Amino acids necessary for DNA contact and dimerization imply novel motifs in the papillomavirus E2 trans-activator

The bovine papillomavirus E2 protein regulates viral transcription by binding as a dimer to the DNA sequence ACCGN4CGGT. The dimerization and DNA-binding properties are localized within its carboxy-terminal 85 amino acids (325-410). Utilizing random mutagenesis coupled with phenotypic selection in yeast, functionally important amino acids in the DNA-binding domain were identified. Four trans-activation defective point mutants within a short segment (amino acids 337-344) were DNA binding defective but dimeric. The mutation of a conserved tryptophan to serine also eliminated DNA binding, but loss of dimerization was implicated because addition of dimeric monoclonal antibody complemented this defect. A simple assay for E2 dimerization was developed using UV irradiation to produce an interchain cross-link within a dimer. No heterodimeric complexes were formed when pools of E2 of varying lengths were mixed, and only proteins with tryptophan at position 360 could be UV cross-linked. Peptide mapping of irradiated E2 protein localized the cross-link to an 18-amino-acid region bracketing this tryptophan. Substitutions for this tryptophan demonstrated the requirement for a hydrophobic residue at this position, but surprisingly, even alanine was functional. Replacement of this tryptophan with three polar amino acids or glycine eliminated DNA-binding activity, but addition of dimeric monoclonal antibody restored this function. The amino acids that were identified as being involved in DNA contact and dimerization imply that these functions are mediated by novel binding motifs.