Control of N-Linked Carbohydrate Unit Synthesis in Thyroid Endoplasmic Reticulum b,y Membrane Organization and Dolichyl Phosphate Availability*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY (c? 1986 by The American Society of Biological Chemists Inc Vol. 261 No. 31 Issue of November 5. pp Printed in U. S. A. Control of NLinked Carbohydrate Unit Synthesis in Thyroid Endoplasmic Reticulum by Membrane Organization and Dolichyl Phosphate Availability* (Received for publication April ) Mary Jane Spiro and Robert G. Spirol From the Departments of Biological Chemistry and Medicine Harvard Medical School and the Elliott P. Josh Research Laboratory Boston Massachus&s 2215 Thyroid rough endoplasmic reticulum (ER) has been shown to contain a highly organized multienzyme system capable of carrying out the Nglycosylation of newly synthesized proteins. These reactions were studied in isolated ER vesicles and found to be controlled to a large extent by the availability of a key substrate dolichyl phosphate (DolP) as well as by the amount of endogenous polypeptide acceptor present. Although in intact vesicles UDPGlc was utilized in an efficient manner to form DolPGlc and glucosylated oligosac charidelipid after disruption of vesicle integrity even with low concentrations of Triton X1 the coupling of DolPGlc formation to lipidlinked oligosaccharide assembly and subsequent Nglycosylation was substantially impaired Increased incubation tem * This work was supported by Grant AM from the National Institutes of Health and Grant 82R335 from the Juvenile Diabetes Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. This paper is dedicated to Luis F. Leloir on the occasion of his 8th birthday. $ To whom correspondence should be addressed: Elliott P. Joslin Laboratory One Joslin Place Boston MA The formation of Nglycosidically linked carbohydrate units is to a large measure carried out in the membranes of the ER which contain the enzymes involved in the assembly of the lipidbound oligosaccharide its transfer to nascent polypeptide and its initial processing by the removal of outer sugar residues (13). It is likely that the numerous reactions leading to Nglycosylation are controlled by the disposition of the enzymes and substrates within the ER membranes and that these interrelationships can best be explored in intact vesicles in which the physiological orientations are preserved. Thyroid ER vesicles represent an attractive system for investigating the regulation of this multienzyme system as they carry out lipidsaccharide synthesis and Nglycosylation in an integrated manner (4). In the present study we have peratures also resulted in a decreased effectiveness of probed the organization of these vesicles with the help of glucose transfer from DolPGlc to lipidoligosaccha radiolabeled sugar nucleotides (UDPGlc GDPMan UDPride presumably because of a decline in the extent of GlcNAc) by modifying the state of the membranes with graded structural organization of the ER membranes. amounts of detergent and altered temperatures as well as by The limited availability of endogenous DolP was circumventing potential control points through the addition demonstrated by the pronounced stimulation in DolP of DolP and a peptide containing the AsnXThr sequence. Glc formation resulting from the addition of this lipid These experiments indicated that the reactions leading from acceptor to Tritondisrupted ER membranes as well as sugar nucleotides to Nlinked carbohydrate units by way of by its generation in intact vesicles. The latter was dolichollinked mono and oligosaccharides occur in a tightly accomplished by stimulating recycling of endogenous coupled manner which depends on membrane integrity for DolP through the addition of a peptide (TyrAsnLeu optimal function. Furthermore the activity of the vesicle ThrSerVal) which is an Nglycosylation substrate. system was found to be restricted by limited DolP and The inhibition of DolPGlc synthesis from UDPGlc polypeptide acceptor pools. Our observations indicated that observed in the presence of elevated levels of GDP all three sugar nucleotide (UDPGlc GDPMan UDP Man which could be relieved in Tritondisrupted or GlcNAc): DolP glycosyltransferases appear to have access to intact ER vesicles by the addition or generation rethe same membrane associated DolP pool as their activities spectively of DolP is considered to be the result of a competing requirement for DolP by the mannosylcould be similarly stimulated by increasing the endogenous transferase. Moreover GTP by selectively inhibiting level of this substrate through the addition of a peptide the mannosyltransferase prevented the decrease of acceptor for the oligosaccharyltransferase. Moreover inhibi DolPGlc formation caused by GDPMan. Since addi tion of UDPGlc utilization brought about by GDPMan in tion of the acceptor peptide to intact vesicles stimulated both intact and disrupted vesicles could be relieved by increas DolPPGlcNAc as well as DolPGlc and DolPMan ing the level of DolP. synthesis it would appear that a pool of DolP available in common to all three enzymes responsible for doli EXPERIMENTAL PROCEDURES chollinked monosaccharide synthesis exists in the ER Preparation of ER VesiclesCalf thyroid glands brought to the membranes. laboratory on ice from a local slaughter house were used to prepare rough ER vesicles by the procedures previously described (4). The membranes obtained by sucrose density gradient centrifugation were washed twice with.25 M sucrose 5 mm Tris acetate ph 7. (1 X g for 6 min) and were then resuspended in this buffer. The protein cont.ent of the membranes was determined by the method of Lowry et al. (5) using bovine serum albumin as a standard. Incubation ofer VesiclesThe incubations of vesicles with radio The abbreviations used are: ER endoplasmic reticulum; DolP dolichyl phosphate; DolPP dolichyl pyrophosphate; DolPPOligo glucosylated dolichol pyrophosphatelinked oligosaccharide; endo H endonacetylglucosaminidase H.

2 14726 Control of NLinked Carbohydrate Synthesis in ER Vesicles labeled sugar nucleotides were carried in a manner similar to that with a wick of Whatman No. 3 paper clamped to the top of the thin previously reported (4). Unless otherwise specified the membranes (18 pg protein) together with 1 pci of sugar nucleotide were incubated at 27 "C for 1 min in 8 pl of 5 mm Tris acetate ph 7. layer plate (11). The individual components after detection by fluorography were scraped from the plates and quantitated by scintillation counting. Acidlabile saccharides from the lower phase of the buffer containing 25 mm potassium acetate 1 mm magnesium ace chloroform/methanol/uffer (3:2:1) extract were identified by chrotate.1 M sucrose 1 mm ATP.5 mm creatine phosphate and 12 pg creatine phosphokinase. The "standard assay" consisted of incubamatography on cellulosecoated plates (.1 mm Merck) in a pyridine/ ethyl acetate/water/acetic acid (5:5:3:1) system (4). tions with 1 pci UDP[3H]glucose (1.45 p ~ under ) the aforemen Radioactivity MeasurementsLiquid scintillation counting was tioned conditions. When Triton X1 was included in the incubations it was added to the tubes in an ice bath and allowed to react performed in Ultrafluor (National Diagnostics) with a Beckman LS 75 instrument. Radioactive components on thin layer plates and with the membranes at this temperature for 5 min prior to the electrophoresis gels were detected by fluorography at 7 "C after addition of the radiolabeled substrate. The radiolabeled sugar nucleotides employed included UDP[63H] glucose (8.6 Ci/mmol) from Amersham and GDP[343H]mannose (33 Ci/mmol) as well as UDPNa~ety1[6~H]glucosamine (6.6 Ci/ spraying with a scintillation cocktail containing 2methylnaphthalene (4) or treating with ENHANCE (New England Nuclear) respectively. For all fluorography XOmat AR film (Eastman Kodak) was employed. mmol) from New England Nuclear while unlabeled GDPMan and UDPGlc which were added in certain incubations were purchased from Behring Diagnostics. RESULTS The commercially obtained DolP (Sigma) which was used as an Integrated Utilization of Glucose by Thyroid ER Vesiclesexogenous acceptor was standardized against tritiumlabeled DolP We have previously reported that intact thyroid ER vesicles of known specific activity (.6 pci/pmol) (prepared by phosphorylaincubated with UDP[3H]Glc bring about a rapid incorporation (6) of a mixture of unlabeled dolichol (Sigma) and [l3h]dolichol (New England Nuclear) and purified by thin layer chromatography) tion of glucose into a lower phase lipid characterized as Dolby comparing the doseresponse curve for the enhancement of Dol PGlc (4). This transfer which has a K of.54 PM is maximal PMan formation from GDP[3H]Man in a thyroid microsomal sys at 1 min under the conditions of our assay and is tightly tem (7). coupled to reactions leading to the formation of oligosaccha Measurement of Lipid and Proteinlinked SaccharidesThe in ridelipid and the Nglycosylation of endogenous protein (Fig. cubations were terminated by the addition of 2 ml of methanol and 1). Indeed these two products reached plateaus at lower UDPlipid extractions were performed with chloroform/methanol/uffer Glc concentrations than the DolPGlc itself (Fig. l) although (3:2:1) and chloroform/methanol/water (113) in a manner previously described to yield dolichollinked monosaccharide and oligo the latter component is known to be responsible for the saccharide fractions respectively (8). The proteinbound oligosaccha glucosylation of the lipidlinked oligosaccharide (12) which rides were obtained as glycopeptides after Pronase digestion of the in turn is transferred to protein. The amount of radioactivity delipidated membrane proteins; Dowex 5 and Dowex 1 columns were recovered in the lipidlinked oligosaccharide which is almost employed to separate the glycopeptides from glucanderived neutral exclusively present as Glc3MangGlcNAc2PPDol in this syscomponents (4). Radioactivity incorporated into the various fractions tem (4) was close to three times that observed in the DolPwas determined by scintillation counting. Glc (Fig. 1) under the usual conditions of assay (1.45 PM Incubation with Exogenous PeptidesTo evaluate the effect of exogenous peptide on the transferases TyrAsnLeuThrSerVal (pep UDPGlc). The radioactivity associated with the protein was tide A) or TyrAsnProThrSerVal (peptide B) supplied by Bio endo H releasable and occurred as revealed by thin layer search Laboratories was incubated at a concentration of 15 pm with chromatography in the form of Glc3MangGlcNAc as well as the ER vesicles for 15 min at 27 "C before the addition of radiolabeled GlczMangGlcNAc and GlcManGlcNAc indicating that some sugar nucleotide. After a further 1min incubation at 27 "C the processing had taken place. samples were fractionated in the usual manner to provide an upper Effect of Vesicle Integrity on Coupling of LipidSacchnride phase from the chloroform/methanol/water (3:2:1) extract from Formation and Nglycosylation of ProteinTo determine the which the glycosylated peptide was recovered after adsorption on a column of Dowex 5X2 (H' form) 24 mesh and elution with 1.5 N NHOH (9). Radioactivity incorporated into the peptides was determined by scintillation counting after removal of the solvent by DolPPOligo lyophilization. Endo H Digestion+Oligosaccharides were released from the glycosylated exogenous hexapeptides or from the Pronaseprepared glycopeptides from endogenous vesicle proteins by treatment with endo n W H from Streptomyces griseus (Miles Laboratories) as previously reported (1); digestion of oligosaccharidelipids with this enzyme was carried out in.1% Triton X1 as recently described (9). After passage of the digests through coupled columns of Dowex 5 (H' form) and Dowex 1 (acetate form) the liberated oligosaccharides were submitted to thin layer chromatography. Evaluation of Vesicle IntegrityAfter radiolabeling vesicle components by incubation with UDP[3H]Glc (5 pci/tube) under standard conditions five tubes were pooled diluted with icecold.25 M sucrose and pelleted by centrifugation at 1 X g for 6 min at 2 "C. Subsequent to a wash with the.25 M sucrose the radiolabeled vesicles (125 +g protein) were incubated for 1 min at 27 "C in 25 pl of.1 M sucrose 5 mm Tris acetate buffer ph 7. containing 25 I 1 I mm potassium acetate 5 mm magnesium acetate and varying con I 2 3 centrations of Triton X1 The incubations were terminated by UDPGLUCOSE placing the tubes in ice and after centrifugation (1 X g for 6 IrM) min 2 "C) the radioactivity in the solubilized material was determined FIG. 1. Effect of UDPGlc concentration on the transfer of by scintillation counting. The residual pellets were dissolved in 4% glucose to endogenous lipid and protein acceptors by thyroid (w/v) sodium dodecyl sulfate and 5% 2mercaptoethanol (v/v) in 8 ER vesicles. Incubations were carried out with 1. pci UDPl3H] mm Tris/HCl ph 6.8 for measurement of radioactivity and exami glucose (1.45 p ~ as ) described under "Experimental Procedures" nation by polyacrylamide gel electrophoresis in a manner previously (standard assay) for 1 min at 27 "C with 18 pg of membrane protein. described (4). Radioactivity was measured in the lower phase after chloroform/ Thin Layer ChromatographyOligosaccharides were resolved by methanol/water (3:2:1) extraction (DolPGlc) in the chloroform/ chromatography in 1propanol/acetic acid/water (3:3:2) for 4 h on methanol/water (11:3) extract (DolPPOligo) and in the glycoplastic sheets precoated with silica gel 6 (.2mm thickness Merck) peptides after Pronase digestion (Protein).

3 Control of NLinked Carbohydrate Synthesis in ER Vesicles effect of vesicle disruption on the reactions leading to the N glycosylation of protein incubations with increasing amounts of Triton X1 were carried out (Fig. 2). Since it has been previously shown that permeabilized vesicles require exogenous lipid acceptor for DolPGlc formation (4) these assays were carried out in the presence of7 pm DolP to assure adequate synthesis of this first reaction product and thereby permit study of the subsequent transfer steps. A comparison of the intact (Fig. 1) and Tritontreated (Fig. 2) vesicles indicates that while there is an efficient transfer of glucose from DolPGlc to oligosaccharidelipid in the former situation this is no longer evident in the latter case; indeed in the presence of the detergent a 7fold increase in DolPGlc was required to achieve the maximal enhancement (1.7fold) of the oligosaccharidelipid. Furthermore the glycosylation of protein showed a decline with increasing concentrations of Triton despite the presence of elevated levels of oligosaccharidelipid donor (Fig. 2). These findings suggested that the coupling of the reactions involving the transfers of glucose from DolPGlc to oligosaccharidelipid and oligosaccharide to protein was substantially impaired by the action of Triton on vesicle integrity. This cannot be attributed to an adverse effect of Triton on the enzymes themselves as previous studies have shown that the DolPG1c:oligosaccharidelipid glucosyltransferase (12) and the oligosaccharidelipidpeptide oligosaccharyltransferase (13) are actually stimulated by this detergent in the concentration range used in the present study. Thin layer chromatography of endo Hreleased oligosaccharides indicated that the Triton had no qualitative effect on the carbohydrate units associated with the lipid or protein. To investigate the state of the ER membranes at various detergent concentrations the solubilization of membrane components from preradiolabeled vesicles was determined. Disassembly of the vesicles already evident at.25% Triton was essentially complete at.1% of the detergent (Fig. 3). Electrophoresis of the material remaining after the Triton treatment indicated by fluorographic examination that the glucosylated proteins and oligosaccharidelipids were solubilized in a nonselective fashion. Effect of Membrane Fluidity on LipidSaccharide SynthsisIn order to further examine the importance of membrane organization on the formation of DolPGlc and glucosylated TRITON (%) FIG. 3. Effect of Triton X1 concentration on the integrity of thyroid ER vesicles. After incubation of vesicles with UDP [3H]Gle followed by centrifugation and washing the solubilization of the radiolabeled membrane components was evaluated at various Triton concentrations as described under Experimental Procedures. > c I I O a a (3 s (C I I I I 1 I I DolPGlc M DolPPOligo U I 4.2 I p a L I I I I I I I I I/Tx IO3 (Kl FIG. 4. Effect of temperature on the formation of lipidsaccharides by thyroid ER vesicles. Incubations with UDP[3H] Glc were carried out at the indicated temperatures under the standard conditions described under Experimental Procedures. The radioactivity incorporated into DolPGlc and DolPPOligo is presented in the form of an Arrhenius plot as well as in a direct manner (inset) TRITON (161 FIG. 2. Effect of Triton X1 concentration on the formation of lipidsaccharides and the Nglycosylation of protein by thyroid ER vesicles in the presence of exogenous DolP. Incubations were carried out with UDP[3H]Glc under standard conditions as described under Experimental Procedures. Each tube also contained DolP (7 p ~ which ) was found on the basis of assays with [3H]DolP to be completely in solution at Triton concentrations above.25%; at.1% Triton the solubility was 75%. The incorporation of radioactivity is expressed relative to the values obtained in the absence of detergent (control) which were and 16 dpm for DolPGlc DolPP12 and Protein respectively. oligosaccharidelipid the effect of incubation temperature was determined. An Arrhenius plot of the data indicated that while DolPGlc synthesis responded to temperature in a linear manner up to 37 C oligosaccharidelipid glucosylation showed a transition point already at 22 C (Fig. 4). This suggested that greater mobility of the membrane components at higher temperature brought about an uncoupling of reactions involving the transfer of glucose which was comparable to that observed with increasing Triton concentrations (Fig. Z) although we cannot exclude the possibility that the individual enzymes may have different temperature responses. Effect of Exogenous DolP on LipidSaccharide Formation and NGlycosylation of ProteinThe addition of DolP to partially disassembled ER vesicles (.4% Triton) resulted in a linear increase in DolPGlc formation with a maximal stimulation at 8 gm (Fig. 5). Although DolPGlc synthesis responded in a sensitive manner to increasing DolP concentration the formation of oligosaccharidelipid again lagged behind and did not reach its maximum (2fold stimulation) until a 1fold increase in DolPGlc had been attained (Fig. 5).

4 14728 Control of NLinked Carbohydrate Synthesis in ER Vesicles 4 t GDPMan 1 z I DolPPOligo Protein DolP (pm) FIG. 5. Effect of DolP concentration on the formation of lipidsaccharides and the Nglycosylation of protein by thyroid ER vesicles. Incubations were carried out with UDP[3H]Glc under standard conditions as described under Experimental Procedures except that.4% Triton X1 and varying amounts of Dol P were included in each sample. The values plotted represent the radioactivity incorporated into DolPGlc DolPPOligo and endogenous Protein I 5 s 4 t a I GDPHon TRITON (X) FIG. 7. Effect of exogenous DolP on the inhibition by GDP Man of DolPGlc formation in thyroid ER at various Triton X1 concentrations. Incubations were performed with UDP [3H]Glc (1.45 NM) under standard conditions as described under Experimental Procedures with varying concentrations of Triton in the presence of unlabeled GDPMan (1 p ~ with ) and without Dol P (7 p ~ as ) well as without either of these additions. The percent inhibition of DolPGlc synthesis by GDPMan in the absence and presence of DolP is shown in the upper panel while in the lower panel DolPGlc formation is expressed as ratio to the value obtained in the absence of GDPMan DolP and Triton (control) which was 327 dpm. I I I I I I GDPMan (pm) FIG. 6. Effect of GDPMan on the utilization of UDPGlc by intact thyroid ER vesicles. Incubations were carried out under standard conditions as described under Experimental Procedures with UDP[3H]Glc (1.45 p ~ and ) varying amounts of unlabeled GDP Man. The values are expressed relative to those obtained in the absence of GDPMan (control) which were 28 and 57 dpm for DolPGk and DolPPOligo respectively. Competition of Glycosyltransferases for DolP PoolSince the formation of lipidmonosaccharide appears to be limited by the DolP content of the vesicle it was of interest to determine whether the ER UDPGlc and GDPMan:DolP glycosyltransferases compete for this lipid acceptor. When the vesicles were incubated with UDP[3H]Glc (1.45 p ~ in ) the presence of increasing concentrations of unlabeled GDP Man pronounced inhibition in the transfer of glucose to both DolP and to oligosaccharidelipid were observed (Fig. 6) with almost complete blockage evident at 1 p~ GDPMan. Although this inhibition by GDPMan of glucose transfer was also evident in the presence of various concentrations of Triton (Fig. 7) it could be overcome by the addition of Dol P; this relief which progressed with increasing concentration of the detergent became complete at.2% Triton (Fig. 7) at which complete disruption of the vesicle membranes occurs (Fig. 3). The inhibition imposed by GDPMan on DolPGlc synthesis could also be relieved by GTP (Fig. 8); since the addition of this nucleotide to the vesicle incubations selectively inhibited the GDPMan:DolP mannosyltransferase it apparently stimulated DolPGlc production by in effect mak c I 2 x W g IO 2 c $ 5 n n o IDolPMan DolP GIC GTP GDPMon + + FIG. 8. Effect of GTP on the inhibition by GDPMan on Dol PGlc formation in thyroid ER vesicles. Incubations with UDP [3H]Glc (1. pci 1.45 p ~ to ) measure DolPGlc formation were carried out under standard conditions as described under Experimental Pracedures in the presence of unlabeled GDPMan (5 p ~ ) and/or GTP (1 mm). For comparison DolPMan synthesis was measured by incubating the vesicles with GDP[3H]Man (.27 YCi.1 pm) in presence or absence of GTP (1 mm); similar inhibitions of DolPMan synthesis were observed at GTP concentrations of.125 mm and.25 mm. ing more of the endogenous DolP available to the glucosyltransferase. In contrast to the inhibitory effect of GDPMan on UDP Glc utilization the latter nucleotide did not interfere with

5 Control of NLinked Carbohydrate Synthesis in ER Vesicles DolPMan synthesis in ER vesicles; addition of increasing amounts of unlabeled UDPGlc (up to 1 PM) to incubations containing.38 PM GDP[3H]Man resulted in no discernible decrease in mannoseincorporation (data not shown). The absence of a mutual inhibition by the two sugar nucleotides inthe DolP: glycosyltransferase reactions is most likely attributable t o different affinitiesof the two enzymes for their substrates; indeed we find that thek for the sugar nucleotide of the GDPMan:DolP mannosyltransferase (.52 PM) in the intact vesicles is one order of magnitude lower than that of the UDPG1c:DolP glucosyltransferase (.54 p M ). Effect of Exogenous Peptides onlipidsaccharide Formation and NGlycosylationSince exogenous DolP can not be made available to the transferases of the ERwithout vesicle disruption we explored the limitation in glycosylation imposed by the level of this substrate by generating endogenous DolP through a stimulation of DolPPoligosaccharide utilization. This was accomplished by adding a membranepermeable Nglycosylation substrate namely TyrAsnLeuThrSerVal (peptide A) to the vesicles prior to carrying out the incubations with UDP[3H]Glc. The presence of peptide A substantially promoted Nglycosylation by the ER vesicles so that despite a progressive decrease in transfer of oligosaccharide to endogenous protein the total saccharide recovered in asparaginelinked carbohydrate (peptide A plus protein) was about 4fold greater than in the absence of this peptide acceptor (Fig. 9). Thin layer chromatography of the oligosaccharides released by endo H from the Nglycosylated peptideaindicated the presence of Glc3MangGlcNAc GlcpMansGlcNAc and GlclMansGlcNAc (Fig. 1). Since the oligosaccharidelipid formed by the ER vesicles is almost exclusively in the fully glucosylated form (4) processing of the transferredcarbohydrate units musthave taken place and indeed this seemed to be more rapid than for the endogenous protein (Fig. 1) FIG. 1. Thin layer chromatographic examination of endo Hreleased oligosaccharides obtained from glycosylated exogenous peptide and endogenous protein radiolabeled byincubations of thyroid ER vesicle with UDP[3H]Glc.The oligosaccharides released from the peptide fraction (Pept) after incubation in thepresence ( 1 ) or absence (2) of 15 pm peptide A (TyrAsnLeuThrSerVal) are compared with those obtained from endogenous pronase digested protein (Prot). The incubations of thyroid vesicles with UDP[3H]Glcwere carried out under standard conditions with a 15min preincubation with and without the exogenous peptide. The conditions for the thin layer chromatography are specified under Experimental Procedures and the components were visualized by fluorography. The migration of standard Glc3MansGlcNAc (G3) GlczMansGlcNAc(G2)and GlclMansGlcNAc ( G I )are indicated. The enhanced oligosaccharide transfer was accompanied by an almost complete depletion of oligosaccharidelipid and more than a %fold increase in the synthesis of DolPGlc (Fig. 9). These changes did not occur (Fig. 9) in the presence of TyrAsnProThrSerVal (peptide B) which differed from peptide A in containing proline in the place of leucine within the AsnXThr sequence an alteration which is known to make it unsuitable as an an acceptor for Nglycosylation (14). Under the conditions of increased oligosaccharide transfer resulting from the addition of peptide A enhanced synthesis of the otherlipidlinked monosaccharides namely DolPMan and DolPPGlcNAc was also brought about (Fig. 11); inthe incubations with GDP[3H]Man and UDP[3H]GlcNAc uni I labeled UDPGlc (1 PM) was added to permit completion of the oligosaccharidelipid and thereby assureefficient transfer P E P T I M (pm) to the peptide. The increases in the three Dollinked monofig. 9. Effect of various concentrations of exogenous pep saccharide products were similar in magnitude and were not tides on the formation of lipidsaccharidesand thetransfer of observed in thepresence of the prolinecontaining peptide B oligosaccharideto the peptides as well as endogenous protein by thyroid ER vesicles. After preincubation (15 min 27 C) with suggesting that a supply of endogenous DolP accessible to the indicated amounts of peptide A ()or peptide B (A) as described the three transferases was produced due to the enhanced under Experimental Procedures the vesicles were incubated with generation of DolPP from the stimulated Nglycosylation UDP[3H]Glc under standard conditions. The radioactivity trans reaction. ferred to lipidsaccharides (DolPGlc and DolPPOligo) Peptide Since the increased oligosaccharide transfer engendered by and Protein was determined by the procedures described under Ex addition of a peptide acceptor to ERvesicles appears to result perimental Procedures and the data are expressed relative to a in the formation of a larger DolP pool it should tend to control in which the vesicles were preincubated in the absence of exogenous peptide; the control values were DolPGlc 313 dpm; relieve the GDPMan inhibitionof DolPGlc formation from DolPPOligo 778 dpm; Peptide 4 dpm; and Protein 148 UDPGlc. To test this possibility we incubated the vesicles with UDP[3H]Glc in the presence of inhibiting amounts of dpm.

6 1473 Control of NLinked Carbohydrate Synthesis in ER Vesicles DolPOh DolPMan DolPPGlcNAc FIG. 11. Effect of exogenous peptides on the formation of Dollinked monosaccharides from several sugar nucleotides. Thyroid ER vesicles were incubated under standard conditions with 1 pci of 3Hlabeled sugar nucleotide (UDPGlc GDPMan or UDP GlcNAc) after a 15min preincubation at 27 C with 15 p~ TyrAsn LeuThrSerVal (Peptide A) or TyrAsnProThrSerVal (Peptide B) as well as in the absence of added peptide (Control). The amount of Dollinked monosaccharide formed was calculated from the radioactivity recovered in the lower phase of the chloroform/methanol/ water (3:2:1) extract + and the specific activity of the sugar nucleotide precursors (see Experimental Procedures ). 5 IO 2 GDPMan (pm) unlabeled GDPMan both with and without peptide A and found that in the presence of the peptide acceptor there was a stimulation of DolPGlc synthesis (Fig. 12); not unexpectedly this stimulation was inversely related to the concentration of the GDPMan. DISCUSSION As our understanding of the intermediates and enzymatic reactions required for the addition of asparaginelinked carbohydrate to newly formed peptide chains has enlarged interest has grown in the manner in which this process is controlled. Since protein Nglycosylation appears to occur at the luminal face of the ER concomitant with the passage of the newly synthesized peptide through the membrane (1517) while the initial steps of the pathway involving formation of dolichollinked monosaccharides occur cytosolically ( ) a high level of organization of the various enzymes and substrates must be required. Indeed our present studies indicate that the thyroid ER vesicles provide an integrated multienzyme system in which Nglycosylation of endogenous proteins from lipidsaccharide precursors takes place in an efficient manner. Disruption of the integrity of the vesicles by detergent led to an uncoupling of the reactions leading to protein glycosylation. Even in the presence of a small amount of Triton (.25%) insufficient to solubilize the membrane components the oligosaccharide transfer step was separated from the preceding reactions and transfer of glucose from DolPGlc to oligosaccharidelipid became much less effective. Since supplementation of the permeabilized vesicles with DolP stimulated DolPGlc production without bringing about a corresponding effect on oligosaccharidelipid synthesis or Nglycosylation it appeared that a tight membrane structural organization is required to assure the concerted action of the various transferases. A similar type of uncoupling effect on the transfer of glucose was observed at incubation temperatures above 22 C where less rigidity of membrane structure would be expected. A major locus of control for Nglycosylation has been postulated to be the level of DolP in ER membranes based on our findings with tissue slices (8) and those of others with cultured cells (19 2) that when incubations are carried out in the presence of inhibitors of protein synthesis the amount of oligosaccharidelipid remained constant presumably because the DolP required for the formation of new lipidlinked oligosaccharide was limiting. The modulation of glycosyltransferase activity which has been observed during organ development (21) or after hormone administration (22) has been attributed to elevated tissue levels of DolP and moreover addition of this lipid to pancreas slices has been found to increase the degree of glycosylation of ribonuclease (23). In the present investigation with thyroid ER vesicles the limitation of the DolP poolwas demonstrated in several ways including the stimulation of DolPGlc formation upon supplementation of permeabilized vesicles with exogenous DolP; the inability of the endogenous supply of DolP to support glucosyltransferase activity in the face of a competing requirement for this lipid acceptor by mannosyltransferase at elevated GDPMan concentrations; and the enhancement of lipidmonosaccharide synthesis from UDPGlc GDPMan and UDPGlcNAc as a result of vesicular DolP generation from DolPPoligosaccharide in the presence of exogenous peptide acceptor. Although addition of DolP to permeabilized ER vesicles FIG. 12. Effect ofan exogenous peptide acceptor on the inhibition by GDPMan of DolPGlc formation in thyroid ER vesicles. Incubations with UDP[3H]Glc (1 pci 1.45 p ~ to ) measure DolPGlc formation were carried out under standard conditions as described under Experimental Procedures in the presence of varying concentrations of unlabeled GDPMan after a 15min preincu brought about a substantial stimulation of DolPGlc synthebation at 27 C with or without peptide A acceptor (15 pm). sis (1fold at 15 PM DolP) this could only be achieved in an extensively disrupted membrane system. A much more effective stimulation of DolPGlc formation could be achieved in intact vesicles by promoting the recycling of endogenous DolP through the addition of a peptide acceptor for the oligosaccharyltransferase reaction. In accord with previous studies employing small peptide acceptors in oviduct microsomes (24) the hexapeptide (peptide A) which we used penetrated the intact ER vesicles to serve as a substrate for Nglycosylation and this permitted a pronounced enhancement of oligosaccharidelipid utilization beyond that which can occur with the limited endogenous polypeptides which are present in ER membranes. In order for DolP to be made available from the DolPP released during the oligosaccharyltransferase reaction cleav age by a DolPP phosphatase must occur; such enzymes have been found in a variety of cell types (2527) and indeed studies with liver microsomes employing the inhibitor bacitracin have

7 Control of NLinked Carbohydrate Synthesis in ER Vesicles suggested that a large portion of DolP which serves as substrate for glycosylation originates from the pyrophosphate derivative (26). The approximate amount of DolP which can be generated and recycled by the addition of saturating amounts of exogenous peptide can be calculated from the decrease in the radioactivity associated with oligosaccharide lipid (7 dpm; Fig. 1) and the specific activity of the UDPGlc employed (19 dpm/pmol) to be about 2.5 pmol per tube based on our previous finding that the molecules of the endogenous thyroid oligosaccharidelipid pool have an average of 1.5 sites per molecule available for attachment of additional glucose residues (28). Making this additional amount of DolP available by incubating with exogenous peptide acceptor resulted in about a 3.3 pmol stimulation of DolPGlc synthesis (Fig. 1) while in contrast 15p~ (12 pmol per tube) of exogenous DolP had to be added to a permeabilized ER vesicle to achieve a similar enhancement in DolPGlc formation (Fig. 4). The greater effectiveness (about 5fold) of endogenous DolP is presumably a function of its insertion into the membrane in a position appropriate for optimal glucosyltransferase action. Our data indicate that the inhibiting effect of GDPMan on DolPGlc synthesis from UDPGlc which has previously been observed by others (29 3) is a function of a limited endogenous DolP pool. This inhibition was observed in intact as well as in Tritondisassembled ER vesicles and in case of the latter situation could be relieved by the addition of Dol P. In the closed vesicles a reduction in the inhibiting effect of GDPMan could be brought about by the additional DolP generated in the presence of an exogenous Nglycosylation substrate. Moreover addition of GTP which inhibits the mannosyltransferase without affecting the transfer of glucose relieved the inhibition by GDPMan providing further evidence for the limitation imposed by a shared DolP pool and arguing against the possibility that the mannose nucleotide produces its effect by reacting directly with the glucosyltransferase as has been suggested (29). Indeed a study of UDP G1c:DolP glucosyltransferase which had been purified from a plasmacytoma cell line so as to remove mannosyltransferase activity indicated that GDPMan even at high concentrations had no inhibiting effect on the action of the enzyme (31). The lack of inhibition by UDPGlc on DolPMan synthesis from GDPMan which we noted suggests that the mannosyltransferase may be more favorably positioned in respect to the endogenous lipid acceptor pool and/or may have a greater synthetic capacity. Indeed physiologically one might anticipate such a situation as in most tissues the concentration of UDPGlc is many times that of GDPMan (32 33) and in thyroid we find that the UDPGlc level is approximately 25 pmol/g compared to 1.6 pmol/g for GDPMan.' Despite this difference in the level of the two sugar nucleotides similar amounts of DolPGlc and DolPMan would have to be produced by the ER membranes in order to assemble a complete lipidlinked oligosaccharide (13). The K values which we find for the two transferases in intact ER vesicles (UDP Glc =.54 pm; GDPMan =.52 p ~ are ) in keeping with the different concentrations of the two nucleotides in the tissue; other published reports of K values for Dol P:glucosyl ( ) and mannosyltransferases ( ) also indicate lower values for the latter enzyme. Our study indicates that the DolP made available by recycling of DolPP released by oligosaccharide transfer to protein becomes part of a common pool which is available to the three enzymes responsible for the synthesis of Dollinked monosaccharides (DolPGlc DolPMan and DolPP M. J. Spiro and R. G. Spiro unpublished observations. GlcNAc) a situation which apparently also applies to DolP formed through the dolichol kinase reaction (38). Since it is believed that oligosaccharide transfer to protein takes place on the luminal face of the ER (16 17) while Dollinked monosaccharide synthesis occurs on the cytoplasmic side ( ) the DolP generated by Nglycosylation and Dol PP phosphatase action must translocate before it can again be available at the outer membrane surface to react with the various DolP:glycosyltransferases. Although the mechanism of movement of DolP and its saccharide derivatives has not yet been defined and represents one of the more challenging aspects of transmembrane movement of polar molecules recent studies (39 4) indicate that DolP molecules by locally inducing a nonbilayer phase in membranes can produce a more fluid environment in which the translocations required for the assembly of the Nlinked carbohydrate units could take place. AcknowledgmentWe would like to thank Kathleen M. Gorski for her skillful assistance. REFERENCES 1. Hubbard S. C. and Ivatt R. J. (1981) Annu. Rev. Biochem Spiro R. G. and Spiro M. J. (1982) Philos. Trans. R. SOC. London Ser. B Kornfeld R. and Kornfeld S. (1985) Annu. Rev. Biochem Spiro M. J. and Spiro R. G. (1985) J. Bwl. Chem Lowry. H. Rosebrough N. J. Farr A. L. and Randall R. J. (1951) J. Biol. Chem Wedgwood J. F. Strominger J. L. and Warren C. D. (1974) J. Biol. Chem Adamany A. M. and Spiro R.G. (1975) J. Biol. Chem Spiro M. J. Spiro R. G. and Bhoyroo V. D. (1976) J. Bwl. Chem Chalifour R. J. and Spiro R. G. (1984) Arch. Biochem. Biophys Annumula K. R. and Spiro R. G. (1983) J. Bwl. Chem Godelaine D. Spiro M. J. and Spiro R. G. (1981) J. Biol. Chem Murphy L. A. and Spiro R. G. (1981) J. Biol. Chem Spiro M. J. Spiro R. G. and Bhoyroo V. D. (1979) J. Bwl. Chem Bause E. and Hettkamp H. (1979) FEBS Lett Hanover J. A. and Lennarz W. J. (1982) J. Biol. Chem Snider M. D. and Robbins P. W. (1982) J. Biol. Chem Snider M. D. Sultzman L. A. and Robbins P. W. (198) Cell Snider M. D. and Rogers. C. (1984) Cell Schmitt J. W. and Elbein A. D. (1979) J. Biol. Chem Hubbard S. C. and Robbins P. W. (198) J. Bid. Chem Harford J. B. and Waechter C. J. (198) Biochem. J Lucas J. J. and Levin E. (1977) J. Biol. Chem Carson D. D. Earles B. J. and Lennarz W. J. (1981) J. Biol. Chem Welply J. K. Shenbagamurthi P. Lennarz W. J. and Naider F. (1983) J. Biol. Chem Wedgwood J. F. and Strominger J. L. (198) J. Bwl. Chem Kato S. Tsuji M. Nakanishi Y. and Suzuki S. (198) Biochem. Biophys. Res. Commun Scher M. G. and Waechter C. J. (1984) J. Bwl. Chem

8 14732 Control of NLinked Carbohydrate Synthesis in ER Vesicles 28. Spiro R. G. Spiro M. J. and Bhoyroo V. D. (1983) J. BWZ Chem Waechter C. J. Kennedy J. L. and Harford J. B. (1976) Arch. 29. Kerr A. K. A. and Hemming F. W. (1978) Eur. J. Biochem. 83 Biochem. Biovhvs Carlo P. L. and Villemez C. L. (1979) Arch. Biochem. Biophys. 3. Ronin C. Caseti C. and Bouchilloux S. (1981) Biochim. Bio phys. Acta Scher 31. Gold P. and Green M. M. G. Devries G. H. and Waechter C. L. (1984) Arch. (1983) J. Biol. Chem Datema R. Schwarz R. T. and Janowski A. W. (198) Eur. J. Biochem. Biophys Biochem Jensen J. W. and Schutzbach J. S. (1984) Biochemistry Spiro M. J. (1984) Diabetologia Villemez. C. L.. and Carlo P. L. (1979) J. Biol. Chem Valtersson C. van Duyn G. Verkleij A. J. Chonjnacki T. de 4819 Kruijff B. and Dallner G. (1985) J. Biol. Chem Scher M. G. Jochen A. and Waechter C. J. (1977) Biochemistry 2751