The gene encoding human glucokinase (ATP:D-hexose 6-phosphotransferase, EC 2.7.1.1), a major component of glucose sensing in pancreatic islet beta-cells, was isolated and characterized. The gene was shown by Southern blotting to exist as a single copy in the genome which mapped to chromosome 7p. It contained 12 exons including two tissue-specific first exons, one active in islet beta-cells (1B), and the other active in liver (1H), and one optional cassette exon which was expressed as a minor form in the liver. Thus the three previously reported isoforms of glucokinase mRNA were the result of tissue-specific activation of separate liver and islet promoters and subsequent alternative splicing events. Eleven exons, including 1H and the optional cassette exon 2A, were scattered over 16 kilobase (kb) in the genome, while exon 1B was separated from the rest by at least 20 kb. Although the islet promoter was found to lack a TATA box, a major transcript from the islet promoter was mapped 486 nucleotides upstream of the translation initiation site. The presence in the islet glucokinase promoter of the potential control element GCCACCAG, a homology of the regulatory element present in both human insulin (GCCACCGG) and rat insulin (GCCATCTG) genes, implied a possible tissue-specific regulatory role of this element. The liver promoter was found to contain a TATA box-like sequence, and transcription was initiated predominantly at 168 nucleotides upstream of the translation initiation site of the major isoform. A new highly polymorphic microsatellite, composed of a compound imperfect dinucleotide repeat [GT]15[GA]8CA[GA]7CA[GA]3AA[GA]2, was mapped 6 kb upstream of islet exon 1. A polymerase chain reaction-based assay was developed, and seven different sized alleles were identified in American Blacks. The sequence information reported here, along with the new polymorphic marker, will make it possible to clarify the molecular basis of potential glucokinase defects in noninsulin-dependent diabetes mellitus patients and may further elucidate the nature of genetic susceptibility to the development of this common metabolic disease.