Processing of the Transforming Growth Factor β Type I and II Receptors

Three cell surface transforming growth factor β (TGFβ) receptor (R) proteins regulate the effects of TGFβ isoforms on growth and differentiation. TGFβ-IR and -IIR are transmembrane serine/threonine kinases directly mediating the signaling across the plasma membrane. Both TGFβ and its receptors are ubiquitously expressed, hence the fine regulation of the multiplicity of responses most likely involves several levels of control including the regulation of expression, complex formation, and down-regulation of the receptor proteins. In mink lung epithelial cells, TGFβ-IIR was first synthesized as a ∼60-kDa endoglycosidase H-sensitive precursor protein, which was converted to a mature ∼70-kDa protein. The half-life of metabolically labeled mature TGFβ-IIR was estimated to be 60 min and was further reduced to ∼45 min in the presence of exogenous TGFβ1. Minimal internalization of 125I-TGFβ1 at 37°C was detected suggesting that the rapid turnover was not due to endocytosis and degradation of the ligand-receptor complexes. TGFβ-IR was synthesized as a ∼53-kDa precursor protein, which was processed to a mature ∼55-kDa receptor protein. The half-life of TGFβ-IR was >12 h. A fraction of tunicamycin-treated type I and II receptors that reach the cell surface was able to associate in the presence of ligand suggesting that heteromeric complexes can form in a post-endoplasmic reticulum compartment before full glycosylation is achieved. These results show differential processing and turnover of TGFβ-IR and TGFβ-IIR providing a potential additional mechanism for modulation of cellular responses to TGFβs.

The TGF␤ 1 family of proteins participates in the regulation of a variety of biological activities including regulation of cellular growth and phenotype (1)(2)(3). Most cells can produce latent forms of TGF␤, and their activation plays an important regulatory role in TGF␤ actions (4,5). In epithelial cells, TGF␤ treatment leads to inhibition of growth, regulation of the production of extracellular matrix proteins, and modulation of proteolysis (4). The cell surface signaling receptor complex is composed of two transmembrane serine/threonine kinases named type I (ϳ55 kDa) (6) and type II (ϳ70 kDa) (7) TGF␤ receptors. TGF␤-IIR binds the ligand first, after which TGF␤-IR is recruited to a heteromeric complex most likely containing several receptor molecules (8,9). Ligand-dependent phosphorylation of the GS-domain of TGF␤-IR leads to the propagation of the signal downstream (10,11). Type III TGF␤ receptor, a proteoglycan also known as betaglycan, functions mostly as a storage protein as well as in presenting the ligand for the signaling receptors (12). Both TGF␤-IR and TGF␤-IIR are needed to mediate the biological effects of TGF␤ ligands. Recent reports, however, suggest separate signaling pathways for the antiproliferative and the matrix modulatory effects of TGF␤ with the latter only requiring TGF␤-IR signaling in some cell systems (13,14).
Since most cell types can produce both TGF␤ receptors and ligand(s), the regulation of cellular responsiveness relies on the production of active TGF␤ and its presentation to signaling receptors. Therefore, the role of receptor protein associations, turnover, and down-regulation is likely critical for the control of TGF␤ signals and the modulation of overall cellular responsiveness. TGF␤-IIR levels have been shown to correlate with TGF␤ responsiveness (15,16). Cancer cells refractory to TGF␤'s antiproliferative action have often lost TGF␤-IIR expression (17)(18)(19). Although TGF␤-IR can bind ligand only in association with TGF␤-IIR, it is indispensable for TGF␤ responses since phosphorylation of its GS domain provides possible binding sites for intracellular substrates (10,20). Interestingly, cellular transformation by Ha-ras oncogene as well as tropic hormones can down-regulate cell surface binding sites for TGF␤, thus altering TGF␤ responsiveness (21)(22)(23). We have examined the biosynthesis and ligand-induced modulation of naturally expressed TGF␤-IR and TGF␤-IIR in CCL-64 mink lung epithelial cells, which are known to express abundant amounts of all three TGF␤ receptors and are potently growth inhibited by exogenous TGF␤. lacking cysteine and methionine. Following labeling the cell monolayers were prepared as described below.
Cell Lysis and Immunoprecipitation-Cell monolayers in 100-mm tissue culture plates were solubilized with 1 ml of lysis buffer (20 mM Tris-HCl, pH 7.4, 2 mM EDTA, 25 mM NaF, 1% Triton X-100, 1 mM dithiothreitol, 2 mM NaMo 4 , 2 mM NaVO 4 , 1 g/ml aprotinin, 1 g/ml leupeptin) for 30 min at 4°C. After a 12,000 ϫ g centrifugation for 15 min, the lysates were precleared with protein A-Sepharose (Sigma) for 30 min at 4°C and precipitated overnight with a polyclonal TGF␤-IR (V-22, Santa Cruz Biotechnology, Santa Cruz, CA), TGF␤-IIR (C-16, Santa Cruz Biotechnology), or TGF␤-IIR 2732 antibodies. The latter was raised against the extracellular domain of the human type II receptor overexpressed in Sf9 cells as a HIS-tagged protein and provided by Dr. Xiao-Fan Wang (Duke University, Durham, NC). This incubation was followed by a 1.5-h incubation with protein A-Sepharose. The Sepharose particles were washed three times with radioimmune precipitation buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS) and the immune complexes dissociated with lysis buffer containing 1% SDS for 5 min at 95°C. The concentration of SDS was diluted to 0.1% with lysis buffer and a second immunoprecipitation with the same antibody was performed for 2.5 h at room temperature, the last 1.5 h in the presence of protein A-Sepharose. The immune complexes were eluted with Laemmli sample buffer, boiled, resolved by 8% SDS-PAGE, and visualized by autoradiography.
Deglycosylation Procedures-Cell lysates were subjected to overnight immunoprecipitation with TGF␤-IR or TGF␤-IIR antibodies as described above. After washes the immune complexes bound to protein A-Sepharose beads were treated with 50 milliunits of endoglycosidase H (Sigma) in 17 mM phosphate-buffered saline, pH 5.5 (170 mM NaCl, 10 mM sodium phosphate buffer) for 24 h at 37°C or deglycosylated using an enzymatic deglycosylation kit containing NANase II, O-glycosidase DS, and PNGase F according to the manufacturer's protocol (Bio-Rad). Control samples were treated with the same buffers lacking enzymes. After deglycosylation, a second immunoprecipitation was performed as described above, and the samples were resolved by 8% SDS-PAGE and visualized by autoradiography.
Internalization of 125 I-TGF␤-Cell monolayers in 6-well plates were preincubated in a binding buffer (128 mM NaCl, 5 mM KCl, 5 mM MgSO 4 , 1.2 mM CaCl 2 , 50 mM Hepes, pH 7.5, 0.2% BSA) for 15 min at 37°C. Fresh binding buffer containing 1 ng/ml 125 I-TGF␤1 (specific activity, 173 Ci/g; DuPont NEN) with or without 100-fold excess unlabeled TGF␤1 (Genentech, South San Francisco, CA) was added and the plates incubated in a 37°C water bath for different times ranging from 1 to 30 min. After incubation at 37°C the plates were quickly put on ice and washed twice with cold phosphate-buffered saline/0.2% BSA. 125 I-TGF␤1 bound to cell surface receptors was acid washed with 50 mM glycine, pH 2.4, 100 mM NaCl, 2 mg/ml polyvinylpyrrolidone, 2 M urea twice for 3 min at 4°C. After two washes with phosphate-buffered saline, 0.2% BSA, cells were lysed with 1 M NaOH for 30 min at 37°C. 125 I measurements from acid-washed and NaOH-lysed cells, representing both the surface bound not internalized and internalized ligand, respectively, were determined in a Gamma 7000 counter (Beckman Instruments). 125 I-TGF␤ Binding and Affinity Cross-linking-Binding was performed on adherent cells in 100-mm tissue culture dishes as described previously (24). Cells were incubated in binding buffer (see above) containing 1 ng/ml 125 I-TGF␤1 with or without 100-fold excess unlabeled TGF␤1 for 4 h at 4°C with gentle rocking. After two washes with ice-cold binding buffer without BSA on ice, the bound 125 I-TGF␤1 was cross-linked to cell surface receptors with 50 M disuccinimidyl suberate (Pierce) for 15 min at 4°C in 10 ml of binding buffer without BSA. Cells were solubilized in lysis buffer as described above (cell lysis and immunoprecipitation) and the lysates subjected to overnight immunoprecipitation with TGF␤-IR or TGF␤-RII antibodies followed by incubation with protein A-Sepharose for 1.5 h. Immune complexes were then resolved by 5-15% gradient SDS-PAGE and visualized by autoradiography. Exposures on PhosphorImager screens and image analysis with ImageQuant software (Molecular Dynamics, Inc., Sunnyvale, CA) were used to visualize labeled proteins in some cases.
Western Blotting-For Western blotting, 100-g aliquots of cellular protein were separated by 8% SDS-PAGE and transferred to nitrocel- Ci/ml Tran 35 S-label for 20 min and chased for the indicated times with growth medium containing unlabeled methionine (300 g/ml) and cysteine (500 g/ml). Lysis of cells was followed by double immunoprecipitation with TGF␤-IIR antibodies as described under "Experimental Procedures." A, chase time ranging from 0 to 4 h. B, after a 30-min chase and immunoprecipitation, the immune complexes bound to protein A-Sepharose were treated with 50 milliunits of endoglycosidase H (endo H) or a combination of NANase II, O-glycosidase, and PNGase F (deglyc) overnight followed by a second immunoprecipitation with the same antibodies. The immune complexes were analyzed by 8% SDS-PAGE and autoradiography. The arrows indicate the migration of receptor proteins. Molecular mass markers in kDa are shown at the left of each panel.

FIG. 2. Biosynthesis of TGF␤-IR. A, cells were labeled with 500
Ci/ml Tran 35 S-label for 20 min and chased for the indicated times with growth medium containing unlabeled methionine (300 g/ml) and cysteine (500 g/ml). Lysis of cells was followed by double immunoprecipitation with TGF␤-IR antibodies. B, cells were labeled with 50 Ci/ml Tran 35 S-label for 2.5 h, lysed, and precipitated with TGF␤-IR antibodies in the presence (ϩ) or absence (Ϫ) of the immunizing peptide. Immune complexes bound to protein A-Sepharose beads were deglycosylated (deglyc) as in Fig. 1B followed by a second immunoprecipitation with the same antibody. For panels A and B, the immune complexes were analyzed by 8% SDS-PAGE and autoradiography. The arrows indicate the migration of receptor proteins. Molecular mass markers in kDa are shown at the left of each panel.
lulose membranes by semi-dry electrophoretic transfer (Bio-Rad). Nonspecific binding was blocked with 5% nonfat milk in TTBS (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween 20) for 1 h at room temperature. The membranes were incubated with antibodies in the same buffer for 1 h at room temperature and washed three times with TTBS for 10 min each. Bound antibodies were detected using peroxidase-conjugated anti-immunoglobulins (Amersham Corp.) and an enhanced chemiluminescence detection system (Kirkegaard & Perry Laboratories, Gaithersburg, MD).

RESULTS AND DISCUSSION
The Biosynthesis and Turnover of TGF␤-IR and -IIR-Recent studies using cycloheximide to block cellular protein synthesis have suggested that the turnover rate of TGF␤ binding sites in bone cells is relatively fast (25). Since TGF␤-IR cannot bind TGF␤1 alone, these affinity binding studies provide clues only on the turnover of TGF␤-IIR protein. We have followed the fate of metabolically labeled receptors in CCL-64 mink lung epithelial cells. Subconfluent cultures were labeled with Tran 35 S-label for 20 min followed by chase with medium lacking label for variable times. The newly synthesized TGF␤-IIR first appeared as a ϳ60-kDa form with a half-life of Ͻ30 min (Fig. 1A). This form was sensitive to treatment with endoglycosidase H (Fig. 1B). Since endoglycosidase H cleaves only high mannose oligosaccharides but not more complex structures, this result suggests that this ϳ60-kDa form represents an ER, pre-Golgi precursor form. Within 15 min of the chase, a ϳ70-kDa endoglycosidase H-resistant smear appeared with a longer half-life of approximately 60 min. This form was sensitive to deglycosylation by enzymes that remove all N-and O-linked oligosaccharides (Fig. 1B), indicating it represents the mature type II receptor.
TGF␤-IR was also first synthesized as a precursor form of molecular mass ϳ53 kDa that was chased to a mature ϳ55-kDa protein ( Fig. 2A). The observed half-life of both precursor and mature forms was considerably longer than that of TGF␤-IIR. The longer persistence of the type I receptor precursor form indicates a less efficient ER processing compared with that of TGF␤-IIR precursor. The steady-state 35 S-labeled mature TGF␤-IR was sensitive to N-and O-linked deglycosylation (Fig.  2B), thus indicating that the 55-kDa form represents the fully processed cell surface receptor. The higher molecular mass band was not blocked by competing immunizing peptide, indicating it is nonspecific (Fig. 2B).
Effect of TGF␤ on TGF␤-IIR and -IR Turnover-We analyzed next whether TGF␤ could further influence the rapid turnover of TGF␤-IIR. For this purpose, subconfluent cells were labeled with Tran 35 S-label for 2.5 h in methionine/cysteine-free medium followed by chase with full medium. At the end of the labeling (time 0) both the precursor and the mature TGF␤-IIR could be immunoprecipitated from the cell lysates (Fig. 3A). The amount of the mature ϳ70-kDa TGF␤-IIR was higher compared with the precursor form indicating that it represents the predominant and/or more stable receptor species. In cells that were treated with 10 ng/ml TGF␤1 for the last 20 min of labeling as well as throughout the chase, two other proteins coprecipitated with TGF␤-IIR: a ϳ50-kDa and a ϳ37-kDa protein, the latter corresponding to the size of TRIP-1 (TGF␤ receptor interacting protein-1 (26)). The identity of these proteins was not further confirmed, but their ability to coprecipitate with TGF␤-IIR antibodies in a ligand-dependent manner under stringent double immunoprecipitation conditions suggests a strong association induced by exogenous ligand. In the presence of TGF␤1 the half-life of the mature TGF␤-IIR was shortened to approximately 45 min (Fig. 3B).
Labeling with Tran 35 S-label for 2.5 h in methionine/cysteinefree medium followed by chase with full medium indicated that the ϳ55-kDa mature TGF␤-IR form was the predominant one (Fig. 4). Some TGF␤-IR protein could still be immunoprecipitated 18 h after the 2.5-h steady-state labeling indicating that TGF␤-IR half-life was considerably longer than that of TGF␤-IIR. Contrary to the results with TGF␤-IIR, the half-life of steady-state 35 S-labeled TGF␤-IR was not considerably affected by TGF␤1 treatment (Fig. 4).
Internalization of 125 I-TGF␤1-We addressed whether the rapid turnover of TGF␤ receptors in the presence of ligand was due to endocytosis and degradation. To measure the rate of 125 I-TGF␤1 internalization, cells were exposed to the radiolabeled ligand for different times at 37°C. Surface-bound ligand was released from the cell surface by an acid wash procedure. This buffer (pH 2.4) was shown to remove ϳ95% of specifically bound 125 I-TGF␤1 after 2 h of binding at 4°C, conditions under which no ligand internalization should occur (Fig. 5A, bars 2). To confirm the efficacy of our receptor stripping procedure, binding was performed in intact cells after acid wash. Cell membranes were not affected by the transient exposure to low pH, since 125 I-TGF␤1 binding was comparable with that in non-pretreated cells (Fig. 5A, bars 3 versus bar 1).
A time-dependent increase in acid-washable specific binding of 125 I-TGF␤1 was observed (Fig. 5B).Very little ligand was internalized during the first 10 min of the experiment, and even after a 30-min incubation at 37°C the ratio of surface/ internalized ligand was 0.25 (Fig. 5B). These results suggest that the rapid turnover of TGF␤ receptors in the presence of ligand may not be explained by endocytosis and subsequent degradation. This was further supported by Western blot analysis of TGF␤1-treated cells. An overnight incubation with exogenous ligand did not alter TGF␤-IIR content in CCL-64 cells (Fig. 5C).
The possibility that the short half-life results from the release of the extracellular domain of TGF␤-IIR into the medium was investigated by immunoprecipitation of the chase medium after labeling with Tran 35 S-label (see above) with the polyclonal TGF␤-IIR antibody (#2732) raised against the receptor's extracellular domain. Even with a long exposure time no proteins were detected in the precipitated chase medium 1-3 h after cell labeling (data not shown), and small proteolytic fragments also were not detected in the precipitated cell lysates (Fig. 1A).
Association of TGF␤-IR with TGF␤-IIR-Chen et al. (27) reported recently that the cytoplasmic domains of type I and II receptors have an inherent affinity for each other even in the absence of the ligand. The interaction was shown to require kinase activity and thus depended on phosphorylation. Part of the receptors at cell surface exist as hetero-oligomers although TGF␤-IIR homo-oligomers predominate (28,29). We studied the stage at which TGF␤-IR can associate with TGF␤-IIR by stripping oligosaccharide chains from receptor proteins with tunicamycin followed by 125 I-TGF␤1 binding at 4°C, covalent cross-linking, and precipitation with TGF␤-IIR or -IR antibodies. Tunicamycin inhibits the formation of N-glycosidic linkages during protein synthesis with the newly synthesized proteins mimicking the precursor/ER form of the receptors. The trafficking of receptors to the cell surface was not eliminated by tunicamycin. Treatment with 5 g/ml tunicamycin for 5 h was enough to deglycosylate TGF␤-IIR, whereas 24 h of treatment was needed for TGF␤-IR (Fig. 6), consistent with the longer half-life and slower processing in the ER of the type I receptor (Fig. 2). Deglycosylated TGF␤-IIR was able to bind exogenous ligand and associate with both fully processed TGF␤-IR (5 h, lane 2) and deglycosylated TGF␤-IR as judged by coprecipitation by TGF␤-IIR antibodies (24 h, lane 2) and TGF␤-IR anti-  5. Internalization of 125 I-TGF␤1. A, 125 I-TGF␤1 binding at 4°C and acid wash procedure. Cell monolayers grown in 6-well plates were incubated in triplicate with 1 ng/ml of 125 I-TGF␤1 Ϯ 100 ng/ml unlabeled TGF␤1 at 4°C. After 2 h incubation, cells were either lysed in 1 M NaOH (bar 1) or acid washed to remove surface bound ligand and then lysed with NaOH (bars 2). In bars 3, an acid wash step preceded binding, which was followed by acid wash and NaOH lysis. B, kinetics of 125 I-TGF␤1 binding and internalization. Cells were incubated with 1 ng/ml of 125 I-TGF␤1 Ϯ 100 ng/ml unlabeled TGF␤1 for the indicated times at 37°C after which surface-bound non-internalized ligand was removed by an acid wash procedure. The remaining internalized ligand was measured by solubilizing cells with 1 M NaOH. C, Western blot of TGF␤-IIR. Cells grown on 100-mm dishes were treated with or without 10 ng/ml TGF␤1 overnight in improved minimal essential medium/10% fetal calf serum. Cell were lysed and 100 g aliquots were subjected to 8% SDS-PAGE followed by a TGF␤-IIR immunoblot (2732 antibody). Molecular mass markers in kDa are shown at left.
bodies (24 h , lane 4). However, deglycosylated TGF␤-IIR did not coprecipitate with TGF␤-IR antibodies (5 h, lane 4). This could well reflect a lower precipitation efficiency of the latter since TGF␤-IIR antibodies coprecipitated both deglycosylated TGF␤-IIR and mature TGF␤-IR (5 h, lane 2). The lesser amounts of deglycosylated TGF␤-IR that could be detected in the presence of tunicamycin (24 h panel) may reflect diminished trafficking to the cell surface, an alternation in the halflife of the ER form of TGF␤-IR, and/or a critical need of Nlinked glycosylation in TGF␤-IR for ligand binding. These not mutually exclusive possibilities will require further study.
In summary, native TGF␤ type I and II receptors are processed differently and separately in mink lung epithelial cells, with the TGF␤-IIR protein exhibiting a more efficient ER processing and a much shorter half-life (approximately 60 min versus Ͼ12 h) than type I receptor. Studies with exogenous labeled ligand suggested that the fast turnover of TGF␤-IIR protein is not due to receptor endocytosis and subsequent degradation. This short metabolic half-life of native TGF␤-IIR, as measured directly by metabolic labeling, agrees with a recent study in osteoblasts. In this study, suppression of protein synthesis with cycloheximide reduced 125 I-TGF␤1 binding to types I and II receptor with a half-life of 2 h (25). The rapid reduction in binding to TGF␤-IR in this study may not have reflected the stability of newly synthesized type I receptor but could be explained by the reduction in type II receptor, critical for TGF␤-IR binding. Our direct biosynthetic studies suggest a more prolonged half-life for TGF␤-IR in epithelial cells. It is possible perhaps that the turnover of TGF␤ receptors may be different in cells of different lineage and/or altered by endogenous secretion of receptor ligands. This speculation requires further study. The short metabolic half-life of TGF␤-IIR may have important implications for the reversible and rapid modulation of the many TGF␤-mediated cellular responses. In addition its different processing with that of TGF␤-IR allows for a possible additional mechanism of regulation of TGF␤s actions.