Transcripts and CpG islands associated with the pro-opiomelanocortin gene and other neurally expressed genes.

DNA sequences of vertebrate genes which code for neural or neuroendocrine peptides were analysed in terms of CpG dinucleotide distribution and G+C content. The vast majority of the genes were found to contain a region with the sequence characteristics of a CpG island surrounding the 5' end. In mammalian species, the gene which codes for the neuroendocrine polypeptide pro-opiomelanocortin (POMC) was shown to be associated with two separate CpG islands: a 5' CpG island which surrounds the POMC transcription start site and a 3' CpG island which lies approximately 5 kb downstream, encompassing the third exon of POMC. Short POMC-related transcripts, known to be transcribed in the germline, were found to initiate from a promoter within the 3' CpG island. The start sites of the short POMC-related transcripts in mouse testis were mapped to the region coding for gamma MSH in exon 3, in a similar location to transcription start sites identified in other mammalian POMC genes. Similar short POMC-related transcripts were identified in both the mouse F9 embryonal carcinoma cell line and mouse embryonic stem cells, suggesting that transcription initiating within the third exon may occur very early in development. No short transcripts were detected by Northern blot hybridization in either Xenopus laevis testis or oocyte poly(A)+ RNA extracts. The Xenopus laevis POMC genes, A and B, were associated with neither a 5' nor a 3' CpG island. Hence, the presence of a 5' CpG island is not required for production of full-length transcripts from the Xenopus laevis POMC gene, but the presence of a 3' CpG island may be required for transcription to occur from the third exon.