First, multiple A/T-rich sequences, called "A elements", each of which contains a "TAAT" reocognized by homeodomain proteins.
[1] Second, two so-called "C elements" – C1 located 107–118 base pairs before the transcription start site; C2 311–317 base pairs before the start site.
C2 (also called the "pancreatic islet cell enhancer sequence" or "PISCES") is bound by PAX6.
[1] Lastly, several "cyclic AMP response elements" with sequence TGACGTCA that binds CREB.
[1] In humans, a "Z-element" resides 243–292 base pairs before the start site and binds a complex called ZaI, as well as PDX-1 and MafA.