ANALYSIS OF MOLECULAR MECHANISMS CONTROLLING NEUROENDOCRINE CELL-SPECIFIC TRANSCRIPTION OF THE CHROMOGRANIN-A GENE

Chromogranin A (CgA), a member of the granin/secretogranin family of a cidic glycoproteins that play multiple roles in the process of regulat ed secretion of peptide hormones and neurotransmitters, is specificall y expressed in endocrine and neuroendocrine cells. We previously clone d and characterized the human (h) CgA gene and showed that nucleotides -55 to +32 relative to the transcriptional start site that contain a consensus cAMP element (CRE) and TATA-box motif were sufficient for ne uroendocrine cell-specific expression. Here, we examined the role of t he well conserved CRE in basal and cAMP-stimulated transcription in ne uroendocrine cells. Transient transfection studies with hCgA gene prom oter/chloroamphenical acetyl transferase (CAT) reporter constructs wer e conducted in a panel of neuroendocrine cell lines as well as in none ndocrine cell lines. Deletion or mutation of the CRE resulted in loss of neuroendocrine cell specific transcriptional activity. Mutation of a well conserved region (the TG-box) located between the CRE and the T ATA box had no effect or resulted in only a modest decrease in activit y. Mutation of the CRE in 5'-extended (-2300 to +32 and -700 to +32) c onstructs resulted in a 50-75% decrease in basal activity in neuroendo crine cells. This emphasized the importance of the CRE in basal transc ription and also suggested that other elements between -700 and -55 ma y act independently of the CRE to contribute to full basal activity in some neuroendocrine cells. Dibutyryl cAMP stimulated transcriptional activity in neuroendocrine cells, and this was abolished by mutation o f the CRE. In the presence of a PKA inhibitor, dibutyryl cAMP-induced activity was completely abolished and basal activity was decreased by up to 85%. Similar protein-DNA complexes were formed in gel retardatio n assays with a CgA-CRE oligonucleotide and nuclear extracts from both neuroendocrine and nonendocrine cells. A predominant complex that was supershifted by addition of a CREB antibody was identical in all cell types. By immunoblot analysis, levels of total CREB protein and phosp horylated (Ser 133) CREB did not differ between neuroendocrine and non endocrine cells. Phosphorylated CREB was increased by forskolin treatm ent, an effect that was blocked by a PEA-inhibitor. Expression of the transcriptional cointegrator, CREB-binding protein (CBP), assessed by both RT-PCR and Western blot analysis, did not differ between neuroend ocrine and nonendocrine cells. In summary, the CRE in the hCgA gene pr oximal promoter is critical for both basal and cAMP-induced expression in neuroendocrine cells via a PKA-mediated pathway. However, the neur oendocrine specificity of hCgA gene transcription mediated by the CRE is not a function of levels of total CREB or phosphorylated CREB or it s cointegrator CBP. Specificity may be achieved by a PEA-responsive CR E-binding protein other than CREB expressed specifically in neuroendoc rine cells, expression of a repressor molecule that binds CREB in none ndocrine cells, or may lie downstream of a CRE-binding protein, e.g. i n the activity or amount of cointegrators other than CBP, which are re quired to couple transactivators to the basal transcriptional machiner y.