Identification and characterization of androgen response elements

The androgen receptor (AR) has a DNA-binding domain that consists of two zinc coordinating modules. While residues of the first module make most of the sequence-specific contacts, the second module functions as a homodimerization interface (1). This explains why the androgen response elements (AREs) are organized as two 5'-AGAACA-3'-like motifs separated by three basepairs (2). AREs can be located near the promoters of androgen-responsive genes, but are also at considerable distances either upstream or downstream, so the initial steps in locating AREs can be challenging. Traditionally, AR-binding sites were identified by DNA cellulose competition assays (3) or by in vitro footprinting (4). However, the advent of the chromatin immunoprecipitation assays made it possible to identify genomic fragments to which the AR binds either directly or indirectly (5). To enable identification of AREs in such genomic fragments, we developed an in silico approach involving a weight matrix based on all known AREs (6). This will point out candidate AREs, which will still need experimental validation involving a direct interaction assay and a transactivation assay. We describe here the methods most fit to describe an ARE: the electrophoretic mobility shift and the transactivation assays.