Using a synthetic switch to regulate insulin receptor activation

One hundred years ago in 1921 Sir Frederick G. Banting and Charles H. Best at the University of Toronto discovered insulin; they successfully isolated an extract from the islets of Langerhans of the pancreas and demonstrated that its injection into a diabetic dog lowered its blood sugar by 40% in 1 h (1). Insulin was subsequently found to regulate the metabolism of carbohydrates, fats, and protein by promoting the absorption of blood glucose into cells of the liver, fat, and skeletal muscle. In 1934 Dorothy Crowfoot, later Hodgkin, obtained crystals of insulin from the Beecham Group plc, a British pharmaceutical company, which was producing the material in crystalline form for treatment of diabetics. Although Dorothy managed to get diffraction patterns of insulin in the mid-1930s (2), the structure determination proved a challenge. Two decades later Fred Sanger defined the sequence of insulin (3), which was an important step required not only for interpreting the electron density defined by X-ray crystallography but also the functional interactions of insulin. The crystal structure of the 2-zinc insulin hexamer was eventually solved in 1969 (4) and its relation to insulin evolution and biological functions discussed soon after (5) by the Hodgkin group in Oxford University. These developments set the stage for analyses of insulin sequence and structure variants over the past five decades. One of the most recent advances, indeed celebrating a century of insulin research at the molecular level, is an intriguing paper in PNAS entitled “Insertion of a synthetic switch into insulin provides metabolite-dependent regulation of hormone–receptor activation” by Michael Weiss and coworkers (6). This paper tests some of the accepted dogma and new ideas about insulin structure–function relationships with the design of a glucose-responsive insulin to protect patients from hypoglycemia.

As the authors comment (6), natural selection has favored protein sequences that are efficiently foldable with sufficient native-state stability to avoid toxic misfolding and amyloidogenesis. In the case of insulin, the folded insulin protomer structure assembles as a 2-zinc-hexameric form with three-fold symmetry relating insulin dimers with two-fold symmetry perpendicular to this, giving rise to 32 symmetry with the two zinc atoms on the three-fold axis (5). This form has evolved as microcrystalline storage granules that protect against aggregation-coupled misfolding and proteolytic degradation in the pancreatic β cells. When released into the circulation the granules take time to dissolve to give first the 2-zinc insulin hexamers. This process increases the lifetime of the active insulin in circulation but eventually allows dissociation into monomers that are the active form (5). These insulin monomers can then adopt the open state that involves the C-terminal region of the B chain—B24 to the C terminus at B30—moving away from the hydrophobic core of the insulin molecule, known as “hinge opening.” This is facilitated by the B23 Gly, which introduces flexibility and allows the B24 Phe, B25 Phe, and B26 Tyr to contribute to receptor activation, involving receptor conformational change to facilitate transmembrane signaling. In this paper Weiss and coworkers seek to test the coupling of the “hinge opening” of the insulin molecule and the activation of the receptor (6).

The design scheme uses a “glucose-cleavable” tether between the C terminus of the B chain and the N terminus of the A chain, shown schematically in Fig. 1, involving a diol binding interaction of phenylboronic acid to a 1,2 diol element of a monosaccharide (6). This leads to an inactive closed form at low glucose concentrations that open ups on increasing concentrations, thus acting as a sensor. The authors have tested this by monitoring receptor autophosphorylation and downstream signaling. They have demonstrated coupling between the hinge opening and receptor signaling. The authors also exploited the resolution revolution in cryo-electron microscopy (7) to define structures of complexes between insulin and the ectodomain of the insulin receptor, confirming a metabolite-dependent “opening” of the closed inactive insulin conformation. They suggest the use of these tools to design “smart” glucose-responsive analogs to treat insulin-dependent diabetes mellitus, a very exciting approach to the design of new pharmaceutical agents.

Fig. 1.

A schematic representation of the “glucose-cleavable” tether between the C terminus of the B chain and the N terminus of the A chain. The tether involves a diol binding interaction of phenylboronic acid to a 1,2 diol element of a monosaccharide.