21. Recognition and Organization of Specific DNA Structures by Human Chromatin-associating Factors

Masahiko Takahagi and Kouichi Tatsumi

Keywords: DNA junction, DNA binding protein, chromatin organization

Specific DNA structures are of interest as transient and activated states in the course of genetic processing. Since they occur at biological active sites, it is important to know their structures and metabolic processes. In particular, during DNA replication, recombination and repair, dynamic changes of DNA structure have been observed in both prokaryotes and eukaryotes. Many activities are known to link with specific DNA structures.

In this study, we focused on a peculiar DNA junction structure, the "Holliday junction", which is an intermediate formed during recombination. It is a well known target not only because of its processing factors but also because of its architectural components like chromatin-relating factors. For example, non-histone proteins HMG1/ HMG2 have a specific affinity for the Holliday-like junction as well as for DNA bending sites, and they possess a cooperative potential with histone H1 or provide substitutive roles for it on chromatin template. Also, histone H1 preferentially binds to Holliday-like DNA junctions. These facts suggest that a sort of DNA junction is topologically equivalent to the putative looping cross-over which HMG1/2 and H1 can recognize at nucleosome linker sites, indicating that Holliday-like junctions are structural analogues to putative DNA looping at nucleosomal linkers. Recently, a specific interaction with Holliday-like junctions has been identified in another functional factor, SWI/SNF complex, which is involved in chromatin organization for transcriptional regulation.

The biological significance of junction types of DNA makes it important to search for interactive proteins from human cell extracts. We separated major DNA binding proteins from cell nuclei through DNA affinity chromatographic procedures. In addition to five major chromatin-associating proteins, i.e. histone H1 nucleolin, hnRNP U, HMG1 and HMG2, DNA-dependent protein kinase catalytic subunit (DNA-PKcs) and Ku were identified as abundant elements. Gel mobility shifl assay with their purified preparations demonstrated that six of the proteins, excluding Ku, specifically bound to common junction types of DNA, suggesting that their targets could be shared and their functions may be associated.

To analyze the putative molecular organization on damaged DNA, we developed a technique using formation of the molecular aggregate as a probe as is the case of matrix attachment region binding proteins including histone H1, Topoisomerase II and lamins. We found that human nuclear extract formed a selective aggregation with damaged types of DNA. Further, to separate the relating proteins from contaminant proteins and nucleic acids, DNA affinity column chromatography was carried out. We detected a series of proteins to co-aggregate with denatured types of DNA. According to the results of amino acid sequencing and western blotting, we identified the major components as DNA-PKcs, Ku and nucleolin.

In order to confirm the potential to aggregate DNAs, their isolated preparation was tested. We observed an ability specific to nucleolin for single stranded DNAs, but not for double-stranded DNAs. Similarly, either DNA-PKcs or hnRNP U was able to aggregate single-stranded DNAs. This evidence emphasized that the three factors may be involved in the organization of common DNA targets.

The existence of a multivalent DNA target for major nucleoproteins leads us to ask how they are functionally specialized and how they interact at the target sites; these are topics for future study.