Diselenide-bond replacement of the external disulfide bond of insulin increases its oligomerization leading to sustained activity

Abstract

Seleno-insulin, a class of artificial insulin analogs, in which one of the three disulfide-bonds (S-S's) of wild-type insulin (Ins) is replaced by a diselenide-bond (Se-Se), is attracting attention for its unique chemical and physiological properties that differ from those of Ins. Previously, we pioneered the development of a [C7U^A,C7U^B] analog of bovine pancreatic insulin (SeIns) as the first example, and demonstrated its high resistance against insulin-degrading enzyme (IDE). In this study, the conditions for the synthesis of SeIns via native chain assembly (NCA) were optimized to attain a maximum yield of 72%, which is comparable to the in vitro folding efficiency for single-chain proinsulin. When the resistance of BPIns to IDE was evaluated in the presence of SeIns, the degradation rate of BPIns became significantly slower than that of BPIns alone. Furthermore, the investigation on the intermolecular association properties of SeIns and BPIns using analytical ultracentrifugation suggested that SeIns readily forms oligomers not only with its own but also with BPIns. The hypoglycemic effect of SeIns on diabetic rats was observed at a dose of 150 μg/300 g rat. The strategy of replacing the solvent-exposed S-S with Se-Se provides new guidance for the design of long-acting insulin formulations.

Fig. 1. Molecular structure of insulins.

Three-dimensional structure of wild-type bovine pancreatic insulin (BPIns; PDB code: 2bn3) (a). Primary amino acid sequences of two seleno-insulins having external (b) or internal (c) Se-Se linkage reported by Arai et al. and Metanis et al. respectively.

Fig. 2. Enzymatic digestion of insulins by human insulin-degrading enzyme (IDE).

a Degradation of BPIns or SeIns catalyzed by IDE (PDB code: 2wby). Monomeric insulin bound into the catalytic chamber of IDE is shown in red. b Time course of insulin degradation observed by HPLC analyses (Supplementary Figs. 3a and 4). Reaction conditions were [inslins]₀ = 5.0 μM and [IDE] = 50 nM (S/E = 100:1) in 0.1 M Tris-HCl at pH 8.0 and 30 °C. Decay rates in the competitive experiments using SeIns-BPIns mixture are shown as total insulin degradation. Data are shown as mean ± SEM (n = 3). c Comparison of apparent first-order rate constant (k_app) for degradation of insulins. Data are shown as mean ± SEM (n = 3). The values were estimated by a single exponential fitting of time course data for insulin degradation with the equation: %insulin = 100(e^kt). Data for S/E = 2.1 μM/21 nM were estimated from previous results in ref. . d HPLC charts obtained from digestion experiment of SeIns-BPIns mixture by IDE at pH 8.0 and 30 °C. Mixture of insulins (5.0 μM [BPIns:SeIns = 1:1]) were incubated with IDE (50 nM) under the same conditions as those in (b).

Fig. 3. Preequilibrium model between oligomers and monomers of insulins during IDE degradation.

a Proposed degradation pathways of BPIns including reversible dissociation of BPIns oligomers with an equilibrium constant (K_d^BPIns) and irreversible degradation of monomeric BPIns with a pseudo first-order rate constant (k_deg^BPIns). b Proposed degradation pathways of SeIns including reversible dissociation of SeIns oligomers with an equilibrium constant (K_d^SeIns) and irreversible degradation of monomeric SeIns with a pseudo first-order rate constant (k_deg^SeIns). c Proposed degradation pathways of BPIns–SeIns mixture including reversible dissociation of BPIns–SeIns oligomers with an equilibrium constant (K_d^mix) and irreversible degradation of monomeric BPIns or SeIns with a pseudo first-order rate constant (k_deg^BPIns or k_deg^SeIns).

Fig. 4. Apparent sedimentation coefficient histograms of insulin monomer and oligomers estimated by analytical ultracentrifugation (AUC) (Supplementary Fig. 5).

Symbols M, D, and H in the panels represent monomer, dimer, and hexamer, respectively. BPIns (a), SeIns (b), or their mixture (c, 1:1) were dissolved in 0.1 M Tris-HCl buffer solution at pH 8.0 and a variety of concentrations (5, 10, and 20 μM). The sums of the histograms obtained from the single-component analyses, i.e., BPIns (5 μM in a) + SeIns (5 μM in b) [bottom] and BPIns (10 μM in a) + SeIns (10 μM in b) [top], are shown in (d).

Fig. 5. In vivo study on the hypoglycemic effect of insulins using normal and diabetic model rats.

a Time course of blood glucose level in normal rat following subcutaneous (s.c.) injection of insulin samples (10 units/mL of saline). The dosage of insulins was 15 μg/300 g rat. The experiments with saline and insulins were repeated four and two times, respectively, and good repeatability was confirmed. Data are shown as the average. b Time course of blood glucose level in STZ-induced diabetic model rat following s.c. injection of insulin samples (10 units/mL of saline). The dosage of insulins was 15 μg/300 g rat. Data are shown as mean ± SEM (n = 3 [insulins] or 5 [saline]). c Time course of blood glucose level in normal rat following s.c. injection of insulin samples (100 units/mL of saline). The dosage of insulins was 150 μg/300 g rat. Data are shown as mean ± SEM (n = 5 [insulins] or 4 [saline]). d Time course of blood glucose level in STZ-induced diabetic model rat following s.c. injection of insulin samples (100 units/mL of saline). The dosage of insulins was 150 μg/300 g rat. Data are shown as mean ± SEM (n = 4 [HIns] or 5 [SeIns and saline]). The blood glucose level was monitored by an i-STAT 1 analyzer with cartridge 6+. The symbols * and ** represent p < 0.05 and p < 0.01, respectively. The p-value was obtained from a t-test.