The nucleotide sequence of the vitellogenin cDNA from the rainbow trout Oncorhynchus mykiss was determined. Analysis of the deduced amino acid sequence (1659 residues) places the lipovitellin I, phosvitin and lipovitellin II domains between amino acids 16 to 1088, 1089 to 1145 and 1146 to 1659, respectively. The general structure is similar to other vertebrate vitellogenins except for the serine rich phosvitin domain which is the shortest identified so far in vertebrates (57 amino acids), being 2 to 4 times smaller than in other species. Sequence comparisons between vertebrate and invertebrate vitellogenins as well as with distantly related proteins allowed to identify two short amino acid motifs particularly well conserved, RGILN and TCGLCG in lipovitellin I and II domains, respectively, and strongly suggest that the lipovitellin II domain is involved in protein interactions via disulfide bridge formation.