HMG box proteins interact with multiple tandemly repeated (GCC)n (GGC)m DNA sequences

A number of tandemly repeated DNA sequences have the ability to form hairpin structures by forming non-standard base pairs. When (GCC)15 and (GGC)15 strands are annealed together, the expected duplex is the only product. However, when (GCC)15 is annealed with (GCC)10, depending on the relative concentrations, up to five complexes can be detected in native gels. Three of these species are susceptible to limited digestion by Exo VII, suggesting they are duplexes containing single stranded tails. The remaining two bands are resistant to the enzyme, and have low mobility on native gels, consistent with branched structures. The latter complexes bind HMG box proteins, members of a highly abundant class of non-histone proteins of the nucleus. These proteins, modeled in this study by the second box fragment from rat HMG1, HMGb, interact strongly with branched or chemically modified DNA, relative to normal duplexes. The expansion of triplet repeats in genomic DNA is associated with tumor formation as well with a variety of heritable neurologiocal disorders. It is our thesis that the stability of branched intermediate structures that arise in replication of these sequences and promote expansion can be influenced directly by the presence of two highly abundant proteins in the cell nucleus: the HMG box proteins, HMG1/2, and the histone H1, which associates with HMG1/2.