Many metabolic processes in cells depend on the self-association of proteins into larger complexes. However, there are currently few techniques for experimentally manipulating protein self-association in a controlled fashion. Fusing proteins to the DNA binding domain of the yeast transcription factor HAP1 (heme activation protein 1) could allow proteins to be dimerized in a DNA-dependent fashion. HAP1 binds as a monomer to DNA sites containing the sequence CGGN₃TA (N is any nucleotide) but binds cooperatively as a dimer to DNA sites containing two CGG sequences spaced by 6 base pairs. Using gel filtration and fluorescence anisotropy, we find that an 11 kDa fragment of HAP1 containing the DNA binding domain causes a DNA-dependent dimerization of the larger 43 kDa maltose binding protein (MBP). In the absence of DNA, the MBP-HAP1 fusion protein (HAP1 fused to the C-terminus of MBP) elutes as a monomer by gel filtration, later than a 68 kDa protein standard. In the presence of DNA containing the CGGN₃TA sequence, MBP-HAP1 and DNA co-elute as a complex with size consistent with 1:1 binding. However, in the presence of DNA containing two appropriately spaced CGG sequences, MBP-HAP1 and DNA co-elute as a complex that is larger than 68 kDa, consistent with a forced dimerization of the MBP. The fluorescence and gel filtration assays are being used to quantitate the effects of ionic strength and steric interference on the DNA binding of HAP1 fusion proteins.