Regulation of N-Glycosylation LONG TERM EFFECT OF CYCLIC AMP MEDIATES ENHANCED SYNTHESIS OF THE DOLICHOL PYROPHOSPHATE CORE OLIGOSACCHARIDE:’’

The influence of 8-bromo-CAMP on N-glycosylation in JEG-3 choriocarcinoma cells was investigated using the octanoyl-tripeptide (OTP; N-octanoyl-asparagyl-1z51- tyrosyl-threonine amide) as glycosyl acceptor. In cells pretreated with 8-bromo-CAMP (2.5 nM to 1 mM), the amount of glycosylated OTP released into the culture medium was increased up to 35-fold. Under the same conditions, a 23-fold higher quantity of the glycoprotein hormone human chorionic gonadotropin was secreted. Preincubation of 10-90 min with 250 PM 8-bromo-CAMP caused only a 2-fold increase of the total amount of gly- cosylated OTP, whereas it was approximately 20-fold higher when the pretreatment was extended to 40 h. This strongly suggests involvement of gene activation rather than CAMP-mediated phosphorylation. The spe- cific activity of the oligosaccharyltransferase, as well as the mRNA levels of ribophorin I and I1 (presumptive subunits of the enzyme), remained unchanged. In pulse-chase experiments, [3H]mannose incorporation into dolichol-linked Glc3Man9(GlcNAc)2 was up to 20-fold higher in cells pretreated with 8-bromo-CAMP (250 PM, 40 h). The radioactivity was chased from the lipid-linked oligosaccharide activity of enzymes that control the synthesis of the lipid-linked core oligosaccharide. Our investigations show clearly that the markedly elevated amount of lipid-linked core oligosaccharide and the flux through this pool in 8-bromo-CAMP-treated choriocarcinoma cells is responsible for the increased N-glycosylation capacity.

The influence of 8-bromo-CAMP on N-glycosylation in JEG-3 choriocarcinoma cells was investigated using the octanoyl-tripeptide (OTP; N-octanoyl-asparagyl-1z51tyrosyl-threonine amide) as glycosyl acceptor. In cells pretreated with 8-bromo-CAMP (2.5 n M to 1 mM), the amount of glycosylated OTP released into the culture medium was increased up to 35-fold. Under the same conditions, a 23-fold higher quantity of the glycoprotein hormone human chorionic gonadotropin was secreted. Preincubation of 10-90 min with 250 PM 8-bromo-CAMP caused only a 2-fold increase of the total amount of glycosylated OTP, whereas it was approximately 20-fold higher when the pretreatment was extended to 40 h. This strongly suggests involvement of gene activation rather than CAMP-mediated phosphorylation. The specific activity of the oligosaccharyltransferase, as well as the mRNA levels of ribophorin I and I1 (presumptive subunits of the enzyme), remained unchanged. In pulsechase experiments, [3H]mannose incorporation into dolichol-linked Glc3Man9(GlcNAc)2 was up to 20-fold higher in cells pretreated with 8-bromo-CAMP (250 PM, 40 h). The radioactivity was chased from the lipid-linked oligosaccharide pool and shifted to the glycoprotein fraction 10 times more rapidly in the pretreated cells. The flux of [3H]mannose through the dolichol phosphate mannose pool was only slightly affected by the 8-bromo-CAMP pretreatment. Our investigations show that the oligosaccharyltransferase activity in JEG-3 cells is not rate-limiting. N-Glycosylation seems to be controlled by the amount of lipid-linked core oligosaccharide. The size of the lipid-linked core oligosaccharide pool, as well as the flux through, is markedly increased by pretreatment with 8-bromo-CAMP.
Glycosylation of proteins in the endoplasmic reticulum (ER)l is a very complex pathway, which has been elucidated to a far extent in the last years (1)(2)(3)(4)(5). It involves a multi-step synthesis of a lipid-linked oligosaccharide precursor (Glc3Man9-(GlcNAc),, core oligosaccharide). This oligosaccharide is coupled to dolichol pyrophosphate and by this linkage anchored '?This work was supported in part by a grant from the Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg (to W. E. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked "aduertisefact. to the ER membrane facing the luminal site. The core oligosaccharide part is transferred as a whole by the oligosaccharyltransferase to an asparagine residue of the Asn-X-Ser/Thr consensus sequence of a protein. Little is known about the regulation of this pathway. Various steps of the biosynthesis of the dolichol pyrophosphate oligosaccharide (Dol-P-P-oligosaccharidei precursor and the transfer to the acceptor protein were considered to be committed reactions concerning enzyme activities and availability of substrate (2, 6, 7). Beyond these endogenous levels of regulation, there are also hormones influencing the N-glycosylation of proteins, e.g. P-adrenergic agonists stimulate glycoprotein synthesis in rat parotid acinar cells (8,9) and lipoprotein lipase in rat heart cells (10). The glycosylation of thyroglobulin in porcine thyroid cells is under the control of thyrotropin (11). Androgens stimulate N-glycosylation in rat epididymis ( 1 2 ) and estrogens in mouse uteri (13,14). In the presence of dexamethasone incorporation of 13H]mannose into dolichol phosphate mannose (Dol-P-Man) was increased (15). Some of these hormonal effects are mediated by CAMP, and it was shown that cAMP itself or membrane permeable cAMP derivatives, like dibutyryl-and 8-bromo-CAMP, have the same effect (16). The steps of the reactions that are the target for that type of regulation are unclear or were discussed controversially. The present paper describes completely different effects of cAMP on N-glycosylation and analyzes the mechanisms of the cAMP action. We show that cAMP increases the amount of Dol-P-P-oligosaccharide available for transfer to the octanoyl-Asn-'""I-Tyr-Thr-NH, model peptide and cellular proteins to be glycosylated.

EXPERIMENTAL PROCEDURES
Cell Culture4EG-3, BeWo, and JAR cells were obtained from the American Type Culture Collection (Rockville, MD). JEG-3 and JAR cells were maintained in monolayer cultures in Dulbecco's modified Eagle's medium (DMEM; Sigma Chemical Co., Deisenhofen, Germany) containing 10% fetal calf serum (Biochrom KG, Berlin, Germany). BeWo cells were grown in Ham's F-12 with 15% fetal calf serum. Culture medium was changed every 2 days. The media were supplemented with sodium bicarbonate (3.7 gfliteri, penicillin (100 units/mli, and streptomycin (100 pg/ml; Biochrom KG, Berlin, Germany). Confluent cell monolayers in 25-cm2 plastic flasks were used. Unless otherwise stated, treatment of cells with 8-bromo-CAMP was started 40 h before incubation with the radioactive tripeptide (see below) or labeling with ['Hlmannose by replacement of the culture medium with medium containing 250 phf 8-bromo-CAMP (sodium salt, Sigma). Control cells obtained fresh culture medium without any additives at the same time as the 8-bromo-CAMP treated cultures. The numbers of cells in the cultures were counted as described (17). Cellular protein content was determined after washing of the cells with phosphate-buffered saline (PBS) as described (18).
Iodination of the Acceptor Peptide-Octanoyl-tripeptide (OTP; octanoyl-Asn-Tyr-Thr-NH,) was a generous gift of Dr. Felix Wieland (University of Heidelberg, Germany). 10 nmol of OTP were iodinated with 1 mCi of NalZ6I (Amersham Buchler, Braunschweig, Germany) and purified as previously'described (19). About 60% of the label was incorporated. The specific activity of the iodinated tripeptide was approxi-8659 mately 1 x lo8 cpdnmol. The purity of the product was characterized by thin-layer chromatography as described (19). lz51-Labeled OTP used in the experiments is designated OTP in the text below.
Incubation of Cells with N-Octanoyl-tripeptide-Monolayers were washed once with buffer B (25 mM Tris-HC1 (pH 7.4), 137 m~ NaCl, 5 mM KCl, 0.7 m~ Na,HPO,). Then 1.8 ml of serum-free medium, containing the OTP (2-5 pCi/ml), was added. 8-bromo-CAMP was supplemented if indicated. After an incubation of 2 h a t 37 "C, the media were removed and kept on ice. The cells were first rinsed with 3 ml of ice-cold DMEM containing 10% fetal calf serum and then with 5 ml of ice-cold buffer B. Finally the cells were lysed and removed from the plate with 1.8 ml of buffer B containing 0.5% (w/v) Triton X-100. The media were supplemented with Triton X-100 to a final concentration of 0.5% (w/v). Cell extracts and media were boiled for 3 min and stored a t -30 "C until further treatment. After thawing, the samples were centrifuged (3 min, 10,000 x g). The supernatants were applied to concanavalin A-Sepharose (Sigma) columns. Concanavalin A-bound glycosylated OTP was washed and eluted as described (19). The carbohydrate part of eluted OTP was further characterized in some experiments by digestion with endo-P-N-acetylglucosaminidase H or peptide glycohydrolase F and separation by thin-layer chromatography as described (19). Oligosaccharyltransferase Assay-Cell extracts were prepared by sonication (Branson sonicator model 250, power setting 3, microtip, for 10 s at 20% charge) in 50 m~ Tris-HC1 (pH 7.51, containing 8.55% sucrose, 25 mM KCl, 5 m~ P-mercaptoethanol, 10 m~ MgCl,, and 2 mM MnCl,. Activity of oligosaccharyltransferase was assayed using OTP a s oligosaccharide acceptor and endogenous Dol-P-P-oligosaccharide as substrate (20, 21). The reaction mixture with a total volume of 20 pl contained 5 p~ OTP (6,000-20,000 cpdpmol), 2-20 pg of cell protein, 250 mM sucrose, 50 mM Tris-HC1 (pH 7.5), 25 mM KC1, 5 m~ P-mercaptoethanol, 10 m~ MgCl,, and 2 mM MnC1,. Assays were incubated for 10 min at 37 "C. Then reaction was terminated by boiling of the samples for 3 min. After adding 750 p1 ofbufferA(l0 m~ Tris-HC1 (pH 7.4),0.15 M NaCl, 1 mM CaCl,, 1 m~ MnCl,, 0.5% (w/v) Triton X-1001, the glycosylated OTP was analyzed by concanavalin A-Sepharose chromatography and characterized by thin-layer chromatography (19).
Western Blot Analysis-Cells were sonicated (see above), and in a first step the cellular debris were removed by centrifugation for 15 min at 5,000 x g. The membranes sedimented at 100,000 x g (90 min, 4 "C) were suspended in a buffer containing 8 g/liter NaCl, 0.2 g h t e r KCl, 1.15 gAiter Na2HP04, 0.2 ghiter KH,P04, 1% (w/v) Triton X-100, 0.5% sodium deoxycholate, 0.1% (w/v) SDS, 2 m~ phenylmethanesulfonyl fluoride, 10 m~ N-ethylmaleimide, and 20 mM EDTA. Samples (corresponding to protein obtained from lo5 cells) were separated by SDSpolyacrylamide gel electrophoresis (12% w/v gels) in a Mini-Protean I1 electrophoresis cell (Bio-Rad, Munich, Germany) (22). Blotting was performed in a Trans-Blot apparatus (Bio-Rad) using nitrocellulose membranes (Schleicher & Schiill, Dassel, Germany). Anti-ribophorin I antiserum from rabbits, generously provided by Dr. Gert Kreibich (Medical Center, University of New York, New York) was used a s first antibody and anti-rabbit immunoglobulin G alkaline phosphatase conjugate (Sigma) a s detection antibody.
Northern Blot Analysis-Cells were washed twice with PBS, and total RNA was isolated using the acid phenol-guanidinium thiocyanate procedure (23). RNA was dissolved in HzO and concentration was determined by measuring optical density at 260 nm. After electrophoretic separation on a n 1% (w/v) agarose gel containing 2.2 M formaldehyde, the RNA was transferred by vacuum blotting to a Hybond-N membrane (Amersham-Buchler, Braunschweig, Germany). Then, the RNA was W cross-linked to the filters a t 1 J/cm2 (Transilluminator, Appligene, Heidelberg, Germany) and sequentially hybridized with ribophorin I and I1 cDNAs, kindly provided by Dr. Gert Kreibich, followed by glyceraldehyde phosphate dehydrogenase, hCG-a, and hCG-p cDNAs labeled with [a-32PldCTP (3,000 Ci/mmol purchased from Amersham Buchler) using random-primed DNA labeling (Boehringer, Mannheim, Germany). The membranes were washed four times (10 min a t room temperature) with 2 x SSC (1 x SSC consists of 0.015 M sodium citrate, pH 7.5, and 0.15 M NaC1) containing 0.1% (w/v) SDS, and once for 15 min at 65 "C with 0.1 x SSC containing 0.1% (w/v) SDS and exposed at -80 "C to a Fuji RX film.
Carbohydrate Labeling, Extraction of Dol-P-Man, Lipid-linked Oligosaccharides, and Glycoprotein-Cultures were washed twice with PBS before addition of 1.5 ml of labeling medium (glucose-free DMEM supplemented with 10 pCi/ml ~-[2,6-~H]mannose, 52 Ci/mmol, Amersham Buchler). The cells were incubated for the indicated time at 37 "C, the radioactive media were removed, and labeling was stopped by washing with ice-cold PBS. After layering with 0.6 ml of ice-cold water, cells were scraped from the plate. The suspension was transferred to a tube containing 1.2 ml of CHCl,:CH,OH (1:l) and mixed thoroughly. The following extraction was carried out as previously described (24). Briefly, after a short centrifugation the CHC13-phase was collected. The remaining aqueous phase and the interphase were extracted again with 0.6 ml of CHC13. Both CHCl, phases were combined and washed once with CHC1,:CH30H:Hz0 (3:48:47) to get the Dol-P-Man fraction. The interphase and the aqueous layer were washed three times with 1 ml of 50% (dv) CH30H. The pellet was extracted with CHC1,:CH30H:H20 (10:10:3) to obtain the lipid-linked oligosaccharides. The remaining pellet (glycoprotein fraction) was washed with 10% (w/v) trichloroacetic acid and water and dissolved in Protosol (NEN, Dreieich, Germany). The radioactivity was counted in a liquid scintillation counter (Tricarb 2450, Packard Instrument, FrankfudM, Germany). Secreted glycoprotein was analyzed by mixing equal volumes of culture medium and 20% (w/v) trichloroacetic acid. After incubation for 1 h on ice, the precipitates were collected by centrifugation. The pellets were washed twice with ice-cold 10% (w/v) trichloroacetic acid and treated further as indicated above. In the chase interval of pulse-chase experiments the pulse medium was removed, the monolayers were washed twice with culture medium supplemented with 1 m~ n-mannose and incubated for the indicated time with 1.5 ml of this culture medium.
Gel Filtration-The 3H-labeled lipid-linked oligosaccharides extracted were dried in a vacuum concentrator (Bachhofer, Reutlingen, Germany) and dissolved in 0.2 ml of 1-propanol and 0.4 ml of 0.01 M HC1. The samples were heated for 20 min a t 100 "C to release the free oligosaccharides (25). The samples were analyzed by passage through a Bio-Gel P4 column (<45 pm), equilibrated with 0.5% acetic acid (26). The column (0.9 x 200 cm) was calibrated with the following 3H-labeled compounds: MangGlcNAc, Man,GlcNAc, Man,GlcNAc, MaQGlcNAc, Man,GlcNAc, and 1~-[2,6-~HImannose, as well as unlabeled bovine serum albumin. The radioactive oligosaccharide standards were a generous gift of Dr. R. Geyer (University of Giessen, Giessen, Germany). The elution volumes of the oligosaccharide standards were plotted according to their glucose units (27). The eluted sample peaks were collected and digested with endo-P-N-acetylglucosaminidase H (Boehringer; 1 milliunitfsample, 100 mM citrate buffer, pH 5.5, 6 h ) to release one GlcNAc residue and thus to allow a direct comparison with the oligosaccharide standards in a second run on the Bio-Gel P4 column. In addition, the Glc3Man,(GlcNAc),-peak was also digested with a-glucosidase (type I11 from yeast, Sigma; 1 unitfsample in 100 mM phosphate buffer, pH 6.8, 16 h). hCG a n d Free Subunits-The concentrations of human chorionic gonadotropin (hCG) and the free subunits were determined with a n enzyme-linked immunosorbent assay (ELISA) as described (28) U). In the medium, glycosylated OTP increased almost with a linear function for at least 6-7 h. After about 6 h of incubation with OTP, the concentration of glycosylated OTP in control cells was only 20% of the level found in 8-bromo-CAMP treated cells and had still not attained an intracellular plateau (Fig. 1B). Pretreatment of the cells with various concentrations of 8-bromo-CAMP increased the formation of glycosylated OTP in a dose-dependent manner (Fig. 2). The most evident stimulation of N-glycosylation by 8-bromo-CAMP in relation to control cells was achieved in the micromolar range, but nanomolar concentrations of 8-bromo-CAMP also exerted a certain effect. The result of a typical experiment is shown in Table I. In the untreated control cells, about 50-75% of the total amount of OTP formed within 2 h accumulated intracellularly. This ratio was gradually reversed with raising 8-bromo-CAMP concentrations. The part of glycosylated OTP released into the culture medium in presence of 1 mM 8-bromo-CAMP rose about 36-fold in comparison to the control. The total amount of glycosylated OTP contained in the cells and released into the medium was 11-fold higher in 8-bromo-CAMP-pretreated cells than in the controls. JEG-3 cells secrete mainly hCG and the free hCG subunits. The biosynthesis and secretion of this glycoprotein hormone are stimulated by 8-bromo-CAMP. The secreted amounts of hCG and glycosylated OTP were increased to a very similar extent in response to various 8-bromo-CAMP concentrations (Fig. 3). This confirms that the stimulatory effect of 8-bromo-CAMP on the glycosylation of the model peptide reflects the effect on the biosynthesis of a glycoprotein.
Time Course of the Action of 8-Bromo-CAMP-An initial hint on the mechanism of the 8-bromo-CAMP-mediated stimulation of N-glycosylation should be obtained from kinetic considerations. It was expected that if CAMP-dependent phosphorylation of glycosylation components is involved, treatment of the cells with 8-bromo-CAMP should increase the intracellular CAMP concentrations within a short time, in the range of minutes or, at most, of a few hours. The other possibility would encounter a CAMP-mediated effect on the gene activity of proteins involved in the regulation of N-glycosylation, which usually takes more time. In the cultures preincubated with 250 VM 8-bromo-CAMP for 0-2 h with 8-bromo-cAMP, a biphasic response curve with one or two peaks of about 2-fold increase in comparison to control cells was observed in all experiments, followed by a decrease of glycosylation capacity between 2 and 5 h regardless of the presence of 8-bromo-CAMP. Prolongation of the pretreatment with 8-bromo-CAMP beyond 20 2 4 h caused a massive increase in the glycosylation capacity, which exceeded the effects of short time preincubation by more than 1 order of magnitude (Fig. 4B). The quantitative differences of these two responses, as well as their time courses, suggest that the influence of 8-bromo-CAMP on N-glycosylation described here is mediated mainly by gene activation rather than by acute CAMP-dependent phosphorylation reactions. Experiments with JAR or BeWo choriocarcinoma cell lines gave similar results (not shown).
Oligosaccharyltransferase-An obvious candidate for the regulation of N-glycosylation is the oligosaccharyltransferase. The activity of the oligosaccharyltransferase, using OTP as substrate, was not increased when the JEG-3 cells were preincubated for 40 h with 8-bromo-CAMP concentrations in the range of 2.5-1,000 p~ (not shown). In contrary, within this range, the higher concentrations seemed to cause a small decrease (by 30%) of the specific activity of the enzyme. When the cells were pretreated with 250 p~ 8-bromo-CAMP for 0-40 h, the activity of the oligosaccharyltransferase was not changed in contrast to the marked increase of glycosylation capacity under these conditions (Fig. 4B). In Western blots carried out with antibodies against ribophorin I, which represents one of the subunits of the oligosaccharyltransferase complex, no difference of the specific staining of ribophorin I of control and 8-bromo-CAMP treated cultures was observed (not shown).
Since ribophorins I and 11 have been identified as parts of the oligosaccharyltransferase complex (211, we also investigated the ribophorin I and I1 mRNA levels in 8-bromo-CAMP-treated and in untreated JEG-3 cells. The ribophorin I and I1 mRNA levels showed only a very small increase in response to 8-bromo-cAMP, if any (not shown). This corresponds very well to the low effects on the amount of ribophorin I in the Western blot of 8-bromo-CAMP treated cells, as well as to the oligosaccharyltransferase activity (see above). In contrary to the ribophorin I and I1 mRNA levels, the hCG-a and p mRNAs showed a marked increase in the 8-bromo-CAMP-pretreated JEG-3 cells. The ratios of the mRNAs hCG-dglyceraldehyde-3-phosphate dehydrogenase in 8-bromo-CAMP-pretreated cells were 10-fold (50 PM 8-bromo-CAMP) up to 30-fold (250 8-bromo-CAMP) higher than the controls (not shown). The corresponding values of hCG-@glyceraldehyde-3-phosphate dehydrogenase mRNAs showed a 20-50-fold increase in comparison to the control cultures.
The Dolichol Pyrophosphate-Oligosaccharide Pool-The lack of a distinct increase of the oligosaccharyltransferase activity in response to treatment of the cells with 8-bromo-CAMP led us investigate the dolichol pyrophosphate-oligosaccharide pool as a parameter possibly committed to the glycosylation capacity of 8-bromo-CAMP treated cells. Control and 8-bromo-CAMPtreated cells were incubated up to 50 min with [3H]mannose. Afterward the fractions containing Dol-P-Man and Dol-P-P-oligosaccharide were isolated by extraction with organic solvents. The efficacy of the extraction method was controlled in a mock experiment. 3H-Labeled Dol-P-Man (2,500 cpm) and 3Hlabeled Dol-P-P-oligosaccharide (34,500 cpm) obtained in a previous experiment were added to the cell suspension and extracted as described under "Experimental Procedures." In the case of 3H-labeled Dol-P-Man, in the fraction that should contain that compound, 68.2 2 5.2% radioactivity was recovered; in the fraction of the lipid-linked oligosaccharides, 13.5 2 1.9% was found (n = 3). When radioactive Dol-P-P-oligosaccharide was given to the cell suspension before the extraction, 0.2 2 0.04% of the radioactivity was found in the Dol-P-Man fraction and 99.8 2 13.5% in the Dol-P-P-oligosaccharide fraction.
The total incorporation of radioactive mannose into the various fractions was markedly higher in 8-bromo-CAMP-treated cells (Fig. 5). In response to the 8-bromo-CAMP concentrations, [3Hlmannose was incorporated into the glycoprotein and the Dol-P-P-oligosaccharide fractions in a very congruent manner (Fig. 5, compare B and C ) . In contrast, an incorporation of [3Hlmannose into the Dol-P-Man fraction more than 1 order of magnitude smaller was observed (Fig. 5 A ) . As a function of time, the incorporation of [3Hlmannose into this compound gave comparable results (Fig. 6A). The incorporation of [3Hlmannose into the glycoprotein and the Dol-P-P-oligosaccharide fractions was accelerated in 8-bromo-CAMP-pretreated cultures in comparison to the corresponding fraction of controls (Fig. 6, B and C).
The extraction procedure of the Dol-P-P-oligosaccharide fraction provides no separation or identification of the different types of lipid-linked oligosaccharides. Therefore, the I3H1mannose-labeled Dol-P-P-oligosaccharide fractions were submitted to mild acid hydrolysis to release the oligosaccharide moieties. The oligosaccharides were separated by gel chromatography (Fig. 7). The bulk of the radioactive material eluted at a position that corresponds to the Glc3Man9(GlcNAc)2 oligosaccharide. Precursor oligosaccharides with lower molecular weights were present only in very minor quantities in that time scale (see Fig. 7, insets).
The amount of G l~~M a n~( G 1 c N A c )~ oligosaccharide formed in 8-bromo-CAMP-treated cells is much higher than in control cells. This is evident particularly at the 10-min value, where it is approximately 20-fold higher. Later, this value dropped to approximately 6-fold. This suggests a more rapid consumption of the lipid-linked oligosaccharide by transfer of the carbohydrate residue to glycosyl acceptor sequences. This led us investigate the turnover of the Dol-P-P-oligosaccharide fraction in pulse-chase experiments. The results of a typical experiment are shown in Fig. 8. During the chase period, the labeled Dol-P-P-oligosaccharide disappeared approximately 10 times faster in the 8-bromo-CAMP-pretreated cells than in the controls most probably caused by a more rapid transfer of the oligosaccharide

Increase of glycosylation of octanoyl-tripeptide (OTP) induced by pretreatment of JEG-3 cultures with 8-bromo-CAMP
JEG-3 monolayer cultures were pretreated (40 h) with 8-bromo-cAMP prior to addition of OTP. After 2 h, glycosylation of OTP was terminated as described under "Experimental Procedures." Total amount comprises glycosylated OTP in medium and cells.  to protein acceptors (Fig. 8, A and D). In addition, the much higher Dol-P-P-oligosaccharide concentration in the 8-bromo-CAMP-treated cells is again obvious. This indicates that the Dol-P-P-oligosaccharide pool, as well as the flux through this pool, is much larger than in the control cells. The incorporation of L3HImannose into intracellular glycoproteins reached a saturation rapidly in controls, as well as in 8-bromo-CAMP-treated cells, but differed clearly in the concentration levels (Fig. 8C). The 8-bromo-CAMP-pretreated cells secreted approximately 15 times more labeled glycoproteins in that interval than the control cells (Fig. 8D). This corresponds very well with the higher flux through the Dol-P-P-oligosaccharide pool (Fig. EL4 1. In contrast to the Dol-P-P-oligosaccharide fraction, the incorporation of the label into Dol-P-Man was 10-80 times smaller, the label was chased more slowly, and it showed no large difference between 8-bromo-CAMP-treated and control cells.

DISCUSSION
The choriocarcinoma cell lines used in this study respond to elevated CAMP levels with a 10-fold increase of hCG biosynthesis and more. The transcription of both hCG subunit genes, a and p, is regulated by CAMP responsive elements (29-32), and the subunit mRNAs are stabilized in the presence of 8-bromo-CAMP (33). The carbohydrate moiety of hCG has a great impact on the subunit interactions (34-37) and on the stimulation of signal transduction by the hormone (38). Recently, we have shown that N-glycosylation of hCG in placenta tissue was accelerated and occurred with higher efficacy in presence of 8-bromo-CAMP (39,40). The marked CAMP-mediated stimulation of hCG biosynthesis, as well as the relevance of the carbohydrate part for physical and biological properties of hCG, prompted us to address the question whether CAMP controls also the capacity of the N-glycosylation pathway in a coordinate manner with its effects on transcription and mRNA stability. We used the model peptide OTP to investigate this question, since it provides the possibility to measure N-glycosylation without any interference with the conformation of the glycosyl acceptor site. 8-bromo-CAMP stimulates N-glycosylation of the model peptide OTP and of the glycoprotein hormone hCG almost to the same extent and in the same dose range (Fig. 3). This strongly suggests that OTP is an appropriate model for these studies. In choriocarcinoma cells pretreated with 8-bromo-CAMP for 40 h the amount of glycosylated OTP increased almost linearly for several hours (Fig. 1) clearly depending on the concentration of 8-bromo-CAMP (Fig. 2) and the duration of the preincubation time (Fig. 4). There were several reports in literature that p-adrenergic stimulation or treatment with dibutyryl-CAMP augmented N-glycosylation (8-11, 16, 41). In all these cases, cellular CAMP levels were increased for several minutes up to a few hours. The effects were explained by CAMP-dependent hosphorylations of enzymes involved in the glycosylation pa P hway. Treatment with dibutyryl-CAMP or P-adrenergic agonists or treatment of microsomes with the catalytic subunit of CAMP-dependent protein kinase caused the activation of the G P mannose:dolichyl phosphate-0-0-D-mannosyltransferase, which catalyzes the synthesis of Dol-P-Man (16,41). This latter compound is substrate, as well as allosteric activator, in the pathway that leads to the synthesis of the Dol-P-P-oligosaccharide. We observed an approximately 2-fold augmentation of the glycosylation of OTP in response to short term pretreatment with 8-bromo-CAMP (Fig.  4A), which correlates quantitatively very well with the effects observed in the literature. The effect of 8-bromo-CAMP on Nglycosylation described here is totally different concerning its magnitude, the kinetics, and the site of action. It is observed exclusively after incubation of the cells with 8-bromo-CAMP for at least 16 h and longer. It exceeds the short term effect (Fig.  4A) by more than 1 order of magnitude (Fig. 4B 1. This suggests a completely distinct mechanism, probably involving gene activation. It is known that the choriocarcinoma cells pass through several steps of differentiation and that hCG, a major product of these cells, is formed predominantly by differentiated cells (42,43). On the other hand, 8-bromo-CAMP initiates differentiation (44). Whether the 8-bromo-CAMP-induced stimulation of N-glycosylation is the result of a differentiation process or whether it is a necessary precondition for differentiation is unclear at present.
Little is known about the regulation of N-glycosylation. A number of steps were considered to be responsible for the regulation of this pathway, e.g. the first step in the pathway, the formation of N-acetylglucosaminylpyrophosphoryl dolichol by the N-acetylglucosamine-1-phosphate transferase (2). The activity of this enzyme is regulated by several factors including Dol-P-Man (2,45,46). The oligosaccharyltransferase may be considered as an additional obvious point of regulation of Nglycosylation. In regenerating rat liver the synthesis of complete carbohydrate chains of glycoproteins are decreased due to a diminished activity of this enzyme (47). The CAMP-mediated stimulation of N-glycosylation of porcine thyroglobulin by thyrotropin was claimed to be caused by an increase of the oligo- saccharyltransferase activity (11). The 8-bromo-CAMP induced stimulation of N-glycosylation was clearly not caused by this enzyme for the following reasons. First, the specific activity of the oligosaccharyltransferase of JEG-3 cells was not increased in response to pretreatment with 8-bromo-CAMP. Second, pretreatment with 8-bromo-CAMP caused only a small increase of the ribophorin I and TI mRNAs, which is in accordance with the results at the protein level. Our investigations show that the oligosaccharyltransferase activity is obviously sufficient for the increased demands for N-glycosylation in 8-bromo-CAMP-pretreated cells.
An additional reason for the enlarged glycosylation capacity of 8-bromo-CAMP-pretreated cells could be a n increase in size of the rough endoplasmic reticulum. It was shown that the amounts of ribophorins I and I1 are closely correlated to the size of rough endoplasmic reticulum (48) and were found in the rough endoplasmic reticulum in a 1:l molar ratio with bound ribosomes (49). The lack of substantial changes in the contents of ribophorin I and 11 mFtNAs and ribophorin I protein argue against the possibility that the 8-bromo-CAMP effects described here are caused by a change in the amount of rough ER.
Long term pretreatment of JEG-3 cells clearly increased the steady-state level of the lipid-linked Glc3Mans(GlcNAc)z fraction (Figs. 5B and 7). The incorporation of l3HImannose into this compound in response to the 8-bromo-CAMP concentration, as well as the kinetics of formation, matches very well with the labeling characteristics of the glycoprotein fraction (Figs. 5 and  6, B and C ) . This suggests that the availability of lipid-linked Glc3Mans(GlcNAc)z oligosaccharide limits the N-glycosylation of OTP and of glycoproteins synthesized by the cells. This hypothesis is supported further by the results of the pulse-chase experiments, which have shown that in 8-bromo-CAMP-pretreated cells the flux through the Dol-P-P-oligosaccharide pool is markedly accelerated (Fig. 8).
It is reasonable to assume that the increase in protein biosynthesis in 8-bromo-CAMP-stimulated cells leads to a higher transfer rate of the core oligosaccharide to acceptor proteins coupled with the enhanced formation of the free lipid anchor.

Time (rnin)
This might cause an accelerated synthesis of new lipid-linked core oligosaccharide since the amount of dolichol phosphate was suggested to be rate-limiting in this pathway (6,50-53). In consequence, the velocity of N-glycosylation could increase simply by mass action driven by the accelerated protein synthesis. There is probably a close tie between protein synthesis and formation of the core oligosaccharide. An inhibition of protein synthesis also caused an inhibition of the incorporation of mannose into the lipid-linked oligosaccharides, although synthesis of Dol-P-Man was only slightly affected (24,54). Our results show a clear increase of mannose incorporation into lipidlinked oligosaccharides and only a slight enhancement of the Dol-P-Man concentration when the protein synthesis was stimulated. The kinetics of the 8-bromo-CAMP effect on the transcription of hCG subunit genes in choriocarcinoma cells and on N-glycosylation argue against the possibility that the elevated protein concentration triggers the stimulation of Nglycosylation. The CAMP-mediated activation of transcription of many proteins is achieved within approximately 1 h, as is the case for the insulin precursor (554 growth hormone (56), tyrosine aminotransferase (571, and hCG-a subunit (31, 58). The transcription of the hCG-P subunit genes is regulated by a more complex mechanism. In this case, the 8-bromo-CAMPinduced stimulation of transcription is retarded and probably requires ongoing protein biosynthesis (31) in contrast to a-subunit gene transcription. Anyhow, hCG secretion rises simultaneously with the increase in mRNA (31) within a few hours after the beginning of the stimulation with 8-bromo-CAMP. The CAMP-responsive stimulation of the N-glycosylation, however, requires a much longer treatment (Fig. 4).
On the basis of these considerations, it is much more likely that 8-bromo-CAMP induces an up-regulation of the gene activity of enzymes that control the synthesis of the lipid-linked core oligosaccharide. Our investigations show clearly that the markedly elevated amount of lipid-linked core oligosaccharide and the flux through this pool in 8-bromo-CAMP-treated choriocarcinoma cells is responsible for the increased N-glycosylation capacity.