Desmosine Peptides: Amino Acid Sequences and the Role of These Sequences in Crosslink Formation

Abstract

We have been involved in defining the biosynthesis of desmosine and isodesmosine. One aspect of this problem was the identification of the lysines which form desmosine and isodesmosine crosslinks. The preparative Edman degradation of desmosine and isodesmosine containing elastolytic peptides of elastin permitted the isolation of peptides C-terminal to the desmosine crosslinks. This approach was used to study the amino acid sequences C-terminal to the desmosine crosslinks in bovine, porcine and human aortic elastin as well as bovine ligamentum nuchae elastin. This identified the lysines in tropoelastin (salt soluble elastin) which give rise to the desmosine crosslinks. The sequences from bovine aortic elastin were identical with those obtained from bovine ligamentum nuchae elastin but differed from those obtained from the other species. The most striking difference involved the occurrence of phenylalanine in bovine elastin and tyrosine in porcine and human elastin C-terminal to the crosslinks. The sequences of the C-terminal peptides fall into two distinct classes, one starting with hydrophobic residues, the other starting with alanine. The study of lysine sequences of tropoelastin, isolated from a lathyritic calf (BAPN), with the use of Myxobacter AL-l Protease II, suggests that essentially all lysines occur in pairs separated by two or three amino acid residues (e.g., Lys-Ala-Ala-Lys; Lys-Ser-Ala-Lys ; and Lys-Ala-Ala-Ala-Lys) and that these pairs are distributed all along the polypeptide chain. This is consistent with the data of Sandberg, Gray, and their colleagues. It is proposed that two such pairs meet to form desmosine or isodesmosine crosslinks and that the hydrophobic residue prevents the enzymic oxidative deamination of the adjacent lysine ε-amino group and this then contributes the nitrogen to the pyridinium ring of the crosslink.