Paired-Homeodomain Transcription Factor PAX4 Acts as a Transcriptional Repressor in Early Pancreatic Development

Abstract

The paired-homeodomain transcription factor PAX4 is expressed in the developing pancreas and along with PAX6 is required for normal development of the endocrine cells. In the absence of PAX4, the numbers of insulin-producing β cells and somatostatin-producing δ cells are drastically reduced, while the numbers of glucagon-producing α cells are increased. To gain insight into PAX4 function, we cloned a full-length Pax4 cDNA from a β-cell cDNA library and identified a bipartite consensus DNA binding sequence consisting of a homeodomain binding site separated from a paired domain binding site by 15 nucleotides. The paired half of this consensus sequence has similarities to the PAX6 paired domain consensus binding site, and the two proteins bind to common sequences in several islet genes, although with different relative affinities. When expressed in an α-cell line, PAX4 represses transcription through the glucagon or insulin promoters or through an isolated PAX4 binding site. This repression is not simply due to competition with the PAX6 transcriptional activator for the same binding site, since PAX4 fused to the unrelated yeast GAL4 DNA binding domain also represses transcription through the GAL4 binding site in the α-cell line and to a lesser degree in β-cell lines and NIH 3T3 cells. Repressor activity maps to more than one domain within the molecule, although the homeodomain and carboxyl terminus give the strongest repression. PAX4 transcriptional regulation apparently plays a role only early in islet development, since Pax4 mRNA as determined by reverse transcriptase PCR peaks at embryonic day 13.5 in the fetal mouse pancreas and is undetectable in adult islets. In summary, PAX4 can function as a transcriptional repressor and is expressed early in pancreatic development, which may allow it to suppress α-cell differentiation and permit β-cell differentiation.

ACKNOWLEDGMENTS

The first two authors contributed equally to this work.

We thank J. Wang, J. Leong, Y. Zhang, and J. Lau for technical assistance, members of the Michael German laboratory for helpful comments and criticisms, and Gabriele Bergers for advice on PCR strategies for cloning Pax4. H.C.E. and S.B.S. are recipients of Juvenile Diabetes Foundation International postdoctoral fellowships. This work was supported by National Institutes of Health grant DK41822.