Overproduction and improved strategies to purify the threenative forms of nuclease-free HU protein

The heterodimeric HU protein was isolated from Escherichia coli as one of the most abundant DNA binding proteins associated with the bacterial nucleoid. HUalphabeta is composed of two very homologous subunits, but HU can also be present in E. coli under its two homodimeric forms, HUalpha(2) and HUbeta(2). This protein is conserved either in its heterodimeric form or in one of its homodimeric forms in all bacteria, in plant chloroplasts and in some viruses. HU can participate, like the histones, in the maintenance of DNA supercoiling and in DNA condensation. This protein which does not recognize any specific sequence on double-stranded DNA, has been shown to bind specifically to cruciform DNA as does the eukaryotic HMG1 protein and to a series of structures which are found as intermediates of DNA repair, e.g., nick, gap, 3'overhang, etc. The strong binding of HU to these diverse DNA structures could explain, in part at least, its pleiotropic role in the bacterial cell. To understand all the facets of its interactions with nucleic acids, it was necessary to develop a procedure which allowed the purification of the three forms of HU under their native form and without the nuclease activity strongly associated with the protein. We describe here such a procedure as well as demonstrating that the three histidine-tagged HUs we have produced, have conserved the binding characteristics of native HUs. Interestingly, by two complementation tests, we show that the histidine-tagged HUs are fully active in vivo.