Glycosylation in human thyroglobulin: location of the N-linked oligosaccharide units and comparison with bovine thyroglobulin.

The amino acid sequence established for human thyroglobulin (hTG) from its cDNA sequence contains 20 putative N-linked glycosylation sites. We have characterized the glycopeptides contained in a tryptic digest of hTG in order to determine which sites are actually linked to carbohydrate. In addition, the distribution of oligosaccharide type(s) at these confirmed sites of N-linked glycosylation has been examined. Glycopeptides were purified using gel permeation chromatography followed by several steps of HPLC. The purified tryptic glycopeptides were characterized by gas phase sequencing and carbohydrate analysis and located within the amino acid sequence of thyroglobulin. Each of the recovered glycopeptides contained a consensus sequence for N-linked glycosylation. Of the 20 putative N-linked glycosylation sites in the human thyroglobulin polypeptide chain, 16 were shown to be actually glycosylated in the mature protein. Eight of these confirmed glycosylation sites (at positions 57, 465, 510, 729, 797, 1696, 1754, and 2230) appear to be linked to complex-type oligosaccharide units containing fucose and galactose in addition to mannose and glucosamine. Five sites (at positions 1200, 1329, 1993, 2275, and 2562) contain high mannose type units and two sites (at positions 179 and 1345) are linked to oligosaccharide units containing galactose in addition to mannose and glucosamine but no fucose and may be either hybrid or complex structures. In addition, position 928 was found to be degenerate in oligosaccharide structure and very different oligosaccharide composition types were found associated with peptides containing the same amino acid sequence. A high probability of a beta turn which would include the glycosylated asparagine residue was predicted for the amino acid sequence found at 13 of the 16 sites. The glycosylation pattern in hTG was also compared with the data recently reported for bovine thyroglobulin (bTG) (27) and as has been recently reported for bTG, no oligosaccharides of the high mannose type were found in the N-terminal portion of hTG. Only four of the 20 putative sites the sequence of hTG, at asparagine residues 91, 477, 1849, and 2102 were not represented in the purified glycopeptide population and are presumed to escape significant glycosylation.

Characterization of hormonogenic sites in an N-terminal, cyanogen bromide fragment of human thyroglobulin.

We have confirmed the observation of Marriq et al. (FEBS Lett. 207, 302-306, 1986) that substantial thyroxine formation occurs on enzymatic iodination of a 171-residue, N-terminal cyanogen bromide (CNBr) fragment of human goiter thyroglobulin. Marriq et al. concluded from their studies that Tyr130 is the donor for the preferential hormonogenic acceptor site Tyr5. However, in the present study we observed that thyroxine formation in the CNBr fragment occurred not at Tyr5, but rather at Tyr130. Moreover, we did not confirm the report of Marriq et al. that formation of thyroxine in the iodinated CNBr fragment involves cleavage of the Val129-Tyr130 bond. Our finding of thyroid hormone at Tyr130 was unexpected, as Tyr5 is known to be the major site of hormone formation in thyroglobulin isolated from the thyroids of humans and other mammals. In the CNBr fragment, however, Tyr5 appeared to act as a donor rather than an acceptor. Thus, while we have confirmed the finding that the N-terminal CNBr fragment of human goiter thyroglobulin is by itself capable of efficient thyroxine formation when enzymatically iodinated in vitro, we observed that the role of specific tyrosines in the hormone-forming process was different from that in intact thyroglobulin. We question, therefore, whether the CNBr fragment is a valid model for defining the role and location of hormonogenic sites in intact thyroglobulin.

Thyroglobulin glycosylation: location and nature of the N-linked oligosaccharide units in bovine thyroglobulin.

The cDNA sequence of bovine thyroglobulin (bTg) predicts 14 putative N-linked glycosylation sites. We have characterized the glycopeptides contained in a tryptic digest of bovine thyroglobulin in order to establish actual glycosylation sites. The distribution of oligosaccharide types within the known sites of N-linked glycosylation has also been examined. Each glycopeptide was subjected to neutral and amino sugar analysis, as well as sialic acid analysis. Thirteen of the 14 putative N-linked glycosylation sites in bTg were confirmed as glycopeptides in the mature protein. Nine of these confirmed glycosylation sites appear to bear complex or hybrid type oligosaccharide units and four contain oligosaccharide structures in which only mannose and glucosamine were seen. Complex or hybrid type glycosylation was confirmed at asparagine residues 91, 464, 476, 834, 928, 1121, 1757, 1851, and 2232, while oligosaccharides containing only mannose and glucosamine were found at asparagine residues 1346, 1995, 2104, and 2277. Analysis of the amino acid sequence in the region of each putative glycosylation site predicts a high probability of a beta turn at 10 of the 13 sites. While glycosylation was distributed through most of the length of the thyroglobulin sequence, no oligosaccharides containing only mannose and glucosamine were found in the N-terminal half of the molecule. Several of the glycosylated peptides showed microheterogeneity and occurred in two or more discrete peaks on HPLC. A single putative site at asparagine residue 179 was not recovered in the purified glycopeptide population.

A neuropeptide Y (NPY)-related peptide is present in the river lamprey CNS.

Pancreatic polypeptide (PP)-like material from brain+spinal cord, and retina extracts of Lampetra fluviatilis was studied by HPLC and RIA. The brain+spinal cord extract showed a complex elution profile with multiple peaks of immunoreactivity. The retina extract showed a much simpler pattern with a single significant peak along with a trace of a second peak corresponding to the latest and penultimate peaks in the brain extract. Twenty-one out of 36 residues could be sequenced from the latest eluting peak in the brain extract. This sequence showed 81% identity with porcine neuropeptide (NPY) suggesting that both the brain/spinal cord and retina of the river lamprey contain a peptide homologous to NPY.

Pancreatic islets of obese hyperglycemic mice (ob/ob).

The development of the obesity-associated hyperglycemic syndrome in ob/ob mice, genetically determined, was observed over time by a combined functional and structural study of pancreatic islets. Islet areas increased with advancing age in ob/ob mice from 2 times at 1 month to 30 times at 6 months of age the size of lean mouse islets. Islet areas apparently increased more than pancreatic insulin content in ob/ob mice. Glucose and insulin tolerance tests were performed to study in vivo responses to glucose and insulin, respectively, in 1-, 3-, and 6-month-old mice. With ob/ob mice, glucose tolerance tests revealed more elevation of plasma glucose than in lean mice, the lean mice revealing more elevated plasma insulin than the obese mice. In insulin tolerance tests, lean mice presented marked hypoglycemia, whereas ob/ob mice revealed slightly higher plasma glucose at 1 month of age but three to four times higher amounts than that of lean mice at 6 months of age. Thus, increasing insulin resistance in ob/ob mice older than 3 months is associated with progressively increasing islet area, which contains proportionally less insulin than that of lean mouse pancreas. The data suggest that insulin resistance in ob/ob mice progressively develops up to 6 months of age and that marked islet hyperplasia is likely in response to sustained hyperglycemia, leading to hyperinsulinemia and eventual marked obesity.

Structure of a porcine relaxin inactive on the mouse pubic ligament.

In addition to B31 (CM-a) and B28 (CM-B) relaxins, acid-acetone extracts of ovaries of pregnant sows contain a third major relaxin species (relaxin C). The major components of relaxin C possess about half the activity of CM-a or CM-B in the guinea pig palpation assay, but are completely inactive in the mouse pubic ligament assay. Its uterotrophic and protein anabolic effects in ovariectomized, estrogen-primed mice, however, are comparable to those of CM-B. Sequence analysis indicates that the two major components of relaxin C, like the other porcine relaxins, consist of two polypeptide chains linked by disulfide bonds. The shorter (A) chains are identical to the A chains of the other porcine relaxins, except for the absence of the N-terminal arginine residue. The B chains display microheterogeneity; the B sequences of the two predominant species are the same as those of the other porcine relaxins through B25, but terminate at valine residue B25 or serine residue B26, respectively.

Isolation and characterization of a variant somatostatin-14 and two related somatostatins of 34 and 37 residues from lamprey (Petromyzon marinus).

Three major forms of somatostatin were isolated from pancreas of the lamprey (Petromyzon marinus). One of the two major forms is a 14-residue somatostatin (SS-14) having the sequence AGCKNFFWKTFSSC. The homologous substitution Ser for Thr in position 12 is the first example of SS-14 from vertebrate preprosomatostatin gene I having a divergent sequence. The longest form is 37 residues in length (SS-37) and has the sequence ALRAAAVAGSPQQLLPLGQRERKAGCKNFFWKTFSSC. A 34-residue form (SS-34) identical in sequence but truncated at a single Arg residue at position 3 of SS-37 was also isolated. The yields of the three forms were SS-37 (0.43 nmol/g), SS-34 (134 nmol/g), and SS-14 (51.5 nmol/g). The identification of this nested series of somatostatins suggests that prosomatostatin processing in lamprey more closely resembles that observed for procholecystokinin than that of mammalian or other piscine prosomatostatins. Somatostatin-producing cells in the lamprey pancreas were identified by immunostaining using antiserum against SS-34 and anti-serum against mammalian SS-14.

Isolation of peptide hormones from the pancreas of the bullfrog (Rana catesbeiana). Amino acid sequences of pancreatic polypeptide, oxyntomodulin, and two glucagon-like peptides.

Insulin, pancreatic polypeptide, glucagon, oxyntomodulin, and two distinct glucagon-like peptides were isolated from acidic ethanol extracts of bullfrog pancreas by gel filtration followed by high pressure liquid chromatography. The amino acid sequences of pancreatic polypeptide, oxyntomodulin, and both glucagon-like peptides were determined. Frog pancreatic polypeptide contains 36 amino acid residues and has a COOH-terminal phenylalaninamide. It is more homologous with human pancreatic polypeptide (61%) than other characterized members of this family of peptides. Frog glucagon has an amino acid composition identical to the NH2-terminal 29 residues of the larger, more abundant oxyntomodulin and was not sequenced. The finding of a single form of glucagon and oxyntomodulin, but two glucagon-like peptides in frog pancreas extract is similar to that found or deduced for mammals.

Endocrine pancreas in the rat model of exocrine pancreatic insufficiency.

A new technique to obstruct the pancreatic ducts was developed by injecting zein solution into the common bile duct of the rat. Four weeks after the injection, the fate of the endocrine pancreas was investigated by studying (a) pancreatic content of four pancreatic hormones, (b) histology and immunohistochemistry of the pancreas, (c) i.v. glucose tolerance and i.v. insulin tolerance tests for monitoring plasma glucose, insulin, and pancreatic polypeptide (PP) levels in vivo, and (d) in vitro perifusion of pancreatic tissue slices to assess insulin and PP secretion. In zein-injected rats, the total pancreatic content of insulin, glucagon, PP, and somatostatin was reduced to 80, 70, 40, and 20%, respectively, of the corresponding controls. In response to insulin-induced hypoglycemia, the plasma PP levels rose to about one-half that of the controls. By contrast, perifused zein-injected rat pancreases released several times more PP than the controls in response to carbachol stimulation. In zein-injected rats, total pancreatic protein was reduced to 20% of the controls and pancreatic amylase was almost absent, reflecting practically complete loss of acinar tissue. Thus, this experimental model appears to be suitable for producing chronic pancreatic insufficiency in the rat and provides a useful model for studying both endocrine and exocrine functions in the small rodent.

Isolation of alligator gar (Lepisosteus spatula) glucagon, oxyntomodulin, and glucagon-like peptide: amino acid sequences of oxyntomodulin and glucagon-like peptide.

Oxyntomodulin, glucagon, and a glucagon-like peptide (GLP) have been isolated from the endocrine pancreas of the alligator gar (Lepisosteus spatula), a ganoid fish. The three peptides were isolated by gel filtration and HPLC and were identified by size, composition, and glucagon-like immunoreactivity. The amino acid sequences of the oxyntomodulin and GLP were determined. The oxyntomodulin contains 36 amino acid residues and its sequence is H S Q G T F T N D Y S K Y L D T R R A Q D F V Q W L M S T K R S G G I T. The composition of the glucagon is identical to the N-terminal 29 residues of the gar oxyntomodulin. The single form of GLP found contains 34 amino acid residues in the following sequence: H A D G T Y T S D V S S Y L Q D Q A A K K F V T W L K Q G Q D R R E. These findings suggest that all three peptides are derived from a common precursor.

Isolation and structure of lamprey (Petromyzon marinus) insulin.

Insulin has been purified to homogeneity from the caudal and cranial pancreas of the adult sea lamprey (Petromyzon marinus). The final yield was 18.6 nmol/g (127.8 micrograms/g). The structures of both A- and B-chains have been determined using amino acid analyses, gas-phase sequence analyses, and proteolytic mapping by fast atom bombardment-mass spectrometry. The sequence of the A-chain was found to be GIVEQCCHRKCSIYDMENYCN. The sequence of the B-chain, extended at the amino terminus, was determined to be SALT-GAGGTHLCGSHLVEALYVVCGDRGFFYTPSKT. Lamprey insulin retains the common features of vertebrate insulins. Sea lamprey insulin has no more homology to hagfish (Myxine glutinosa) insulin than it has to the teleost fish or to mammalian insulins.

Influence of avian pancreatic polypeptide on pancreatic and biliary secretion in laying hens.

White Leghorn hens, 14 to 29 wk of age, were surgically prepared with cannulae for collecting secretions from the cystic duct and the duct draining the ventral pancreatic lobe and for infusing the jugular vein with avian pancreatic polypeptide (aPP) or saline. A plasma infusion rate that produced a plasma level of 15 ng of aPP/mL was used. A comparison of values obtained during saline infusion with those obtained during aPP infusion indicated that pancreatic and biliary secretory volumes and pancreatic total protein concentration were significantly depressed by aPP. The pH of pancreatic and biliary secretions were not significantly affected by aPP. Because aPP also depresses gastric secretion and motility in hens, it is proposed that its physiological role may be to oppose or modulate the actions of other, stimulatory gastrointestinal hormones.

Isolation and structures of alligator gar (Lepisosteus spatula) insulin and pancreatic polypeptide.

Insulin and a 36-residue peptide with homology to pancreatic polypeptide (PP) were isolated from the endocrine pancreas of the alligator gar (Lepisosteus spatula), a ganoid fish, by gel filtration and HPLC. Heterologous radioimmunoassays were used to detect insulin-like and PP-like immunoreactivities during purification of the two peptides. The sequence of the 36-amino acid peptide containing a C-terminal tyrosinamide was identical at 31 of 36 positions to porcine neuropeptide Y (NPY). The amino acid sequence of this peptide is YPPKPENPGEDAPPEELAKYYSALRHYINLITRQRY-NH2. The second peptide, gar insulin, contains 52 amino acid residues and is composed of a 21-residue A chain and a 31-residue B chain. The sequence of the A chain is GIVEQCCHKPCTIYELENYCN. The sequence of the B chain is AANQHLCGSHLVEALYLVCGEKGFFYNPNKV.

Structure of a peptide from coho salmon endocrine pancreas with homology to neuropeptide Y.

The amino acid sequence of a peptide isolated from the Pacific salmon (Oncorhynchus kisutch) endocrine pancreas has been determined. This simple 36 residue peptide is a member of the pancreatic polypeptide family. It contains a C-terminal tyrosinamide and is more homologous with porcine neuropeptide Y (NPY) (83%) and peptide YY (75%) than any of the previously characterized pancreatic polypeptides (PP). This peptide appears to be the major but not the only representative of this family of peptides present in the endocrine pancreas of this fish. This peptide is referred to as salmon pancreatic polypeptide (salmon PP).

Characterization of coho salmon (Oncorhynchus kisutch) islet somatostatins.

Three different somatostatins have been isolated from the pancreatic islet tissue of the coho salmon (Oncorhynchus kisutch) by gel filtration and HPLC. Two of these peptides contain 14 amino acids and the larger third peptide consists of 25 amino acids. The sequence of the salmon SST-25 is Ser-Val-Asp-Asn-Leu-Pro-Pro-Arg-Glu-Arg-Lys-Ala-Gly -Cys-Lys-Asn-Phe-Tyr-Trp-Lys-Gly-Phe-Thr-Ser-Cys. The sequence of the salmon SST-14-I is Ala-Gly-Cys-Lys-Asn-Phe-Phe-Trp-Lys-Thr-Phe-Thr-Ser-Cys. The other small somatostatin (SST-14-II) which was not sequenced has an amino acid composition identical to the C-terminal 14 amino acids of the SST-25 and it is probably derived from this larger form. Evidence for low levels of a somatostatin containing 28 amino acids is also presented. This SST-28 appears to be an N-terminal extended precursor of SST-25 or a peptide derived via alternative processing of a common preprosomatostatin. Injected into juvenile salmon, SST-25 caused a decline in circulating levels of plasma insulin, depletion of liver glycogen, and activation of lipolytic pathways. Juvenile salmon treated with anti-SST-25 serum revealed elevated levels of plasma insulin as well as an increase of the glycogen content of the liver.

Isolation and structures of coho salmon (Oncorhynchus kisutch) glucagon and glucagon-like peptide.

Glucagon and glucagon-like peptide (GLP) containing 31 amino acids have been isolated from the principal islet of coho salmon (Oncorhynchus kisutch) by gel filtration of acid alcohol extracts followed by HPLC, and the complete amino acid sequence of both peptides has been determined. Salmon glucagon is a simple 29 residue peptide differing at 3 positions when compared to catfish glucagon and at 8 positions when compared to porcine glucagon. Salmon GLP differs at 6 positions when compared with the N-terminal 31 amino acids of the 34 amino acid catfish GLP. Both coho salmon glucagon and GLP cross-react weakly in our mammalian glucagon radioimmunoassay and therefore this technique could not be used to determine tissue content. Glucagon and GLP isolated amounted to 156 micrograms/g and 350 micrograms/g wet tissue, respectively.

Influence of cholecystokinin, vasoactive intestinal peptide and secretin on plasma concentration of avian pancreatic polypeptide in laying hens.

The influence of saline infusion (i.v.) followed by infusion of either cholecystokinin, vasoactive intestinal peptide, or secretin on plasma concentration of avian pancreatic polypeptide (aPP) was studied in sixteen 18-26-week old Single Comb White Leghorn hens. Three concentrations were used for each hormone. Blood was drawn after both saline and hormone infusions and assayed for aPP content. No significant influence of any of the three hormones on plasma aPP level was found in either fed or fasted hens.

Pancreatic polypeptide in islet cell tumors. Morphologic and functional correlations.

Twelve islet cell tumors and one islet cell hyperplasia were studied with immunocytochemical and radioimmunoassay methods. With immunocytochemical staining, all six insulinomas, one mixed insulinoma-glucagonoma, and four gastrinomas were positive for insulin, insulin and glucagon, and gastrin, respectively. Pancreatic polypeptide (PP) was positive in three insulinomas and one mixed insulinoma-glucagonoma. All of the tumors were positive for neuron-specific enolase (NSE). Radioimmunoassays of tissue extracts further disclosed that all functioning tumors contained more than one pancreatic hormone. PP concentrations of two insulinomas and one mixed insulinoma-glucagonoma were higher than that of normal control pancreases. A study of protein meal-stimulated PP secretion revealed that three of the insulinoma cases and two gastrinoma cases exhibited higher plasma PP levels than the age-matched controls. The findings suggest that: both functioning and nonfunctioning islet cell tumors derive from neuroendocrine cells positive for NSE; all functioning islet cell tumors appear to contain PP in the tumor tissue as a minor component; as many as 70% of the patients with islet cell tumors present with abnormally higher plasma PP levels after protein meals; and a study of meal-stimulated PP secretion may well be used as a marker for the presence of functional islet cell tumors.

The influence of avian pancreatic polypeptide on gastric secretion and motility in laying hens.

The effects of pancreatic polypeptide (PP) on digestive functions have been studied extensively in mammals. The only two previous studies on the effects of PP on avian (aPP) gastrointestinal function employed bolus injections (i.v.) of the hormone, which may have resulted in nonphysiological plasma levels of aPP. The present study was to determine the effect of aPP infused (iv) at levels simulating postprandial plasma concentrations. Young laying hens were prepared with chronic proventricular and jugular cannulas and with strain gauge transducers sutured to the serosa of the gizzard. In each experiment, a control saline infusion was followed by infusion sufficient to achieve either 7.5 or 15 ng aPP/ml of plasma. Volume, pH, and pepsin content of gastric secretions and frequency and amplitude of gastric contractions were determined. At 15 ng aPP/ml of plasma, secretory volume (P less than .01) and contractile frequency (P less than .001) were significantly depressed as compared to similar data during the saline infusion. At 7.5 ng aPP/ml of plasma, only contractile frequency (P less than .01) was depressed.

Characterization of coho salmon (Oncorhynchus kisutch) insulin.

Insulin has been isolated from islet tissue of coho salmon (Oncorhynchus kisutch) by gel filtration and HPLC and the complete amino acid sequence has been determined. The sequence differs from bovine insulin at 14 sites but all interchanges are conservative from the viewpoint of preservation of conformation. A comparison of insulin sequences from other fish is presented. Salmon insulin cross-reacts very weakly with antiserum to bovine insulin and vice versa. A completely homologous radioimmunoassay has been developed and used to estimate the insulin in salmon islet tissue and in plasma. The hypoglycemic effect of salmon insulin in salmon was more pronounced and persisted longer than that caused by identical doses of bovine insulin.