It was recently shown that the two transcripts encoding the isoforms of 2',3'-cyclic nucleotide 3'-phosphodiesterase (CNP1 and CNP2) are differentially regulated during the process of oligodendrocyte maturation. In oligodendrocyte precursors, only CNP2 mRNA is present, whereas in differentiating oligodendrocytes, both CNP1 and CNP2 mRNAs are expressed. This pattern of CNP expression is likely due to stage-specific transcriptional regulation of the two CNP promoters during the process of oligodendrocyte differentiation. Here, we report the influence of increased intracellular cyclic AMP (cAMP) levels on the transcription of both CNP1 and CNP2 mRNAs in rat C6 glioma cells. We found that the transcription of CNP1 mRNA was significantly increased in comparison with that of CNP2 mRNA in cells treated with cAMP analogues to elevate intracellular cAMP levels. This up-regulation of CNP1 expression (a) is due to an increase of transcription, (b) requires de novo protein synthesis, and (c) requires the activity of protein kinase A. These results are physiologically significant and support the idea that a cAMP-mediated pathway is part of the molecular mechanisms regulating the expression of CNP1 in oligodendrocytes. The regulation of CNP1 promoter activity by cAMP was then investigated in stably transfected C6 cell lines containing various deletions of the CNP promoter directing the bacterial chloramphenicol acetyltransferase gene. We showed that the sequence between nucleotides -126 and -102 was essential for the cAMP-dependent induction of CNP1 expression. Gel retardation analysis showed that two protein-DNA complexes are formed between this sequence and nuclear factors from C6 cells treated or not treated with cAMP. This suggests that the induction of CNP1 mRNA transcription is not mediated by changes in binding of nuclear factors that interact directly with the -126/-102 sequence. Sequence analysis of this region revealed the presence of a putative activator protein-2 (AP-2) binding site. It is interesting that mutagenesis of this region resulted in a significant reduction in transcriptional responses to cAMP, implying a possible role for the AP-2 factor in the expression of CNP1. In addition, we have shown that putative binding sites for activator protein-4 and nuclear factor-1 adjacent to the AP-2 site are required for efficient induction of CNP1 expression by cAMP. Taken together, our results show that the cAMP-dependent accumulation of CNP1 mRNA appears to depend on the synergistic interaction of several regulatory elements.