We did the domain of HAP2 that mediates subunit asiation In the heteromeric CCAAT-binding complex, first by genetic mutational analysis and then by structural studies. The mutational data suggest that a very short region in HAP2 mediates protein-protein association and that the structure of this domain is likely to be an a-helix. The CD analyses ofa 15-residue synthetic oligopeptide covering this region confirm this surmise. The oligopeptide indeed formed an unusuafly thermal stable a-helix in aqueous solution. Eight amino acids that lie along one face of this helix, including three arginines, are found to be critical for protein-protein association. The partner that interacts with this helical motif is likely to be another subunit in the HAP complex, since the CCAATbinding factor is shown to contain one molecule of HAP2. Our results suggest that very short regions in proteins can encode precise structures and mediate stable and specific proteinprotein recognition and interactions.The specificity and stability of protein-protein interactions are critical determinants in the assembly of macromolecular complexes consisting of protein subunits. One well-studied example of these complexes involves the transcription factors Fos and Jun (for review, see refs. 1 and 2). Jun-Jun homodimers and Fos-Jun heterodimers will form readily and are active in cells, whereas Fos-Fos dimers do not form (3-5). Inspection of a family of transcription factors with DNA-binding domains like those of Fos and Jun led to the leucine zipper hypothesis (6).A heteromeric transcriptional activator that was identified early on is the yeast (Saccharomyces cerevisiae) CCAATbox binding factor HAP2/3/4 (7-9). Both HAP2 and HAP3were present in a gel-shift complex with a CCAAT boxcontaining probe (9), and HAP2 and HAP3 remained stably associated in the absence of DNA (8) vivo (10, 14).Here we describe the dissection of the SAD of HAP2 by genetic mutational and biophysical structural analyses. Our findings delimit the region and pinpoint which residues in HAP2 are critical for its function. The mutational data suggest that a very short region in HAP2 is likely to form an a-helix which mediates assembly. Our findings show that 8 aa along one face of this helix dictate the specificity and at least part of the stability of interaction between HAP2 and its interacting partner. Further, a 15-aa synthetic peptide covering this region showed a typical a-helical circular dichroism (CD) spectrum. We conclude that very short peptide regions can encode precise structures and mediate stable and specific protein-protein recognition and interactions in macromolecular complexes.
MATERIALS AND METHODSStrains, Plasmids, and f-Galactodase Assays. The S. cerevisiae strain BWG 1-7a (MAT a leu2-3,112 his4-519 adel-100 ura3-52) or isogenic derivatives were used in all experiments described. HAP2 complementation was assayed by using strain JO1-la (14). Vectors pRB1155 (16), pJO323 (14), and pLGA265UP1 (17); plasmid pXY3236 (10); and the assay for /3-galactosidase ac...