A nuclear factor that binds to a dyad-symmetric sequence with a CGTCA motif in the 5'-upstream region of the sweet potato beta-amylase gene.

A nuclear extract from petioles of sweet potato protected several sites in the 5'-upstream region of a gene for beta-amylase from DNase I digestion. One of these sites, located at a region around 800-base pairs upstream from the transcription start site, having an imperfect palindromic sequence of CGTCACGTCACG, was designated the R-box. The site contained tandemly duplicated CGTCA sequences, referred to below as 5'- and 3'-CGTCA. Competition experiments in gel mobility shift assays with mutant R-box oligonucleotides indicated that mutations in bases outside the 3'-CGTCA of the R-box do not severely affect the binding. By contrast, single-base substitutions in any one base of the 3'-CGTCA greatly abolished the binding even when the mutated R-boxes contained intact 5'-CGTCA. However, oligonucleotides with mutations in the 3'-CGTCA had the ability to bind the nuclear factor when additional mutations were introduced to create a partially palindromic sequence containing the CGTCA sequence in its 3'-half on the opposite strand. These results indicate that the CGTCA sequence alone is not sufficient for the binding of the R-box binding factor (RBF) and that the RBF binds to the sequence with partial dyad symmetry that contains the CGTCA motif in its 3'-half. The optimum sequence for the binding of the RBF is suggested to be a palindromic octameric sequence TGACGTCA, which is identical to the consensus sequence of the cAMP-responsive element (CRE) of animal genes. Bacterially produced HBP-1b of wheat bound to the R-box, and its binding to mutated R-boxes was similar to that of RBF, suggesting that the RBF belongs to a family of bZIP-type plant nuclear factors that bind to CGTCA-related sequences. However, several differences between the RBF and HBP-1b were also noted.