We have isolated and characterized four collagen-related c-DNA clones (N-COL 1, N-COL 2, N-COL 3, N-COL 4) that are highly expressed in developing nematocytes in hydra. All four c-DNAs as well as their corresponding transcripts are small in size (600-1,000 bp). The deduced amino acid sequences show that they contain a central region consisting of 14 to 16 Gly-X-Y triplets. This region is flanked amino-terminal by a stretch of 14-23 proline residues and carboxy-terminal by a stretch of 6-9 prolines. At the NH2- and COOH-termini are repeated patterns of cysteine residues that are highly conserved between the molecules. A model is proposed which consists of a central stable collagen triple helix of 12-14 nm length from which three 9-22 nm long polyproline II type helices emerge at both ends. Disulfide linkage between cysteine-rich segments in these helices could lead to the formation of oligomeric network structures. Electrophoretic characterization of nematocyst extracts allows resolution of small proline-rich polypeptides that correspond in size to the cloned sequences.