Assembly requirements for lactose repressor protein have been examined by substitutions in the leucine heptad repeat (LHR) tetramerization domain. All a and d positions of the LHR were substituted with Ile (II) or a positions with Leu and d positions with Ile (ILI). Additionally, identical substitutions made in the previously-constructed LacI/GCN4 LHR chimera, Izip were termed IzI and IzLI. To express these proteins, a system was designed employing the popular T7&phis; promoter without usual repression by lactose repressor protein. Cloning T7 gene 1 under arabinose repressor control on pBAD33 (Guzman et al., 1995) generated the plasmid pTara. Following co-transformation of pTara and expression plasmid, production of both T7 RNA polymerase and target protein was induced with L-arabinose. High expression levels of lactose repressor mutant proteins were obtained, and the general applicability of this system was confirmed using human p53 and Drosophila Ubx proteins. Proteins II and ILI demonstrated elution profiles from column chromatography typical of predominantly monomeric assembly, whereas IzI and IzLI eluted at comparable ionic strengths to wild type LacI. Gel filtration chromatography indicated that IzI is of the correct size for hexameric protein, and IzLI is in an equilibrium between tetramer and monomer. IzI and IzLI bound neither wild type operator nor completely palindromic operator sequence with specific affinity, and no decrease in operator binding was observed for either protein with inducer. IzI and IZLI bound inducer with wild type affinity, but did not exhibit the wild type reduction in affinity normally observed at pH 9.2. Trypsin digestion studies indicated that both of these proteins were reduced to lower molecular weight bands than wild type lactose repressor and that IzLI is digested to these lower weight states more rapidly than IzI. Urea denaturation studies yielded midpoints of denaturation for both proteins lower than wild type lactose repressor protein, and less steep slopes for the unfolding/dissociation transitions. These data are collectively consistent with folding to a wild type monomeric tertiary structure but anomalous assembly to dimer and higher order oligomers. The changes in dimer assembly apparently have consequences for operator recognition and cooperativity of inducer binding.