Glycosylation of thyroid-stimulating hormone in pituitary tumor cells: influence of high mannose oligosaccharide units on subunit aggregation, combination, and intracellular degradation.

We have studied the glycosylation of TSH in cell culture and have examined the influence of carbohydrate on subunit aggregation, intracellular degradation, and combination. Dispersed mouse thyrotropic tumor cells were labeled by pulse-chase methods with [35S]methionine and various 3H-labeled carbohydrates; cell lysates and media were precipitated with antisera to TSH alpha and TSH beta, and the products were analyzed by sodium dodecyl sulfate gradient gel electrophoresis without or with preexposure to Endoglycosidase (Endo) H. At early pulses, both intracellular alpha and beta were mainly composed of one Endo H-sensitive (high mannose) carbohydrate unit and a small amount of nonglycosylated forms; alpha only had the posttranslational addition of a second high mannose unit. With increasing chase times up to 18 h, intracellular subunits showed a slow but progressive increase in Endo H-resistant (complex) forms, and media subunits were completely resistant. Preincubation of cells with tunicamycin caused production of nonglycosylated subunits that showed a high degree of aggregation, especially after heating at 37 C under nonreducing conditions. Unlike glycosylated subunits, which were not degraded, nonglycosylated subunits were 50-65% degraded intracellularly before secretion; the degradation caused by tunicamycin was specific for TSH subunits and not noted for other 35S-labeled proteins. Incubation of various 35S-labeled alpha forms with excess unlabeled TSH beta showed high combining activity for intracellular alpha with two high mannose units, intermediate activity for media alpha with two complex units, and low activity for intracellular alpha with one high mannose unit or nonglycosylated media alpha. These data suggest that the initial glycosylation with high mannose carbohydrate units prevents intracellular aggregation and degradation of TSH subunits and enhances attainment of the conformation necessary for alpha- and beta-subunit combination.

Mitogen-activated protein kinase and phosphatidylinositol 3-kinase pathways are not sufficient for insulin-like growth factor I-induced mitogenesis and tumorigenesis.

Insulin-like growth factor-I (IGF-I) and insulin are known to activate a signaling cascade involving ras --> kappa raf-1 --> mitogen-activated protein (MAP) kinase kinase (MEK) --> p42/p44 MAP kinase (Erk-1 and -2). Recent reports suggest that activation of this ras/MAP kinase pathway is involved in mitogenesis and c-fos transcription but is not required for insulin action on metabolic processes such as glycogen synthesis, lipogenesis, and GLUT-4-mediated glucose transport. Previously we and others have demonstrated that substitution of both tyrosines at positions 1250 and 1251 in the carboxy-terminal region of the human IGF-I receptor has relatively small effects on receptor and endogenous substrate phosphorylation but completely abrogated the ability of these cells to form tumors in nude mice or proliferate in response to IGF-I in culture. Replacement of the tyrosine at position 1316 also did not affect the kinase activity of the receptor with respect to autophosphorylation or phosphorylation of endogenous substrates but did reduce the ability of the receptor to mediate mitogenic or tumorigenic signals. To further characterize the role of these tyrosines in IGF-I receptor function, we have used three distinct approaches to examine the ras/MAP kinase pathway in IGF-I-induced mitogenesis and tumorigenesis in NIH-3T3 cells overexpressing wild-type and mutated IGF-I receptors: 1) tyrosine phosphorylation of the MAP kinases Erk-1 and -2; 2), mobility shifts indicative of MAP kinase phosphorylation; and 3) in vitro MAP kinase activation. We have also examined IGF-I-induced phosphatidylinositol (PI) 3-kinase activation in the same cell lines. By each method we show that the IGF-I-induced MAP kinase phosphorylation/activation and PI 3-kinase activation, are not different between cells overexpressing wild-type IGF-I receptors and cells carrying IGF-I receptors having tyrosine motifs replaced at positions 1250 and 1251. We conclude that mitogenic and tumorigenic signals involving tyrosine residues in the C-terminal domain of the IGF-I-receptor include pathways other than the MAP kinase and PI 3-kinase pathways.

Characterization and charge distribution of the asparagine-linked oligosaccharides on secreted mouse thyrotropin and free alpha-subunits.

Mouse hemipituitaries in vitro secrete TSH, composed of an alpha-beta heterodimer, as well as excess (free) alpha-subunits. By dual metabolic labeling with [35S]sulfate and [3H]mannose, we have characterized oligosaccharides from secreted TSH alpha, TSH beta, and free alpha-subunits released from the apoprotein by enzymatic deglycosylation. Oligosaccharides from each subunit displayed a distinct anion exchange HPLC profile due to a specific pattern of sialylation and sulfation. Six species were obtained from TSH alpha (with two glycosylation sites), including neutral oligosaccharides as well as those with one or two negative charges. For TSH beta (with one glycosylation site) at least eight oligosaccharide species were noted, representing nearly every permutation of sialylation and sulfation; approximately 30% contained three or more negative charges. Analysis of [3H]mannose-labeled oligosaccharides on Concanavalin-A-agarose showed 85% binding for those from TSH alpha, 70% for free alpha, and 50% for those from TSH beta. These data demonstrate that oligosaccharides from secreted TSH beta were more sialylated and sulfated, consistent with a more complex branching pattern, than those from TSH alpha. Oligosaccharides from free alpha-subunit were more sialylated than those from TSH alpha, and the net negative charge was intermediate between those of TSH alpha and TSH beta. Although great microheterogeneity is present even at the single glycosylation site on the beta-subunit of secreted TSH, a pattern of sialylation and sulfation could be discerned. If one assigns probabilities of sialylation [p(N)] and sulfation [p(S)] based on the observed distribution within monoacidic (charge -1) species, the proportion of diacidic (charge -2) oligosaccharides could be predicted for each subunit by [p(N)]2, 2[p(N)] [p(S)], [p(S)]2, corresponding to species containing two sialic acid, one sialic acid and one sulfate, and two sulfate residues, respectively. This suggests that the probability of sialylation or sulfation at a second site on these oligosaccharides is similar to that at the first and that anionic oligosaccharides in secreted TSH and free alpha are distributed binomially with regard to sialic acid and sulfate residues.

Differential processing and regulation of thyroid-stimulating hormone subunit carbohydrate chains in thyrotropic tumors and in normal and hypothyroid pituitaries.

Thyroid-stimulating hormone (TSH) alpha- and beta-subunit glycosylation was investigated in mouse thyrotropic tumor and in normal and hypothyroid pituitary cells for various periods of time in the presence of [3H]mannose or [3H]galactose. After sequential precipitation with anti-alpha and anti-beta sera, subunits were treated with Pronase followed by endo-beta-N-acetylglucosaminidase H (Endo H) and analyzed by paper chromatography. In primary cultures of thyrotropic tumor cells incubated for 60 min with [3H]mannose, primarily Man9GlcNAc and Man8GlcNAc were found on TSH + alpha subunits, whereas Glc1Man9GlcNAc and Man9GlcNAc were prominent on free beta subunits. After preincubation of cells for 16 h in the presence or absence of glucose followed by a 60-min pulse of [3H]mannose, there was an 8-fold increase in labeled TSH + alpha but only a minimal change in free beta or total proteins. In the absence of glucose, there was a selective accumulation of Man8GlcNAc on TSH + alpha but not on free beta or total proteins; however, there was no detectable accumulation of Endo H resistant forms during glucose starvation on TSH subunits or total proteins. Normal mouse and rat pituitary minces incubated for 60 min with either [3H]mannose or [3H]galactose showed no glucose-containing species on TSH subunits, but equal amounts of Man9GlcNAc and Man8GlcNAc on TSH + alpha, and mostly Man9GlcNAc on free beta subunits. In contrast, hypothyroid mouse and rat pituitaries exhibited an increase in Glc1Man9NAc and Glc1Man8GlcNAc on free beta but not on TSH + alpha or total proteins.(ABSTRACT TRUNCATED AT 250 WORDS)

Differential carbohydrate processing and secretion of thyrotropin and free alpha subunit. Effects of 1-deoxynojirimycin.

In pulse-chase experiments we compared the kinetics of early carbohydrate processing and subsequent secretion of thyroid-stimulating hormone (TSH) and free alpha subunit under control conditions and after treatment with 1-deoxynojirimycin, an inhibitor of glucosidases I and II. Under control conditions TSH achieved resistance to endo-beta-N-acetylglucosaminidase H (endo H) more rapidly than free alpha (t1/2 0.3 h versus 0.9 h); however, free alpha was secreted more rapidly than TSH (t1/2 2.2 h versus 3.4 h). With 1-deoxynojirimycin, oligosaccharides co-migrating with G3Man9GlcNAc and G2Man9GlcNAc were demonstrated on TSH for the first time, suggesting that previous pulse-chase studies did not disclose these intermediates due to rapid removal of glucose residues from the common G3Man9GlcNAc2 precursor. 1-Deoxynojirimycin delayed the rate of attainment of endo H resistance for both TSH and free alpha, but there was no effect on subunit combination. With 5 mM 1-deoxynojirimycin the amount of secreted free alpha was reduced to 65% of control; secreted TSH was reduced markedly to 17% of control without intracellular accumulation, suggesting increased intracellular degradation. There was no significant toxicity from these doses of 1-deoxynojirimycin on the production or secretion of the two major nonglycosylated pituitary proteins, growth hormone and prolactin, or on at least 10 other secretory proteins. Basal differences in the relative rates of TSH and free alpha processing and secretion as well as differential sensitivity to 1-deoxynojirimycin suggest separate secretory pathways for these two closely related proteins.

Alterations in the glycosylation of secreted thyrotropin during ontogenesis. Analysis of sialylated and sulfated oligosaccharides.

We have examined the carbohydrate structure of thyrotropin (TSH) secreted in vitro by pituitaries from prenatal, perinatal, and mature rats using concanavalin A (ConA)-agarose chromatography and anion-exchange high performance liquid chromatography (HPLC). [3H]Glucosamine-labeled TSH was immuno-precipitated and treated with either Pronase to generate glycopeptides or a mixture of endo-beta-N-acetyl-glucosaminidase F and peptide:N-glycosidase F to release oligosaccharides. The percentage of secreted TSH glycopeptides not bound to ConA was greater in mature animals (47 +/- 3%) than in either prenatal (29 +/- 3%) or perinatal animals (29 +/- 6%), suggesting more multiantennary oligosaccharides in the older animals. These structural changes were characterized further by performing anion-exchange HPLC on released oligosaccharides. Secreted TSH from prenatal rats predominantly contained oligosaccharides with 1 and 2 negative charges, while TSH from mature rats contained these structures as well as 15% with 3 negative charges. In addition, the ratio of sialylated to sulfated oligosaccharides was greater in mature compared to prenatal animals for species with 1 negative charge (1.9-fold) as well as for species with 2 negative charges (4.3-fold). We also correlated the structural alterations noted by ConA analysis with anion-exchange HPLC. Oligosaccharides that bound to ConA and were eluted with alpha-methylglucoside, when analyzed by anion-exchange HPLC, consisted of species with 1 and 2 negative charges, whereas oligosaccharides that were unbound to ConA were predominantly species with three negative charges. Together, these data suggest that with maturation of the hypothalamic-pituitary-thyroid axis secreted TSH contains more negatively charged multiantennary oligosaccharides with increased sialylation and decreased sulfation.

Replacement of tyrosine 1251 in the carboxyl terminus of the insulin-like growth factor-I receptor disrupts the actin cytoskeleton and inhibits proliferation and anchorage-independent growth.

Insulin-like growth factor (IGF)-I signaling through the IGF-I receptor modulates cellular adhesion and proliferation and the transforming ability of cells overexpressing the IGF-I receptor. Tyrosine phosphorylation of intracellular proteins is essential for this transduction of the IGF-I-induced mitogenic and tumorigenic signals. IGF-I induces specific cytoskeletal structure and the phosphorylation of proteins in the associated focal adhesion complexes. The determination of the exact pathways emanating from the IGF-I receptor that are involved in mediating these signals will contribute greatly to the understanding of IGF-I action. We have previously shown that replacement of tyrosine residues 1250 and 1251 in the carboxyl terminus of the IGF-I receptor abrogates IGF-I-induced cellular proliferation and tumor formation in nude mice. In this study, replacement of either tyrosine 1250 or 1251 similarly reduces the cells ability to grow in an anchorage-independent manner. The actin cytoskeleton and cellular localization of vinculin are disrupted by replacement of tyrosine 1251. Tyrosine residues 1250 and 1251 are not essential for tyrosine phosphorylation of two known substrates; insulin receptor substrate-1 and SHC, nor association of known downstream adaptor proteins to these substrates. In addition, these mutant IGF-I receptors do not affect IGF-I-stimulated p42/p44 mitogen-activated protein kinase activation or phosphatidylinositol (PI) 3'-kinase activity. Thus, it appears that in fibroblasts expressing tyrosine 1250 and 1251 mutant IGF-I receptors, the signal transduction pathways impacting on mitogenesis and tumorigenesis do not occur exclusively through the PI 3'-kinase or mitogen-activated protein kinase pathways.

Glycosylation and processing of high-mannose oligosaccharides of thyroid-stimulating hormone subunits: comparison to nonsecretory cell glycoproteins.

Thyroid-stimulating hormone (TSH) subunit glycosylation was compared to that of total cell glycoproteins in mouse thyrotropic tumors. Lipid-linked oligosaccharides, total cell glycoproteins, and TSH subunits were labeled with either [3H]mannose, [3H]galactose, or [3H]glucose in pulse and pulse-chase experiments. The various oligosaccharides were isolated respectively by lipid extraction and mild acid hydrolysis, by selective immunoprecipitation, or by acid precipitation followed by trypsin and endoglycosidase H treatment. The nature of the oligosaccharides was assessed by their migration in paper chromatography, their relative incorporation of different precursors, and also their resistance to alpha-mannosidase. At 60 min, lipid-linked oligosaccharides were found to be composed of Glc3-2Man9GlcNAc2, Man9-8GlcNAc2, and Man5GlcNAc2. At 10 or 60 min of labeling, total cell proteins contained Glc3Man9GlcNAc2, Glc1Man9GlcNAc2, Man9GlcNAc2, Glc1Man8GlcNAc2, Man8GlcNAc2, and Man7GlcNAc2. The largest oligosaccharide, Glc3Man9GlcNAc2, had an unusually long half-life of about 2 h. In contrast, no Glc3Man9GlcNAc2 was found either on TSH + alpha subunits or on free beta subunits isolated either by immunoprecipitation or by sodium dodecyl sulfate gel electrophoresis. Instead, primarily Man9GlcNAc2 was found after a 10-min pulse both on TSH + alpha subunits and on beta subunits. When the pulse was followed by a chase up to 2 h, there was a progressive increase in Man8GlcNAc2 in higher amounts on TSH + alpha-subunit carbohydrate chains than on beta subunits.(ABSTRACT TRUNCATED AT 250 WORDS)

Differential effect of inhibitors of oligosaccharide processing on the secretion of thyrotropin from dispersed rodent pituitary cells.

We examined the effect of various inhibitors of oligosaccharide processing on the content and secretion of newly synthesized thyroid-stimulating hormone (TSH) from dispersed hypothyroid rodent pituitary cells. 1-deoxynojirimycin and N-methyl-1-deoxynojirimycin, both inhibitors of glucosidases I and II, decreased intracellular TSH (to 60-76% of control) and secreted TSH (to 60-63% of control) after a 1-hour incubation (pulse) with [35S]methionine and an 8-hour incubation (chase) in isotope-free media. In contrast, deoxymannojirimycin and swainsonine, inhibitors of mannosidase I and II, respectively, increased both intracellular TSH (to 267-309% of control) and secreted TSH (to 192% of control) at 8 hours. TSH oligosaccharides synthesized in the presence of these glucosidase and mannosidase inhibitors were largely sensitive to endo-beta-N-acetylglucosaminidase H (endo H), confirming inhibition of processing. Despite differences in oligosaccharide structure, the in vitro bioactivities of these secreted TSH isoforms were nearly identical. These data confirm and extend previous work performed with 1-deoxynojirimycin suggesting that glucosylated high mannose forms of TSH are more susceptible to intracellular degradation. The novel finding that deoxymannojirimycin and swainsonine increase secreted and total TSH above control levels suggests that non-glucosylated high mannose forms as well as hybrid-type oligosaccharides may facilitate secretion and direct TSH away from a natural degradation pathway.

Changes in the sialylation and sulfation of secreted thyrotropin in congenital hypothyroidism.

We have examined the oligosaccharide structure of secreted thyrotropin (TSH) in perinatal and mature rats with congenital primary hypothyroidism. Rat pituitaries from euthyroid control animals and those rendered hypothyroid by methimazole treatment were incubated with [3H]glucosamine in vitro. Secreted TSH was purified, and oligosaccharides were enzymatically released and characterized by anion-exchange HPLC. In perinatal hypothyroid animals compared with control animals, oligosaccharides from TSH alpha and beta subunits contained more species with three or more negative charges. Moreover, perinatal hypothyroid animals demonstrated a dramatic increase in the ratio of sialylated to sulfated species within oligosaccharides of the same negative charge (2.9- to 7.4-fold increase for TSH-alpha; 15.1- to 25.5-fold increase for TSH-beta). In mature hypothyroid 9-week-old animals compared with control animals, changes were less pronounced, suggesting that endocrine regulation of oligosaccharide structure is dependent upon the maturational state of the animal. These changes were specific for TSH because glycosylation of free alpha subunit (synthesized by the thyrotroph and gonadotroph) and of total glycoproteins was minimally altered by hypothyroidism. Together, these data provide direct evidence and characterization of specific changes in the structure of a secreted pituitary glycoprotein hormone occurring as a result of in vivo endocrine alterations during early development. Moreover, they provide a potential structural basis to explain the delayed clearance of both TSH and the gonadotropins with end-organ deficiency, which may have important implications for the in vivo biological activities of these hormones. Specifically, such posttranslational changes may be an important adaptive response to prevent the consequences of endocrine deficiency during early development.

Sphingosine 1-phosphate stimulates tyrosine phosphorylation of Crk.

The proto-oncogene molecule c-Crk plays a role in growth factor-induced activation of Ras. Sphingosine 1-phosphate (SPP), a metabolite of cellular sphingolipids, has previously been shown to play a role in growth factor receptor signaling (Olivera, A., and Spiegel, S. (1993) Nature 365, 557-560). SPP was found to strongly induce tyrosine phosphorylation of Crk, but not Shc, in NIH-3T3 parental, insulin-like growth factor-I receptor-overexpressing and Crk-overexpressing (3T3-Crk) fibroblasts. Sphingosine, a metabolic precursor of SPP, also produced a slight increase in tyrosine phosphorylation of Crk. In contrast, other sphingolipid metabolites including ceramide did not alter Crk tyrosine phosphorylation. Furthermore, Crk enhanced SPP-induced mitogenesis, as measured by SPP-stimulated [3H]thymidine incorporation in a manner proportional to the level of Crk expression in 3T3-Crk cells. This stimulation appears to be Ras-dependent, whereas SPP stimulation of MAP kinase activity is Ras-independent. These data indicate that SPP activates a tyrosine kinase that phosphorylates Crk and that Crk is a positive effector of SPP-induced mitogenesis.