are otherwise used to isolate the nuclear matrix. Chromosome-painting techniques clearly demonstrated that territories remained intact up to the point where a minor subset of acidic nuclear matrix proteins was released - potentially those proteins that governed their association with a nuclear skeleton.

The existence of a nuclear skeleton was first proposed about 60 years ago (Zbarsky and Debov 1948), and methods for the preparation of such an entity have been developed and refined ever since (reviewed by Martelli et al. 1996). We will maintain the idea that a nuclear skeleton acts as a dynamic support for many specialized reactions as the most suggestive model to guide the reader through this review. This concept will rationalize current efforts that are dedicated to inhibitors affecting the interaction of distinct transcription factors with the protein components of such a matrix and its associated DNA elements.

After a brief overview of the architectural principles of eukaryotic genomes, our discussion will deal with the properties of certain DNA regions that can serve as scaffold/matrix attachment regions (S/MARs), which are DNA elements with a well-established spectrum of biological activities. In this context, we will address the dynamic properties of prominent constitutive and facultative fibre-forming protein scaffold constituents and continue with factors that associate with the relevant protein or DNA interaction partners. Regarding the first class, we emphasize the lamins and hnRNPs and their functional interactions. For the second class, we will focus our attention on those examples that hold promise to either assist diagnosis or to lead to administrable pharmaceuticals. Here the ubiquitous poly(ADP-ribosyl) polymerase (PARP-1) and the cell-specific factor SATB1 will serve as the cores within networks of multiple interacting factors with a relation to the scaffold, to S/MARs or to both. It is anticipated that these paradigms will strengthen work at the verge of in vivo and in vitro studies all the more as these projects can now be guided and coordinated by up-to-date system biology approaches (examples are the inserts in Figs. 1 and 2).

0 0

Post a comment