Genetic reporter systems are necessarily limited in the kinds of information they can provide. The most obvious limitation is that the insides of cells are complex. The phenotype of a gene fusion protein will be determined not only by the property we wish to assay, but also by other effects of the intracellular environment on the chimeric gene product, including the following: protein folding in the cells used (usually E. coli or Saccharomyces cerevisiae), sensitivity to intracellular proteolysis, appropriate localization within the cell, and interactions with other cellular components. The strength of interactions needed to confer repressor activity is clearly dependent on the level of expression of the fusion protein. For repressor fusion proteins, our ability to detect assembled or weakly bound oligomers is limited by the fact that the N-terminal domain by itself will confer repressor activity at high expression levels. At the other extreme, for very tight dimers or tetramers only the most drastic mutations might lose enough activity to give a detectable change of phenotype.The significance of activity in the λ repressor system, or in similar genetic oligomerization reporters based on LexA repressor [4xConstruction, purification, and characterization of a hybrid protein comprising the DNA binding domain of the LexA repressor and the Jun leucine zipper: a circular dichroism and mutagenesis study. Schmidt-Dorr, T., Oertel, B.P., Pernelle, C., Bracco, L., Schnarr, M., and Granger, S.M. Biochemistry. 1991; 30: 9657–9664Crossref | PubMedSee all References][4], 434 repressor [15xDimerization of leucine zippers analyzed by random selection. Pu, W.T. and Struhl, K. Nucleic Acids Res. 1993; 21: 4348–4355Crossref | PubMed | Scopus (20)See all References][15], AraC protein [5xFunctional domains of the AraC protein. Bustos, S. and Schleif, R. Proc. Natl. Acad. Sci. USA. 1993; 90: 5638–5642Crossref | PubMedSee all References][5] and LuxR protein (DM Sitnikov, JC Hu and TO Baldwin, unpublished data) is necessarily limited. Genetic methods are not meant to be a substitute for rigorous biochemistry in determining oligomerization states and binding constants. That said, the ‘awesome power’ of genetics lies in the speed with which one can test the properties of astronomically large numbers of amino acid sequences. The need to test such numbers of sequences arises in searching for specific genes in large genomes, and our results suggest that dominant negative selections can be used to find novel partners for known dimers. Moreover, the ability to test large numbers of sequences has interesting implications for protein design. Fusion methods such as the ones described here will allow us to test populations of sequences that represent a family of guesses as to what is needed to form a desired structure. Imitating the evolution of naturally occurring proteins, we cannot only allow selection to identify the best candidates from such populations, but also allow additional rounds of mutagenesis and selection to refine their design.