In order to investigate the mechanism(s) of plant responses to short-wavelength light, the regulation of chalcone synthase (CHS) expression has been analyzed. CHS catalyzes the first committed enzymatic step of flavonoid biosynthesis and is encoded in parsley (Petroselinum crispum) by a single gene whose expression is tightly controlled at the transcriptional level. Light is the primary external stimulus regulating the activity of the chs gene in leaf epidermis as well as suspension-cultured parsley cells. Analysis of
the chs promoter by in viva footprinting revealed four short sequences, designated Boxes I, II, Ill, and IV, that displayed light-induced protein contacts. Transient expression experiments in parsley protoplasts demonstrated that the four sequences are functionally relevant components of the chs promoter. These cis-acting elements are arranged in two light-regulatory units which are about 50 bp in length (LRU 1 containing Boxes I and II, LRU 2 containing Boxes Ill and IV). Each of them was shown to be sufficient for light responsiveness. Point mutation experiments defined a critical nucleotide sequence of seven bases (5'-ACGTGGC-3') within Box II of LRU 1. This heptameric sequence is also present in a closely related form in Box Ill of LRU 2. Nuclear extracts from suspension-cultured parsley cells contain a set of proteins which recognize the heptamer and related sequences. We isolated three parsley cDNAs encoding proteins which specifically bind to the 5'-ACGTGGC-3' sequence. Related sequences recognized by these "common plant regulatory factors" (CPRF-1, 2 and 3) contain an ACGT core motif which is present in similar sequence contexts in many cis-acting elements. Such ACGT elements (ACEs) are also of functional significance in a variety of other plant promoters, where they are involved in abscisic acid regulation, tissue- and development-specific gene expression as well as light responsiveness of rbcS promoters. The deduced amino acid sequences of all three ACGT-binding proteins revealed conserved basic and leucine-zipper domains characteristic of bZIP type DNA-binding proteins.