THE JOURNAL OF BIOLOGICAL. CHEMISTRY Vol. 265, No. 34, Issue of December 5, pp. 21269-21278,199O 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

Endocytic Membrane Traffic to the Golgi Apparatus in a Regulated Secretory Cell Line*

(Received for publication, March 15, 1990)

Samuel A. Green+ and Regis B. Kelly From the Department of Biochemistry and Biophysics, University of California, San Francisco, California 94143-0448

We have established a ricin-resistant glycosylation- defective PC12 pheochromocytoma cell line to study biochemically glycoprotein traffic from the cell surface to the Golgi apparatus in regulated secretory cells. The strategy employed in this study is a modification of that used previously (Duncan, J. R., and Kornfeld, S. (1988) J. Cell Biol. 106, 617-628) to demonstrate transport of the cation-independent and -dependent mannose g-phosphate receptors from the cell surface to the trans-Golgi network in nonsecretory cell types. In ricin-resistant PC12 cells, radiolabeled galactose was incorporated enzymatically into surface glycocon- jugates, primarily glycoproteins. Resistance to B-ga- lactosidase was acquired upon reculture at 37 *C due to further terminal glycosylation of the galactose res- idues. Treatment of N-linked oligosaccharides isolated from recultured cells with a variety of glycosidases in conjunction with fl-galactosidase demonstrated the ad- dition of sialic acid N-acetylglucosamine and fucose residues to the galactose residues in recultured cells. Resistance to /3-galactosidase was not acquired in cells recultured at 19 “C, indicating that subsequent glyco- sylation of galactose residues did not occur at the cell surface or in endosomes. While glycosylation of galac- tose incorporated into asparagine oligosaccharides in Chinese hamster ovary clone 13 cells was not signifi- cant (~1%) after 6 h of reculture, approximately 10% of the galactose incorporated into surface oligosaccha- rides was further glycosylated in PC12 cells in this time. Analysis of total labeled versus &galactosidase- resistant proteins by sodium dodecyl sulfate-polyacryl- amide gel electrophoresis demonstrated that endocytic traffic to the site of glycosylation activity in mutant PC12 cells was highly selective, but included a much greater number of proteins than were detected in Chinese hamster ovary clone 13 fibroblasts.

Since the initial demonstration by Snider and Rogers (1) that the transferrin receptor can be transported from the cell surface to the Golgi apparatus, it has been established using a variety of methods that both the cation-dependent and the -independent mannose 6-phosphate receptors can follow the same endocytic route. These proteins are transported from

* This work was supported by National Institutes of Health Grant DK33937 (to R. B. K.) and was carried out during the tenure of a Postdoctoral Fellowship from the Muscular Dystrophy Association (to S. A. G.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed. Tel.: 4%476- 4805.

the cell surface to the truns Golgi network (TGN),’ the site of sialytransferase activity (2), with half times of 1-4 h (1, 3, 4, 5). In contrast, bulk transport of plasma membrane glyco- proteins (3) or glycoconjugates (5) to the TGN is extremely inefficient. Similarly, in the epithelial MDCK cell line, only a few proteins recycle efficiently between the apical plasma membrane and the TGN (6), while the cation-independent mannose B-phosphate receptor cycles with relatively high efficiency between the basolateral plasma membrane and the TGN (7). Thus, in the cell types examined, glycoprotein transport from the plasma membrane to the TGN is highly selective, and proceeds at rates that are much slower than the recycling of receptors between the cell surface and endosomes (8).

In secretory cells, which have a large number of vesicles devoted to transporting secretory products from the Golgi apparatus to the plasma membrane, vesicle membrane must be rapidly internalized from the cell surface following exocy- tosis in order to maintain the steady state distribution of membrane area between the compartments (9). There is bio- chemical evidence that at least some of that membrane is recycled to the Golgi apparatus (10, 11). This is presumed to be true in cells exhibiting constitutive secretion, in which the Golgi-derived secretory vesicles containing the secretory cargo fuse with the plasma membrane soon after they are made, as well as in regulated secretory cells, in which the Golgi-derived vesicles (granules) accumulate in the cytoplasm until an ex- ternal stimulus triggers rapid exocytosis (12, 13). Morpholog- ical studies documenting the transport of bulk membrane or fluid phase tracers to the Golgi apparatus in secretory cells indicate that traffic from the cell surface to the Golgi appa- ratus is a significant pathway in these specialized cell types (12, 14-17). Bulk endocytic tracers cannot readily be traced to the Golgi apparatus in cells that are not specialized for secretion (18-20) (which for simplicity we will refer to as “nonsecretory” cells), presumably because these cells have relatively less membrane and fluid flow through this pathway. These results suggest that secretory cells transport more fluid and membrane area from the cell surface to the Golgi appa- ratus than nonsecretory cell types, perhaps as a consequence of having to recycle secretory vesicle membrane components.

Since morphological studies of endocytosis to the Golgi apparatus have relied primarily on nonspecific fluid and mem- brane tracers and/or detection of enzymatic reaction products, quantitative analysis of this pathway, and identification of its components, has been limited. The biochemical methods used to study the transport of transferrin receptors and mannose

’ The abbreviations used in this paper are: TGN, truns Golgi network; BSA, bovine serum albumin- E&II, Bandeiruea simplicifo& lectin II; PBS. phosphate buffered saline: SDS. sodium dodecvl sul- fate; CHO, Chinese hamster ovary; Hepes, 4-(2-hydroxye&yl)-l- piperazineethanesulfonic acid; MES, 4-morpholineethansulfonic acid.

21269

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

21270 Endocytic Go&i Traffic in Regulated Secretory Cells

6-phosphate receptors can be used to obtain quantitative information. Using cell lines that are deficient in the addition of a specific sugar residue, or by experimentally modifying exposed oligosaccharides on the cell surface, substrates for Golgi glycosyltransferases can be created and radiolabeled at the cell surface. The subsequent activity of the Golgi glyco- syltransferases can be created and radiolabeled at the cell surface. The subsequent activity of the Golgi enzymes on these proteins can be monitored quantitatively. These meth- ods have thus far been applied only to nonsecretory cell types.

In order to develop a system to study quantitatively the traffic of cell surface glycoproteins to the Golgi apparatus in regulated secretory cells, we have selected glycosylation-de- fective mutants of PC12 pheochromocytoma cells for growth in the presence of a galactose-binding toxin, ricin. Here we describe the properties of the ricin-resistant PC12 cells, and our use of these cells to measure glycoprotein transport from the cell surface to the Golgi apparatus. We found that in PC12 cells, as in nonsecretory cell types, transport of cell surface proteins to the Golgi apparatus is highly selective. In contrast to the nonsecretory cell types, transport from the cell surface to the Golgi apparatus in PC12 cells appears to represent a major endocytic pathway for a greater number of glycoproteins.

MATERIALS AND METHODS

Chemicakr-Except where noted, chemicals were purchased from Sigma.

Enzymes-Fhobacterium meningosepticum peptide:N glycosidase F, Diplococcus pneeumoniae /3-galactosidase and @-N-acetylglucosa- minidase, bovine kidney or-fucosidase, Bacteroides fragilis endo-@- galactosidase, Helix pomatia arylsulfatase and Vibrio &&roe neura- minidase were purchased from Boehringer Mannheim. Bovine milk galactosyltransferase and Clostridium perfringens neuraminidase were from Sigma.

Cells-PCi2 cells were grown in Dulbecco’s minimal essential medium (Cell Culture Facilitv. Universitv of California. San Fran-

“ I

cisco) containing 5% fetal bovine serum (HyClone Laboratories, Logan, UT) and 10% horse serum (HyClone Laboratories), and were passaged by trituration in divalent cation-free phosphate buffered saline (PBS) containing 10 mM Hepes, pH 7.2, and 1 mM EDTA. Chinese hamster ovary (CHO) clone-13 cells (21), kindly provided by Dr. Stuart Kornfeld (Washinaton Universitv, St. Louis, MO), were grown in a-modified‘Eagle’s medium containing 10% fetal ‘bovine serum, and were passaged by trypsinization. For some experiments cells were plated on dishes precoated with poly-D-lysine. CHO clone 13 cells have a defect in the ability to transport UDP-galactose from the cytoplasm to the lumen of the Golgi vesicles, accounting for their deficiency in incorporating galactose metabolically (21,34).

Selection of Ricin-resistant Cells-PC12 cells were grown in me- dium containing various concentrations of Ricin communk toxin (RCAGo, Sigma). 10 rig/ml ricin toxin was sufficient to kill >99% of the cells within 4 days. For selection of ricin-resistant clones, 5 X lo6 cells were plated in each of three lo-cm dishes and fed with medium containing 10 rig/ml ricin. After 4 days, at which time most of the cells had died, cells were refed with 20% PClL-conditioned medium containing ricin. Clones were visible after 5 days, and were picked using cloning rings after 19 days. Large stocks of selected clones were grown up without ricin for storage in liquid nitrogen. After the initial freezing, clones were passaged not more than 25 times, and were generally reselected after thawing for one passage (3 days) in 20 ng/ ml ricin. Selection was performed on mutagenized PC12 cells, how- ever similar numbers of colonies were obtained without mutagenesis.

Lectin Binding Assays-Ricin toxin (RCA& and Bandeiraea sim- plicifolia lectin II (BSII) (Sigma) were iodinated to low specific activitv using Iodobeads (Pierce Chemical Co.). Tvpicallv, 1 mg of protein in 350 ~1 of PBS was iodinated for 45-60 s at roomtempera- ture in the presence of 0.5 mCi of NalZ51 (Amersham Corp.) and one iodobead. Free iodine was removed by ion-exchange chromatography over 250-300 ~1 of HCl-charged AG-IX8 resin (Bio-Rad) prewashed with PBS containing 10 mg/ml bovine serum albumin (PBS-BSA). The column was waihed o&e with 150 ~1 PBS containing 4% BSA following the addition of the iodinated sample. The assays were

performed in plastic tubes blocked for at least 1 h with PBS contain- ing 4% BSA. The cells were washed twice with ice-cold PBS, har- vested in PBS without divalent cations (and containing 5 mM EDTA in the case of CHO cells), and resuspended at 1 X 10’ cells per ml in PBS. Aliquots (300 ~1) of cells were added to the coated tubes containing 100 ~1 of PBS, 4% BSA or to uncoated tubes for protein determination. Lectins were added to the cells in 100 pl PBS, 1% BSA. Competing sugars were mixed with PBS, 4% BSA at 200 mg/ ml (final concentration in the assay of 40 mg/ml). After incubation on ice for 2 h with frequent agitation, cells were pelleted, washed once with PBS, 1% BSA, and the cell pellets counted in a Beckman Gamma 4000 counter.

Glycosyltransferose Assays-Galactosyltransferase was measured as described (22) usine 2 uCi of UDP?Hlealactose (Amersham Corn.) per sample and twice the published‘concentratin ‘of UDP-galactose (Sigma) and ATP. Sialytransferase was measured as described (23) using 1.2 &i of CMP-[3H]neuraminic acid (Du Pont-New England Nuclear) per sample with a final concentration of 0.18 mM CMP- neuraminic acid (Sigma). For both assays, samples were precipitated for 10 min with ice-cold 20% trichloroacetic acid, 1% phosphotungstic acid, 0.5 M HCl, pelleted by centrifugation at 16,000 x g for 10 min, and washed once with the acid mixture. The pellets were solubilized in 300 pl of NCS tissue solubilizer (Amersham Corp.), neutralized with 15 ~1 of glacial acetic acid, and counted in 4 ml of Ecolume scintillation cocktail (ICN Schwarz/Mann). Protein was quantitated by dye binding (24) using BSA as a standard.

Uptake and Release of Norepinephrine from PC12 Cells-The con- ditions for loading PC12 cells and the glycosylation mutants with [3H]norepinephrine (Amersham Corp.) in the presence of pargyline and stimulation of exocytosis by membrane depolarization with 55 mM KC1 in the presence of carbamylcholine chloride were exactly as described (25). In this study, cells were washed twice after uptake of labeled norepinephrine and incubated for 8 min in 1 ml of culture medium at 37 ‘C. This medium was collected, and an additional 1 ml of medium with or without added KC1 (55 riM final concentration) and 5 mM carbamylcholine chloride was added to each dish. The final incubation medium was collected after 8 min, and 100~~1 aliquots of all collected media were counted in 4 ml of scintillation fluid.

Exogulactosylation-Exogalactosylation of cells was carried out as described (3, 23, 26). Cells on polylysine-coated 3.5~cm or 6-cm tissue culture plates were labeled at 75-90% confluence, usually 1-2 (PC12 cells) or 2-3 days (CHO clone 13 cells) after plating. Cells were cooled to 0 “C and washed twice with ice-cold exogalactosylation buffer (23) (minimal essential medium without glucose or bicarbonate containing 1% fetal bovine serum, 2 mM glutamine, and 10 mM Hepes, pH 7.3). The labeling medium contained exogalactosylation buffer with the following additions: 4-12 mM MnCl*, 0.5-1.0 units/ml galactosyl- transferase, and 35-100 pCi/ml UDP-[3H]galactose. Cells were la- beled with 0.25 ml per 3.5-cm dish or 0.5 ml per 6-cm dish for 45-50 min on ice. At the end of the labeling all cells were washed twice with ice-cold PBS containing 1 mg/ml glucose. Cells that were recultured were washed once with ice cold culture medium, then incubated in culture medium at 37 or 19 “C. Cycloheximide was included in some cultures at 10 rg/ml. All cells were washed with PBS lacking divalent cations just prior to lysis. For nonradioactive exogalactosylation prior to lectin binding assays, cells were washed with exogalactosylation buffer in suspension and exogalactosylated at 2 X 10’ cells/ml for 60 min in 0.5 ml of labeling buffer containing 1 unit/ml galactosyltrans- ferase and 10 mM UDP-galactose in place of the radioactive substrate.

Lipid Extraction-PC12 Al cells exogalactosylated with UDP-[3H] galactose were washed four times with PBS, and then harvested in PBS lacking divalent cations. 1.2 X lo6 cells were resuspended in 600 ~1 of distilled water and disrupted by passage through a 25-gauge needle. The homogenate was aliquoted into three 180-rl samples and extracted according to the method of Dahms and Schnarr (27). Methanol was added followed by chloroform to a chloroform: methanol:water ratio of 4:8:3. After vigorous mixing the extract was centrifuged at 12,000 x g for 8 min, and the pellet washed three times with 150 ~1 of solvent mixture. The remaining insoluble material was designated the protein fraction. The supernatants were pooled and water was added to a final solvent ratio of 4:8:5.6 (chloro- form:methanol:water). The mixture was vortexed vigorously and the aqueous and organic phases separated by centrifugation at 12,000 X g for 4 min. Both phases were reextracted with solvent phases prepared in parallel. The pooled aqueous and organic phases were dried under vacuum at 40 “C. Organic phase material was dissolved in scintillation fluid and counted, and was found to contain <0.02% of the total radioactivity recovered in the extraction. Aqueous phase

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

Endocytic Golgi Traffic in Regulated Secretory Cells 21271

material, which is expected to include glycolipids, was resuspended in 600 ml of distilled water. To mimic the composition of the detergent lysates used in this study (see below), a 400+1 portion was adjusted to 0.15% Nonidet P-40; 50 mM NaCl. 100 fig of bovine serum albumin was added, and the sample was acid precipitated with an equal volume of 20% trichloroacetic acid, 1% phosphotungstic acid, 0.5 M HCl on ice for 10 min. The 16,000 x g pellet from this precipitation was washed with acetone. The resulting acid precipitate and the protein pellet from the extraction were dissolved in 40 pl of 8 M urea, 2% SDS. 20 mM dithiothreitol, 5 mM EDTA in a boiling water bath. The$e and the remaining iO0 ~1 of the redissolved aqueous phase material were counted in 4 ml of scintillation fluid.

Detection of Terminally Glycosylated Proteins by Acid Precipita- tion-Radioactively exogalactosylated cells were lysed on the plates by the addition of 100 pl of lysis buffer I (0.5% Nonidet P-40, 50 mM sodium phosphate, pH 6.2) containing freshly added proteinase in- hibitor cocktail (added from a 1000 X stock containing 10 pg/ml each of pepstatin, chymostatin, leupeptin, and aprotinin in dimethyl sulf- oxide and a 200 X stock containing 200 mM phenylmethylsulfonyl fluoride, 2 mM benzamidine, and 200 pg/ml o-phenanthroline in ethanol). The dishes were incubated 5 min on ice, the lysate collected, and 6-cm plates washed with an additional 100 ~1 of lysis buffer. Lysates were centrifuged at 30,000 X g for 15 min to remove nuclei and insoluble material. Portions of the supernatant were incubated for 14-16 h at 37 “C with or without the addition of 25-45 milliunits of fl-galactosidase. An additional lo-15 milliunits of enzyme was then added, and the incubation continued for 2 h. Samples were precipi- tated with ice-cold 20% trichoroacetic acid, 1% phosphotungstic acid in 0.5 M HCl, and then washed with ice-cold acetone. The second addition of P-galactosidase was necessary to reduce the background, which was defined as acid precipitable counts resistant to digestion from cells that were lysed immediately after labeling, below 2% of the counts incorporated. In preliminary experiments, the precipitates were solubilized in 300 ~1 of NCS tissue solubilizer, neutralized with 15 ~1 of glacial acetic acid, and counted in 4 ml of scintillation cocktail. For subsequent experiments precipitates were dissolved by heating in a boiling-water b&h in 40-80 &of 8 M urea, 2% SDS, 20 mM dithiothreitol, 5 mM EDTA. For exneriments that included analysis by SDS-polyacrylamide gel electrophoresis, the precipitates were resolubilized by sequential addition of 20-50 ~1 of 7 M urea, 20 mM dithiothreitol, followed by an equal volume of 2 X concentrated electrophoresis sample buffer (28) and finally sample buffer to a final volume of loo-150 pl, with heating in a boiling water bath after each addition. Portions (lo-20 ~1) of the solubilized samples were counted in 4 ml of scintillation fluid.

Analysis of N-Linked Oligosacchurides-Radioactively exogalacto- sylated cells were lysed on the plates by the additin of 50-80 ~1 lysis buffer II (1% Nonidet P-40; 40 mM Tris, pH 8.0, 5 mM EDTA) containing freshly added proteinase inhibitor cocktail. The dishes were incubated 5 min on ice, the lysate collected, and the plates washed with an additional 50 ~1 of lysis buffer. Lysates were centri- fuged at 30,000 x g for 15 min. The supernatants were adjusted to 0.2% SDS and 1% @-mercaptoethanol and heated in a boiling water bath for 3 min. After cooling to room temperature, 10 milliunits of peptide:N glycosidase F (29) was added, and the lysates were incu- bated for Is-20 h at 37 “C. The samples were then cooled to 0 “C, nrecipitated with 100 ul of 20% trichloroacetic acid. 1% nhosnho- tungitic acid, 0.5 M HCl for 5 min, and centrifuged ai 16,OdO x g for 10 min at 4 “C. To minimize loss of a-linked sugar residues by acid hydrolysis, the supernatants were immediately adjusted to pH 8-9 with 20 ~1 of 10 N NaOH and 1.5 ~1 of saturated aqueous Tris base, and desalted into 2 mM Tris, pH 7.5, by gel filtration chromatography on a 0.6- X 22-cm Sephadex G25 (Pharmacia LKB Biotechnology, Inc.) column. Radioactive fractions were pooled, dried under vacuum, and dissolved in 150-200 pl of 50 mM N-morpholinoethanesulfonic acid (MES) pH 6.0. Redissolved oligosaccharides (approximately 2 x 10' cpm in 40 gl) were incubated for 14-16 h at 37 “C with 10 milliunits of @-galactosidase. An additional 5 milliunits of P-galacto- sidase was then added, and the incubation continued for 3-5 h. Samples were boiled for 3 min to inactivate the enzyme, made lo- 20% (w/v) sucrose, and analyzed by gel filtration chromatography on Sephadex G-25. 350-pl fractions were counted in 4 ml of scintillation fluid.

To analyze terminal additions to galactose residues in recultured cells, some of the following glycosidases were added to oligosaccharide samples, either singly or in combination, in addition to i0 milliunits of P-galactosidase; 8-10 milliunits each of Vibrio cholera and Clostri- dium perfringens neuraminidases, 8 milliunits of /3-N-acetylglucosa-

minidase, I5 pg of a-fucosidase, 8-10 milliunits of endo-p-galactosid- ase, 40 Kg of arysulfatase. Digestions were carried out in parallel with fl-galactosidase digestions, and half of the original amount of each enzyme was added after 14-16 h. In some experiments portions of desalted oligosaccharides were dried and dissolved in 10 ~1 of 2 M acetic acid, incubated for 60 min in a boiling water bath, neutralized with NaOH, and diluted into 80 ~1 of 50 mM MES prior to glycosidase treatment.

Electrophoresis and Fluorography-Sodium dodecyl sulfate (SDS)- polyacrylamide gels were run as described (28). After fixing in 10% methanol, 5% acetic acid, gels were rinsed with water and impreg- nated with 0.5 M sodium salicylate according to the method of Chamberlain (30). The dried gels were exposed to preflashed (31) Kodak XAR5 film at -70 “C.

RESULTS

Experimental Stategy-One approach that has been used to study endocytic traffic to the Golgi apparatus entails se- lecting cell lines that are deficient in addition of terminal galactose residues to N-acetylglucosamine residues on nascent glycoproteins. Glycoproteins synthesized in cells with such a defect do not incorporate sialic acid or other terminal sugar residues into their glycoproteins since the normal substrate for these additions, galactose, is not incorporated. This phe- notype allows the enzymatic addition of radiolabeled galactose to the exposed cell surface glycoproteins followed by incuba- tion of the cells to allow recycling to the Golgi apparatus to occur. If glycoproteins radiolabeled in this way are recycled to the Golgi apparatus, they will be sustrates for glycosyltrans- ferases that normally act on terminal galactose residues, which were not incorporated when the proteins were synthe- sized. Addition of sugar residues to a radiolabeled galactose residue will render that galactose residue resistant to digestion with fi-galactosidase, since this enzyme removes only terminal galactose residues from complex oligosaccharides (3, 26). Quantitation of P-galactosidase-resistant galactose residues on glycoproteins can then be accomplished either by analysis of isolated oligosaccharides by treatment with glycosidase and gel filtration chromatography (3), or by acid precipitation of glycosidase-treated glycoproteins. The strategy is summarized in Fig. 1.

Selection of Ricin-resistant PC12 Cells-To select PC12 cells with defective galactosylation of glycoproteins, cells were grown in the presence of ricin toxin. Ricin toxin binds to terminal galactose residues through its B chain, which nor- mally facilitates the efficient internalization of the toxin (32). It was found that 10 rig/ml toxin killed >99% of wild type PC12 cells within 4 days of culture. This concentration of toxin was used to select PC12 cells from two lo-cm dishes (approximately 1 X lo7 cells per dish). Nine clones were obtained from these plates, seven of which were extensively characterized.

Ricin-resistant PC12 Cells Exhibited Pectin Binding Prop- erties Consistent with Defective Galactosylation-Ricin-resist- ant clones were initially screened for their capacity to bind radiolabeled ricin toxin. Of seven clones characterized, three bound >20% as much ricin per cell as wild type PC12 cells. The remaining four bound 13-16% as much as wild type cells (not shown). A more detailed analysis of lectin binding was performed on two of these four clones, Al and A3. These cells were analyzed for their glycosylation phenotype by testing for binding of both ricin toxin (galactose specific) and Bandieraea simplicifolia lectin II (BSII), a lectin specific for terminal N- acetylglucosamine residues (33), either with or without prior exogalactosylation with unlabeled UDP-galactose (see “Ma- terials and Methods”). Both clones bound approximately 6 times less ricin toxin and 16 times more BSII than wild type PC12 cells (Table I?. Prior exogalactosylation increased ricin

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

21272 Endocytic Golgi Traffic in Regulated Secretory Cells

EXPERIMENTAL STRATEGY

CELLSURFACE SURFACE u*u GLYCOPEUTEN

EXCGALACTOSYUITE. RECULTURE 37%

0. Yii

AClD PRECIPTTATE. SOLUBILIZE

l *

B ISOLATE N-LINKED CLIGQSACCHARIDES. DIGEST WI-T” B GALACTOSrnASE

.*

u ti

FRACWNATE BY GEL FILTRATION

FIG. 1. Schematic representation of the experimental strat- egy. N-linked oligosaccharides on cell surface glycoproteins that terminate in N-acetylglucosamine (the mannose core is not shown) can be labeled with [3H]galactose upon the addition of galactosyl- transferase and UDP-[3H]galactose at 0 “C (exogalactosylation). Upon reculture of the cells at 37 “C, glycoproteins may remain on the cell surface or may be transported to a variety of endocytic compart- ments. If galactosylated glycoproteins are transported to the Golgi apparatus, some or all of the galactose incorporated at the cell surface can be further glycosylated by endogenous glycosyltransferases. Gly- cosylation can be measured in two ways as follows. A, acid precipi- tation from detergent lysates. Following reculture, detergent lysates are incubated with or without P-galactosidase. Terminal galactose residues are removed by the enzyme; additionally glycosylated galac- tose residues are resistant to digestion. Following digestion, the glycoproteins are acid-precipitated to separate them from free galac- tose. The precipitates are solubilized, and tritium counts recovered from digested and undigested samples are compared to quantitate glycosylation of galactose residues, or analyzed qualitatively by SDS- polyacrylamide gel electrophoresis and fluorography. B, analysis of isolated N-linked oligosaccharides. N-linked oligosaccharides are iso- lated from recultured cells by digestion of cell lysates with peptides:N- glycosidase F, acid precipitation, and desalting of the oligosaccharides in the supernatant. The isolated oligosaccharides are digested ex- haustively with P-galactosidase. Oligosaccharides are then separated from free galactose by gel filtration chromatography. n , N-acetylglu- cosamine; 0, [3H]galactose; +, additional terminal sugar residues (sialic acid, N-acetylglucosamine and/or fucose).

TABLE I

Lectin binding properties of PC12 Al and A3 cells (ng of lectin bound per mg of cell protein)

Approximately 3 x lo6 cells were incubated in 0.5 ml of PBS, 1% BSA with 40 pg of radioiodinated lectin for 2 h on ice with frequent mixing. Cells were washed once and counted.

Rich Rich + Gal” BSII BSII + GlcNAc”

PC12 wild type 197 4.0 5.7 3.1 PC12 A1 35 3.4 90 3.6 PC12 A3 35 5.6 96 4.4 Al/exogalb 261 ND’ 24 ND A3/exogal* 245 ND 19 ND

LI Competing sugars (galactose or N-acetylglucosamine) were in- cluded at a final concentration of 40 mg/ml.

b Cells were exogalactosylated with galactosyltransferase and 10 mM UDP-galactose for 60 min at 0 “C prior to washing and addition of radiolabeled lectin.

’ ND, not determined.

binding above wild type levels and reduced BSII binding to approximately one quarter of control levels in both clones. Lectin binding was virtually abolished in the presence of the appropriate competing sugar (Table I). These studies dem- onstrate the presence of many more terminal N-acetylglucos- amine residues, and many fewer galactose residues, on the cell surface glycoproteins of the PC12 Al and A3 cells than on wild type PC12 cells, and that the selected clones cells are good substrates for the enzymatic addition of galactose.

CHO clone 13 cells, previously used for detection of glyco- protein traffic from cell surface to the TGN (3), were used as a representative nonsecretory cell type throughout this study. These cells are deficient in the ability to transport UDP- galactose into the lumen of the Golgi apparatus from the cytoplasm, resulting in the synthesis of glycoproteins with terminal N-acetylglucosamine residues (21, 34). Wild type PC12 cells and PC12 Al cells were compared with CHO clone 13 cells for their capacity to bind ricin toxin and BSII lectin. Washed cells were incubated in suspension with various con- centrations of radiolabeled lectin at 0 “C for 2 h prior to washing and counting. The lectin binding profiles of PC12 Al cells and CHO clone 13 cells were similar for both ricin toxin (Fig. 2A) and BSII (Fig. 2B). As expected, wild type PC12 cells bound significantly more ricin toxin and less BSII than PC12 Al cells or CHO clone 13 cells.

Ricin-resistant PC12 Cells Retained Normal Levels of Gly- cosyltransferase Actiuities-One possible defect leading to the

Dif,. , , . , 0 10 20 30 40

A wg Rick Added

B ~9 BSll Added

FIG. 2. Lectin binding profiles of wild type and ricin-re- sistant PC12 cells and CHO clone 13 cells. Approximately 3 x lo6 cells were incubated in 0.5 ml of PBS, 1% BSA with the indicated quantities of radioiodinated lectin for 2 h on ice. Cells were washed once and counted. Ricin-resistant PC12 Al cells and CHO clone 13 cells bound similar amounts of ricin toxin (specific for terminal galactose and galactosamine residues), while wild type PC12 cells bound approximately 5 times more. Wild type cells bound very little Bandeiraea simplicifolia lectin II (BSII, specific for terminal N- acetylglucosamine residues) while PC12 Al and CHO clone 13 cells bound approximately 25 and 40 times more, respectively.

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

Endocytic Golgi Traffic in Regulated Secretory Cells 21273

phenotype observed in ricin-resistant cells is a lack of galac- tosyltransferase activity. This was tested in several of the ricin-resistant PC12 clones. Most of the clones showed spe- cific activities of galactosyltransferase approximately twice that of the wild type cells. None was found deficient in galactosyltransferase activity. The results of clones Al and A3 are shown in Table II.

An essential requirement for the detection of glycoprotein recycling in galactosylation-deficient cells is that the cells be capable of adding additional sugar residues to experimentally galactosylated oligosaccharides when they do reach the Golgi apparatus. The most common addition to galactose residues on glycoproteins is sialic acid. Therefore, we tested the ricin- resistant clones for sialyltransferase activity. Of seven clones tested, the specific activities ranged from approximately half to approximately twice the wild type level of enzyme activity. Clones Al and A3 exhibited specific activities slightly greater than wild type cells (Table II).

Ricin-resistant PC12 Cells Exhibited Regulated Secretion- The final property required of the cells for this study is that they retained the regulated secretory pathway normally found in PC12 cells (35, 36). This was tested by monitoring the release of stored [3H]norepinephrine in response to membrane depolarization with potassium in combination with carba- mylcholine chloride (25). Most of the clones tested exhibited increased secretion of [3H]norepinephrine in response to stim- ulation similar in magnitude to that observed in wild type cells, confirming that the specializations required for regu- lated exocytosis are retained in the ricin-resistant cells (Table II).

Galactose Was Incorporated into Glycoproteins and Glyco- lipids in Ricin-resistant PC12 Cells-Preliminary experiments using detergent lysates of labeled cells suggested that a frac- tion of the radiolabeled galactose was incorporated into gly- colipid acceptors in ricin-resistant PC12 cells, and that the labeled glycolipids were acid-precipitable under the experi- mental conditions (see below). To characterize the nature of cell surface acceptors for galactose, PC12 Al cells were exo- galactosylated with radiolabeled galactose, washed thor- oughly, and harvested. Cell pellets were homogenized and subjected to extraction in chloroform:methanol:water as de- scribed under “Materials and Methods” (27). The extract was separated into insoluble (protein and nucleic acid), organic, and aqueous phases. In triplicate samples, 84 + 3% (mean f S.D.) of the label was recovered as insoluble material (pre- sumed to be glycoprotein) in the extraction while 16 f 3% of the label was recovered in the aqueous phase, which would contain glycolipids as well as any unincorporated label. Pro- tein, detergent, and salt were added to a portion of the aqueous phase to approximate the composition of detergent lysates, and the mixture was acid-precipitated (see “Materials and

TABLE II Specific activities of ricin-resistant PC12 cells relative to wild type

PC 12 wlls

PC12 Al PC12 A3

Galactosyltransferase” 2.1 2.2 Sialyltransferase” 1.4 1.8 Regulated secretion* 0.83 1.9

a Measured as cpm transferred to acceptor per mg of cell protein. * Cells incubated in medium containing [3H]norepinephrine and

pargyline for 1 h were washed twice, then incubated for two sequential 8-min periods. The second incubation contained 55 mM KC1 and 5 mM carbachol. [3H]Norepinephrine released into the medium during each incubation was measured by scintillation counting. Wild type PC12 cells released 4.0 times more [3H]norepinephrine in the second incubation than in the first.

Methods”). Of the radioactivity recovered in the aqueous phase, 70 + 5% was acid-precipitable (approximately 11% of the total radioactivity). These results show that most of the galactose was incorporated into protein acceptors, while a significant fraction was incorporated into organic solvent- extractable, acid-insoluble acceptors, assumed to include gly- clipids.

Cell Surface Glycoproteins Acquired Resistance to f3-Galac- tosidase upon Reculture at 37 “C-In order to begin to char- acterize the traffic of cell surface glycoproteins to the Golgi apparatus, PC12 Al and A3 cells and CHO clone 13 cells were exogalactosylated with radiolabeled galactose as described under “Materials and Methods.” Following the labeling, cells were either harvested and lysed in detergent immediately or were incubated in growth medium for up to 12 h at 37 “C prior to lysis. Radiolabeled galactose was analyzed either by digestion of intact glycoproteins (and glycolipids) in the ly- sates with /3-galactosidase followed by acid precipitation to separate released galactose from &galactosidase-resistant in- corporated galactose, or by isolating N-linked oligosaccha- rides, digesting these with @-galactosidase, and separating the products by gel filtration chromatography.

For analysis of intact glycoproteins, portions of the lysates were digested exhaustively with /3-galactosidase prior to acid precipitation and solubilization. Aliquots of the solubilized precipitates were counted (see Fig. 1). The results are sum- marized in Table III. Galactose incorporated into acid-precip- itable material in cells harvested immediately after labeling was almost entirely sensitive to &galactosidase digestion. This level of acid-precipitable radioactivity (<2% of the input) was defined as the background in the assay. Acid-precipitable P-galactosidase-resistant radioactivity increased to approxi- mately 1% of the total radioactivity in CHO clone 13 cells upon reculture for 6 h at 37 “C. However, in lysates from PC12 Al cells recultured at 37 “C! for 6 h approximately 16% of the incorporated galactose was resistant to @-galactosidase. PC12 A3 cells yielded essentially identical results (not shown), indicating that the findings are not peculiar to a single clone. After correcting for background, approximately 14% of the galactose incorporated into PC12 Al surface glycoconjugates, and approximately 1% of the galactose incorporated into CHO clone 13 cells, acquired resistance to P-galactosidase after 6 h of reculture. Kinetic experiments indicated that acquisition

TABLE III Percentage of PHlgalactose resistant to fi-galactosidase digestion in

detergent lysates Cells were exogalactosylated with UDP-[3H]galactose and galac-

tosyltransferase for 45-50 min at 0 “C. Cells were washed and lysed immediately (control) or were recultured for 6 h at the indicated temperatures prior to lysis. Lysates were divided unequally (at ratios ranging from 1:3 to 1:7), and the larger fraction was digested exhaus- tively with /3-galactosidase. Lysate samples were then acid-precipi- tated, the precipitates were washed with acetone, and solubilized by the addition of 7 M urea followed by 2 X concentrated electrophoresis sample buffer. Aliquots of the solubilized precipitates were counted. The percentage of galactose protected in each sample was calculated based on the acid-precipitable radioactivity in the smaller, undigested fraction of the same lysate. Values represent the mean k standard deviation, and were determined based on three to five independent experiments.

Control 6 h 37 “C 6 h 18 “C Neuraminidase”

PC12 Al 1.5 + 0.5 16 -c 0.4 2.3 f 0.2 11 * 0.5 CHO clone 13 0.7 + 0.5 1.6 + 0.2 NDb ND

’ Lysates from cells recultured for 6 h were digested with neura- minidase in addition to p-galactosidase (see “Materials and Methods” and “Results”).

* ND, not determined.

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

21274 Endocytic Go@ Traffic in Regulated Secretory Cells

of fi-galactosidase resistance measured by acid precipitation was linear for at least 8 h in PC12 Al cells, and reached 23% of the total acid-precipitable radioactivity by 12 h of chase (Fig. 4).

N-linked oligosaccharides were isolated from lysates by digestion with peptide:N-glycosidase F followed by acid pre- cipitation and desalting of the supernatant. Approximately 85-90% of the incorporated radioactivity was recovered in the acid supernatant, and approximately 95% of this radioactive material eluted in the void volume and separated from free galactose, salts, and buffers on Sephadex G-25 gel filtration chromatography. These oligosaccharides were digested with fi-galactosidase, and the resulting products analyzed by Seph- adex G-25 chromatography. Radioactivity in P-galactosidase- treated oligosaccharides from cells harvested immediately after labeling eluted as a single peak corresponding to the elution position of free galactose (Fig. 3). In oligosaccharides from cells recultured for 2 h, some of the radioactivity eluted in a peak beginning at the void volume, and corresponding to the position of oligosaccharides isolated in the desalting step. With increasing times of reculture, a greater fraction of the radioactivity incorporated into oligosaccharides eluted in the faster peak after @galactosidase treatment (Fig. 3). The in- crease in &galactosidase resistance as measured by this method showed kinetics very similar to those seen by digestion of lysates followed by acid precipitation, but reached a lower level of 16% of the incorporated galactose after 12 h of chase (Fig. 4).

Resistance to P-Galactosidase Was Not Acquired upon Re- culture at 19 “C-The effect of reduced temperature on ac-

Fraction Number

FIG. 3. Galactose incorporated into N-linked oligosaccha- rides in PC12 Al cells acquired resistance to fi-galactosidase upon reculture at 37 ‘C. Confluent 35-mm plates of PC12 Al cells exogalactosylated with UDP-[3H]galactose were recultured for up to 12 h at 37 “C, as indicated. N-linked oligosaccharides were isolated from cell lysates by digestion with peptide:N-glycosidase F, acid precipitation, and desalting of the supernatant. Isolated oligosaccha- rides were digested with @-galactosidase, and the products separated by gel filtration chromatography. The peak at fraction 11 corresponds to the elution volume of free galactose. The peak at fraction 5 corresponds to the peak of radioactivity obtained by desalting the acid-precipitated, peptide:N-glycosidase F-digested lysate samples. All of the galactose from cells that were not recultured was sensitive to release from the oligosaccharides with @galactosidase. With in- creasing time of reculture, an increasing fraction of the galactose eluted in the oligosaccharide peak. By 12 h of reculture, 16% of the radioactivity eluted in the first peak.

4 8 12

Chase (h)

FIG. 4. Kinetics of acquisition of &galactosidase resistance in total acid-precipitable material and in isolated iv-linked oligosaccharides from PC12 Al cells. Acquisition of resistance to fl-galactosidase upon reculture of exogalactosylated PC12 Al cells was measured either by acid precipitation (see Table III) or gel filtration analysis of isolated N-linked oligosaccharides (see Fig. 3). Resistance to p-galactosidase as measured by either method increased linearly for at least 8 h, and less rapidly between 8 and 12 h. At 12 h of chase approximately 23% of acid-precipitable galactose, and ap- proximately 16% of galactose in N-linked oligosaccharides was re- sistant to P-galactosidase.

quisition of P-galactosidase resistance was investigated to begin to define the site of terminal glycosylation of surface- labeled glycoproteins in PC12 cells. Vesicular traffic between certain organelles is sensitive to reduce temperature. While endocytosis continues at a reduced rate at temperatures below 20 “C, delivery of endocytic contents to lysosomes is inhibited (37, 38). More importantly, the transport of ricin toxin from endosomes to the Golgi apparatus (39, 40) and the transport of cell surface glycoproteins to the site of sialyltransferase activity (1, 3, 4, 6) are also efficiently blocked below 20 “C. When ricin-resistant PC12 cells were incubated at 19 “C for 6 h following exogalactosylation, the fraction of galactose resistant to P-galactosidase, either in total acid-precipitable material (Table III), or isolated N-linked oligosaccharides (not shown) was only slightly greater than the fraction found in cells harvested immediately following labeling. These ex- periments show that galactose residues were not modified on the cell surface or in endosomes in ricin-resistant PC12 cells, and are consistent with a requirement for transport to the Golgi apparatus. Finally, PC12 Al cells recultured at 37 “C in medium containing 10 pg/ml cycloheximide exhibited the same level of P-galactosidase-resistant radiolabel in isolated oligosaccharides and in acid-precipitable products from ly- sates as cells recultured without cycloheximide (not shown). These results preclude the possibility that radiolabeled galac- tose was incorporated biosynthetically from a pool liberated from glycoproteins that were degraded in lysosomes during reculture.

The experiments described above demonstrate that the galactose incorporated into glycoconjugates at the cell surface in ricin-resistant PC12 cells becomes resistant to @-galacto- sidase upon reculture as a result of additional terminal gly- cosylation of the galactose residues, and that the glycosylation occurs in an endocytic compartment not accessible at 19 “C. This compartment is most likely the Golgi apparatus (2, 41)

PC12 Al Cells Added Several Different Sugar Residues to Galactose Incorporated into N-linked Oligosaccharides-Pub- lished reports (1,3,4,5,6) on transport of glycoproteins from the cell surface to the Golgi apparatus have demonstrated the addition of sialic acid residues. If sialic acid is the only sugar residue incorporated into exogalactosylated glycoproteins, ad- dition of neuraminidase to the P-galactosidase digestion should result in the removal of all of the radiolabeled galac- tose. When this experiment was performed by acid precipi- tation of digested lysate samples from PC12 Al or A3 cells,

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

Endocytic Go&i Traffic in Regulated Secretory Cells

using Vibrio chderae and Clostridium perfringens neuramini- dases, approximately one-third of the &galactosidase-resist- ant galactose was removed (Table III). Thus, in contrast to earlier biochemical studies of endocytic Golgi traffic, PC12 cells appeared to add sugar residues in addition to sialic acid to galactosylated glycoproteins.

In order to determine the nature of the modifications ac- counting for resistance to &galactosidase, radiolabeled N- linked oligosaccharides were isolated from PC12 Al cells recultured for 6 h after exogalactosylation. These oligosaccha- rides were digested with p-galactosidase with or without a variety of additional glycosidases, and with or without prior mild acid hydrolysis to remove completely sialic acid residues. The products were separated by gel filtration chromatogra- phy, and the radioactivity in the two peaks counted as de- scribed above. The percentage of counts eluting in the first peak (B-galactosidase resistant) in each sample was normal- ized to the percentage in the sample digested with P-galacto- sidase alone. The results are summarized in Table IV. For simplicity, it will be understood in the following description that all samples were digested with @-galactosidase.

Removal of sialic acid residues by neuraminidase digestion or acid hydrolysis alone converted approximately one-quarter of the ,&galactosidase-resistant galactose in isolated oligosac- charides to /3-galactosidase-sensitive. Addition of p-N-acetyl- glucosaminidase alone exposed approximately one-quarter of the protected galactose. Removal of sialic acid residues and N-acetylglucosamine residues simultaneously by combining these treatments removed approximately half of the protected galactose residues. The additive effects of these two treat- ments suggest that upon reculturing the cells sialic acid was added to some of the galactose residues incorporated at the cell surface, and N-acetylglucosamine was added to others. Addition of N-acetylglucosamine to galactose to form N- acetyllactosamine disaccharide units can occur both on gly-

TABLE IV

Glycosidase analysis of asparagine-linked oligosacckarides isolated from recultured PC12 Al cells: Fraction of rH]galactose residues

incorporated into asparagine-linked oligosaccharides that were resistant to digestion with P-galactosidase in combination with other

glycosidases, relative to the fraction resistant to P-galactosidase alone PC12 Al cells were exogalactosylated with UDP-[3H]galactose and

recultured for 6 h at 37 “C. Asparagine-linked oligosaccharides were isolated, aliquots dissolved in 50 mM MES, pH 6.0, and digested exhaustively with the indicated glycosidases as described under “Ma- terials and Methods.” For acid hydrolysis, isolated oligosaccharides were dissolved in 10 ~1 of 2 M acetic acid, incubated 60 min in a boiling water bath, neutralized with 2 ~1 of 10 N NaOH, and diluted lo-fold in 50 mM MES, pH 6.0, prior to addition of glycosidases. Following digestion, samples were boiled for 3 min to inactivate the enzymes and analyzed by gel filtration chromatography on Sephadex G-25. Counts eluting in a peak immediately following the void volume were counted as resistant to digestion, and the large peak of counts eluting in later fractions counted as sensitive to digestion (see Fig. 3). The percentages of counts eluting in the first peak were normalized to the percentage obtained when oligosaccharides were digested with @galactosidase alone.

+ P-Galactosidsse + @Galsctosidase and a-fucosidase

No additional treatment Acid hydrolysis Neuraminidases P-N-Acetylglucosaminidase Acid hydrolysis, and then P-N-

acetylglucosaminidase ,B-N-Acetylglucosaminidase

and neuraminidases

1.00 0.66 0.77 ND” 0.77 ND 0.76 0.40 0.51 ND

0.46 0.13

Endo-&galactosidase

’ ND, not determined.

0.59 0.30

coproteins and glycolipids, and can exist as tandem repeating disaccharide units (polylactosamine) terminating either in sialic acid or in a-linked galactose (42). Such polylactosamine repeats have been described in PC12 cells (43). The generation of lactosamine structures on these oligosaccharides was con- firmed by treatment with endo-p-galactosidase, which was more effective than P-N-acetylglucosaminidase at unblocking and removing protected galactose. (Assuming that the lacto- samine units in these cells terminated in N-acetylglucosa- mine, the product of endo-Sgalactosidase digestion would be a disaccharide, which would elute in the second peak on gel filtration chromatography.)

A well characterized modification of N-acetylglucosamine and galactose residues is the addition of fucose residues. Addition of fucose or other sugar residues renders both N- acetylglucosamine and lactosamine relatively resistant to their cognate glycosidases (44,45). To assess the role of fucose addition in protecting galactose residues on the isolated oli- gosaccharides, cu-fucosidase was included in some digestions. Including a-fucosidase substantially decreased the fraction of protected galactose in the presence or absence of P-N-acetyl- glucosaminidase or endo+galactosidase (Table IV). Since the a-fucosidase used in this study cleaves both (u1,2 and &,3- linked fucose residues (46), fucose residues linked al,3 to terminal N-acetylglucosamine (47-49) or cu1,2 to galactose (47) would have been removed. The presence of significant p- N-acetylglucosaminidase activity in the a-fucosidase used in this work raises the possibility that the effect of cy-fucosidase in the absence of added ,8-N-acetylglucosaminidase may have been at least partially due to this contamininating activity. Thus, we conclude that fucose residues were added to N- acetylglucosamine residues and/or to galactose residues in recultured PC12 Al cells.

Combining several of the treatments described above in order to remove sialic acid, N-acetylglucosamine, and fucose residues resulted in the conversion of approximately 87% of the resistant galactose to P-galactosidase sensitive. The small fraction of P-galactosidase-resistant galactose remaining un- der these conditions could be blocked by other covalent mod- ifications, or could be due to decreased efficiency in some of the enzymes in an increasingly dilute and complex solution.

In summary, it was found that the presence of sialic acid residues alone accounted for approximately one-quarter of the observed fi-galactosidase resistance in isolated oligosaccha- rides from recultured PC12 Al cells. Terminal N-acetylglu- cosamine accounted for an additional one-quarter, and fucose addition to N-acetylglucosamine and/or galactose residues acccounted for at least 40% of the observed resistance. All of these modifications to experimentally added galactose resi- dues upon reculture of the cells are consistent with transport of glycoproteins from the cell surface to the Golgi apparatus.

PC12 Al Cells Transported Many More Proteins to the Site of Sialytransferase Than Did CHO clone 13 Cells-It has been clearly demonstrated that transport to the Golgi apparatus from the cell surface is highly selective in nonsecretory cell types (3, 5, 6). This aspect of glycoprotein transport was addressed by comparing the proteins that were terminally glycosylated after exogalactosylation in PC12 Al cells and CHO clone 13 cells. Samples from the quantitation experi- ments presented in Table III were analyzed by SDS-polyacryl- amide gel electrophoresis and fluorography. Based on the data obtained by counting radioactivity in aliquots of the solubi- lized acid precipitates, equal counts of radioactivity were loaded on the gels where possible. Samples that contained too few counts were analyzed in their entirety. An important consequence of loading equal counts of radioactivity is this:

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

21276 Endocytic Go&i Traffic in Regulated Secretory Cells

Where equal counts of /3-galactosidase-resistant and undi- gested proteins were analyzed (PC12 Al cells), a band repre- senting a glycoprotein that is additionally glycosylated with an efficiency equal to the average for the total (14% in 6 h) would appear with equal intensity in the digested and undi- gested samples. Bands of greater intensity in the P-galactosid- ase-digested samples relative to the undigested samples rep- resent glycoproteins glycosylated with greater efficiency than the average for the total sample, and fainter bands correspond to glycoproteins glycosylated with efficiencies less than the average.

The results of this analysis are shown in Fig. 5, in which the relative protein loads in each lane are indicated. A similar small number of proteins incorporated the bulk of the labeled galactose in both CHO and PC12 clones (Fig. 5,0 chase lanes). A comparison of the patterns from undigested samples after 0 or 6 h of chase revealed minor differences in only a small number of bands in PC12 Al cells (arrowheads; Fig. 5, - p- galactosidase). In both cell types, the pattern observed for the P-galactosidase-resistant radioactivity after 6 h of culture was quite distinct from the pattern of undigested samples. In CHO

M, CLONE 13 PC12 Al x10- 3 205-

97.4-

66-

29-

18- 14-

DYE - FRONT

Chase (h) 0 6 6 0 6 6 O-gal. - + - - + -

Protein 1 10.5 1.5 1 7 1

FIG. 5. Comparison of total labeled glycoproteins and sial- vlated rrlvcooroteins in CHO clone 13 and PC12 Al cells. Samples-from-the quantitation experiments presented in Table III were analyzed by electrophoresis on a 10% polyacrylamide gel. For each cell type, equal counts were loaded in each lane of the gel, except in the caes of /3-galactosidase digested proteins from CHO clone 13 cells, which yielded insufficient counts. The relative protein load within each panel is indicated. A fluorograph of the dried gel is shown. CHO clone 13 cells glycosylated very little of the exogenous galactose during 6 h of reculture (three low mobility bands were detected when the protein load in the digested sample was 85 times that in the undigested sample instead of 10.5 times). In contrast, PC12 Al cells glycosylated a significant fraction of the galactose incorporated into a large number of proteins (6-h chase, + P-gulactosidase lane). Few of these glycosylated proteins comigrated with the major labeled proteins, and since the protein loads were unequal, bands that are fainter in the digested sample than in the undigested sample were glycosylated with less efficiency than the average (14%) for the total sample (see text). Reculturing for 6 h affected the abundance or mobility of only two of the major labeled glycoproteins (arrowheads, 0 chase, and 6-h chase, - fi-galuctosidase).

clone 13 cells, 85 times more /?-galactosidase-digested protein than undigested protein was required to detect a few faint bands of low mobility (in the experiment shown, only 10.5 times as much digested protein was loaded on the gel). The bands were much more numerous and prominent in the PC12 Al samples under the same conditions (Fig. 5, 6-h chase, + P-galactosiduse). Most of the bands that were prominent in the undigested samples from PC12 Al cells were faint or undetected in the digested sample, indicating that these pro- teins were not glycosylated efficiently. Other bands that were not prominent in the undigested samples were considerably enhanced in the digested sample (arrows). Analysis of PC12 A3 cells yielded similar results (not shown). Thus, the trans- port of cell surface glycoproteins to the site of glycosyltrans- ferase activities in PC12 cells was highly selective, efficiently including few if any of the major labeled glycoproteins. In contrast to CHO clone 13 cells, many more proteins that were not among the major labeled proteins acquired significant resistance to fl-galactosidase in PC12 Al cells.

In PC12 Al cells that had been recultured for 6 h, acid precipitable radioactivity was incorporated into material that migrated ahead of the dye front (Fig. 5, 6-h chase, + p- guluctosidase). This material was not seen in samples contain- ing equal total radioactivity (one-seventh as much precipi- tated lysate) that were not digested, nor was it detected in any samples from CHO clone 13 cells. This radioactivity may represent glycolipids that are efficiently acid-precipitated, and resistant to P-galactosidase digestion under the conditions employed in this experiment. If this radiolabeled material proves to be glycolipid, the ricin-resistant PC12 cells can be used to monitor transport of endogenous glycolipid as well as glycoprotein to the Golgi apparatus.

DISCUSSION

The ricin-resistant PC12 cells described in this report con- stitute the first system for measuring glycoprotein traffic from the cell surface to the Golgi apparatus biochemically in regu- lated secretory cells. As described previously for the mannose 6-phosphate receptors in nonsecretory cell lines (3, 7), some of the radiolabeled galactose incorporated into a subset of PC12 cell surface glycoproteins acquired additional terminal sugar residues upon reculturing the cells. Terminal glycosyl- ation was not observed when ricin-resistant PC12 cells were recultured at 19 “C, a condition that allows endocytic trans- port from the cell surface to endosomes, but prevents transfer to lysosomes (37, 38) and the Golgi apparatus (4, 6, 39, 40). These results strongly suggest that the acquisition of resist- ance to P-galactosidase we have measured reflects transport of cell surface glycoproteins to the Golgi apparatus in endo- crine cells.

The nature of the defect in ricin-resistant PC12 cells has not yet been defined. The PC12 mutants tested all had normal or slightly elevated levels of galactosyltransferase activity. Since the nonspecific galactose acceptor ovalbumin was used in this assay, a defect in fi-N-acetylglucosaminide pl,4-gal- actosyltransferase, the enzyme normally active in glycosyla- tion of N-acetylglucosamine residues in N-linked oligosaccha- rides, cannot be definitively ruled out (50, 51). However, the resistance of these cells to ricin toxin strongly suggests that they incorporated few galactose residues into any acceptors by any mechanism. At least two cell lines, CHO clone 13 (34) and MDCK (23), exhibiting the same glycosylation phenotype have been shown to be deficient in the transport of UDP- galactose from the cytoplasm to the lumen of the Golgi apparatus, and several other mutant cell lines display similar glycosylation phenotypes (52). A different genotype with the

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

Endocytic Golgi Traffic in Regulated Secretory Cells 21277

same phenotype is a defect in the UDP-galactose, UDP- galactosamine 4-epimerase (53). While the nature of the de- fect in PC12 mutants awaits definitive identification, the phenotype is clearly appropriate for the studies reported here.

In contrast to earlier results in nonsecretory cell types, galactose incorporated into N-linked oligosaccharides ac- quired N-acetylglucosamine and fucose residues in addition to sialic acid residues in ricin-resistant PC12 cells. In many cell types, sialytransferase activity is localized primarily to the tram Golgi cisterna and truns Golgi network (TGN), although several cell types have sialyltransferase in other Golgi cisternae (2). Therefore, addition of sialic acid residues alone to galactose incorporated at the cell surface has been interpreted as evidence for endocytic transport to the late Golgi. The site of synthesis of polylactosamine structures has not been as rigorously localized, but requires the simultaneous presence of galactoside$l,3-N-acetylglucosaminyltransferase and of N-acetylglucosaminide:pl,4-galactosyltransferase. The latter is generally localized to the tram Golgi cisterna (2, 42). Similarly, addition of fucose-linked al,6 to core N-ace- tylglucosamine residues occurs in medial Golgi compartments (42), but addition of fucose to galactose and to terminal N- acetylglucosamine residues in polylactosamine repeats cannot precede the addition of galactose in the truns Golgi cisterna( Thus, while the Golgi subcompartment that receives endocy- tic traffic in PC12 cells cannot yet be identified with certainty, it is likely a late Golgi compartment, presumably the trams cisterna and/or the TGN. Whether this represents a sig- nificant difference in membrane targeting between non-secre- tory cell types and PC12 (and possibly other secretory) cells, or simply reflects differences in glycosyltransferase distribu- tion in PC12 cells remains to be determined.

Several assumptions are inherent in the design and inter- pretation of these experiments. It is assumed that the range of structures of complex N-linked oligosaccharides are similar on all of the glycoproteins produced by a given cell type, that the exogalactosylation reaction labels all cell surface oligosac- charides with equal efficiency, and that all galactosylated oligosaccharides are equally good substrates for additional glycosylation, if they are transported to the Golgi apparatus. While it is likely that none of these assumptions is strictly true (42), they probably represent a reasonable approximation for many glycoproteins. However, Margolis et al. (43), using endo+galactosidase to expose acceptor sites for exogalacto- sylation, have presented data suggesting that many, but not all glycoproteins in PC12 cells carrypolylactosamine-contain- ing oligosaccharides. Finally, it is not known whether a single round of transport to the Golgi apparatus is sufficient to glycosylate fully all of the labeled galactose incorporated into oligosaccharides at the cell surface, or whether all of the galactose residues are substrates for additional glycosylation. If either is not the case, the percentage of galactose protected from P-galactosidase would be an underestimate of the per- centage of galactosylated molecules that have returned to the Golgi apparatus.

Transport of cell surface glycoproteins to the Golgi appa- ratus appeared to be quantitatively much greater in ricin- resistant PC12 cells than in the nonendocrine CHO clone 13 cells and involved many more glycoproteins. While a signifi- cant fraction of the mannose 6-phosphate receptors at the cell surface are transported to the TGN in CHO and other nonsecretory cells (3,4), no significant transport of total cell surface glycoproteins to the Golgi apparatus has been detected in several cell types using several different methods (3, 5, 6, 18). In ricin-resistant PC12 cells at least 16% of the cell surface N-linked oligosaccharides were transported to the

Golgi apparatus after 12 h. The dramatic difference between PC12 cells and other cell types in the magnitude of glycosyl- ation of cell surface galactose residues may merely reflect a more efficient action of a greater variety of glycosyltransfer- ases on the same number of substrates. If this is so, the measurements made in other cell types may have significantly underestimated the actual efficiency of this pathway. An alternative explanation is that the presence of a regulated secretory pathway in PC12 cells may amplify the endocytic Golgi pathway due to the recycling of secretory granule mem- brane components. The latter hypothesis is consistent with the demonstration that secretory vesicle membrane is turned over more slowly than the vesicle contents are secreted (ll), implying that the vesicle proteins are reutilized. The current data cannot definitively distinguish between these alterna- tives.

A fraction of the galactose incorporated into acceptors other than N-linked oligosaccharides in ricin-resistant PC12 cells exhibited properties suggesting that the acceptors were gly- colipids. This included up to 7% of the acid-precipitable /3- galactosidase-resistant galactose detected after 12 h of recul- ture, a fraction of which migrated ahead of the dye front on SDS-polyacrylamide gel electrophoresis. (It should be noted that the conditions employed in this study are not optimal for glycosidase treatment of glycolipids.) The little that is known about glycolipid traffic is consistent with its recycling from the cell surface to the Golgi apparatus. It has recently been shown that Shiga toxin, which binds to cell surface glycolipids, is internalized via clathrin-coated pits and tar- geted in part to the Golgi apparatus (54). Cholera toxin, which binds to the ganglioside GM1, has also been traced to the Golgi region (55). Glycolipid and sphingomyelin analogs incorpo- rated into the cell surface can be internalized and recycled along the same pathway as transferrin (56,57), and exogenous or endogenous glycolipids are sialylated at least in part by pathways independent of lysosomal turnover (58-60). If the galactose that is not in N-linked oligosaccharides in PC12 cells is in glycolipids, the ricin-resistant PC12 cells would represent a powerful system for the study of glycolipid traffic.

Part of the impetus for the current work came from a comparison of morphological studies of traffic from the cell surface to the Golgi apparatus in secretory and nonsecretory cell types. Due to the nature of the markers used in these studies, it was not possible to identify the specific elements of the pathway. Using the ricin-resistant PC12 cells described in this report, a quantitative analysis of endocytic transport to the Golgi apparatus for specific membrane markers is now feasible. Such studies may elucidate the relative contributions of different organelles, and the importance of the regulated secretory pathway, in generating this membrane transport pathway. These studies will complement both the morpholog- ical studies in regulated secretory cells and the quantitative biochemical data previously obtained in nonsecretory cell types.

Acknowledgments-We thank Stuart Kornfeld for providing CHO clone 13 cells and for many helpful suggestions and discussions, Pamela Stanley for sharing her extensive knowledge of glycosylation, and Judy White, Ira Mellman, Stephen Doxsey, Andre Brandli, and members of the Kelly lab for comments on the manuscript.

REFERENCES

1. Snider, M. D., and Rogers, 0. (1985) J. Cell Biol. 100, 826-834 2. Roth, J. (1987) Biochim. Biophys. Actu 906,405-436 3. Duncan, J. R., and Kornfeld, S. (1988) J. Cell Biol. 106,617-628 4. Jin, M., Sahagian, G. G., and Snider, M. D. (1989) J. Biol. Chem.

264,7675-7680 5. Reichner, J. S., Whiteheart, S. W., and Hart, G. W. (1988) J.

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

21278 Endocytic Go& Traffic in Regulated Secretory Cells

6. I.

8.

9. 10. 11. 12.

13. 14. 15.

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

26. 21.

28. 29.

30. 31.

32.

33.

34.

Biol. Chem. 263,16316-16326 Brandli, A. W., and Simons, K. (1989) EMBO J. 8, 3207-3213 Prydz, K., Brandli, A. W., Bomsel, M., and Simons, K. (1990) J.

Biol. Chem. 265, 12629-12635 Steinman, R. M., Mellman, I. S., Muller, W. A., and Cohn, Z. A.

(1983) J. Cell Biol. 96, l-27 Palade, G. E. (1975) Science 189,347-358 Winkler, H. (1977) Neuroscience 2, 657-683 Meldolesi. J. (1974) J. Cell Biol. 61. l-13 Farquhar,‘M. G., and Palade, G. E. (1981) J. Cell Biol. 91, (suppl.)

77-103 Kelly, R. B. (1985) Science 230, 25-32 Farquhar, M. G. (1981) Methods Cell Biol. 23,399-427 Herzog, V., and Farquhar, M. G. (1977) Proc. Natl. Acad. Sci. U.

S. A. 74,5073-5077 Orci, L., Ravazzola, M., Amherdt, M., Brown, D., and Perrelet,

A. (1986) EMBO J. 5,2097-2101 Ottosen, P. D., Courtoy, P. J., and Farquhar, M. G. (1980) J. Erp.

Med. 152, 1-19 Rauh, T. J., and Roberts, P. M. (1986) Exp. Cell Res. 165, 92-

106 von Bonsdorff, C.-H., Fuller, S. D., and Simons, K. (1985) EMBO

J. 4,2781-2792 Stieber, A., Gonatas, J. O., and Gonatas, N. K. (1984) J. Cell.

Physiol. 119,71-76 Briles, E. B., Li, E., and Kornfeld, S. (1977) J. Biol. Chem. 252,

1107-1116 Brew, K., Shaper, J. H., Olsen, K. W., Thayer, I. P., and Hill, R.

L. (1975) J. Biol. Chem. 250 1434-1444 Bran&i, A. W., Hansson, G. C., Rodriguez-Boulan, E., and Si-

mons, K. (1988) J. Biol. Chem. 263,16283-16290 Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 Lowe, A. W., Maddedu, L., and Kelly, R. B. (1988) J. Cell Biol.

106.51-59 Thilo, L. (1983) Methods Enzymol. 98, 415-421 Dahms, N. M., and Schnaar, R. L. (1983) J. Neurosci. 3, 806-

817 Laemmli, U. K. (1970) Nature 227,680-685 Tarentino. A.. Gomez. C. M.. and Plummer. T. H.. Jr. (1985)

Bioche&& 24,46k5-467i Chamberlain, J. P. (1979) Anal. Biochem. 98, 132-135 Laskey, R. A., and Mills, A. D. (1975) Eur. J. Biochem. 56,335-

341 Baenziger, J. U., and Fiete, D. (1979) J. Biol. Chem. 254,9795-

9799 Ebisu, S., Shankar Iyer, P. N., and Goldstein, I. J. (1978) Carbo-

hydr. Res. 61,129-138 Deutscher, S. L., and Hirschberg, C. B. (1986) J. Biol. Chem.

261,96-100

35. Schweitzer, E. S., and Kelly, R. B. (1985) J. Cell Biol. 101, 667- 676

36. Greene, L. A., and Rein, G. (19’77) Brain Res. 129, 247-263 37. Dunn, W. A., and Hubbard, A. L. (1984) J. Cell Biol. 98, 2148-

2159 38. Marsh, M., Bolzau, E., and Helenius, A. (1983) Cell 32,931-940 39. van Deurs, B., Sand& K., Petersen, 0. W., Olsnes, S., Simons,

K.. and Griffiths. G. (1988) J. Cell Biol. 106.253-267 40. van Deurs, B., Petersen; 0. W., Olsnes, S., and Sandvig, K. (1987)

Exp. Cell Res. 171,137-152 41. Griffiths, G., and Simons, K. (1986) Science 234,438-443 42. Kornfeld, R., and Kornfeld, S. (1985) Annu. Rev. Biochem. 54,

631-664 43. Margolis, R. K., Greene, L. A., and Margolis, R. U. (1986)

Biochemistry 25,3463-3468 44. Yamashita, K., Ohkura, T., Yoshima, H., and Kobata, A. (1981)

Biochem. Biophys. Res. Commun. 100,226-232 45. Scudder, P., Hanfland, P., Uemura, K., and Feizi, T. (1984). J.

Biol. Chem.259,6586-6592 46. Wiederschain, G. Y., and Rosenfeld, E. L. (1971) Biochem. Bio-

phys. Res. Commun. 44,1008-1014 47. Kobata, A. (1984) in Biology of Carbohydrates (Ginsburg, V., and

Robbins, P. W., ed) Vol. 2, pp. 87-161, Wiley Interscience, New York

48. Fukuda, M., Bothner, B., Ramsamooj, P., Dell, A., Tiller, P. R., Varki, A., and Klock, J. C. (1985). J. Biol. Chem. 260, 12957- 12967

49. Stanley, P., and Atkinson, P. H. (1988) J. Biol. Chem. 263, 11374-11381

50. Beyer, T. A., Sadler, J. E., Rearick, J. I., Paulson, J. C., and Hill, R. L. (1981) Adu. Enzymol. Relat. Areas Mol. Biol. 52, 23-175

51. Yamashita, K., Tachibana, Y., and Kobata, A. (1978). J. Biol. Chem. 253,3862-3869

52. Stanley, P. (1984) Annu. Reu. Genet. 18, 525-552 53. Kingsley, D. M., Kozarsky, K. F., Hobbie, L., and Krieger, M.

(1988) Cell 44,749-759 54. Sandvig, K., Olsnes, S., Brown, J. E., Petersen, 0. W., and van

Deurs, B. (1989) J. Cell Biol. 108.1331-1343 55. Gonatas, N. K., Stieber, A., Gonatas, J., Mommoi, T., and Fish-

man. P. H. (1983) Mol. Cell Biol. 3.91-101 56. Kok, J: W., Eskelinen, S., Hoekstra, K., and Hoekstra, D. (1989)

Proc. N&l. Acad. Sci. U. S. A. 86,9896-9900 57. Koval, M., and Pagano, R. E. (1989) J. Cell Biol. 108,2169-2181 58. Sonderfeld, S., Conzelmann, E., Schwarzmann, G., Burg, J.,

Hinrichs, U., and Sandhoff, K. (1985) Eur. J. Biochem. 149, 247-255

59. Ghidoni, M., Trinchera, M., Venerando, B., Fiorilli, A., Sonnino, S., and Tettamanti, G. (1986) Biochem. J. 237,147-155

60. Usuki, S., Lyu, S.-C., and Sweeley, C. (1988) J. Biol. Chem. 263, 6847-6853

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from

S A Green and R B Kellyline.

Endocytic membrane traffic to the Golgi apparatus in a regulated secretory cell

1990, 265:21269-21278.J. Biol. Chem. 

  http://www.jbc.org/content/265/34/21269Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/265/34/21269.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on Novem

ber 18, 2020http://w

ww

.jbc.org/D

ownloaded from