Analysis of eight cDNAs and six genes for intermediate filament (IF) proteins in the cephalochordate Branchiostoma reveals differences in the IF multigene families of lower chordates and the vertebrates.

We report the sequences of seven new cytoplasmic intermediate filament (IF) proteins of the cephalochordate Branchiostoma. The eight sequences currently known describe four subfamilies (A, B, C and D). All eight IF proteins show the short-length version of the coil 1b subdomain found in vertebrates and lack the additional 42 residues present in all nuclear lamins and the protostomic IF proteins. Although the lancelet is considered to be the closest relative of the vertebrates, it is difficult to relate its IF subfamilies unambiguously to a particular type I-IV subfamily of vertebrates. C1 and C2 have tail domains with two 64 residue repeats of coiled coil-forming ability, a structural feature unknown for IF proteins from vertebrates or protostomia. The epidermal protein D1 shows only a slightly better identity score with vertebrate type II keratins than with type III proteins, but the D1 gene organization is that of type III proteins. The same holds for A1, A2, B1, B2 and C2 genes, although the latter has an additional and uniquely positioned intron. Antibodies (Ab) raised against recombinant C2 and D1 proteins reveal these proteins in epidermis, some internal epithelia and parts of the spinal cord. The results on exonic sequences, gene organization and expression suggest that Branchiostoma IF proteins may retain a largely archetypal condition, whereas the vertebrates have established the well-known type I-IV IF system.