SRBC/cavin-3 is a caveolin adapter protein that regulates caveolae function
2009; Springer Nature; Volume: 28; Issue: 8 Linguagem: Inglês
10.1038/emboj.2009.46
ISSN1460-2075
AutoresKerrie‐Ann McMahon, Hubert K. Zajicek, Wei-Ping Li, Michael J. Peyton, John D. Minna, V. James Hernandez, Katherine Luby‐Phelps, Richard G.W. Anderson,
Tópico(s)RNA Research and Splicing
ResumoArticle5 March 2009free access SRBC/cavin-3 is a caveolin adapter protein that regulates caveolae function Kerrie-Ann McMahon Kerrie-Ann McMahon Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Hubert Zajicek Hubert Zajicek Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Wei-Ping Li Wei-Ping Li Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Michael J Peyton Michael J Peyton Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, USA Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author John D Minna John D Minna Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, USA Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author V James Hernandez V James Hernandez Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Katherine Luby-Phelps Katherine Luby-Phelps Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Richard G W Anderson Corresponding Author Richard G W Anderson Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Kerrie-Ann McMahon Kerrie-Ann McMahon Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Hubert Zajicek Hubert Zajicek Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Wei-Ping Li Wei-Ping Li Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Michael J Peyton Michael J Peyton Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, USA Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author John D Minna John D Minna Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, USA Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author V James Hernandez V James Hernandez Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Katherine Luby-Phelps Katherine Luby-Phelps Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Richard G W Anderson Corresponding Author Richard G W Anderson Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA Search for more papers by this author Author Information Kerrie-Ann McMahon1, Hubert Zajicek1, Wei-Ping Li1, Michael J Peyton2,3, John D Minna2,3, V James Hernandez1, Katherine Luby-Phelps1 and Richard G W Anderson 1 1Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas, TX, USA 2Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, USA 3Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, TX, USA *Corresponding author. Department of Cell Biology, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd, Dallas, TX 75390-9039, USA. Tel.: +1 214 648 2346; Fax: +1 214 648 7577; E-mail: [email protected] The EMBO Journal (2009)28:1001-1015https://doi.org/10.1038/emboj.2009.46 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Caveolae are a major membrane domain common to most cells. One of the defining features of this domain is the protein caveolin. The exact function of caveolin, however, is not clear. One possible function is to attract adapter molecules to caveolae in a manner similar to how clathrin attracts molecules to coated pits. Here, we characterize a candidate adapter molecule called SRBC. SRBC binds PKCδ and is a member of the STICK (substrates that interact with C-kinase) superfamily of PKC-binding proteins. We also show it co-immunoprecipitates with caveolin-1. A leucine zipper in SRBC is essential for both co-precipitation with caveolin and localization to caveolae. SRBC remains associated with caveolin when caveolae bud to form vesicles (cavicles) that travel on microtubules to different regions of the cell. In the absence of SRBC, intracellular cavicle traffic is markedly impaired. We conclude that SRBC (sdr-related gene product that binds to c-kinase) and two other family members [PTRF (Pol I and transcription release factor) and SDPR] function as caveolin adapter molecules that regulate caveolae function. Introduction Caveolae are the most abundant membrane domain on the cell surface devoted to endocytosis. They can occupy 18–40% of the plasma membrane surface (Chang et al, 1994) and engage in at least three different types of dynamic activities (Mundy et al, 2002; White and Anderson, 2005). Type 1 caveolae are able to pinch off the membrane to form endocytic vesicles called cavicles (Mundy et al, 2002) that can carry cargo such as SV40 virus to specialized endosomes called caveosomes (Pelkmans et al, 2001). Cavicles can also travel to the recycling endosome region of the cell (Mundy et al, 2002). Cavicle movement to these sites is microtubule dependent. Type 2 caveolae, which are the most abundant type, superficially appear not to be dynamic (Thomsen et al, 2002). Recent studies suggest, however, that quantal units of caveolae are engaged in continuous rounds of fission and fusion at the cell surface (Pelkmans and Zerial, 2005). Quantal caveolae behaviour exhibits the same kinetics and characteristics as folate receptor internalization and recycling by potocytosis (Anderson et al, 1992), which is a caveolae-dependent mechanism for delivering vitamins and ions such as calcium (Isshiki et al, 2002) to the cell interior. Type 2 cavicles can also move rapidly just beneath the membrane. The third type of caveolae behaviour (type 3) is the relatively uncharacterized ability of these domains to form long, thin caveolin-1 (Cav1)-positive tubules that project from the plasma membrane deep into the cell interior (Mundy et al, 2002). There are no known functions for these tubules, but GPI-anchored prions have been localized to these structures by immunogold electron microscopy (EM) (Peters et al, 2003). Type 3 caveolae may mediate Cav3-rich tubule formation during skeletal muscle cell development (Parton et al, 1997). All three types of caveolae behaviours can occur in the same cell. The three pathways depend on the ability of caveolae to form flask or tubular invaginations and, at least for types 1 and 2, seal off from the extracellular space. Several years ago we initiated a search for molecules that control caveolae internalization and identified PKCα as a regulator of type 2 caveolae internalization (Smart et al, 1995a). PKCα is highly enriched in caveolae (Smart et al, 1995a), and deletion of PKCα from cells or displacement from caveolae by AlF4 or inactivation with inhibitors (Mineo and Anderson, 2001) blocks uptake of folate receptors and flattens the flask-shaped morphology of caveolae. PKCα is probably a natural regulator of caveolae internalization because activation of histamine H1 receptors, which causes the dissociation of PKCα from caveolae, transiently blocks folate internalization (Smart et al, 1995a). Phosphorylation of an unidentified 90-kDa protein by PKCα has been linked to caveolae internalization. On the basis of these observations, we hypothesized that caveolae contain a protein(s) that recruits PKCα to this domain. Subsequently, we identified SDR (serum deprivation response) protein as a PKCα-binding protein that is enriched in caveolae (Mineo et al, 1998). PKCα binding to SDR in caveolae is dependent on amino acids 145–250 of SDR, calcium and an intact PKC regulatory domain. Addition of phosphatidylserine (PS) is not required. SDR (also called PS-p68 and SDPR) is a 412-amino-acid modular protein that contains a leucine zipper (LZ), a PKC-binding site, a PKC phosphorylation site, a PS-binding site and two PEST domains. These structural motifs are also found in PTRF [Pol I and transcription release factor (Jansa et al, 1998), also called, cavin/cav-p60 (Voldstedlund et al, 2003; Vinten et al, 2005)] and SRBC [sdr-related gene product that binds to c-kinase (Izumi et al, 1997)]. On the basis of these motifs, these three proteins form a subset of the STICK (substrates that interact with C-kinase) superfamily of PKC-binding proteins (Chapline et al, 1998; Jaken and Parker, 2000)). In addition to PKC-binding sites, all STICK proteins contain PS-binding sites and are found at the interface between membranes and the cytoskeleton. They include MARCKS, annexin I and II, desmoykin, vinculin/talin, α- and β-adducin, GAP43 and AKAP79. Recently, PTRF was reported to be concentrated in adipocyte (Vinten et al, 2005; Aboulaich et al, 2006; Hill et al, 2008; Liu and Pilch, 2008) and smooth muscle caveolae (Voldstedlund et al, 2003), even though it was originally identified and studied as a transcription release factor (Jansa et al, 1998, 2001; Jansa and Grummt, 1999; Hasegawa et al, 2000). PTRF has also been implicated in regulating caveolae function (Hill et al, 2008; Liu and Pilch, 2008). In contrast to SDPR and PTRF, a relationship between SRBC and caveolae function has not been established. SRBC was initially identified in overlay assays as a PKCδ-binding protein (Izumi et al, 1997). PKCδ binds and phosphorylates SRBC in vitro. Similar to SDPR, the mRNA for SRBC is induced in response to serum deprivation as well as during retinoic acid-induced differentiation of P19 cells. SRBC was also identified in a yeast two-hybrid screen as a BRCA1-interacting protein (Xu et al, 2001). The human SRBC gene maps to 11p15.5–15.4, which is a suppressor region that is mutated in a number of sporadic breast, lung and ovarian cancers. The loss of SRBC in these cancers appears to be caused by DNA methylation and subsequent gene silencing. A follow-up study (Zochbauer-Muller et al, 2005) suggests, however, that DNA methylation is not the only mechanism of SRBC loss in lung cancer cells. SRBC is also downregulated in gastric cancer (Lee et al, 2007). A naturally occurring fusion protein isolated from colon cancer cells, which consists of the first 184 amino acids of SRBC linked to c-Raf, has cell-transforming activity (Tahira et al, 1987). Interestingly, activation of c-Raf occurs in caveolae (Mineo et al, 1996) and Raf fused to the C-terminal consensus sequence for prenylation is constitutively active in this domain (Mineo et al, 1997). It is possible, therefore, that the SRBC–Raf fusion protein inappropriately targets c-Raf to caveolae where it is constitutively active. Here, we report that SRBC is highly localized to caveolae. Targeting of SRBC to caveolae depends on both the LZ domain of SRBC and the expression of Cav1. SRBC is downregulated in response to depletion of Cav1, suggesting that Cav1 stabilizes SRBC. Spinning disc confocal microscopy shows that SRBC is in caveolae, cavicles and caveosomes and traffics with Cav1 to different locations in the cell. In the absence of SRBC, however, intracellular cavicle traffic is markedly impaired. Results SRBC was identified in a proteomic screen for proteins enriched in detergent-free caveolae, purified from various cell types (Aboulaich et al, 2004; McMahon et al, 2006). Alternative splicing of the mRNA for SRBC can potentially produce five different transcripts. The domain organization of these five proteins is shown in Figure 1A. The isoform used in the current study is hSRBCβ, which codes for a 261-amino-acid-long protein designated hSRBC (Izumi et al, 1997). All five isoforms are predicted to be soluble proteins. Figure 1.hSRBC is localized to human fibroblast caveolae. (A) The domain structure of five predicted isoforms of human SRBC. Only hSRBCβ was used in this study. The region used to produce the mAb α-SRBC is indicated in brackets. (B–F) Immortalized human fibroblasts were fixed, permeabilized and processed for immunofluorescence to colocalize endogenous hSRBC and Cav1 using an mAb α-hSRBC IgG (B) and pAb α-Cav1 (C). The merged image (E) and the scatter plot (D) show that greater than 88% of α-SRBC and α-Cav1 IgG spots coincide. The colocalization channel (F) is a z-axis projection showing voxels with statistically significant colocalization. Bar=10 μm. (G) Normal human fibroblasts were processed for immunogold staining with mAb α-SRBC IgG. The majority of the gold particles was associated with invaginated caveolae at the cell surface or cavicles/caveosomes within the cell (arrow). Bar=0.2 μm. Download figure Download PowerPoint We used immunofluorescence and immunogold EM to verify that SRBC is enriched in caveolae. The SRBC mAb we used is directed against the C-terminal end of hSRBCα,β (Figure 1A, bracket, hSRBCβ) and recognizes only human SRBC. This antibody, which detects a single 45-kDa band on immunoblots, colocalizes with Cav1 (Figure 1B–F). The scatter plot (Figure 1D) shows the red and green intensities at each pixel. From these data, we calculated the Pearson's coefficient of correlation (PCC), which measures how well two fluorescent foci coincide (Costes et al, 2004), to be 0.88 where 1.0 represents perfect colocalization. Another measure of coincidences of the two signals is the colocalization channel (Figure 1F), which is a z-axis projection showing voxels that have a statistically significant colocalization. Finally, immunogold labelling (Figure 1G) showed a high degree of gold localization to invaginated caveolae as well as endosomal structures that could be either caveosomes or cavicles. Substitution of an irrelevant mAb for α-SRBC IgG did not give labelling (data not shown). Therefore, hSRBC is preferentially associated with caveolae as well as caveolae-derived endo-membranes. hSRBC is localized to caveolae in many different human tissue culture cell lines (data not shown). Tissue distribution of SRBC We used immunofluorescence to identify tissues that express both hSRBC and Cav1 and determine whether they were both in the same cell type. Samples of human skeletal muscle, liver, stomach, lung, kidney and heart were processed for indirect immunofluorescence (Figure 2). Both proteins were expressed in all the tissues we examined, but not in every cell type within an individual tissue. For example, blood vessels expressed both proteins (yellow in the merge) in skeletal muscle tissue, whereas muscle cells per se expressed Cav1 but not SRBC (red in merge). In stomach, connective tissue cells expressed both proteins but neither protein was expressed in epithelium. Importantly, even though many cells expressed Cav1 but not hSRBC (red in the merge images of skeletal, liver, lung and kidney), rarely did we find cells that expressed only SRBC. This raises the possibility that SRBC expression may be linked to Cav1 expression. Figure 2.hSRBC and Cav1 expression in tissue cells. Sections of paraffin-embedded normal human tissues were incubated overnight at 4°C in the presence of mAb α-hSRBC IgG and pAb α-Cav1 IgG. The sections were then washed and the primary antibody was detected by incubating the sections with the appropriate Alexa-Fluor IgG. Images were taken using a Leica TCS SP confocal microscope (Leica, Bannockburn, IL). Bar=100 μm. Download figure Download PowerPoint LZ is required for targeting to caveolae We used site-directed mutagenesis to identify regions of SRBC that might be necessary for targeting the protein to caveolae (Figure 3). There are at least four regions that could be involved in targeting (Figure 3A); a LZ (aa 22–70), two PEST domains (aa 140–173 and 225–241), a putative PS-binding site (aa 113–124) and a PKC-interacting region (aa 175–194). We constructed five cDNAs coding for wild type (WT) and mutant hSRBC tagged with HA. Each cDNA was expressed in human fibroblasts and the samples were processed to localize HA (green) and endogenous Cav1 (red). As expected, the WT HA–SRBC colocalized with Cav1 (Figure 3B, WT coloc) with a PCC of 0.74. Deletion of the LZ shifted hSRBC from caveolae to the cytoplasm and nucleus of the cell (Figure 3C), giving a PCC of –0.08, which indicates no specific association with Cav1. By contrast, deletion of the PS-binding site (Figure 3D) or the PKC-binding region (Figure 3E) or the PEST domain between amino acids 215–241 (Figure 3F) had little affect on colocalization (coloc) with caveolin. The PCC for these constructs was 0.63, 0.61 and 0.78, respectively. Figure 3.Leucine zipper required for hSRBC targeting to caveolae. (A) A diagram showing schematically the wild type and four different hSRBCs with deletions of regions that might be involved in targeting to caveolae. (B–F) The indicated HA-tagged versions of each construct were transiently expressed in immortalized human fibroblasts. Cells were fixed, permeabilized and processed for colocalization of Cav1 and HA using mAb α-HA and pAb α-Cav1. In (B, D–F), the fluorescence signal for the two antibodies was analysed to map voxels that overlapped between channels (coloc) as well as determine the Pearson's coefficient of correlation of the two images (PCC). Instead of using the colocalization channel for Δ22–70, we show the merge because of the diffuse distribution of the expressed protein. Bar=10 μm. Download figure Download PowerPoint To determine the general importance of LZ in targeting this family of proteins to caveolae, we studied SDPR (Figure 4A). Similar to hSRBC, SDPR colocalizes with Cav1 (Figure 4B; Supplementary Figure 1SA) and both hSRBC and SDPR colocalize with each other (Supplementary Figure 1SB). The LZ occupies the same relative position in SDPR (amino acids 52–100), PTRF (amino acids 50–98) and hSRBC (Figure 4A). First, we constructed cDNAs that code for amino acids 1–337 (WT), 1–168 and Δ52–100, all tagged with HA, and expressed them in human fibroblasts. The WT (Figure 4B, WT) and the 1–168 construct (Figure 4C, 1–168) localized to caveolae. The SDPR construct lacking the LZ domain, by contrast, was found in the cytoplasm and nucleus of the cell but did not colocalize with Cav1 (Figure 4D, 52–100, PCC=–0.236). Likewise an SDPR with leucines 86, 93 and 100 changed to glutamic acid did not colocalize with Cav1 (Supplementary Figure 1SC). These results suggest that the LZ has a critical function in directing SDPR, hSRBC and probably PTRF to caveolae and raises the possibility that other LZ-containing molecules spend time in caveolae. Dimerization of the LZ may be required for targeting. Figure 4.LZ in SDPR also required for targeting to caveolae. (A) A diagram showing schematically the wild type and two different deletions we made to determine the part of SDPR that was required for targeting to caveolae. (B–D) The indicated Myc-tagged versions of each construct were transiently expressed in immortalized human fibroblasts. Cells were fixed, permeabilized and processed for colocalization of endogenous Cav1 and Myc using mAb α-Myc and pAb α-Cav1. Instead of using the colocalization channel for Δ52–100, we show the merge because of the diffuse distribution of the SDPR protein. Bar=10 μm. Download figure Download PowerPoint Caveolin 1 or 3 is required for localization of hSRBC to the cell surface Cav1 is an obvious candidate for the protein that attracts hSRBC to caveolae. We used immunofluorescence to explore the relationship between these two molecules (Figure 5). Immortalized fibroblasts derived from Cav1−/− mice were transfected with HA–hSRBC alone (Figure 5A and B) and processed for indirect immunofluorescence detection of HA. HA–hSRBC was diffusely distributed in the cytoplasm (Figure 5B) of cells that lack Cav1 (Figure 5A). By contrast, when HA–hSRBC was co-transfected with Cav1–GFP, the two proteins colocalized (Figure 5C and D) and the distribution precisely matched the caveolae staining-pattern usually seen in fibroblasts. Other experiments showed that Cav3–Myc effectively substituted for Cav1 (Figure 5E and F), whereas hSRBC did not colocalize with Cav2–GFP that appeared to have a Golgi apparatus staining pattern (Figure 5G and H). A similar result was obtained with SDPR (data not shown). We conclude that all three members of this family depend on either Cav1 or Cav3 to reach the cell surface. Figure 5.Caveolin-dependent recruitment of hSRBC to caveolae. (A, B) Mouse Cav1−/− fibroblasts were transiently transfected with HA-tagged hSRBC for 24 h and processed for immunofluorescence colocalization of caveolin (A) and HA (B). (C, D) Mouse Cav1−/− fibroblasts were transiently co-transfected with cDNAs coding for Cav1–GFP (C) and HA-tagged hSRBC (D) for 24 h, fixed and processed to localize GFP (C) and HA (D). (E, F) Mouse Cav1−/− fibroblasts were transiently co-transfected with cDNAs coding for Myc–Cav3 and HA-tagged hSRBC for 24 h, fixed and processed for immunofluorescence to localize Myc (E) and HA (F). (G, H) Mouse Cav1−/− fibroblasts were transiently co-transfected with cDNAs coding for Cav2–GFP and HA-tagged hSRBC for 24 h, fixed and processed for localization of GFP (G) and HA (H). (I) The PNS fraction from immortalized human fibroblasts was solubilized with Triton X-100–octylglucosidase and incubated with beads conjugated with pAb α-rabbit IgG (control) or pAb α-caveolin-1 IgG. The beads were washed stringently before the bound proteins were eluted from the beads, separated by SDS–polyacrylamide gel electrophoresis and processed for immunoblotting using mAb α-hSRBC IgG (left panel). Another sample of PNS was processed in the same way substituting mAb α-SRBC for mAb α-Cav1 IgG and immunoblotting with mAb α-Cav1 IgG (right panel). (J) Immortalized human fibroblasts were incubated in the presence of siRNAs directed against an irrelevant RNA (lanes 2 and 4), Cav1 RNA (lane 1) and hSRBC RNA (lane 3). After 72 h, the postnuclear supernatant fractions were prepared and processed for immunoblotting using pAb α-Cav1 IgG, mAb α-hSRBC IgG or pAb α-cyclophilin-A IgG as a loading control. Bar=10 μm. Download figure Download PowerPoint A requirement for Cav1 in hSRBC targeting to caveolae suggests the two may physically and functionally interact. To test for a physical interaction, we used an immunoprecipitation protocol to see whether the two proteins would reciprocally co-precipitate from solubilized SV589 human fibroblasts. We solubilized the postnuclear supernatant fraction of these cells using a mixture of Triton X-100 and octylglucoside to ensure the complete solubilization of the caveolae, and then we incubated the soluble material with beads containing either α-Cav1 IgG (Figure 5I, left panel) or α-SRBC IgG (Figure 5I, right panel). The beads were washed before the bound proteins were processed for immunoblotting with either α-SRBC IgG (Figure 5I, IB, left panel) or α-Cav1 IgG (Figure 5I, IB, right panel). The immunoblots show that α-Cav1 immunoprecipitated hSRBC and α-SRBC immunoprecipitated Cav1. Therefore, under conditions when both proteins are in caveolae the two proteins are directly or indirectly interacting. In the tissue survey, we carried out to localize Cav1- and hSRBC-expressing cells (Figure 2), we rarely found Cav1-negative cells that expressed hSRBC. To explore the possible significance of this observation, we used small interfering RNA (siRNA) to reduce the level of Cav1 in fibroblasts (Figure 5J). When we depleted fibroblast Cav1 (lane 1), the amount of hSRBC in the cell markedly declined. A control siRNA did not affect the level of either hSRBC or Cav1 (lanes 2 and 4). By contrast, reducing hSRBC levels did not affect the amount of detectable Cav1 (lane 3). Therefore, the presence of hSRBC appears to be linked to the presence of Cav1, which may explain why we rarely found cells in tissues that expressed hSRBC alone. If the LZ is important for targeting hSRBC to caveolae, then the interaction of Cav1 with ΔLZ–SRBC may be impaired. Immortalized human fibroblasts were transfected with a cDNA coding for HA-tagged WT or ΔLZ–SRBC. The cells were then processed to immunoprecipitate HA and both the immunoprecipitate and the lysate proteins were separated on polyacrylamide gels (Figure 6A). A silver stain of the IP gel shows that the transfected cells express proteins with the molecular weight expected for WT and ΔLZ–SRBC (ΔLZ). The identity of these proteins was verified by immunoblotting (Figure 6B). We noticed in the silver-stained gel of the immunoprecipitates (A) that the WT lane contained high molecular weight (*) and a low molecular weight band (**) that were not present in the control lanes. Mass spectrometry identified the top protein as EPS15R. We confirmed that EPS15R co-precipitated with hSRBC by immunoblotting a duplicate gel with antibodies directed against SRBC, Cav1 and EPS15R (Figure 6B, lanes 2 and 3). Cav1 and EPS15R were not present in the immunoprecipitate of ΔLZ–SRBC. Lysates used for the immunoprecipitation contained similar amounts of the expressed (SRBC and ΔLZ–SRBC) as well as endogenous (Cav1 and EPS15R) protein (data not shown). Figure 6.Immunoprecipitation of hSRBC co-precipitates Cav1 and EPS15R. (A) The PNS fraction from immortalized human fibroblasts transfected with vectors containing no cDNA (lane 1) or the cDNA for HA-tagged WT hSRBC (lane 2) or the cDNA for HA-tagged Δ22–70 hSRBC was solubilized with a detergent mixture of Triton X-100–octylglucosidase. The solubilized material was incubated with beads conjugated with mAb α-HA IgG. The beads were washed before the bound proteins were processed for separation by polyacrylamide electrophoresis and silver stained. (B) The same beads were processed for immunoblotting using α-HA IgG (lanes 1–3) or α-Cav1 IgG (α-Cav1) or α-EPS15R IgG (α-EPS15R). Download figure Download PowerPoint Cav1 and hSRBC traffic together in cells Live cell imaging of Cav–GFP has shown that Cav1-rich membranes engage in three types of membrane traffic (types 1–3). To see whether hSRBC traffics with all three types of Cav1-rich membranes, we co-expressed in HeLa cells either hSRBC–tdTomato and Cav1–GFP (Figure 7A–C) or Cav1–tdTomato and hSRBC–GFP (Figure 7D–F) and recorded the behaviour of the two proteins with a spinning disc confocal microscope using excitation at 488 nm (GFP) and 568 nm (tdTomato) and the appropriate band-pass filters to separate the two fluorescence signals (Supplementary Movies S1, S2, S3). In agreement with our immunofluorescence data, hSRBC highly colocalized with caveolin (Figure 7A, yellow), both on the cell surface and on internal vesicles. Live cell imaging detected all three types of caveolin-rich membranes and in each case SRBC was present. Still images of types 1 and 2 caveolae movement are analysed in Figure 7. Type 1 hSRBC/Cav-positive cavicles moved at ∼0.8–2.03 μm/s for several micrometres as if travelling along microtubules (Figure 7B and C, still images; Supplementary Movie S1). Interestingly, cavicles displayed saltatory movement (Supplementary Movie S1). They paused 11–33 s (Figure 7A, arrowheads) between segments and often changed direction when movement resumed. Type 2 hSRBC/Cav1-positive cavicles were seen travelling (Supplementary Movie S2) beneath the plasma membrane at an average rate of 2.6 μm/s (Figure 7D–F, arrow and circles). This population of cavicles did not exhibit long pauses nor did they change direction. Type 3 tubular caveolae were also positive for both hSRBC and caveolin (Supplementary Movie S3). These results indicate that hSRBC is intimately associated with caveolae as well as caveolae-derived membranes such as cavicles and caveosomes, and they both remain associated with these membranes as they travel in the cell. Figure 7.hSRBC and Cav1 traffic together on type 1 and 2 cavicles. A colour composite and individual frames from a time-lapse sequence of HeLa cells co-expressing red SRBC–tdTomato and green Cav1–GFP (see Supplementary Movies S1, S2, S3). (A–C) The path followed by a single hSRBC–tdTomato/caveolin–GFP-positive type 1 cavicle is traced using Adobe Premiere. The trajectory of the cavicle was divided into four segments. At each of the positions indicated by the arrowheads, the cavicle stopped and hovered on average 19 s (33, 11 and 15 s) before moving on in a new direction. The mean velocity of the moving cavicle, averaged over the four segments, was 1.3 μm/s (1.22, 1.17, 0.8 and 2.03 μm/s). Selected frames from the time-lapse sequence, cropped to the area indicated by the white box in (A). (B) hSRBC–tdTomato; (C) caveolin–GFP. (D–F) The path followed by a single caveolin–tdTomato/hSRBC–GFP-positive type 2 cavicle is traced using Adobe Premiere. Unlike type 1 cavicles, these cavicles do not pause for significant times and have a trajectory that follows the contour of the membrane. The average speed was 2.6 μm/s. (E) Caveolin–tdTomato; (F) hSRBC–GFP. In each frame, the cavicle being tracked is
Referência(s)