The assembly of the type I procollagen heterotrimer is initiated by an interaction between the carboxyl propeptides, with triple helix folding proceeding in the C→N direction. The pro-α1-(I)-C-propeptides can interact with self to form the homotrimer or with pro-α2(I)-C-propeptide to establish the heterotrimer. The two propeptides are similar in length and have about 65% identity in sequence. Nevertheless, we proposed that differences in interaction between propeptides might account for thein vivo selection of heterotrimer formation rather than homotrimer formation. To test this hypothesis we have determined the probable structures of the human C-propeptides by molecular modeling and energy minimization using Molecular Simulations Insight, Discover 95.0/3.0, and Biopolymer programs. The propeptide structures were constrained with the two known intrachain disulfide bonds in each case. The two structures were globally similar, with three distinct structural domains (G-I, L, G-II) in each case. A few crucial Pro residues and other sequence differences, however, produced different structures in each domain. The different interaction profiles of the three domains may be of crucial importance for heterotrimer selection.

The amino-telopeptides of type I collagen have been implicated to have a crucial role in the events of fibril formation. The sequences have been highly conserved in a variety of species. Several proposals have been presented to correlate the sequence with structure and role in fibril formation, but no definite information on telopeptide structure has been deduced. The infrared spectrum of the amide I band, due almost entirely to the C=O stretch vibration of peptide carbonyls, has been a most useful probe for determining the secondary structures of proteins in solution. In the present study, the secondary structure of a synthetic ratα1(I) amino-telopeptide has been investigated in aqueous solution by Fourier transform infrared spectroscopy (FTIRS) using a 9-pass internal reflectance ZnSe prism cell. Conformational changes were monitored as the aqueous solution was heated from 4 to 50°C by observing changes in the frequency position. The amide I band frequency shifted by about 10 cm⁻¹ when the aqueous telopeptide solution was heated from 4–50°C. Deconvolution of the amide I band showed that the major component could be best represented as in a random configuration at 4°C but changed to aβ sheet withβ turns around 30°C. To support these experimental data the telopeptide region was modeled using BIOSYM/MSI software on a Silicon Graphics R-4000,X/Z graphics workstation. The proposed telopeptide structure was energy minimized using DISCOVER CVFF repetitive build and minimize process to reduce steric hindrance and maximize H-bonding. The potential energy surface was quite low and the conformation was stabilized by only 3 H-bonds. This model suggests a telopeptide structure that can be induced to assume a conformation favorable for binding during its interchain interaction at the collagen helix cross-link (N-telopeptide) receptor domain around collagen residue 930.