The structure and function of the 5'-flanking region of the mouse and human serotonin 1a receptor gene have been analyzed by RNA 5' end mapping, DNA-protein interaction, and transient expression assays. A large number of mRNA 5' termini, detected by mapping 5' ends from mouse brain RNA, were found dispersed over a region of about 700 base pairs flanking the receptor coding sequence. Consistent with the apparently heterogeneous pattern of transcription initiation, the flanking DNA sequence lacked typical TATA box elements and was rich in guanine and cytosine. The mouse and human 5'-flanking sequences were 63% homologus and similarly organized. A guanine-cytosine-rich DNA sequence motif related to the sequence 5'-GGGG(C/A)GGGG-3' was repeated within the 5'-flanking region and located at or near several mRNA 5' ends. This DNA sequence motif bound to proteins in a crude HeLa cell nuclear extract. A cDNA encoding a protein that interacts with this sequence was cloned and found to be the MAZ (Pur-1, Zif87) protein. The interaction between MAZ and the receptor gene 5'-flanking region proximal to the protein coding sequence was examined by DNase I footprinting, and four sites of MAZ interaction were identified. Three of the four MAZ binding sites also were shown to interact with transcription factor Sp1. Overproduction of MAZ or Sp1 in transient transfection assays increased expression directed by the human 5'-flanking sequence, although MAZ was substantially more effective. This result suggests that MAZ and Sp1 both participate in regulating expression from the serotonin 1a receptor gene promoter, and it raises the possibility that MAZ may act at a variety of promoters through the guanosine-cytosine-rich sequences generally thought to serve as binding sites for the Sp1 family of transcription factors. Analysis of one of the guanosine-cytosine-rich DNA sequences also revealed that it can serve as a transcription initiator sequence in vitro. This initiator sequence differs from previously characterized initiators and may represent a new class of this transcriptional control sequence.