Antigen in Simian Virus 40-Transformed Mouse Tumor Cells - NCBI

Vol. 63, No. 9

JOURNAL OF VIROLOGY, Sept. 1989, p. 3926-3933

0022-538X/89/093926-08$02.00/0 Copyright © 1989, American Society for Microbiology

Only a Minor Fraction of Plasma Membrane-Associated Large T Antigen in Simian Virus 40-Transformed Mouse Tumor Cells (mKSA) Is Exposed on the Cell Surface ANTON WALSER,' YVONNE RINKE,2 AND WOLFGANG DEPPERT1* Heinrich-Pette-Institut fur Experimentelle Virologie und Immunologie an der Universitat Hamburg, Hamburg,' and Abteilung Biochemie der Universitat Ulm, Ulm,2 Federal Republic of Germany Received 8 March 1989/Accepted 17 May 1989

The bulk of simian virus 40 (SV40) large T antigen in SV40-infected and -transformed cells localizes within the cell nucleus, while a minor fraction specifically associates with the plasma membrane (PM) and is exposed on the cell surface. PM-associated large T seems to span the lipid bilayer but, on the other hand, does not display typical features of a transmembrane protein. To further characterize the postulated transmembrane orientation of large T, we asked whether all large T molecules associated with the plasma membrane indeed are exposed on the cell surface. We compared the amount of cell surface-exposed large T, determined on living cells by a sensitive 3H-protein A-binding assay and by external immunoprecipitation, with that of total PMassociated large T extracted from isolated PM. We demonstrate that in mKSA cells (SV40-transformed BALB/c mouse fibroblasts), total PM-associated large T accounted for a substantial portion (ca. 2%) of total cellular large T. However, only 0.1 to 0.2% of it could be detected on the cell surface. Thus, only a minor fraction of PM-associated large T (less than 10%) is exposed on the surface of these cells. Interior PM-associated large T is stably associated with the plasma membrane, while the small fraction of surface-exposed large T is rapidly released from the cell surface.

The transforming protein of the DNA tumor virus simian virus 40 (SV40), the large tumor antigen (large T), is a multidomained protein exhibiting multiple functions during infection of permissive cells and during transformation of nonpermissive cells (for a review, see reference 31). This multifunctionality may be reflected by the unusual dual subcellular distribution of large T. Driven by a very efficient nuclear transport signal, the bulk of large T localizes within the cell nucleus, while a minor fraction escapes nuclear transport and specifically associates with the plasma membrane (reviewed in reference 4). While there is growing information on the functional roles of nuclear large T (reviewed in reference 29), very little is known about the plasma membrane-associated (PMA) subclass of large T, regarding its biogenesis (16, 27) and its biological function(s) (4). PMA large T can be divided into two subfractions by sequentially extracting plasma membranes isolated from cells of the SV40-transformed BALB/c mouse cell line mKSA: large T recovered from the Nonidet P-40 (NP40)soluble plasma membrane fraction and large T tightly associated with the NP40-insoluble cytoskeletal framework of the plasma membrane, the plasma membrane lamina (PML) (18, 19). The PML, by way of its biogenesis, is part of the cytoskeleton (22), but functionally and structurally is an integral part of the plasma membrane (1, 5). The PML directly underlies the lipid bilayer and is thought to connect the plasma membrane with the cytoskeleton (1, 5). Only PML-associated (PMLA) large T was found to be fatty acid acylated (18), and in mKSA cells grown in suspension culture, it was predominantly this subclass of large T that was exposed on the cell surface (19, 22a). We therefore concluded that PMLA, cell surface-exposed large T exhib*

ited properties characteristic of a transmembrane protein (19). However, the evidence for a transmembrane orientation of large T resulted from two different and not directly comparable observations. Cell surface-exposed large T could be labeled by cell surface iodination, and the labeled molecules then were found to be associated with the PML in purified plasma membranes after cell fractionation (19). PMLA large T also could be demonstrated by cell fractionation of metabolically labeled cells (18, 19). However, it was not clear whether all large T molecules associated with the PML were also exposed on the cell surface, since quantitative analyses of cell surface-exposed and total PMA large T are not available. On the other hand, the design of further experiments characterizing the plasma membrane association of large T critically depend on the understanding of whether PMA large T in its majority indeed is a transmembrane protein or, alternatively, whether it is located on the interior side of the plasma membrane, with cell surfaceexposed PMA large T constituting only a minor fraction of it. Recently, Rinke and Deppert (22a) determined the number of large T molecules actually exposed on the surface of living mKSA cells by using a sensitive 3H-protein A-binding assay. They found that a maximum of about 2,000 large T molecules per cell could be detected on the surface of these cells, nearly all of them associated with the PML. This small number of cell surface-exposed large T molecules suggested to us that possibly only a minor portion of PMA large T is exposed on the cell surface, whereas the majority of it might be an internal membrane protein. This interpretation might explain why it had been so difficult to detect cell surfaceexposed large T by standard immunofluorescence microscopy of untreated, living cells (7) and why, on the other hand, PMA large T could be visualized by immunofluorescence microscopy after treatment of the cells with agents like paraformaldehyde (7) and EDTA (20, 28), i.e., agents

Corresponding author. 3926

VOL. 63, 1989

SURFACE EXPOSURE OF T IN SV40-TRANSFORMED MOUSE TUMOR CELLS

which are known to make interior membrane proteins more accessible to antibody binding by altering the structure of the plasma membrane (10, 14; unpublished data). To further explore this hypothesis, we compared the amount of total PMA large T, extracted from isolated plasma membranes of mKSA cells, with the amount of large T exposed on the surface of these cells, determined by a sensitive 3H-protein A-binding assay (22a) and by external immunoprecipitation (23, 24). We demonstrate that PMA large T accounts for about 2% of total cellular large T in mKSA cells, whereas only a minor fraction (0.1 to 0.2%) is exposed on the surface of these cells. This finding is not restricted to mKSA cells, since large T expressed in the human cell line 293 (13) after infection with the adenovirus vector Ad5-SVR111 (11) exhibited a similar interaction with the plasma membrane. In mKSA cells, interior and cell surface-exposed PMA large T constitute different subclasses, since interior PMA large T is stably associated with the plasma membrane, whereas surface-exposed large T is rapidly released from the cell surface. MATERIALS AND METHODS Cells. mKSA cells (SV40-transformed BALB/c mouse fibroblasts [17]), 293 cells (adenovirus serotype 5-transformed human kidney cells [13]), and Meth A cells (methylcholanthrene-transformed BALB/c mouse fibroblasts [6]) were grown in suspension culture at a density of ca. 5 x 105 cells per ml in Dulbecco modified Eagle medium (DMEM) supplemented with 5% fetal calf serum (Boehringer, Mannheim, Federal Republic of Germany). Infection of cells. 293 cells were infected with the adenovirus vector AdS-SVR111 (11) at a multiplicity of infection of 10 at 5 x 106 cells per ml. After an adsorption period of 30 min, the cells were diluted with medium to a final concentration of 5 x 105/ml. Virus-infected 293 cells were harvested at 24 h postinfection (p.i.). Labeling of cells. A total of 108 mKSA cells were harvested and suspended at a concentration of 5 x 106 cells per ml in methionine-free DMEM supplemented with 5% fetal calf serum and 10 ,uCi of [35S]methionine (1,200 Ci/mmol; New England Nuclear Corp.) per ml. For pulse-chase experiments, the labeling medium was removed after 1 h, and the cells were split into two parts and washed several times with medium. One part (pulse) was subjected to cell fractionation immediately, while the remaining cells (chase) were diluted with growth medium (chase period, 2 h). 3H-protein A assay. A detailed description of the 3Hprotein A assay has been given elsewhere (22a). Briefly, 2 x 107 to 3 x 107 living mKSA or Ad5-SVR111-infected 293 cells were harvested and washed with phosphate-buffered saline (PBS). In a first step, the cells were incubated with an anti-T monoclonal antibody (PAb 108 [15] or KT3 [21]; 10 ,ul of ascites fluid). After unbound antibodies were removed, the cells were incubated with secondary antibody (rabbit anti-mouse immunoglobulins, RAM 7S [5 ,ug]). Then 3Hprotein A was added (0.2 R,g, corresponding to 2.4 x 105 cpm). All incubation steps were performed in PBS for 30 min on ice. Finally the cells were extensively washed and sequentially lysed with NP40 and Empigen BB. The 3H-protein A content in the NP40 and Empigen BB fractions was measured in a liquid scintillation counter after protein precipitation with 10% trichloroacetic acid. External immunoprecipitation. External immunoprecipitation was performed by the method of Santos and Butel (23, 24). Living mKSA or AdS-SVR111-infected 293 cells were

3927

incubated with PAb 108 and then with secondary antibody (RAM 7S), extensively washed, and lysed with NP40 and Empigen BB. Cell surface-exposed large T bound to antibody was precipitated with protein A-Sepharose and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Preparation of plasma membranes and cell fractionation. A detailed description of the fractionation procedure has been given elsewhere (8, 18). Briefly, the cells were washed once with 10 ml of KM buffer {10 mM MES [2-(N-morpholino)ethane sulfonic acid], 10 mM NaCl, 1.5 mM MgCl2, 5 mM dithiothreitol [DTT], 1 mM EGTA [ethylene glycol-bis(paminoethyl ether)-N,N,N',N'-tetraacetic acid], 30 p.g [200 KIU] of aprotinin [Trasylol, Bayer] per ml, pH 6.2} and suspended in 2.5 ml of swelling buffer (10 mM MES, 5 mM MgCl2, 5 mM DTT, 1 mM EGTA, 30 jig of aprotinin per ml, pH 6.5). After being swelled for about 10 min, the cells were gently broken by Dounce homogenization. The homogenate was centrifuged for 15 min at 800 x g in a CRU-5000 centrifuge (International Equipment Company, Needham Heights, Mass.) (4°C). The supernatant of this low-speed spin was cleared by a 16,000 x g spin (Sorvall, SS34, 30 min, 4°C) to remove contaminating nuclei or small plasma membrane fragments. The cleared cytoplasm then was centrifuged at 130,000 x g for 30 min and thereby subfractionated into a soluble (CJ) and a particulate (Cp) cytoplasmic fraction. The Cp fraction was solubilized with 1 ml of lysis buffer (0.5% NP40, 120 mM NaCl, 10% glycerol, 5 mM DTT, 1 mM EGTA, 30 p.g of aprotinin per ml, 50 mM Tris hydrochloride, pH 8.0). The pellet of the low-speed centrifugation containing nuclei and plasma membranes was suspended in an aqueous two-phase polymer system containing polyethylene glycol 6000 (Carbowax) and Dextran T 500 (Pharmacia). Plasma membranes accumulated in the interphase of the two-phase system, while nuclei were found in the pellet after centrifugation at 5,000 x g (10 min, 4°C). Plasma membranes were purified three times by centrifugation in the two-phase system. Harvested plasma membranes, washed once with distilled water, were suspended in 1 ml of PBS (140 mM NaCl, 3 mM KCl, 8 mM Na2HPO4 2H20, 1.5 mM KH2PO4, pH 7.4) containing 5 mM DTT, 1 mM EGTA, 30 ,ug of aprotinin per ml, and 1% NP40 and kept for 30 min on ice. The NP40-treated plasma membranes then were subjected to centrifugation (12,000 x g, 4°C) for 30 min. The supernatant represented the NP40-soluble plasma membrane fraction (NP40-PM). The NP40-insoluble PML was suspended by sonication in 1 ml of TK buffer (25 mM KCl, 5 mM MgCl2, 5 mM DTT, 1 mM EGTA, 30 ,ug of aprotinin per ml, 50 mM Tris hydrochloride, pH 9.0) containing 1% Empigen BB (Albright and Wilson, Ltd.) and solubilized for 1 h on ice. Nuclei from the pellet fraction of the two-phase system were extracted with 1 ml of lysis buffer for 30 min on ice (N1), and NP40-insoluble nuclear structures were removed by centrifugation (12,000 x g, 30 min, 4°C). The residual nuclear structures were solubilized with 1 ml of TK buffer containing 1% Empigen BB for 1 h on ice (N2). Immediately after fractionation, all extracts were adjusted to 150 mM NaCl and pH 9.0 and phenylmethylsulfonyl fluoride was added to a concentration of 1 mM. To the N2 and PML extracts, NP40 was also added (final concentra-

tion, 2%). Immunoprecipitation and SDS-PAGE. The cell extracts were cleared in a Beckman ultracentrifuge (30 min, 130,000 x g, 4°C), and large T was immunoprecipitated by using 10

pJ of the monoclonal antibody PAb 108 ascites fluid (15) and

3928

J. VIROL.

WALSER ET AL.

200 p.l of settled protein A-Sepharose preswollen in PBS. Immune complexes bound to protein A-Sepharose were removed and processed for SDS-PAGE. SDS-PAGE and Western immunoblotting were performed as described previously (8, 18).

TABLE 1. Comparison of nuclear, cell surface-exposed, and total PMLA large T in mKSA and AdS-SVR111infected 293 cells at 24 h pi."

Cells

RESULTS Use of mKSA suspension cultures for analyzing PMA large T. For the quantitative analysis of cell surface-exposed and total PMA large T molecules, we chose the well-characterized SV40-transformed cell line mKSA (BALB/c mouse fibroblasts [17]) grown in suspension culture. The use of mKSA cells grown in suspension culture offers several advantages for analyzing PMA large T. Only cells able to grow in suspension culture can be used for preparation of isolated plasma membranes without any pretreatment, such as trypsination or scraping them off the dish. Furthermore, as shown previously by cell surface iodination (19) and by a 3H-protein A-binding assay (22a), the majority of cell surface-exposed large T in these cells was associated with the PML, whereas in mKSA cells grown on culture dishes a considerable amount of NP40-soluble large T was also detected on the cell surface (19). Therefore, by using mKSA cells grown in suspension, only the PMLA subclass of large T had to be considered for the comparison of cell surfaceexposed versus total PMA large T. The focus on PMLA large T had the additional advantage that its quantitation is much more accurate than the quantitation of NP40-soluble PMA large T. The PML fraction is preparatively defined as a residual structure of the plasma membrane remaining after extraction of isolated plasma membranes with nonionic detergents (1, 5, 22). PMLA large T thus is not contaminated with large T originating from other NP40-soluble subcellular compartments, e.g., internal membranes (see below). Furthermore, mKSA cells also grow as suspension cells in a mouse as an ascites tumor. Analysis of such ascites tumor cells previously had shown that in these cells also, predominantly PMLA large T was exposed on the cell surface (19). Thus, analysis of mKSA cells grown in suspension culture provides a suitable model system for analyzing the properties of PMA large T in a tumor cell. Last, this cell line was used for many analyses characterizing PMA large T (24-26, 30), thereby allowing a comparison of the data provided here with other studies. Quantitation of cell surface-exposed large T. (i) Cell surface protein A-binding assay. For the quantitation of cell surfaceexposed large T, Rinke and Deppert (22a) recently developed a sensitive 3H-protein A-binding assay, in which living cells were analyzed for cell surface-exposed large T. Briefly, the cells are sequentially incubated with large T-specific monoclonal antibodies, secondary antibody (rabbit antimouse immunoglobulins), and 3H-protein A. After extensive washes, the cells are sequentially extracted with NP40, followed by extraction with the zwitterionic detergent Empigen BB, to release the NP40-soluble and the NP40-insoluble subfractions of large T. Since the specific activity of 3H-protein A is known, the amount of cell surface-exposed large T in each fraction then can be estimated by quantitating the amount of 3H-protein A. Large T is inserted into the plasma membrane in a specific manner, with the carboxy and the amino termini exposed on the cell surface and the midportion buried in the lipid bilayer (9, 12, 33). Thus, we first used the anti-T monoclonal antibody PAb 108, directed against the amino terminus of large T (15). mKSA cells grown in suspension culture were

mKSA Ad5-SVR111-infected 293

PMLA large T (ng/107 cells) External Total 3H-protein immuno- plasma (ngI107 cells) A assay precipi- membrane tation preparation

Nuclear large T

2,400 22,000

1.5-3 2-4

4-6 3-4

50 50

" The amount of cell surface-exposed PMLA large T was determined by either a 'H-protein A-binding assay or by external immunoprecipitation. Total PMLA large T was extracted from purified plasma membranes. The results presented were obtained with the monoclonal antibody PAb 108, directed against the amino terminus of large T.

sequentially incubated for 30 min on ice with PAb 108, with secondary antibody, and then with 3H-protein A. After each incubation step, the cells were washed extensively. The cells were then extracted first with NP40, followed by extraction with Empigen BB. Taking into account a 1:1 stoichiometry for binding of secondary to primary antibody and binding of 2 molecules of protein A to only the secondary antibody (for details, see Rinke and Deppert [22a]), we estimate that PAb 108 recognized approximately 1.5 to 3 ng of cell surfaceexposed PMLA large T per 107 cells (Table 1). Only a few labeled large T molecules were found in the NP40 extract, confirming previous results (19, 22a) that the NP40-soluble plasma membrane fraction of these cells did not contain a significant amount of cell surface-exposed large T. To exclude the possibility that amino-terminal determinants of large T recognized by PAb 108 are masked on the cell surface, we quantitated the number of cell surface-exposed large T in mKSA cells by using an antibody directed against the carboxy terminus of large T. We chose the monoclonal antibody KT3, directed against a synthetic peptide corresponding to the carboxy-terminal 11 amino acids of SV40 large T antigen (21). KT3, however, recognized a maximum of only about 1 ng of PMLA large T molecules on the surface of 107 mKSA cells, probably reflecting that more aminoterminal than carboxy-terminal determinants of surface T are accessible to antibody binding (for further discussion, see Rinke and Deppert [22a]). Again, NP40-soluble cell surface-exposed large T was barely detectable. (ii) External immunoprecipitation of cell surface-exposed large T. To confirm the results obtained with the 3H-protein A-binding assay, we quantitated cell surface-exposed large T by external immunoprecipitation (23, 24). Living mKSA cells were incubated with PAb 108 and then with secondary antibody for 30 min on ice, washed extensively, and then extracted sequentially with NP40 and Empigen BB. Antibodies bound to cell surface-exposed large T were precipitated with protein A-Sepharose and analyzed for large T by SDS-PAGE. The small amount of large T precipitated by external immunoprecipitation could not be detected by Coomassie staining of the gel (Fig. 1A). This large T, however, could be visualized after Western blotting (Fig. 1B) and was quantitated by comparing it with various dilutions of a large T preparation of known protein content. Evaluation of these Western blots resulted in a similar amount of PMLA large T molecules exposed on the surface of 107 mKSA cells as were detected by the 3H-protein A-binding assay (ca. 4 to 6 ng; Table 1). External immunoprecipitation followed by Western blotting also revealed the

VOL. 63, 1989

*-S.w

SURFACE EXPOSURE OF T IN SV40-TRANSFORMED MOUSE TUMOR CELLS A

B T-

"Im

-T

3929

:. w.

lgG:

IgG-

Cs cp N1 a b a b FIG. 1. External immunoprecipitation of cell surface-exposed large T. mKSA cells were subjected to external immunoprecipitation as described in Materials and Methods. After lysis of the antibody-treated cells with NP40 (lanes a). followed by extraction with Empigen BB (lanes b), the extracts were analyzed for antibodylarge T complexes by immunoprecipitation with protein A-Sepharose. The immunoprecipitates were subjected to SDS-PAGE and analyzed by Coomassie blue staining (A) or Western blotting (B).

small amount of NP40-soluble cell surface-exposed large T (less than 1 ng/107 cells; Fig. 1B, lane a). Quantitation of total PMLA large T. (i) Preparation and characterization of purified plasma membranes. To preparatively purify plasma membranes from mKSA cells, we used a cell fractionation scheme which had been developed and characterized in our laboratory (8, 18). Briefly, after hypotonic swelling, the cells are disrupted by gentle Dounce homogenization. Nuclei and plasma membrane fractions are separated from the cytoplasm by low-speed centrifugation. The supernatant cytoplasmic fraction then is further subfractionated by high-speed centrifugation into a soluble (Cs) and a particulate (Cp) fraction, the latter containing internal membranes of the endoplasmic reticulum and the Golgi apparatus. Plasma membranes and nuclei are separated by using an aqueous two-phase polymer system. Purified plasma membranes are first treated with NP40 to extract NP40-soluble constituents of the plasma membrane (NP40PM). The NP40-resistant cytoskeletal framework of the plasma membrane, the PML, is then solubilized with Empigen BB. Nuclei are first extracted with NP40 lysis buffer to yield an NP40-soluble nuclear extract (N1), followed by solubilization of the residual NP40-insoluble nuclear structures with Empigen BB (N). To evaluate the approximate yield of purified plasma membranes, the cells were subfractionated and the amount of isolated plasma membranes was determined as a measure of protein content. Purified plasma membranes from mKSA cells contained about 6% of total cellular protein. Since it has been reported that plasma membranes from mouse L-cells contain 4 to 6% of total cellular protein (3, 32), this value seems to indicate an almost quantitative recovery of cellular plasma membranes. However, one has to consider that our plasma membrane fraction was still contaminated with intracellular membranes, as had been previously demonstrated by electron microscopic analysis of plasma membranes isolated from mKSA cells (18). Thus, the amount of purified

N2NP40PML

FIG. 2. Quantitative evaluation of large T in subcellular fractions of mKSA cells. mKSA cells were subfractionated as described in Materials and Methods, and large T was immunoprecipitated from the soluble cytoplasmic fraction (C), the NP40 extract of the particulate cytoplasmic fraction (Cp), the NP40-soluble nuclear fraction (N1), the Empigen BB extract of the NP40-insoluble nuclear fraction (N.), the NP40-soluble plasma membrane fraction (NP40PM), and the Empigen BB extract of the NP40-insoluble PML fraction (PML). After SDS-PAGE and Coomassie blue staining of the proteins, the amount of large T in each subcellular fraction was determined by comparison with BSA standards of known protein content.

plasma membranes is probably lower. Any contamination of our plasma membrane preparation with internal membranes, however, does not affect the quantitation of PMLA large T, since microsomal vesicles are NP40 soluble. Thus, large T associated with these membranes will not be found in the NP40-insoluble PML fraction. (ii) Quantitation of PMLA large T. For the quantitative analysis of PMLA large T from isolated plasma membranes, mKSA cells were fractionated as described above. Large T from each fraction was immunoprecipitated with PAb 108 and protein A-Sepharose and analyzed by SDS-PAGE. Bovine serum albumin (BSA) standards of known protein content were run on parallel gels. After SDS-PAGE and Coomassie blue staining of the proteins, the amount of large T in each fraction was determined by comparison with BSA standards. Figure 2 shows the steady-state levels of large T in the various subcellular fractions of mKSA cells. The bulk of large T was located within the two subfractions of the cell nucleus (N1 and N,), while minor amounts were found in the cytoplasmic fractions (Cs and C, fractions) and in the plasma membrane fractions (NP40-PM and PML fractions). Large T extracted from purified plasma membranes of mKSA cells could be clearly demonstrated on the Coomassie-stained gel (Fig. 2). This is in contrast to cell surface-exposed large T, detected by external immunoprecipitation, which could not be visualized by Coomassie staining (cf. Fig. 1A). Thus, the amount of total PMA large T in mKSA cells seems to be significantly higher than that of cell surface-exposed large T. Evaluation of the amount of PMLA large T in mKSA cells then demonstrated that this fraction contained about 50 ng of large T per 107 cells. In contrast, depending on the assay, only between 1.5 and 6 ng of large T could be detected on the cell surface of 107 mKSA cells (Table 1). Thus, at most only

WALSER ET AL.

3930

J. VIROL.

TABLE 2. Analysis of plasma membranes isolated from mKSA cells for nuclear contamination" Subcellular fraction

C

N PM

[3H]thymidine (cpm) 516

181,285 1,056

c

A

% of total

[3Hlthymidine 0.3 99.1 0.6

"To test plasma membranes of mKSA cells for nuclear contamination. the cells were labeled with [3H]thymidine for 3 h. After a chase period of 12 h, the cells were fractionated into a cytoplasmic (C), a nuclear (N), and a plasma membrane (PM) fraction as described in Materials and Methods. Each fraction was lysed in sample buffer (3% SDS, 1% 2-mercaptoethanol, 62.5 mM Tris hydrochloride, pH 6.8), and samples were counted by liquid scintillation analysis.

10% of total PMLA large T was detectable by antibody binding on the cell surface. Total PMLA large T (50 ng/107 cells) accounted for about 2% of total cellular large T (2.5 [Lg/107 cells) in mKSA cells and thus constitutes a considerable fraction of large T in these cells. Control experiments. The bulk of large T is located within the cell nucleus. To substantiate our finding that cell surfaceexposed large T is only a minor fraction of total PMA large T, it was vital to demonstrate that our isolated plasma membranes were free of contaminating nuclei. This was achieved by labeling the cells with [3H]thymidine prior to cell fractionation and analyzing the [3H]thymidine content of the individual subcellular fractions as a measure of DNA content and thus for nuclear contamination. mKSA cells were pulse-labeled with [3H]thymidine for 3 h, followed by a chase in normal growth medium for 12 h before cell fractionation. The cells then were fractionated as described above. Isolated plasma membranes of mKSA cells were almost free of nuclear contamination (Table 2). In addition, prior to preparation of PMLA large T, the membranes first were extracted with NP40, removing most of the large T present in the small amount of nuclei maximally contaminating the plasma membrane fraction. Thus, large T recovered from the NP40-insoluble PMLA fraction represented almost pure PMLA large T. Another possibility for artificially localizing large T in the PMLA fraction might also be envisioned. After fractionation of mKSA cells, large T was also found in the Cs and Cp fractions (Fig. 2). Thus, some cytoplasmic large T might have associated with the plasma membrane fractions in vitro, i.e., during cell fractionation. To demonstrate that large T from the cytoplasmic fraction does not bind to the PML of mKSA cells during cell fractionation, we drastically increased the amount of solubilized large T by extensive Dounce homogenization of the cells, in order to break up nuclei and induce the leakage of nuclear large T into the cytoplasmic fraction. The result of this experiment is shown in Fig. 3. Although a significant portion of nuclear large T had been released into the soluble cytoplasmic fraction (Cs fraction) during vigorous Dounce homogenization (Fig. 3B), no increase of PMLA large T was detectable, demonstrating that large T did not associate with the PML in vitro (cf. Fig. 3A and B). To test whether isolated plasma membranes of transformed BALB/c mouse fibroblasts have any capacity to bind large T in vitro, we analyzed whether plasma membranes from cells not containing SV40 large T would bind large T in a reconstitution experiment. For this experiment, purified plasma membranes from Meth A cells (methylcholanthrenetransformed BALB/c mouse fibroblasts [6]) were used, because these cells are similar to mKSA cells in many biolog-

N

! i-;N

4'N

Dt

,

.-.

.

I%P

-l

PW 1

FIG. 3. Analysis of in vitro binding of large T to plasma membranes during cell fractionation. (A and B) mKSA cells labeled with [35S]methionine for 1 h were Dounce-homogenized either gently (A) or vigorously (B) before cell fractionation. Large T was then immunoprecipitated from each subcellular fraction and analyzed by SDS-PAGE, followed by Western blotting and fluorography as described in Materials and Methods. Designation of individual fractions is as described in the legend to Fig. 2. (C) Purified plasma membranes from Meth A cells were incubated with the Cs extract of

mKSA cells gained by vigorous Dounce homogenization (panel B). After an incubation period of 30 min, the plasma membranes were repurified and subfractionated into an NP40-soluble plasma membrane fraction (NP40-PM) and an NP40-insoluble plasma membrane fraction (PML). Fractions were analyzed for large T by immunoprecipitation, followed by SDS-PAGE and Western blotting.

ical properties (unpublished observations). Purified plasma membranes from Meth A cells were incubated for 30 min in a soluble cytoplasmic extract obtained by vigorous Dounce homogenization (see above). The membranes then were repurified by our standard fractionation protocol, sequentially extracted with NP40 and Empigen BB, and analyzed for associated large T. Large T present in the soluble cytoplasmic fraction did not bind to purified plasma membranes of Meth A cells (Fig. 3C). Thus, an in vitro association of large T with plasma membranes during cell fractionation is extremely unlikely. An interesting observation was made by evaluating the Western blots in Fig. 3A and B. The amount of large T in the CS fractions varied according to the different fractionation conditions applied in the reconstitution experiments. The amount of large T in the Cp fraction, however, remained constant. The fact that the amount of Cp-associated large T was independent of the amount of solubilized large T might indicate that the observed Cp localization of large T represents a genuine association of large T with internal membranes in vivo as well. Comparison of cell surface-exposed and total PMLA large T in Ad5-SVR111-infected 293 cells. Our results so far demonstrated that in mKSA cells only a minor fraction of total PMA large T is exposed on the cell surface. To analyze whether this unexpected finding is restricted to this particular cell line, we wanted to analyze the plasma membrane association of large T in other cells. Such an analysis, however, is restricted by technical reasons, since only cells able to grow in suspension culture are suitable for these assays (see above), and the number of SV40-transformed cell lines that grow in suspension is very limited. We therefore chose to analyze the plasma membrane association and cell surface expression of large T in the adenovirustransformed human cell line 293 (13), infected with the adenovirus vector Ad5-SVR111 (11). In Ad5-SVR111, the early region of SV40 is placed under the control of an


VOL. 63, 1989

3H-cpm x 103

A

3931

B p c

C

ii# 0,

= -

0 '0 0

NP40 Empigen

-

-

T

..P,A

p53 -

x

v

0,6

._

§ 0,4 0

Ss

I 0,2

.8 0,0 0,0

1,5 2,0 1,0 2,5 reincubaton FIG. 4. Release of cell surface-exposed large T. mKSA cells were treated with Fab fragments of PAb 108. The cells were then washed with PBS and reincubated for the indicated times (in hours) in normal growth medium at 37°C. After various times, portions were harvested, incubated sequentially with rabbit anti-mouse immunoglobulins (RAM 7S; 5 ,ug) and with 3H-protein A, and subjected to sequential extraction with NP40 and Empigen BB. 0,5

adenovirus major late promoter after deleting early regions 1 and 3 (El and E3) of the adenovirus genome. Deletion of most of El renders the virus defective for growth in normal cells and requires that the vector be grown on a permissive human cell line (e.g., 293 [13]) that constitutively expresses El and thereby complements the deletion (11). In AdSSVR111-infected 293 cells, large T is overproduced during the late phase of infection. AdS-SVR111-infected 293 cells fulfill the technical requirements, since 293 cells grow in suspension culture and can therefore be used in our assays. We analyzed Ad5-SVR111-infected 293 cells during the late phase of infection (24 h p.i.). The number of cell surface-exposed large T molecules was determined by our 3H-protein A assay and by external immunoprecipitation with PAb 108. Approximately the same number of PMLA large T molecules were detectable on the surface of these cells as on mKSA cells (2 to 4 ng/107 cells; Table 1). Also in this system, virtually no NP40-soluble cell surface-exposed large T was detectable, further arguing for a typical plasma membrane association of large T in Ad5-SVR111-infected 293 cells. To determine the amount of total PMLA large T from purified plasma membranes of Ad5-SVR111-infected 293 cells, plasma membrane fractions were prepared and characterized as described for mKSA cells (see above). The amount of total PMLA large T extracted from isolated plasma membranes from AdS-SVR111-infected 293 cells was about 50 ng/107 cells. In this system also, only a minor fraction of total PMLA large T was exposed on the cell surface. It is noteworthy that AdS-SVR111-infected 293 cells contained a similar amount of PMLA large T as mKSA cells, although the amount of nuclear large T was increased about 10-fold (Table 1). This may indicate that the amount of PMA large T is independent of the amount of nuclear large T. Biological properties of PMLA and cell surface-exposed large T in mKSA cells. Santos and Butel (24) previously demonstrated that cell surface-exposed large T was rapidly

CCN1

92 N P4i P ML

-

NL p P41 2pM

1

L

FIG. 5. Analysis of the metabolic stability of PMA large T from mKSA cells after pulse-chase labeling. (A) mKSA cells were pulse labeled (lane p) for 1 h or pulse-chase labeled (lane c) (1-h pulse, 2-h chase) with [355]methionine (see Materials and Methods). The cells were then subfractionated, and large T was immunoprecipitated from each subcellular fraction and analyzed by SDS-PAGE and fluorography. Designation of individual fractions is as described in the legend to Fig. 2. (B) The pulse and the pulse-chase medium was analyzed for large T released into the medium.

released from the cell surface. These data seemed to be in contrast to experiments by Klockmann and Deppert (19), who showed by pulse-chase labeling that PMA large T was metabolically stable. The seeming discrepancy between these two results might be explained by considering that Santos and Butel analyzed only the minor fraction of large T exposed on the cell surface, while Klockmann and Deppert analyzed total PMA large T. To test this assumption, we compared the dynamics of the cell surface association of large T by using our 3H-protein A-binding assay, with the metabolic stability of total PMA large T in mKSA cells grown in suspension culture. To analyze the release of cell surface-exposed large T, mKSA cells were treated with antibody (Fab fragments of PAb 108), transferred into normal growth medium, and then labeled with secondary antibody and 3H-protein A after various times. The cells then were sequentially extracted with NP40 and Empigen BB and analyzed for the amount of radiolabeled 3H-protein A. Cell surface-exposed PMLA large T was rapidly released from the cell surface (Fig. 4). Incubation of the antibody-treated cells for 30 min in normal growth medium revealed that only 10 to 20% of the large T molecules were still detectable on the cell surface by the 3H-protein A assay. For analysis of the metabolic stability of total PMLA large T, mKSA cells were pulse labeled for 1 h (pulse, p) or pulse-chase labeled (1-h pulse, 2-h chase; c) with [35S] methionine and then subfractionated. Large T was immunoprecipitated from subcellular fractions and analyzed by SDS-PAGE. Large T was also precipitated from the pulse and the pulse-chase medium to analyze the amount of large T possibly released into the medium. Figure 5 shows the result of this experiment. After pulse labeling, the bulk of [35S]methionine-labeled large T was found in the nucleus; minor amounts were present in the C, and Cp fractions. NP40-soluble PMA large T and PMLA large T were also present after the pulse label. After a 2-h chase period, similar amounts of radiolabeled large T were found in the C ,, the nuclear, and the plasma membrane fractions as after the pulse label (Fig. SA). PMA large T in mKSA cells thus is metabolically stable. This result was further supported by

3932

WALSER ET AL.

analyzing large T precipitated from the pulse and pulsechase media (Fig. 5B). Only barely detectable amounts of large T were found in both media, further supporting the notion that the bulk of PMLA large T had not been released from the cell surface. DISCUSSION In this study we provide evidence that only a minor fraction (less than 10%) of large T associated with the plasma membrane of cells of the SV40-transformed BALB/c mouse tumor line mKSA is exposed on the cell surface. This conclusion was reached by comparing the amount of cell surface-exposed large T determined by a sensitive 3Hprotein A-binding assay (22a) as well as by external immunoprecipitation (23, 24) with the amount of total PMA large T extracted from isolated plasma membranes. The number of cell surface-exposed large T molecules was determined by using monoclonal antibodies directed against the amino terminus (PAb 108) or the carboxy terminus (KT3) of large T. We restricted our analysis to PMA large T associated with a substructure of the plasma membrane, the PML, since only PMLA large T can be clearly defined by its distinct association with this plasma membrane substructure. PMLA large T thus can be prepared largely without being contaminated with large T from other cellular compartments. Furthermore, only PMLA large T had to be considered for comparison with cell surface-exposed large T, since in mKSA cells grown in suspension culture, only this subclass of large T is exposed on the cell surface (19). Since the PML structurally is part of the cytoskeleton and directly underlies the lipid bilayer (1, 5), the simplest way of interpreting our results is that at least the bulk of PMLA large T is located at the cytoplasmic face of the plasma membrane. Plasma membrane preparations of mKSA cells also contain an NP40-soluble fraction of PMA large T (19; this study). This fraction contained an approximately equal amount of large T as the PMLA fraction (see Fig. 2, 3A, and 3B). However, the NP40-soluble subclass of PMA large T has so far only been preparatively defined, and we do not know to what extent the NP40-soluble PMA large T fraction represents large T genuinely associated with the plasma membrane or, at least in part, contains large T associating with interior membranes of the endoplasmic reticulum and the Golgi complex contaminating our plasma membrane preparation (18). With regard to our analysis, however, the portion of surface-exposed PMA large T would even be smaller by also considering NP40-soluble PMA large T as genuine PMA large T. To analyze whether the results obtained reflect a peculiarity of mKSA cells, we compared the amount of total PMA large T with cell surface-exposed large T in the adenovirustransformed human cell line 293 (13) infected with the adenovirus vector Ad5-SVR111 (11) containing the early region of SV40 and producing large T in large amounts during the late phase of infection. We chose this system because, for technical reasons, our analyses were restricted to the use of suspension cells. Furthermore, Ad5-SVR111infected 293 cells should reflect typical intrinsic properties of PMA large T, since large T in adenovirus-infected cells does not exhibit any biological functions. Therefore, the analysis of the plasma membrane association and of the cell surface expression in this system should compensate for the analysis of many different SV40-transformed cell lines. In AdSSVR111-infected 293 cells, the amount of PMA large T was similar to that in mKSA cells, although they contained a ca.

J. VIROL.

10-fold-higher amount of nuclear large T. However, in these cells the bulk of PMA large T also could not be detected on the cell surface. The fact that the actual amounts as well as the ratio of interior to cell surface-exposed PMA large T in these two different systems were comparable may allow the generalization that the bulk of PMA large T in SV40-infected and -transformed cells is not exposed on the cell surface. In mKSA cells, we found that only the minor fraction of surface-exposed large T was released from the cell surface, whereas interior PMA large T remained stably associated with the plasma membrane. The different stability of cell surface-exposed and of interior PMA large T supports our finding that only a minor fraction of total PMA large T is exposed on the cell surface and suggests the existence of at least two subclasses of PMA large T. (i) PMLA, cell surfaceexposed large T, which seems to traverse the lipid bilayer, was found to be rapidly released from the cell surface. Nothing so far is known about the mechanisms allowing a transmembrane orientation of this subclass. (ii) Interior PMA large T, which is stably associated with the cytoskeletal framework of the plasma membrane, exhibits properties of a peripheral membrane protein. It is tempting to speculate that interior PMLA large T might be associated with the plasma membrane in a similar fashion as several other plasma membrane-associated oncogene products found on the cytoplasmic face of the plasma membrane, e.g., polyomavirus m-T, p6osrc and p21ras (2). It amounts to about 50 ng/107 mKSA cells and thus accounts for a substantial proportion of total cellular large T (ca. 2%). PMLA large T is relatively easy to prepare in high purity and is then ready for further biochemical and biological analyses. Such analyses should provide insight into the biogenesis and possible biological functions of this subclass of large T. ACKNOWLEDGMENTS This study was supported by grants De 212/6-1 and Fa 138/3-1 from the Deutsche Forschungsgemeinschaft and a grant from the Fonds der Chemischen Industrie. The Heinrich-Pette-Institut is supported by the Freie und Hansestadt Hamburg and the Bundesministerium fur Jugend, Familie, Frauen und Gesundheit. LITERATURE CITED 1. Ben-Ze'ev, A., A. Duerr, F. Solomon, and S. Penman. 1979. The outer boundary of the cytoskeleton: a lamina derived from plasma membrane proteins. Cell 17:859-865. 2. Bishop, J. M. 1985. Viral oncogenes. Cell 42:23-38. 3. Brunette, D. M., and J. E. Till. 1971. A rapid method for isolation of L-cell surface membranes using an aqueous twophase polymer system. J. Membrane Biol. 5:215-224. 4. Butel, J. S., and D. L. Jarvis. 1986. The plasma-membraneassociated form of SV40 large tumor antigen: biochemical and biological properties. Biochim. Biophys. Acta 865:171-195. 5. Carter, W. G., and S. Hakomori. 1981. A new cell surface, detergent-insoluble glycoprotein matrix of human and hamster fibroblasts. J. Biol. Chem. 256:6953-6960. 6. DeLeo, A. B., H. Shiku, T. Takahashi, M. John, and L. J. Old. 1977. Cell surface antigens of chemically induced sarcomas of the mouse. J. Exp. Med. 146:720-734. 7. Deppert, W., K. Hanke, and R. Henning. 1980. Simian virus 40 T-antigen-related cell surface antigen: serological demonstration on simian virus 40-transformed monolayer cells in situ. J. Virol. 35:505-518. 8. Deppert, W., A. Walser, and U. Klockmann. 1988. A subclass of the adenovirus 72K DNA binding protein specifically associating with the cytoskeletal framework of the plasma membrane. Virology 165:457-468. 9. Deppert, W., and G. Walter. 1982. Domains of simian virus 40 large T-antigen exposed on the cell surface. Virology 122:56-70.

VOL. 63, 1989


10. Flaherty, L., and D. Zimmermann. 1979. Surface mapping of mouse thymocytes. Proc. Nati. Acad. Sci. USA 76:1990-1993. 11. Gluzman, Y., H. Reichl, and D. Solnick. 1982. Helper-free adenovirus type 5 vectors, p. 187-192. In Y. Gluzman (ed.). Eukaryotic viral vectors. Cold Spring Harbor Laboratory. Cold Spring Harbor. N.Y. 12. Gooding, L. R., R. W. Geib, K. A. O'Connel, and E. Harlow. 1984. Antibody and cellular detection of SV40 T-antigenic determinants on the surface of transformed cells. In A. J. Levine, G. F. Vande Wounde, W. C. Topp, and J. D. Watson (ed.), Cancer cells 1: the transformed phenotype. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 13. Graham, F. L., and J. Smiley. 1977. Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol. 36:59-72. 14. Gupta, S. L., G. Goldstein, and E. A. Boyse. 1977. Accessibility of plasma membrane antigens. Immunogenetics 5:379-387. 15. Gurney, E. G., S. Tamowski, and W. Deppert. 1986. Antigenic binding sites of monoclonal antibodies specific for simian virus 40 large T antigen. J. Virol. 57:1168-1172. 16. Jarvis, D. L., W.-K. Chan, M. K. Estes, and J. S. Butel. 1987. The cellular secretory pathway is not utilized for biosynthesis, modification, or intracellular transport of simian virus 40 large tumor antigen. J. Virol. 61:3950-3959. 17. Kit, S., T. Kurimura, and D. R. Dubbs. 1969. Transplantable mouse tumor line induced by injection of SV40 transformed mouse kidney cells. Int. J. Cancer 4:384-392. 18. Klockmann, U., and W. Deppert. 1983. Acylated simian virus 40 large T-antigen: a new subclass associated with a detergentresistant lamina of the plasma membrane. EMBO J. 7:11511157. 19. Klockmann, U., and W. Deppert. 1985. Evidence for transmembrane orientation of acylated simian virus 40 large T antigen. J. Virol. 56:541-548. 20. Lanford, R. E., and J. S. Butel. 1979. Antigenic relationship of SV40 early proteins to purified large T polypeptides. Virology 97:295-306. 21. McArthur, H., and G. Walter. 1984. Monoclonal antibodies specific for the carboxy terminus of simian virus 40 large T antigen. J. Virol. 52:483-491. 22. Mescher, M. F., M. J. L. Jose, and S. P. Balk. 1981. Actin-

3933

containing matrix associated with the plasma membrane of murine tumor and lymphoid cells. Nature (London) 289:139144. 22a.Rinke, Y., and W. Deppert. 1989. Quantitative analysis of cell surface associated SV40 large T antigen using a newly developed 3H-protein A binding assay. Virology 170:424-432. 23. Santos, M., and J. S. Butel. 1982. Association of SV40 large tumor antigen and cellular proteins on the surface of SV40transformed mouse cells. Virology 120:1-17. 24. Santos, M., and J. S. Butel. 1984. Dynamic nature of the association of large tumor antigen and p53 cellular protein with the surfaces of simian virus 40-transformed cells. J. Virol. 49:50-56. 25. Santos, M., and J. S. Butel. 1984. Antigenic structure of simian virus 40 large tumor antigen and association with cellular protein p53 on the surfaces of simian virus 40-infected and -transformed cells. J. Virol. 51:376-383. 26. Santos, M., and J. S. Butel. 1985. Surface T-antigen expression in simian virus 40-transformed mouse cells: correlation with cell growth rate. Mol. Cell. Biol. 5:1051-1057. 27. Sharma, S., L. Rodgers, J. Brandsma, M.-J. Gething, and J. Sambrook. 1985. SV40 T antigen and the exocytic pathway. EMBO J. 4:1479-1489. 28. Soule, H. R., and J. S. Butel. 1979. Subcellular localization of simian virus 40 large tumor antigen. J. Virol. 30:523-532. 29. Stahl, H., and R. Knippers. 1987. The simian virus 40 large tumor antigen. Biochim. Biophys. Acta 910:1-10. 30. Tevethia, M. J., R. W. Anderson, S. S. Tevethia, D. Simmons, J. Feunteun, and C. Coles. 1986. Influence of amino acids encoded in the 3' open reading frame of the SV40 early region on transformation and antigenicity of large T antigen. Virology 150:361-372. 31. Tooze, J. (ed.). 1981. Molecular biology of tumor viruses, part Il: DNA tumor viruses, 2nd rev. ed. Cold Spring Harbor Laboratory. Cold Spring Harbor, N.Y. 32. Warren, L., and M. C. Glick. 1968. The metabolic turnover of the surface membrane of the L cell, p. 3. In L. A. Manson (ed.), Biological properties of the mammalian surface membrane. Wistar Institute, Philadelphia. 33. Whittaker, L., A. Fuks, and R. Hand. 1985. Plasma membrane orientation of simian virus 40 T antigen in three transformed cell lines mapped with monoclonal antibodies. J. Virol. 53:366-373.