Biosynthesis, Processing, and Secretion of the Islet Hormones: Insulin, Islet Amyloid Polypeptide (Amylin), Glucagon, Somatostatin, and Pancreatic Polypeptide

Published on 28/03/2015 by admin

Filed under Endocrinology, Diabetes and Metabolism

Last modified 28/03/2015

Print this page

This article have been viewed 5704 times

Chapter 6

Biosynthesis, Processing, and Secretion of the Islet Hormones

Insulin, Islet Amyloid Polypeptide (Amylin), Glucagon, Somatostatin, and Pancreatic Polypeptide

Donald F. Steiner, Christopher J. Rhodes, Louis H. Philipson, Graeme I. Bell and Shu Jin Chan

Although diabetes had long been recognized as a serious metabolic disorder, it was the classic experiments of Von Mering and Minkowski in 1890 that demonstrated the important role of the pancreas in its prevention.¹ A number of investigators attempted to prepare antidiabetic pancreatic extracts in the early 20th century, but it was not until the work of the Canadians—Frederick Banting, Charles Best, and John Collip, working in the laboratory of Professor J.J.R. Macleod at the University of Toronto in 1921-1922—that potent preparations of insulin were routinely made.² The name insulin was based on the belief that the hormone was derived from the islets of Langerhans and actually was suggested as early as 1909 by deMayer and by Sir Edward Sharpey-Schafer in 1917. Preparative methods based on the early work of Scott,³ as well as refinements by Banting et al.,⁴ were rapidly developed for the commercial preparation of the hormone, and within about 1 year, insulin began to be administered to patients with diabetes, often with dramatic effects.² The chemical nature of insulin was a more elusive problem, although the fact that it was destroyed by proteolytic enzymes suggested that it was a protein.⁵ When J.J. Abel first crystallized insulin in 1926, there was great controversy as to whether the crystals of protein he obtained were actually the active biological principle or merely the vehicle for a smaller active moiety.⁶ Today we know that insulin occurs throughout the vertebrate kingdom, and insulin-like substances are present in the brains and/or digestive systems of many invertebrates.⁷ Many of the modern techniques of protein chemistry were developed in part through the clinical need for insulin, which made it abundantly available to biochemists as a model protein for study.⁸^,⁹

This chapter reviews the isolation, structure, and properties of insulin and other islet hormones, including islet amyloid polypeptide (IAPP or amylin), glucagon, somatostatin, pancreatic polypeptide (PP) and their precursor forms, and the mechanisms of their biosynthesis and secretion. Studies on the mechanism of biosynthesis of insulin via preproinsulin and proinsulin in the pancreatic β cells have provided a useful paradigm for understanding the production of many other neuroendocrine peptides and a basis for advances in the diagnosis and treatment of diabetes and related endocrine disorders. The recent identification of the enzymes that catalyze the proteolytic cleavage of prohormones and proneuropeptides—prohormone convertases—has revealed a new family of processing enzymes that act on a larger subset of precursor proteins, including growth factors, their receptors, and many other proteins that traverse the secretory pathway.¹⁰

The Biosynthesis of Insulin

General Aspects

The two-chain structure of insulin was the first complete protein structure to be elucidated through the pioneering work of Fred Sanger in 1955.⁹ In 1967, the discovery of the single-chain insulin precursor molecule, proinsulin,¹¹ then solved the problem of assembly of the insulin chains (Fig. 6-1). The human proinsulin polypeptide begins with the 30-amino-acid B chain domain, extends through a 35-amino-acid connecting segment, and ends with the 21-amino-acid A chain domain, comprising a single 9000-D polypeptide chain.^11–14 It contains paired basic residue cleavage sites (Arg-Arg or Lys-Arg) at each end of the connecting segment which are removed during the proteolytic conversion of proinsulin to insulin within maturing secretory granules. However, cell-free translation of the initial polypeptide products encoded in insulin messenger ribonucleic acid (mRNA) extracted from islets or islet cell tumors led in 1976 to the discovery of preproinsulin (Fig. 6-2). This extended form of the prohormone has a hydrophobic N-terminal 24-residue signal peptide.¹⁵ Such prepeptide extensions interact with the signal recognition particle (SRP), a cytosolic ribonucleoprotein particle that assists in transfer of the nascent secretory protein from the cytosolic compartment, where its synthesis is initiated, into the secretory pathway via a complex series of molecular interactions that result in the translocation of the nascent peptide across the membrane of the rough endoplasmic reticulum (RER) into its internal compartments, or cisternae.¹⁶ During or shortly after translocation, the signal sequence is cleaved by the signal peptidase, which is located on the inner surface of the RER membrane, and the signal peptide is rapidly degraded.¹⁷ The proinsulin molecule then folds and undergoes rapid formation of disulfide bonds to achieve its native structure. Evidence suggests that this process is catalyzed by protein disulfide isomerase,¹⁸ a resident protein of the RER having a C-terminal Lys-Glu-Asp-Leu (KDEL) localization sequence.¹⁹ The folded proinsulin is then transferred in small coated vesicles from the endoplasmic reticulum (ER) to the cis region of the Golgi apparatus.²⁰^,²¹ It then passes from the cis to the trans Golgi, where it is sorted into immature secretory vesicles, or progranules, where it is proteolytically cleaved to release insulin and the connecting segment, minus the four basic residues, from the cleavage sites. The connecting peptide fragment released by cleavage of proinsulin is designated the C peptide. Both insulin and C peptide are stored in the secretory granules along with small amounts of residual proinsulin and partially cleaved intermediate forms²² as well as a variety of other minor β cell secretory products.²³ Stimulation of secretion results in the release of the granule contents into the hepatic portal circulation.

FIGURE 6-1 Amino acid sequence of human and hagfish insulins, with substitutions occurring at each position in 70 other known vertebrate insulins shown below for comparison. Invariant residues are enclosed in boxes.

FIGURE 6-2 Schematic structure of human preproinsulin. Rapid removal of the first 24 amino acids (signal peptide) gives rise to proinsulin, which then folds to form the native insulin domain stabilized by three disulfide bonds (black bars). The connecting segment is not structurally organized. Cleavage after B30 (T) and before A1 (G) then gives rise to insulin and the free C peptide. (See text for details.)

Morphologic Organization of the Insulin Biosynthetic Machinery

The β cells of the islets of Langerhans share many features with other endocrine and neurosecretory cells. Preproinsulin is the initial translational product of insulin mRNA, as discussed earlier, and is rapidly cleaved to proinsulin in the RER. After folding and disulfide oxidation, proinsulin is transported from the ER to the Golgi apparatus in small coated vesicles ²⁰ (Fig. 6-3) in an energy-dependent process that requires about 20 minutes.^22,24,25 The biochemical basis of the energy requirement for the intracellular transport of secretory proteins is associated with the budding and/or fusion of small vesicles that transfer secretory products from the ER to and through the cis to trans Golgi cisternae.²⁶^,²⁷

FIGURE 6-3 Schematic model of the subcellular transport of preproinsulin after its synthesis on membrane-bound ribosomes of the rough endoplasmic reticulum (RER) and rapid cleavage to proinsulin (within 1 to 2 minutes). Proinsulin is then released into the intracisternal spaces of the RER, where it folds and forms the native disulfide bonds of insulin. It is then transported to the Golgi apparatus by an energy-dependent process. The clathrin-coated early granules budding from the trans-Golgi cisternae are rich in proinsulin and contain the converting proteases, PC2 and PC1/3. Processing occurs mainly, if not exclusively, in the early secretory granules ²⁴^,³⁰ giving rise to the more condensed mature granules. Fractionation studies have confirmed that the mature granule-dense cores consist almost entirely of insulin, often in crystalline arrays (see Fig. 6-9), whereas the granule-soluble phase that surrounds the inclusion consists mainly of C peptide and small amounts of proinsulin.⁸⁴ The release of newly synthesized proinsulin and insulin begins only about 1 hour after synthesis in the RER; hence, granules must undergo a maturation process that renders them competent for secretion. There is no evidence for significant nongranular routes of secretion of either proinsulin or insulin in normal islets. Exocytosis of granules is regulated by glucose and many other factors (see section on insulin secretion), and in humans and dogs, it results in the release of insulin and C peptide in approximately equimolar proportions under both basal and stimulated conditions.⁹⁵ The mechanism of recycling of the granule membrane and its components is not well understood.

The role of the Golgi apparatus in forming insulin-containing secretory vesicles, or β-granules, was first recognized in the 1940s and subsequently confirmed and extended by Munger,²⁸ who used electron microscopy to identify pale, homogeneous-appearing, immature “progranules” near the Golgi body. The structural organization and various functions of this complex organelle continue to engage the attention of cell biologists and biochemists.²⁹ Orci and coworkers demonstrated that these newly formed clathrin-clad vesicles are derived from the trans-Golgi network (TGN) cisternae and contain a high proportion of proinsulin, definitively confirming their proposed role as the major sites of proinsulin processing to insulin.³⁰ Conversion of proinsulin to insulin in vivo behaves kinetically as a pseudo–first-order reaction having a half-time ranging from about 20 minutes to 1 hour.^22,31,32 Peak labeling of proteins in the Golgi apparatus occurs 30 to 40 minutes after pulse labeling islets with ³H-amino acids, and relatively little radioactivity remains in this region after 1 hour.³³ A similar pattern for proinsulin is observed by means of electron microscopic immunocytochemistry. Thus, although small amounts of proinsulin conversion may be initiated in the TGN, most of it occurs in newly formed secretory vesicles, or progranules, which mature biochemically in the cytosol³⁰ (see Fig. 6-3). Numerous subsequent studies have confirmed that newly synthesized neuroendocrine precursor peptides pass via the Golgi apparatus into secretory granules of the regulated secretory pathway, where they are cleaved and processed to their mature products and then stored until secretion in response to signals such as glucose.¹⁰

Structure and Properties of Proinsulin

Mammalian proinsulins range in size from 81 (cow) to 86 (human, horse, rat) amino acid residues, owing to variations in the length of the C peptide. Various C-peptide sequences are compared in Fig. 6-4. Proinsulin is closely similar to insulin in many properties, including solubility, isoelectric point,²² self-associative properties,³⁴ and reactivity with insulin antisera.¹¹^,³⁵ Indeed, the conformation of the insulin moiety in proinsulin is essentially identical to that of insulin itself.³⁶ It is noteworthy that the highly flexible C peptide is much longer than necessary simply to bridge the short 8-Å gap between the ends of the B and A chains as they exist in the folded insulin molecule (see Fig. 6-2). The C-peptide segment overlays a portion of the surface of the insulin monomer, reducing its potency to about 3% to 5% of that of insulin in vitro,²²^,³⁷ and it undergoes little if any proteolysis or activation in the circulation or tissues.³⁸ The presence of the C peptide in proinsulin also does not prevent the dimerization or hexamerization of its insulin moiety³⁹ (Fig. 6-5). It lies outside the zinc-stabilized globular insulin hexamer in a disordered loop, as assessed by nuclear magnetic resonance (NMR) spectroscopy. Since low levels of proinsulin can form mixed hexamers with insulin, it can be efficiently incorporated into insulin crystals.⁴⁰ This property accounts for the presence of 1% to 2% of proinsulin in most crystalline preparations of insulin prepared from animal pancreata.⁴¹ However, x-ray analysis of crystals of proinsulin have failed to reveal any details regarding the structure of the C-peptide segment.⁴²^,⁴³

FIGURE 6-4 Compilation of amino acid sequences of proinsulin C peptides in vertebrates, amphioxus, and in a mollusk, Lymnae stagnalis (for sources see Ref. 96).

FIGURE 6-5 Hypothetical 2-Zn proinsulin hexamer as viewed along the threefold axis. The connecting peptide is shown in light gray and white around the periphery of the darker outline of an insulin hexamer arranged according to the structural data of Blundell et al.⁸⁸ The central density represents two zinc atoms in coordination linkage to the six (three above and three below hexamer plane) histidine side chains at position 10 in the B chain.

The major intermediate cleavage products of proinsulin consist mainly of forms cleaved at only one site and lacking the exposed basic residues (i.e., des-31,32 or des-64,65 intermediates).12,22,41,44,45 Their relative ratios in the circulation differ among species, depending on differences in the amino acid sequences surrounding the cleavage sites.⁴⁶ Proinsulin-like proteins having comparably sized C peptides have been found throughout the vertebrates and also in a variety of invertebrates. Although it seems likely that the single-chain insulin-like growth factors, IGF-1 and IGF-2, arose from a single proinsulin-like gene in protochordates,⁴⁷ some invertebrates, such as Caenorhabditis elegans,⁴⁸ have numerous insulin-like peptides that lack defined C peptides and typical paired basic residue sites for cleavage by the evolutionarily well-conserved prohormone convertases.

Enzymatic Basis of the Conversion of Proinsulin to Insulin

The major proteolytic cleavages required for converting proinsulin to insulin are summarized in Fig. 6-6. In early studies, it was shown that the joint action of pancreatic trypsin and carboxypeptidase B gives rise to the naturally occurring conversion products—C peptide and native insulin in high yields—quantitatively converting proinsulin to insulin in vitro.⁴⁹ This model also explains the major intermediate forms found in pancreatic extracts,⁴⁴^,⁵⁰ and it led to a search for a cellular trypsin-like convertase, assumed to be a serine protease related to trypsin,²² and for carboxypeptidases related to carboxypeptidase B, an exocrine pancreatic exopeptidase with specificity for C-terminal basic residues.⁴⁹ Early studies with isolated islet secretory granules revealed that these were major sites of proinsulin conversion, and that conversion resulted in the release of insulin along with free arginine, rather than basic dipeptides,⁵¹ confirming the likely participation of both trypsin-like and carboxypeptidase B–like proteases in maturing secretory granules.⁵² Carboxypeptidase E (or H), a homolog of carboxypeptidase B but with a more acidic pH optimum, was first identified in islets⁵² and subsequently by Fricker in brain.⁵³ Molecular cloning confirmed its structural and evolutionary relationship to the pancreatic carboxypeptidases. Additional processing carboxypeptidases have been uncovered in brain and other tissues.⁵⁴

FIGURE 6-6 The cleavage of proinsulin to insulin by the combined action of the subtilisin-like prohormone convertases, PC2 and PC1/PC3, and carboxypeptidase E. (See text for details.)

The first authentic processing endoprotease, kexin, was identified through studies on the processing of the yeast alpha mating–factor precursor and proved to be a calcium-dependent serine protease related to bacterial subtilisin and the subtilases rather than to trypsin.55–57 This information enabled the discovery of mammalian homologs, which form a family of seven closely related endoproteases with differing points of action within the secretory pathway.^10,58,59 These have been designated as subtilisin-like proprotein convertases (SPCs), or more simply as PCs. Two members of this family, PC1/3 and PC2, have proven to be the principal effectors of neuroendocrine peptide precursor processing and are largely confined in their expression to brain and endocrine tissues in most animal species.¹⁰ For further information on the extended SPC/PC family and its multiple functions, please see the reviews in Refs. 60 to 62.

Role of PC2 and PC1/3 in Proinsulin Processing

Both PC1/3 and PC2 are required for proinsulin processing and have been shown to be identical with the calcium-dependent type I and type II insulinoma granule-processing activities, respectively, originally described by Davidson et al.⁶³ The importance of calcium for the transport and proteolytic maturation of proinsulin has been well documented.⁶⁴ The role and order of action of PC2 and PC1/3 in proinsulin conversion has also been carefully studied.⁶⁵ Rhodes et al.⁶⁶ demonstrated that the type II convertase (PC2) prefers the proinsulin intermediate that has already been processed at the B chain–C peptide junction (des-31,32 intermediate) as a substrate. This observation has led to the scheme for conversion outlined in (Fig. 6-7), in which PC1/3 acts first to generate the des-31,32 intermediate, which is then cleaved at the C peptide–A chain junction by PC2. This order of cleavage is consistent with observations that PC1/3 achieves an enzymatically active form more rapidly than PC2 and has a somewhat higher pH optimum. Thus, PC1/3 may begin cleaving proinsulin as it is concentrated into immature secretory vesicles in the TGN, whereas PC2 acts only in the maturing granules as the pH decreases to around 5.5. According to this scheme, PC1/3 plays a more important role in proinsulin processing, as born out by observations on islets from mice lacking PC2 or PC1/3.^46,67–69

FIGURE 6-7 Routes of processing of proinsulin in the pancreatic β cell. The pathway on the right is probably more dominant under normal conditions because des-31,32 proinsulin is a preferred substrate for PC2,⁶⁶ and the more acidic pH optimum and slower maturation of this enzyme may delay its action during the initial phases of secretory granule maturation. The C-terminal basic residues are removed by CPE (CPH) after endoproteolytic cleavage by the PCs. (Modified from Rouillé Y, Duguay SJ, Lund K et al: Proteolytic processing mechanisms in the biosynthesis of neuroendocrine peptides: the subtilisin-like proprotein convertases, Front Neuroendocrinol 16:322–361, 1995.)

PC2-null mice are not diabetic, but they exhibit significant hyperproinsulinemia, with plasma proinsulin levels in the range of 60%.⁶⁷ Pancreatic extracts also show increased proinsulin levels, but only in the homozygous nulls, as indicated in Table 6-1. Pulse-chase studies of insulin biosynthesis in isolated islets, comparing PC2^(−/−) mice with wild-type (WT) controls, also confirm significantly slower processing of proinsulin to insulin, with accumulation of significant amounts of des-31,32 intermediate proinsulin.⁴⁶^,⁶⁷ Approximately a third of the labeled proinsulin remains after a 3- or 4-hour chase, consistent with the levels found by radioimmunoassay in pancreatic extracts. Thus in normal mice, PC2 converts at most about a third of the available proinsulin, while PC1/3 is responsible for processing the remaining two-thirds.⁶⁹

Table 6-1

Percent Proinsulin-Like Immunoreactivity in Pancreatic Extracts of Wild-Type, PC2-Null, and PC1/3-Null Mice

A human subject with inactivating mutations in both copies of the PC1/3 gene has been reported.⁷⁰ This 43-year-old woman is obese and had gestational diabetes. Examination of the subject’s serum revealed no detectable circulating insulin, in association with greatly elevated intact proinsulin and significant amounts of des-64,65 intermediate proinsulin but little or no des-31,32 intermediate proinsulin. Similar results with respect to proinsulin processing have been reported for the PC1/3-null mouse, which has very high levels of proinsulin and elevated des-64,65 intermediate proinsulin⁶⁸^,⁶⁹ (see Table 6-1). This picture is consistent with the conversion scheme outlined in Fig. 6-7 and indicates that the pathway shown on the right side of the diagram is the predominant one. It also confirms the likelihood that PC1/3 is the more important convertase in the processing of proinsulin. The PC1/3-null mice, however, are not obese like the human subject; instead, they exhibit a severe growth defect, most likely due to lack of normally processed growth hormone–releasing hormone.⁶⁹ These mice are also unable to produce GLP-1 but did not have altered intraperitoneal glucose tolerance. Altered proopiomelanocortin processing and other endocrine abnormalities, as well as intestinal malfunction, also are present in these mice.⁶⁸

Significance of Circulating Proinsulin and Des-31,32 Intermediate in Man

An important clinical issue concerns observations that in man, des-31,32 intermediate proinsulin is a major intermediate, making up a very significant proportion of the circulating proinsulin-like material.⁷¹ It has been suggested that the elevated accumulation of proinsulin and des-31,32 proinsulin in diabetics might be due to defective action of PC2.⁷² Isolated normal human islets of Langerhans also have been reported to convert proinsulin to insulin with significant accumulation of des-31,32 intermediate, despite the presence of normal levels of PC2.⁷³ However, des-31,32 intermediate only reaches levels of 15% to 20% of total immunoreactive insulin-like material during biosynthetic pulse-chase studies in the islets of PC2-null mice.⁴⁶ Thus, even the complete absence of PC2 should not in itself give rise to such high levels of this intermediate as are seen in human serum samples. It has recently been clarified that this phenomenon arises due to the faster rate of cleavage by the PCs of substrates having a basic residue in the fourth position upstream (P4 position) of the usual dibasic cleavage sites. Such an upstream basic residue is present at the B chain–C peptide junction in human proinsulin (B29 is Lys) but is lacking at the A chain–C peptide junction (C62 of proinsulin is Leu) (see Fig. 6-6). This causes an imbalance in the relative susceptibility of these two sites to either of the two convertases and tends to favor the accumulation of the des-31,32 intermediate.⁴⁶ On the other hand, in the dual proinsulins of rats and mice,²² other P4 sequence changes in these cleavage sites produce different ratios of intermediates than those seen in man.⁴⁶

Studies on the regulation of the biosynthesis of PC2 and PC1/3 in β cells suggest that the rates of translation of both of these enzymes are up-regulated by glucose, similar to that of proinsulin on glucose stimulation.⁷⁴ Also both PC2 and PC1/3 mRNA levels are equally elevated along with insulin mRNA during more prolonged stimulation of islets with glucose.⁷⁵ However, under conditions of chronic stimulation, a relative deficiency of PC2 could develop that could exaggerate the abnormalities in circulating proinsulin intermediates seen in diabetics. Accordingly, the normal accumulation of des-31,32 intermediate proinsulin in human subjects merely reflects sequence differences in the two cleavage sites of human proinsulin, whereas increased levels of both proinsulin and des-31,32 proinsulin in prediabetics and diabetics may result from a deficiency of convertase action when islets are stressed by hyperglycemia. Genetic studies have not yet indicated a major role for mutations in the PC1/3,⁷⁶ PC2,⁷⁷ or CPE⁷⁸ genes in susceptibility to any form of diabetes, although common amino acid polymorphisms in PC1/3 may increase risk of obesity.⁷⁹ However, loss-of-function mutations in the transcription factor PAX6 have recently been described which result in reduced expression of PC1/3 in the islets and are associated with hyperproinsulinemia and mild age-related diabetes.⁸⁰

Secretory Granule Formation and Maturation

One of the unsolved puzzles of neuroendocrine and other secretory cells is the nature of the mechanism underlying the efficient sorting of proteins destined for regulated secretion into immature secretory vesicles in the TGN. In the β cell, this process is remarkably efficient, resulting in very low levels of unregulated or “constitutive” release of proinsulin (<1% to 2%). However, the newly formed secretory granules have a clathrin coat (Fig. 6-8) which appears to be involved in some reorganization of the granule contents after and/or during their formation in the TGN.⁸¹ This passive sorting presumably occurs via small clathrin-clad vesicles that transport some proteins that are excluded from the condensing granule cores into endosomal pathways that either recycle to the TGN or to the cell surface. As a consequence of this “constitutive-like” pathway, proteins such as furin, procathepsin B, and possibly others briefly pass through the immature granule compartment, where they may play an active albeit transient role (e.g., furin may participate in processing of some prohormones).⁸¹^,⁸² Also, small amounts of abundant soluble granule components such as proinsulin and/or C peptide may exit the granules within these vesicles.⁸³ Passive sorting may also play a role in maintaining synchrony between granule membrane area and granule volume as maturation proceeds.⁸¹

FIGURE 6-8 A, Clathrin immunolabeling; B, 3-(2,4-dinitroanilino)-3′-amino-N-methyldipropylamine (DAMP) immunolabeling; and C, proinsulin immunolabeling of Golgi areas (G) of B cells (protein A–gold techniques). These electron micrographs show that the population of secretory granules with tightly fitting cores (black arrowheads) is clathrin coated (arrows in A), DAMP poor (B), and proinsulin rich (C). These granules correspond to the maturing coated secretory granules freshly released from the Golgi complex. Conversely, secretory granules with wide, clear halos (white arrowheads) are deprived of clathrin (not shown here), are DAMP rich (B). and are proinsulin poor (C). These correspond to the noncoated mature (storage) insulin-containing secretory granules. Because DAMP immunoreactivity is assumed to represent an indirect measure of intraorganelle acidity, this may indicate a decreasing pH gradient between proinsulin-rich and insulin-rich granules (i.e., between the converting and the storage compartments) (A, ×28,000; B and C, ×27,000). (From Orci L: The insulin cell: its cellular environment and how it processes proinsulin, Diabetes Metab Rev 2:71, 1986.)

Newly formed secretory vesicles in neuroendocrine cells undergo biochemical and morphologic maturation to typical dense-core granules. In β cells, the progranules are larger, less dense, and more uniform in appearance.²⁹^,³⁰ The most important biochemical change taking place as these progranules mature is the proteolytic conversion of proinsulin to insulin, accompanied by reorganization of the products.³⁰ Electron microscopic studies indicate that they progressively acquire a crystalline dense central core (Fig. 6-9). High magnification reveals repeat-unit spacings in the cores that are closely similar to those observed in ordinary zinc insulin crystals.⁸⁴ Thus, as insulin is liberated from proinsulin, it tends to crystallize with zinc that is concentrated by the β cells. Recent studies have demonstrated that zinc is taken up into maturing granules via the ZnT8 zinc transporter, and they do not assume the normal dense-core appearance in its absence, even though proinsulin conversion is not impaired.⁸⁵ Biochemical fractionation of mature islet secretory granules confirms that the cores contain only insulin, whereas the C peptide liberated in the conversion process remains in solution in the clear fluid space surrounding the dense crystalline core.⁸⁴ There is no evidence for co-crystallization of the C peptide with insulin under these conditions or in vitro.

FIGURE 6-9 A, Photomicrograph of normal rat β cells (×28,000) showing morphology of mature granules (bar = 1 µm). B, Isolated rat β-granule cores (×17,000)²³⁶ (bar = 1 µm). C, High-magnification view (×250,000) showing repeat unit structure of a crystalline core (bar = 0.1 µm). The cores are made up of both rat insulin I and II in approximately equal proportions (Michael J and Steiner DF, unpublished data). Samples were fixed with Karnovsky’s solution and stained with osmium tetroxide. (Electron micrographs courtesy Hewson H. Swift.)

Most of the zinc in islets is present in the β granules and is liberated proportionately to insulin during secretion, in keeping with its role in crystallization of the hormone.⁸⁶ The insulins of some species, including the guinea pig, coypu, and other hystricomorph rodents⁴² and the primitive hagfish,⁸⁷ lack the histidine residue at position 10 of the B chain required for zinc binding during the association of insulin dimers into hexamers.⁸⁸ A human subject heterozygous for a mutation at B10 (His B10 Asp) had elevated levels of this proinsulin in the blood because it tended to be released constitutively at much higher levels than normal proinsulin.⁸⁹ However, in a transgenic mouse model of this mutation, there was no defect in its conversion to insulin in the granules, despite its altered self-association and sorting properties.⁹⁰

The pH of the interior of the mature secretory granule appears to be between pH 5.0 and 6.0,⁵¹^,⁹¹ an optimal pH range for insulin crystallization in vitro. The neutral or slightly alkaline pH in the cisternal spaces of the RER favors proinsulin folding and sulfhydryl oxidation. The pH remains near neutral throughout the Golgi apparatus but becomes more acidic (pH 6.1) in the TGN as the secretory products are sorted into granules and proteolytic processing begins. Vesicular proton pumps may begin to increase the uptake of protons, which then displace the cationic arginine and lysine residues liberated during conversion. As these move out of the granules and are replaced by hydrogen ions, a downward shift in intragranular pH may occur. Thus the initially mildly acidic progranules⁹¹ undergo gradual acidification as they mature in the cytosol (see Fig. 6-8), creating appropriate conditions for the conversion and crystallization of the newly formed insulin. One recently discovered role of the constitutive pathway convertase, furin, in the β cell is to activate the proton pump, which acidifies the secretory granules. As a result, proinsulin processing is impaired in β cells lacking furin.⁹²

The cellular processes related to the biosynthesis of insulin via preproinsulin and proinsulin and their intracellular transport, sorting, proteolysis, and ultimate storage in secretory granules are remarkably well integrated, both topologically and biochemically. This delicately poised integration of processes leading to the formation and storage of insulin is disturbed in islet cell tumors, which often show unregulated release of insulin together with large amounts of proinsulin; measurements of the latter can provide a useful diagnostic indicator ⁹³ (see Chapter 21).

Biosynthetic Role and Biological Actions of the C Peptide

Because of its cosecretion with insulin in essentially equimolar amounts,22,94,95 the C peptide has been of great value as a marker of insulin secretion in humans under a variety of conditions. Representative vertebrate C-peptide amino acid sequences are compared in Fig. 6-4.⁹⁶ These peptides exhibit a 15-fold higher rate of mutation acceptance than do the corresponding insulins, a finding that has often been interpreted as indicating that this region in the proinsulin molecule is unlikely to have any specific hormonal function. Nonetheless, several acidic residues are consistently present at certain positions in mammalian C peptides. These offset the added cationic charges due to the pairs of basic residues at the proinsulin cleavage sites, such that the isoelectric pH of proinsulin is nearly the same as that of insulin (i.e., in the range of pH 5.1 to 5.5).²²

A large body of evidence supports a biosynthetic role for the C-peptide segment in proinsulin. Primarily, it converts the insulin A and B chain interaction from an inefficient bimolecular reaction to a highly efficient and concentration-independent unimolecular reaction.⁹⁷ Certain regions of the connecting peptide may also facilitate the folding of the proinsulin polypeptide chain and the formation of the correct disulfide bonds, or guide the enzymatic cleavage of proinsulin to insulin by helping to orient the basic residue pairs for efficient binding and cleavage by the convertases.⁹⁸^,⁹⁹ Recent molecular modeling studies indicate that the C peptide may assist in achieving important conformational changes in the N-terminal helical region of the A-chain segment, aiding its interaction with the convertases during processing.¹⁰⁰

The conservation of a length of 30 to 35 amino acids in almost all proinsulin C peptides, including those of insects, mollusks, and nematodes, despite the short 8- to 10-angstrom separation between A1 and B30 in the native insulin molecule ⁸⁸ (see Fig. 6-2), supports the existence of an important length-related function. Since much shorter connecting segments do not impair sulfhydryl oxidation and formation of the native structure,⁹⁹^,¹⁰¹ it is clear that this is not a critical constraint. A highly plausible reason for the evolutionary retention of a relatively long C peptide in proinsulin may be to facilitate the translocation of proinsulin across the RER membrane during its synthesis. The length of polypeptide chain required to span the large ribosomal subunit and the RER membrane has been estimated to be about 65 residues in extended configuration.¹⁰² Moreover, the arrest of translation by the signal recognition particle (SRP) occurs only after synthesis of a nascent chain of about this length and thus may play a role in translational control of insulin biosynthesis.¹⁰² Efficient intracellular transport and correct targeting to secretory granules may impose additional demands on the primary (and tertiary) structure of proinsulin and other precursor proteins. However, miniproinsulins with deleted or greatly truncated C peptides do not fold efficiently but are correctly targeted to the regulated secretory pathway.^98,101,103

Another very important function of the C-peptide is to facilitate its removal from proinsulin after it has served its biosynthetic functions, as mentioned earlier. The kinetics of insulin metabolism in vivo, as well as its biological activity, are both dependent upon its high affinity for the insulin receptor. Proinsulin’s much lower receptor-binding affinity leads to a much slower rate of clearance, which explains its relatively higher apparent biological activity in vivo. The rapid clearance of insulin is adaptively as important a characteristic as its bioactivity. To assure its interaction with the convertases, proinsulin must present more than just a pair of basic residues to the convertases. Their active sites require an extended β strand that is at least eight residues in length for recognition at each cleavage site. Molecular modeling studies have suggested how the full length of the connecting segment contributes to achieving the necessary conformations for effective proinsulin binding and cleavage by the convertases.¹⁰⁰

Since the early 1990s, a number of reports have appeared describing a variety of biological actions of the C peptide and/or smaller peptides derived from it.¹⁰⁴ These putative effects include enhancement of glucose transport and utilization; improvements in microcirculation in muscle, skin, retina, and nerve in diabetics; and stimulation of renal tubular Na⁺,K⁺-adenosine triphosphatase (ATPase) activity and other parameters of renal function. It has been suggested that tissue receptors for C peptide might exist, and that the circulating C peptide may contribute to improved glycemic control and help to slow the development of the vascular and neural complications of diabetes.¹⁰⁵ Further study is needed in this area in view of reports that some C-peptide effects do not follow the usual rules of ligand-receptor chemistry (i.e., chirality does not matter). Thus, a C peptide synthesized entirely of d amino acids is equally as active as a peptide made with l amino acids in reversed order,¹⁰⁶ whereas a random sequence of the same amino acids leads to loss of activity. These tantalizing findings suggest the need for a controlled clinical trial of combined insulin and C-peptide therapy in diabetics to fully evaluate its therapeutic potential.

The availability of biosynthetic human proinsulin and insulin, as well as of human C peptide, has opened many new possibilities for studies of the role, metabolism, and antigenicity of these peptides.107–109 In recent years, active insulin analogs with shortened linking peptides have also been developed. Such miniproinsulins and/or “single-chain insulins” appear to readily oxidize to form correct disulfide bridges and either can be used intact or after cleavage to insulin.^98,99,101 Other efforts have been directed towards synthesis of modified insulins with improved absorption, aggregation, and/or pharmacokinetic properties for therapeutic use.

Regulation of Insulin Production

Under normal circumstances the pancreatic β cell retains a remarkable state where insulin secreted in response to a stimulus is rapidly replenished at the biosynthetic level in parallel, thus keeping intracellular insulin stores optimal. The vast majority (98% to 99%) of insulin is secreted via the “regulated” secretory pathway ¹¹⁰ and only occurs in response to increases in specific nutrients, certain hormones, neuropeptides, and some pharmacologic reagents.¹¹¹^,¹¹² In general, most nutrients that regulate insulin secretion also control proinsulin biosynthesis,¹¹¹ and the most physiologically relevant of these is glucose.¹¹² Glucose metabolism is required for both glucose-induced insulin secretion and production,¹¹¹ but the secondary stimulus-coupling mechanisms are quite distinct. For example, stimulated insulin secretion is Ca²⁺ dependent, but glucose-induced proinsulin biosynthesis is Ca²⁺ independent.¹¹²^,¹¹³ There are effective inhibitors of glucose-induced insulin secretion, such as diazoxide, epinephrine, and somatostatin, but these have no effect on glucose-induced proinsulin biosynthesis.¹¹¹ In contrast, effective stimulators of insulin secretion, such as sulfonylureas and fatty acids, have no effect on proinsulin biosynthesis.¹¹¹^,¹¹⁴ This differential regulation of insulin secretion and proinsulin biosynthesis indicates that glucose-induced proinsulin biosynthesis is not regulated by a positive feedback of locally secreted insulin.¹¹³ For stimulated insulin secretion, the intracellular secondary signaling mechanisms are relatively well defined,¹¹⁵ but those for glucose-induced proinsulin biosynthesis, other than a need for mitochondrial metabolism and subsequent succinate production,¹¹⁶ are not currently well understood. However, it is known that both glucose-induced insulin secretion and production can be potentiated by an elevation in [cAMP]_i secondary to the action of incretins such as GLP-1.¹¹⁷ The glucose-dependent effect of GLP-1 to increase proinsulin biosynthesis is predominately at the translational level, not at the transcriptional level.¹¹⁸

Glucose regulation of proinsulin biosynthesis is relatively specific and occurs independently of the control of general protein synthesis in the β cell.¹¹⁹ There is a minor subset of β-cell proteins (~50) that are also specifically regulated by glucose at the biosynthetic level in parallel to proinsulin, with these mostly being insulin secretory-granule proteins, including the proinsulin-processing endopeptidases, PC1/3 and PC2.^{74,75,120,121} This provides a mechanism whereby the enzymes that process proinsulin and the intracellular storage compartment in which insulin is stored are coordinately increased in response to the up-regulation of proinsulin biosynthesis.

The mechanism for specific regulation of glucose-induced proinsulin biosynthesis is complex and can occur at multiple levels depending on the temporal exposure to elevated glucose levels. Under normal fluctuations in glucose concentrations that occur over periods of ≤3 hours, glucose-induced proinsulin biosynthesis is entirely mediated at the translational level.111,113,122,123 For periods of ≥12 hours, this translational regulation is supplemented by glucose promoting preproinsulin mRNA stability¹²⁴ to provide more preproinsulin mRNA template for translational regulation.¹¹³ Longer-term increased proinsulin biosynthesis can also be contributed to by increased insulin gene transcription.¹²⁵^,¹²⁶

However, even under these longer-term circumstances, the primary control of glucose-induced proinsulin biosynthesis is mediated at the translational level. Yet despite glucose-induced translational regulation of proinsulin biosynthesis being known for nearly 40 years,22,122,127 the molecular mechanism remains unresolved. There are adaptive responses of the general translational machinery to up-regulate preproinsulin mRNA translation that have been shown to occur at the initiation and elongation phases of translation,^22,123,127 as well as reduction in the duration of SRP signal peptide–mediated arrest of nascent preproinsulin-chain elongation prior to ribosome docking in the early phases of RER translocation.¹²³ However, the specific translational regulation of proinsulin biosynthesis lies in the preproinsulin mRNA itself, particularly in the untranslated regions (UTRs).¹²⁸ The 3′ UTR of preproinsulin mRNA contains elements such as a polypyrimidine-rich tract (PRT)¹²⁹ and a highly conserved UUGAA element¹¹¹^,¹²⁸ that likely control its stability. The control of preproinsulin mRNA stability via the 3′ UTR PRT also involves a glucose-responsive association of the polypyrimidine tract binding protein (PTB).¹²⁸ The 5′ UTR of preproinsulin mRNA contains a conserved translational control cis element, ppIGE (for preproinsulin glucose element), that is required for glucose-induced translational control of proinsulin biosynthesis.¹²⁸^,¹³⁰ The ppIGE has a highly conserved ppIGE “mirrored” core of GUCx_nCUG or GUUx_nUUG (where n ≤ 4 bases). There is a protein trans-acting factor (ppIE-BP) found in islet cytosolic extracts that binds to this translational control ppIGE cis element of preproinsulin mRNA in a glucose-dependent manner that parallels the characteristics of glucose-regulated proinsulin biosynthesis.¹³⁰ The identity of the ppIGE-BP is currently unknown. Nonetheless, it is intriguing that the ppIGE is also conserved in the 5′ UTR of the mRNA templates of other secretory granule proteins whose biosynthesis is specifically controlled at the translational level in β cells, including PC1/3 and PC2.⁹⁸^,⁹⁹ As such, the glucose-induced specific translational control of proinsulin biosynthesis and that for other insulin secretory-granule protein is likely via the same mechanism. For the longer-term transcriptional regulation of insulin gene expression by glucose, this is known to be a Ca²⁺-dependent mechanism¹³¹ involving the putative phosphorylation of key β-cell transcription factors, Pdx-1 and MafA.¹²⁵ Prolonged glucose stimulation can also lead to increased β-cell hyperplasia and hypertrophy,¹³² which might also be additionally influenced by a wide variety of growth factors and other nutrients.¹³³ Increased numbers of β cells will also increase insulin production, as occurs in nondiabetic obesity and pregnancy.¹³³