Abstract
PURPOSE: Heterozygous mutations in the insulin gene can give rise to a monogenic diabetes syndrome due to toxic misfolding of the variant proinsulin in the endoplasmic reticulum (ER) of pancreatic β-cells. Clinical mutations are widely distributed in the sequence (86 amino acids). Misfolding induces chronic ER stress and interferes in trans with wildtype biosynthesis and secretion. In the present work we sought to study relative folding efficiencies of proinsulin variants in relation to age of disease onset. METHODS: To enable efficient preparation of non-foldable variants, we developed a four-segment native chemical-ligation scheme that exploits two native cysteines (Cys(B19) and Cys(A6); residues 19 and 71 in proinsulin) and an alanine in the connecting domain (Ala(C20); residue 50). From N- to C terminus, the four segments have respective lengths 18, 31, 22 and 15 residues-convenient to "mix and match" native and variant synthetic segments as a platform technology. RESULTS: Folding of the reduced and unfolded polypeptides was investigated under three conditions: pH 10.6 (which promotes disulfide pairing as in the pharmaceutical manufacture of insulin) and pH 7.4 in the absence or presence of "foldase" protein disulfide isomerase. Whereas wild-type proinsulin efficiently folds to form a single dominant product (in accordance with classical studies), the clinical variants exhibited marked impairment, especially at neutral pH. CONCLUSION: Among representative clinical variants, relative folding yields correlated with both degree of ER stress in cell culture and ages of clinical diabetes onset (neonatal, adolescence or adulthood). Implications for the native mechanism of nascent protein folding are discussed.