Among the various intermediate filament (IF) proteins, cytokeratin 8 (CK8) is especially remarkable as it is produced early in embryogenesis, is the only type-II CK occurring in many simple epithelial cells, and can also be synthesized in certain non-epithelial cells. Using a cDNA probe specific for human CK8 we have isolated an approx. 14-kb genomic clone (in phage lambda EMBL3) which contains the gene encoding human CK8. The gene comprising a total of 7766 nucleotides (nt) from the transcription start point, determined by primer extension analysis, to the polyadenylation site, determined from cDNA sequencing, and a 1030-nt 5' upstream region have been sequenced. The sequence of 485 amino acids (aa) deduced from the exon sequences (Mr 53,532) shows strong homology with the corresponding gene products of bovine, murine and amphibian (Xenopus laevis) origins. Surprisingly, the ck8 gene contains only seven introns instead of eight as found in all other genes of the same (type-II) CK subfamily; while all seven introns occur in identical positions as in the other type-II CK-encoding genes, intron V of these genes is missing in the ck8 gene. Intron I of ck8 is remarkably long (2534 nt) and contains a cluster of potential regulatory sequences, including three Sp1 sites, and an extended Alu-element. In Southern-blot analyses, we found only one intron-containing gene encoding CK8 in the human genome, and by heterologous transfection experiments we showed that this gene is correctly transcribed in non-human cells expressing the orthologous gene.(ABSTRACT TRUNCATED AT 250 WORDS)