A spectrum of cysteine mutations in fibulin-3 highlight its disulfide bonding complexity and potential to induce ER stress

Fibulin-3 (FBLN3), also known as EFEMP1, is a secreted extracellular matrix (ECM) glycoprotein that contains forty cysteine residues. These cysteines, which are distributed across one atypical and five canonical calcium-binding epidermal growth factor (EGF) domains, are important for regulating FBLN3 structure, secretion, and presumably function. As evidence of this importance, a rare homozygous p.C55R mutation in FBLN3 negates its function, alters disulfide bonding, and causes marfanoid syndrome. Additional studies suggest that heterozygous premature stop codon mutations in FBLN3 may also cause similar, albeit less severe, connective tissue disorders. Interestingly, a series of twenty-four cysteine mutations in FBLN3 have been identified in the human population and published in the Clinical Variation (ClinVar) and gnomAD databases. We tested how seven of these cysteine mutants (five loss-of-cysteine variants: C42Y, C190R, C218R, C252F, and C365S, two gain-of-cysteine variants: R358C, Y369C) and two newly-developed mutations (G57C and Y397C) altered FBLN3 secretion and disulfide bonding. We selected three variants to test alterations in MMP2 zymography and stress response activation. Surprisingly, we found a wide variety of biochemical behaviors: i) loss-of-cysteine variants correlated with an increased likelihood of disulfide dimer formation, ii) N-terminal mutations were less likely to disrupt secretion, and were less prone to aggregation, iii) in contrast to wild-type FBLN3, multiple, but not all variants failed to induce MMP2 levels in cell culture, and iv) C-terminal mutations (either loss or gain of cysteines) were more prone to significant secretion defects, intracellular accumulation/misfolding, and stress response activation. These results provide molecular and biochemical insight into FBLN3 folding, secretion, and function for many cysteine mutations found in the human population, some of which may increase the likelihood of subclinical connective tissue or other FBLN3-associated haploinsufficiency diseases.