Domain organization of eukaryotic genome

INTRODUCTION It is a well-established fact that transcription of eukaryotic genes is regulated at several levels. One of these levels is represented by mechanisms that operate within relatively large genomic domains. Active domains are preferentially sensitive to DNase I (for review see Gross and Garrard, 1988). They could be several dozens kb long and usually include a group of functionally related genes (e.g. globin genes). The borders of the domains are quite sharp. Recent fmdings indicate that transcrip- tional status of a whole domain can be determined by a so-called dominant control region (DCR). The best known example is the dominant control region of the domain of human ~-globin genes (Grosveld et al., 1987; Groodin et al., 1989). A relatively short region located in the upstream area of this domain drastically influences the functional status of a genomic region that is more than 100 Kbp long (in particular: transcriptional status of this region, sensitivity to DNase I, methylation pattern and replication thning). Existence of functional domains in eukaryotic genome raises a question whether these domains correspond to some stmctural units in chromatin. There are at least three levels of DNA packaging in chromatin, namely nucleosomal level, solenoid level and DNA loops level (for review see Reeves, 1984). Only the structural units of the third level of DNA compactization (i.e. DNA loops anchored at the chromosomal skeleton) have the size comparable with the size of the above discussed functional genomic domains. The DNA loops were first recognized when sedimen- tation properties of high-salt extracted nuclei were studied (Cook and BrazeU, 1980). It turned out that after extraction of histones the DNA was topologically constrained and could be either supercofled or relaxed depending on some external parameters. The average size of constrained domains determined by several independent methods was found to be of the order of 100 kb (for review see Hancock and Huges, 1982). An attractive hypothesis has been suggested that individual loops correspond to the functional genomic domains (Cook, 1989) In the frame of this hypothesis it was easy to explain how the preferential sensitivity to DNase I could be maintained within the limited though large areas of the genome and why the borders of the sensitive regions are rather sharp. The most simple suggestion appeals to the possibility of reversible distraction of solenoid level of DNP packaging within the individual loops. All the above mentioned considerations draw particular attention to the problem of sequence specificity of DNA interaction with the chromosomal skeletal elements. Indeed, if loops correspond to functional genomic domains, one could expect that borders of the loops have specific positions on the DNA chain and and are located