Disruptions in Intracellular Membrane Trafficking and Structure Preclude the Glucocorticoid-dependent Maturation of Mouse Mammary Tumor Virus Proteins in Rat Hepatoma Cells*

We have previously shown that glucocorticoids reg- ulate the trafficking and processing of mouse mammary tumor virus (MMTV) proteins in viral-infected M1.64 rat hepatoma cells. To examine the role of intracellular membrane integrity on MMTV protein maturation, brefeldin A (BFA) was utilized to disrupt membrane flow between the endoplasmic reticulum and Golgi. Immunoprecipitation and immunofluores- cence microscopy revealed that in the presence of dexamethasone, BFA inhibited the proteolytic processing, cell surface delivery, and externalization of MMTV glycoproteins. Glycosidase digestion and inhibitors of protein glycosylation confirmed that the observed dif-ferences in apparent sizes of MMTV glycoprotein products are due to BFA-induced changes in oligosaccha- ride processing. BFA treatment inhibited the proteolytic processing of the MMTV phosphoprotein pre- cursor, which normally associates with the cytoplasmic face of intracellular membranes. Similarities in salt extraction efficiency revealed that BFA did not affect the membrane affinity of the uncleaved phosphoryl- ated precursor. In a complementary approach, proteolytic processing of the phosphorylated polyprotein did not occur in glucocorticoid-treated HTC cells transfected with a mutant MMTV provirus encoding a nor- mal phosphorylated precursor, but which express a truncated MMTV glycoprotein missing its transmem- brane domain and cytoplasmic tail. These results suggest that the MMTV glycoproteins and phosphoproteins may interact at a late step in the transport path- way in a manner required for their mutual processing in response to glucocorticoids and establishes the im- portance of functional interactions with intracellular membranes for maturation of the cytoplasmic MMTV phosphoproteins.


Disruptions in Intracellular Membrane Trafficking and Structure Preclude the Glucocorticoid-dependent Maturation of Mouse Mammary Tumor Virus Proteins in Rat Hepatoma Cells*
(Received for publication, July 8,1991) Steven R. KainS, Emily J. PlattB, Karen S . Brown, Nicholas Black, and Gary L. Firestonen From the DeDartment of Molecular and Cell Biology and the Cancer Research Laboratory, University of California, Berkeley, Caiifornia 94720 " We have previously shown that glucocorticoids regulate the trafficking and processing of mouse mammary tumor virus (MMTV) proteins in viral-infected M1.64 rat hepatoma cells. To examine the role of intracellular membrane integrity on MMTV protein maturation, brefeldin A (BFA) was utilized to disrupt membrane flow between the endoplasmic reticulum and Golgi. Immunoprecipitation and immunofluorescence microscopy revealed that in the presence of dexamethasone, BFA inhibited the proteolytic processing, cell surface delivery, and externalization of MMTV glycoproteins. Glycosidase digestion and inhibitors of protein glycosylation confirmed that the observed differences in apparent sizes of MMTV glycoprotein products are due to BFA-induced changes in oligosaccharide processing. BFA treatment inhibited the proteolytic processing of the MMTV phosphoprotein precursor, which normally associates with the cytoplasmic face of intracellular membranes. Similarities in salt extraction efficiency revealed that BFA did not affect the membrane affinity of the uncleaved phosphorylated precursor. In a complementary approach, proteolytic processing of the phosphorylated polyprotein did not occur in glucocorticoid-treated HTC cells transfected with a mutant MMTV provirus encoding a normal phosphorylated precursor, but which express a truncated MMTV glycoprotein missing its transmembrane domain and cytoplasmic tail. These results suggest that the MMTV glycoproteins and phosphoproteins may interact at a late step in the transport pathway in a manner required for their mutual processing in response to glucocorticoids and establishes the importance of functional interactions with intracellular membranes for maturation of the cytoplasmic MMTV phosphoproteins. *This research was supported by National Institutes of Health Grant DK-42799 and, in part, by funds awarded from the Lucille P. Markey Program in Biomolecular Structure and Design. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement'' in accordance with 18 U.S.C. Section 1734 solely to ifidicate this fact.
$ Supported by a Postdoctoral Research Fellowship F32 CA09053 awarded by the National Cancer Institute of the Department of Health and Human Services and supplemental salary support awarded by the American Liver Foundation. § Predoctoral trainee supported by National Research Service Grant CA-09041 awarded by the National Institutes of Health.
ll To whom correspondence and reprint requests should be addressed Dept. of Molecular and Cell Biology, Box 591 LSA, University of California, Berkeley, CA 94720. Tel.: 510-642-8319; Fax: 510-643-6791. Glucocorticoid hormones elicit their biological effects in a variety of target tissues by selectively modulating the transcription of glucocorticoid-responsive genes (1)(2)(3)(4)(5). Conceivably, regulatory factors under the direct transcriptional control of this steroid could modulate the expression, synthesis, transport, or processing of other biologically active proteins. By utilizing the expression of mouse mammary tumor virus (MMTV)' polypeptides in viral-infected M1.54 rat hepatoma cells as traceable molecular probes, we have uncovered two distinct posttranslational regulatory circuits under the control of glucocorticoids (6)(7)(8)(9)(10)(11)(12). The core MMTV phosphoproteins (encoded by the gag gene) and the MMTV envelope glycoproteins (encoded by the enu gene) are each produced as precursor polyproteins that require a series of processing steps to yield stable maturation products. Treatment of M1.54 cells with dexamethasone, a synthetic glucocorticoid, selectively stimulates the glycosylated precursor to enter a sorting route resulting in the generation of three new cell surface-associated glycoproteins (two uncleaved forms of the precursor, gp78 and gp70, and the gp32 carboxyl-terminal fragment), and an extracellular 70-kDa glycoprotein (gp70) (reviewed in Ref. 8). The amino-terminal fragment, gp50, is detected in the presence or absence of dexamethasone. Glucocorticoids also regulate maturation of the MMTV phosphorylated precursor, yielding two new proteolytic products designated p35 and p24 (6,8,9,12). Only steroid hormones with glucocorticoid biological activity elicit these two responses in M1.54 cells; each posttranslational response is abrogated by the glucocorticoid antagonist RU38486 (10,121. Additionally, the glucocorticoidregulated trafficking and processing events require a functional receptor protein (13) and de nouo synthesis of cellular proteins (14). Detailed analysis of oligosaccharide-processing kinetics and subcellular fractionation in M1.54 cells have shown that glucocorticoids exert their regulatory effects on MMTV glycoprotein trafficking within the medial to transcisternae of the Golgi (11,15,16), suggesting that hormonemodulated cellular and/or viral regulatory factors function within or target components to this organelle.
Although the glucocorticoid-dependent maturation of the viral phosphoproteins is generally observed under the same conditions as those required for trafficking of integral membrane-associated MMTV glycoproteins, virtually nothing is known about the relationship between membrane components and the regulated processing of the cytoplasmic-residing MMTV phosphoproteins. As one test of the role of membrane integrity in MMTV protein maturation, we have utilized the Polyprotein Processing ana' Membrane Trafficking fungal metabolite brefeldin A (BFA) to reversibly disrupt protein trafficking past the endoplasmic reticulum (ER) transitional elements in viral-infected cells. BFA has been shown to rapidly induce vesiculation of the Golgi cisternae (17,18), leading to the selective disassembly of this organelle and movement of Golgi membrane components to the ER via a microtubule-dependent retrograde transport pathway (19,20). When BFA is removed from cells, this process is reversed, leading to reformation of the Golgi apparatus and restoration of ER to Golgi protein transport (17,21). This reagent has therefore allowed us to investigate whether a global disruption in membrane trafficking affects the glucocorticoid-dependent maturation of the cytoplasmic MMTV phosphoproteins and the membrane-associated MMTV glycoproteins. In a complementary approach that modifies only the MMTV glycoprotein component of the membrane, HTC cells were transfected with a mutated MMTV provirus that encodes a wild type phosphoprotein precursor but a truncated glycoprotein precursor missing its transmembrane domain and cytoplasmic tail (23). In cells transfected with this mutant provirus, the phosphoprotein precursor, despite having a wild type sequence, is not processed in the presence of glucocorticoids. Our results suggest that global and specific disruptions in intracellular membrane trafficking and structure can preclude the hormonedependent processing of cytoplasmic MMTV phosphoproteins in rat hepatoma cells.

EXPERIMENTAL PROCEDURES
Materials-All media and sera used for tissue culture were purchased from Whittaker M. A. Bioproducts (Walkersville, MD). L-[""SIMethionine (1000 Ci/mmol) and [32P]orthophosphate (8500 Ci/ mmol) were obtained from Du Pont-New England Nuclear. Dexamethasone, swainsonine, and 1-deoxymannojirimycin (DMM) were purchased from Sigma. Endoglycosidase H, N-glycosidase F, and noctyl glucoside were obtained from Boehringer Mannheim. Staphylococcus aureus cells (Pansorbin) were acquired from Calbiochem. The anti-gp52 antiserum used in the immunofluorescence studies was obtained from the National Institutes of Health repository (Bethesda, MD), and the fluorescein isothiocyanate-conjugated F(ab')Z fragment of anti-goat IgG was purchased from Organon Teknika-Cappel (Malvern, PA). Total anti-MMTV (24) and preimmune sera used for immunoprecipitations were generously provided by L. J. T. Young and R. D. Cardiff (Department of Pathology, University of California, Davis). The plasmid pGR16, containing an MMTV provirus with a premature termination codon in the enu gene (22), was provided by H. Ponta (Institute for Genetics and Toxicology, Karlsruhe, Germany). Brefeldin A was a generous gift of D. Roemer (Sandoz LTD, Basel, Switzerland). A stock solution of 1.0 mg/ml BFA in methanol was prepared and stored at -20 "C. All additional reagents were of the highest quality available.
Cells and Method of Culture"M1.54 is a clonal cell line derived from MMTV-infected rat HTC hepatoma cells and contains 10 stably integrated MMTV proviruses (6,25). HGR16.5 cells are single cellderived clones selected from rat HTC hepatoma cells that have been transfected with the plasmid pGR16 encoding a truncated MMTV glycoprotein (23,26). Each cell line was propagated as monolayer cultures at 37 'C on Corning Tissue Culture plates in Dulbecco's modified Eagle's medium supplemented with 10% horse serum in a humidified atmosphere of 95% air, 5% COz. HGR16.5 cultures were additionally supplemented with 1 mg/ml Geneticin (G418; Gibco Laboratories, Santa Clara, CA).
formed essentially as described previously (14,27). Briefly, M1.54 Indirect Immunofluorescence-Immunofluorescence was percells were seeded on glass coverslips and cultured for 16 h in the presence of 1 p M dexamethasone alone or in addition to 2.0 pg/ml BFA. The cells were washed with phosphate-buffered saline (PBS) and fixed by incubation with 3.7% formaldehyde in PBS containing 0.1 M glycine (PBS/glycine). To stain intracellular proteins, cells were permeabilized with 0.5% Triton X-100,300 mM sucrose in PBS for 10 min at room temperature. This step was omitted for cell surface staining. The cells were then washed extensively with PBS/glycine prior to incubation with 1:350 serum dilution of polyclonal anti-gp52 IgG for 30 min at room temperature. After washing in PBS/glycine, the cells were incubated with a 1:400 dilution of fluorescein isothiocyanate-conjugated F(ab'), fragment of anti-goat IgG. The cells were washed a final time in PBS/glycine and mounted on microscope slides in 90% glycerol, 10% 100 mM Tris-HC1, pH 7.5. Slides were stored in darkness at -20 "C until visualized and photographed using a Zeiss fluorescent microscope. Transfection and Expression of a Plasmid Containing a Truncated MMTV Glycoprotein Gene-The plasmid pGR16 was cotransfected with pSV2neo into rat HTC hepatoma cells using the calcium phosphate procedure (28); 48 h after transfection, the cells were replated into selective media containing 1 mg/ml G418. After 3 weeks, individual cell colonies were harvested, expanded in culture, and tested for expression of MMTV RNA by cytoplasmic dot blot analysis (29). Cell clones expressing high amounts of MMTV-specific RNA were further screened for the production of MMTV protein by immunoprecipitation of [35S]methionine-labeled cell extracts with total anti-MMTV sera as described below. One such clone, designated HGR16.5, was used for the experiments described in this study.
Radiolabeling and Collection of Cellular and Secreted Fractions-Monolayer cultures of M1.54 cells were treated with dexamethasone in the presence or absence of BFA, as described in appropriate figure legends. Prior to radiolabeling, the cells were incubated with methionine-free or phosphate-free Dulbecco's modified Eagle's medium in the absence of serum for 30 min, followed by the addition of 50 pCi/ ml [35S]methionine or 200 pCi/ml [32P]orthophosphate in the appropriate medium for an additional 4 h. For cells treated with glycosylation inhibitors, cultures were preincubated with either 1 mM DMM or 20 pg/ml swainsonine in methionine-free Dulbecco's modified Eagle's medium for 1 h at 37 "C prior to the addition of [35S] methionine. The secreted fractions were harvested from radiolabeled cells by centrifugation of the culture medium at 2000 X g for 10 min. The supernatant fractions, containing the secreted MMTV proteins, were brought to 1% Triton X-100 and 5 mM EDTA prior to immunoprecipitation. The cellular fractions were obtained by washing the radiolabeled monolayers three times in ice-cold PBS, harvesting the cells in 10 mM Tris, 1 mM EDTA, in PBS, pH 7.5, and centrifugation at 1000 X g for 10 min. Samples were then solubilized in immunoprecipitation buffer (1% Triton X-100, 5 mM EDTA, 250 mM NaCl, 25 mM Tris-HC1, pH 7.5) and centrifuged in an Eppendorf microcentrifuge at maximum speed for 10 min at 4 "C. The supernatant fraction was used for subsequent analysis. The total level of radiolabeled protein was determined by precipitation onto filter discs using 10% trichloroacetic acid and the values used to normalize samples for immunoprecipitation.
Immurwprecipitation and Electrophoresis of MMTV Proteins-Immunopreciptations of cellular and secreted MMTV proteins using either total anti-MMTV or preimmune sera was performed as described previously (30, 31). The immunoprecipitation of cell surfaceassociated MMTV glycoproteins was accomplished by incubating intact radiolabeled cells with anti-MMTV antiserum by a modification (10,11) of the method described by Krangel et al. (32). Immunoprecipitated proteins were solubilized in electrophoresis sample buffer containing 62.5 mM Tris-HCI, pH 6.8, 2% sodium dodecyl sulfate (SDS), 10% glycerol, and 80 mM dithiothreitol for 5 min at 90 "C. Electrophoretic analysis of immunoprecipitated proteins was performed with samples containing equivalent trichloroacetic acidprecipitable counts. SDS-polyacrylamide gel electrophoresis was performed according to the method of Laemmli (33), using a 10% resolving gel to separate MMTV proteins. After electrophoresis, the gels were prepared for fluorography by soaking in 100 mM salicylic acid for 1 h prior to drying. The dried gels were exposed to Kodak X-Omat AR film (Eastman Kodak) at -80 "C.
Sodium Chloride Extraction of Peripheral Membrane Proteins-Salt extraction of peripheral membrane proteins was performed essentially as described previously (34). Briefly, the cellular fractions from radiolabeled monolayers were obtained as described above, and the cells were resuspended in ice-cold hypotonic buffer containing 20 mM PBS, pH 7.5, 5 mM KCI, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 0.3 p~ aprotinin. The cells were allowed to swell on ice for 30 min and then homogenized using 100 strokes in a tight fitting Dounce homogenizer. The samples were centrifuged in an Eppendorf microcentrifuge for 10 min, and supernatant fractions were adjusted to the appropriate salt concentration using 5 M NaCl in hypotonic buffer. The samples were incubated on ice for 20 min with occasional gentle mixing, followed by centrifugation at 100,000 X g for 60 min at 4 "C. The resultant high speed pellet and supernatant fractions were solubilized, and the samples were immunoprecipitated using total anti-MMTV antibodies as described above. and Membrane Trafficking Digestion with Endoglycosidase H and N-Glycosidase F-For endoglycosidase H (endo H) digestions, the final Pansorbin pellets from immunoprecipitation of cell extracts were resuspended in 45 pl of 100 m M sodium citrate buffer, pH 5.5, containing 1% SDS, boiled for 3 min, and then centrifuged in an Eppendorf microcentrifuge for 10 min. The supernatant fractions were collected and divided into 2O-pl aliquots, and 20 p1 of sodium citrate alone or 20 pl of sodium citrate containing 0.2 milliunits of endo H added to each sample. The reaction mixture was incubated for 18 h at 37 "C, and the reaction was terminated by evaporation in a Savant vacuum centrifuge. N-Glycosidase F (endo F) digestions were performed by the same protocol, except Pansorbin pellets were resuspended in PBS containing 1% SDS, and the reaction mixture consisted of 20 pl of sample supernatant, 15 pl of PBS, 4 pl of n-octyl glucoside, in the presence or absence of 0.1 units of endo F. The samples from each digestion were resuspended in electrophoresis sample buffer and resolved on 10% gels as described above.

RESULTS
BFA Disrupts the Glucocorticoid-dependent Processing, Cell Surface Delivery, and Externalization of MMTV Glycoproteins-Along with the rapid changes in Golgi morphology brought about by BFA, there is a coincidental arrest of protein transport to this organelle (17,18,21,35,36). To verify that BFA appropriately prevents the transport of MMTV glycoproteins in viral-infected M1.54 hepatoma cells, the localization of the cell surface viral glycoproteins in glucocorticoidtreated cells was examined by immunofluorescence microscopy. Monolayer cultures of M1.54 cells were treated with either 1 PM dexamethasone alone or in combination with 2.0 pg/ml BFA for 16 h and incubated with primary total anti-MMTV antiserum and secondary fluorescein isothiocyanateconjugated anti-goat IgG. As shown in Fig  suggesting an association with the Golgi, whereas in BFAtreated cells, the staining is more diffuse and spread throughout the cytoplasm (panel F ) . Such dispersion of intracellular MMTV glycoproteins is consistent with the actions of BFA in other cell types (17,18,21). These results indicate that BFA is acting much like in other systems to prevent delivery of immunoreactive MMTV glycoproteins to the cell surface of M1.54 cells.
To test whether the observed BFA-mediated block in transport of MMTV glycoproteins was accompanied by alterations in proteolytic processing, a time course of BFA effects on MMTV protein maturation was examined by immunoprecipitation of [35S]methionine-labeled cell extracts. As shown in Fig. 2, the addition of BFA to dexamethasone-treated M1.54 cells resulted in both qualitative and quantitative changes in the overall pattern of MMTV maturation products. During shorter incubations with BFA, the 70-kDa MMTV glycoprotein precursor (gp70), and the 50-kDa (gp50) and 32-kDa (gp32) proteolytic products were each shifted to an apparent lower molecular mass following treatment with BFA (referred to as gp70*, gp50*, and gp32*, respectively). Additionally, the sialyated glycosylated polyprotein, gp78; was absent from BFA-treated cells, The change in the molecular mass of these glycoproteins was maintained throughout the BFA time course up to approximately 8 h of treatment and was accompanied by a progressive decrease in the cellular concentration of each protein. The cellular loss of these proteins could not be accounted for by their externalization into the culture medium, as immunoreactive MMTV proteins were completely absent from this fraction under all conditions of BFA treatment up to 16 h (Fig. 2, lanes L and M ) . Similar to the 1-h time point (lune M ) , the culture medium from cells treated with BFA from 2 up to 16 h lacked any immunoprecipitable MMTV glycoproteins (data not shown). After 16 h of exposure to BFA, the processing of gp70* to form gp50* and gp32* was completely inhibited, and the low levels of remaining precursor suggested that gp70* is rapidly turned over. When cells were released from the inhibitory effects of BFA by incubating in the absence of this compound, the hormone-dependent trafficking of MMTV glycoproteins and processing of viral phosphoproteins were restored, along with a complete recovery of total secretory capacity (data not shown). This recovery suggests that the BFA-mediated inhibition of MMTV protein processing resumes as the intracellular membranes reform into their functional structures.
Effects of BFA on MMTV Glycoprotein Oligosaccharide Processing-To test whether BFA disrupts the maturation of oligosaccharide side chains attached to MMTV glycoproteins, radiolabeled extracts from BFA-treated and untreated cells were immunoprecipitated with anti-MMTV antibodies, and the resultant proteins digested with either endo H or endo F. Endo H cleaves N-linked oligosaccharide side chains of the high mannose type characteristic of glycoproteins in transit between the ER and Golgi apparatus (37); glycoproteins acquire resistance to endo H only after reaching the medial Golgi. MMTV glycoproteins expressed from glucocorticoidtreated cells contain a mixture of endo H sensitivities. The carboxyl-terminal fragment, gp32, was completely resistant to endo H, whereas gp50 contains both endo H-resistant and -sensitive sugar side chains (Fig. 3, lanes A and B ) . In the presence of BFA, each of these glycoprotein products is ren- dered completely sensitive to endo H digestion (Fig. 3, lunes C and D). Production of the gp50* and gp32* maturation products suggests that the normally Golgi-associated proteolytic enzymes redistribute into the ER in the presence of BFA. Alternatively, it is formally possible that the proteases newly synthesized in the ER are responsible for the generation of gp50* and gp32* in BFA-treated cells. However, digestion of immunoprecipitated proteins with endo F to remove all the oligosaccharide side chains from MMTV glycoproteins yielded an equivalent array of proteins in the presence or absence of BFA (Fig. 3, lane F versus lane H), demonstrating that BFA does not alter the structure of the polypeptide backbone of the glycoprotein derivatives.
To further confirm that the BFA-mediated alteration in electrophoretic mobilities of MMTV glycoproteins can be accounted for by an alteration in carbohydrate side chain maturation, M1.54 cells were treated with known inhibitors of oligosaccharide processing within the Golgi. Glucocorticoid-treated cells were exposed to either DMM, which inhibits the Golgi mannosidase I and results in the formation of oligosaccharide side chains with a Man9GlcNAc2 high mannose structure (38), or to swainsonine, an inhibitor of Golgi mannosidase I1 (39), which causes the formation of carbohydrate side chains with a GlcNAcMan5GlcNAc2structure. Intracellular and cell surface MMTV proteins were immunoprecipitated from ["Slmethionine-labeled cells and analyzed by electrophoretic fractionation. In the presence of either oligosaccharide processing inhibitor, both of the glycosylated proteolytic maturation products were stably produced. The carboxyl-terminal maturation product, gp32, migrated with the same apparent molecular mass as observed for gp32* in BFAtreated cells (Fig. 4, lunes C-F), which is consistent with BFA preventing the MMTV glycoproteins from being accessible to the Golgi oligosaccharide-processing enzyme. The gp32 produced in DMM-treated cells also migrated as a sharper band in the gel, most likely reflecting a decreased heterogeneity of the carbohydrate side chains. As expected, all of the oligosac-  (lanes A , C, E, and G ) , as described in the text.
Glycosidase-treated and untreated samples were analyzed by SDSpolyacrylamide gel electrophoresis and radiolabeled proteins visualized by fluorography. The molecular mass standards are described in the legend to Fig. 2. The apparent size of gp50 is unaffected by either DMM or swainsonine treatment, which is consistent with our recent observations that the proteolytic cleavage event normally releases a gp50 with unprocessed oligosaccharides? Thus, changes in oligosaccharide structure per se, brought about by swainsonine or DMM, do not impair the ability of MMTV glycoproteins to enter and transit through the exocytic pathway. Taken together, these results demonstrate that BFA perturbs the Golgi membrane structure in such a way as to preclude delivery of the MMTV glycoproteins to intracellular compartments containing the oligosaccharide-processing enzymes required for normal glycoprotein maturation.

Effects of BFA on MMTV Phosphoprotein Maturation-
The BFA-mediated disruption of the Golgi was used to examine the role of membrane trafficking on the glucocorticoiddependent maturation of the cytoplasmic viral phosphoproteins. Immunoprecipitations were carried out with extracts isolated from cells radiolabeled for 4 h with either [35S] methionine or [32P]orthophosphate; the 32P-labeled extracts were utilized to selectively monitor the processing pattern of MMTV phosphoproteins. 32P-Labeled p35 and p24 were observed only in dexamethasone-treated cells (Fig. 5, lane F   versus lane G), whereas the phosphorylated MMTV precursor (Pr74g*K) was stably produced in both dexamethasone-treated and untreated cells. Treatment of M1.54 cells with dexamethasone in the presence of BFA inhibited processing of the MMTV phosphoprotein precursor, resulting in diminished levels of p35 and the total absence of p24 (Fig. 5, lanes G and I ) . As expected, no processing of MMTV precursors was evident in the absence of dexamethasone. These results demonstrate that the BFA disruption of the Golgi causes a concomitant inhibition of the glucocorticoid-dependent maturation of MMTV phosphoproteins.
The Membrane Association of MMTV gag Phosphoproteins Is Not Impaired by BFA-Our previous results suggest that the MMTV phosphoproteins expressed in hepatoma cells reside as cytoplasmic peripheral membrane proteins (8). To L. J. Goodman, S. R. Kain, and G. L. Firestone, manuscript in preparation.
determine whether BFA inhibits MMTV phosphoprotein maturation by disruption of membrane-phosphoprotein interactions, microsomes isolated from radiolabeled cells treated with and without BFA were salt-extracted at differing concentrations of NaCl to remove peripheral membrane proteins. Pretreatment conditions that allow detection of salt-extracted processing products were chosen based on the results shown in Fig. 2, which indicate that 0.5 h of pretreatment with BFA results in altered electrophoretic mobilities of the MMTV glycoproteins and detectable phosphoproteins (lanes G-L).
Sixteen hours of pretreatment largely eliminates all processing, which allowed the selective monitoring of the salt extractability of the phosphoprotein precursor Pr74K"g ( Fig. 2, lanes  M-R). The pellets and supernatant fractions after a 100,000 x g centrifugation were immunoprecipitated with anti-MMTV antibodies. In the absence of NaC1, all of the MMTV proteins expressed in BFA-treated and untreated cells reside predominantly in the membrane pellet (Fig. 6, lane A versus   lane G versus lane M ) . In microsomes isolated from cells not treated with BFA, the addition of NaCl prior to centrifugation resulted in the elution of Pr74gag, as well as the phosphoprotein maturation products, p35 and p24, from the membrane pellet into the soluble fraction (Fig. 6, lane B versus D uersus F). In microsomes isolated from BFA-treated cells, Pr74g*g and p35 can be salt-extracted into the soluble fraction in the absence of detectable p24. The efficiency of salt extraction of individual MMTV proteins was quantitated by densitometry of the autoradiograph displayed in Fig. 6. As shown in Table  I, the MMTV phosphoprotein precursor Pr74gag was saltextracted from microsomes isolated from BFA-treated and untreated cells to a similar extent. Efficient extraction of Pr74gag occurred in both microsome samples at 0.2 M NaC1, which indicated that these viral proteins are weakly associated with membranes. Interestingly, p35 appeared to be extracted somewhat more efficiently in the absence of BFA, suggesting that this MMTV protein is more tightly associated with the   (Table I). As expected for integral membrane proteins, the MMTV glycoproteins remained exclusively associated with the high speed pellet in the presence or in the absence of BFA (Table I) by virtue of their transmembrane domain (23,40,41). These results demonstrate that in the presence of BFA, binding of the MMTV phosphoprotein precursor to the intracellular membranes occurs with approximately the same affinity as in untreated cells, suggesting that a disruption of the inherent ability of this viral phosphoprotein to bind to membranes cannot account for its altered proteolytic processing in the BFA-treated cells.

TABLE I Salt extraction of MMTV proteins from microsomal membranes isolated from BFA-treated and untreated cells Microsomal membranes were isolated from dexamethasone-treated M1.54 cells preincubated with or without BFA for 0.5 h, followed by radiolabeling with [35S]methionine for 4 h in the continued presence or absence of BFA. After incubation with the indicated concentrations of NaCl, salt-extracted (supernatant fraction) and membrane-associated (pellet) proteins were isolated by centrifugation, and the
Proteolytic Processing of Both MMTV Polyproteins Is Absent in Cells Expressing a Truncated Viral Glycoprotein-The BFA-mediated inhibition of MMTV phosphoprotein maturation suggests that overall membrane integrity is a requirement for this regulated processing reaction. To determine whether specific structural domains of the MMTV glycoprotein also need to be accessible for MMTV phosphoprotein maturation, HTC hepatoma cells were transfected with a mutant provirus encoding a wild type phosphoprotein precursor (Pr74g"g) but containing a point mutation in the viral glycoprotein gene that results in a premature termination codon (22). The resulting truncated MMTV glycoprotein (trgp67) is missing its transmembrane domain and cytoplasmic tail and remains completely within the lumen of the microsomes (23). Transfected glucocorticoid-treated and untreated hepatoma cells (designated HGR16.5) were radiolabeled with [35S]methionine and immunoprecipitated with anti-MMTV antibodies. As shown in Fig. 7, in contrast to viral-infected M1.54 cells (lane B), both the viral phosphorylated precursor and the truncated MMTV glycoprotein remained unprocessed in dexamethasone-treated HTC cells transfected with the mutant provirus (Fig. 7, lane D). This result suggests that the transmembrane or cytoplasmic tail of the MMTV glycoprotein may be necessary for maturation of both classes of MMTV proteins. Thus, subtle interactions between the MMTV phosphoproteins and glycoproteins may be important for efficient polyprotein processing in the presence of glucocorticoids.

DISCUSSION
We have previously established that glucocorticoids concurrently regulate the trafficking of MMTV glycoproteins and the proteolytic maturation of MMTV phosphoproteins in viral-infected rat hepatoma cells (6)(7)(8)(9)(10)(11)(12). Salt extractions revealed that the cytoplasmic-residing viral phosphoproteins are loosely associated with the intracellular membranes, suggesting that a specific interaction could occur with the membrane-associated viral glycoproteins that may be required for their mutual transport and/or maturation. Consistent with this notion, our current results have demonstrated that proteolytic processing of both the MMTV glycoproteins and phosphoproteins observed in glucocorticoid-treated hepatoma cells were inhibited after global and selective disruptions in MMTV glycoprotein trafficking. The fungal metabolite BFA was used to inhibit general protein trafficking from the ER to the Golgi complex by inducing the disassembly of the Golgi cisternae (17)(18)(19)(20). In a second approach, transfection of a mutant MMTV provirus, which encodes a truncated MMTV glycoprotein lacking its transmembrane domain and cytoplasmic tail, expresses MMTV glycoproteins and phosphoproteins that reside in different cellular compartments. These complementary strategies have provided the first evidence that proteolytic processing of the MMTV phosphopolyprotein is dependent on the integrity of intracellular membranes and the accessibility of the MMTV glycoprotein. These results further suggest that specific glycoprotein/phosphoprotein interactions may be required for intracellular trafficking and processing of MMTV proteins in viral-infected cells.
Several studies have established that the MMTV glycoproteins and phosphoproteins interact with each other in the virion (42, 43). For example, chemical cross-linking studies have suggested that a heterodimer composed of the aminoterminal gag protein region p10 and gp36 (equivalent to gp32 expressed in hepatoma cells) exists in mature MMTV virions (42). Furthermore, p10 has been shown to have an affinity for the MMTV envelope, most likely by binding to the viral glycoproteins (44,45). This interaction has been suggested to occur between the p10 portion of the uncleaved MMTV phosphopolyprotein and hydrophilic membrane component prior to the endoproteolytic cleavage of the phosphoprotein (45). In general, however, the role of viral glycoprotein-phosphoprotein interactions in exocytic trafficking is poorly understood. It is intriguing to consider that since the relative orientations of the MMTV glycoproteins and phosphoproteins are maintained during their exocytic trafficking and viral assembly, certain contact points between these viral components may be responsible for their mutual processing and/or transport. Consistent with our results, other investigators have postulated a role for the cytoplasmic-associated nonglycosylated viral proteins in promoting the transport of HIV glycoproteins in transfected CHO cells (46). There are also several examples of mutant viral glycoproteins inhibiting virion formation and/or phosphoprotein proteolysis (22,47), suggesting the possibility that specific structural domains within the glycoproteins interact with the phosphoprotein precursors. To determine the precise MMTV glycoproteinphosphoprotein interactions involved in exocytic trafficking and processing, we are currently constructing and expressing a series of mutant MMTV proviral sequences that contain specific mutations in either the viral phosphoprotein or glycoprotein genes.
BFA has been utilized to examine the effects of blocking ER to &-Golgi movement on the posttranslational trafficking of a variety of viral and cellular glycoproteins. For example, BFA prevents the transport and secretion of vesicular stomatitis virus VSV G protein (21), E l and E2 envelope proteins of sindbis virus (48), and the pseudorabies virus glycoproteins (49), as well as inhibiting the release of thyrotropin from pituitary cells (50). BFA has also been shown to prevent the processing of viral envelope glycoproteins in murine erythroleukemia cells (51) and to prevent maturation of glycophorins in murine erythroleukemia cells (18). Aside from the secreted proteins, virtually all the BFA studies have examined integral membrane proteins. The unique feature of our results, therefore, is that the proteolytic processing of a protein (MMTV phosphoprotein) that resides in equilibrium between the cytosol and the cytoplasmic face of the membranes can be concurrently inhibited by a global disruption in membrane structure induced by BFA. Salt extractions further revealed that the peripheral association of MMTV phosphoprotein precursor with isolated microsomes was only minimally affected by BFA treatment, suggesting that more subtle interactions are affected in a way that precludes the proteolytic processing of the phosphorylated MMTV polyprotein.
Exposure to BFA has been shown to dissociate a 110-kDa peripheral membrane protein from the Golgi stacks referred to as /%COP (35, 52). The displacement of this protein from Golgi membrane is very rapid following BFA treatment and precedes all morphological changes in the distribution of Golgi membrane components (53). We hypothesize that the BFAmediated disruption in membrane structure may be masking a critical MMTV glycoprotein-phosphoprotein interaction that occurs relatively late in the exocytic pathway. It is intriguing to consider that the gag-encoded MMTV phosphorylated polyprotein, either alone or in a complex with the viral core, may be transported as a peripheral membrane protein via vesicular transport and proteolytically processed only when in the cellular location that activates or releases the viral protease to yield stable derivatives such as p24. Consistent with this notion, BFA disrupts transport from the ER, whereas the formation of p24-like products in other retroviral systems appears to be a late event in the exocytic pathway leading to or during viral assembly and budding (54, 55). Moreover, other investigators have found that a significant portion of Moloney murine leukemia virus-myristylated gag proteins are transported by vesicles to the plasma membrane in a process that requires specific matrix proteins and the myristic acid modification (56). It is also possible that the retroviral proteases contained within the intracellular viral A particles reside in an inappropriate physical environment that precludes their ability to act on the viral polyprotein substrates. For example, pH and ionic conditions greatly influence viral polyprotein processing (57). Alternatively, it is conceivable that in BFA-treated cells, the MMTV viral core with associated uncleaved phosphopolyproteins reaches the cytoplasmic side of the plasma membrane, whereas the viral glycoproteins remain in the ER. The final viral phosphoprotein maturation events may presumably fail to occur due to a lack of glycoproteins associated with the viral core proteins at the plasma membrane.
Exposure to BFA causes a drastic recompartmentalization of membrane-associated processing enzymes and cellular transport machinery (58). This suggests that the inhibition of MMTV phosphoprotein maturation by BFA may be mechanistically distinct from the lack of processing observed in cells expressing the mutant MMTV glycoprotein precursor. However, regardless of the precise mechanism by which BFA inhibits MMTV phosphopolyprotein maturation, the concurrent disruption of MMTV glycoprotein trafficking and phosphoprotein processing suggests that the glucocorticoid-regulated events affecting each viral component may be functionally interrelated. In this regard, we have recently observed that glucocorticoids exert a significantly reduced effect on MMTV glycoprotein trafficking in rat hepatoma cells transfected with the glycoprotein gene in the absence of expressed viral phosphoproteins (59). We are currently extending our use of BFA as an important reagent to examine the relative contributions of the MMTV phosphoproteins and glycoproteins and membrane remodeling to the glucocorticoid-regulated trafficking and processing reactions in hepatoma cells transfected with both wild type and mutated viral genes.