Advertisement
Research Article Free access | 10.1172/JCI2242
Laboratory of Molecular Endocrinology, Massachusetts General Hospital, Harvard Medical School, and Howard Hughes Medical Institute, Boston, Massachusetts 02114, USA.
Find articles by Stoffers, D. in: JCI | PubMed | Google Scholar
Laboratory of Molecular Endocrinology, Massachusetts General Hospital, Harvard Medical School, and Howard Hughes Medical Institute, Boston, Massachusetts 02114, USA.
Find articles by Stanojevic, V. in: JCI | PubMed | Google Scholar
Laboratory of Molecular Endocrinology, Massachusetts General Hospital, Harvard Medical School, and Howard Hughes Medical Institute, Boston, Massachusetts 02114, USA.
Find articles by Habener, J. in: JCI | PubMed | Google Scholar
Published July 1, 1998 - More info
The homeodomain transcription factor insulin promoter factor-1 (IPF-1) is required for development of the pancreas and also mediates glucose-responsive stimulation of insulin gene transcription. Earlier we described a human subject with pancreatic agenesis attributable to homozygosity for a cytosine deletion in codon 63 of the IPF-1 gene (Pro63fsdelC). Pro63fsdelC resulted in the premature truncation of an IPF-1 protein which lacked the homeodomain required for DNA binding and nuclear localization. Subsequently, we linked the heterozygous state of this mutation with type 2 diabetes mellitus in the extended family of the pancreatic agenesis proband. In the course of expressing the mutant IPF-1 protein in eukaryotic cells, we detected a second IPF-1 isoform, recognized by COOH- but not NH2-terminal-specific antisera. This isoform localizes to the nucleus and retains DNA-binding functions. We provide evidence that internal translation initiating at an out-of-frame AUG accounts for the appearance of this protein. The reading frame crosses over to the wild-type IPF-1 reading frame at the site of the point deletion just carboxy proximal to the transactivation domain. Thus, the single mutated allele results in the translation of two IPF-1 isoproteins, one of which consists of the NH2-terminal transactivation domain and is sequestered in the cytoplasm and the second of which contains the COOH-terminal DNA-binding domain, but lacks the transactivation domain. Further, the COOH-terminal mutant IPF-1 isoform does not activate transcription and inhibits the transactivation functions of wild-type IPF-1. This circumstance suggests that the mechanism of diabetes in these individuals may be due not only to reduced gene dosage, but also to a dominant negative inhibition of transcription of the insulin gene and other beta cell-specific genes regulated by the mutant IPF-1.