Search

Introduction: Insulin icodec is a novel, long-acting insulin analog designed to cover basal insulin requirements with once-weekly subcutaneous administration. Here we describe the molecular engineering and the biological and pharmacological properties of insulin icodec., Research Design and Methods: A number of in vitro assays measuring receptor binding, intracellular signaling as well as cellular metabolic and mitogenic responses were used to characterize the biological properties of insulin icodec. To evaluate the pharmacological properties of insulin icodec in individuals with type 2 diabetes, a randomized, double-blind, double-dummy, active-controlled, multiple-dose, dose escalation trial was conducted., Results: The long half-life of insulin icodec was achieved by introducing modifications to the insulin molecule aiming to obtain a safe, albumin-bound circulating depot of insulin icodec, providing protracted insulin action and clearance. Addition of a C20 fatty diacid-containing side chain imparts strong, reversible albumin binding, while three amino acid substitutions (A14E, B16H and B25H) provide molecular stability and contribute to attenuating insulin receptor (IR) binding and clearance, further prolonging the half-life. In vitro cell-based studies showed that insulin icodec activates the same dose-dependent IR-mediated signaling and metabolic responses as native human insulin (HI). The affinity of insulin icodec for the insulin-like growth factor-1 receptor was proportionately lower than its binding to the IR, and the in vitro mitogenic effect of insulin icodec in various human cells was low relative to HI. The clinical pharmacology trial in people with type 2 diabetes showed that insulin icodec was well tolerated and has pharmacokinetic/pharmacodynamic properties that are suited for once-weekly dosing, with a mean half-life of 196 hours and close to even distribution of glucose-lowering effect over the entire dosing interval of 1 week., Conclusions: The molecular modifications introduced into insulin icodec provide a novel basal insulin with biological and pharmacokinetic/pharmacodynamic properties suitable for once-weekly dosing., Trial Registration Number: NCT02964104., Competing Interests: Competing interests: All authors are current or past employees of Novo Nordisk A/S, Denmark., (© Author(s) (or their employer(s)) 2021. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.)

The life-saving discovery of insulin in Toronto in 1921 is one of the most impactful achievements in medical history, at the time being hailed as a miracle treatment for diabetes. The insulin molecule itself, however, is poorly amenable as a pharmacological intervention, and the formidable challenge of optimizing insulin therapy has been ongoing for a century. We review early academic insights into insulin structure and its relation to self-association and receptor binding, as well as recombinant biotechnology, which have all been seminal for drug design. Recent developments have focused on combining genetic and chemical engineering with pharmaceutical optimization to generate ultra-rapid and ultra-long-acting, tissue-selective, or orally delivered insulin analogs. We further discuss these developments and propose that future scientific efforts in molecular engineering include realizing the dream of glucose-responsive insulin delivery., Competing Interests: Declaration of interests P.K., E.N., H.H., T.H-J., E.J., J.S., and T.K. are employees and shareholders of Novo Nordisk. P.M. was an employee and shareholder of Novo Nordisk at the time of preparation of this manuscript., (Copyright © 2021 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

Recently, the clinical proof of concept for the first ultra-long oral insulin was reported, showing efficacy and safety similar to subcutaneously administered insulin glargine. Here, we report the molecular engineering as well as biological and pharmacological properties of these insulin analogues. Molecules were designed to have ultra-long pharmacokinetic profile to minimize variability in plasma exposure. Elimination plasma half-life of ~20 h in dogs and ~70 h in man is achieved by a strong albumin binding, and by lowering the insulin receptor affinity 500-fold to slow down receptor mediated clearance. These insulin analogues still stimulate efficient glucose disposal in rats, pigs and dogs during constant intravenous infusion and euglycemic clamp conditions. The albumin binding facilitates initial high plasma exposure with a concomitant delay in distribution to peripheral tissues. This slow appearance in the periphery mediates an early transient hepato-centric insulin action and blunts hypoglycaemia in dogs in response to overdosing.

Peripheral hyperinsulinemia resulting from subcutaneous insulin injection is associated with metabolic defects which include abnormal glucose metabolism. The first aim of this study was to quantify the impairments in liver and muscle glucose metabolism that occur when insulin is delivered via a peripheral vein compared to when it is given through its endogenous secretory route (the hepatic portal vein) in overnight fasted conscious dogs. The second aim was to determine if peripheral delivery of a hepato-preferential insulin analog could restore the physiologic response to insulin that occurs under meal feeding conditions. This study is the first to show that hepatic glucose uptake correlates with insulin's direct effects on the liver under hyperinsulinemic-hyperglycemic conditions. In addition, glucose uptake was equally divided between the liver and muscle when insulin was infused into the portal vein, but when it was delivered into a peripheral vein the percentage of glucose taken up by muscle was 4-times greater than that going to the liver, with liver glucose uptake being less than half of normal. These defects could not be corrected by adjusting the dose of peripheral insulin. On the other hand, hepatic and non-hepatic glucose metabolism could be fully normalized by a hepato-preferential insulin analog.

The structure of insulin, a glucose homeostasis-controlling hormone, is highly conserved in all vertebrates and stabilized by three disulfide bonds. Recently, we designed a novel insulin analogue containing a fourth disulfide bond located between positions A10-B4. The N-terminus of insulin's B-chain is flexible and can adapt multiple conformations. We examined how well disulfide bond predictions algorithms could identify disulfide bonds in this region of insulin. In order to identify stable insulin analogues with additional disulfide bonds, which could be expressed, the Cβ cut-off distance had to be increased in many instances and single X-ray structures as well as structures from MD simulations had to be used. The analogues that were identified by the algorithm without extensive adjustments of the prediction parameters were more thermally stable as assessed by DSC and CD and expressed in higher yields in comparison to analogues with additional disulfide bonds that were more difficult to predict. In contrast, addition of the fourth disulfide bond rendered all analogues resistant to fibrillation under stress conditions and all stable analogues bound to the insulin receptor with picomolar affinities. Thus activity and fibrillation propensity did not correlate with the results from the prediction algorithm. Statement: A fourth disulfide bond has recently been introduced into insulin, a small two-chain protein containing three native disulfide bonds. Here we show that a prediction algorithm predicts four additional four disulfide insulin analogues which could be expressed. Although the location of the additional disulfide bonds is only slightly shifted, this shift impacts both stability and activity of the resulting insulin analogues., (© 2015 The Protein Society.)

Background: Insulin analogues comprising acidic amino acid substitutions at position B10 have previously been shown to display increased mitogenic potencies compared to human insulin and the underlying molecular mechanisms have been subject to much scrutiny and debate. However, B10 is still an attractive position for amino acid substitutions given its important role in hexamer formation. The aim of this study was to investigate the relationships between the receptor binding properties as well as the metabolic and mitogenic potencies of a series of insulin analogues with different amino acid substitutions at position B10 and to identify a B10-substituted insulin analogue without an increased mitogenic to metabolic potency ratio., Methodology/principal Findings: A panel of ten singly-substituted B10 insulin analogues with different amino acid side chain characteristics were prepared and insulin receptor (both isoforms) and IGF-I receptor binding affinities using purified receptors, insulin receptor dissociation rates using BHK cells over-expressing the human insulin receptor, metabolic potencies by lipogenesis in isolated rat adipocytes, and mitogenic potencies using two different cell types predominantly expressing either the insulin or the IGF-I receptor were systematically investigated. Only analogues B10D and B10E with significantly increased insulin and IGF-I receptor affinities as well as decreased insulin receptor dissociation rates displayed enhanced mitogenic potencies in both cell types employed. For the remaining analogues with less pronounced changes in receptor affinities and insulin receptor dissociation rates, no apparent correlation between insulin receptor occupancy time and mitogenicity was observed., Conclusions/significance: Several B10-substituted insulin analogues devoid of disproportionate increases in mitogenic compared to metabolic potencies were identified. In the present study, receptor binding affinity rather than insulin receptor off-rate appears to be the major determinant of both metabolic and mitogenic potency. Our results also suggest that the increased mitogenic potency is attributable to both insulin and IGF-I receptor activation.

An ingenious system evolved to facilitate insulin binding to the insulin receptor as a monomer and at the same time ensure sufficient stability of insulin during storage. Insulin dimer is the cornerstone of this system. Insulin dimer is relatively weak, which ensures dissociation into monomers in the circulation, and it is stabilized by hexamer formation in the presence of zinc ions during storage in the pancreatic β-cell. Due to the transient nature of insulin dimer, direct investigation of this important form is inherently difficult. To address the relationship between insulin oligomerization and insulin stability and function, we engineered a covalently linked insulin dimer in which two monomers were linked by a disulfide bond. The structure of this covalent dimer was identical to the self-association dimer of human insulin. Importantly, this covalent dimer was capable of further oligomerization to form the structural equivalent of the classical hexamer. The covalently linked dimer neither bound to the insulin receptor, nor induced a metabolic response in vitro. However, it was extremely thermodynamically stable and did not form amyloid fibrils when subjected to mechanical stress, underlining the importance of oligomerization for insulin stability.

The relative expression patterns of the two IR (insulin receptor) isoforms, +/- exon 11 (IR-B/IR-A respectively), are tissue-dependent. Therefore we have developed insulin analogues with different binding affinities for the two isoforms to test whether tissue-preferential biological effects can be attained. In rats and mice, IR-B is the most prominent isoform in the liver (> 95%) and fat (> 90%), whereas in muscles IR-A is the dominant isoform (> 95%). As a consequence, the insulin analogue INS-A, which has a higher relative affinity for human IR-A, had a higher relative potency [compared with HI (human insulin)] for glycogen synthesis in rat muscle strips (26%) than for glycogen accumulation in rat hepatocytes (5%) and for lipogenesis in rat adipocytes (4%). In contrast, the INS-B analogue, which has an increased affinity for human IR-B, had higher relative potencies (compared with HI) for inducing glycogen accumulation (75%) and lipogenesis (130%) than for affecting muscle (45%). For the same blood-glucose-lowering effect upon acute intravenous dosing of mice, INS-B gave a significantly higher degree of IR phosphorylation in liver than HI. These in vitro and in vivo results indicate that insulin analogues with IR-isoform-preferential binding affinity are able to elicit tissue-selective biological responses, depending on IR-A/IR-B expression.

Background: The insulin receptor (IR) exists in two isoforms, A and B, and the isoform expression pattern is tissue-specific. The C-terminus of the insulin B chain is important for receptor binding and has been shown to contact the IR just adjacent to the region where the A and B isoforms differ. The aim of this study was to investigate the importance of the C-terminus of the B chain in IR isoform binding in order to explore the possibility of engineering tissue-specific/liver-specific insulin analogues., Methodology/principal Findings: Insulin analogue libraries were constructed by total amino acid scanning mutagenesis. The relative binding affinities for the A and B isoform of the IR were determined by competition assays using scintillation proximity assay technology. Structural information was obtained by X-ray crystallography. Introduction of B25A or B25N mutations resulted in analogues with a 2-fold preference for the B compared to the A isoform, whereas the opposite was observed with a B25Y substitution. An acidic amino acid residue at position B27 caused an additional 2-fold selective increase in affinity for the receptor B isoform for analogues bearing a B25N mutation. Furthermore, the combination of B25H with either B27D or B27E also resulted in B isoform-preferential analogues (2-fold preference) even though the corresponding single mutation analogues displayed no differences in relative isoform binding affinity., Conclusions/significance: We have discovered a new class of IR isoform-selective insulin analogues with 2-4-fold differences in relative binding affinities for either the A or the B isoform of the IR compared to human insulin. Our results demonstrate that a mutation at position B25 alone or in combination with a mutation at position B27 in the insulin molecule confers IR isoform selectivity. Isoform-preferential analogues may provide new opportunities for developing insulin analogues with improved clinical benefits.

The insulin from the Atlantic hagfish (Myxine glutinosa) has been one of the most studied insulins from both a structural and a biological viewpoint; however, some aspects of its biology remain controversial, and there has been no satisfying structural explanation for its low biological potency. We have re-examined the receptor binding kinetics, as well as the metabolic and mitogenic properties, of this phylogenetically ancient insulin, as well as that from another extant representative of the ancient chordates, the river lamprey (Lampetra fluviatilis). Both insulins share unusual binding kinetics and biological properties with insulin analogues that have single mutations at residues that contribute to the hexamerization surface. We propose and demonstrate by reciprocal amino acid substitutions between hagfish and human insulins that the reduced biological activity of hagfish insulin results from unfavorable substitutions, namely, A10 (Ile to Arg), B4 (Glu to Gly), B13 (Glu to Asn), and B21 (Glu to Val). We likewise suggest that the altered biological activity of lamprey insulin may reflect substitutions at A10 (Ile to Lys), B4 (Glu to Thr), and B17 (Leu to Val). The substitution of Asp at residue B10 in hagfish insulin and of His at residue A8 in both hagfish and lamprey insulins may help compensate for unfavorable changes in other regions of the molecules. The data support the concept that the set of unusual properties of insulins bearing certain mutations in the hexamerization surface may reflect those of the insulins evolutionarily closer to the ancestral insulin gene product.

Conjointly, the solvent-exposed residues of the central alpha-helix of the B chain form a well-defined ridge, which is flanked and partly overlapped by the two described insulin receptor binding surfaces on either side of the insulin molecule. To evaluate the importance of this interface in insulin receptor binding, we developed a new powerful method that allows us to introduce all the naturally occurring amino acids into a given position and subsequently determine the receptor binding affinities of the resulting insulin analogues. The total amino acid scanning mutagenesis was performed at positions B9, B10, B12, B13, B16, and B17, and the vast majority of the insulin analogue precursors were expressed and secreted in amounts close to that of the wild-type (human insulin) precursor. The analogue binding data revealed that positions B12 and B16 were the two positions most affected by the amino acid substitutions. Interestingly, the receptor binding affinities of the B13 analogues were also markedly affected by the amino acid substitutions, suggesting that GluB13 indeed is a part of insulin's binding surface. The B10 library screen generated analogues covering a wide range of (20-340%) of relative binding affinities, and the results indicated that a structural stabilization of the central alpha-helix and thereby a more rigid presentation of the binding epitope at the insulin receptor is important for receptor recognition. In conclusion, systematic amino acid scanning mutagenesis allowed us to confirm the importance of the B chain alpha-helix as a central recognition element serving as a linker of a continual binding surface.

The folding stability of insulin is positively correlated with the expression yield of the precursor expressed in yeast. Insulin assembles into dimers and hexamers in a concentration-dependent manner and amino acid substitutions that impair the ability of insulin to associate into dimers concomitantly decrease the expression yield (excluding substitutions that enhance folding stability). In contrast, introduction of an amino substitution that enhances the self-association of insulin improved the yeast expression yield. In the monomeric state the majority of the non-polar residues of insulin are exposed to the solvent and assembly into dimers and hexamers shields these from contact with the solvent. It is proposed that self-association enhances the flux of insulin through the secretory pathway by increasing the hydrophilicity, decreasing the surface area as well as decreasing the molar concentration in the secretory pathway., (Copyright 2002 Elsevier Science (USA))

Adaptation to efficient heterologous expression is a prerequisite for recombinant proteins to fulfill their clinical and biotechnological potential. We describe a rational strategy to optimize the secretion efficiency in yeast of an insulin precursor by structure-based engineering of the folding stability. The yield of a fast-acting insulin analogue (Asp(B28)) expressed in yeast was enhanced 5-fold by engineering a specific interaction between an aromatic amino acid in the connecting peptide and a phenol binding site in the hydrophobic core of the molecule. This insulin precursor is characterized by significantly enhanced folding stability. The improved folding properties enhanced the secretion efficiency of the insulin precursor from 10 to 50%. The precursor remains fully in vitro convertible to mature fast-acting insulin.

An insulin-containing fusion protein (ICFP, encoding the yeast prepro-alpha factor leader peptide fused via a lysine-arginine cleavage site to a single chain insulin) has been expressed in Saccharomyces cerevisiae where it is inefficiently secreted. Single gene disruptions have been identified that cause enhanced immunoreactive insulin secretion (eis). Five out of six eis mutants prove to be vacuolar protein sorting (vps)8, vps35, vps13, vps4, and vps36, which affect Golgi<-->endosome trafficking. Indeed, in wild-type yeast insulin is ultimately delivered to the vacuole, whereas vps mutants secrete primarily unprocessed ICFP. Disruption of KEX2, which blocks intracellular processing to insulin, quantitatively reroutes ICFP to the cell surface, whereas loss of the Vps10p sorting receptor is without effect. Secretion of unprocessed ICFP is not based on a dominant secretion signal in the alpha-leader peptide. Although insulin sorting mediated by Kex2p is saturable, Kex2p functions not as a sorting receptor but as a protease: replacement of Kex2p by truncated secretory Kex2p (which travels from Golgi to cell surface) still causes endoproteolytic processing and intracellular insulin retention. Endoproteolysis promotes a change in insulin's biophysical properties. B5His residues normally participate in multimeric insulin packing; a point mutation at this position permits ICFP processing but causes the majority of processed insulin to be secreted. The data argue that multimeric assembly consequent to endoproteolytic maturation regulates insulin sorting in the secretory pathway.

Pulse-chase analysis of folded and misfolded insulin precursor (IP) expressed in Saccharomyces cerevisiae was performed to establish the requirements for intracellular transport and the influence of the secretory pathway quality control mechanisms on secretion. Metabolic labelling of the IP expressed in S. cerevisiae showed that the effect of a leader was to stabilise the IP in the endoplasmic reticulum (ER), and facilitate intracellular transport of the fusion protein and rapid secretion. The first metabolically labelled IP appeared in the culture supernatant within 2-4 min of chase, and most of the secreted IP appeared within the first 15 min of chase. After enzymatic removal of the leader in a late Golgi apparatus compartment, the IP followed one of two routes: (1) to the plasma membrane and hence to the culture supernatant, or (2) to a Golgi or post-Golgi compartment from which secretion was restricted. Combined secretion and intracellular retention of the IP reflected either saturation of a Golgi or post-Golgi compartment and secretion as a consequence of overexpression, or competition between secretion and intracellular retention. IP which was misfolded, either due to amino acid substitution or because disulphide bond formation had been prevented with dithiothreitol (DTT), was transported from the ER to the Golgi apparatus but then retained in a Golgi or post-Golgi compartment and not exported to the culture supernatant.

The yeasts Pichia pastoris and Saccharomyces cerevisiae have similar overall features regarding the secretory expression of insulin. The S. cerevisiae mating factor alpha (alpha-factor) prepro-leader facilitated the secretion of an insulin precursor, but not proinsulin expressed in P. pastoris. Synthetic prepro-leaders developed for the secretory expression of the insulin precursor in S. cerevisiae also facilitated the secretion of the insulin precursor expressed in P. pastoris. In contrast with S. cerevisiae, only insulin precursor and no unprocessed hyperglycosylated alpha-factor pro-leader/insulin precursor fusion protein was secreted from P. pastoris. A spacer peptide in the fusion protein increased the fermentation yield of the insulin precursor in P. pastoris. A synthetic prepro-leader, but not an alpha-factor prepro-leader lacking N-glycosylation sites, facilitated the secretion of the insulin precursor in P. pastoris. P. pastoris has a capacity for secretory expression of the insulin precursor that is equal to or better than that of S. cerevisiae. Peptide mapping and MS indicated a structure of the insulin precursor expressed in P. pastoris identical with that of human insulin.

Synthetic prepro-leaders lacking consensus N-linked glycosylation sites confers secretion competence of correctly folded insulin precursor expressed in the yeast species Saccharomyces cerevisiae with a yield comparable to, or better than the alpha-factor prepro-leader. In contrast, the S. cerevisiae alpha-factor prepro-leader's three N-linked oligosaccharide chains are necessary for the ability to facilitate secretion of the insulin precursor from S. cerevisiae (T. Kjeldsen et al., Biotechnol. Appl. Biochem. 27, 109-115, 1998). Synthetic prepro-leader lacking both N-glycosylation and the dibasic Kex2 endoprotease processing site also efficiently facilitated secretion of a pro-leader/insulin precursor fusion protein in which the insulin precursor was correctly folded. The unprocessed pro-leader/insulin-precursor fusion protein was purified from culture medium and matured in vitro to desB30 insulin by Achromobacter lyticus lysyl-specific protease providing an alternative yeast expression system not dependent on the Kex2 endoprotease. The synthetic prepro-leader lacking N-linked glycosylation provides the opportunity for secretory expression in yeast utilizing either in vivo Kex2 endoprotease maturation of the fusion protein during secretion or in vitro maturation of the purified fusion protein with a suitable enzyme., (Copyright 1998 Academic Press.)

Albumin is organized in three homologous domains formed by double loops stabilized by disulfide bonds. Utilizing a secretory expression system based on a synthetic secretory prepro-leader, the three human serum albumin domains were expressed in the yeast Saccharomyces cerevisiae. Human serum albumin domains I and III were efficiently expressed and secreted, indicating that these domains can form independent structural units capable of folding into stable tertiary structures. In contrast, albumin domain II was not secreted and disappeared early in the secretory pathway. Human serum albumin has the ability to bind a large number of small molecule ligands, including fatty acids, presumably due to its structure and structural flexibility. Purified albumin domain III bound myristic acid, whereas purified albumin domain I did not bind myristic acid. A new soluble long-acting insulin an alogue acylated with myristic acid (Markussen J., et al., Diabetologia 39, 281-288, 1996) bound to domain III and bound markedly more weakly to domain I., (Copyright 1998 Academic Press.)

To evaluate the possible relationship between N-linked glycosylation of the Saccharomyces cerevisiae alpha-factor pro-peptide and transport of the alpha-factor pro-peptide/insulin precursor fusion protein through the Saccharomyces cerevisiae secretory pathway, we analysed secretion of insulin precursor facilitated by alpha-factor pro-peptides with one or more of the three N-linked glycosylation sites removed. Mutation of the three alpha-factor pro-peptide N-linked glycosylation sites drastically decreased insulin precursor secretion. The three alpha-factor pro-peptide N-linked glycosylation sites differ in their ability to facilitate secretion of the insulin precursor. The two alpha-factor pro-peptide N-linked glycosylation sites localized closest to the insulin precursor contributed significantly to secretion, whereas the most N-terminally linked glycosylation site did not appear to facilitate secretion. Only correctly folded insulin precursor was found in the culture supernatant, regardless of the pro-peptide used for secretion, indicating that alpha-factor pro-peptide N-linked oligosaccharide chains are not necessary for correct folding of the insulin precursor. Thus, N-linked glycosylation facilitates intracellular transport of the alpha-factor propeptide/insulin precursor fusion protein through the Saccharomyces cerevisiae secretory pathway and secretion of the insulin precursor. N-linked glycosylation per se is not sufficient to facilitate secretion of the insulin precursor; the position of the N-linked oligosaccharide chain on the alpha-factor pro-peptide is important for facilitating efficient secretion.

Alanine scanning mutagenesis has been used to identify specific side chains of insulin which strongly influence binding to the insulin receptor. A total of 21 new insulin analog constructs were made, and in addition 7 high pressure liquid chromatography-purified analogs were tested, covering alanine substitutions in positions B1, B2, B3, B4, B8, B9, B10, B11, B12, B13, B16, B17, B18, B20, B21, B22, B26, A4, A8, A9, A12, A13, A14, A15, A16, A17, A19, and A21. Binding data on the analogs revealed that the alanine mutations that were most disruptive for binding were at positions TyrA19, GlyB8, LeuB11, and GluB13, resulting in decreases in affinity of 1,000-, 33-, 14-, and 8-fold, respectively, relative to wild-type insulin. In contrast, alanine substitutions at positions GlyB20, ArgB22, and SerA9 resulted in an increase in affinity for the insulin receptor. The most striking finding is that B20Ala insulin retains high affinity binding to the receptor. GlyB20 is conserved in insulins from different species, and in the structure of the B-chain it appears to be essential for the shift from the alpha-helix B8-B19 to the beta-turn B20-B22. Thus, replacing GlyB20 with alanine most likely modifies the structure of the B-chain in this region, but this structural change appears to enhance binding to the insulin receptor.

Secretion leaders are essential for expression of many heterologous proteins including insulin in yeast. The function of secretion leaders and their interaction with the secretory pathway is not clear. To determine what constitutes functional pre-pro-leader sequences in Saccharomyces cerevisiae, synthetic leader sequences for secretion of the insulin precursor were developed by a combination of rational design and stepwise systematic optimization. The synthetic leaders efficiently facilitate secretion of the insulin precursor from S. cerevisiae when compared with the alpha-factor leader, leading to a high yield of correctly folded insulin precursor in the culture supernatant. The synthetic leaders feature two potential N-linked glycosylation sites which are efficiently glycosylated during secretion. Pulse-chase analysis indicates that the synthetic leaders/insulin precursor fusion protein have a prolonged residence in the endoplasmic reticulum compared to the alpha-factor leader/insulin precursor fusion protein. The longer transition time in the endoplasmic reticulum mediated by the synthetic leaders might provide additional time for correct folding of the insulin precursor and account for the increased fermentation yield.

1. To investigate the structure/function relationship of the interaction between ligand and receptor in the insulin-like growth factor I (IGF-I) and insulin receptor systems we have prepared and characterized a single-chain insulin/IGF-I hybrid. The single-chain hybrid consists of the insulin molecule combined with the C domain of IGF-I. The single-chain hybrid was found to bind with high affinity to both truncated soluble insulin receptors and membrane-bound holoreceptors. The affinity for interacting with the soluble truncated insulin receptors was 55-94% relative to insulin, and affinity for membrane-bound insulin receptors was 113% of that of insulin. Furthermore we found that the affinity of the single-chain hybrid molecule for IGF-I receptors was 19-28% relative to IGF-I. 2. The affinity of the single-chain hybrid for chimeric insulin/IGF-I receptors exceeded that of either natural ligand. This indicates that coordinately changing domains of the receptors and the ligands can induce higher affinity of ligand for receptor, supporting the idea that these receptors have a common ligand-binding site [Kjeldsen, Andersen, Wiberg, Rasmussen, Schäffer, Balschmidt, Møller and Møller (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 4404-4408]. 3. In contrast with what was generally assumed about the ligand structure required for binding to the insulin receptor we demonstrate the first single-chain insulin analogue that can bind with high affinity to the insulin receptor.

The exact nature of how the insulin molecule interacts with the insulin receptor is obscure although chimeric receptors have shown that the ligand specificity of the insulin receptor and the IGF-I receptor (i.e. the sequences that discriminate between insulin and insulin-like growth factor I) reside in different regions of a common binding site and that the N-terminal 68 amino acids of the insulin receptor are involved in conferring specificity for insulin on this receptor (Kjeldsen, T., Andersen, A. S., Wiberg, F. C., Rasmussen, J. S., Schäffer, L., Balschmidt, P., Møller, K. B., and Møller, N. P. H. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 4404-4408). Using chimeric insulin/IGF-I receptors to elucidate how the insulin receptor interacts with the insulin molecule we identified phenylalanine 39 of the insulin receptor as a major contributor in determining the receptor specificity for insulin, increasing insulin affinity 15-fold when replacing the corresponding amino acid in the insulin-like growth factor I receptor. Furthermore, replacement of the insulin receptor amino acid phenylalanine 39 with the corresponding IGF-I receptor amino acid, serine 35, decreased insulin affinity 8-fold.

We have examined, by use of a hybrid insulin/insulin-like growth factor-I analog and chimeric insulin/type I insulin-like growth factor receptors, the interplay between ligand and receptor structure in determining the affinity and specificity of hormone-receptor interactions in the insulin and insulin-like growth factor-I systems. Our findings, obtained through the study of radiolabeled peptide binding to detergent-solubilized full-length receptors and to soluble truncated receptors, show that (a) the two-chain hybrid analog exhibits significant cross-reactivity with both receptor systems, (b) the exchange of appropriate domains in chimeric receptors enhances the receptor binding affinity of the analog by 3.5-21-fold, and (c) the affinity of the hybrid analog for the chimeric receptors actually exceeds that of either natural insulin or natural insulin-like growth factor-I. We conclude that the specificity-conferring domains of the insulin and type I insulin-like growth factor receptors reside in different regions of a common binding site, and that the exchange of domains between pairs of related hormones and between pairs of related receptors can yield new ligand-receptor systems with significantly altered affinities and selectivities.