The Holliday junction is a central intermediate in the process of genetic recombination. The position of its branch-point can relocate through an isomerization known as branch migration. This migration occurs because the branch-point is flanked by homologous symmetry. All attempts at modeling the kinetics of branch migration have relied on the assumption that branch migration minima are sequence-independent. We have tested that assumption here, using a competition assay based on symmetric immobile branched junctions; these are junctions that cannot undergo branch migration, despite the fact that they are flanked by homology. The assay used is predicated on the non-association of strands displaced in the assay; we have tested this assumption, and have performed our experiments under conditions where we know that it is true. We have measured the free energy of relocating a branched junction from a fixed non-homologous sequence to all possible dimeric symmetric sequences. We find that the assumption of sequence-independence is often valid, but that it is not universally true. We find that the flanking sequences can have a marked effect on the free energy measured, both for extensions of symmetry and for reversals of flanking nucleotides. We have varied the temperature in our experiments, and have derived both enthalpies and entropies for the different sequences. The entropies are largely unfavorable, whereas the enthalpies are largely favorable; regardless of the signs of these quantities, we see that this is another system where enthalpy-entropy compensation is operative.
Copyright 1998 Academic Press.