Fig. 2. Mad and Med directly repress sal expression in the haltere through binding sites in the sal1.1 CRE. (A) Schematic of the Drosophila sal1.1 CRE with Ubx binding sites 1-7 (red circles). Putative Med (green box) and Mad (yellow box) binding sites are located between Ubx binding sites 5 and 6 (red boxes). Binding site mutations introduced into the M1 site are indicated below the wild-type sal1.1 sequence. The binding of proteins to the probes was quantitated with ImageQuant software to assess the effect of mutations on their affinity for specific sites. (B-I) The effect of mutations in the putative M1 site on Mad and Med binding in vitro and their effect on reporter gene expression in vivo are aligned in columns. (B) Electrophoretic mobility shift assays (EMSAs) with GST-MedMH1. In each set of lanes, the protein concentration increases from left to right. The point mutation at bp 808 (lanes 5-8) eliminates Med binding as compared with the wild-type M1 site (lanes 1-4). Mutations at positions 812 (lanes 9-12), 813 (lanes 13-16) and 814 (lanes 17-20) have little to no effect on Med binding. (C) EMSAs with GST-MadN on wild-type and mutant M1 probes. In each set of lanes, the protein concentration increases from left to right. Each point mutation (808, 812, 813 and 814) causes a decrease in the strength of Mad binding (2.2-, 2.7-, 2.6- and 3.6-fold, respectively) as compared with the wild-type sequence (compare lanes 25-29 with 30-49). Combining the four point mutations in kM1 causes an 8.6-fold reduction in Mad binding affinity (lanes 50-54). (D-I) Haltere imaginal discs of transgenic sal1.1 reporter lines that are either wild-type or that carry mutations in the M1 site, immunolabeled for lacZ expression.