It is important to appreciate the major features of normal type I collagen to understand the pathogenesis of OI. The type I procollagen molecule is formed from two proα1(I) chains (encoded by COL1A1) and one similar but distinct proα2(I) chain (encoded by COL1A2) (Fig. 1).

Fig1. The structure of type I procollagen. Each collagen chain is made as a procollagen triple helix that is secreted into the extracellular space. The amino- and carboxyl-terminal domains are cleaved extracellularly to form collagen; mature collagen fibrils are then assembled and, in bone, mineralized. Note that type I procollagen is composed of two proα1(I) chains and one proα2(I) chain. (Redrawn from Byers PH: Disorders of collagen biosynthesis and structure. In Scriver CR, Beaudet Al, Sly WS, et al, editors: The metabolic bases of inherited disease, ed 6, New York, 1989, McGraw-Hill, pp 2805–2842.)
Proteins composed of subunits, like collagen, are often subject to variants that prevent subunit association by altering the subunit interfaces. The triple helical (collagen) section is composed of 338 tandemly arranged Gly-X-Y repeats; proline is often in the X position, and hydroxyproline or hydroxylysine is often in the Y position. Glycine, the smallest amino acid, is the only residue compact enough to occupy the axial position of the helix, and consequently, variants that substitute other residues for those glycines are highly disruptive to the helical structure.
Several features of procollagen maturation are of special significance to the pathophysiology of OI. First, the assembly of the individual proα chains into the trimer begins at the carboxy terminus, and triple helix formation progresses toward the amino terminus. Consequently, variants that alter residues in the carboxy-terminal part of the triple helical domain are more disruptive because they interfere earlier with the propagation of the triple helix (Fig. 2). Second, the posttranslational modification (e.g., proline or lysine hydroxylation; hydroxylysyl glycosylation) of procollagen continues on any part of a chain not assembled into the triple helix. Thus, when triple helix assembly is slowed by a change, the unassembled sections of the chains amino-terminal to the defect are modified excessively, which slows their secretion into the extracellular space. Overmodification may also interfere with the formation of collagen fibrils. As a result of all these abnormalities, the number of secreted collagen molecules is reduced, and many of them are abnormal. In bone, the abnormal chains and their reduced number lead to defective mineralization of collagen fibrils.

Fig2. The pathogenesis of the major classes of type I procollagen mutants. (Column 1) The types of procollagen chains available for assembly into a triple helix. Although there are two α1 and two α2 collagen genes/genome, as implied in the left column, twice as many α1 collagen molecules are produced, compared to α2 collagen molecules, as shown in the central column. (Column 2) The effect of type I procollagen stoichiometry on the ratio of normal to defective collagen molecules formed in mutants with proα1 chain versus proα2 chain variants. The small vertical bars on each procollagen chain indicate posttranslational modifications (see text). (Column 3) The effect of variants on the biochemical processing of collagen. OI, Osteogenesis imperfecta; Proα1M, a proα1 chain with a missense variant; Proα2M, a proα2 chain with a missense variant; Proα10, a proα1 chain null allele.