H-Ras Resides on Clathrin-Independent ARF6 Vesicles That Harbor Little RAF-1, but Not on Clathrin-Dependent Endosomes

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1813 (2011) 298–307

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

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

H-ras resides on clathrin-independent ARF6 vesicles that harbor little RAF-1, but not on clathrin-dependent endosomes

Jodi McKay 1,2, Xing Wang 1,3, Jian Ding 4, Janice E. Buss 5, Linda Ambrosio ⁎

Department of Biochemistry, Biophysics and Molecular Biology, Iowa State University, Ames, IA 50011-3260, USA

article info abstract

Article history: Internalization of H-Ras from the cell surface onto endomembranes through vesicular endocytic pathways may Received 23 June 2010 play a significant role(s) in regulating the outcome of Ras signaling. However, the identity of Ras-associated Received in revised form 2 November 2010 subcellular vesicles and the means by which Ras localize to these internal sites remain elusive. In this study, we Accepted 29 November 2010 show that H-Ras is absent from endosomes initially derived from a clathrin-dependent endocytic pathway. Available online 8 December 2010 Instead, both oncogenic H-Ras-61L and wild type H-Ras (basal or EGF-stimulated) bind Arf6-associated clathrin- independent endosomes and vesicles of the endosomal-recycling center (ERC). K-Ras4B-12V can also be Keywords: fi Ras internalized via Arf6 endosomes, and the C-terminal tails of both H-Ras and K-Ras4B are suf cient to mediate Endocytosis localization of GFP chimeras to Arf6-associated vesicles. Interestingly, little Raf-1 was found on these Arf6- Arf6 associated endosomes even when active H-Ras was present. Instead, endogenous Raf-1 distributed primarily on Protein trafficking EEA1-containing vesicles, suggesting that this H-Ras effector, although accessible for H-Ras interaction on the Endosomes plasma membrane, appears to separate from its regulator during early stages of endocytosis. The discrete and Clathrin dynamic distribution of Ras pathway components with spatio-temporal complexity may contribute to the Raf-1 specificity of Ras:effector interaction. Endosomal-recycling center © 2010 Elsevier B.V. All rights reserved.

1. Introduction identifiable by the presence of the small GTPase Arf6 that participates in vesicular trafficking [14,15]. Although the site of Ras signaling is widely depicted as the plasma The internalization pathways of the H-Ras and K-Ras4B proteins membrane, endocytosis of Ras pathway components is required to appear to differ. In neurons, calcium signals trigger Ca++/calmodulin achieve efficient and specific spatio-temporal activation of downstream binding to K-Ras4B and cause the release of the K-Ras/calmodulin cascades [1–3]. Ras proteins have now been shown to be distributed on, complex from the plasma membrane [16]. This transient, cytosolic and to produce significant signals from, internal membranes [4–6].The complex is subsequently targeted to Golgi or recycling endosomes means by which Ras proteins localize to these internal sites are varied [17]. Internalization of K-Ras can also be mediated through vesicular and complicated [7]. Some of these routes may depend on release of Ras endocytic pathways. Lu et al. [18] found K-Ras4B associated with from the plasma membrane, while others likely require transfer of EEA1/Rab7-bearing early endosomes, as well as membranes of late membrane tethered Ras between membranous compartments via endosomes, lysosomes and multivesicular bodies in COS-1 cells. vesicular budding or fusion mechanisms. The classic vesicular internal- For newly synthesized H-Ras, a small portion of the protein is ization route is via clathrin-mediated endocytosis. Formation of delivered to Golgi membranes, and from there, these molecules board clathrin-dependent traffic machineries requires involvement of clathrin outward-bound exocytic vesicles [19,20]. However, the majority of coat and adapter proteins, which are the key components of this newly synthesized H-Ras appears to initially traffic to the plasma pathway. Although not as well characterized as clathrin-mediated membrane via a non-conventional pathway that may be non- endocytosis, several pathways for clathrin-independent endosomal vesicular in nature [21]. Additional studies indicate that palmitoylated entry of proteins have been described [8–13]. One of these is a pathway H-Ras that has already reached the plasma membrane can be depalmitoylated and released into the cytosol. The farnesylated but non- acylated protein can then be re-palmitoylated, and thus captured on ⁎ Corresponding author. Tel./fax: +1 515 294 0453. E-mail address: [email protected] (L. Ambrosio). internal membranes, to join with those H-Ras molecules traveling to 1 These authors contributed equally. the plasma membrane via the biosynthetic vesicular pathway [20,22]. 2 Current address: Department of Biology and Chemistry, Morningside College, Sioux City, Like K-Ras, there is evidence that H-Ras can leave the plasma IA 51104, USA. membrane and enter vesicular endocytic pathways. H-Ras has been 3 Current address: Fermentas Life Sciences, Dengliang Road, Shenzhen 518054, China. repeatedly found on internal vesicles with all the characteristics of 4 Current address: Children's Hospital Boston, Harvard Medical School, Boston, MA 02115, USA. endosomes [23–25]. A previous study has shown that both Rab5 and 5 Current address: TechLab, 2001 Kraft Drive, Blacksburg, VA 24060, USA. Rab11 regulate translocation of H-Ras to the pericentriolar recycling

center [26]. However, Rab5 functions on endosomal membranes and NIH 3T3 cells were cultured at 37 °C and 10% CO2 in Dulbecco's sorting endosomes that are derived from both clathrin-mediated and modified Eagle's medium supplemented with 10% calf serum plus clathrin-independent pathways [11,26], thus, its presence does not penicillin, streptomycin, glutamine and sodium pyruvate. For immu- define a specific class of endosomes. nofluorescence and immuno-isolation, NIH 3T3 cells were transfected Several clues to the identity of H-Ras endocytic pathways have with Effectene (Qiagen) and used after 24 hours or as mentioned in been derived from the identification of components from H-Ras the results. Stable lines contained H-Ras-61L or H-Ras-WT (pZIPneo associated vesicles. One study demonstrated that H-Ras appears to be vector), or GFP-H-Ras-61L cDNA (pEGFP-C3 vector). Cells treated with present on endosomes containing epidermal growth factor receptor EGF (Gibco BRL) were cultured as described above but incubated with (EGFR) [24]. Vesicles captured by magnetic beads coated with EGF 5 or 10 ng/ml of EGF at 37 °C for the time as indicated, followed by contained EGFR along with endogenous Ras isoforms. These endo- washing with phosphate-buffered saline and fixation. some preparations also contained clathrin protein [27], suggesting cargo, including Ras, on these captured vesicles may be derived from a 2.2. Antibodies clathrin-dependent pathway. However, EGFR can be internalized via both clathrin-dependent and independent mechanisms. Thus, the Primary antibodies used for immunofluorescence included rat identity and types of H-Ras associated vesicles remained undefined. monoclonal anti-Ras sc34 (Santa Cruz Biotechnology); the mouse Recently, Porat-Shliom et al. [28] has reported that endosomes in a monoclonals anti-Ras 610010 (BD Bioscience; Fig. 4) and anti-Ras clathrin-independent pathway associated with Arf6 contained H-Ras. LA069 (Quality Biotech); and the rabbit polyclonals anti-c-Raf-1 sc227 They proposed an elegant model for sequential production of and anti-clathrin heavy chain sc9069 (Santa Cruz Biotechnology), phosphoinositides on these endosomes accompanied by signaling of anti-EEA1P1-063(Affinity BioReagents), anti-hemagluttinin H-Ras to the phosphatidyl inositol 3-kinase (PI3-K) effector pathway. PRB101P (Convance) and anti-giantin PRB114C (Convance). Primary However, this study did not, in a systematic way, rule out the possibility antibodies used for immuno-isolation included mouse monoclonal that H-Ras can also enter cells via clathrin-mediated endocytosis. anti-Ras LA069 (Quality Biotech), rat monoclonal anti-Ras sc34 (Santa Ras effector proteins can be internalized by clathrin-dependent Cruz Biotechnology), and rabbit polyclonals anti-EEA1 P1-063 and clathrin-independent means. Effector components that are (Affinity BioReagents), and anti-hemagluttinin (HA) PRB101P (Con- involved in PI3-K signaling, as well as downstream components in vance). Primary antibodies used for immunoblotting included mouse the Raf effector branch of the Ras pathway, including KSR1 and ERK, monoclonals anti-Ras 610010 and anti-c-Raf 61051 (BD Biosciences), have been found on clathrin-independent endosomes [28,29]. anti-GFP 1814460 (Roche), and rabbit polyclonals anti-hemagluttinin However, it is not clear if Raf-1 is also present on these vesicles, as PRB101P (Convance) and anti-EEA1 P1-063 (Affinity BioReagents). Raf-1 has been observed on early endosomes, co-localizing with the Secondary antibodies from Molecular Probes included goat anti- EEA1 marker [25]. It has also been demonstrated that expression of mouse IgG coupled to Alexa Fluor 350, goat anti-rat IgG coupled to the constitutively active Rab5-Q79L protein causes some H-Ras, but, Alexa Fluor 488, and the Alexa Fluor 594 coupled goat anti-mouse IgG, notably, not Raf-1, to accumulate on enlarged endosomes [30]. Thus, goat anti-rabbit IgG, and donkey anti-rat IgG. there is evidence that Raf-1 disengages from H-Ras associated endosomes, and it has become important to test whether or not H- 2.3. Live cell labeling with transferrin antibody Ras and Raf might be internalized together on newly generated vesicles produced via clathrin-independent means. To mark early endosomes of the clathrin-dependent pathway, Here, we characterize the endocytic itinerary of Ras, and cells were transfected and incubated for 18 hours, then treated with demonstrate that H-Ras is absent from clathrin-dependent endo- cycloheximide (10 μg/ml) for 5.25 hours, to allow time for previously somes and in particular, from EEA1-associated endosomes that harbor synthesized protein to vacate Golgi membranes. Cells were then chilled Raf-1 proteins. Instead, we find both H-Ras and K-Ras4B bind to Arf6- for 45 minutes at 4 °C in medium containing 50 mM HEPES, pH 7.4 associated vesicles and can be internalized by an Arf6 mediated, buffer, cycloheximide (10 μg/ml), and AlexaFluor 594 transferrin clathrin-independent pathway. Although active H-Ras on these Arf6- T-13343 (Molecular Probes, 10 μg/ml). Finally, the cold medium was associated vesicles appears to signal to PI3-K, as previously suggested removed and pre-warmed (37 °C) medium containing 50 mM HEPES, by Porat-Shliom et al. [28], we found that these vesicles bound very pH 7.4 was added and cells were incubated at 37 °C in a water bath for 1 little Raf-1. Our studies suggest that co-localization with or disen- minute before fixation with cold 4% paraformaldehyde and processed gagement from different effectors at discrete subcellular locations for immunofluorescence. To mark the endosomal recycling center may provide a means for spatio-temporal regulation of Ras signaling (ERC), cells were transfected with various cDNAs, incubated for 18 in cells. hours, and then treated for 5.25 hours with cycloheximide (10 μg/ml). AlexaFluor-transferrin T-13343 (Molecular Probes, 10 μg/ml) was 2. Materials and methods then added for 45 minutes. Cells were washed, fixed and processed for immunofluorescence. 2.1. Plasmids, cell culture and DNA transfection 2.4. Immuno-isolation of EEA1 and Arf6 associated membranes The H-Ras-WT, H-RasQ61L, and ExtRas cDNA plasmids in the pcDNA3 vector were described previously [31]. Plasmids GFP-H-Ras- Twenty-four hours after transfection, cells were scraped into WT, GFP-H-Ras-61L, GFP-K-Ras4B-12V, GFP-ExtRas-WT, and GFP- 300 μl of homogenization buffer (3 mM imidazole, pH 7.4, 1 mM

ExtRas-61L were generated by PCR ampliﬁcation and cloned into the Na3VO4, 5 mM MgCl2, 1% Protease Inhibitor Cocktail (Sigma), and 5% pEGFP-C3 vector (Clontech Labs). The GFP-H-tail plasmid was sucrose) and homogenized with 15 strokes in a Dounce homogenizer constructed by attaching the bases encoding the 10 C-terminal followed by 20 passages through a 26-gauge needle. The nuclei and amino acids of H-Ras (GCMSCKCVLS), a termination codon, and a unbroken cells were removed by centrifugation at 1300g (4 °C) for Bam H1 restriction enzyme site to pEGFP-C3 by PCR. We obtained the 10 minutes. This post-nuclear supernatant (PNS) was pre-cleared by following plasmids from Dr. J.G. Donaldson, National Institutes of incubation for 1 hour at 4 °C with protein A (for rabbit anti-EEA1 Health: Arf6-WT-GFP, Arf6-Q67L-GFP and GFP-K-tail (with the last 20 antibody) or G (for mouse anti-HA to capture Arf6-Q67L-HA) agarose amino acids of K-Ras4B fused to the C-terminus of EGFP) in the beads (Invitrogen). The supernatant was incubated with rabbit EEA1 pEGFP-N1 vector; and Arf6-WT-HA, Arf6-Q67L-HA (with a C-terminal or mouse anti-HA antibody overnight for immunoprecipitation (IP). hemagluttinin (HA) epitope tag) in the pXS vector. The following day fresh protein A or G beads were added for 2 hours at 300 J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307

4 °C. The beads were washed 3 times in cold homogenization buffer 3. Results and boiled in SDS-PAGE sample buffer. Proteins associated with EEA1 or Arf6 coupled membranes were separated by SDS-PAGE and 3.1. H-Ras does not bind endosomes of the clathrin-dependent pathway detected on immunoblots with the indicated antibodies. Ninety-five percent of the sample was loaded in the IP lane, and 5% in the PNS Although numerous studies have demonstrated that H-Ras is lane. For control blot (Fig. 2E) supernant was incubated with rat anti- present on vesicles that harbor specific endosomal markers, the H-Ras antibody. endocytic itinerary of H-Ras has not been systemically characterized, and the identity of vesicles that bind and internalize H-Ras, as well as its effector Raf-1, remains uncertain. 2.5. Immunoblotting Several reports have suggested that H-Ras might be internalized on endosomes of the dynamin-sensitive, clathrin-dependent pathway Samples collected by immuno-isolation were separated by SDS- [24,27]. To test this idea, we used multiple methods to assess the PAGE. Proteins were transferred to polyvinylidene difluoride (PVDF) possible association of H-Ras with endosomes generated by the membranes (Pall Life Science). The PVDF membranes were incubated clathrin-dependent pathway using NIH3T3 cells. In this pathway, in blocking buffer (Vector Labs 10× casein diluted to 1× in TTBS, clathrin coat proteins are required for formation and budding of (0.15 mol/L NaCl, 0.05 mol/L Tris–HCl, 0.05% Tween-20, pH 7.4) for 20 endosomal vesicles from the plasma membrane, and detection of minutes at room temperature, then washed in TTBS and incubated these coat proteins has been used as an early marker of clathrin- with primary antibodies overnight at 4 °C. The proteins of interest dependent endocytosis. In order to inspect specifically only plasma were detected using horseradish peroxidase-conjugated goat anti- membranes for possible H-Ras and clathrin co-localization, adherent mouse or goat anti-rabbit secondary antibodies and an enhanced basal plasma membranes of transfected NIH 3T3 cells were prepared chemiluminescence kit from Pierce. by sonication, fixed, and examined by indirect immunofluorescence [32]. At least 30 membrane sheets from cells transfected with either 2.6. Immunofluorescence and preparation of sonicated membrane the activated H-Ras-61L or H-Ras-WT were assayed. However, neither fragments H-Ras-61L (Fig. 1A) nor H-Ras-WT (data not shown) overlapped with clathrin heavy chain. NIH 3T3 cells were plated on coverslips coated with 100 μg/ml At a later step of the clathrin-dependent pathway, endosomes lose poly-L-lysine (Sigma) and 500 μg/ml fibronectin (Sigma) and trans- their clathrin coats, but can bind the early endosome antigen-1 protein fected with DNA. After 24 hours (or as mentioned in results), cells (EEA1) [18]. Next, we tested whether H-Ras was present on these EEA1- were fixed with fresh 4% formaldehyde in phosphate-buffered saline, associated vesicles. On adherent fragments of basal membranes (N30 pH 7.4, at room temperature for 1 hour then permeabilized with membrane sheets) from sonicated cells, the population of EEA1-positive −20 °C methanol and quenched with 50 mM ammonium chloride in vesicles was clearly resolved, but again H-Ras (H-Ras-61L) was not phosphate-buffered saline. The fixed cells were incubated with present on these vesicles (Fig. 1B). Furthermore, upon examination of blocking buffer (2% horse serum and 0.4% bovine serum albumin in EEA1-positive vesicles in at least 30 intact cells, no overlap of H-Ras-WT phosphate-buffered saline, pH 7.4) then sequentially with primary and EEA1 was detected (Fig. 1C). These results indicated that neither an and secondary antibodies diluted in blocking buffer. For the oncogenic nor the normal form of H-Ras were present on early experiment shown in Fig. 3D, cells were co-transfected with cDNAs endosomes derived from the clathrin-dependent pathway. for H-Ras-61L and Arf6-Q67L-GFP. Cells were fixed, and stained with We also tested if H-Ras-WT that became transiently GTP-bound mouse anti-Ras (BD Bioscience 610010) and rabbit anti-Raf (Santa after EGF stimulation might now join the clathrin-dependent pathway Cruz Biotechnology sc227) antibodies. The Ras antibody was detected along with the EGF receptor [24]. This was a relevant question, as with a goat anti-mouse IgG secondary antibody coupled to Alexa- the serum-activated form of cellular H-Ras is reported to shift from Fluor350 (blue), and its co-expression with Arf6 was visually raft-like lipid microdomains, to a more general membrane distribu- confirmed, but blue images were not captured. To produce adherent tion from which clathrin-mediated internalization might occur [33]. plasma membrane fragments, experiments were carried out as However, after stimulation of cells for 10 minutes with a moderate, described previously [32]. 10 ng/ml dose of EGF [34], no co-alignment of the EEA1 marker of Processed coverslips were mounted in Vectashield mounting early endosomes and H-Ras-WT was detected (Fig. 1D) for N30 cells medium (Vector Laboratories). All experimental findings reported examined. here are based on observations of more than 30 NIH3T3 cells (or 30 As an additional test, a fluorescent derivative of transferrin was adherent membrane fragments) that showed co-expression of Ras, added to live cells to enable the first minutes of clathrin-mediated Arf6, Raf-1 or markers described in the text. For experiments imaging endocytosis to be tracked. Transferrin protein can localize to internal adherent membrane fragments, cells were analyzed using a 63× oil vesicular compartments via the clathrin-dependent route [35]. For these immersion lens on a Nikon Eclipse E800 microscope with Spot RT experiments, a 6-hour cycloheximide pre-treatment was included to digital camera. All confocal images were taken with a Leica TCS-NT inhibit synthesis of protein and thereby to avoid imaging newly confocal system with a 63× oil immersion lens with an aperture synthesized H-Ras on Golgi-derived vesicles that might mingle with adjustable from 1.32 to 0.6, using an argon laser (488 nm excitation) the endocytic vesicles. Cells were then chilled and incubated with and a krypton laser (568 nm excitation). Other images were taken transferrin, so as to label, but prevent internalization of, transferrin with Leica DMIRE2 microscope using a 100× oil immersion lens with receptors. Cells were warmed very briefly (1 minute) to allow an aperture adjustable from 1.3 to 0.6 and Q Imaging Retiga 1300 internalization to re-start. Although H-Ras-positive vesicles were camera. Images were exported to Adobe Photoshop 7.0 or, for the present in the re-warmed cells, the transferrin-containing early images taken with the Leica DMIRE2 microscope; images were taken endosomes did not overlap with GFP-H-Ras-61L (Fig. 1E). Thus, for all in black and white, processed in Openlab3.5.1, displayed in colors and the cells we examined (N30), although H-Ras was present on internal then exported to Adobe Photoshop 7.0 to equalize image intensity vesicles, it was notably absent from endosomes associated with these prior to merging. For some experiments, as indicated in the Figure characteristic steps of the clathrin-dependent endocytic route. legends, a series of z-stack images were acquired using a motorized To supplement the imaging results presented above, we used a stage on the Leica DMIRE2 microscope. These images were processed biochemical approach to examine whether H-Ras in its GTP-bound using Openlab's nearest neighbor deconvolution module and pro- form might be present on EEA1-containing endosomes. An antibody to cessed as above. EEA1 was used to capture detergent-free membrane fragments and J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307 301

3.2. Endosomes associated with Arf6 bind H-Ras

A prominent class of non-clathrin plasma membrane structures in fibroblasts and endothelial cells is comprised of caveolae. However, the ability of caveolae to internalize (as “caveosomes”) is generally thought to be restricted to special situations, such as viral infection [37,38]. H-Ras, although present on membranes with a low buoyant density similar to that of caveolae, is largely absent from these structures [32,33]. An additional type of clathrin-independent endosome is the type that arises from the Arf6-associated constitutive pathway by which the endogenous major histocompatibility complex I (MHC-I) is internalized [39]. Previously, it was reported using HeLa, MCF7, HT1080 and COS-7 cells that H-Ras can bind these endosomes and be internalized via this clathrin-independent pathway [28]. To determine if H-Ras is internalized via this pathway in our NIH3T3 cells, we co-transfected these cells with H-Ras-WT and a GFP-tagged version of the cellular form of Arf6 (Arf6-WT-GFP). H-Ras-WT was observed on the plasma membrane, as expected, and also on both small endosomes and larger, apparently fluid-filled, macropinosomes that were also populated by Arf6-WT-GFP (N30 co-transfected cells assayed, Fig. 2A–A′). At high resolution (in a separate cell), the Arf6 and H-Ras proteins appeared to inhabit adjacent but distinct domains of many of these larger vesicles, producing a beaded appearance in cross-section (Fig. 2B). For these resting NIH3T3 cells, WT H-Ras proteins were predominantly in a GDP-bound state. Next, we tested the subcellular localization of a constitutively active, GTP-bound H-Ras (H-Ras-61L) and found that H-Ras-61L was also present on Arf6-associated endosomes after its transient expression (N30 cells assayed, data not shown). These results indicated that endosomes containing Arf6 could bind both wild type and oncogenic H-Ras, regardless of its activation state in NIH3T3 cells, and are similar to those found by Porat-Shliom et al. [28] using human and African green monkey cell lines. Together, these findings suggest that Fig. 1. H-Ras is absent from endocytic vesicles of the clathrin-dependent pathway in Arf6-containing endosomes may be a type of traffic machinery broadly NIH 3T3 cells. Adherent membrane fragments from sonicated cells (A–B) allow utilized for H-Ras internalization in mammalian cells. – detection of nascent and newly emerged endosomes, while intact cells (C E) provide Furthermore, we co-expressed the constitutively active, GTP- views of internalized vesicles. (A) A non-overlapping H-Ras-61L (green) and clathrin heavy chain (red) localization pattern is apparent in this merged image of a plasma bound form of Arf6 (Arf6-Q67L) with H-Ras-61L and assessed their membrane fragment. (B) The accumulation of EEA1 (red) is also distinct from that of localization in cells. The inability of Arf6-Q67L to hydrolyze its bound H-Ras-61L (green) on this membrane fragment and marks a later step in clathrin- GTP prevents nascent endosomes from further progress, leading to an dependent vesicular maturation. (C) Discrete internalized vesicles are associated with endocytic cul-de-sac and the accumulation of cargo on multiple, H-Ras-WT (green) or EEA1 (red) in quiescent cells and (D) in those stimulated with enlarged macropinosomes [28,40]. In cells that co-expressed Arf6- EGF. (E) GFP-H-Ras-61L (green) and AlexaFluor-transferrin (red) are associated with distinct vesicles in this merged image. Transferrin receptors internalize via clathrin- Q67L-GFP and H-Ras-61L, the distinctive groups of these large, Arf6- coated vesicles. Vertical z-stack images of each color were deconvolved before merging. Q67L-GFP endosomal vesicles were brightly outlined by the H-Ras- Scale bars=5 μm. (F) H-Ras-61L was not detected on EEA1 associated vesicles. 61L protein (N30 cells assayed, Fig. 2C–C"). Membranes from post-nuclear supernatant (PNS) of NIH 3T3 cells stably expressing These H-Ras containing vesicles were further characterized using a H-Ras-61L were captured with EEA1 antibody (IP). Proteins were separated by SDS- PAGE and identified by immunoblotting with indicated antibodies. 95% of the sample biochemical approach. Cells were transfected with a hemagluttinin was loaded in the IP lane, and 5% in the PNS lane. (HA)-tagged Arf6-Q67L along with GFP-H-Ras-61L so that fragments of Arf6-containing membranes could be immuno-isolated using HA antibody. On these isolated, detergent-free fragments of Arf6-HA- containing membranes, GFP-H-Ras-61L was present (Fig. 2D). A reciprocal experiment, using H-Ras-directed antibody capture, showed vesicles by immuno-isolation [25] and the presence of H-Ras and that Arf6-Q67L-GFP could also be detected on H-Ras-containing endogenous Raf-1 on EEA1-containing endosomes was assessed. In membrane fragments (Fig. 2E). In addition, a control experiment, in agreement with the previous imaging results, neither cellular H-Ras, which anti-HA antibody-directed membrane isolation was performed nor activated H-Ras-61L (Fig. 1F) could be detected on the captured using cells that expressed only GFP-H-Ras, but had no Arf6-HA to EEA1-containing membranes. Thus, by multiple techniques, we found capture, showed that no GFP-H-Ras was present in the HA-antibody- no evidence of H-Ras on newly generated endocytic vesicles that were captured sample (Fig. 2F). This indicated that the presence of H-Ras in derived from the clathrin-dependent pathway. In spite of this absence, the sample required Arf6, and was not the result of incomplete endogenous Raf-1 was easily observed in these EEA1 antibody- membrane washing or non-specific binding of H-Ras to the HA antibody. captured endosomes (Fig. 1F). This ability to capture a protein that did associate with EEA1-containing membranes verified that the 3.3. Very little Raf-1 binds to H-Ras-61L on Arf6 endosomes membrane isolation protocol was successful, and validated the absence of H-Ras on those membranes. This result also confirmed Previous studies demonstrated that clathrin-independent vesicles previous work that Raf-1 could be found on clathrin- and EEA1- harbor GTP-bound H-Ras [28]. Consistent with this finding, we also containing early endosomes [18,25,36]. All these studies indicated that detected active GTP-bound H-Ras on Arf6-associated endosomes endosomes to which H-Ras was bound had not entered the cells via a upon EGF stimulation (data not shown). This raised the question clathrin-mediated pathway. whether Raf-1, which can bind GTP-bound H-Ras, might also be 302 J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307

Fig. 2. H-Ras is present on Arf6-associated vesicles in NIH 3T3 cells. Arf6 is often associated with small endosomes and larger, ﬂuid-ﬁlled macropinosomes [52]. Confocal images of a cell expressing (A) H-Ras-WT (red) and (A′) Arf6-WT-GFP (green) proteins showing co-localization on small and large Arf6-WT-GFP associated vesicles. (B) Enlarged and merged images of another cell showing the beaded appearance (arrows) of endosomal membranes with H-Ras-WT (red) and Arf6-WT-GFP (green) proteins bound. (C–C″) Confocal images of a cell transfected with cDNAs encoding activated H-Ras-61L (red) and activated Arf6-Q67L-GFP (green) proteins. The merged image (C″) indicates Arf6-Q67L-GFP-associated vesicles harbor H-Ras-61L proteins. Scale bars=5 μm. (D) Membrane fragments of NIH3T3 cells expressing GFP-H-Ras-61L and Arf6-Q67L-HA were prepared from post-nuclear supernatant (PNS), and captured with anti-HA antibody (IP). Both Arf6-HA and GFP-H-Ras were present on these membranes and detected by immunoblotting using anti-HA (for Arf6-HA) and anti-GFP (for GFP-H-Ras) antibodies. Asterisk (*) shows position of the HA-antibody heavy chain used for the immunoprecipitation. (E) NIH 3T3 cells stably expressing H-Ras-61L were transfected with cDNA encoding Arf6-Q67L-GFP protein, membranes were prepared from the post-nuclear supernatant (PNS), and captured with anti-Ras antibody (IP). Both H-Ras-61L and Arf6-Q67L-GFP proteins were detected using anti-Ras and anti-GFP antibodies, respectively. (C) Control experiment with anti-HA antibody isolation performed on membrane fragments from the post-nuclear supernatant (PNS) of NIH 3T3 cells stably expressing GFP-H-Ras-61L proteins, without Arf6-Q67L-HA transfection (IP). Although GFP-H-Ras-61L could be detected in the PNS, it was absent for the HA captured sample. Asterisk (*) shows the position of the HA-antibody heavy chain used for the immunoprecipitation.

present on Arf6-associated vesicles. Because Arf6 is present both on the early stage of Arf6-Q67L-stimulated endocytosis, these Arf6 the plasma membrane (where Raf-1 is known to associate with active containing vesicles bound very little Raf-1, despite harboring H-Ras) as well as on endosomal membranes, immuno-isolation from a oncogenic H-Ras-61L proteins. mixture of plasma and endosomal membrane fragments could not provide this answer; immunofluorescence was used so that endo- 3.4. H-Ras is present in the perinuclear endosomal recycling center somes could be examined specifically. As observed in studies by Porat- Shliom et al. [28] and this study (see Fig. 2C"), co-transfection of cells Several studies suggest that Arf6-associated vesicles converge on with cDNAs for Arf6-Q67L and H-Ras-61L results in enlarged easily the endosomal recycling center (ERC) [39,41], where they mingle identifiable endosomally derived vesicles. Next, we determined the among other endosomes that are derived, independently, from location of Raf-1 in such double-transfected NIH3T3 cells (N30 cell clathrin-mediated endocytosis [35,42]. These Arf6-associated vesicles assayed). Consistently, the characteristic Arf6-Q67L vesicles were can also participate in recycling to return proteins to the plasma brightly rimmed with GFP-H-Ras-61L (Fig. 3A and B). However, the membrane [43,44]. Thus, the potential for an additional contribution Raf-1 antibody did not illuminate those clustered endosomes of endosomal pathways, and in particular the perinuclear endosomal (Fig. 3A′–B), and for this typical example, in only a few spots did recycling center (ERC), to H-Ras localization was investigated. endogenous Raf-1 and H-Ras-61L co-localize on endosomes. As Newly synthesized H-Ras has been found on Golgi membranes expected, some endogenous Raf-1 did co-localize with H-Ras-61L at [19]. In addition, other studies have demonstrated that H-Ras also the plasma membrane (Fig. 3C). A second experiment was performed persists in the perinuclear area despite clearing the H-Ras biosyn- to directly examine whether smaller endosomes that displayed thetic pathway through exposure of cells to cycloheximide [20,22,26]. transfected Arf6-Q67L-GFP and H-Ras-61L could bind Raf-1. Cells In order to specifically distinguish the ERC from Golgi membranes, that expressed exogenous H-Ras-61L and Arf6-Q67L-GFP on small transferrin, a protein known to enter the ERC area (via the clathrin- endosomes were identified (N30 cells assayed, see Materials and mediated route), was used. In cells pre-treated with cycloheximide for methods). The staining for Raf-1 in these cells showed that very few of 6 hours, live-cell uptake of AlexaFluor-transferrin for 45 minutes at the dots of endogenous Raf-1 aligned with the small Arf6-Q67L-GFP 37 °C produced a strong perinuclear accumulation of transferrin in the endosomes, or with the clustered, larger, Arf6-Q67L-stimulated ERC, along with numerous peripheral transferrin-laden vesicles endosomal structures (Fig. 3D). This further suggested that during traversing the cytoplasm. In these cycloheximide-treated cells (N30 J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307 303

Fig. 3. Very little Raf-1 is present on Arf6-containing endosomes. (A–C) NIH 3T3 cells were transfected with cDNAs for GFP-H-Ras-61L (green) and Arf6-Q67L-HA (unstained). Cells were ﬁxed and stained with anti-Raf-1 antibody and AlexaFluor-labeled secondary antibody (red). Arrows in (A) indicate enlarged, Arf6-Q67L stimulated endosomal vesicles that contain GFP-H-Ras-61L and chevrons in (A′) indicate the absence of endogenous Raf-1 on those vesicles. (B) Merged and enlarged boxed area of (A–A') with arrows indicating enlarged endosomal vesicles that contain GFP-H-Ras-61L but lack Raf-1 proteins. (C) Merged images of H-Ras-61L (green) and endogenous Raf-1 (red) in a NIH 3T3 cell, co-localization of H-Ras-61L and Raf-1 on plasma membrane is indicted (arrow). (D) Merged images of Arf6-Q67L-GFP (green) and endogenous Raf-1 (red) in a NIH 3T3 cell that was co-transfected with H-Ras-61L and Arf6-Q67L-GFP cDNAs. Arrow indicates an enlarged endosomal vesicle with Arf6-Q67L-GFP that lacks Raf-1; chevron indicates Raf-1 on a separate structure that lacks Arf6. Scale bars=10 μm. Vertical z-stack images of each color were deconvolved and merged where indicated.

cells assayed), the previously synthesized H-Ras-61L that was located above, we asked whether a GFP chimera having only the last twenty on internal membranes co-localized entirely with the perinuclear C-terminal residues of K-Ras4B, GFP-K-tail, could mirror the ability of concentration of AlexaFluor-transferrin (Fig. 4A–A′), and not with a the activated full-length K-Ras4B protein to bind Arf6-Q67L vesicles. marker of the cis-Golgi, giantin (Fig. 4B–C). Transfected H-Ras-WT As they did with full length K-Ras protein, the Arf6-Q67L macropino- also co-aligned well with the transferrin-marked ERC (N30 cells somes bound this GFP-K-tail protein (N30 cells assayed, Fig. 5B). In a assayed, data not shown). This indicated that the internal membranes parallel experiment, enlarged Arf6-Q67L associated vesicles also populated by previously synthesized H-Ras were predominantly bound a second chimeric version of GFP, GFP-H-tail, with the last those of the ERC, and that, with this protocol, very little H-Ras was ten residues of H-Ras (N30 cells assayed, Fig. 5C). located or captured on Golgi membranes. This data complement the Normal anchoring of K-Ras4B in the plasma membrane by its report that H-Ras is associated with Rab11-positive vesicles in the C-terminal tail involves a strongly polybasic sequence that includes (pericentriolar) recycling center of CHO-K1 cells [26]. Notably, in six lysine residues, and an immediately adjoining farnesylated agreement with our findings described above for transfected cells, at terminal cysteine [17]. In contrast, H-Ras membrane targeting the periphery of cells that had not been transfected with Arf6 the depends on two neighboring cysteines that acquire palmitate groups previously synthesized GFP-H-Ras-61L was present in small vesicles together with the farnesylated terminal cysteine [50]. To learn more that were separate from those that contained AlexaFluor-transferrin about the properties of these C-terminal residues in Arf6 vesicular (N30 cells assayed, Fig. 4D–E). Together, these results indicated that, targeting, a chimeric H-Ras protein with C-terminal characteristics of in addition to Arf6-associated endosomal membranes and mem- K-Ras4B (lysine residues) and H-Ras (palmitoylated residues) was branes of the Golgi, H-Ras could distribute in NIH3T3 cells to a third examined. This protein (ExtRas) associates with the plasma mem- location and type of internal membrane—the membranes of the ERC. brane and becomes palmitoylated, but fails to undergo C-terminal farnesylation and does not interact with ER or Golgi membranes 3.5. K-Ras4B can bind to Arf6 vesicles [31,46]. Interestingly, ExtRas differed from K-Ras4B and H-Ras, and was not detected on Arf6-associated macropinosomes (N30 cells The route of the outward, biosynthetic trafficking of the non- assayed, Fig. 5D–D′). This was an important control because, although palmitoylated K-Ras4B protein is unknown, but appears to differ from endocytosis via Arf6-associated vesicles can occur continuously, Ras the vesicular mechanism used (occasionally) by H-Ras [21].We protein binding to these endosomes was neither universal nor therefore examined if these two Ras proteins also differed in their inevitable. presence on Arf6-associated vesicles. In N30 transfected cells assayed, activated GFP-K-Ras4B-12V protein could be clearly seen on the 4. Discussion plasma membrane (Fig. 5A) and on Arf6-Q67L-enlarged macropinosomes (Fig. 5A′). This result indicated that Arf6-Q67L-associated Ras small GTPases can reside on and signal from both the plasma vesicles could also bind activated K-Ras4B, and that this form of Ras membrane and numerous types of endomembranes. Studies that seek was not excluded from these membranes. to reveal the identity of Ras-associated subcellular compartments, as The C-terminal tails of Ras proteins can guide their association well as Ras trafficking pathways can provide important insight to with cellular membranes. GFP chimeras of these tails have been understand the role(s) of spatio-temporal regulation on Ras signaling. readily used as probes to track the subcellular location of Ras in A previous study by Porat-Shliom et al. [28] elegantly showed that H- previous studies [17,47–49]. To further test our observation described Ras bound clathrin-independent, Arf6-associated vesicles in HeLa, 304 J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307

Fig. 4. H-Ras is present in the perinuclear endosomal-recycling center. (A–A′) Confocal images of NIH 3T3 cells transfected with H-Ras-61L cDNA, treated with cycloheximide for 6 hours, exposed to AlexaFluor-transferrin (red) for 45 minutes, then fixed and stained with H-Ras antibodies (green). Arrow (A) indicates perinuclear accumulation of H-Ras-61L and arrowhead (A′) indicates transferrin. (B–C) Merged image of a cell transfected with H-Ras-61L cDNA, treated with cycloheximide for 6 hours, fixed, and stained with antibodies for H-Ras (green) and giantin (red) with boxed area of (B) enlarged in (C). Arrow denotes accumulated H-Ras-61L on perinuclear membranes that do not contain giantin (arrowhead). (D) A cell treated as in (A–A') with boxed area enlarged in (E) showing separate vesicles containing transferrin (red; arrowhead) and H-Ras (green, arrow). Fig. 5. K-Ras4B binds Arf6-associated vesicles. A confocal image of a NIH 3T3 cell showing (A) GFP-K-Ras4B-12V (green; arrow) and (A′) Arf6-Q67L-HA (red; arrowhead) proteins on enlarged macropinosomes. (B–C) C-terminal residues of K-Ras4B or HT1080, MCF7 and COS-7 cells. The possibility that H-Ras could also H-Ras can target GFP to enlarged Arf6 macropinosomes. (B) Localization of GFP-K-tail (green) on Arf6-Q67L-HA (red) associated vesicles (arrow) is shown in a merged enter clathrin-dependent pathways was investigated using the image. (C) Merged images of a NIH 3T3 cell expressing GFP-H-tail (green) and Arf6- transferrin receptor as a marker. Here, we characterized the endocytic Q67L-HA (red) proteins with arrow indicating co-localization. (D–D') Confocal images itinerary of Ras in NIH3T3 cells using multiple, additional markers and indicate that (D) ExtRasWT (red) does not accumulate on (D') Arf6-WT-GFP enlarged approaches to fully explore this possibility and found that from the endosomal vesicles (green; arrowheads). Chevrons indicate absence of ExtRas on Arf6- WT-coated vesicles. Vertical z-stack images of each color were deconvolved, and then very earliest stages of endocytosis, H-Ras was absent from clathrin- merged. Scale bars=10 μm. dependent endosomes. Our study also provides evidence that in NIH3T3 fibroblasts Arf6- containing, clathrin-independent endosomes binds normal, growth factor stimulated, and oncogenic forms of H-Ras. These vesicles are defined [8,9,52–55]. Another member of the Arf GTPase family, Arf1, has separate from, and in addition to, exocytic vesicles of the H-Ras recently been found to regulate a dynamin- and clathrin-independent biosynthetic pathway. Thus, an Arf6-associated means of H-Ras endocytic pathway [53]. However, this path is distinct from the Arf6 endocytosis has been defined in cell lines derived from at least three pathway, and does not appear to contain H-Ras [52]. It will now be different mammals and is likely a general means for H-Ras internali- important to learn if H-Ras also utilizes other clathrin-independent zation in higher organisms. Together, our results and those of Porat- endosomes, or only those associated with Arf6. Shliom et al. [28] show that Arf6 endosomes provide a second, Arf6 endosomes bind both inactive and active forms of H-Ras, and membrane-based mechanism for binding H-Ras to internal membranes thus provide a new platform for H-Ras signaling within the cell. The that is distinct from the deacylation-controlled cycle that shuttles H-Ras nucleotide-independence of H-Ras endosome binding is consistent with between the cell surface, cytosol and internal membranes [5,20,22,51]. previous reports and with the constitutive nature of Arf6-mediated Several other clathrin-independent endocytic routes are rapidly being endocytosis [25,56–58]. Additionally, although Arf6 endocytosis occurs J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307 305 continuously, it can also be regulated by other factors including the TBC amino acid residues important for targeting H-Ras and K-Ras4B to (Tre-2/Bub2/Cdc16) domain protein TRE17, the Arf6 nucleotide Arf6 vesicles remains a goal for future investigation. exchange factor ARNO/cytohesin and the EFA6 family of Arf6 guanine nucleotide exchange factors (GEFs) [56,59,60]. Internal localization of 4.2. The role of endosomes in H-Ras signaling H-Ras via deacylation/re-acylation also operates for both GDP- and GTP- bound forms [20,22]. Interestingly, palmitate turnover is accelerated The separation of H-Ras from the clathrin-dependent pathway when H-Ras is in the GTP-bound form [32], suggesting that the acylation has special importance for interactions between H-Ras and its effector, cycle of relocation may also be subject to control. While the plasma Raf-1. Both in this work and in studies with insulin-stimulated [25], membrane is widely viewed as a static final destination for Ras proteins, angiotensin-stimulated [63], or EGF-stimulated cells [18],activatedRaf-1 this report adds support to the idea that the amount of Ras at the plasma has been found primarily on EEA1-containing endosomes. The presence of membrane (not just its activation state) may be dynamic and Raf-1 on endosomes derived in part from the clathrin-dependent pathway continuously adjustable, by both deacylation and endocytosis. is also consistent with the report that H-Ras-stimulated Raf-1 signaling is This study also confirms the presence of H-Ras in the perinuclear decreased by down-regulation of clathrin-dependent endocytosis by a recycling center as presented by Gomez and Daniotti using CHO-K1 cells dominant negative dynamin mutant [30]. Together, these reports suggest [26], and reveals an additional challenge for studies that wish to that at least a portion of H-Ras-triggered, Raf-1 signals may occur on measure H-Ras signaling from internal sites. Since the endocytic classical early endosomes, without the presence of H-Ras. What prevents recycling center is often embedded within or near the Golgi, it may be Raf-1 from retaining its association with H-Ras as it leaves the plasma necessary to distinguish H-Ras on ERC membranes from newly membrane? This potential disengagement of Raf-1 from activated H-Ras is synthesized or re-acylated H-Ras on Golgi membranes [22]. The also consistent with the findings that H-Ras:Raf interactions at the plasma methods developed in this study may be useful for such studies. membrane are actually quite brief [64]. Future studies may help to link Interestingly, we also found that K-Ras4B could bind Arf6-associated endosomal trafficking, H-Ras:effector interaction and the signaling output endosomes. In addition, Lu et al. [18] demonstrated that some K-Ras4B of isoform specific Ras mediated pathways. Since it appears that the H-Ras co-localizes with clathrin- and EEA1/Rab5-bearing endosomes in EGF- and K-Ras4B proteins can utilize physically distinct endocytic routes, the stimulated COS-1 cells. Thus, it seems K-Ras4B can be internalized K-Ras4B that co-localizes with Raf-1 on endomembranes may have through both clathrin-dependent and independent mechanisms, and its preferential access to this effector pathway. signaling may be regulated not only qualitatively but also quantitatively In contrast to the abundance of Raf-1 on classical early endosomes, by endocytosis. It will be interesting to characterize how much K-Ras4B very little Raf-1 was found on Arf6-associated vesicles that contained is capable of entering cells via either of these pathways. activated H-Ras-61L. Such a limitation in a major Ras effector pathway Since H-Ras on Arf6-associated vesicles can be transported to the suggests that the spectrum of signals arising from endosomal H-Ras endosomal recycling center, we also tested whether K-Ras4B could may differ from that of H-Ras on the plasma membrane. There is localize onto this perinuclear compartment. When the ERC was already precedent for selective association of nucleotide exchange marked by loading NIH3T3 cells with AlexaFluor-transferrin (without factors (GEFs), effectors, and GTPase activating proteins (GAPs) with Arf6-Q67L), activated K-Ras4B was not observed in the ERC (N30 cells H-Ras that is present on Golgi membranes [4,7,65]. Furthermore, Arf6- assayed, data not shown). Thus, in NIH3T3 cells, K-Ras4B differed associated endosomes bind other signaling proteins, such as Src and from H-Ras and did not bind either ERC or Golgi membranes. Gomez Rac [35,59], which could influence or contribute to H-Ras signals that and Daniotti [26] reported a similar result for K-Ras using CHO-K1 might arise from endosomes. cells. However, in neurons stimulated with glutamate or the calcium ionophore, ionomycin, both full-length K-Ras4B and a YFP-K-Ras-tail 4.3. Reciprocal interaction between Ras and Arf6 construct have been observed on ERC and Golgi membranes [17,45]. Together, these results indicate that it will be important to further Ras signaling pathways have previously been linked functionally study possible cell- and stimulus-specific variations in the internal- with Arf6. Several studies show that Arf6 can activate members of the ization of Ras proteins. Ras signaling pathway, the ERK and Akt protein kinases [28,29,66]. Thus, a dynamic view of Ras proteins emerge with their movement Reciprocally, H-Ras stimulates phosphatidylinositol 3-kinase activity, to the plasma membrane from biosynthetic pathways, release and which leads to membrane recruitment and activation of the Arf6 return via constitutive endosomal pathways and in the case of acylated nucleotide exchange factor, ARNO [39]. Inhibition of the extracellular Ras proteins (H-Ras, N-Ras and K-Ras-4A) membrane association signal-regulated kinases (ERKs) that are downstream of Ras proteins, dependent on the acylation cycle. Upon activation by growth factors, a specifically impedes Arf6-mediated, but not clathrin-mediated, new layer of dynamic complexity emerges with Ras interactions with endocytosis [29]. Arf6 and H-Ras collaborate in vesicular transport, effector molecules at the plasma membrane and those that occur or are through co-stimulation of phospholipase D activity needed for vesicle excluded from endocytic membranes. budding [67]. Both Arf6 and H-Ras proteins are also central to the pinocytosis and membrane ruffling that are caused by growth factors. 4.1. Properties of Ras proteins that bind to Arf6-associated vesicles The identification of a specific class of endosomes to which H-Ras can bind in this and the study by Porat-Shliom et al. together with the Our results show that the seven, C-terminal residues of mature H- previous findings that H-Ras is found on recycling endosomes [26,28], Ras, with the accompanying farnesyl, methyl, and palmitoyl groups, completes a framework for a new cycle by which membrane-bound H- are sufficient to induce endosome binding, just as they do for its Ras traffics from plasma membrane, through internal sites, back to the association with Golgi membranes [19,22]. Similarly, the C-terminal plasma membrane. These reports provide further pieces of the physical tail of K-Ras4B can mediate localization to Arf6-associated endo- foundation that will allow the reciprocal interactions of Arf6 and H-Ras, somes. Since the GFP-H-tail protein resides in low-density (“raft”) and the impact of those interactions on signaling, to be unraveled. membrane microdomains [61], its internalization probably takes place from these domains. The GFP-K-tail protein and the full-length Acknowledgments KRas4B-G12V protein, as well as GTP-bound forms of H-Ras, all of which reside outside of the raft-like domains [47], were also observed We thank Janina Brandt, Joy Walker and Jon Coloff for wonderful on Arf6-Q67L endosomal vesicles. These results are consistent with technical assistance, and Dr. Julie Donaldson for Arf6 cDNAs. This reports that Arf6-associated endosomes can internalize proteins from work was supported by the National Science Foundation, Grant No. both types of domains [35,62]. The identification of key C-terminal 0110114 (J. E. Buss). 306 J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307

References [33] I.A. Prior, A. Harding, J. Yan, J. Sluimer, R.G. Parton, J.F. Hancock, GTP-dependent segregation of H-ras from lipid rafts is required for biological activity, Nat. Cell [1] O. Kranenburg, I. Verlaan, W.H. Moolenaar, Dynamin is required for the activation Biol. 3 (2001) 368–375. of mitogen-activated protein (MAP) kinase by MAP kinase kinase, J. Biol. Chem. [34] C. Puri, D. Tosoni, R. Comai, A. Rabellino, D. Segat, F. Caneva, P. Luzzi, P.P. Di Fiore, 274 (1999) 35301–35304. C. Tacchetti, Relationships between EGFR signaling-competent and endocytosis- [2] Y. Shen, L. Xu, D.A. Foster, Role for phospholipase D in receptor-mediated competent membrane microdomains, Mol. Biol. Cell 16 (2005) 2704–2718. endocytosis, Mol. Cell Biol. 21 (2001) 595–602. [35] N. Naslavsky, R. Weigert, J.G. Donaldson, Characterization of a nonclathrin [3] B.T. Andresen, M.A. Rizzo, K. Shome, G. Romero, The role of phosphatidic acid in endocytic pathway: membrane cargo and lipid requirements, Mol. Biol. Cell 15 the regulation of the Ras/MEK/Erk signaling cascade, FEBS Lett. 531 (2002) 65–68. (2004) 3542–3552. [4] T.G. Bivona, M.R. Philips, Ras pathway signaling on endomembranes, Curr. Opin. [36] Y. Daaka, L.M. Luttrell, S. Ahn, G.J. Della Rocca, S.S.G. Ferguson, M.G. Caron, R.J. Cell Biol. 15 (2003) 136–142. Lefkowitz, Essential role for G protein-coupled receptor endocytosis in the [5] V.K. Chiu, T. Bivona, A. Hach, J.B. Sajous, J. Silletti, H. Wiener, R.L.J. II, A.D. Cox, M.R. activation of mitogen-activated protein kinase, J. Biol. Chem. 273 (1998) Philips, Ras signalling on the endoplasmic reticulum and the Golgi, Nat. Cell Biol. 4 685–688. (2002) 343–350. [37] M. Kirkham, R.G. Parton, Clathrin-independent endocytosis: new insights into [6] I.P. de Castro, T.G. Bivona, M.R. Philips, A. Pellicer, Ras activation in Jurkat T cells caveolae and non-caveolar lipid raft carriers, Biochim. Biophys. Acta 1746 (2005) following low-grade stimulation of the T-cell receptor is specific to N-Ras and 350–363. occurs only on the Golgi apparatus, Mol. Cell. Biol. 24 (2004) 3485–3496. [38] L. Pelkmans, D. Puntener, A. Helenius, Local actin polymerization and dynamin [7] U. Ashery, O. Yizhar, B. Rotblat, Y. Kloog, Nonconventional trafficking of Ras recruitment in SV40-induced internalization of caveolae, Science 296 (2002) associated with Ras signal organization, Traffic 7 (2006) 1119–1126. 535–539. [8] O.O. Glebov, N.A. Bright, B.J. Nichols, Flotillin-1 defines a clathrin-independent [39] N. Naslavsky, R. Weigert, J.G. Donaldson, Convergence of non-clathrin- and endocytic pathway in mammalian cells, Nat. Cell Biol. 8 (2006) 46–54. clathrin-derived endosomes involves Arf6 inactivation and changes in phosphoi- [9] M. Kirkham, A. Fujita, R. Chadda, S.J. Nixon, T.V. Kurzchalia, D.K. Sharma, R.E. Pagano, nositides, Mol. Biol. Cell 14 (2003) 417–431. J.F. Hancock, S. Mayor, R.G. Parton, Ultrastructural identification of uncoated [40] H. Radhakrishna, J.G. Donaldson, ADP-ribosylation factor 6 regulates a novel caveolin-independent early endocytic vehicles, J. Cell Biol. 168 (2005) 465–476. plasma membrane recycling pathway, J. Cell Biol. 139 (1997) 49–61. [10] C. Lamaze, A. Dujeancourt, T. Baba, C.B. Lo, A. Benmerah, A. Dautry-Varsat, [41] N. Naslavsky, M. Boehm, P.S. Backlund Jr., S. Caplan, Rabenosyn-5 and EHD1 Interleukin 2 receptors and detergent-resistant membrane domains define a interact and sequentially regulate protein recycling to the plasma membrane, clathrin-independent endocytic pathway, Mol. Cell 7 (2001) 661–671. Mol. Biol. Cell 15 (2004) 2410–2422. [11] C. LeRoy, J.L. Wrana, Clathrin- and nonclathrin-mediated endocytic regulation of [42] D.K. Sharma, A. Choudhury, R.D. Singh, C.L. Wheatley, D.L. Marks, R.E. Pagano, cell signalling, Nat. Rev. Mol. Cell Biol. 6 (2005) 112–126. Glycosphingolipids internalized via caveolar-related endocytosis rapidly merge [12] I.R. Nabi, P.U. Le, Caveolae/raft-dependent endocytosis, J. Cell Biol. 161 (2003) with the clathrin pathway in early endosomes and form microdomains for 673–677. recycling, J. Biol. Chem. 278 (2003) 7564–7572. [13] B.J. Nichols, J. Lippincott-Schwartz, Endocytosis without clathrin coats, Trends Cell [43] C. D'Souza-Schorey, E. van Donselaar, V.W. Hsu, C. Yang, P.D. Stahl, P.J. Peters, Biol. 11 (2001) 406–412. ARF6 targets recycling vesicles to the plasma membrane: insights from an [14] C. D'Souza-Schorey, P. Chavrier, ARF proteins: roles in membrane traffic and ultrastructural investigation, J. Cell Biol. 140 (1998) 603–615. beyond, Nat. Rev. Mol. Cell Biol. 7 (2006) 347–358. [44] M. Prigent, T. Dubois, G. Raposo, V. Derrien, D. Tenza, C. Rosse, J. Camonis, P. [15] J.G. Donaldson, C.L. Jackson, Regulators and effectors of the ARF GTPases, Curr. Chavrier, ARF6 controls post-endocytic recycling through its downstream exocyst Opin. Cell Biol. 12 (2000) 475–482. complex effector, J. Cell Biol. 163 (2003) 1111–1121. [16] R.S. Sidhu, R.R. Clough, R.P. Bhullar, Ca2+/calmodulin binds and dissociates K- [45] C. Lopez-Alcala, B. Alvarez-Moya, P. Villalonga, M. Calvo, O. Bachs, N. Agell, RasB from membrane, Biochem. Biophys. Res. Commun. 304 (2003) 655–660. Identification of essential interacting elements in K-Ras/Calmodulin binding and [17] M. Fivaz, T. Meyer, Reversible intracellular translocation of KRas but not HRas in its role in K-Ras localization, J. Biol. Chem. 283 (2008) 10621–10631. hippocampal neurons regulated by Ca2+/calmodulin, J. Cell Biol. 170 (2005) [46] C. Figueroa, J. Taylor, A.B. Vojtek, Prenylated Rab acceptor protein is a receptor for 429–441. prenylated small GTPases, J. Biol. Chem. 276 (2001) 28219–28225. [18] A. Lu, F. Tebar, B. Alvarez-Moya, C. Lopez-Alcala, M. Calvo, C. Enrich, N. Agell, T. [47] I.A. Prior, C. Muncke, R.G. Parton, J.F. Hancock, Direct visualization of Ras Nakamura, M. Matsuda, O. Bachs, A clathrin-dependent pathway leads to KRas proteins in spatially distinct cell surface microdomains, J. Cell Biol. 160 (2003) signaling on late endosomes en route to lysosomes, J. Cell Biol. 184 (2009) 165–170. 863–879. [48] E. Choy, V.K. Chiu, J. Silletti, M. Feoktistov, T. Morimoto, D. Michaelson, I.E. Ivanov, [19] E. Choy, V.K. Chiu, J. Silletti, M. Feoktistov, T. Morimoto, D. Michaelson, I.E. Ivanov, M.R. Philips, Endomembrane trafficking of ras: the CAAX motif targets proteins to M.R. Philips, Endomembrane trafficking of Ras: the CAAX motif targets proteins to the ER and Golgi, Cell. 98 (1999) 69–80. the ER and Golgi, Cell 98 (1999) 69–80. [49] E. Choy, M. Philips, Green fluorescent protein-tagged Ras proteins for intracellular [20] J.S. Goodwin, K.R. Drake, C. Rogers, L. Wright, J. Lippincott-Schwartz, M.R. Philips, localization, Methods Enzymol. 332 (2001) 50–64. A.K. Kenworthy, Depalmitoylated Ras traffics to and from the Golgi complex via a [50] J.F. Hancock, A.I. Magee, J.E. Childs, C.J. Marshall, All ras proteins are poly- nonvesicular pathway, J. Cell Biol. 170 (2005) 261–272. isoprenylated but only some are palmitoylated, Cell 57 (1989) 1167–1177. [21] H. Zheng, J. McKay, J.E. Buss, H-Ras does not need COP I- or COP II-dependent [51] O. Rocks, M. Gerauer, N. Vartak, S. Koch, Z.P. Huang, M. Pechlivanis, J. Kuhlmann, L. vesicular transport to reach the plasma membrane, J. Biol. Chem. 282 (2007) Brunsveld, A. Chandra, B. Ellinger, H. Waldmann, P.I. Bastiaens, The palmitoylation 25760–25768. machinery is a spatially organizing system for peripheral membrane proteins, Cell [22] O. Rocks, A. Peyker, M. Kahms, P.J. Verveer, C. Koerner, M. Lumbierres, J. 141 (2010) 458–471. Kuhlmann, H. Waldmann, A. Wittinghofer, P.I.H. Bastiaens, An acylation cycle [52] M. Kalia, S. Kumari, R. Chadda, M.M. Hill, R.G. Parton, S. Mayor, Arf6-independent regulates localization and activity of palmitoylated Ras isoforms, Science 307 GPI-anchored protein-enriched early endosomal compartments fuse with sorting (2005) 1746–1752. endosomes via a Rab5/Phosphatidylinositol-3'-Kinase-dependent machinery, [23] D. Bar-Sagi, J. Feramisco, Induction of membrane ruffling and fluid-phase Mol. Biol. Cell 17 (2006) 3689–3704. pinocytosis in quiescent fibroblasts by Ras proteins, Science 233 (1986) [53] S. Kumari, S. Mayor, Arf1 is directly involved in dynamin-independent 1061–1068. endocytosis, Nat. Cell Biol. 10 (2008) 30–41. [24] X. Jiang, A. Sorkin, Coordinated traffic of Grb2 and Ras during epidermal growth factor [54] S. Mayor, R.E. Pagano, Pathways of clathrin-independent endocytosis, Nat. Rev. receptor endocytosis visualized in living cells, Mol. Biol. Cell 13 (2002) 1522–1535. Mol. Cell Biol. 8 (2007) 603–612. [25] M.A. Rizzo, C.A. Kraft, S.C. Watkins, E.S. Levitan, G. Romero, Agonist-dependent [55] S. Sabharanjak, P. Sharma, R.G. Parton, S. Mayor, GPI-anchored proteins are traffic of raft-associated Ras and Raf-1 is required for activation of the MAPK delivered to recycling endosomes via a distinct cdc42-regulated, clathrin- cascade, J. Biol. Chem. 276 (2001) 34928–34933. independent pinocytic pathway, Dev. Cell 2 (2002) 411–423. [26] G.A. Gomez, J.L. Daniotti, H-Ras dynamically interacts with recycling endosomes [56] A.M. Powelka, J. Sun, J. Li, M. Gao, L.M. Shaw, A. Sonnenberg, V.W. Hsu, in CHO-K1 Cells: involvement of Rab5 and Rab11 in the trafficking of H-Ras to Stimulation-dependent recycling of integrin beta1 regulated by ARF6 and this pericentriolar endocytic compartment, J. Biol. Chem. 280 (2005) Rab11, Traffic 5 (2004) 20–36. 34997–35010. [57] N. Vitale, S. Chasserot-Golaz, Y. Bailly, N. Morinaga, M.A. Frohman, M.F. Bader, [27] H.S. Li, D.B. Stolz, G. Romero, Characterization of endocytic vesicles using Calcium-regulated exocytosis of dense-core vesicles requires the activation of magnetic microbeads coated with signalling ligands, Traffic 6 (2005) 324–334. ADP-ribosylation factor (ARF)6 by ARF nucleotide binding site opener at the [28] N. Porat-Shliom, Y. Kloog, J.G. Donaldson, A unique platform for H-Ras signaling plasma membrane, J. Cell Biol. 159 (2002) 79–89. involving clathrin-independent endocytosis, Mol. Biol. Cell 19 (2008) 765–775. [58] Q. Zhang, J. Calafat, H. Janssen, S. Greenberg, ARF6 is required for growth factor- [29] S.E. Robertson, S.R.G. Setty, A. Sitaram, M.S. Marks, R.E. Lewis, M.M. Chou, and Rac-mediated membrane ruffling in macrophages at a stage distal to Rac Extracellular signal-regulated kinase regulates clathrin-independent endosomal membrane targeting, Mol. Cell. Biol. 19 (1999) 8158–8168. trafficking, Mol. Biol. Cell 17 (2006) 645–657. [59] J.G. Donaldson, Multiple roles for Arf6: sorting, structuring, and signaling at the [30] S. Roy, B. Wyse, J.F. Hancock, H-Ras signaling and K-Ras signaling are differentially plasma membrane, J. Biol. Chem 278 (2003) 41573–41576. dependent on endocytosis, Mol. Cell. Biol. 22 (2002) 5128–5140. [60] L. Martinu, J.M. Masuda-Robens, S.E. Robertson, L.C. Santy, J.E. Casanova, M.M. [31] M.A. Booden, T.L. Baker, P.A. Solski, C.J. Der, S.G. Punke, J.E. Buss, A non- Chou, The TBC (Tre-2/Bub2/Cdc16) domain protein TRE17 regulates plasma farnesylated HRas protein can be palmitoylated and trigger potent differentiation membrane-endosomal trafficking through activation of Arf6, Mol. Cell. Biol. 24 and transformation, J. Biol. Chem. 274 (1999) 1423–1431. (2004) 9752–9762. [32] T.L. Baker, H. Zheng, J. Walker, J.L. Coloff, J.E. Buss, Distinct rates of palmitate [61] B. Rotblat, I.A. Prior, C. Muncke, R.G. Parton, Y. Kloog, Y.I. Henis, J.F. Hancock, Three turnover on membrane-bound cellular and oncogenic H-Ras, J. Biol. Chem. 278 separable domains regulate GTP-dependent association of H-Ras with the plasma (2003) 19292–19300. membrane, Mol. Cell. Biol. 24 (2004) 6799–6810. J. McKay et al. / Biochimica et Biophysica Acta 1813 (2011) 298–307 307

[62] P. Sharma, S. Sabharanjak, S. Mayor, Endocytosis of lipid rafts: an identity crisis, [65] N.Y. Mitin, M.B. Ramocki, A.J. Zullo, C.J. Der, S.F. Konieczny, E.J. Taparowsky, Cell Dev. Biol. 13 (2002) 205–214. Identiﬁcation and characterization of Rain, a novel Ras-interacting protein with a [63] L.M. Luttrell, F.L. Roudabush, E.W. Choy, W.E. Miller, M.E. Field, K.L. Pierce, R.J. unique subcellular localization, J. Biol. Chem. 279 (2004) 22353–22361. Lefkowitz, Activation and targeting of extracellular signal-regulated kinases by b- [66] S.E. Tague, V. Muralidharan, C. D'Souza-Schorey, ADP-ribosylation factor 6 arrestin scaffolds, Proc. Natl. Acad. Sci. USA 98 (2001) 2449–2454. regulates tumor cell invasion through the activation of the MEK/ERK signaling [64] K. Hibino, T.M. Watanabe, J. Kozuka, A.H. Iwane, T. Okada, T. Kataoka, T. Yanagida, pathway, Proc. Natl. Acad. Sci. 101 (2004) 9671–9676. Y. Sako, Single- and multiple-molecule dynamics of the signaling from H-Ras to [67] O.A. Jovanovic, F.D. Brown, J.G. Donaldson, An effector domain mutant of Arf6 cRaf-1 visualized on the plasma membrane of living cells, Chem. Phys. Chem. 4 implicates phospholipase D in endosomal membrane recycling, Mol. Biol. Cell 17 (2003) 748–753. (2006) 327–335.