Disulfide Mispairing During Proinsulin Folding in the Endoplasmic Reticulum

Abstract

Proinsulin folding within the endoplasmic reticulum (ER) remains incompletely understood, but it is clear that in mutant INS gene-induced diabetes of youth (MIDY), progression of the (three) native disulfide bonds of proinsulin becomes derailed, causing insulin deficiency, β-cell ER stress, and onset of diabetes. Herein, we have undertaken a molecular dissection of proinsulin disulfide bond formation, using bioengineered proinsulins that can form only two (or even only one) of the native proinsulin disulfide bonds. In the absence of preexisting proinsulin disulfide pairing, Cys(B19)-Cys(A20) (a major determinant of ER stress response activation and proinsulin stability) preferentially initiates B-A chain disulfide bond formation, whereas Cys(B7)-Cys(A7) can initiate only under oxidizing conditions beyond that existing within the ER of β-cells. Interestingly, formation of these two "interchain" disulfide bonds demonstrates cooperativity, and together, they are sufficient to confer intracellular transport competence to proinsulin. The three most common proinsulin disulfide mispairings in the ER appear to involve Cys(A11)-Cys(A20), Cys(A7)-Cys(A20), and Cys(B19)-Cys(A11), each disrupting the critical Cys(B19)-Cys(A20) pairing. MIDY mutations inhibit Cys(B19)-Cys(A20) formation, but treatment to force oxidation of this disulfide bond improves folding and results in a small but detectable increase of proinsulin export. These data suggest possible therapeutic avenues to ameliorate ER stress and diabetes.

Figure 1

Expression of proinsulin “lose one bond” mutants in pancreatic β-cells. A: Schematic of the “lose mutants”; all constructs were myc tagged within the C-peptide (shown in red). B–D: Indirect immunofluorescence localization of lose mutants in INS1 cells (anti-myc [in red]) compared with that observed for anti-insulin that detects primarily the endogenous Ins1 and Ins2 gene products (in green)—a merged image is shown at right. The red arrows indicate the subplasmalemmal distribution of insulin secretory granules. Scale bar = 20 µm.

Figure 2

Secretion of proinsulin “lose mutants” from pancreatic β-cells. A: 293T cells were transfected with myc-tagged proinsulin bearing wild-type B and A chains (WT), lose-A6/A11, lose-B7/A7, lose-B19/A20, or single C(A6)S or C(A11)S. At 30 h posttransfection, cell culture media were collected overnight, and cells were lysed. Left: Secretion and cellular content of myc-tagged proinsulin were analyzed by Western blotting using anti-myc (for proinsulin) and anti-tubulin antibodies. Right: Quantitation of secretion as measured by Western blotting (from 8 individual samples in 3 independent experiments. Additional proinsulin secretion results, with the identical conclusion, were obtained as measured by rat insulin radioimmunoassay [not shown]). B and C: 293T cells transfected with myc-tagged proinsulin bearing wild-type B and A chains or lose-A6/A11 were pulse labeled with ³⁵S-labeled amino acids for 30 min and chased as indicated. The media were collected and cells lysed; both samples were immunoprecipitated with anti-insulin and then analyzed by nonreducing (B) and reducing (C) Tris–tricine–urea–SDS-PAGE and autoradiography. In this gel system, lose-A6/A11 migrates more slowly than wild type even under reducing conditions, but this does not confuse identification of disulfide isomers that clearly migrate much more slowly under nonreducing conditions.

Figure 3

Expression of proinsulin “keep one bond” mutants in pancreatic β-cells. A: Schematic of the “keep mutants”; all constructs were myc tagged within the C-peptide (shown in red). B–D: Immunofluorescence localization of keep mutants in INS1 cells (anti-myc [red]) compared with that observed for anti-insulin that detects primarily the endogenous Ins1 and Ins2 gene products (in green)—a merged image is shown at right. Scale bar = 20 μm.

Figure 4

Oxidation and stability of proinsulin keep mutants in pancreatic β-cells. A and B: INS1 cells were transiently transfected with myc-tagged proinsulin bearing wild-type B and A chains (WT), keep-B19/A20, keep-B7/A7, or keep-A6/A11. At 48 h, cells were lysed and analyzed by reducing SDS-PAGE and Western blotting (WB) with anti-myc and anti-tubulin (A), with bands quantified by ImageJ and normalized to the synthesis of newly made proinsulin (Supplementary Fig. 1) (B) (the numbers in white reflect independent replicates); data represent the mean ± SEM (*P < 0.05 relative to wild type). C: INS1 cells were transiently cotransfected with a BiP promoter–firefly luciferase plasmid and plasmid encoding myc-tagged proinsulin C(A7)Y (Akita), lose mutants, or keep mutants at a ratio of 1:10, respectively. At 48 h posttransfection, cells were lysed and firefly luciferase was measured per cellular content of myc-tagged proinsulin. The numbers in white reflect independent replicates; data represent the mean ± SEM (*P < 0.05 relative to Akita proinsulin). D and E: INS1 cells were transiently transfected with keep-B19/A20, keep-B7/A7, or keep-A6/A11. At 48 h, cells were pulse labeled with ³⁵S-Met for 15 min (D) and treated with diamide for 2 min at the concentrations shown (E). The cells were washed with ice-cold PBS containing 20 mmol/L NEM and lysed in the presence of 2 mmol/L NEM. Cell lysates were immunoprecipitated with anti-myc antibodies, and newly synthesized proinsulin was analyzed by nonreducing Tris–tricine–urea–SDS-PAGE and phosphorimaging. The positions of both reduced (red) and oxidized (ox) forms of newly synthesized keep-B19/A20 and keep-B7/A7 are indicated. F: Oxidized and reduced proinsulin expressions from three independent experiments similar to that shown in E were quantified by densitometry.

Figure 5

Cooperativity in proinsulin interchain disulfide pairing. INS1 cells transfected with lose-A6/A11 or keep-B20/A20 (or empty vector) were pulse labeled with ³⁵S-labeled amino acids for 15 min and either lysed without chase (0) or chased for 90 min. After washing of cells in ice-cold PBS containing 20 mmol/L NEM, cells were lysed in the presence of 2 mmol/L NEM. After immunoprecipitation with anti-myc, cell lysates (C) and media (M) were analyzed by nonreducing (upper) or reducing (lower) Tris–tricine–urea–SDS-PAGE and autoradiography.

Figure 6

Potential infidelity of proinsulin disulfide bond formation. A: INS1 cells were transiently transfected with proinsulin mutants bearing only two Cys residues that can potentially form only nonnative disulfide partnerships. At 24 h after transfection, cells were pulse labeled with ³⁵S-amino acids for 10 min, lysed as in Fig. 5, and immunoprecipitated with anti-insulin followed by nonreducing Tris–tricine–urea–SDS-PAGE and phosphorimaging. Lane 1 (separated by black line) is a wild-type proinsulin control analyzed separately. Lane 2 (separated by black line) is from the same gel but was underloaded and was intentionally overexposed to facilitate comparison. Small green arrows denote minor misfolded species formed by Cys(B7), as described in the text. B: 293T were transiently transfected to express keep-B19/A20 or selected novel keep mutants bearing only two Cys residues that can potentially form only nonnative disulfide partnerships. After pulse labeling for 10 min as in A, cells were either untreated (–) or treated with 1 mmol/L diamide (+) for 5 min, lysed, and immunoprecipitated and analyzed as in A. C: INS1 cells were transiently transfected with myc-tagged keep mutants, each bearing one additional interloper cysteine. At 48 h after transfection, cells were labeled with ³⁵S-amino acids for 15 min, lysed as in Fig. 5, immunoprecipitated with anti-myc, and resolved by nonreducing Tris–tricine–urea–SDS-PAGE and phosphorimaging.

Figure 7

Effects of MIDY mutants on proinsulin disulfide bond formation. A: INS1 cells were transfected to express myc-tagged proinsulin wild-type (WT) or missense mutants as indicated. At 42 h posttransfection, the cells were pulse labeled with ³⁵S-amino acids for 15 min, lysed as in Fig. 5, immunoprecipitated with anti-myc, and analyzed by nonreducing Tris–tricine–urea–SDS-PAGE and phosphorimaging. Control, untransfected cells. B: 293T cells were transiently transfected with keep-B19/A20 (–) or that bearing an additional MIDY mutation as indicated. At 48 h, cells were labeled with ³⁵S-amino acids for 15 min, lysed as in Fig. 5, immunoprecipitated with anti-myc, and analyzed by nonreducing SDS-PAGE and phosphorimaging, and the oxidized and reduced proinsulin bands were quantified by scanning densitometry (n = 3) (*P < 0.05 compared with keep-B19/A20). C: 293T cells transfected with lose-A6/A11 or lose-A6/A11 bearing the B22Q MIDY mutation were pulse labeled with ³⁵S-amino acids for 30 min and chased for the times indicated. After washing of cells in ice-cold PBS containing 20 mmol/L NEM, cells were lysed in the presence of 2 mmol/L NEM and immunoprecipitated with anti-insulin. Both cell lysates (C) and media (M) were analyzed by nonreducing (upper) and reducing (lower) Tris–tricine–urea–SDS-PAGE and autoradiography.

Figure 8

Partial rescue of the native proinsulin disulfide bond formation that is perturbed in MIDY. 293T cells transfected with lose-A6/A11 bearing the B22Q MIDY mutation were pulse labeled with ³⁵S-amino acids for 30 min and either untreated (control) or treated with 1 mmol/L diamide for 5 min (diamide) and chased for the times indicated. After washing of cells in ice-cold PBS containing 20 mmol/L NEM, cells were lysed in the presence of 2 mmol/L NEM and immunoprecipitated with anti-insulin. Both cell lysates (C) and media (M) were analyzed by nonreducing (upper) and reducing (lower) Tris–tricine–urea–SDS-PAGE and autoradiography.