Biochimica et Biophysica Acta 1813 (2011) 742–753

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamcr

The TGF-β co-receptor, CD109, promotes internalization and degradation of TGF-β receptors

Albane A. Bizet a, Kai Liu a,1, Nicolas Tran-Khanh b, Anshuman Saksena a, Joshua Vorstenbosch a, Kenneth W. Finnson a, Michael D. Buschmann b, Anie Philip a,⁎ a Division of Plastic Surgery, Department of Surgery, McGill University, Montreal, QC, Canada b Department of Biomedical and Chemical Engineering, Ecole Polytechnique, Montreal, QC, Canada article info abstract

Article history: Transforming growth factor-β (TGF-β) is implicated in numerous pathological disorders, including cancer and Received 11 August 2010 mediates a broad range of biological responses by signaling through the type I and II TGF-β receptors. Received in revised form 19 January 2011 Internalization of these receptors via the clathrin-coated pits pathway facilitates SMAD-mediated signaling, Accepted 24 January 2011 whereas internalization via the caveolae pathway is associated with receptor degradation. Thus, molecules that Available online 2 February 2011 modulate receptor endocytosis are likely to play a critical role in regulating TGF-β action. We previously identified CD109, a GPI-anchored , as a TGF-β co-receptor and a negative regulator of TGF-β signaling. Here, we Keywords: TGF-β receptor demonstrate that CD109 associates with caveolin-1, a major component of the caveolae. Moreover, CD109 CD109 increases binding of TGF-β to its receptors and enhances their internalization via the caveolae. In addition, CD109 Internalization promotes localization of the TGF-β receptors into the caveolar compartment in the presence of ligand and Degradation facilitates TGF-β-receptor degradation. Thus, CD109 regulates TGF-β receptor endocytosis and degradation to Caveolae inhibit TGF-β signaling. © 2011 Elsevier B.V. All rights reserved.

1. Introduction In addition to the type I and II receptors, many cell types also express the TGF-β co-receptors betaglycan and [4]. Our TGF-β5, a multifunctional growth factor with three different group has recently identified another TGF-β co-receptor, CD109, in subtypes, TGF-β1, β2, β3, controls numerous cellular processes such skin cells [5]. CD109 is a 180 kDa glycosylphosphatidylinositol (GPI)- as growth, differentiation, migration and extracellular matrix depo- anchored protein that belongs to the α2-macroglobulin/complement sition, and thus plays a crucial role in development and homeostasis family [6]. CD109 binds the TGF-β1 subtype with high affinity, forms a [1]. Dysregulation of its signaling pathway has been implicated in heteromeric complex with the TGF-β signaling receptors and inhibits tissue fibrosis and cancer, underscoring the critical importance of a TGF-β signaling and responses in a variety of cell types [5]. Moreover, tight regulation of TGF-β action. recent reports have shown that CD109 expression is deregulated in TGF-β signalingistransducedbyapair of transmembrane serine/ basal-like breast carcinoma [7], squamous cell carcinoma [8–11] and threonine kinases known as type I (TGFBR1) and II (TGFBR2) receptors [2]. melanoma [5] and that CD109 is mutated in colorectal cancer [12]. The binding of TGF-β to TGFBR2, a constitutively active kinase, results in Despite its important role in TGF-β signaling and its potential the phosphorylation of TGFBR1. The activated TGFBR1 then propagates significance in cancer progression, its mechanism of action is not yet the signal by phosphorylating its intracellular substrates SMAD2 and understood. SMAD3,whichinturnformcomplexeswithSMAD4.Thecomplexes Receptor endocytosis is a pivotal regulatory mechanism in signal accumulate in the nucleus, where they regulate expression [3]. transduction. TGF-β receptors are internalized via both clathrin- and caveolae-dependent pathways [13]. Internalization of the TGF-β receptors via the clathrin-coated pits has been linked with signaling via SMAD2/3 and receptor recycling [14–17]. In contrast, TGF-β Abbreviations: TGF-β, transforming growth factor-β; Smurf, SMAD ubiquitination regulatory factor; GPI, glycosylphosphatidylinositol; MβCD, methyl-β-cyclodextrin; receptor localization in caveolae is associated with downregulation of EEA-1, early endosome antigen-1; HA, hemagglutinin SMAD2/3 signaling and receptor degradation following ubiquitination ⁎ Corresponding author at: Montreal General Hospital, 1650 Cedar Ave., Rm C9-158, by the E3-ubiquitin ligase Smurf2 [13,18]. McGill University, Montreal, QC, H3G 1A4, Canada. Tel.: +1 514 934 1934 x44535; Because CD109 is a negative regulator of TGF-β signaling, we fax: +1 514 934 8289. investigated whether CD109 exerts this effect by modulating E-mail address: [email protected] (A. Philip). 1 Present address: Department of Colorectal Surgery, Tianjin Medical University caveolae-mediated internalization of the TGF-β receptors or Cancer Hospital, Tianjin, PR China. receptor degradation.

0167-4889/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamcr.2011.01.028 A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753 743

2. Materials and methods 2.6. Sucrose gradient fractionation

2.1. Cell lines Preparation of membrane fraction from HaCaT cells was per- formed as described in Ref. [21]. The twelve fractions collected were The human keratinocyte cell line HaCaT, kindly provided by P. analyzed by western blot for EEA-1 and caveolin-1. The fractions Boukamp (Heidelberg, Germany), Mv1Lu cells and human embryonic containing EEA-1 or caveolin-1 were pooled and equal amounts kidney 293 cells purchased from the American Type Culture Collection, of protein were analyzed by SDS-PAGE/autoradiography or western were cultured as described previously [5]. HaCaT clones stably blot. expressing CD109 (or its empty vector, EV) were selected and cultured in the presence of 0.5 mg/ml Geneticin (Invitrogen, Carlsbad, CA, USA). 2.7. Immunofluorescence microscopy

293 cells were incubated with 250 pM biotinylated-TGF-β1 (R&D, 2.2. Transient transfections and siRNA treatment Minneapolis, MN, USA) for 2 h at 4 °C, then with 10 μg/ml streptavi- din-AF647 for 1 h at 4 °C and placed at 37 °C in DMEM, 10% FBS, for Transfections of different combinations of the following plasmids: 30 min. Cells were fixed with 4% paraformaldehyde and permeabi- CD109 or its empty vector (pCMVSport6), caveolin-1 (gift from C. lized with methanol. After blocking non-specific sites with 10% Hardin, University of Missouri), Dyn2K44A (gift from S. Egan, normal goat serum, cells were stained using an anti-CD109 antibody University of Toronto), TGFBR2 and TGFBR1-WT (gifts from J. Wrana, followed by an Alexa Fluor 488 (AF488) conjugated anti-mouse IgG University of Toronto) were performed using Superfect (Qiagen, (Invitrogen) or by a Cy5-conjugated Fab fragment goat anti-mouse Mississauga, ON, Canada). Alternatively, cells were transfected with IgG (Jackson ImmunoResearch Lab, West Grove, PA, USA). Cells were CD109 siRNA (cat.#129083), caveolin-1 siRNA (cat.#10297) or a then stained with a Cy3 conjugated rabbit anti-caveolin-1 or AF488 negative control siRNA (cat.#4611) (Ambion, Austin, TX, USA) using conjugated mouse anti-HA antibodies (Invitrogen). Images were Lipofectamine 2000 (Invitrogen). acquired with a LSM 510 META Axioplan 2 confocal laser scanning microscope (Carl Zeiss, Toronto, ON, Canada) as described previously [22]. Background was removed by median filtering [23] or by 2.3. Affinity labeling deconvolution using Huygens Professional Software (Fig. 1D and

125 Supplementary Fig. S2D). 293 cells or HaCaT cells were incubated with 100 pM I-TGF-β1 for 3 h at 4 °C, the ligand was cross-linked to the receptors with a non- 2.8. Plasma membrane preparation permeable cross-linker BS3 (Pierce, Rockford, IL) and the cells were incubated at 37 °C for 0 to 8 h prior to lysis as previously described Cell surface were extracted using the Hook Cell Surface [19]. Protein concentrations were determined using a Lowry protein Protein Isolation (G-Biosciences, St. Louis, MO, USA). Briefly, cell assay (Bio-Rad, Mississauga, ON, Canada) and the lysates were surface proteins were labeled with Hook-sulfo-NHS-SS-Biotin and analyzed by SDS-PAGE/autoradiography. Densitometry was per- lysed. An aliquot was analyzed by western blot for actin to make sure formed using the ImageJ software. that equal amount of protein was used. Biotinylated proteins from the remaining lysates were isolated using streptavidin-agarose column, 2.4. Immunoprecipitation and western blot eluted and analyzed by western blot.

Lysates from 293 cells were immunoprecipitated with a mouse 2.9. Statistical methods monoclonal anti-caveolin-1 antibody (clone 2297, BD Biosciences, ≥ Mississauga, ON, Canada). Lysates from HaCaT cells were immuno- Numerical results are represented as means of n 3 independent fi precipitated with mouse monoclonal anti-CD109 (gift from R&D experiments±SEM. Statistical signi cance between two groups was systems, Minneapolis, MN, USA) or a rabbit polyclonal anti-caveolin-1 determined by two-tailed Student t test. Comparisons within more antibody (BD Biosciences), while a mouse monoclonal anti-histone than two groups were made by one-way analysis of variance. Multiple – H1 or a rabbit polyclonal anti-HA antibodies (both from Santa Cruz comparisons were made by Holm Sidak test (post-hoc). Biotechnology, Santa Cruz, CA, USA) were used as control IgG. Western blot analyses were conducted with the following antibodies: 3. Results mouse monoclonal anti-CD109 (TEA 2/16, BD Biosciences), anti-EEA- 1(BD Biosciences), anti-SMAD3 (FL-425) and anti-actin (H-300) 3.1. CD109 associates with caveolin-1 at endogenous concentration antibodies (both form Santa Cruz Biotechnology), rabbit polyclonal anti-TGFBR1 and anti-phospho-SMAD3 antibodies (both from Cell We sought to determine whether CD109 exerts its negative effects β β Signaling Technologies, Danvers, MA, USA). on TGF- signaling by modulating TGF- receptor localization into the caveolae, a process shown to enhance TGF-β receptor degradation. We first examined whether CD109, a GPI-anchored protein, can associate 2.5. Internalization assay with caveolin-1, a major component of the caveolae. Co-immunopre- cipitation experiments reveal that CD109 associates with caveolin-1 in Internalization assay was performed as described in Ref. 293 cells co-transfected with both CD109 and caveolin-1 (Fig. 1A). In [20] with some modifications. Briefly, cells were treated with the absence of caveolin-1 transfection, (i.e. when caveolin-1 expres- 2.5 mM MβCD (Sigma Aldrich, Oakville, ON, Canada) for 1 h at 37 °C sion is undetectable [24]), CD109 is not co-immunoprecipitated prior to internalization assay. Cells were then incubated with (Fig. 1A), demonstrating the specificity of the association. 125 100 pM I-TGF-β1 (Perkin Elmer, Waltham, MA, USA) with or In untransfected HaCaT cells, caveolin-1 is co-immunoprecipitated without 10 nM unlabeled TGF-β1 (to determine non-specific binding) by an anti-CD109 antibody (Fig. 1B), confirming that the interaction for 2 h at 4 °C and transferred to 37 °C for the indicated time. between CD109 and caveolin-1 is not an artifact of overexpression Surface ligand was removed with 150 mM NaCl, 0.1% acetic acid, 2 M and that it occurs at endogenous concentrations of caveolin-1 and urea. Cells were then solubilized in order to extract internalized CD109 (Fig. 1B). Similarly, an anti-caveolin-1 antibody co-immuno- ligand. precipitates small but detectable amounts of CD109 (Supplementary 744 A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753

Fig. S1 and Fig. 1C). Importantly, the association between CD109 and 3.3. CD109 enhances binding of TGF-β to TGF-β receptors and TGF-β-bound caveolin-1 is enhanced after TGF-β treatment (Fig. 1C). receptor internalization Immunofluorescence microscopy of HaCaT cells (Fig. 1D) shows that caveolin-1 (in red) displays a punctate staining pattern at the We then asked if CD109 affects TGF-β-bound receptor internaliza- plasma membrane, with some punctate staining in the cytoplasm, tion. CD109 overexpression significantly (pb0.05, n=5 independent which may represent neo-synthesized or internalized caveolin-1. experiments) increases the number of cells that have internalized Endogenous CD109 (blue) localizes mainly at the plasma membrane biotin-TGF-β (i.e. TGF-β-bound receptors) as compared to empty vector in both non-caveolar and caveolin-1 positive compartments. The (EV) transfection in 293 cells (Fig. 2A and B). After normalization for colocalization of CD109 with caveolin-1 (in purple, Fig. 1D) suggests transfection efficiency (as determined by caveolin-1 expression), we that CD109 is able to associate with caveolin-1 in the same cellular observe that 64% of caveolin-1 positive cells internalize biotin-TGF-β in compartment, supporting the results obtained in Fig. 1A–C. the presence of CD109, whereas only 27% incorporate biotin-TGF-β in EV transfected cells (Fig. 2B), suggesting that CD109 increases the internalization of TGF-β-bound receptors. The amount of internalized 3.2. CD109 and TGF-β receptors colocalize in caveolae biotin-TGF-β reflects the levels of receptors that underwent endocyto- sis, and also any recycling and/or degradation that may have occurred We have previously shown that CD109 binds directly to TGFBR1 [5]. during that time period. Notably, CD109 transfection does not alter the We therefore sought to determine whether TGF-β receptors co-localize level of TGF-β receptors or caveolin-1, as compared to EV transfection with CD109 in caveolae. Because the endogenous levels of TGF-β (Fig. 3B and data not shown), suggesting that CD109 only affects the receptors in HaCaT cells are too low to allow detection of the amount of internalized TGF-β-bound receptors. internalized ligand, 293 cells were transfected with TGFBR1/TGFBR2 We next examined the effect of CD109 on the internalization of 125 and caveolin-1 and labeled with biotin-TGF-β, followed by streptavidin- I-TGF-β at endogenous TGF-β receptor concentration in HaCaT cells. 125 AF647 treatment. After 30 min at 37 °C to allow receptor and ligand The I-TGF-β internalization assay has been shown to reliably reflect internalization, the cells were immunostained for CD109 and caveolin-1 receptor internalization [20,25]. At early time points (15 and 30 min), 125 (Fig. 1E). Importantly, a significant amount of triple colocalization CD109 enhances I-TGF-β internalization (Fig. 2C). The increase of 125 (visualized in white) is observed between the internalized biotin-TGF-β I-TGF-β internalization in CD109 overexpressing cells is followed by a (in green), CD109 (in blue), and caveolin-1(in red)(Fig. 1E). The decline after 30 min whereas a decline occurs only after 60 min in EV 125 proportion of internalized biotin-TGF-β that is not associated with transfected cells (Fig. 2C). The decline in internalized I-TGF-β has been CD109 and caveolin-1 (Fig. 1E) may represent molecules that are reported previously and might be due to ligand depletion and/or internalized via the clathrin pathway (associated with SMAD2/3 degradation [20]. Importantly, our results suggest that CD109 enhances signaling) or that are in recycling compartments. In the absence of ligand-bound receptor internalization. TGF-β treatment, CD109 localizes mainly at the plasma membrane, Interestingly, CD109 overexpression is associated not only with an 125 where it partially co-localizes with caveolin-1 (Fig. 1F), as observed in increase of I-TGF-β internalization, but also with an increase of 125 untransfected HaCaT cells (Fig. 1D). When the cells were incubated at surface-bound I-TGF-β at 30 min (Fig. 2D). The ratio of internalized/ 125 4 °C instead of 37 °C, no biotin-TGF-β staining was detected (data not surface-bound I-TGF-β (I/S) is higher in CD109 transfected cells, as shown), indicating that at 4 °C, the receptors are not internalized. compared to EV cells at both 15 and 30 min (Fig. 2E), suggesting that Interestingly, in the absence of TGF-β receptor transfection, low levels of CD109 accelerates TGF-β-bound receptor endocytosis. Moreover, co-localization between CD109 and caveolin-1 could be observed knocking down CD109 expression using CD109 siRNA or CD109 (Fig. 1G and Supplementary Fig. S2A). It is possible that the interaction antisense morpholino oligos leads to a decrease in both internalized 125 between the GPI-anchored CD109 and the integral membrane protein and surface-bound I-TGF-β at 30 min (Fig. 2F and Supplementary caveolin-1 is mediated via endogenous TGF-β receptors in that case, Fig. S3). These results indicate that endogenous CD109 increases surface although the contribution of other transmembrane proteins or lipid binding and internalization of TGF-β. molecules cannot be ruled out. To determine whether the increase in TGF-β binding to the cell The notion that the biotin-TGF-β staining is specific for biotin-TGF- surface is due to an increase in TGF-β binding to the TGF-β receptors and β bound to TGFBR1/TGFBR2 complex and not to other TGF-β binding not solely due to an increase in TGF-β binding to CD109, we performed molecules under our detection settings is supported by the following affinity labeling of cell surface receptors at 4 °C followed by SDS-PAGE results: (1) when TGFBR1/TGFBR2 are not transfected, no staining is analysis. Our results show that CD109 is able to enhance TGF-β binding observed for biotin-TGF-β, even in the presence of CD109 (Fig. 1G); to the receptors (Fig. 3A and B). CD109 siRNA transfection in HaCaT cells (2) biotin-TGF-β colocalizes with HA-tagged TGFBR1 (Supplementary decreases and CD109 overexpression in 293 cells increases the amount 125 Fig. S2C). This suggests that TGF-β follows the same internalization of cell surface I-TGF-β-bound receptors (Fig. 3A and B). Notably, route as TGFBR1, as reported by others [13,20]; (3) HA-TGFBR1 co- CD109 does not affect the total level of TGFBR1, as demonstrated by localizes with CD109 in caveolae vesicles and at the plasma western blot (Fig. 3A and B). Moreover, overexpression of CD109 does membrane (Supplementary Fig. S2D). Together, these data indicate not alter the cell-surface level of TGF-β receptors (Fig. 3B). These results, that biotin-TGF-β can be used to assess TGF-β receptor localization together with our previous findings demonstrating that CD109 forms a and suggest that CD109 is able to form a complex with TGF-β-bound complex with the TGF-β receptors [5], suggest that such association is an TGF-β receptors (activated receptors) in caveolae. important step in CD109's ability to increase TGF-β binding to this

Fig. 1. CD109 associates with caveolin-1 and colocalizes with the TGF-β receptors in caveolae. (A–C) Co-immunoprecipitation: (A) Cell lysates from 293 cells transfected with CD109 and caveolin-1 or their corresponding empty vector were immunoprecipitated with an anti-caveolin-1 (cav-1) antibody followed by western blot with an anti-CD109 antibody. The IgG band demonstrates equal loading. (B) Cell lysates from non-transfected HaCaT cells were immunoprecipitated with an anti-CD109 antibody or with an anti-histone antibody (used as control IgG), followed by western blot for CD109 and caveolin-1. (C) Non-transfected HaCaT cells were treated with or without 25 pM TGF-β for 30 min. Cell lysates were then subjected to immunoprecipitation with an anti-caveolin-1 antibody or with an anti-HA antibody (used as control IgG), followed by western blot for CD109 and caveolin-1. The densitometry of CD109 levels in the immunoprecipitate is indicated. (D–G) Immunofluorescence: (D) HaCaT cells were immunostained for endogenous caveolin-1 (red) and endogenous CD109 (blue). Co-localization of CD109 with caveolin-1 appears in purple. (E) 293 cells transfected with CD109, caveolin-1, TGFBR1/TGFBR2 were stained with biotin- TGF-β/streptavidin (green) and incubated for 30 min at 37 °C to allow receptor internalization. After fixation, cells were stained for CD109 (blue) and caveolin (red) and analyzed by confocal microscopy. Triple co-localization of caveolin-1, CD109, and biotin-TGF-β appears in white. A representative cell of n=5 independent experiments is shown. (F) 293 cells transfected as in (E) were left untreated and stained for CD109 (blue) and caveolin-1 (red). Co-localization of CD109 with caveolin-1 appears in purple. A representative cell of n=3 experiments is shown. (G) 293 cells transfected with CD109, caveolin-1, TGFBR1/TGFBR2 or their corresponding empty vectors were stained as in (E). The arrow points at a cell where colocalization between CD109 and caveolin-1 occurs. A representative panel of cells is shown (n=3 experiments) A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753 745 multimeric TGF-β receptor complex and enhances the internalization of Internalization of biotin-TGF-β in EV transfected 293 cells is this TGF-β bound-complex. prevented by co-transfection of the Dynamin2-K44A mutant (Dyn2K44A) (Fig. 4A), which blocks endocytosis via both clathrin 3.4. CD109 enhances TGF-β receptor internalization via the caveolar and caveolar pathways [26] and is consistent with previous reports pathway demonstrating that TGF-β receptors are internalized via clathrin and/ or caveolae pathways [13,27]. Importantly, internalization of biotin- Next, we assessed if the CD109-induced increase in TGF-β receptor TGF-β is also blocked by co-transfection of the Dynamin2-K44A internalization could be mediated by the caveolar compartment. mutant (Dyn2K44A) when CD109 is overexpressed (Fig. 4A). In the 746 A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753

Fig. 2. CD109 promotes TGF-β internalization. (A–B) Immunofluorescence: 293 cells were stained with biotin-TGF-β/streptavidin (green) and incubated for 30 min at 37 °C to allow receptor internalization. Cells were fixed and immunostained for CD109 (blue) and caveolin (red) and analyzed by confocal microscopy. (A) Comparison of populations of CD109 and control empty vector (EV) transfected cells. A representative cell showing internalization of biotin-TGF-β is enlarged in the left corner of the overlay. (B) Percentage of transfected 125 (caveolin-1 positive) cells that have internalized biotin-TGF-β (mean of n=5 independent experiments±SEM, *: pb0.05). (C–F) I-TGF-β internalization assay: (C) Left panel: 125 125 HaCaT cells transfected with CD109 or EV were treated with 100 pM I-TGF-β and incubated at 37 °C for the indicated times. The graph shows the amount of internalized I-TGF-β1 125 (expressed as specific cpm). Right panel: Western blot showing CD109 overexpression in HaCaT cells. (D) Amount of surface-bound and internalized I-TGF-β after 30 min incubation at 37 °C in HaCaT cells transfected with CD109 or EV (normalized to surface-bound EV). (E) Ratio of internalized/surface-bound 125I-TGF-β (I/S) in HaCaT cells transfected 125 with CD109 or EV. (F) Left panel: Amount of surface-bound and internalized I-TGF-β after 30 min incubation at 37 °C in HaCaT cells transfected with CD109 siRNA or control siRNA (normalized to surface-bound control siRNA). Right panel: Western blot showing CD109 siRNA efficacy. All graphs show the mean of nN3 independent experiments, ± SEM, *: pb0.05. presence of Dyn2K44A, the biotin-TGF-β staining appears at the caveolae, a route that has been shown to lead to receptor degradation plasma membrane in 81 (± 5.2)% of the cells and co-localizes with and signaling downregulation. We demonstrate that pre-treatment CD109 staining, whereas the biotin-TGF-β staining is found at the with methyl-β-cyclodextrin (MβCD), an inhibitor of the caveolae/ plasma membrane in only 5.1 (± 4.6)% of the cells in the absence of lipid raft-dependent endocytosis [17,26] blocks CD109's effect on 125 Dyn2K44A (Fig. 4A). This indicates that CD109 promotes dynamin- I-TGF-β internalization in HaCaT cells (Fig. 4B). The increase in dependent internalization of the TGF-β-bound receptors. TGF-β internalization mediated by CD109 is also reversed by Because CD109 is an inhibitor of TGF-β signaling [5], we looked caveolin-1 siRNA (Fig. 4C), confirming that CD109 is able to enhance more specifically at CD109's ability to increase internalization via caveolae-mediated endocytosis. The slight (and non-significant) A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753 747

Fig. 3. CD109 enhances TGF-β binding to the TGF-β receptors without altering the total or cell surface receptor levels. (A) HaCaT cells transfected with control or CD109 siRNA were 125 affinity labeled with 100 pM I-TGF-β. Left panel: The amount of cell surface receptors bound to TGF-β is visualized by autoradiography. CD109 is detected as two bands of 180 and 150 kDa, as described previously [5]. Right panel: The same lysates were analyzed for TGFBR1 by western blot. (B) 293 cells transfected with TGFBR1, TGFBR2 and CD109 or its control 125 EV were affinity labeled with 100 pM I-TGF-β. Left panel: Autoradiography showing the amount of cell surface TGF-β-bound receptors. Middle panel: The same lysates were analyzed for TGFBR1 by western blot. Right panel: Cell surface proteins from 293 cells transfected with TGFBR1/TGFBR2, CD109 or EV were labeled with Hook-sulfo-NHS-SS-Biotin, pulled down with streptavidin and analyzed by western blot with CD109 and TGFBR1 antibodies. Similar amounts of protein were used in the experiment as demonstrated by the actin blot. All results are representative of n=3 independent experiments and all lanes from a panel come from the same blot.

decrease observed after MβCD treatment or caveolin-1 siRNA CD109 on TGFBR1 trafficking occurs in a ligand-independent manner. transfection in EV transfected cells suggests that the internalization By sucrose gradient centrifugation, the raft fractions (fractions 5 and of TGF-β via the caveolar pathway at basal conditions is modest in 6) were separated from the non-raft fractions (fractions 8 to 12). The HaCaT cells. This is not consistent with the report showing that purity of the raft and the non-raft fraction was verified using the early about 50% of TGF-β receptors localized in caveolae in Mv1Lu cells endosome antigen-1 (EEA-1) and caveolin-1, two markers conven- [13]. To determine whether the discrepancy is due to cell-type tionally used for the non-raft/endosomal and the raft/caveolar specific differences, we determined whether CD109 regulates TGF-β compartment, respectively [13,21] (Fig. 5A). The analysis of TGFBR1 internalization into caveolae in Mv1Lu cells. In this cell type, CD109 and CD109 localization in the pooled raft and pooled non-raft 125 siRNA decreases I-TGF-β internalization as expected, in the fractions were performed on the same gel by western blot (Fig. 5B). absence of MβCD (Fig. 4D). Importantly, treatment of MβCD leads In the absence of exogenous TGF-β, TGFBR1 is present predominantly 125 to a reduction of I-TGF-β internalization in control siRNA but not in in the non-raft fraction in both CD109 and EV transfected cells. In CD109 siRNA transfected cells, indicating that, in Mv1Lu cells, a large contrast, TGF-β treatment leads to a marked increase in the amount of proportion of TGF-β is internalized via the caveolae and that TGFBR1 in the raft fraction of CD109 transfected cells as compared endogenous CD109 mediates this effect. In addition, these results to EV transfected cells (Fig. 5B). In addition, TGF-β treatment of support the argument that the amount of TGF-β internalizing via the EV transfected cells results in a modest increase in the amount of caveolae varies between cell-types. Also, the difference observed in TGFBR1 in the raft fraction (Fig. 5B), presumably due to the presence control siRNA Mv1Lu or EV HaCaT cells in presence of MβCD is due to of endogenous CD109. The low amount of TGFBR1 detected in the cell type specificity since no difference was observed between EV raft fraction in HaCaT cells is consistent with our previous results in and control siRNA transfection in the absence of MβCD (data not this cell type (Fig. 4B) showing that only a small amount of TGF-β- shown). Together, our results suggest that CD109 increases the bound receptors is internalized via the lipid raft/caveolar pathway proportion of TGF-β-bound receptors that internalize via the caveolar in EV transfected HaCaT cells. Under basal conditions, CD109 pathway. localizes mainly in the non-raft compartment, but is also present in the lipid raft compartment (visualized on a more exposed film), which 3.5. CD109 promotes TGFBR1 compartmentalization into the caveolae in agrees with our immunofluorescence data (Fig. 1D). The effect of TGF- a ligand-dependent manner β on the pattern of CD109 localization is similar to the one observed for TGFBR1: TGF-β treatment increases the localization of CD109 in the raft TGF-β receptors have been shown to traffic independently of compartment, which is consistent with our co-immunoprecipitation ligand addition [13,17]. Thus, we determined whether the effect of data (Fig. 1C). Together, these results suggest that CD109 promotes 748 A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753

Fig. 4. CD109 enhances TGF-β internalization via the caveolar pathway. (A) Immunofluorescence: 293 cells transfected with TGFBR1/TGFBR2, CD109 (or its EV) and Dyn2K44A (or its EV) were treated with biotin-TGF-β/streptavidin (green), incubated for 30 min at 37 °C and stained for CD109 (blue). Colocalization of CD109 and biotin-TGF-β appears in turquoise 125 (overlay). Representative cells are shown. (B–D) I-TGF-β internalization assay: (B) HaCaT cells transfected with CD109 or EV were treated with or without MβCD prior to incubation 125 125 125 with I-TGF-β. The graph shows the amounts of internalized I-TGF-β after 30 min incubation at 37 °C (normalized to EV, no treatment). (C) Left panel: Amount of internalized I- TGF-β after 30 min incubation at 37 °C in HaCaT cells co-transfected with CD109 or EV and with caveolin-1 or control siRNA (normalized to EV, control siRNA). Right panel: Western 125 blot showing caveolin-1 siRNA efficacy. (D) Left panel: Mv1Lu cells transfected with CD109 siRNA or control siRNA were treated with or without MβCD prior to incubation with I- TGF-β. The graph shows the amounts of internalized 125I-TGF-β after 30 min incubation at 37 °C (normalized to CD109 siRNA, no treatment). Right panel: Western blot showing CD109 siRNA efficacy in Mv1Lu cells. All graphs show the mean of n≥3 independent experiments±SEM, *: pb0.05.

TGFBR1 entry into the lipid raft and that this effect is enhanced by TGF-β 3.6. Inhibition of TGF-β signaling by CD109 involves caveolae-mediated treatment. internalization We further examined the ability of CD109 to induce TGF-β-bound receptor localization to the caveolar compartment by first affinity The results presented so far show that CD109 promotes TGF-β 125 labeling the receptors on HaCaT cells with I-TGF-β and then receptor internalization via the caveolae. We next investigated if this performing sucrose gradient fractionation as describe above. In step is involved in mediating CD109's inhibitory effect on TGF-β CD109 transfected cells, the majority of TGF-β-bound TGFBR1 signaling. Our results show that overexpression of CD109 inhibits localizes into the caveolin-1 positive/lipid raft fraction, whereas in TGF-β-induced SMAD3 phosphorylation (Fig. 6A), as expected [5]. EV transfectants, TGFBR1 resides mainly in the non-raft fraction This inhibitory effect of CD109 is abrogated by co-transfection of (Fig. 5C). Similar results were obtained with TGF-β-bound TGFBR2 caveolin-1 siRNA (Fig. 6A), indicating that CD109' effects on TGF-β (data not shown). These results indicate that CD109 leads to signaling involve the caveolar pathway. Similarly, knocking down preferential localization of TGFBR1 into the lipid raft compartment, CD109 expression using siRNA results in an increase in TGF-β-induced in the presence of TGF-β. SMAD3 phosphorylation, and this effect is blocked by co-transfection A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753 749

Fig. 5. CD109 promotes TGFBR1 entry into the caveolar compartment in the presence of TGF-β. (A) Lysates from HaCaT cells transfected with CD109 or EV and treated without (−)or with (+) 25 pM TGF-β for 30 min were subjected to sucrose gradient fractionation. The resulting fractions were analyzed by western blot for EEA1 and caveolin-1. (B) Left panel: The pooled raft (fractions 5–6) and pooled non-raft (fractions 8–12) fractions were analyzed by western blot for CD109 and TGFBR1. A representative western blot is shown. Right panel: 125 Densitometric analysis of TGFBR1 levels. (C) Left panel: HaCaT cells transfected with CD109 or EV were affinity labeled with I-TGF-β prior sucrose gradient fractionation. Labeled TGFBR1 is detected by autoradiography. Coomassie blue staining shows equal protein loading. The same samples were analyzed for EEA-1 and caveolin-1 by western blot. Right panel: Densitometric analysis of the autoradiogram.

of caveolin-1 siRNA (Fig. 6B). This result suggests that downregulation inhibitor, indicating that CD109 enhances proteasomal degradation of of TGF-β signaling by endogenous CD109 involves the caveolar TGF-β receptors (Fig. 7B). Moreover, HaCaT cells transfected with pathway. CD109 siRNA and treated with TGF-β display a significant (pb0.05) increase in TGFBR1 levels, as compared to control siRNA transfection 3.7. CD109 accelerates TGF-β receptor degradation and this effect of CD109 can be blocked by addition of MG132, but not chloroquine, an inhibitor of lysosomal degradation (Fig. 7C). These Localization of TGF-β receptors to caveolae has been shown to be results indicate that endogenous CD109 facilitates proteasomal associated not only with a decrease in SMAD2/3 phosphorylation but degradation of TGF-β receptors and confirm that the previous results also with receptor degradation [13]. Therefore, we tested if CD109 (Fig. 7A and B) were not an artifact due to CD109 overexpression. modulates TGF-β receptor degradation, by monitoring the levels of Collectively, our results suggest that CD109 negatively regulates affinity labeled receptors (Fig. 7A). Consistent with our previous TGF-β action by promoting TGF-β receptor internalization to the 125 findings, CD109 overexpression increases the binding of I-TGF-β to caveolar compartment and by enhancing TGFBR1 degradation via the its receptors, compared to EV transfection (see time 0). TGFBR1, proteasomal pathway (Fig. 8). TGFBR2 and CD109 levels decrease over time, suggesting that the CD109-TGF-β-bound-receptor complex is being degraded (Fig. 7A). 4. Discussion Quantification of TGFBR1 levels by densitometry (expressed as a percentage of time 0) reveals that CD109 accelerates degradation of Our group has previously identified CD109, a GPI-anchored TGF-β-bound receptor (Fig. 7A, right panel, pb0.05). In addition, protein, as a novel TGF-β co-receptor [5]. CD109 has high affinity stable transfection of CD109 in HaCaT cells (clone CD3-3 and CD2-4) for TGF-β1, forms a complex with the TGF-β signaling receptors and accelerates the degradation of endogenous TGFBR1 and TGFBR2, as negatively regulates TGF-β signaling in vitro [5]. Mutation of CD109 or compared to stable transfection of EV (clone EV3-3 and EV1-6) deregulation of its expression has been reported in many cancers [7– (Fig. 7B and Supplementary Fig. S4, pb0.05). Furthermore, the 12,28]. Because of CD109's potential relevance as a regulator of TGF-β increased rate of degradation is prevented by MG132, a proteasome action in vivo, we examined the mechanism by which it regulates TGF- 750 A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753

Fig. 6. CD109's inhibition of TGF-β signaling involves the caveolar pathway. (A–B) Western blot: (A) HaCaT cells transfected with CD109 or its EV, and, caveolin-1 siRNA or control siRNA were treated with 15 pM TGF-β. Cell lysates were analyzed by western blot for phosphorylated-SMAD3 (P-SMAD3) and total SMAD3 (T-SMAD3). Bottom panel: Densitometry of P-SMAD3 after 30 min TGF-β treatment (mean of n=4 independent experiments±SEM, *pb0.05). (B) HaCaT cells transfected with CD109 siRNA, caveolin-1 siRNA or control siRNA were treated with 15 pM TGF-β for 15 min. Cell lysates were analyzed by western blot as in (A). Bottom panel: Densitometry of P-SMAD3 (mean of n=4 independent experiments±SEM, *pb0.05).

β signaling. Here, we show that CD109 increases TGF-β binding to its and that this event is regulated by CD109. Thus, an alteration in CD109 receptors, promotes TGF-β receptor compartmentalization and concentration, as seen in many cancers, may lead to a change in the internalization into the caveolae and accelerates TGF-β receptor proportion of TGF-β receptor internalizing via caveolae, resulting in proteasomal degradation. Together, these findings suggest that aberrant TGF-β signaling and action. downregulation of TGF-β signaling by CD109 involves these mech- Several molecules have been implicated in the regulation of TGF-β anisms (Fig. 8). In addition, our results suggest that ligand addition receptor compartmentalization. Hyaluronan has been shown to may regulate CD109's effect on TGF-β receptor localization to lipid modulate TGF-β receptor partitioning to caveolae [21], but the raft/caveolar compartment. question of hyaluronan controlling TGF-β receptor degradation Caveolae serve as platforms for signaling molecule assembly, remains open. The TGF-β co-receptors betaglycan and its homologue endocytosis of receptors, and modulation of signaling [29]. They are a endoglin have been shown to facilitate TGF-β receptor endocytosis subset of lipid rafts characterized by the insertion of caveolin-1 in the through their interaction with β-arrestin2, a component of the membrane bilayer and are enriched in GPI-anchored proteins. Here, clathrin-coated pits [31,32]. However, there is a discrepancy in the we report that caveolin-1 associates with CD109, suggesting that results reported, since internalization via the clathrin pathway leads CD109 can localize to caveolae. This association is likely indirect and to inhibition of TGF-β signaling in one study [31], and delay in may involve lipid molecules or transmembrane proteins, such as receptor degradation in another study [33]. The proportion of TGF-β TGFBR1, which has been shown to interact directly with both CD109 receptors that internalize via the caveolae may vary depending on the [5] and caveolin-1 [18]. Interestingly, immunofluorescence and cell type (possibly due to a difference in the levels of caveolin, sucrose gradient analyses reveal that CD109 is present in both receptors and/or co-receptors), as observed in the current study, and caveolar and non-caveolar compartments. Whether the non-caveolar may explain the discrepancy. Here, we demonstrate that, in contrast compartment represents a storage area for CD109 or whether CD109 to other TGF-β co-receptors, CD109 may regulate TGF-β receptor has other roles in that compartment is not known. internalization and degradation in a ligand-dependent manner. The ability of CD109 to enhance the binding of TGF-β to its Although the events downstream of TGF-β receptor internalization receptors on one hand and to inhibit TGF-β signaling on the other are ligand-dependent [34,35], studies on TGF-β receptor endocytosis suggests that CD109 may direct the ligand and its receptors into a have been unable to demonstrate an effect of ligand on TGF-β receptor compartment where they cannot signal. Although our results do not compartmentalization or trafficking [13,36]. Our results reveal that rule out the possibility that CD109 may also play a role in clathrin- ligand may play a role in CD109's effect on TGF-β receptor compart- mediated endocytosis of TGF-β receptor, our data showing that mentalization. By sucrose gradient fractionation, we show that CD109 inhibitors of caveolae-mediated endocytosis reverse CD109's effect on promotes localization of TGFBR1 into the caveolae, in the presence of TGF-β internalization suggest that CD109 facilitates TGF-β-bound TGF-β.Ourfindings are in line with the results obtained for several receptor internalization via the caveolae, a compartment associated receptors, such as insulin, β2adrenergic and EGF receptors, that localize with TGF-β receptor degradation and signaling downregulation into caveolae in a ligand-dependent manner [37–39]. Although the [13,18]. Accumulation of receptors in caveolae has been shown to molecular mechanisms involved in TGF-β receptor internalization are trigger the rapid internalization via caveolae [30]. Thus, CD109 may less well understood than those of the receptors above, our study facilitate receptor clustering in the presence of ligand to promote their provides evidence that ligand addition can modulate TGF-β receptor internalization via the caveolar route. Together, our study demon- localization into caveolae in the presence of CD109. These findings strates that activated TGF-β receptors can internalize via the caveolae implicate a link between ligand-induced TGF-β receptor A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753 751

Fig. 7. CD109 accelerates TGFBR1 proteasomal degradation. (A–B) Affinity labeling: (A) 293 cells transfected with CD109 or EV and TGFBR1/TGFBR2 were affinity labeled at 4 °C with 125 I-TGF-β, which was then covalently cross-linked to the receptors, and incubated at 37 °C for 0 to 8 h to allow receptor internalization. Left panel: The labeled receptors were visualized by autoradiography. Coomassie blue staining shows equal protein loading. Right panel: The amount of labeled TGFBR1 was quantified by densitometry (mean of n=4 independent experiments±SEM, * difference between EV and CD109 groups is statistically significant: pb0.05). (B) HaCaT cells stably transfected with CD109 (clone CD3-3) or EV (clone EV3-3) were affinity labeled as in (A), in the presence or absence of 20 μM MG132. Left panel: Labeled TGFBR1 and TGFBR2 were visualized by autoradiography. Coomassie blue staining shows equal protein loading. Middle panel: The amount of labeled TGFBR1 was quantified by densitometry (mean of n=3 independent experiments±SEM, *: pb0.05). Right panel: Western blot showing CD109 overexpression in CD3-3 clone, compared to EV3-3 clone. (C) Western blot: HaCaT cells transfected with control or CD109 siRNA were treated with 100 pM TGF-β for 16 h, in the presence or absence of MG132 or chloroquine. Cell lysates were analyzed by western blot using the indicated antibodies (left panel). Right panel: Densitometry of TGFBR1 (mean of n=3 independent experiments, ± SEM, *: pb0.05).

compartmentalization/trafficking and ligand-induced receptor degra- into the caveolae, sequesters the receptors away from SMAD2/3. dation. Our results demonstrating that CD109 facilitates TGF-β receptor Furthermore, CD109 promotes TGF-β receptor degradation, which in degradation is consistent with CD109's ability to direct TGF-β receptor turn can lead to inhibition of signaling. CD109 may thus dampen TGF-β localization into the caveolae. Although TGF-β receptors have been responses in order to avoid a deleterious effect of excess TGF-β that shown to be degraded by both lysosomal and proteasomal machineries often leads to many human diseases such as tissue fibrosis and cancer following receptor ubiquitination [35,40], limited data are available metastasis. Because deregulation of CD109 expression may result in on factors controlling TGFBR1 proteasomal degradation. Our results aberrant TGF-β receptor internalization and degradation, CD109 may showing that CD109 enhancement of TGFBR1 degradation is blocked by represent a novel therapeutic target for treatment of diseases in which the proteasome inhibitor MG132 but not by the lysosome inhibitor TGF-β is known to play a pathophysiological role. chloroquine, suggest that CD109 may regulate TGFBR1 proteasomal degradation. 5. Conflict of interest An important finding in the present study is that CD109 inhibits SMAD3 phosphorylation in a caveolin-dependent manner. One poten- The authors have no competing financial interests in relation to the tial explanation is that CD109, by promoting TGF-β receptor endocytosis work described. 752 A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753

Fig. 8. Schematic model of the potential mechanism by which CD109 may regulate TGF-β receptor internalization and degradation. TGF-β receptors can internalize via the clathrin-coated pits or the caveolar pathway [13]. CD109 increases TGF-β binding to TGF-β receptors and promotes TGF-β receptor localization to the caveolae, thereby facilitating TGF-β receptor endocytosis via the caveolar pathway and/or enhancing proteasomal degradation of TGFBR1. These effects of CD109 ultimately lead to downregulation of TGF-β signaling.

Supplementary materials related to this article can be found online [11] J.-M. Zhang, M. Hashimoto, K. Kawai, Y. Murakumo, T. Sato, M. Ichihara, S. Nakamura, M. Takahashi, CD109 expression in squamous cell carcinoma of the at doi:10.1016/j.bbamcr.2011.01.028. uterine cervix, Pathol. Int. 55 (2005) 165–169. [12] T. Sjoblom, S. Jones, L.D. Wood, D.W. Parsons, J. Lin, T.D. Barber, D. Mandelker, R.J. Leary, J. Ptak, N. Silliman, S. Szabo, P. Buckhaults, C. Farrell, P. Meeh, S.D. Acknowledgments Markowitz, J. Willis, D. Dawson, J.K. Willson, A.F. Gazdar, J. Hartigan, L. Wu, C. Liu, G. Parmigiani, B.H. Park, K.E. Bachman, N. Papadopoulos, B. Vogelstein, K.W. Kinzler, V.E. Velculescu, The consensus coding sequences of human breast and We thank Hahn Soe-Lin for generation of HaCaT stable transfected colorectal cancers, Science 314 (2006) 268–274. cells, Erik Meulmeester, Peter ten Dijke and Daniel Bernard for critical [13] G.M. Di Guglielmo, C. Le Roy, A.F. Goodfellow, J.L. Wrana, Distinct endocytic pathways regulate TGF-β receptor signalling and turnover, Nat. Cell Biol. 5 (2003) review of the manuscript, P. Boukamp, J. Wrana, C. Hardin, S. Egan, T. 410–421. Imamura and Genzyme Corporation for reagents. This work was [14] S. Hayes, A. Chawla, S. Corvera, TGF-β receptor internalization into EEA1-enriched – supported by CIHR operating grant to A.P (FRN13732). early endosomes: role in signaling to SMAD2, J. Cell Biol. 158 (2002) 1239 1249. [15] C.E. Runyan, H.W. Schnaper, A.-C. Poncelet, The role of internalization in TGF-β1- induced SMAD2 association with SARA and SMAD2-dependent signaling in human mesangial cells, J. Biol. Chem. 280 (2005) 8300–8308. [16] F. Itoh, N. Divecha, L. Brocks, L. Oomen, H. Janssen, J. Calafat, S. Itoh, P. Dijke Pt, The References FYVE domain in SMAD anchor for receptor activation (SARA) is sufficient for localization of SARA in early endosomes and regulates TGF-β/SMAD signalling, [1] K.J. Gordon, G.C. Blobe, Role of TGF-β superfamily signaling pathways in human Cells 7 (2002) 321–331. disease, Biochim. Biophys. Acta: Mol. Basis Dis. 1782 (2008) 197–228. [17]H.Mitchell,A.Choudhury,R.E.Pagano,E.B.Leof,Ligand-dependentand [2] Y. Shi, J. Massagué, Mechanisms of TGF-β signaling from cell membrane to the -independent TGF-β receptor recycling regulated by clathrin-mediated endocy- nucleus, Cell 113 (2003) 685–700. tosis and Rab11, Mol. Biol. Cell 15 (2004) 4166–4178. [3] B. Schmierer, C.S. Hill, TGF-β-SMAD signal transduction: molecular specificity and [18] B. Razani, X.L. Zhang, M. Bitzer, G. von Gersdorff, E.P. Bottinger, M.P. Lisanti, functional flexibility, Nat. Rev. Mol. Cell Biol. 8 (2007) 970–982. Caveolin-1 regulates TGF-β/SMAD signaling through an interaction with the [4] K.C. Kirkbride, B.N. Ray, G.C. Blobe, Cell-surface co-receptors: emerging roles in TGF-β type I receptor, J. Biol. Chem. 276 (2001) 6727–6738. signaling and human disease, Trends Biochem. Sci. 30 (2005) 611–621. [19] B.Y.Y. Tam, D. Larouche, L. Germain, N.M. Hooper, A. Philip, Characterization of a [5] K.W. Finnson, B.Y. Tam, K. Liu, A. Marcoux, P. Lepage, S. Roy, A.A. Bizet, A. Philip, 150 kDa accessory receptor for TGF-β1 on keratinocytes: direct evidence for a GPI Identification of CD109 as part of the TGF-β receptor system in human anchor and ligand binding of the released form, J. Cell. Biochem. 83 (2001) keratinocytes, FASEB J. 20 (2006) 1525–1527. 494–507. [6] M. Lin, D. Sutherland, W. Horsfall, N. Totty, E. Yeo, R. Nayar, X. Wu, A. Schuh, Cell [20] J.C. Zwaagstra, M. El Alfy, M.D. O'Connor-McCourt, TGF-β1 Internalization: surface antigen CD109 is a novel member of the alpha 2 macroglobulin/C3, C4, C5 modulation by ligand interaction with TGF-β receptors types I and II and a family of thioester-containing proteins, Blood 99 (2002) 1683–1691. mechanism that is distinct from clathrin-mediated endocytosis, J. Biol. Chem. 276 [7] M. Hasegawa, S. Hagiwara, T. Sato, M. Jijiwa, Y. Murakumo, M. Maeda, S. Moritani, (2001) 27237–27245. S. Ichihara, M. Takahashi, CD109, a new marker for myoepithelial cells of [21] T. Ito, J.D. Williams, D.J. Fraser, A.O. Phillips, Hyaluronan regulates transforming mammary, salivary, and lacrimal glands and prostate basal cells, Pathol. Int. 57 growth factor-β1 receptor compartmentalization, J. Biol. Chem. 279 (2004) (2007) 245–250. 25326–25332. [8] S. Hagiwara, Y. Murakumo, T. Sato, T. Shigetomi, K. Mitsudo, I. Tohnai, M. Ueda, M. [22] A. Blanc, N. Tran-Khanh, D. Filion, M.D. Buschmann, Optimal processing method to Takahashi, Up-regulation of CD109 expression is associated with carcinogenesis obtain four-color confocal fluorescent images of the cytoskeleton and nucleus in of the squamous epithelium of the oral cavity, Cancer Sci. 99 (2008) 1916–1923. three-dimensional chondrocyte cultures, J. Histochem. Cytochem. 53 (2005) [9] M. Hashimoto, M. Ichihara, T. Watanabe, K. Kawai, K. Koshikawa, N. Yuasa, T. 1171–1175. Takahashi, Y. Yatabe, Y. Murakumo, J.M. Zhang, Y. Nimura, M. Takahashi, [23] L. Landmann, P. Marbet, Colocalization analysis yields superior results after image Expression of CD109 in human cancer, Oncogene 23 (2004) 3716–3720. restoration, Microsc. Res. Tech. 64 (2004) 103–112. [10] T. Sato, Y. Murakumo, S. Hagiwara, M. Jijiwa, C. Suzuki, Y. Yatabe, M. Takahashi, [24] J. Wharton, T. Meshulamy, G. Vallega, P. Pilch, Dissociation of High-level expression of CD109 is frequently detected in lung squamous cell expression and signaling from caveolin-1 expression, J. Biol. Chem. 280 (2005) carcinomas, Pathol. Int. 57 (2007) 719–724. 13483–13486. A.A. Bizet et al. / Biochimica et Biophysica Acta 1813 (2011) 742–753 753

[25] A. Guimond, T. Sulea, J.C. Zwaagstra, I. Ekiel, M.D. O'Connor-McCourt, Identifica- [33] S. McLean, G. Di Guglielmo, TβRIII directs clathrin-mediated endocytosis of TGF-β tion of a functional site on the type I TGF-[beta] receptor by mutational analysis of type I and II receptors, Biochem. J. (2010). its ectodomain, FEBS Lett. 513 (2002) 147–152. [34] R.A. Anders, S.L. Arline, J.J.E. Dore, E.B. Leof, Distinct endocytic responses of [26] V. Puri, R. Watanabe, R.D. Singh, M. Dominguez, J.C. Brown, C.L. Wheatley, D.L. heteromeric and homomeric transforming growth factor beta receptors, Mol. Biol. Marks, R.E. Pagano, Clathrin-dependent and -independent internalization of Cell 8 (1997) 2133–2143. plasma membrane sphingolipids initiates two Golgi targeting pathways, J. Cell [35] P. Kavsak, R.K. Rasmussen, C.G. Causing, S. Bonni, H. Zhu, G.H. Thomsen, J.L. Wrana, Biol. 154 (2001) 535–548. SMAD7 binds to Smurf2 to form an E3 ubiquitin ligase that targets the TGF-β [27] V. Luga, S. McLean, C. Le Roy, M.D. O'Connor-McCourt, J.L. Wrana, G. Di Guglielmo, receptor for degradation, Mol. Cell 6 (2000) 1365–1375. The extracellular domain of the TGFbeta type II receptor regulates membrane raft [36] J.J.E. Dore Jr., D. Yao, M. Edens, N. Garamszegi, E.L. Sholl, E.B. Leof, Mechanisms partitioning, Biochem. J. 421 (2009) 119–131. of transforming growth factor-β receptor endocytosis and intracellular sorting [28] M. Hasegawa, S. Moritani, Y. Murakumo, T. Sato, S. Hagiwara, C. Suzuki, S. Mii, M. differ between fibroblasts and epithelial cells, Mol. Biol. Cell 12 (2001) Jijiwa, A. Enomoto, N. Asai, S. Ichihara, M. Takahashi, CD109 expression in basal- 675–684. like breast carcinoma, Pathol. Int. 58 (2008) 288–294. [37] V.O. Rybin, X. Xu, M.P. Lisanti, S.F. Steinberg, Differential targeting of β-adrenergic [29] R.G. Parton, K. Simons, The multiple faces of caveolae, Nat. Rev. Mol. Cell Biol. receptor subtypes and adenylyl cyclase to cardiomyocyte caveolae, J. Biol. Chem. 8 (2007) 185–194. 275 (2000) 41447–41457. [30] A.L. Kiss, E. Botos, Endocytosis via caveolae: alternative pathway with distinct [38] S. Sigismund, T. Woelk, C. Puri, E. Maspero, C. Tacchetti, P. Transidico, P.P. Di Fiore, cellular compartments to avoid lysosomal degradation? J. Cell. Mol. Med. 13 S. Polo, From the cover: clathrin-independent endocytosis of ubiquitinated cargos, (2009) 1228–1237. Proc. Natl Acad. Sci. 102 (2005) 2760–2765. [31] W. Chen, K.C. Kirkbridge, T. How, C.D. Nelson, J. Mo, J.P. Frederick, Z.-F. Wang, R.J. [39] S. Vainio, S. Heino, J.-E. Månsson, P. Fredman, E. Kuismanen, O. Vaarala, E. Ikonen, Lefkowitz, G.C. Blobe, β-arrestin 2 mediates endocytosis of type III TGF-β receptor Dynamic association of human insulin receptor with lipid rafts in cells lacking and down-regulation of its signaling, Science 301 (2003) 1394–1397. caveolae, EMBO 3 (2002) 95–100. [32] N.Y. Lee, G.C. Blobe, The interaction of endoglin with β-arrestin2 regulates TGF-β- [40] M. Kowanetz, P. Lonn, M. Vanlandewijck, K. Kowanetz, C.-H. Heldin, A. Moustakas, mediated ERK activation and migration in endothelial cells, J. Biol. Chem. (2007)8 TGF-β induces SIK to negatively regulate type I receptor kinase signaling, J. Cell M700176200. Biol. 182 (2008) 655–662.