Glucose stimulates the activity of the guanine nucleotide-exchange factor eIF-2B in isolated rat islets of Langerhans

Over short time periods glucose controls insulin biosynthesis predominantly through effects on preexisting mRNA. However, the mechanisms underlying the translational control of insulin synthesis are unknown. The present study was carried out to determine the effect of glucose on the activity and/or phosphorylation status of eukaryotic initiation and elongation factors in islets. Glucose was found to increase the activity of the guanine nucleotide-exchange factor eIF-2B over a rapid time course (within 15 min) and over the same range of glucose concentrations as those that stimulate insulin synthesis (3-20 mM). A nonmetabolizable analogue of glucose (mannoheptulose), which does not stimulate insulin synthesis, failed to activate eIF-2B. The best characterized mechanism for modulating eIF-2B activity involves changes in the phosphorylation of the alpha-subunit of its substrate eIF-2. However, in islets, no change in eIF-2 alpha phosphorylation was seen under conditions where eIF-2B activity was increased, implying that glucose regulates eIF-2B via an alternative pathway. Glucose also did not affect the phosphorylation states of three other regulatory translation factors. These are the cap-binding factor eIF-4E, 4E-binding protein-1, and elongation factor eEF-2, which do not therefore seem likely to be involved modulating the translation of the preproinsulin mRNA under these conditions.

Processing of pro-islet amyloid polypeptide (proIAPP) by the prohormone convertase PC2

Islet amyloid polypeptide (IAPP), 'amylin', is the component peptide of islet amyloid formed in Type 2 diabetes. IAPP is expressed in islet beta-cells and is derived from a larger precursor, proIAPP, by proteolysis. An in vitro translation/translocation system was used to separately examine processing of human proIAPP by the beta-cell endopeptidases PC2, PC3 or furin. ProIAPP was converted to mature IAPP by PC2 but there was little conversion by furin or PC3. These data are consistent with processing of proIAPP in beta-cell secretory granules. Abnormal cellular proteolysis associated with type 2 diabetes could contribute to IAPP amyloidosis.

A silencer and an adjacent positive element interact to modulate the activity of the human insulin promoter

A negative regulatory element (NRE) is located between positions -279 and -261 relative to the transcription start site of the human insulin gene. The NRE contains at least three distinct overlapping binding sites for several nuclear proteins. These proteins may be distinguished by their interaction with mutant variants of the NRE. Mutagenesis of two of these protein-binding sites within an insulin gene fragment containing sequences from position -361 to position +112 attenuated the negative activity of the NRE, confirming the importance of these sites in the function of the NRE. When placed in isolation upstream of the herpes-simplex-virus thymidine-kinase promoter, the NRE exhibited stimulatory activity in non-beta (BHK) and beta (HIT) cells. The positive activity within the NRE was mapped to a sequence that resembled the binding site for the ubiquitous factor Oct-1. These results indicated that the negative activity of the NRE was dependent on its interaction with other regulatory sites within the insulin gene. Thus, the NRE exhibited negative activity in transfected HIT cells when placed upstream of an insulin gene fragment (positions -261 to +112). However, its activity was modulated in a positive manner by inclusion of additional sequences from position -279 to -341. This region contains the CT3 box that binds the homeodomain protein IUF1 (insulin upstream factor 1). The NRE resembled a silencer, being at least partly independent of precise location and orientation, and being able to operate upon a variety of promoters. The role of the NRE is unclear, although it may be involved in restricting expression of the insulin gene to cells of the islets of Langerhans.

Identification and characterization of a functional retinoic acid/thyroid hormone-response element upstream of the human insulin gene enhancer

A deletion analysis of the human insulin gene extending to 2 kb upstream of the transcription start site provided evidence of regulatory sequences located upstream of the insulin-linked polymorphic region (ILPR). Within this ILPR-distal region is a sequence (Ink, for insulin kilobase upstream) which contains three potential nuclear hormone-receptor half-sites, closely matching the consensus sequence AGGTCA. These sequences are arranged as a palindromic element with zero spacing over-lapping a direct repeat with 2 bp spacing. The Ink sequence was used in electrophoretic mobility-shift assays within nuclear extracts from COS-7 cells overexpressing the vitamin D, thyroid hormone or retinoic acid receptors, or from an insulin-expressing hamster cell line, HIT-T15. These studies suggest that the insulin-expressing cell line contains thyroid hormone and retinoic acid receptors at least, and that these receptors are able to recognize the Ink sequence. Three copies of the Ink sequence were placed upstream of the thymidine kinase promoter and firefly luciferase reporter gene. In COS-7 cells expressing the appropriate nuclear hormone receptor, this construct was responsive to both thyroid hormone (18-fold) and all-trans-retinoic acid (31-fold). In HIT-T15 cells the same construct responded to all-trans-retinoic acid, but not to thyroid hormone. Within the context of a 2 kb insulin gene fragment, the Ink sequence was shown to be activated by retinoic acid and by the retinoic acid receptor, but acted as a negative element in the presence of both retinoic acid and the retinoic acid receptor. Mutagenesis studies demonstrated that the palindromic sequence was important for the retinoic acid response, and for binding of complexes containing retinoic acid receptor. In human islets of Langerhans, retinoic acid was shown to stimulate insulin mRNA levels. These results demonstrate that a functional nuclear hormone-receptor-response element is located upstream of the human ILPR. As retinoic acid and thyroid hormone are frequently involved in developmental regulatory processes, it is possible that this element may be important in the process of islet cell differentiation.

The insulin gene promoter. A simplified nomenclature

The tools of molecular biology have rapidly expanded our knowledge of how β-cells regulate insulin gene expression. As this work has progressed in parallel in different laboratories, alternate nomenclature systems have been developed to describe the functionally important elements of the insulin gene. This jumble of names is confusing to those outside the field and intimidating to neophytes. Therefore, we have agreed to a simple, uniform set of names for the major insulin gene promoter elements.

Differences between the catalytic properties of recombinant human PC2 and endogenous rat PC2

Human prohormone convertase PC2 was expressed in Xenopus oocytes and its properties were compared with those of the Type-2 endopeptidase of rat insulin secretory granules, previously identified as PC2 [Bennett, Bailyes, Nielson, Guest, Rutherford, Arden and Hutton (1992) J. Biol. Chem. 267, 15229-15236]. Recombinant PC2 had the same substrate specificity as the Type-2 endopeptidase, cleaving at the CA-junction (Lys64, Arg65) of human des-31,32-proinsulin to generate insulin; little activity was found toward human des-64,65-proinsulin or proinsulin itself. Recombinant PC2 was maximally active in 5-7 mM Ca2+ (K0.5 = 1.6 mM) whereas the Type-2 endopeptidase was maximally active in 0.5-1 mM Ca2+ (K0.5 = 40 microM). Both enzymes had a pH optimum of 5.0-5.5 but the Type-2 endopeptidase was active over a wider pH range. Two molecular forms of recombinant PC2 (71 kDa and 68 kDa) were found, both had an intact C-terminus but differed by the presence of the propeptide. The endogenous PC2 comprised several overlapping forms (size range 64-68 kDa), approximately two-thirds of which lacked C-terminal immunoreactivity. Part of the size difference between recombinant and endogenous PC2 was attributable to differences in N-glycosylation. The different post-translational proteolytic modifications of recombinant and endogenous PC2 did not account for the different pH and Ca2+ sensitivities shown by the enzymes. A modulating effect of carbohydrate on enzyme activity could not be excluded.

The insulin enhancer binding site 2 (IEB2; FAR) box of the insulin gene regulatory region binds at least three factors that can be distinguished by their DNA binding characteristics

Located at approximately 230 bp upstream from the transcription start site, the insulin enhancer binding site 2 (IEB2) or FAR region of the insulin gene is one of several important sequences involved in regulating transcription of the gene. The present study was undertaken to characterize the transcription factors binding at the IEB2/FAR region of the rat insulin II gene and to compare these with factors known to bind to the equivalent sequence in the rat I and human insulin genes. An endocrine-enriched factor, EFD3, was identified, which bound to the sequence CAGGAG. A second factor (D4) was identified as the widely expressed factor USF (upstream stimulating factor), while a third factor (D5) remained largely uncharacterized. The binding affinities of these three factors differed in the three genes, suggesting that the role of the IEB2/FAR sequence may vary subtly between the rat insulin II, rat insulin I and human insulin genes.

Differences in pH optima and calcium requirements for maturation of the prohormone convertases PC2 and PC3 indicates different intracellular locations for these events

PC2 and PC3, which is also known as PC1, are subtilisin-like proteases that are involved in the intracellular processing of prohormones and proneuropeptides. Both enzymes are synthesized as propolypeptides that undergo proteolytic maturation within the secretory pathway. An in vitro translation/translocation system from Xenopus egg extracts was used to investigate mechanisms in the maturation of pro-PC3 and pro-PC2. Pro-PC3 underwent rapid (t1/2 < 10 min) processing of the 88-kDa propolypeptide at the sequence RSKR83 to generate the 80-kDa active form of the enzyme. This processing was blocked when the active site aspartate was changed to asparagine, suggesting that an autocatalytic mechanism was involved. In this system, processing of pro-PC3 was optimal between pH 7.0 and 8.0 and was not dependent on additional calcium. These results are consistent with pro-PC3 maturation occurring at an early stage in the secretory pathway, possibly within the endoplasmic reticulum, where the pH would be close to neutral and the calcium concentration less than that observed in later compartments. Processing of pro-PC2 in the Xenopus egg extract was much slower than that of pro-PC3 (t1/2 = 8 h). It exhibited a pH optimum of 5.5-6.0 and was dependent on calcium (K0.5 = 2-4 mM). The enzymatic properties of pro-PC2 processing were similar to that of the mature enzyme. Further studies using mutant pro-PC2 constructs suggested that cleavage of pro-PC2 was catalyzed by the mature 68-kDa PC2 molecule. The results were consistent with pro-PC2 maturation occurring within a late compartment of the secretory pathway that contains a high calcium concentration and low pH.

Glucose modulates the binding activity of the beta-cell transcription factor IUF1 in a phosphorylation-dependent manner

In the human insulin gene, three regulatory sequences upstream of the transcription start site at -77 (the CT1 box), -210 (the CT2 box), and -315 (the CT3 box) bind a beta-cell-specific transcription factor, IUF1. Recent studies have mapped a glucose response element to a CT-like sequence in the rat insulin I gene. The present study was therefore undertaken to ascertain the role of IUF1 in glucose-stimulated insulin gene transcription. IUF1-binding activity was measured by electrophoretic mobility shift assay using the CT2 box as probe. When freshly isolated rat islets of Langerhans were incubated in medium containing low concentrations (3 mM) of glucose IUF1 activity fell to undetectable levels within 6 h. In high (20 mM) glucose IUF1 activity remained constant over a 24 h period. The loss of IUF1 activity was reversible. Thus when islets were incubated for 4 h in low glucose and transferred to high glucose, IUF1 levels recovered within 15 min. This effect was dependent on glucose metabolism as it was inhibited by mannoheptulose. Incubation of islets for 4 h in low concentrations of glucose supplemented with phosphatase inhibitors prevented the fall in IUF1 activity. No recovery in IUF1 activity was observed when islets were treated for 4 h with low glucose and then for a further 1 h with low glucose and dibutyryl cyclic AMP, or forskolin, or the phorbol ester phorbol 12-myristate 13-acetate. These results demonstrate that the IUF1-binding activity in islets of Langerhans is modulated by glucose in a phosphorylation-dependent manner, and that protein kinase A or protein kinase C are not involved. Finally, IUF1 was shown to be immunologically related to a recently cloned factor, IPF1, that binds to a CT-like sequence in the rat insulin I gene promoter.

Insulin-releasing pituitary cells as a model for somatic cell gene therapy in diabetes mellitus

Insulin delivery by somatic cell gene therapy was evaluated using murine pituitary AtT20MtIns-1.4 cells. These cells have been stably transfected to release human insulin by the introduction of a recombinant plasmid bearing a human preproinsulin cDNA under the control of a zinc-sensitive metallothionein promoter. 6 x 10(7) AtT20MtIns-1.4 cells were implanted subcutaneously into streptozotocin-diabetic mice immunosuppressed with cyclosporin A. Release of human insulin was assessed using a specific plasma human C-peptide assay. On days 1 and 2 after implantation human C-peptide concentrations were about 0.02 pmol/ml. Consumption of zinc sulphate solution (500 mg/l) as drinking fluid for days 3-5 increased plasma human C-peptide concentrations to 0.11 +/- 0.01 pmol/ml (mean +/- S.E.M.), n = 11, P < 0.01, and concentrations declined when zinc was discontinued. The extent of hyperglycaemia was slightly lower (P < 0.05) than in a group implanted with non-transfected AtT20 cells. The study was terminated after 9 days, and tumour-like aggregations of implanted cells were identified at autopsy. These comprised a large necrotic core with insulin-containing cells at the periphery. The study provides support for the view that somatic cell gene therapy offers a potential approach to insulin delivery in diabetes mellitus.

Calcium- and pH-dependent aggregation and membrane association of the precursor of the prohormone convertase PC2

PC2 is a subtilisin-like protease which is thought to be involved in the processing of prohormones and proneuropeptides in neuroendocrine cells. The mature 68-kDa enzyme is generated by intracellular proteolytic processing of a 75-kDa pro-PC2 polypeptide. Neuroendocrine cells contain at least two secretory pathways: the regulated pathway whereby secreted products are concentrated, stored in granules, and released in response to external stimulation of the cell, and the constitutive pathway, whereby secretory and plasma membrane proteins are continuously transported to the cell surface without prior concentration or storage. An important step in the sorting of proteins into these pathways is thought to involve the aggregation of proteins destined for storage granules. To define the mechanisms in the intracellular sorting of PC2 to secretory vesicles, the present study was undertaken to investigate the aggregation and membrane association properties of precursor and mature forms of PC2. Using material expressed in microinjected Xenopus oocytes, it was demonstrated that the 75-kDa pro-PC2 polypeptide undergoes calcium- and acid pH-dependent aggregation. Calcium exhibited an effect on the aggregation of pro-PC2 at pH 7.0 and 6.5, whereas below 6.5 aggregation was independent of calcium. Association of pro-PC2 with membranes was observed at pH 5.5, but not at pH 7.0, 6.5, nor 6.0. The mature 68-kDa PC2 polypeptide remained soluble under conditions that caused aggregation and membrane association of the 75-kDa propolypeptide. Deletion of COOH-terminal sequences had no effect on the association of pro-PC2 with membranes, whereas a peptide corresponding to amino acids 57-85 of the propeptide was able to partially compete the membrane associated pro-PC2 away from the membranes.

Synthesis and processing in vivo of the novel mouse thyrotropin beta-presubunit that contains an NH2-terminal extension sequence

Expression of the single mouse TSH beta gene gives rise to multiple mRNAs, and we have previously shown that in vitro, one of these mRNAs gives rise to a novel TSH beta-presubunit due to initiation of translation at an in-frame start site unique to this mRNA which is up-stream of the normal start site. The novel presubunit contains a 17-amino acid NH2-terminal extension sequence compared to the normal presubunit. Although this extension sequence does not have the characteristics of a normal signal sequence, the novel TSH beta-presubunit was processed in vitro by microsomal membranes. In this study we have examined the translation product of this mRNA in intact cells and whether in vivo it gives rise to a processed secreted TSH beta-subunit that has an NH2-terminal sequence different from that of the established TSH beta-subunit. Firstly, mRNAs encoding alpha-presubunit and either the normal or novel TSH beta-presubunit were microinjected into Xenopus oocytes, and it was found that immunoprecipitable TSH dimer was secreted into the medium regardless of the mRNA used for TSH beta-subunit synthesis. However, less TSH was obtained when the TSH beta-subunit was derived from the extended TSH beta-presubunit. Secondly, when COS cells were transiently transfected with plasmids expressing alpha-presubunit and either the normal or novel TSH beta-presubunit, secreted TSH was obtained when the TSH beta-subunit was derived from either presubunit. TSH dimer was also obtained when the TSH beta-presubunit was derived from a mRNA encoding the extended presubunit in which the down-stream AUG had been eliminated by site-specific mutagenesis. This demonstrated that the up-stream translation start site was used in the intact cell and that secreted TSH beta-subunit was derived from the extended presubunit and not from normal presubunit resulting from translational readthrough to the down-stream AUG. When secreted TSH beta-subunits derived from the normal and extended TSH beta-presubunits were digested with endoproteinase LysC, the NH2-terminal fragments were similar in size, suggesting that the NH2-terminal extension had little if any effect on the site of cleavage by signal peptidase. Our data, therefore, demonstrate that the longer TSH beta-presubunit is synthesized in vivo and strongly suggest that it is processed in the intact cell to give a mature secreted TSH beta-subunit indistinguishable from that derived from the normal TSH beta-presubunit.

Autocatalytic maturation of the prohormone convertase PC2

PC2 is a member of the eukaryotic family of subtilisin-like proteases, which is thought to participate in the processing of prohormones and proneuropeptides in neuroendocrine cells. PC2 is synthesized as a 69-kDa prepropolypeptide. The NH2-terminal signal sequence is removed during segregation within the endoplasmic reticulum, where glycosylation occurs to generate a 75-kDa propolypeptide. A combination of site-directed mutagenesis and a cell-free translation/translocation system from Xenopus eggs was used to investigate the processing of the pro-PC2 precursor. The 75-kDa polypeptide underwent slow cleavage after the sequence Arg-Lys-Lys-Arg84 to generate a 68-kDa mature enzyme. Cleavage was blocked when the tetrabasic sequence was deleted (PC2M3) or when the active site Asp142 was changed to Asn (PC2M4). This latter observation suggested that cleavage of the 75-kDa propolypeptide to the mature 68-kDa enzyme was autocatalytic. Incubation of the PC2M4 mutant with the wild type PC2 precursor resulted in cleavage of both the wild type polypeptide and the catalytically inactive PC2M4 mutant. This indicates that cleavage could occur through an intermolecular reaction. The results also demonstrate that the novel Xenopus egg extract translation/translocation system represents a powerful cell-free method for studying proteolytic processing of propolypeptides.

Nutrient regulation of insulin gene expression

The beta cell of the islets of Langerhans contributes along with other factors to glucose homeostasis by sensing changes in the plasma glucose concentrations and adjusting the rate of insulin production and release. Over short periods of time, insulin production is controlled principally through translation of pre-existing mRNA. Over longer periods, insulin mRNA levels are modulated through effects on the rate of transcription of the insulin gene, and also through changes in the rate of decay of insulin mRNA. These long-term effects may be important in allowing the beta cell to adapt to changes in diet or periods of fasting. Several mechanisms involved in the control of the rate of translation of insulin mRNA have been described. Effects of glucose metabolism on the turnover of insulin mRNA have yet to be characterized in detail. At the level of transcription, cis-acting DNA elements and trans-acting factors involved in the transient response of the insulin gene to changes in intracellular cAMP levels, or to signals generated as a result of glucose metabolism, have been identified.

Negative regulation of transcription in eukaryotes

Images

Insulin delivery by somatic cell gene therapy

The feasibility of somatic cell gene therapy as a method of insulin delivery has been studied in mice. Murine pituitary AtT20 cells were transfected with a human preproinsulin DNA in a plasmid containing a metallothionein promoter and a gene conferring resistance to the antibiotic G418. The AtT20MtIns-1.4 clone of cells was selected because of its higher insulin-releasing activity compared with other clones. After culturing for 24 h in Dulbecco's medium containing 10 mM glucose, the AtT20MtIns-1.4 cells released human insulin at about 5 ng/10(6) cells per 24 h. Insulin release was not significantly altered by raised concentrations of glucose, potassium or calcium, but insulin release was increased by 20 mM arginine, 5 mM isomethylbutylxanthine and 90 microM zinc. AtT20MtIns-1.4 cells (2 x 10(6)) were implanted intraperitoneally into non-diabetic athymic nude (nu/nu) mice, and the mice were made diabetic by injection of streptozotocin after 7 days. Release of human insulin in vivo was assessed using a specific plasma human C-peptide assay. Human C-peptide concentrations were maintained at about 0.1 pmol/ml throughout the 29 days of the study. The development of streptozotocin-induced hyperglycaemia was delayed in recipients of the cells releasing human insulin, compared with a control group receiving an implant of non-transfected cells. At autopsy the implanted AtT20MtIns-1.4 cells in each recipient had formed a tumour-like aggregation, with an outer region of insulin-containing cells.(ABSTRACT TRUNCATED AT 250 WORDS)

The helix-loop-helix transcription factor USF (upstream stimulating factor) binds to a regulatory sequence of the human insulin gene enhancer

Two important sequence elements, designated insulin enhancer binding site 1 (IEB1) or NIR and IEB2 or FAR, are involved in regulating expression of the rat insulin I gene. These elements bind a helix-loop-helix transcription factor, insulin enhancer factor 1 (IEF1). The IEB1 site is highly conserved among insulin genes but the IEB2 site is not conserved. To investigate the factors binding at the equivalent IEB1 and IEB2 sites in the human insulin gene enhancer, electrophoretic mobility shift assays were performed using a variety of cell extracts and probes specific for the homologous IEB1 and IEB2 sites. The results indicate that a factor with similar tissue distribution and binding characteristics to those of IEF1 binds to the IEB1 site in the human insulin gene, but that a separate factor, identified as the adenovirus major late transcription factor [MLTF, or upstream stimulating factor (USF)] binds to the IEB2 site.

Human insulin gene enhancer-binding proteins in pancreatic alpha and beta cell lines

Electrophoretic mobility shift assays were performed using oligonucleotides corresponding to known protein binding sites within the human insulin gene enhancer and nuclear extracts from mouse pancreatic alpha and beta cell lines. The results demonstrate that a previously described factor, IUF-1, binds to three sites at -82 (the CT1 box), -215 (the CT2 box), and -319 (the CT3 box) in the human insulin gene enhancer. IUF-1 was present only in beta but not in alpha cells, while all other DNA-binding proteins were present in both cell lines. IUF-1 may therefore be an important determinant of insulin gene beta cell-specific expression.

The human insulin gene-linked polymorphic region adopts a G-quartet structure in chromatin assembled in vitro

The insulin gene-linked polymorphic region (ILPR), located 363 bp upstream of the human insulin gene, is composed of tandem repeats of the consensus sequence ACAGGGGT(G/C)(T/C)GGGG. It has previously been shown that an insulin gene fragment containing the ILPR adopts an altered DNA structure in vitro. Furthermore, oligonucleotides containing the consensus repeat sequence exhibit multiple quadriplex DNA structures. The present study was undertaken to determine whether such altered DNA structures existed within the ILPR when the insulin gene was assembled into chromatin in vitro. Chromatin assembly was achieved using histones and an extract from unfertilized eggs from Xenopus laevis. The presence of altered DNA conformations within the 5' region of the human insulin gene was investigated using the structural probe nuclease P1. Nuclease P1 recognized multiple distinct sites in the 5' flanking region of the human insulin gene in naked DNA. Most of these sites disappeared when the recombinant plasmid DNA was treated with histones and unfertilized egg extract. In the assembled DNA, the ILPR appeared as the major site of nuclease P1 hypersensitivity. Fine-mapping of the multiple reactive sites within the ILPR showed a pattern characteristic of G-quartet foldback structures similar to those that have been observed for telomeric DNA.

Analysis of DNA structure in the human insulin gene-linked polymorphic region in vivo

An altered DNA structure exists within the hypervariable region located 360 bp upstream of the human insulin gene. The aim of the present study was to determine whether this structure exists in the insulin gene in vivo, and whether its presence is related to the expression of the insulin gene. However, since there were no clonal human beta-cell lines available for such studies, the human insulin gene was transfected into a rat insulinoma-derived beta-cell line and several human insulin-expressing clones were selected. One such cell line was treated in vivo with the DNA structural probe bromoacetaldehyde and the chromosomal DNA was extracted. Following digestion with TaqI and subsequent digestion with S1-nuclease to cleave at the bromoacetaldehyde-reactive sites, the DNA was subjected to agarose gel electrophoresis, and insulin gene fragments were detected by Southern blot analysis. Bromoacetaldehyde generated subfragments of 2500, 1700 and 800 bp in the human insulin gene isolated from the rat beta-cell line, while the human insulin gene in the non-expressing HeLa cell line was unreactive to bromoacetaldehyde. These results suggest that an altered structure might exist in the insulin gene-linked polymorphic region of the human insulin gene in vivo, and that this structure may play a role in the expression of the insulin gene.