Fig. 6. The ISL sites in the Mef2c anterior heart field enhancer are specifically bound by ISL1. Isl1 was transcribed and translated in vitro and incubated with radiolabeled, double-stranded oligonucleotides spanning the Mef2c ISL-p site (lanes 9-14), the Mef2c ISL-d site (lanes 15-20) or a consensus ISL1 binding site from the rat Insulin I gene as a control (lanes 1-8). ISL1 efficiently bound to all three ISL sites (lanes 2, 10 and 16). Binding of ISL1 to the ISL-p site was specifically competed by excess unlabeled Mef2c ISL-p (I-p) (lane 11) and by excess unlabeled Insulin I ISL1 site (In) (lane 13) but not by a 100-fold excess of mutant ISL-p site (mI-p) (lane 12) or by a 100-fold excess of a mutant Insulin I site (mIn) (lane 14). Likewise, the binding of ISL1 to the ISL-d site was specifically competed by excess unlabeled ISL-d (I-d) (lane 17) and by excess unlabeled Insulin I site (lane 19) but not by a 100-fold excess of either mutant ISL-d (mI-d) (lane 18) or mIn control (lane 20). The Mef2c ISL-p and ISL-d sites also efficiently competed for ISL1 binding to the control ISL1 site from the Insulin I promoter (lanes 5 and 7), but mutant versions of the ISL-p and ISL-d sites were unable to compete for binding to the Insulin I site (lanes 6 and 8). In samples where in vitro translated ISL1 protein was not included (denoted by a minus sign in lanes 1, 9 and 15), an equal amount of unprogrammed reticulocyte lysate was included.