Sequence analysis of murine cytokeratin endo A (no. 8) cDNA. Evidence for mRNA species initiated upstream of the normal 5' end in PCC4 cells.

Keratin 8 (Endo A) is expressed in simple epithelia, together with keratin 18 (Endo B). Filaments formed by this keratin pair are the first cytoskeletal elements induced during mouse embryogenesis. We have isolated Endo A cDNA clones from lambda gt11 libraries prepared with mRNA isolated from PCC4 embryonal carcinoma (EC) cells. Sequencing of three overlapping cDNAs and of a genomic clone allowed us to determine the complete sequence of the Endo A message. Analysis of the protein sequence deduced showed that the Endo A protein presented all the characteristics of intermediate filaments, including an alpha-helical central rod domain and nonhelical N- and C-termini. In the rod domain, the degree of similarity to the other members of the basic keratin family was high. A high degree of homology to keratin 8 of other species was observed, even in the non-helical domains. During these analyses, we found clones extending upstream of the normal 5' end of the mRNA. Sequence comparison between these cDNAs and the 5' upstream region of the Endo A gene suggested that they corresponded to transcripts initiated at an upstream alternative promoter. These observations supported previous results showing the presence of Endo A transcripts initiated upstream of the normal 5' end in mouse morulae and blastocysts.