Mutagenesis and gene transfer define site-specific roles of the gonadotropin oligosaccharides

Human chorionic gonadotropin (hCG), luteinizing hormone (LH), follicle-stimulating hormone and thyroid-stimulating hormone are a family of glycoprotein hormones that share a common alpha subunit but differ in their hormone-specific beta subunits. Using site-directed mutagenesis and gene-transfer, we analyzed the role of the N-linked oligosaccharides of alpha and chorionic gonadotropin (CG)beta in the secretion, assembly, and biologic activity of hCG. Absence of carbohydrate at alpha asparagine (Asn) 52 decreased combination with CG beta but did not alter monomer secretion. Absence of the alpha Asn78 oligosaccharide increased the degradation of the alpha subunit, but the presence of CG beta stabilized this alpha mutant in an efficiently formed dimer complex. Alternatively, absence of both alpha oligosaccharides slowed both secretion and dimer formation but allowed an intermediate level of alpha secreted or dimerized compared to the single-site mutants. Analysis of the CG beta glycosylation mutants revealed that absence of the Asn30 oligosaccharide, but not Asn13, slowed secretion but not assembly, whereas absence of both oligosaccharides slowed both secretion and dimer formation. Analysis of the receptor binding of the hCG glycosylation mutants showed that absence of any or all of the hCG N-linked oligosaccharides had only a minor effect on receptor affinity of the derivatives. However, the absence of alpha Asn52, but not the alpha Asn78 or the CG beta carbohydrate units, reduced the steroidogenic effect, unmasked differences in the beta oligosaccharides, and converted the deglycosylated derivatives into antagonists.     

The role of the asparagine-linked oligosaccharides of the alpha subunit in the secretion and assembly of human chorionic gonadotrophin

Human chorionic gonadotropin (hCG) is a member of a family of heterodimeric glycoprotein hormones that have a common alpha subunit but differ in their hormone-specific beta subunit. Site-directed mutagenesis of the two asparagine-linked glycosylation sites of hCG alpha was used to study the function of the individual oligosaccharide chains in secretion and subunit assembly. Expression vectors for the alpha genes (wild-type and mutant) and the hCG beta gene were constructed and transfected into Chinese hamster ovary cells. Loss of the oligosaccharide at position 78 causes the mutant subunit to be degraded quickly and less than 20% is secreted. However, the presence of hCG beta stabilizes this mutant and allows approximately 45% of the subunit in the form of a dimer to exit the cell. Absence of carbohydrate at asparagine 52 does not perturb the stability or transport of the alpha subunit but does affect dimer secretion; under conditions where this mutant or hCG beta was in excess, less than 30% is secreted in the form of a dimer. Mutagenesis of both glycosylation sites affects monomer and dimer secretion but at levels intermediate between the single-site mutants. We conclude that there are site-specific functions of the hCG alpha asparagine-linked oligosaccharides with respect to the stability and assembly of hCG.     

Gonadotropin beta subunits determine the rate of assembly and the oligosaccharide processing of hormone dimer in transfected cells

The glycoprotein hormones lutropin (LH) and chorionic gonadotropin (CG) share a common structure consisting of an identical alpha subunit noncovalently linked to a hormone-specific beta subunit. While LH is produced in the anterior pituitary, CG is synthesized in placenta. To compare the assembly, processing, and secretion of human LH and CG in the same cell type, we have expressed their subunits, individually and together, in mouse C-127 mammary tumor cells. Analysis of transfected clones revealed an unexpected difference in the secretion of individually expressed subunits. Whereas alpha and CG beta subunits were rapidly and quantitatively secreted, only 10% of newly synthesized LH beta subunit reached the medium. The remaining subunit was found in an intracellular, endoglycosidase H (endo H)-sensitive pool that had a turnover rate of approximately 8 h. Coexpression with alpha subunit resulted in "rescue" of LH beta subunit by formation of LH dimer, which was efficiently secreted. However, combination of LH beta with alpha was slow, with an overall efficiency of only 50% despite the presence of excess alpha. In contrast, CG beta was rapidly assembled with the alpha subunit after synthesis. The two beta subunits also differed in their influence on the N-linked oligosaccharide processing of combined alpha. The oligosaccharides of LH dimer were endo H resistant, while those of CG dimer remained partially endo H sensitive. Thus, despite a high degree of homology between LH beta and CG beta, the two subunits differ in their secretion as free subunits, their rate of assembly with alpha subunit, and in their effect on the N-linked oligosaccharide processing of combined alpha.     

Site-specific processing of the N-linked oligosaccharides of the human chorionic gonadotropin alpha subunit

Two forms of the gonadotropin alpha subunit are synthesized in placenta and in human chorionic gonadotropin (hCG)-producing tumors: an uncombined (monomer) form and a combined (dimer) form. These forms show differences in their migration on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The slower migration of the monomeric form on sodium dodecyl sulfate-polyacrylamide gel electrophoresis has been attributed to a different glycosylation pattern. Previous studies demonstrated different roles of each of the two alpha N-linked glycosylation sites (Asn-52 and Asn-78) in secretion of the uncombined subunit and the biologic activity of hCG dimer. To assess the influence of formation of dimer on the processing pattern at the individual sites, we characterized the N-linked oligosaccharides of monomer and dimer forms of recombinant human choriogonadotropin alpha subunit. Two approaches were employed. First, site-directed mutagenesis was used to alter the two N-linked oligosaccharide attachment sites, thus allowing the expression of alpha subunits containing only one glycosylation site. Second, tryptic glycopeptides of the wild-type subunits were examined. Concanavalin A (ConA) binding and sialic acid content indicated that the oligosaccharides at each glycosylation site of the uncombined alpha subunit are processed differently. Oligosaccharides present at Asn-52 are almost exclusively ConA-unbound and contain three sialic acid residues. The majority of Asn-78-linked oligosaccharides are ConA-bound and disialylated. Both sites are processed independently because no significant differences were observed between the oligosaccharides at the same sites in wild-type and mutant monomeric alpha subunits. By contrast, the majority of the oligosaccharides at both glycosylation sites of the dimer alpha are bound to ConA. Thus, combination primarily affects the processing pattern of the Asn-52-linked species. Because glycosylation at this site is essential for hCG assembly and signal transduction, these data imply a critical link between the site-specific processing and hormone function.     

Site specificity of the chorionic gonadotropin N-linked oligosaccharides in signal transduction

The role of the human chorionic gonadotropin (hCG) N-linked oligosaccharides in receptor binding and signal transduction was analyzed using site-directed mutagenesis and transfection studies. hCG derivatives with alterations at individual glycosylation sites were expressed in Chinese hamster ovary cells. Receptor binding studies showed that absence of any or all of the hCG N-linked oligosaccharides had only a minor effect on the receptor affinity of the derivatives. Similarly, absence of the N-linked oligosaccharides from the beta subunit or a single oligosaccharide from Asn-78 of alpha had no effect on the production of cAMP or on steroidogenesis. However, the absence of carbohydrate at Asn-52 of alpha decreases both the steroidogenic and cAMP responses. Furthermore, absence of this critical oligosaccharide unit on alpha unmasks differences in the two N-linked oligosaccharides on beta; the beta Asn-13 oligosaccharide but not the beta Asn-30 oligosaccharide plays a more important role in steroidogenesis. Dimers containing deglycosylated beta subunit and an alpha subunit lacking either the Asn-52 oligosaccharide or both oligosaccharides fail to stimulate cAMP or steroid formation. Moreover, these derivatives bind to receptor and behave as competitive antagonists. The use of site-directed mutagenesis was critical in uncovering site-specific functions of the hCG N-linked oligosaccharides in signal transduction and reveals the importance of the Asn-52 oligosaccharide in this process.     

N-linked oligosaccharides direct the differential assembly and secretion of inhibin alpha- and betaA-subunit dimers

The biosynthetic pathway governing inhibin heterodimer (alpha/beta) and activin homodimer (beta/beta) assembly and secretion from ovarian granulosa cells is not fully understood. Here, we examined the role of inhibin subunit glycosylation in the assembly and secretion of mature inhibin A and activin A. Inhibition of subunit glycosylation by tunicamycin treatment of alpha- and beta(A)-expressing CHO cell lines reduced inhibin but not activin secretion. Dimeric inhibin A is preferentially secreted from parental isogenic wild-type (wt) cell lines (alpha(wt)beta(wt)). Mutation of a single glycosylation site at asparagine 268 (alpha(Delta268)beta(wt)) reduces inhibin secretion by 78% and permits beta/beta assembly and secretion. Conversely, gain of a glycosylation (GOG) site in the analogous region of the beta(A)-subunit (alpha(wt)beta(GOG327)) enhances inhibin A secretion. The present study demonstrates that N-linked glycan sites direct heterodimer vs. homodimer assembly, and prevention of glycosylation abrogates inhibin secretion. These data support a definitive role for site-specific N-glycosylation in governing inhibin/activin dimer assembly and secretion.     

N-Linked oligosaccharides on the meprin A metalloprotease are important for secretion and enzymatic activity, but not for apical targeting

The alpha and beta subunits of meprins, mammalian zinc metalloendopeptidases, are extensively glycosylated; approximately 25% of the total molecular mass of the subunits is carbohydrate. The aim of this study was to investigate the roles of the N-linked oligosaccharides on the secreted form of mouse meprin A. Recombinant meprin alpha and mutants in which one of the 10 potential Asn glycosylation sites was mutated to Gln were all secreted and sorted exclusively into the apical medium of polarized Madin-Darby canine kidney cells, indicating that no specific N-linked oligosaccharide acts as a determinant for apical targeting of meprin alpha. Several of the mutant proteins had decreased enzymatic activity using a bradykinin analog as substrate, and deglycosylation of the wild-type protein resulted in loss of 75-100% activity. Some of the mutants were also more sensitive to heat inactivation. In studies with agents that inhibit glycosylation processes in vivo, tunicamycin markedly decreased secretion of meprin, whereas castanospermine and swainsonine had little effect on secretion, sorting, or enzymatic properties of meprin. When all the potential glycosylation sites on a truncated form of meprin alpha (alpha-(1-445)) were mutated, the protein was not secreted into the medium, but was retained within the cells even after 10 h. These results indicate that there is no one specific glycosylation site or type of oligosaccharide (high mannose- or complex-type) that determines apical sorting, but that core N-linked carbohydrates are required for optimal enzymatic activity and for secretion of meprin alpha.     

Functional analysis of N-linked glycosylation mutants of the measles virus fusion protein synthesized by recombinant vaccinia virus vectors.

The role of N-linked glycosylation in the biological activity of the measles virus (MV) fusion (F) protein was analyzed by expressing glycosylation mutants with recombinant vaccinia virus vectors. There are three potential N-linked glycosylation sites located on the F2 subunit polypeptide of MV F, at asparagine residues 29, 61, and 67. Each of the three potential glycosylation sites was mutated separately as well as in combination with the other sites. Expression of mutant proteins in mammalian cells showed that all three sites are used for the addition of N-linked oligosaccharides. Cell surface expression of mutant proteins was reduced by 50% relative to the wild-type level when glycosylation at either Asn-29 or Asn-61 was abolished. Despite the similar levels of cell surface expression, the Asn-29 and Asn-61 mutant proteins had different biological activities. While the Asn-61 mutant was capable of inducing syncytium formation, the Asn-29 mutant protein did not exhibit any significant cell fusion activity. Inactivation of the Asn-67 glycosylation site also reduced cell surface transport of mutant protein but had little effect on its ability to cause cell fusion. However, when the Asn-67 mutation was combined with mutations at either of the other two sites, cleavage-dependent activation, cell surface expression, and cell fusion activity were completely abolished. Our data show that the loss of N-linked oligosaccharides markedly impaired the proteolytic cleavage, stability, and biological activity of the MV F protein. The oligosaccharide side chains in MV F are thus essential for optimum conformation of the extracellular F2 subunit that is presumed to bind cellular membranes.     

N-linked oligosaccharide chains of the insulin receptor beta subunit are essential for transmembrane signaling.

Insulin receptor (IR) is a glycoprotein possessing N-linked oligosaccharide side chains on both alpha and beta subunits. The present study focuses for the first time on the potential contribution of N-linked oligosaccharides of the beta subunit in the processing, structure, and function of the insulin receptor. To investigate this point, a receptor mutant (IR beta N1234) was obtained by stable transfection into Chinese hamster ovary cells of an IR cDNA modified by site-directed mutagenesis on the four potential N-glycosylation sites (Asn-X-Ser/Thr) of the beta subunit. The mutated receptor presents an alpha subunit of 135 kDa, indistinguishable from the wild type alpha subunit, but the beta subunit has a reduced molecular mass (80 kDa instead of 95 kDa) most likely due to the absence of N-glycosylation. Metabolic labeling experiments indicate a normal processing and maturation of this mutated receptor which is normally expressed at the surface of the cells as demonstrated by indirect immunofluorescence. The affinity of the mutant for insulin (Kd = 0.12 nM) is similar to that of the wild type receptor (Kd = 0.12 nM). However, a major defect of the mutated IR tyrosine kinase was assessed both in vitro and in vivo by (i) the absence of insulin-stimulated phosphorylation of the poly(Glu-Tyr) substrate in vitro; (ii) the reduction of the insulin maximal stimulation of the mutated IR autophosphorylation in vitro (2-fold stimulation for the mutant receptor as compared to a 7-fold stimulation for the wild type); and (iii) a more complex alteration of the mutated receptor tyrosine autophosphorylation in vivo (3-fold increase of the basal phosphorylation and a 4-fold simulation of this phosphorylation as compared to the wild type receptor, the phosphorylation of which is stimulated 14-fold by insulin). The physiological consequences of this defect were tested on three classical insulin cellular actions; in Chinese hamster ovary IR beta N1234, glucose transport, glycogen synthesis, and DNA synthesis were all unable to be stimulated by insulin indicating the absence of insulin transduction through this mutated receptor. These data provide the first direct evidence for a critical role of oligosaccharide side chains of the beta subunit in the molecular events responsible for the IR enzymatic activation and signal transduction.     

Mutagenesis and chimeric genes define determinants in the beta subunits of human chorionic gonadotropin and lutropin for secretion and assembly

Chorionic gonadotropin (CG) and lutropin (LH) are members of a family of glycoprotein hormones that share a common alpha subunit but differ in their hormone-specific beta subunits. The glycoprotein hormone beta subunits share a high degree of amino acid homology that is most evident for the LH beta and CG beta subunits having greater than 80% sequence similarity. However, transfection studies have shown that human CG beta and alpha can be secreted as monomers and can combine efficiently to form dimer, whereas secretion and assembly of human LH beta is less efficient. To determine which specific regions of the LH beta and CG beta subunits are responsible for these differences, mutant and chimeric LH beta-CG beta genes were constructed and transfected into CHO cells. Expression of these subunits showed that both the hydrophobic carboxy-terminal seven amino acids and amino acids Trp8, Ile15, Met42, and Asp77 together inhibit the secretion of LH beta. The carboxy-terminal amino acids, along with Trp8, Ile15, Met42, and Thr58 are implicated in the delayed assembly of LH beta. These unique features of LH beta may also play an important role in pituitary intracellular events and may be responsible for the differential glycosylation and sorting of LH and FSH in gonadotrophs.     

Expression of recombinant human choriogonadotropin in Chinese hamster ovary glycosylation mutants

Human CG, a member of the glycoprotein hormone family that includes LH, FSH, and TSH, is composed of two nonidentical subunits each containing two asparagine linked (N-linked) oligosaccharides. The role of the oligosaccharides in the action of these hormones is unclear. To examine the structure-activity relationships of the glycoprotein hormone oligosaccharides using nonenzymatic and nonchemical methods, we transfected CG subunit genes into mutant cell lines derived from Chinese hamster ovary cells. Two mutant cell lines that synthesize truncated oligosaccharides were used. Cell line 15B, lacking N-acetylglucosaminyltransferase I, synthesizes N-linked carbohydrates containing Man5 oligomannosyl structures, and 1021, defective in transporting CMP-sialic acid into the Golgi, results in sialic-acid deficient glycoproteins. The binding of these derivatives to the LH/CG receptor did not differ significantly from purified CG (CR119), but the ability of the mutant hormones to stimulate cAMP biosynthesis in vitro is reduced compared to wild-type CG or CR119. Since the amino acid sequence of CG from the mutant and wild-type cells is identical, these data indicate that oligosaccharide structures, while not influencing receptor binding, directly affect signal transduction.     

Molecular characterization of two point mutants in the chloroplast atpB gene of the green alga Chlamydomonas reinhardtii defective in assembly of the ATP synthase complex

Two point mutants of Chlamydomonas reinhardtii, previously found by recombination and complementation analysis to map in the chloroplast atpB gene encoding the beta subunit of the CF1/CF0 ATP synthase, are here shown to be missense alterations near the 5' end of that gene. One mutant (ac-u-c-2-9) has a change at amino acid position 47 of the beta subunit from leucine (CTA) to arginine (CGA). In the second mutant (ac-u-c-2-29), the codon AAA (lysine) is changed to AAC (asparagine) at position 154. Spontaneous revertants of each mutant were isolated that restore the original wild type base pair. Northern analysis of total RNA and in vivo pulse labeling followed by immunoprecipitation reveals that both mutant atpB genes are transcribed and translated normally. However, immunoblots show that the amount of beta subunit associated with mutant thylakoids is only approximately 3% of that seen in wild type and that the CF1 alpha and gamma subunits are missing entirely. The disruption of ATP synthase complex assembly in these mutants is much more severe than in Escherichia coli beta subunit gene point mutants, which retain significant amounts of alpha and beta subunits on their membranes (Noumi, T., Oka, N., Kanazawa, H., and Futai, M. (1986) J. Biol. Chem. 261, 7070-7075). These results support the hypothesis that there are differences in assembly of the ATP synthase between E. coli and chloroplasts. In particular they indicate that beta must be present for assembly of the alpha and gamma subunits of CF1 onto chloroplast membranes.     

Selective removal of alpha heavy-chain glycosylation sites causes immunoglobulin A degradation and reduced secretion.

The importance of carbohydrate in the secretion of immunoglobulin A (IgA) has previously been suggested by results of studies with tunicamycin, which prevents N-linked glycosylation of all cell glycoproteins. To directly evaluate the role of individual oligosaccharides in the secretion of IgA, we have used site-directed mutagenesis to selectively eliminate the two N-linked attachment sites reported to be glycosylated in alpha heavy chains. Transfected wild-type and mutant alpha genes were expressed in kappa light-chain-producing MPC-11 variant myeloma cells, and secretion kinetics of the IgAs were compared. Removal of either or both glycosylation sites led to intracellular alpha heavy-chain degradation and a 90 to 95% inhibition of IgA secretion. These results reveal that both N-linked oligosaccharides of the alpha heavy chain are essential for intracellular stability and normal secretion of IgA. This suggests that the key function of carbohydrate here is to maintain proper conformation of the glycoprotein. We also found that when expressed in the MPC-11 variant cells, alpha heavy chains were glycosylated at a third, normally unused site.     

Gonadotropin alpha subunit. Differential processing of free and combined forms in human trophoblast and transfected mouse cells

The gonadotropins luteinizing hormone, follicle-stimulating hormone, and human chorionic gonadotropin are composed of two noncovalently linked subunits, alpha and beta. The alpha subunit, identical in all three hormones, is produced in excess over the unique beta subunits by pituitary and placenta, and is secreted as uncombined, or free subunit. Free alpha subunit from both tissues has a larger molecular weight than the dimer form. In bovine pituitary an extra O-linked oligosaccharide is added to free alpha subunit, and this modification has recently been detected at an analogous position (threonine 39) on human alpha subunit secreted by choriocarcinoma cells. To assess the contribution of N-linked and O-linked oligosaccharides to the heterogeneity of human free alpha subunit, we have compared free alpha with human chorionic gonadotropin alpha secreted by explants and cultured cytotrophoblasts of human first trimester placenta. We have also examined the free and combined forms of human alpha subunit expressed in transfected C-127 mouse mammary tumor cells. Processing of the alpha subunit in placental and C-127 cells was similar. Tryptic mapping of placental-derived and transfected alpha subunits indicated that O-glycosylation at threonine 39 was not a major modification. In the presence of the oligosaccharide processing inhibitor swainsonine the difference in size between the free and combined forms of alpha was eliminated in both placental and C-127 cells, indicating that the two forms of alpha differed in their N-linked oligosaccharides. Furthermore, the oligosaccharides of free alpha subunits from placental and transfected cells were resistant to endoglycosidase H, but the combined forms of alpha were partially sensitive to the enzyme. Thus, in human first trimester placenta and mouse C-127 cells, combination of alpha with human chorionic gonadotropin beta alters the processing of N-linked oligosaccharides on alpha subunit.     

Elimination of disulfide bonds affects assembly and secretion of the human chorionic gonadotropin beta subunit

Human chorionic gonadotropin (hCG) consists of two noncovalently joined alpha and beta subunits similar to the other glycoprotein hormones. To study the function of the individual disulfide bonds in subunit assembly and secretion, site-directed mutagenesis was used to convert the 12 cysteine (Cys) residues in the beta subunit of hCG to either alanine or serine. Both cysteines of proposed disulfide pairs were also mutated. These mutant hCG beta genes were transfected alone or together with the wild-type alpha gene into Chinese hamster ovary cells. Only 3-10% assembly could be achieved with derivatives containing single Cys mutations at positions 26, 110, 72, and 90, whereas no assembly was detected with the other 8 mutants. However, double mutations of pairs 26-110 or 23-72 showed increased dimer formation (11 and 36%, respectively). The secretion rate of individual mutants varied significantly. Whereas the Cys-23 and 72 mutants were secreted normally (t1/2 = 140-190 min), the Cys-26 mutant was secreted faster (t1/2 = 70 min), and the other 9 mutants were secreted slower (t1/2 = 280-440 min); mutations of both Cys at 26 and 110 caused much faster secretion (t1/2 = 34 min). Although the secretion rate of these mutants differed, they were quantitatively recovered in the medium except for mutant Cys-88, Cys-23-72, and Cys-34-88 (40, 55, and 10% secreted, respectively). Thus, interruption of any disulfide bond in the hCG beta subunit alters the structure sufficiently to block dimer formation and in some cases slow secretion, although the stability for most of the mutant hCG beta subunits is not greatly affected. The data indicate that interruption of any hCG beta disulfide bond generates different structural forms that are unable to assemble with the alpha subunit, and that the structural requirements for stability and assembly are different.     

Role of N-linked oligosaccharides in processing and intracellular transport of E2 glycoprotein of rubella virus.

The role of N-linked glycosylation in processing and intracellular transport of rubella virus glycoprotein E2 has been studied by expressing glycosylation mutants of E2 in COS cells. A panel of E2 glycosylation mutants were generated by oligonucleotide-directed mutagenesis. Each of the three potential N-linked glycosylation sites was eliminated separately as well as in combination with the other two sites. Expression of the E2 mutant proteins in COS cells indicated that in rubella virus M33 strain, all three sites are used for the addition of N-linked oligosaccharides. Removal of any of the glycosylation sites resulted in slower glycan processing, lower stability, and aberrant disulfide bonding of the mutant proteins, with the severity of defect depending on the number of deleted carbohydrate sites. The mutant proteins were transported to the endoplasmic reticulum and Golgi complex but were not detected on the cell surface. However, the secretion of the anchor-free form of E2 into the medium was not completely blocked by the removal of any one of its glycosylation sites. This effect was dependent on the position of the deleted glycosylation site.     

Effect of N-linked glycosylation on hepatic lipase activity

Hepatic lipase (HL) is a secretory protein synthesized in hepatocytes and bound to liver endothelium. Previous studies have suggested that HL N-linked glycans are required for catalytic activity. To directly test this hypothesis, Xenopus laevis oocytes were used to express native rat HL or HL lacking one or both N-linked glycosylation sites. The expressed and secreted native HL had an apparent molecular mass of 53 kDa, consistent with purified rat liver HL. The mutant lacking both glycosylation sites, while poorly secreted, had an apparent molecular mass of 48 kDa, the same size observed for HL after enzymatic removal of N-linked oligosaccharides. Mutants lacking one of the two sites were intermediate in size and showed reduced secretion. Each of these expressed and secreted proteins had full catalytic activity that was inhibited by antisera to rat HL. Thus, N-linked glycosylation of rat HL, while important to lipase secretion, is not essential for the expression of lipase activity.     

N-linked oligosaccharide processing and autocrine stimulation of tumor cell proliferation

Somatic mutations which impair complex-type N-linked oligosaccharide processing and chemical inhibitors of processing have been shown to reduce metastatic potential in several experimental tumor models. In this report, we demonstrate that glycosylation mutants of the metastatic MDAY-D2 tumor cell line with either truncated glycans lacking sialic acid and galactose or a mutant with less branched N-linked oligosaccharides grow more slowly in serum-free medium (SFM) than do MDAY-D2 cells. In medium containing fetal calf serum, growth rates of the cell lines were similar. A revertant of the former mutation showed a return to a more rapid growth rate in SFM. The N-linked processing inhibitor swainsonine also reduced cell growth rate in SFM but not in serum-containing medium. One of five randomly selected clones of the MDAY-D2 tumor cell line showed a slower growth rate in SFM and also showed decreased expression of branched N-linked oligosaccharides. These observations suggest that in MDAY-D2 cells, optimal factor-independent stimulation is dependent upon expression of branched complex-type N-linked oligosaccharides. The growth rate of MDAY-D2 cells in SFM was dependent on the initial seeding density of the cultures, and medium conditioned by the cells accelerated the growth of low-density cultures, suggesting that the cells respond to an autocrine factor. Culture supernatants conditioned by mutant and wild-type cells had similar levels of growth-stimulating activity. However, both mutants and swainsonine-treated cells were less responsive to this growth-stimulating activity. The growth rates of the MDAY-D2 tumor cell lines in vivo as subcutaneous tumors correlated with their relative growth rates in SFM in vitro. The results suggest that branched complex-type N-linked oligosaccharides commonly expressed in malignant cells are required for optimal autocrine-dependent growth in vitro and may be a significant factor in tumor progression in vivo.     

Reciprocal modulation between the alpha and beta 4 subunits of hSlo calcium-dependent potassium channels

Large conductance Ca(2+)-dependent potassium (K(Ca) or maxi K) channels are composed of a pore-forming alpha subunit and an auxiliary beta subunit. We have shown that the brain-specific beta4 subunit modulates the voltage dependence, activation kinetics, and toxin sensitivity of the hSlo channel (Weiger, T. M., Holmqvist, M. H., Levitan, I. B., Clark, F. T., Sprague, S., Huang, W. J., Ge, P., Wang, C., Lawson, D., Jurman, M. E., Glucksmann, M. A., Silos-Santiago, I., DiStefano, P. S., and Curtis, R. (2000) J. Neurosci. 20, 3563-3570). We investigated here the N-linked glycosylation of the beta4 subunit and its effect on the modulation of the hSlo alpha subunit. When expressed alone in HEK293 cells, the beta4 subunit runs as a single molecular weight band on an SDS gel. However, when coexpressed with the hSlo alpha subunit, the beta4 subunit appears as two different molecular weight bands. Enzymatic deglycosylation or mutation of the N-linked glycosylation residues in beta4 converts it to a single lower molecular weight band, even in the presence of the hSlo alpha subunit, suggesting that the beta4 subunit can be present as an immature, core glycosylated form and a mature, highly glycosylated form. Blockage of protein transport from the endoplasmic reticulum to the Golgi compartment with brefeldin A abolishes the mature, highly glycosylated beta4 band. Glycosylation of the beta4 subunit is not required for its binding to the hSlo channel alpha subunit. It also is not necessary for cell membrane targeting of the beta4 subunit, as demonstrated by surface biotinylation experiments. However, the double glycosylation site mutant beta4 (beta4 N53A/N90A) protects the channel less against toxin blockade, as compared with the hSlo channel coexpressed with wild type beta4 subunit. Taken together, these data show that the pore-forming alpha subunit of the hSlo channel promotes N-linked glycosylation of its auxiliary beta4 subunit, and this in turn influences the modulation of the channel by the beta4 subunit.     

Decorin core protein secretion is regulated by N-linked oligosaccharide and glycosaminoglycan additions

Expression of decorin using the vaccinia virus/T7 expression system resulted in secretion of two distinct glycoforms: a proteoglycan substituted with a single chondroitin sulfate chain and N-linked oligosaccharides and a core protein glycoform substituted with N-linked glycans but without a glycosaminoglycan chain. In this report, we have addressed two distinct questions. What is the rate-limiting step in glycosaminoglycan synthesis? Is glycosylation with either N-linked oligosaccharides or glycosaminoglycan required for secretion of decorin? N-terminal sequencing of the core protein glycoform, the addition of benzyl-beta-d-xyloside, and a UDP-xylose: core protein beta-d-xylosyltransferase activity assay show that xylosylation is a rate-limiting step in chondroitin sulfate biosynthesis. Decorin can be efficiently secreted with N-linked oligosaccharides alone or with a single chondroitin sulfate chain alone; however, there is severely impaired secretion of core protein devoid of any glycosylation. A decorin core protein mutant devoid of N-linked oligosaccharide attachment sites will not be secreted by Chinese hamster ovary cells deficient in xylosyltransferase or by parental Chinese hamster ovary wild type cells if the xylosyltransferase recognition sequence is disrupted. This finding suggests that quality control mechanisms sensitive to an absence of N-linked oligosaccharides can be abrogated by interaction of the core protein with the glycosaminoglycan synthetic machinery. We propose a model of regulation of decorin secretion that has several components, including appropriate substitution with N-linked oligosaccharides and factors involved in glycosaminoglycan synthesis.