Analysis of glycosaminoglycan substitution in decorin by site-directed mutagenesis

Posttranslational glycosaminoglycan attachment to decorin, a chondroitin/dermatan sulfate proteoglycan, was studied by expression of a wild-type decorin cDNA and several mutagenized forms in two types of mammalian cells. Transfection of the wild-type cDNA resulted in the synthesis of an authentic chondroitin/dermatan sulfate proteoglycan similar to the decorin molecule synthesized by cultured human fibroblasts. Conversion of the serine residue that serves as the attachment site for the sole glycosaminoglycan chain in decorin to a threonine residue greatly reduced the efficiency of the glycosaminoglycan substitution. Less than 10% of the threonine-mutated core protein acquired a glycosaminoglycan chain, whereas most of the core protein was secreted without such substitution. Expression of cDNA in which an alanine residue had been introduced into the substituted serine position resulted in the secretion of core protein with no detectable glycosaminoglycan. Conversion to alanine of either one of the glycine residues that are adjacent to the substituted serine yielded the proteoglycan form of decorin. These results show that the xylosyltransferase responsible for the initiation of the glycosaminoglycan chain on the core protein can use a threonine residue for this substitution instead of a serine residue, but that such substitution is only partial, creating a "part-time" proteoglycan. Moreover, variations are possible in the sequence context of a glycosaminoglycan-substituted serine residue without loss of glycosaminoglycan substitution. The conformation of the substitution site may therefore be important for xylosyltransferase recognition.     

Different usage of the glycosaminoglycan attachment sites of biglycan

Biglycan is a member of the small leucine-rich proteoglycan family. Its core protein comprises two chondroitin/dermatan sulfate attachment sites on serine 42 and serine 47, respectively, which are the fifth and tenth amino acid residues, respectively, after removal of the prepro peptide. Because the regulation of glycosaminoglycan chain assembly is not fully understood and because of the in vivo existence of monoglycanated biglycan, mutant core proteins were stably expressed in human 293 and Chinese hamster ovary cells in which i) either one or both serine residues were converted into alanine or threonine residues, ii) the number of acidic amino acids N-terminal of the respective serine residues was altered, and iii) a hexapeptide was inserted between the mutated site 1 and the unaltered site 2. Labeling experiments with [(35)S]sulfate and [(35)S]methionine indicated that serine 42 was almost fully used as the glycosaminoglycan attachment site regardless of whether site 2 was available or not for chain assembly. In contrast, substitution of site 2 was greatly influenced by the presence or absence of serine 42, although additional mutations demonstrated a direct influence of the amino acid sequence between the two sites. When site 2 was not substituted with a glycosaminoglycan chain, there was also no assembly of the linkage region. These results indicate that xylosyltransferase is the rate-limiting enzyme in glycosaminoglycan chain assembly and implicate a cooperative effect on the xylosyl transfer to site 2 by xylosylation of site 1, which probably becomes manifest before the removal of the propeptide. It is shown additionally that biglycan expressed in 293 cells may still contain the propeptide sequence and may carry heparan sulfate chains as well as sulfated N-linked oligosaccharides.     

Synthesis and structure of proteoglycan core protein

Studies of the structure and synthesis of cartilage proteoglycan core protein have been carried out. Deglycosylation of completed, secreted proteoglycan by HF-pyridine treatment yielded an intact homogeneous core protein of approximately 210,000 daltons, with a blocked amino-terminus. Greater than 95% of chondroitin sulfate chains and 80% of N- and O-linked oligosaccharides were removed by the procedure, which made the product an excellent xylosyltransferase acceptor. Little alteration of core protein structure occurred during the HF-pyridine treatment as shown by complete immunoreactivity with antiserums prepared against hyaluronidase-digested proteoglycan. In other studies, the initially synthesized precursor for proteoglycan core protein was found to be approximately 376,000 daltons and localized to the rough membrane fractions. This precursor already contained N-linked oligosaccharides, and was also able to accept xylose, thereby initiating chondroitin sulfate chains. The precursor was translocated intact in an energy-dependent manner to smooth membrane-Golgi fractions where further processing of high mannose type of oligosaccharides and addition of glycosaminoglycan chains occurred. The subcellular distribution pattern of the chondroitin sulfate-synthesizing enzymes corroborated the proposed topological modifications of the proteoglycan core protein precursor.     

Biosynthesis of chondroitin and heparan sulfate in chinese hamster ovary cells depends on xylosyltransferase II

Xylosyltransferase (XylT) catalyzes the initial enzymatic reaction in the glycosaminoglycan assembly pathway for proteoglycan biosynthesis. Its activity is thought to be rate-limiting. Two xylosyltransferases have been found using genomic analyses, and one of these, XylT1, has been shown to have xylosyltransferase activity. On the other hand, the less studied XylT2 in recombinant form lacks xylosyltransferase activity and has no known function. Wild-type Chinese hamster ovary cells express abundant Xylt2 mRNA levels and lack detectable Xylt1 mRNA levels. Analysis of a previously described Chinese hamster ovary cell xylosyltransferase mutant (psgA-745) shows that it harbors an Xylt2 nonsense mutation and fails to assemble glycosaminoglycans onto recombinant biglycan. Transfection of this cell line with a murine Xylt2 minigene results in the production of recombinant chondroitin sulfate-modified biglycan core protein and restoration of fibroblast growth factor binding to cell surface-associated heparan sulfate. Expression analyses on 10 different human transformed cell lines detect exclusive XYLT2 expression in two and co-expression of XYLT1 and XYLT2 in the others but at disparate ratios where XYLT2 expression is greater than XYLT1 in most cell lines. These results indicate that XylT2 has a significant role in proteoglycan biosynthesis and that cell type may control which family member is utilized.     

Initiation of the decorin glycosaminoglycan chain in the endoplasmic reticulum-Golgi intermediate compartment

We have transiently expressed decorin with a C-terminal KDEL endoplasmic reticulum retention signal peptide in COS-7 cells to study initiation of galactosaminoglycan synthesis in the endoplasmic reticulum-Golgi intermediate compartment. All decorin-KDEL molecules were substituted with N-linked oligosaccharides sensitive to endoglycosidase H, indicating that the core protein was located proximal to the medial-Golgi. O-Linked glycosylation was only initiated in a minor fraction of the molecules. The O-linked saccharides were characterized by gel filtration after stepwise degradations using chondroitin ABC/AC-I lyases, beta1-3-glycuronidase, beta-galactosidase, and alkaline phosphatase. The major O-linked saccharide was the linkage region pentasaccharide GalNAcbeta1-4GlcUAbeta1-3Galbeta1-3Galbeta1-4Xyl-2-phosphate, demonstrating initiation of chondroitin synthesis in the endoplasmic reticulum-Golgi intermediate compartment. In the presence of brefeldin A, partial elongation of a chondroitin chain took place, indicating retrieval of polymerases but not of sulfotransferases.     

Biosynthesis and properties of a further member of the small chondroitin/dermatan sulfate proteoglycan family

Human osteosarcoma cells express a 78-kDa proteoglycan core protein to which an asparagine-bound oligosaccharide, O-glycosidically linked oligosaccharides and probably only a single chondroitin 6-sulfate chain of 29-kDa are bound. Prior to O-glycosylation, the N-glycosylated core protein exhibits a mass of 83 kDa. Upon digestion of the secreted proteoglycan with chondroitin ABC lyase a mature core protein with an apparent molecular mass of 106 kDa is obtained. Smaller amounts of core proteins of 101 and 115 kDa can be detected occasionally. The glycosaminoglycan composition and the relative molecular mass of the glycosaminoglycan chain distinguish this proteoglycan, tentatively named proteoglycan 100 (PG-100), from biglycan (small proteoglycan I) and decorin (small proteoglycan II) which are also expressed by osteosarcoma cells. An antiserum against PG-100 shows partial cross-reactivity with decorin, but in contrast to the latter proteoglycan it does not bind to type I collagen fibrils. PG-100 is not a unique product of osteosarcoma cells. It has also been found in the secretions of human skin fibroblasts.     

Chinese hamster ovary cell mutants defective in glycosaminoglycan assembly and glucuronosyltransferase I

The proteoglycans of animal cells typically contain one or more heparan sulfate or chondroitin sulfate chains. These glycosaminoglycans assemble on a tetrasaccharide primer, -GlcAbeta1, 3Galbeta1,3Galbeta1,4Xylbeta-O-, attached to specific serine residues in the core protein. Studies of Chinese hamster ovary cell mutants defective in the first or second enzymes of the pathway (xylosyltransferase and galactosyltransferase I) show that the assembly of the primer occurs by sequential transfer of single monosaccharide residues from the corresponding high energy nucleotide sugar donor to the non-reducing end of the growing chain. In order to study the other reactions involved in linkage tetrasaccharide assembly, we have devised a powerful selection method based on induced resistance to a mitotoxin composed of basic fibroblast growth factor-saporin. One class of mutants does not incorporate 35SO4 and [6-3H]GlcN into glycosaminoglycan chains. Incubation of these cells with naphthol-beta-D-xyloside (Xylbeta-O-Np) resulted in accumulation of linkage region intermediates containing 1 or 2 mol of galactose (Galbeta1, 4Xylbeta-O-Np and Galbeta1, 3Galbeta1, 4Xylbeta-O-Np) and sialic acid (Siaalpha2,3Galbeta1, 3Galbeta1, 4Xylbeta-O-Np) but not any GlcA-containing oligosaccharides. Extracts of the mutants completely lacked UDP-glucuronic acid:Galbeta1,3Gal-R glucuronosyltransferase (GlcAT-I) activity, as measured by the transfer of GlcA from UDP-GlcA to Galbeta1,3Galbeta-O-naphthalenemethanol (<0.2 versus 3.6 pmol/min/mg). The mutation most likely lies in the structural gene encoding GlcAT-I since transfection of the mutant with a cDNA for GlcAT-I completely restored enzyme activity and glycosaminoglycan synthesis. These findings suggest that a single GlcAT effects the biosynthesis of common linkage region of both heparan sulfate and chondroitin sulfate in Chinese hamster ovary cells.     

Site-specific mutagenesis defines the intracellular role of the asparagine-linked oligosaccharides of chorionic gonadotropin beta subunit

The beta subunit of human chorionic gonadotropin contains two asparagine (N)-linked oligosaccharides. To examine the structural and functional roles of these oligosaccharide units in vivo, we constructed mutant genes containing alterations in either the asparagine or threonine codons of the two glycosylation consensus sequences and inserted them into a eukaryotic expression vector. Wild-type and mutant CG beta proteins were expressed in Chinese hamster ovary cells alone or in the presence of native alpha subunit. Pulse-chase analysis of the beta-expressing clones showed that absence of the second N-linked sugar but not the first slows secretion 1.6-1.8-fold; absence of both N-linked units slows secretion 2-2.4-fold. Analysis of dimer clones reveals that greater than 80% of the native and glycosylation mutant CG beta subunits are secreted as dimer. However, pulse-chase analysis of these clones also reveals that the mutants completely devoid of N-linked sugars but not the single site mutants are slow to assemble with the alpha subunit. Thus, in vivo the two N-linked oligosaccharides of CG beta are critical for efficient secretion and assembly with the alpha subunit and are likely important for proper folding of the CG beta subunit.     

Identification of a high molecular weight macrophage colony-stimulating factor as a glycosaminoglycan-containing species.

Chinese hamster ovary cells transfected with a 4.0-kilobase macrophage colony-stimulating factor (M-CSF) cDNA express two different M-CSF species; one has an apparent molecular weight of 85,000 and is identified as a homodimer of a 43-kDa subunit, and the other has an indeterminate structure greater than 200 kDa. In this study, we investigated the structure of the high molecular weight M-CSF by immunochemical procedures. The high molecular weight M-CSF was easily purified, since it bound tightly to DEAE-Sephacel and eluted at a characteristically high salt concentration. The high molecular weight M-CSF migrated as a diffuse band of over than 200,000 on nonreducing sodium dodecyl sulfate-polyacrylamide gels. Analysis of the same samples under reducing conditions revealed that the larger species consisted of a heteromer of the 43- and 150-200-kDa M-CSF subunits. Digestion of the 150-200-kDa M-CSF subunit with chondroitinase, which degrades the chondroitin sulfate glycosaminoglycan chain, yielded a 100 kDa band. This species was secreted instead of 150-200-kDa species when the cells were cultured in the presence of beta-D-xyloside, which inhibits the elongation of the chondroitin sulfate glycosaminoglycan chain in proteoglycans, providing additional evidence for the existence of a chondroitin sulfate chain in the 150-200-kDa M-CSF subunit. Removal of O- and N-linked carbohydrate from the 150-200-kDa subunit yielded a polypeptide chain with a larger molecular mass (approximately 45 kDa) than that of the 43-kDa subunit (approximately 25 kDa). Collectively, these results indicate that the 150-200-kDa M-CSF subunit is a proteoglycan with a core protein that may be an alternatively processed form of M-CSF.     

Estradiol beta-D-xyloside, an efficient primer for heparan sulfate biosynthesis

Animal cells utilize beta-D-xylosides as primers for glycosaminoglycan synthesis. However, most xylosides preferentially stimulate chondroitin sulfate synthesis and only weakly prime heparan sulfate synthesis. To test if the structure of the aglycone determines the type of glycosaminoglycan made, the priming activity of methyl, n-octyl, p-nitrophenyl, 4-methylumbelliferyl, trans,trans-farnesyl, cholesteryl, and estradiol beta-D-xylosides was compared. Their potency was tested in pgsA-745 cells, a Chinese hamster ovary cell mutant unable to initiate glycosaminoglycan synthesis due to a defect in xylosyltransferase. All of the xylosides stimulated chondroitin sulfate synthesis in the mutant, but only estradiol beta-D-xyloside primed heparan sulfate synthesis efficiently. When incubated with 30 microM estradiol beta-D-xyloside, mutant cells made about 3-fold more glycosaminoglycan than untreated wild-type cells and as much as 50% was heparan sulfate. Estradiol beta-D-xyloside also induced heparan sulfate synthesis in cycloheximide-treated wild-type Chinese hamster ovary cells, bovine aortic endothelial cells, baby hamster kidney cells, and Balb/c 3T3 fibroblasts. In addition to stimulating heparan sulfate synthesis, low concentrations of estradiol beta-D-xyloside inhibited the formation of endogenous heparan sulfate proteoglycans.     

The relationship of N-linked glycosylation and heavy chain-binding protein association with the secretion of glycoproteins

The relationship of N-linked glycosylation and association with heavy chain binding protein (BiP) to the secretion of Factor VIII (FVIII), von Willebrand Factor (vWF), and tissue plasminogen activator (tPA) was studied in Chinese hamster ovary (CHO) cells. FVIII has a heavily glycosylated region containing 20 clustered potential N-linked glycosylation sites. A significant proportion of FVIII was detected in a stable complex with BiP and not secreted. Deletion of the heavily glycosylated region resulted in reduced association with BiP and more efficient secretion. Tunicamycin treatment of cells producing this deleted form of FVIII resulted in stable association of unglycosylated FVIII with BiP and inhibition of efficient secretion. vWF contains 17 potential N-linked glycosylation sites scattered throughout the molecule. vWF was transiently associated with BiP and efficiently secreted demonstrating that CHO cells are competent to secrete a highly glycosylated protein. tPA, which has three utilized N-linked glycosylation sites, exhibited low level association with BiP and was efficiently secreted. Disruption of N-linked glycosylation of tPA by either site-directed mutagenesis or tunicamycin treatment resulted in reduced levels of secretion and increased association with BiP. This effect was enhanced by high levels of tPA expression. The glycosylation state and extent of association with BiP could be correlated with secretion efficiency.     

Post-translational addition of chondroitin sulfate glycosaminoglycans. Role of N-linked oligosaccharide addition, trimming, and processing

A melanoma proteoglycan model system has been used to examine the role of core protein asparagine-linked (N-linked) oligosaccharides in the transport and assembly of proteoglycan molecules. The use of agents which block discrete steps in the trimming and processing of core oligosaccharides (castanospermine, 1-deoxynojirimycin, N-methyldeoxynojirimycin, 1-deoxymannojirimycin, and swainsonine) demonstrates that removal of glucose residues from the N-linked oligosaccharides is required for the cell surface expression of a melanoma proteoglycan core protein and for the conversion of the core protein to a chondroitin sulfate proteoglycan. However, complete maturation of the oligosaccharides to a "complex" form is not required for these events. Treatment of M21 human melanoma cells with the glucosidase inhibitors castanospermine, 1-deoxynojirimycin, or N-methyldeoxynojirimycin results in a dose-dependent inhibition of glycosaminoglycan (GAG) addition to the melanoma antigen recognized by monoclonal antibody 9.2.27. In contrast, treatment with the mannosidase inhibitors 1-deoxymannojirimycin and swainsonine does not effect GAG addition. Identical results are obtained when the major histocompatibility complex class II antigen gamma chain proteoglycan is examined in inhibitor-treated melanoma and B-lymphoblastoid cells. These data, in conjunction with the known effects of the glucosidase and mannosidase inhibitors on the transport and secretion of other glycoproteins support the hypothesis that the addition, trimming, and processing of N-linked oligosaccharides is involved in the transport of certain proteoglycan core proteins to the site of GAG addition and to the cell surface.     

Sequence Analysis and Domain Motifs in the Porcine Skin Decorin Glycosaminoglycan Chain

Decorin proteoglycan is comprised of a core protein containing a single O-linked dermatan sulfate/chondroitin sulfate glycosaminoglycan (GAG) chain. Although the sequence of the decorin core protein is determined by the gene encoding its structure, the structure of its GAG chain is determined in the Golgi. The recent application of modern MS to bikunin, a far simpler chondroitin sulfate proteoglycans, suggests that it has a single or small number of defined sequences. On this basis, a similar approach to sequence the decorin of porcine skin much larger and more structurally complex dermatan sulfate/chondroitin sulfate GAG chain was undertaken. This approach resulted in information on the consistency/variability of its linkage region at the reducing end of the GAG chain, its iduronic acid-rich domain, glucuronic acid-rich domain, and non-reducing end. A general motif for the porcine skin decorin GAG chain was established. A single small decorin GAG chain was sequenced using MS/MS analysis. The data obtained in the study suggest that the decorin GAG chain has a small or a limited number of sequences.     

Uncoupling of chondroitin sulfate glycosaminoglycan synthesis by brefeldin A

Brefeldin A has dramatic, well-documented, effects on the structural and functional organization of the Golgi complex. We have examined the effects of brefeldin A (BFA) on the Golgi-localized synthesis and addition of chondroitin sulfate glycosaminoglycan carbohydrate side chains. BFA caused a dose-dependent inhibition of chondroitin sulfate glycosaminoglycan elongation and sulfation onto the core proteins of the melanoma-associated proteoglycan and the major histocompatibility complex class II-associated invariant chain. In the presence of BFA, the melanoma proteoglycan core protein was retained in the ER but still acquired complex, sialylated, N-linked oligosaccharides, as measured by digestion with endoglycosidase H and neuraminidase. The initiation of glycosaminoglycan synthesis was not affected by BFA, as shown by the incorporation of [6-3H]galactose into a protein-carbohydrate linkage region that was sensitive to beta-elimination. The ability of cells to use an exogenous acceptor, p-nitrophenyl-beta-D-xyloside, to elongate and sulfate core protein-free glycosaminoglycans, was completely inhibited by BFA. The effects of BFA were completely reversible in the absence of new protein synthesis. These experiments indicate that BFA effectively uncouples chondroitin sulfate glycosaminoglycan synthesis by segregating initiation reactions from elongation and sulfation events. Our findings support the proposal that glycosaminoglycan elongation and sulfation reactions are associated with the trans-Golgi network, a BFA-resistant, Golgi subcompartment.     

Isolation and characterization of bovine gingival proteoglycans versican and decorin.

1. We have isolated, chemically and immunologically characterized versican and decorin from bovine gingiva. 2. Versican was of large molecular weight and the molecular size of the core protein was estimated to be greater than 200 kDa. 3. The glycosaminoglycan chains were susceptible to chondroitinase ABC and N-linked oligosaccharides were present on the protein core of the molecule. 4. Immunological studies provided evidence that a hyaluronic acid binding region was present in the core protein of versican. 5. The overall structure was similar to that of versican isolated from bovine sclera. 6. Decorin had a molecular weight of 102 kDa and its glycosaminoglycan chain was completely digested by specific glycosidases. 7. The partially deglycosylated core protein had a molecular weight of 55 kDa and N-linked oligosaccharides were present on the molecule.     

Biosynthesis of small proteoglycan II (decorin) by chondrocytes and evidence for a procore protein

We have studied the biosynthesis of cartilage dermatan sulfate proteoglycan II (DS-PGII) (decorin) using in vitro translation of mRNA to determine the size of the primary gene product and by radiolabeling the protein in the presence of tunicamycin to inhibit the addition of Asn-linked oligosaccharides. Pulse-chase experiments were performed to examine post-translational processing and secretion. Inhibitors of oligosaccharide processing were used to determine whether DS-PGII molecules containing partially processed oligosaccharides could become proteoglycans and be secreted. Cell-free translation of sucrose gradient-fractionated RNA and subsequent immunoprecipitation of the core protein confirmed that the functional translated mRNA is in the size range of the two mRNA species observed by hybridization of chondrocyte RNA with a bone PGII cloned probe and that the translation product is a single protein with an apparent molecular mass of 42 kDa. Digestion of the intact proteoglycan (average molecular mass = 103 kDa) with chondroitinase ABC or AC results in an approximately 48-49-kDa product. Chondrocytes treated with tunicamycin to inhibit Asn-linked oligosaccharide addition synthesize and secrete a glycosaminoglycan (GAG)-substituted proteoglycan (average molecular mass = 86 kDa), yielding a 42-kDa core protein after chondroitinase ABC digestion, showing that Asn-linked oligosaccharides are not required for the addition of GAG chains or secretion. Following a short pulse (10 min) of [3H]leucine, three glycosylated forms of the DS-PGII core protein were observed, one of which is likely to be the precursor form of PGII predicted by the implied protein sequence of both bovine and human cDNA clones. Following the apparent cleavage of the propeptide, GAG-substituted intracellular core protein is detectable. Susceptibility to endoglycosidase H indicates that approximately one-third of the secreted core protein contains exclusively complex-type Asn-linked oligosaccharides and approximately two-thirds contain high mannose as well as complex-type oligosaccharides. Secreted DS-PGII appears to be fully substituted with three Asn-linked oligosaccharide chains. Inhibitors of oligosaccharide processing, however, permitted secretion of GAG-substituted DS-PGII that was fully (three chains) or incompletely (one or two chains) substituted with partially processed Asn-linked carbohydrate chains. By comparison of chondrocyte DS-PGII with fibroblast DS-PGII, we conclude that the addition and processing of Asn-linked carbohydrate chains are directed by the amino acid sequence of the core protein. The results reported here also suggest that the addition of xylose, the initial step in GAG chain synthesis, occurs early in biosynthesis and is determined by the primary amino acid sequence of the core protein.(ABSTRACT TRUNCATED AT 400 WORDS)     

Biosynthesis of proteodermatan sulfate in cultured human fibroblasts

Biosynthesis and secretion of proteodermatan sulfate produced by cultured human skin fibroblasts were investigated employing immunological procedures. During an incubation period of 10 min in the presence of [3H]leucine, two core protein forms of Mr = 46,000 and 44,000, respectively, were synthesized. They were converted to mature proteodermatan sulfate with a half-time of approximately 12 min. Fifty per cent of total mature proteodermatan sulfate were found in the culture medium after a 35-min chase. Six to eight per cent remained associated with the cell layer after a chase of 6 h. In the presence of tunicamycin, fibroblasts synthesized a single core protein of Mr = 38,000 that was converted to mature proteodermatan sulfate and secreted with similar kinetics as the N-glycosylated species. Subtle differences in the molecular size of core proteins were noted when cell-associated and secreted proteodermatan sulfate were degraded with chondroitin ABC lyase, but core proteins free of N-linked oligosaccharides were identical. Labeling with [3H]mannose revealed that secreted proteodermatan sulfate contains two or three complex-type or two complex-type and one high-mannose-type N-linked oligosaccharide chains. The N-glycans are bound to a 21-kDa fragment of the core protein. After incubation in the presence of [3H]glucosamine, the [3H]galactosamine/[3H]glucosamine ratio was 3.76 and 3.30 for secreted and cell-associated proteodermatan sulfate, respectively. Evidence for the presence of O-linked oligosaccharides could not be obtained. Small amounts of core protein free of dermatan sulfate chains were secreted when the cultures were treated with p-nitrophenyl-beta-D-xyloside.     

Glycosylation and secretion of human tissue plasminogen activator in recombinant baculovirus-infected insect cells.

Cell lines established from the lepidopteran insect Spodoptera frugiperda (fall armyworm; Sf9) are used routinely as hosts for the expression of foreign proteins by recombinant baculovirus vectors. We have examined the pathway of protein glycosylation and secretion in these cells, using human tissue plasminogen activator (t-PA) as a model. t-PA expressed in Sf9 cells was both N glycosylated and secreted. At least a subset of the N-linked oligosaccharides in extracellular t-PA was resistant to endo-beta-N-acetyl-D-glucosaminidase H, which removes immature, high-mannose-type oligosaccharides. This refutes the general conclusion from previous studies that Sf9 cells cannot process immature N-linked oligosaccharides to an endo-beta-N-acetyl-D-glucosaminidase H-resistant form. A nonglycosylated t-PA precursor was not detected in Sf9 cells, even with very short pulse-labeling times. This suggests that the mammalian signal sequence of t-PA is efficiently recognized in Sf9 cells and that it can mediate rapid translocation across the membrane of the rough endoplasmic reticulum, where cotranslational N glycosylation takes place. However, t-PA was secreted rather slowly, with a half-time of about 1.6 h. Thus, a rate-limiting step(s) in secretion occurs subsequent to translocation and N glycosylation of the t-PA polypeptide. Treatment of Sf9 cells with tunicamycin, but not with inhibitors of oligosaccharide processing, prevented the appearance of t-PA in the extracellular medium. This suggests that N glycosylation per se, but not processing of the N-linked oligosaccharides, is required directly or indirectly in baculovirus-infected Sf9 cells for the secretion of t-PA. Finally, the relative efficiency of secretion decreased dramatically with time of infection, suggesting that the Sf9 host cell secretory pathway is compromised during the later stages of baculovirus infection.