Understanding DNA Deformability by DNA Bending Studies with the IHF/HU Family of Architectural Proteins

Sequence-dependent DNA shape and deformability is critical to how proteins bind to DNA and kink, bend, or twist it to facilitate various cellular processes. However, many of the rules that govern local DNA shape/deformability remain largely untested, especially for severely deformed DNA. This thesis describes studies aimed at characterizing DNA deformability, utilizing a prokaryotic, architectural, DNA-bending protein Integration Host Factor (IHF) that binds to DNA both nonspecifically, to facilitate DNA packaging, and specifically, to form higher-order protein-DNA complexes for specialized functions. IHF has a remarkable ability to bend specific target sites on DNA by nearly 180 degrees over 35 base pairs, which it accomplishes by inducing two sharp kinks in the DNA. How flexible or rigid are these kinks was not known and the competition between the two binding modes was poorly characterized. Using fluorescence lifetime studies to measure Forster resonance energy transfer (FRET) efficiency between fluorescent probes attached to the ends of 35-mer DNA constructs, we showed that IHF-DNA complexes sample different conformations with varying degrees of DNA bends, indicating a highly dynamic complex. The distribution of conformations is modulated by DNA sequence variations, highlighting the finely tuned interactions between IHF and its many biological targets that keep the DNA bent (or not). Next, we used a “lattice-binding” model to simultaneously describe binding (measured by fluorescence anisotropy) and bending (measured by FRET) of IHF to specific and nonspecific DNA. These studies provided accurate measurements of the relative affinities in the two binding modes and revealed that specific and nonspecific binding likely occur simultaneously to the same DNA site at sufficiently high protein concentrations. Using laser temperature-jump spectroscopy, we established that transient Hoogsteen base pairing – a thermodynamically less favorable alternative to Watson-Crick pairing prevalent at flexible sites in duplex DNA – plays no role in facilitating DNA kinking by IHF. Finally, we performed high-throughput “SELEX” studies designed to identify highly “kinkable” DNA sequences using in vitro selection of high-affinity sequences for IHF from a pool of random DNA sequences. These SELEX studies provide a database of highly deformable sequences that can be used to improve current models of sequence-dependent DNA mechanics.