Ion concentration and temperature dependence of DNA binding: Comparison ofPurR and LacI repressor proteins

Purine repressor (PurR) binding to specific DNA is enhanced by complexing w ith purines, whereas lactose repressor (LacI) binding is diminished by inte raction with inducer sugars despite 30% identity in their protein sequences and highly homologous tertiary structures. Nonetheless, in switching from low- to high-affinity DNA binding, these proteins undergo a similar structu ral change in which the hinge region connecting the DNA and effector bindin g domains folds into an a-helix and contacts the DNA minor groove. The diff erences in response to effector for these proteins should be manifest in th e polyelectrolyte effect which arises from cations displaced from DNA by in teraction with positively charged side chains on a protein and is quantitat ed by measurement of DNA binding affinity as a function of ion concentratio n. Consistent with structural data for these proteins, high-affinity operat or DNA binding by the PurR purine complex involved similar to 15 ion pairs, a value significantly greater than that for the corresponding state of Lac I (similar to6 ion pairs). For both proteins. however, conversion to the lo w-affinity state results in a decrease of similar to2-fold in the number of cations released per dimeric DNA binding site. Heat capacity changes (Delt aC(p)) that accompany DNA binding, derived from buried apolar surface area, coupled folding, and restriction of motional freedom of polar groups in th e interface, also reflect the differences between these homologous represso r proteins. DNA binding of the PurR guanine complex is accompanied by a Del taC(p), (-2.8 kcal mol(-1) K-1) more negative than that observed previously for LacI (-0.9 to -1.5 kcal mol(-1) K-1), suggesting that more extensive p rotein folding and/or enhanced structural rigidity may occur upon DNA bindi ng for PurR compared to DNA binding for LacI. The differences between these proteins illustrate plasticity of function despite high-level sequence and structural homology and undermine efforts to predict protein behavior on t he basis of such similarities.