Fig. 3. Identification of protein binding sites in the AP-1/Cx activator region. A, The DNA sequences used to make probes for EMSAs are indicated by bars below the sequence for the first 60 bp of the 90 bp AP-1/Cx activator region. The AP-1 consensus sequence is boxed, and mutations introduced into the probes are shown in lower case letters. The sequence alteration used to mutate the AP-1 site (
TGACTAA to
GTACTAA on the antisense strand) has been demonstrated previously to disrupt AP-1 binding and activity in another promoter (Lee et al., 1991). The location of novel putative transcription factor binding sites are indicated by barslabeled as Cx1, Liv1, and Cx2.B, EMSAs with radiolabeled probe A and nuclear extract from postnatal day 5 rat cerebral cortex. The first lane contains only the radiolabeled probe and nuclear extract, whereas the additional lanes include either a 100- or 500-fold molar excess of unlabeled wild-type probe A or probe A with mutations 1, 2, or the combination of mutations 1 and 2. P stands for a polylinker DNA with no similarity to the probe A sequence. If a mutation affects protein binding, then the competitor with that mutation should compete less effectively or not at all. AP-1 specific-bands are labeled withlarge arrows, and smaller arrows indicate additional sequence specific bands that do not appear to depend on the AP-1 sequence. C, EMSAs with probe A and liver nuclear extract. D, EMSAs with probe B and nuclear extract from postnatal cerebral cortex. Numbers designate probe B with mutations 2, 3, 4, 5, or 6. Competitor probes were used at a 500-fold molar excess. Cx1 refers to the binding site defined by mutations 2 and 3, whereas Cx2 refers to the binding site defined by mutations 5 and 6. E, EMSAs with probe B and liver nuclear extract. Competitors were used at a 50-fold molar excess in this case. Liv1 refers to the binding site defined by mutations 3, 4, and 5.