Fig. 1.

The MARE sites are crucial for mammalian insulin gene activity. (A) The MARE-related sequences within human and rat insulin II gene enhancer are shown in relation to the consensus MARE (see ref. 31) binding site. Underlined letters indicate sequences that deviate from the consensus. Strong binding (++++) to no binding (NB) was found in gel-shift experiments conducted with the labeled insulin MARE probes and βTC3 nuclear extracts (see Fig. 5, which is published as supporting information on the PNAS web site). The authenticity of the MafA-containing complex was determined by αMaf antisera supershift analysis. (B) βTC3 cells were transfected with the wild-type and MARE site mutants of human (-251-LUC) and rat II (-238-LUC) insulin enhancer- and promoter-driven luciferase reporters. Enhanced BETA2:E47 activator-binding at the E1 site contiguous to MARE1 presumably masks the effect on rat -238 mutant activation (data not shown); although, because we were unable to generate mutants at this site that did not influence BETA2:E47 binding, it is possible also that this element is not an important MafA regulatory site. Shown are normalized activity values ± SEM, compared with wild-type insulin-reporter activity.