DNA binding site preferences and transcriptional activation properties of the Arabidopsis transcription factor GBF1.

Abstract

The G-box is a cis-acting element found within the promoters of many plant genes where it mediates expression in response to a variety of different stimuli. This palindromic DNA motif (CCACGTGG) is composed of two identical half sites, the base pairs of which we have numbered -4 to +4 (numbering from 5' to 3'). Both half sites are involved in the binding of the bZIP protein GBF1, a member of the GBF family of Arabidopsis thaliana. Here we demonstrate using the random binding site selection method that GBF1 interacts with, in addition to the palindromic G-box, other DNA motifs that fall into seven distinct groups. All groups share the ACGT core sequence, common to most DNA motifs bound by plant bZIP proteins so far characterized. Our studies demonstrate that a high affinity GBF1 binding site is further defined by the following two parameters: first, all sites contain a G residue at position +3 (as in ACGTG) and secondly, only certain base pair combinations are allowed at positions -4, -3 and +4. Two of the identified groups (TGACGTGG and TGACGTGT) contain the base pairs TG at positions -4 and -3 and hence resemble the binding sites of another class of plant bZIP proteins (TGACGT/C binding proteins). However, GBF1 only interacts with the TGACGT sequence if the two 3' distal nucleotides (positions +3 and +4) are occupied by GG or GT. These data define the differences between a G-box binding protein and TGACGT/C binding proteins. The N-terminal domain of GBF1 is defined by a high proline content. Such regions were also identified in proteins related to GBF1. We demonstrate that this N-terminal proline-rich domain of GBF1, when fused to a heterologous DNA binding domain, stimulates transcription in both plant protoplasts and mammalian cells. These extensive DNA binding studies and the characterization of the GBF1 activation domain will facilitate both the identification of regulatory elements and the in vivo function of GBF1.