Regulation of α1-acid glycoprotein externalization and intracellular accumulation in glucocorticoid-induced rat hepatoma cells

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 246, No. 1, April, pp. 449-459, 1986

Regulation of q-Acid Glycoprotein Externalization and Intracellular Accumulation in Glucocorticoid-Induced Rat Hepatoma Cells’ OMAR K. HAFFAR,2 CAROLINE P. EDWARDS, Department of Physiology-Anatmy,

AND

University of California,

GARY L. FIRESTONE3 Berhzley, California 94720

Received August 21,1985, and in revised form November

27,1985

al-Acid glycoprotein (CQ-AGP) is a glucocorticoid inducible gene product that is synthesized and secreted by certain rat hepatoma tissue culture (HTC) cell lines such as M1.54. Exposure to monensin, a Na+-K+ ionophore, causes a significant redistribution of ai-AGP into two distinct fractions; immunoprecipitation of [?S]methionine-labeled proteins revealed that a 27% decrease in secretion accounts for a sixfold increase in accumulation of a stable intracellular species. The new intracellular cui-AGP is more heterogeneous than normal while the extra-cellular form is 6000 Da smaller than normal. These effects are due to selective alterations in carbohydrate maturation; endo-P-Nacetylglucosaminidase H (endo H) digestion demonstrated that both (Y~-AGP species contain variable numbers of endo H-resistant oligosaccharide side chains ranging between zero and five. Ricin affinity chromatography revealed that the attachment of galactose residues is strikingly correlated with (Y~-AGP externalization while neuraminidase digestions demonstrated that sialic acid attachment appears unessential for its secretion. Taken together, our results suggest that in the presence of monensin the cellular transport of intracellular destined and externalized (Y~-AGPproceeds in common through the early segments of the Golgi and at a point prior to or at the compartment containing galactosyl transferase, q-AGP becomes committed for secretion. o I986 Academic Press. Inc.

q-Acid glycoprotein (cu~-AGP)~is one of a series of acute phase reactants that is secreted from the liver in response to vari This paper is dedicated to the memory of Edward C. Heath, who supervised G.L.F.‘s doctoral research from 1975 to 1980 and whose guidance and friendship will be remembered always. This investigation was supported by the Public Health Service Grant CA 35547 awarded by the National Cancer Institute, DHHS, G.L.F. is a recipient of a National Science Foundation Presidential Young Investigator Award (PCM 8351884). a This work will be submitted by O.K.H. in partial fulfillment of the requirements for the Ph.D. degree. a To whom reprint requests and all correspondence should be addressed. ’ Abbreviations used: a,-AGP, al-acid glycoprotein; endo H, endo-@-N-acetylglucosaminidase H, PBS, phosphate-buffered saline; ppm, parts per million; DTT, dithiothreitol; SDS, sodium dodecyl sulfate; RER, rough endoplasmic reticulum.

ious types of biological stress such as inflammation, surgery, and bacterial infections (l-5). Under these conditions, the absolute level of al-AGP produced by the liver significantly increases due to a higher steady-state concentration of mRNA (4,6). Glucocorticoid hormones mediate this effect in whole animals, isolated hepatocytes, and in certain hepatoma cell lines (4,6-8); al-AGP mRNA appears to be stabilized in the presence of hormone with relatively little effect on the rate of transcript synthesis (9). Similar to other secreted and membrane-associated glycoproteins (10, ll), al-AGP is translated on rough endoplasmic reticulum bound polysomes, the signal peptide cleaved during synthesis and the polypeptide backbone modified by the cotranslational attachment of either five or six “high-mannose” oligosaccharide side chains depending on the species or cell line 449

0003-9861/86 $3.00 Copyright 0 1986 by Academic Press. Inc. All rights of reproduction in any form reserved.

450

HAFFAR.

EDWARDS. AND FIRESTONE

(4, 12-14). Biochemical analysis demonstrated that the carbohydrate portion of this glycoprotein represents between 40 and 55% of its apparent molecular weight (15, 16). As newly synthesized (rI-AGP is transported through discrete cisternae segments of the Golgi, the oligosaccharide side chains are processed in sequential fashion into complex structures (15-19) while the polypeptide backbone remains intact (20). The resulting mature form of +AGP is presumably carried by secretory vesicles from the Golgi and then externalized after the vesicle fuses with the plasma membrane. The “high mannose” oligosaccharide side chains that are stably attached to aI-AGP have the general structure (G~cNAc)~(Man)g while secreted aI-AGP contains complex carbohydrate moieties composed of less mannosyl residues but with varying levels of the terminal sugars, N-acetylglucosamine, galactose, sialic acid, and fucose (15, 16, 18). The selective processing of these oligosaccharides suggests a precise intracellular sorting route from the RER through the Golgi and finally to the extracellular environment. However, virtually nothing is known about structural features on a1-AGP that regulate or facilitate its intracellular trafficking. Conceivably, vesicle-associated receptors recognize structural signals within the polypeptide backbone (11,21), while other lectin-like receptors may recognize specific portions of the carbohydrate moiety similar to the role of mannose 6-phosphate in the compartmentalization of certain lysosomal hydrolases (22-24).

To begin to investigate oligosaccharide modification reactions that are correlated with or perhaps responsible for the intracellular transport of glycoproteins, protein trafficking through the Golgi was disrupted by exposure to the Na+-K+ ionophore, monensin (25, 26), and the externalization and processing of al-AGP examined. Monensin causes a significant redistribution in aI-AGP compartmentalization resulting in expression of a new stable intracellular form of this glycoprotein. Partial analysis of oligosaccharide structures of both the intracellular and secreted

species suggests a correlation between the attachment of certain terminal sugar residues and the commitment for externalization of aI-AGP. EXPERIMENTAL

PROCEDURES

Materials. All media and sera were purchased from the UCSF Tissue Culture Facility. L-[?S]Methionine (1000 Ci/mmol) was obtained from Amersham Corp. (Arlington Heights, Ill.) and En*Hance from New England Nuclear (Boston, Mass.). Dexamethasone, monensin, and galactose were purchased from Sigma Chemical Company (St. Louis, MO.), Pansorbin from Calbiochem (La Jolla, Calif.), agarose derivatized with castor bean lectin (Ricin-120) and neuraminidase from P-L Biochemicals, Inc./Pharmacia (Milwaukee, Wis.), and X-ray film was purchased from Merry X-Ray Chemical Corp. (Burlingame, Calif.). The anti-al-acid glycoprotein antisera (4) was a generous gift from John Taylor (Molecular Biology Unit, Gladstone Foundation Laboratories, San Francisco, Calif.). All other reagents were of highest available purity. Cells and method of culture. Ml.54 cells, a clonal MTV-infected rat hepatoma (HTC) cell line containing 10 integrated proviruses, were described previously (2’7, 28). Cultures were propagated as monolayer in Dulbecco’s modified Eagle’s medium supplemented with 10% horse serum at 3’7’C in a humid atmosphere of 5% COz:95% air. Cells were radiolabeled with [S6Sjmethionine in methionine-deficient medium containing 0.5% dialyzed fetal calf serum; the amino acid free media was employed 30 min prior to the addition of the radiolabel. Unless otherwise mentioned, incubations were carried out in the presence of 1 *M dexamethasone and the indicated concentrations of inhibitor. Steady-state radiolabeling and separation of wllular and secreted fradions. Cells pretreated with the indicated concentrations of monensin for 1 h were incubated with 1 PM dexamethasone for 16 h in the presence of ionophore, and subsequently radiolabeled with 30 &i/ml pS]methionine for 10 h. Both hormone and the appropriate concentration of monensin were present during the radiolabeling period. Labeling media was gently collected, centrifuged at 2000g for 10 min and the supernate harvested as the secreted fraction. The cells were washed 3X with phosphatebuffered saline (PBS), dislodged from the plate with 5 mM EDTA, 10 mM Tris-HCl, pH 7.5, in PBS and cellular fraction collected by centrifugation at 6OOg for 10 min. Both cellular and secreted fractions were solubilized in detergents and then either precipitated with 10% trichloroacetic acid or immunoprecipitated with appropriate antibodies as outlined below. Pulse-chase experiment Cell cultures were preincubated in the presence or absence of 1 PM monensin for 1 h, then supplemented with 1 pM dexamethasone.

GLYCOPROTEIN

TRAFFICKING

After 16 h, the cells were incubated for 30 min in methionine-free media and then pulsed with 50 pCi/ ml pS]methionine for 10 min. After quickly rinsing the cells several times in PBS, the radiolabel was chased by incubating the cells in medium containing 0.5% fetal calf serum and 0.3 mg/ml unlabeled methionine; appropriate concentrations of dexamethasone and monensin were included in the pulse and chase medium throughout the time course. After 0, 30,60,120, and 180 min of chase, cellular and secreted fractions were harvested and expression of ai-AGP examined as described below. Immunoaakorption of deteqwnt sdutrilized cells. Cell pellets (10’ cells) were homogenized in 1 ml of solubilization buffer (1% Triton X-100,0.5% deoxycholate, 5 mM EDTA, 250 mM NaCl, 25 mM Tris-HCl, pH 7.5) and centrifuged 10 min at 20,OOOgand supernatant fractions were utilized as sources for solubilized material. Secreted fractions were brought to 1% Triton X-100,0.5% deoxycholate, and 5 mM EDTA just prior to immunoprecipitation. The solubilized fractions (cellular and secreted) were analyzed for total radiolabeled proteins by precipitating a small aliquot (10 ~1) with 10% trichloroacetic acid. Pansorbin, formaldehyde fixed Cowen 1 strain of the bacterium Stuphykmccus aureas containing cell surface associated protein A (Staph A), was utilized as an adsorbant for antibody:antigen complexes. Solubilized cell extracts (300-500 ~1) were 6rs.t preadsorbed without antibodies with 10 pl of a 10% suspension (w/v) of fixed Staph A in solubilization buffer; simultaneously the Staph A employed for the immunoadsorption was preadsorbed with unlabeled detergent solubilized cell extracts. After 15 min at room temperature, preadsorbed radiolabeled extracts and Staph A were each harvested by centrifugation for 3 min in an Eppendorf centrifuge. Subsequently, the radiolabeled extracts were added to 100 ~1 solubilization buffer containing 50 mg/ml BSA and 10 ~1 of a l/10 dilution (in PBS) of either immune or preimmune serum. After 10 min at room temperature, 10 ~1 of 10% Staph A (preincubated with unlabeled cell extracts and resuspended in solubilization buffer) was added and the mixture incubated for an additional 5 min. The entire reaction mixture was layered over 600 ~1 of 1 M sucrose in solubilization buffer, centrifuged for 2 min in an Eppendorf centrifuge and the nonadsorbed material removed by aspirating to the sucrose interface. The tubes were washed by overlaying the sucrose cushion with 2 M urea in solubilization buffer containing 500 mM NaCl and aspirating to the Staph A pellets. The pellets were washed first with solubilization buffer, and then with 5 mM EDTA, 10 mM Tris-HCl, pH 7.5. Final pellets were then either stored at -2O’C or immediately reimmunoprecipitated as described below. To further reduce nonspecific adsorption of radiolabeled proteins, the antigen-antibody complexes were released from the Staph A in 30 pl of 1% SDS by

AND

EXTERNALIZATION

451

incubating 20 min at 37°C followed by 2 min Eppendorf centrifugation; the supernatant fractions were added to 500 pl of solubilization buffer containing 10 mg/ml BSA and 10 ~1 of a l/l0 dilution of the appropriate immune or preimmune serum. After 15 min at room temperature, 10 pl of 10% preadsorbed pansorbin was added and the mixture incubated 5 min at room temperature. Staph A-immunocomplex pellets were collected by centrifugation in an Eppendorfcentrifuge, washed in solubilization buffer and then in 10 mM Tris-HCl, pH 7.5,5 mM EDTA. Final Staph A pellets were either stored at -20°C or prepared immediately for SDS gel electrophoresis. With the second immunoadsorption, backgrounds are reduced from 2500 parts per million (ppm) to approximately 25 ppm while the levels of specifically bound material are reduced by only 30%. SDS-gel electrophoresis and autoradiography. Immunoadsorbed ai-AGP was prepared for electrophoretie analysis by incubating the final Staph A pellets with 25 ~1 SDS-Gel sample buffer (29) and 5 ~1 of 0.5 M DTT at 100°C for 2 min. After a 2 min Eppendorf centrifugation, the supernatant fractions were electrophoresed in SDS polyacrylamide gels (30) containing 10% acrylamide unless indicated otherwise. Proteins were subsequently fixed and stained by incubation overnight with 0.4% Coomassie brilliant blue in 10% acetic acid, 25% isopropyl alcohol. Gels were destained in 10% acetic acid for 4 h, impregnated with En’Hance for 1 h, incubated in water for 1 h, dried under vacuum at 6O”C, and analyzed by fluorography on Kodak RP-Royal X-Omat film at -70°C. R&in-agarose column chromatography. Ricin-agarose was poured to a bed volume of 0.5 ml and washed extensively with 0.1% Triton X-100, 150 mM NaCl, 10 mM Tris, pH 7.5, (ricin binding buffer). Solubilized cell extracts were diluted to a final concentration of 0.1% Triton X-100 by the addition of 10 mM Tris-HCl, pH 7.5, 150 mM NaCl while secreted fractions were brought to 0.1% Triton X-100. The solubilized cellular and secreted samples were applied to the ricinagarose column, washed extensively in ricin binding buffer, and ricin-bound material eluted with 3% galactose in ricin-binding buffer. Bound, nonbound, and starting fractions were immunoprecipitated and analyzed by SDS-polyacrylamide gel electrophoresis. Neuraminidose digestion To cleave sialie acid residues, final Staph A pellets were resuspended in 500 ~1 incubation buffer (50 mM sodium acetate, pH 4.9,2 mM CaClp, 0.2 mM EDTA) in the presence or absence of 1.4 mg/ml neuraminidase, and incubated for 24 h at 37°C with continuous shaking. The reaction was terminated by sedimenting the Staph A pellets in an Eppendorf centrifuge for 3 min and the pellets prepared for electrophoresis by washing first in 10 mM Tris-HCl, pH 7.5,5 mM EDTA before the addition of SDS-polyacrylamide gel sample buffer (see above). End.o H digestion Immunoprecipitated (Y,-AGP was released from the antibody-Staph A complex in 1%

452

HAFFAR,

EDWARDS,

SDS at 37°C for 10 min. After a 3-min Eppendorf centrifugation, the supernatants were added to 150 ~1 100 mM sodium citrate buffer, pH 5.5, containing 80 mu/ml endo H (31); control incubations did not contain enzyme. After incubating 24 h at 37°C with continuous shaking, the reactions were terminated by adding 400 ~1 of immunoprecipitation (solubilization) buffer containing 10 pl of l/10 diluted ai-AGP antibody. The antigen-antibody complexes were precipitated with Staph A, washed, and subsequently fractionated by SDS-polyacrylamide gel electrophoresis as mentioned previously. RESULTS

Dose-dependmt effects of wumens in on crlAGP expression. To begin to assess poten-

tial relationships between oligosaccharide processing and the intracellular sorting and secretion of al-acid glycoprotein ((~iAGP), glucocorticoid-induced Ml.54 cells were radiolabeled with [35S]methionine in the presence of various concentrations of monensin and the absolute level of extracellular and intracellular glycoprotein quantitated by immunoprecipitation. Monensin treatment causes a significant shift in the steady-state distribution of (pi-AGP with essentially no effect on the total level of expressed glycoprotein (Fig. 1). At the highest tested concentration of ionophore, intracellular ai-AGP is increased by sixfold with a concomitant decrease (approximately 27%) in its externalization. At steady state, the vast majority of (pi-AGP is secreted and as also shown in Fig. 1 (graph insert) the loss in extracellular glycoprotein accounts for the sixfold accumulation of intracellular Q-AGP. Consistent with this notion, monensin does not affect the production of functional ai-AGP mRNA as monitored by in vitro translation of RNA isolated from ionophore-treated and untreated cells (data not shown). Monensin also alters the nature of both the intracellular and secreted forms of cui-AGP. Electrophoretic analysis of [““S]methionine-labeled ai-AGP revealed that thespeciessecretedfrommonensin-treated cells is approximately 6000 Da smaller in apparent molecular weight (Fig. 1, gel insert lane C vs D) whereas the accumulated intracellular form is strikingly more heterogeneous and on the average approxi-

AND

FIRESTONE

mately 4000 Da smaller than normal (lane A vs B). As shown in later sections of this study, these differences reflect selective disruptions in oligosaccharide side chain maturation. Kinetics of al-AGP expressicm, At steady state, monensin causes a stable redistribution of a portion of extracellular (u~-AGP into an intracellular compartment. Therefore, to examine the effects of monensin on the initial rate of intracellular accumulation and externalization of ai-AGP, ionophore-treated and untreated cells were pulse-labeled with [35S]methionine for 10 min and subsequently incubated with excessunlabeled methionine for the indicated time periods. As shown in Fig. 2 (lanes AE), in the absence of monensin, newly synthesized ai-AGP can be completely chased from the intracellular fraction into the extracellular environment (Fig. 3, lanes AE), leaving essentially no ai-AGP remaining in the cell. Densitometric analysis of this gel revealed that the half-time of this process is 63 min. In contrast, monensin causes a significant accumulation of a new, more heterogeneous ai-AGP species that displays an apparent molecular weight between 36,000 and 41,000 Da (Fig. 2, lanes F-J). Furthermore, this heterogeneity and accumulation result from altered posttranslational reactions since the (Y~-AGP species initially synthesized in monensintreated and untreated cells are identical in size and quantity (Fig. 2, lane A vs F). Analysis of the secreted material (Fig. 3, lanes F-J) revealed that monensin significantly reduces the initial rate of externalization of (hi-AGP. A low level of WAGP is first detected after 120 min chase, whereas, untreated cells competently secrete this glycoprotein within 60 min. (Fig. 3, lanes A-E). The overall externalization remains reduced at steady state since ionophore-treated cells produce 27% less extracellular (wi-AGP over a 24-h period (Fig. 1). Taken together, our results demonstrate that in monensin-treated cells, one heterogeneous (Y~-AGP species accumulates in an intracellular compartment while a second form of lower size is secreted at reduced levels. This redistribution of CX~-AGP into two distinct fractions was exploited to

GLYCOPROTEIN

TRAFFICKING

453

AND EXTERNALIZATION

secmled

MON

MONENSIN

cellubr

-

(JIM)

FIG. 1. Dose-dependent effects of monensin on the expression of cellular and secreted ai-AGP. Ml.54 cells were incubated with the indicated concentrations of monensin and 1 PM dexamethasone for 16 h and subsequently radiolabeled with 125 pCi [35S]methionine for 10 h in the presence of both ionophore and hormone. Labeling media was collected, centrifuged at 2OOOgfor 10 min, and the supernatant harvested as the secreted fraction. The cells were rinsed 3X with PBS, dislodged from the plate, and harvested by centrifugation at 600s for 10 min. Both cellular and secreted fractions were detergent solubilized, immunoprecipitated with either anti-oli-AGP or preimmune sera and immunoprecipitated proteins directly counted by liquid scintillation. The relative levels of cellular and secreted (Y~-AGPwere quantitated at each concentration of monensin by calculating the level of specifically immunoprecipitated glycoprotein (immune minus preimmune radioactivity). Total expression of radiolabeled (Y~-AGP(graph insert) is based on immunoprecipitable protein per total radiolabeled material precipitated with 10% trichloroacetic acid, 100% expression is the amount detected in the absence of ionophore. Gel insert: Dexamethasone-induced cells were incubated in the presence (lanes B, D, and F) or absence (lanes A, C, and E) of 1 pM monensin as described above and radiolabeled with 125 pCi p”S] meth’lonine for 10 h in the presence of both hormone and ionophore. Harvested cellular (lanes A, B, E, and F) and secreted (lanes C and D) fractions were detergent solubilized, immunoprecipitated with either anti-a,-AGP (lanes A-D) or preimmune (lanes E and F) sera, immunoprecipitated material fractionated by SDS-polyacrylamide gel electrophoresis and analyzed by fluorography. Molecular weight standards are conalbumin (76,000 M,), bovine serum albumin (68,000Mr), bovine y-globulin (50,000&f,), ovalbumin (43,000Af,), glyceraldehyde phosphate dehydrogenase (36,000 Af,), and trypsin inhibitor (20,000 A&).

examine oligosaccharide processing reac- weight in SDS polyacrylamide gels. In fact, tions and other structural features that are the sialic acid residues account for a sigassociated with its externalization. nificant portion of ai-AGPs electrophoretic Neuraminidase digestion of aI-AGP. mobility. Therefore, this phenomenon was Studies in our laboratory (data not shown) utilized to examine potential correlations as well as by other investigators (15, 16, between the attachment of sialic acid and 18) have shown that the carbohydrate secretion of al-AGP. Immunoprecipitated moiety of al-AGP represents approxiglycoprotein was incubated in the presence mately 40-50% of its apparent molecular or absence of neuraminidase and analyzed

454

HAFFAR,

EDWARDS,

AND

FIRESTONE

Ricin a&&y

*

-9 -*OWENSIN

+ YOWENSIN

+YON

FIG. 2. Pulse-chase analysis of cellular a,-AGP produced in the presence or absence of monensin. Hormone-induced Ml.54 cells treated in the presence (lanes F-L) or absence (lanes A-E) of 1 pM monensin were pulse labeled with 200 &i [%]methionine for 10 min followed by further incubation with excess unlabeled methionine. Dexamethasone, and when appropriate monensin, were included in the culture media throughout the experiment. The cellular fractions were harvested immediately after the pulse and after 30,60,120, and 180 min of subsequent chase. Detergent solubilized extracts were immunoprecipitated with either anti-or-AGP (lanes A-J) or preimmune (lanes K and L) sera and immunoprecipitated material fractionated by SDS-polyacrylamide gel electrophoresis and analyzed by fluorography. The molecular weight markers are described in Fig. 1 with the addition of phosphorylase b (97,000 M,) and /I-galactosidase (120,000 1M,).

electrophoretically. As shown in Fig. 4, neuraminidase treatment did not alter the gel mobility of CQ-AGP secreted from monensin-treated cells (lanes G vs H) while reducing the apparent size of extracellular ai-AGP in untreated cells by 3000 Da (lane E vs F). In addition, the migration of cellular forms of al-AGP expressed in either ionophore-treated or untreated cells is not affected by neuraminidase (lanes A-D). Thus, the failure of al-AGP to acquire sialic acid in monensin-treated cells accounts for 3000 Da of the 6000-Da shift in apparent size. More importantly the attachment of sialic acid appears unessential for its externalization.

chromatography

of cellular

and extracellular aI-AGP. Rein affinity chromatography revealed that the attachment of galactose (32), a terminal sugar residue added prior to sialic acid, can be correlated with the externalization of alAGP in monensin-treated cells. Detergent solubilized cellular and secreted r”S]methionine-labeled material was applied to ricin-agarose columns, eluted with 3% galactose, and the starting (total), nonbound (pass through), and bound fractions then assayed for al-AGP by immunoprecipitation. As shown in Fig. 5, most ai-AGP secreted in the presence or in the absence of monensin contains galactose as evident by binding to ricin (lanes K and L). Approximately 5% of the al-AGP externalized from monensin-treated cells (Fig. 5, lane J) fails to bind ricin and migrates slightly faster than the ricin-binding species. Thus, the vast majority of extracellular CQ-AGP 0

30

60

120 160

0

30

60

120

180

0

66

chase

Mr

-36

-

w -MONENSIN

+MONENSIN

+NJN

FIG. 3. Pulse-chase analysis of (or-AGP secreted in the presence or absence of monensin. Hormone-induced cells were incubated in the presence (lanes FL) or absence (lanes A-E) of 1 PM monensin and radiolabeled exactly as described in Fig. 2. At the indicated times of chase, harvested secreted fractions were detergent solubilized, immunoprecipitated with anti-Lur-AGP (lanes A-J) or preimmune (lanes K and L) sera, and immunoprecipitated material electrophoretically fractionated and analyzed by fluorography. The molecular weight markers are described in Fig. 1.

GLYCOPROTEIN

TRAFFICKING

secreted E

F

G

* H

4 76 66 50 43

36

20 M NEUR

AND

455

EXTERNALIZATION

cellular fractions were digested with endofi-N-acetylglucosaminidase H (endo H) and analyzed electrophoretically. Endo H probes side chain structures that had undergone early Golgi processing events; complex oligosaccharides initially trimmed of most of their mannose residues and with some newly acquired N-acetylglucosamine become resistant to this enzyme (10, 3335). As shown in Fig. 6, intracellular alAGP expressed in the absence of ionophore is completely sensitive to endo H digestion (lanes A and B); the resulting 23,000 M, species comigrates with endo H-treated QAGP that had been radiolabeled only during a lo-min pulse of [35S]methionine (data

--++ -

+

-

+

FIG. 4. Neuraminidase digestion of cellular and secreted (u~-AGP. Dexamethasone-treated Ml.54 cells were exposed in the presence (lanes C, D, G, and H) or in the absence (lanes A, B, E, and F) of 1 PM monensin and radiolabeled with 125 &i [?S]methionine as described in the text. Cellular and secreted fractions were individually harvested, detergent solubilized, and immunoprecipitated with anti-ai-AGP serum. The Iinal Staph A pellets were suspended in the presence (lanes B, D, F, and H) or absence (lanes A, C, E, and G) of 1.4 mg/ml neuraminidase and digestions continued at 37°C for 24 h (see Experimental Procedures). Digested and undigested samples were fractionated by SDS-polyacrylamide gel electrophoresis and radioactive bands visualized by fluorography. The acrylamide concentration was 11% and the molecular weight markers are described in Figs. 1 and 2.

contains terminal nonreducing galactosyl residues in its oligosaccharide side chains. In contrast, essentially all of the intracellular al-AGP that accumulates in the presence of monensin fails to bind ricin (Fig. 5, lane D vs F). Thus, two distinct forms of CY~-AGPare detected in monensintreated cells, a stable cellular form lacking galactosyl residues and an externalized species that acquires galactose. Endo H sensitivity of aI-AGP oligosaccharide sidechains. To examine at what point in glycoprotein trafficking al-AGP becomes committed to an intracellular compartment or the extracellular environment, [35S]methionine-labeled al-AGP immunoprecipitated from extracellular or

Cellular

Secreted

FIG. 5. Ricin-agarose chromatography of al-AGP expressed in monensin-treated and untreated cells. Dexamethasone-induced Ml.54 cells treated in the presence (lanes B, D, F, H, J, and L) or absence (lanes A, C, E, G, I, and K) of 1 PM monensin were radiolabeled with 250 PCi [86S]methionine as described in the text. Both cellular (lanes A-F) and secreted (lanes GL) fractions were individually harvested, detergent solubilized, diluted with buffer until the final concentration of Triton X-100 was 0.1% and applied to separate 0.5 ml ricin-agarose columns. After collecting the pass through (PT) fractions, the ricin bound (B) material was eluted with 3% gala&se in ricin-binding buffer. The peak fractions from the pass through (lanes C, D, I, and J) and bound (lanes E, F, K, and L) material were pooled separately and immunoprecipitated with anti-cY1-AGP serum in parallel with starting samples (total: lanes A, B, G, and H). Immunoprecipitated proteins were fractionated eleetrophoretieally and analyzed by fluorography. The molecular weight markers are described in Figs. 1 and 2.

456

HAFFAR, *

cellular

*

secreted

*

EDWARDS,

t

Mr

-36

MONENDOH-

-++

--++

+-+

-+-+

FIG. 6. Endo H digestion of ai-AGP expressed in monensin-treated and untreated cells. ori-AGP was immunoprecipitated from both cellular (lanes A-D) and secreted (lanes E-H) fractions harvested from [?S]methionine labeled cells that had been treated with (lanes C, D, G, and H) and without (lanes A, B, E, and F) 1 ELMmonensin. Immunoprecipitated a,-AGP was released from the Staph A pellets in 1% SDS and digested in the presence (lanes B, D, F, and H) or absence (lanes A, C, E, and G) of endo H as described under Experimental Procedures. Digested and undigested samples were fractionated by SDS-polyacrylamide gel electrophoresis and analyzed by fluorography. The molecular weight standards are described in Figs. 1 and 2.

not shown). In contrast, endo H digestion of intracellular (Y~-AGP expressed in monensin-treated cells revealed a ladder of bands which corresponds to individual species that contain variable numbers of endo H-resistant oligosaccharides between zero and five side chains (lanes C and D). Thus, although its oligosaccharide processing is clearly heterogeneous and incomplete, densitometric analysis of Fig. 6, lane D, revealed that 60% of cellular ai-AGP contains at least one endo H-resistant carbohydrate side chain and therefore migrated in the Golgi beyond the step where the carbohydrate moieties become resistant to this glycosidase (Fig. 7). Intracellular ai-AGP appears equally distributed between species with one to five endo Hresistant oligosaccharides. Glycosidase

AND

FIRESTONE

digestion of (u~-AGP secreted from monensin-treated cells, revealed the same overall ladder of expressed species (Fig. 6, lanes G and H) even though a higher percentage (40%) is completely endo H resistant. Moreover, the high percentage (60%)of secreted al-AGP containing endo H sensitive carbohydrate units and the overall size heterogeneity suggests the existence of unusual oligosaccharides. In fact, Peters and co-workers have shown that human chorionic gonadotropin secreted in the presence of monensin contains several types of incompletely processed oligosaccharides (36). Taken together, our data suggest that the early Golgi transport steps of the accumulated intracellular and externalized ai-AGP appear similar, while terminal oligosaccharide processing suggest that a rate limiting step or divergence in their sorting occurs at a later Golgi step. DISCUSSION

A sequential series of post-translational processing and compartmentalization reactions function during the externalization of (u~-AGP after its synthesis in the RER.

1 II 0

I 1

I 2

NUMBER OF END0 OLIGOSACCHARIDE

I 3

I 4

II 5

H RESISTANT SIDE CHAINS

FIG. 7. Heterogeneity of oligosaccharide side chains attached to intracellular and secreted (pi-AGP. Fluorographs representing endo H digested cellular (Fig. 6, lane D) and extracellular (Fig. 6, lane H) (Y,-AGP were scanned with a soft laser densitometer and the peaks integrated. The percentage of (u,-AGP with each number (zero to five) of endo H-resistant oligosaccharide side chains was determined and plotted as indicated.

GLYCOPROTEIN

TRAFFICKING

Immunocytochemical, genetic, and carbohydrate structural evidence suggest that glycoproteins such as ai-AGP may enter any one of several potential transport routes through the Golgi (37-42). al-AGP’s precise pathway through this organelle and other vesicle components is unknown. Two distinct forms of al-AGP were detected and analyzed in monensin-treated cells, a stable species that accumulates in an intracellular fraction and a second form that is secreted at a slightly lower rate. This unique redistribution of al-AGP localization was exploited to examine oligosaccharide structural features that are closely associated with and perhaps regulate its externalization. The primary correlation between alAGP externalization and Golgi-mediated oligosaccharide processing is the attachment of terminal sugar residues, in particular galactose. Ricin lectin affinity chromatography revealed that the accumulated intracellular form of cul-AGP lacks terminal nonreducing galactose residues; whereas, most secreted ai-AGP is recognized by ricin and eluted with galactose (Fig. 5). Conceivably, a branch point or rate limiting step in glycoprotein trafficking regulates al-AGP sorting at or prior to the Golgi compartment containing galactosyl transferase. The form of cul-AGP that becomes committed to externalization continues through or enters a transport pathway that eventually allows the acquisition of galactose. This putative branch point or rate limiting step in CQ-AGP sorting most likely occurs in the Golgi at a step distal to the trimming of mannosyl residues from high mannose oligosaccharides and subsequent attachment of some N-acetylglucosamine. Both the intracellular and extracellular species contain approximately the same level of oligosaccharide heterogeneity as monitored by digestion with endo H; each form of (Y~-AGPis a mixture of individual species containing variable numbers of endo H-resistant oligosaccharide side chains ranging from zero to five. Thus, the accumulated intracellular and secreted alAGP appear to follow a common route into and through the cis-Golgi cisternae.

AND

EXTERNALIZATION

457

Neuraminidase digestion revealed that (u~-AGP secreted from monensin-treated cells does not contain sialic acid suggesting that the attachment of this sugar residue is unessential for its eventual externalization. This extracellular form is 6000 Da smaller than normal and the failure to attach sialic acid accounts for approximately 3000 Da of this difference in molecular weight; the remaining difference is most likely due to incomplete oligosaccharide processing. Not surprisingly, nongalactosylated intracellular al-AGP also fails to acquire sialic acid since terminal galactosyl residues provide acceptor sites for sialic acid. The lack of sialic acid does not reflect secondary effects on the sialyl transferase since certain viral glycoproteins acquire sialic acid in the presence of monensin (data not shown). Many studies have suggested the existence of vesicle- and organelle-associated receptors that bind to and help selectively sort membrane-associated and secreted polypeptides during transit to their final cellular locations (11,21,42). The receptors may recognize various structural features on transported glycoproteins, such as sites on oligosaccharide side chains analogous to the mannose-&phosphate receptors (2224), or sites on the polypeptide backbone. In fact, for a given glycoprotein, in a given cell type, a hierarchy of sorting signals may function at different cellular locations as these glycoproteins come in contact with distinct receptors in the various vesicle populations along their intracellular route. In this regard, it is interesting to speculate that in its glycosylated state the externalization of al-AGP may be regulated in part by lectin-like receptors that recognize discrete oligosaccharide structures since the primary differences between the accumulated intracellular and externalized species are in the carbohydrate moiety. One possible candidate may be galactosyl residues that are recognized by carrier proteins analogous to the asialoglycoprotein receptors (44, 45) although clearly other oligosaccharide features may be important. Consistent with this view, Aronson, Jr., and co-workers have shown that the rate of (YeAGP secretion from rat hepatocytes can

458

HAFFAR,

EDWARDS,

be correlated with the number of attached oligosaccharide side chains produced in the presence of a threonine analog (14). Also, in a more general sense, certain glycoproteins display an unusually stringent dependence upon normal protein glycosylation for their intracellular sorting and externalization (46, 47). Other investigators have suggested that combinations of carrier-mediated and passive (nonspecific) mechanisms may regulate the flow of proteins through intracellular membranes (42). Passive sorting appears particularly important for bulk movement of proteins trapped inside of secretory vesicles or when appropriate carriers are unavailable such as in the case of saturated or mutated transport pathways (39). Perhaps the small fraction of nongalactosylated (ri-AGP is externalized by a passive mechanism after saturation of the intracellular pathway. Monensin has been shown to selectively raise the pH of the trans-cisternae of the Golgi (48) suggesting that certain receptormediated trafficking events can be disrupted while leaving others intact. In glucocorticoid-induced Ml.54 cells, this phenomenon results in the stable redistribution of q-AGP into secreted and cellular compartments and allowed us to examine Golgi-mediated oligosaccharide processing events that are closely associated with its externalization. Further characterization of al-AGP sorting will help identify those carrier molecules that regulate its posttranslational transport and elucidate the biological significance of this pathway. ACKNOWLEDGMENTS We thank our colleagues in the laboratory for their helpful comments and suggestions during the course of this study. We also thank Judy Cantwell and Aileen Kim for their superb typing and preparation of this manuscript and John Underhill for his excellent photography. REFERENCES 1. K~J, A. (1974) in Structure and Function of Plasma Proteins (Allison, A, C., ed.), pp. 74-131, Plenum, New York. 2. JAMIESON, J. C., ASHTON, F. E., FRIESEN, A. D.,

AND

FIRESTONE

AND CHOU, B. (19’72) Can&. J. Biodwm 50,871880. 3. JAMIESON, J. C., KAPLAN, H. A., WOLOSKI, B. M. R. N. J., HELLMAN, M., AND HAM, K. (1983) Canad J. Biochem 61, 1041-1048. 4. RICCA, G. A., MCLEAN, J. W., AND TAYLOR, J. M. (1982) Ann N. I: Acad Sci 389,88-105. 5. BAUMANN, H., HELD, W. A., AND BERBER, F. G. (1984) J. Biol. Chem. 259,566-573. 6. BAUMANN, H., FIRESTONE, G. L., BURGESS, T. L., GROSS, K. W., YAMAMOTO, K. R., AND HELD, W. A. (1983) J. Biol Chem. 258,563-570. 7. VANNICE, J. L., RINGOLD, G. M., MCLEAN, J. W., AND TAYLOR, J. M. (1983) DNA 2,204-212. 8. GROSS, V., ANDUS, T., TRAN-THI, T.-A., BAUER, J., DECKER, K., AND HEINRICH, P. C. (1984) Exp. Cell Res. 151, 46-54. 9. VANNICE, J. L., TAYLOR, J. M., AND RINGOLD, G. M. (1984) Proc. NatL Acad Sci. USA 81,4241-4245. 10. HUBBARD, C., AND IVATT, R. J. (1981) Annu. Rev. B&hem. 50,555-583. 11. SABATINI, D. D., KREIBICH, G., MORIMOTO, T., AND ADESNIK, M. (1982) J. Cell Biol 92,1-22. 12. RICCA, G. A., AND TAYLOR, J. M. (1981) J. Bi01 Chem. 256, 11,199-11,202. 13. YOSHIMA, H., MATSJMOTO, A., MIZOUCHI, T., KAWASAKI, T., AND KOBATA, A. (1981) J. Bid C&m. 256,8476-8484. 14. DOCHERTY, P. A., AND ARONSON, JR., N. N. (1985) J. BioZ. Chem. 260,10847-10855. 15. SCHMID, K. (1985) in The Plasma Proteins (Putnam, F. W., ed.) Vol. I, pp. 183-228, Academic Press, New York. 16. FOURNET, B., MONTREUIL, J., STECKER, G., DORLAND, L., HAVERKAMP, J., VLIEGENTHART, J. F. G., BINETTE, J. P., AND SCHMID, K. (1978) Biochemistry, 17,5206-5214. 17. JAMIESON, J. C., AND ASHTON, F. E. (1973) Canad J. Biochem. 51,1281-1291. 18. PAULSON, J. C., PRIEELS, J. P., GLASGLOW, L. R., AND HILL, R. L. (1978) J. Biol Chem 253,56175624. 19. NAGASHIMA, M., URBAN, J., AND SCHREIBER, G. (1980) J. Biol Chem 255,4951-4956. 20. NAGASHIMA, M., URBAN, J., AND SCHREIBER, G. (1981) J. Biol Chem 256,2091-2093. 21. BLOBEL, G. (1980) Proc. Nat1 Ad Sci. USA 77, 1496-1500. 22. TABAS, I., AND KORNFELD, S. (1980) J. Biol. Chem. 255,6633-6639. 23. HASILIK, A., AND NEUEFELD, E. F. (1980) J. Biol Chem 255,4946-4950. 24. BROWN, W. J., AND FARQUHAR, M. G. (1984) CeU 36,295-307. 25. PRESSMAN, B. C. (1976) Annu. Rev. Biochem 45, 501-530. 26. TARTAKOFF, A. (1983) CelE 32,1026-1028. 27. RINGOLD, G. M., SHANK, P. R., VARMUS, H. E., RING,

GLYCOPROTEIN

28. 29. 30. 31.

32. 33. 34. 35. 36.

37. 38.

TRAFFICKING

J., AND YAMAMOTO, K. R. (1979) Proc. Nat1 Acad Sci USA 76,665-669. FIRESTONE, G. L., AND YAMAMOTO, K. R. (1983) Mel Cell. BioL 3,149-160. LAEMMLI, U. K. (1970) Nature &wdon) 227,680685. O’FARRELL, P. H. (1975) J. Biol. Chevz. 250,400’74021. BAROUKI, R., FINIDORI, J., CHOBERT, M.-N., AGGERBECK, M., LAPERCHE, Y., AND HANOUNE, J. (1984) J. BioL Chem 259, ‘7970-7974. NICOLSON, G. L., AND BLALISTEIN, J. (1972) B&him. Biophys. Ada 266,543-547. TABAS, I., AND KORNFELD, S. (1978) J. BioL Chem. 253,7779-7786. HARPAZ, N., AND SCHACHTER, H. (1980) J. BioL Chem. 255,4894-4902. TARENTINO, A., PLUMMER, T. H., JR., AND MALEY, F. (1974) J. BioL Chem. 241,818-824. PETERS, B. P., BROOKS, M., HARTLE, R. J., KRZESICKI, R. F., PERINI, F., AND RUDD~N, R. W. (1983) J. BioL Chem 258,14,505-14,515. HATTON, M. W. C., MARZ, L., AND REGOECZI, E. (1983) Trends B&hem. Sci. 8,287-291. COLMAN, A. (1982) Trends B&hem. Sci 7, 435437.

AND

EXTERNALIZATION

459

39. TARTAKOFF, A. M. (1983) B&hem. J 216,1-9. 40. SARASTE, J., AND HEDMAN, K. (1983) EMBO J. 2, 2001-2006. 41. RODRIGUEZ-BOULAN, E., PASKIET, K. T., SALES, P. J. I., AND BARD, E. (1984) J. Cell BioL 98,308318. 42. GETHING, M.-J. (ed.) (1985) Current Communications in Molecular Biology: Protein Transport and Secretion, Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. 43. ROTHMAN, J. E. (1981) Science 213,1212-1219. 44. HUDGIN, R. L., PRICER, W. E., JR., ASHWELL, G., STOCKERT, R.J., AND MORELL, A. G.(1974)J. BioL Chem. 249,5536-5543. 45. GEUZE, H. J., SLOT, J. W., STROUS, G. J. A. M., LODISH, H. F., AND SCHWARTZ, A. L. (1983) Cell 32,277-287. 46. FIRESTONE, G. L. (1983) J BioL Chem 258, 61556161. 47. OLDEN, K., BERNARD, B.A., HUMPHRIES, M. J., YEO, T.-K., YEO, K.-T., WHITE, S. L., NEWTON, S. A., BAUER, H. C., AND PARENT, J. B. (1985) Trends Biochem. Sci. 10,78-82. 48. ANDERSON, R. G. W., AND PATHAK, R. K. (1985) Cell 40,635~643.