Protein Delivery with Transportans Is Mediated by Caveolae Rather

Apr 6, 2009 - Data from the FACS experiments and western band quantification are the mean ± SD (n = 4−6). The down-regulation of flotillin-1 was qu...
0 downloads 3 Views 9MB Size
Bioconjugate Chem. 2009, 20, 877–887

877

Protein Delivery with Transportans Is Mediated by Caveolae Rather Than Flotillin-Dependent Pathways Pille Sa¨a¨lik,†,⊥ Ka¨rt Padari,† Aira Niinep,† Annely Lorents,† Mats Hansen,‡,| Eija Jokitalo,§ Ülo Langel,‡,| and Margus Pooga*,†,⊥ Institute of Molecular and Cell Biology, University of Tartu, Estonia, Department of Neurochemistry, Stockholm University, Sweden, Institute of Biotechnology, Electron Microscopy Unit, University of Helsinki, Finland, Institute of Technology, University of Tartu, Estonia, and Estonian Biocentre, Tartu, Estonia. Received September 30, 2008; Revised Manuscript Received March 3, 2009

Delivery of large bioactive cargoes into cells with the help of cell-penetrating peptides (CPPs) is mostly based on endocytic processes. Here we map the cellular pathways used by transportan and transportan 10 (TP10) for protein transduction in HeLa cells. CPP-mediated cellular delivery is often suggested to be lipid-raft-dependent; therefore, we used flotillin-1, caveolin, Rab5, and PI3P as markers to elucidate the involvement of these particular endosomal pathways in the protein uptake process. Confocal laser scanning and electron microscopy reveal only a negligible overlap of avidin/neutravidin conveyed into cells by transportans with the raft marker flotillin-1 or early endosomal markers Rab5 and PI3P. However, about 20% of protein-CPP complexes colocalize with the caveolar/caveosomal marker caveolin, and down-regulation of caveolin-1 by siRNA treatment leads to the inhibition of the CPPmediated protein uptake by 30-50%. On the contrary, the lack of flotillin-1 increases rather than decreases the CPP-mediated protein transport. The participation of the caveolin-1-dependent pathway in CPP-mediated protein delivery was also corroborated by using caveolin-1 knockout mouse embryonic fibroblasts.

INTRODUCTION The cellular transport of bioactive but low-permeable molecules using cell-penetrating peptides (CPPs) has been under intense investigation as an alternative mode for drug delivery for the past 15 years. During this period, the number of peptides with putative membrane-translocating abilities has grown dramatically and has also led to the broadening of the meaning of the term CPP toward less explicit characteristics (1). Currently CPP is considered a relatively short peptide being capable of gaining access to the cell interior by means of different mechanisms, including endocytosis, and having the capacity to promote the intracellular delivery of covalently or noncovalently attached bioactive cargoes. The main focus of CPP studies has now shifted to improving the specificity of targeting and characterization of the functionality of delivered cargo in cells. However, knowledge about the cell entry mechanism(s) is still of high importance, as the mode of internalization determines addressing of the cargo and thereby its residual activity. Some CPPs are capable of using a nonvesicular cellular entry mechanism under certain conditions (2-6). However, in these studies the peptide was carrying a small cargo, a fluorescent tag (2, 3, 6) or a short peptide (4, 5), and the observed diffuse cellular staining, which is thought to reflect a nonvesicular internalization mode, was detected at high peptide concentrations or low temperature. Still, most peptide-cargo constructions, especially large ones, gain entry to the cell interior by inducing endocytic processes, i.e., within various types of plasmamembrane-derived vesicles (7-9). The rapidly accumulating * To whom correspondence should be addressed. Tel: (372)7 375 049, fax: (372)7 420 286, e-mail: [email protected]. † Institute of Molecular and Cell Biology, University of Tartu. ⊥ Estonian Biocentre. ‡ Stockholm University. | Institute of Technology, University of Tartu. § University of Helsinki.

knowledge about endocytosis mechanisms and intracellular targeting pathways also necessitates a more detailed analysis of routes involved in CPP-mediated cargo delivery and a better definition of the transport vesicles used. CPPs as relatively unspecific ligands have been associated with many endocytosis routes. Clathrin-mediated endocytosis (CME1) is the most common for down-regulation of a variety of plasma membrane receptors. Although different receptors may utilize highly specifically regulated mechanisms, the CME seems to be the general route of uptake (10-12). Without the specification of the exact mode of CME, this pathway has been shown to be active in CPP-mediated cargo delivery in several experimental systems, although with different efficacy (13, 14). In addition, caveolae may participate in CPP-mediated cargo delivery. Caveolae are abundant in some cell types, and their functions extend from endocytosis to cell signaling (15). Interestingly, internalization of Tat-GFP fusion protein has been demonstrated to proceed using the caveolar pathway (16, 17) but has also been reported to use macropinocytosis (7). Macropinocytosis is initiated by certain stimuli (18), and it is reported that CPPs activate the cellular signaling cascades and subsequent macropinocytosis in analogy with growth factors as stimuli (19, 20). The least characterized/understood type of vesicular transport, clathrin- and caveolin-independent endocytic pathway, is active in pathogen internalization and also in some physiological processes. Simian virus 40 (SV40) has been shown to enter even caveolin-1 (Cav-1) knockout cells, which express dominant-negative Eps15 that interferes with the assembly of the clathrin-coated pits and the formation of caveolae. Accord1 Abbreviations: AF, Alexa Fluor; Cav, caveolin; CLSM, confocal laser scanning microscopy; CME, clathrin-mediated endocytosis; CtxB, cholera toxin B subunit; FACS, fluorescence-activated cell sorter; Flot1, flotillin-1; GFP, green fluorescent protein; PI3P, phosphatidyl inositol3-phosphate; pTat, Tat peptide; siRNA, small interfering RNA; TEM, transmission electron microscopy; TxR, Texas Red; TP, transportan; TP10, transportan 10.

10.1021/bc800416f CCC: $40.75  2009 American Chemical Society Published on Web 04/06/2009

878 Bioconjugate Chem., Vol. 20, No. 5, 2009

ingly the internalization of SV40 is dependent on cholesterol, microtubules, and Arf1 (21). In addition, the internalization of cholera toxin B subunit (CtxB) is not inhibited in cells depleted of functional dynamin and Arf6, and it takes place by inducing formation of a tubular network that is not present in cells with active dynamin and Arf6 (22). The network formation was dependent on cholesterol (22), suggesting that lipid rafts play a significant role in this type of cellular trafficking. However, the lack of adequate methods for assessing the dynamics of lipids in membranes of living cells limits the studies to using mainly the raft-associated proteins (23). Recently, flotillin was suggested to be a membrane-organizing marker protein of lipid rafts. Endocytosis of CtxB subunit via flotillin-positive domains takes place independently of dynamin, and the formed vesicles do not contain proteins characteristic to CME or caveosomes (24). We have described the cellular routing of complexes of transportan and TP10 with protein earlier based on morphological features of used endocytic vehicles (8, 14). However, we could not define the nature of a substantial fraction of vesicles that contained complexes of protein and transportan/TP10 by any of the used cellular markers and endocytic inhibitors (8). As the recently mapped flotillin-1-marked endocytic pathway and the CPP-mediated cellular delivery are both cholesteroldependent and flotillin-1-defined membrane areas and endocytic vesicles do not overlap with the clathrin- or caveolin-dependent route (24, 25), we analyzed the cellular traffic of complexes of transportan/TP10 and avidin/neutravidin in relation to this novel lipid raft marker. We also included caveolin as a marker of specific type of membrane rafts in cells lacking the typical caveolae. The role of cholesterol rich membrane domains and the respective proteins in transportan/TP10-mediated protein was elucidated by down-regulation of flotillin-1 and caveolin-1 using siRNA. In addition, the cellular localization of complexes in relation to early endosomal markers Rab5 and phosphatidylinositol-3 phosphate (PI3P) was mapped in order to disclose the early steps of transportan-mediated delivery. Our data demonstrate that flotillin-containing lipid rafts do not participate in CPP-mediated delivery, and their absence rather increases translocation efficiency.

EXPERIMENTAL PROCEDURES Reagents. For immunomicroscopy experiments, the following antibodies (Ab) were used: mouse monoclonal anti-PI3P (Echelon Biosciences, UT), rabbit polyclonal anticaveolin (cat no 610059, BD Transduction Laboratories, Belgium), anti-Rab5 (AbCam, UK), polyclonal antiflotillin-1 (Santa Cruz Biotechnology, Heidelberg, Germany), and mouse monoclonal anti-β1 integrin (anti-CD29, BD Transduction Laboratories). Alexa Fluor (AF) 488- or AF 594-conjugated antimouse and antirabbit antibodies, AF 594-conjugated transferrin and cholera toxin B subunit (CtxB), avidin-Texas Red, and Oligofectamine were bought from Invitrogen (UK). FITC-labeled avidin, saponin, and sodium thiosulfate pentahydrate were from Sigma-Aldrich and Nanogold-Fab′ fragments of goat antirabbit IgG and HQ Silver kit were purchased from Nanoprobes (Yaphank, NY). siRNA against flotillin-1 (sc-35391), caveolin-1 (sc-44202) and nonspecific siRNA (sc-37007) were acquired from Santa Cruz Biotechnology. Peptide Synthesis. Transportan (GWTLNSAGYLLGKINLKALAALAKKIL-NH2), TP10 (AGYLLGKINLKALAALAKKIL-NH2), and pTat (GRKKRRQRRRPPQ-NH2) were synthesized stepwise on a 0.1 mmol scale on an automated peptide synthesizer (Applied Biosystems Model 431A) using the t-Boc solid-phase peptide synthesis strategy. tert-Butyloxycarbonyl amino acids were coupled as hydroxybenzotriazole esters to a p-methylbenzylhydrylamine resin (Neosystem, Strasbourg, France) to obtain C-terminally amidated peptides. Biotin was coupled

Sa¨a¨lik et al.

manually to the side chain of Lys13 or Lys7 in transportan and TP10, respectively, or to N-terminus of pTat. The final cleavage and purification of peptides was performed, as described earlier (14). The molecular weights of biotinylated peptides were determined by MALDI-TOF mass spectroscopy (prO-TOF 2000, PE Biosciences) and calculated molecular weights were obtained each time. Cell Culture. Human cervical carcinoma cell line HeLa (ATCC CCL-2), wild type mouse (C57bl6/J) embryonic fibroblasts, and embryonic fibroblasts from caveolin-1 knockout 3T3 mouse (ATCC-CRL-2753) were cultured in a humidified atmosphere containing 5% CO2 at 37 °C in Iscove’s Modified Dulbecco’s Medium (IMDM) (Gibco, Invitrogen, UK) supplemented with 10% fetal bovine serum (FBS) (PAA Laboratories, Austria), 100 IU/mL penicillin, and 100 µg/mL streptomycin (Gibco, Invitrogen, UK). Confocal Laser Scanning Microscopy. An amount of 5 × 104 cells were seeded one day before experiments onto round glass coverslips (Ø 12 mm, Menzel-Glazer, Germany) on a 24well plate. Cells were incubated with 0.15 µM Texas Red (TxR)or FITC-labeled avidin and 0.5 µM biotinyl-CPP in 400 µL of serum-free IMDM at 37 °C for 30 min. Biotinylated transportan and TP10 were dissociated with 10 mg/mL leucine before formation of complexes with avidin in order to reduce the aggregation of peptides and the size of subsequently forming complexes. After incubation, the medium with CPP-avidin complexes was removed, the cells were rinsed twice with PBS and fixed for 2 h either with PLP fixative (2% paraformaldehyde, 75 mM lysine-HCl, and 10 mM sodium periodate in 75 mM phosphate buffer, pH 7.4) or with 4% paraformaldehyde for 30 min for staining with antibodies against flotillin-1/Rab5/β1integrin or PI3P/caveolin, respectively. The cells were permeabilized either with 0.1% Triton X-100 (anticaveolin), 0.5% saponin (anti-PI3P) for 5 min, or with a solution of 0.01% saponin and 0.1% BSA in sodium phosphate buffer for 8 min, which also acted as a blocking solution (flotillin-1, Rab5, β1integrin). Nonspecific binding sites were blocked with 10% w/v nonfat dry milk solution in PBS for 1 h (anticaveolin) or with 10% heat-inactivated goat serum (anti-PI3P). After treatment with the respective AF-labeled secondary antibodies, the coverslips were mounted in 30% glycerol in PBS. For control experiments, AF 594-labeled transferrin and CtxB were applied to cells at 20 µg/mL and 3 µg/mL, respectively, for indicated time periods. For low temperature experiments, cells were precooled for 15 min followed by uptake of peptide-protein complexes/CtxB during 30 min on ice and treated for Flot-1 visualization as indicated above. Images were obtained with Nikon EZ-Z1 confocal microscope using excitation at 488 nm (for Alexa Fluor 488) and 543 nm (for Texas Red, AF 594) or with Olympus IX81 inverted microscope equipped with FluoView 1000 confocal system using excitation at 488 nm (for AF 488) and 559 nm (for Texas Red, AF 594). In both cases the systems were run in sequential scanning mode, where only one laser was active at a time, to avoid spectral overlap. Obtained images were processed with Adobe Photoshop 7.0. Down-Regulation of Flotillin-1 and Caveolin-1. An amount of 4 × 104 cells were seeded onto 12-well culture plates one day before siRNA transfections. Transfection was performed with duplex RNA oligonucleotides against flotillin-1 or caveolin1, or nonspecific siRNA using Oligofectamine (Invitrogen), according to the manufacturer’s instructions in two consecutive days. Concentrations of siRNA for single transfection were 0.1 µM for flotillin-1 (Flot-1) and nonspecific siRNA. Caveolin-1 (Cav-1) siRNA was applied at 0.1 and 0.2 µM on consecutive days. For the cotransfection of Flot-1 and Cav-1, the concentration of both siRNA was 0.1 µM per transfection. On day 4 the

Protein Delivery with Transportans

cells were incubated 30 min in 700 µL of serum-free medium containing CPP-protein complexes at the same concentrations as in confocal microscopy experiments. The complex-treated cells were trypsinized (0.025% trypsin supplemented with 0.01% EDTA in PBS) for 10 min prior to FACS analysis to remove the membrane-bound but noninternalized avidin. The remaining extracellular fluorescence was quenched by addition of trypan blue into the cell suspension (final concentration 0.015%), and 3 × 104 cells were counted per sample. Measurements were performed in duplicates, and the results are presented as a mean of three separate experiments. Western Blot. 105 HeLa cells were seeded onto six-well plates and treated with Flot-1 and/or Cav-1 siRNA at the same concentrations as described above. Forty eight hours after the second siRNA treatment, the cells were scraped off the dishes and pelleted by centrifugation, and the wet cell mass was weighed for determination of the protein amount. Cells were lysed in 2 × SDS sample buffer and sonicated, and 0.5 µg of protein per lane was separated on 15% SDS-PAGE. Proteins were transferred to polyvinyl difluoride (PVDF, Applied Biosystems) membrane, blocked in 5% solution of nonfat dry milk in PBS, and treated with primary antibodies against Cav, Flot-1 (0.5 and 0.4 µg/mL, respectively), or β-actin as internal control (0.6 µg/mL). After incubation with alkaline phosphatase (AP)conjugated antirabbit (Flot-1, Cav) or antimouse (anti-β-actin) secondary antibody (1:2500), the protein bands were visualized with AP substrate BCIP/NBT treatment. Intensities of protein bands of specific size (47 kDa for Flot-1, 21-24 kDa for Cav, and 42 kDa for β-actin) were analyzed by Image J 1.38x. Electron Microscopy. HeLa cells were seeded onto glass coverslips in 24-well plates and grown to 90-100% confluency. Cells were incubated with complexes of biotinylated transportan, TP10, or pTat and colloidal gold-labeled neutravidin at 37 °C for 1 h. CPP-protein complexes were preformed by incubating neutravidin-gold (d 10 nm, 1:100 or 1:200 dilution) (8) with biotinyl-CPP at room temperature in a minimal volume for 5 min. Solution with complexes was diluted with culture medium to a final peptide concentration of 1.5-3 µM according to experiment and applied to cells. Pre-embedding Immunolabeling. The cells treated with CPP-neutravidin complexes were fixed for immunoelectron microscopy with PLP fixative at room temperature for 2 h and permeabilized with 0.01% saponin and 0.1% BSA in 0.1 M sodium phosphate buffer for 8 min. After permeabilization, the cells were incubated with anticaveolin (1:50) or antiflotillin-1 antibodies (1:30) for 1 h and treated with 1.4 nm nanogoldconjugated Fab′ fragments of secondary antibodies (1:60) for an additional 1 h. The nanogold label was magnified by silver enhancement according to manufacturer’s protocol for 2-5 min, followed by gold-toning with 0.05% gold chloride (26). After postfixation with 1% osmium tetroxide in the 0.1 M sodium cacodylate buffer (pH 7.4), cells were dehydrated and embedded in epoxy resin. Ultrathin sections were cut parallel to the coverslip, poststained with uranyl acetate and lead citrate, and examined with a JEM-100S (JEOL, Japan) or FEI Tecnai 10 (Philips, Netherlands) transmission electron microscope at 80 kV. The scanned electron microphotos were analyzed and processed with Adobe Photoshop 7.0. Statistical Analyses. Confocal fluorescence figures are representative of at least two independent experiments. Data from the FACS experiments and western band quantification are the mean ( SD (n ) 4-6). The down-regulation of flotillin-1 was quantified in CLSM images by using FluoView10-ASW 1.6 software. Images of cells treated either with nonspecific or flotillin-1 specific siRNA were recorded with identical microscope settings. The regions of interest (ROI) that were chosen at the plasma membrane based on the β1-integrin

Bioconjugate Chem., Vol. 20, No. 5, 2009 879 Table 1. Quantitative Analysis of Colocalization of Cell-Delivered Neutravidin and Lipid Raft Marker Proteins in TEM Imagesa flotillin-1

caveolin

content of vesicle

TPb

TP10b

TPb

TP10b

CPP-protein complexes Flot-1/Cav CPP-protein and Flot-1/Cav % of colocalization

158 522 23 14.5b

242 644 33 13.6b

467 942 93 20

183 1396 28 15.3

a The caveolinor flotillin-1-positive and CPP-protein complex-containing vesicles were counted in 20 randomly selected cells, and colocalization is presented in percents. b Colocalization of CPP-protein complexes with flotillin-1 was detected almost exclusively in the perinuclear region. The degree of colocalization at the plasma membrane was lower than 5%.

staining and in the cytoplasmic areas were analyzed separately on 6-10 z-sections from seven cells (>40 sections for each pool). The results are shown as the percentage of flotillinpositive pixels of all pixels from the analyzed ROI. The colocalization of CPP-protein complexes with caveolin or flotillin-1 was quantified by counting vesicles in randomly selected 20 cells in thin sections by TEM. The number of vesicles containing respectively CPP-protein complexes, caveolin, or flotillin-1 or both, the CPP-protein complexes, and endocytic markers was counted and the percentage of colocalization of CPP-protein with Cav/Flot-1 was calculated (Table 1).

RESULTS Transportan and TP10 Do Not Direct the Cargo Protein to Flotillin-1-Dependent Endosomal Pathways. To unravel the internalization mode of transportan-mediated cargo delivery, we mapped the localization of the peptide-protein complexes in relation to flotillin-1, previously described as a marker of a specific type of lipid rafts (24). The flotillin-1 (Flot-1) positive areas showed almost no overlap with complexes of avidin and biotinylated transportan or TP10 when evaluated by confocal microscopy (Figure 1A,B). Almost no colocalization could be detected inside cells whereas some overlap was observed at the plasma membrane after 30 min incubation with peptide-protein complexes. At the plasma membrane, colocalization was mostly detected as smaller flotillin-marked focuses covered with a bigger irregular peptide-protein aggregate (Figure 1A-C). The loci of flotillin-1 were detected as a dense punctate staining on the plasma membrane, in the cytoplasm, and on the cellular projections while the majority of complexes assembled into larger structures, which were associated with the outer membrane, as detected by confocal microscopy (Figure 1A-C). As reported previously, cholera toxin B subunit uses (CtxB) uses flotillin-dependent endocytosis as one route for cell entry (24); therefore, we used CtxB for proving the relevance of this pathway in our system. After 30 min incubation of HeLa cells with the toxin, we detected CtxB in perinuclear structures showing partial overlap with Flot-1 (Figure 4A), confirming the functionality of this pathway in HeLa cells and corroborating the results published by Glebov et al. In parallel to confocal microscopy (CLSM) we used transmission electron microscopy (TEM), which allows identification of cellular structures by morphology at a higher resolution level. The complexes of biotinylated CPP and neutravidin (bigger gold particles) were detected as bigger aggregates, which localized mostly to the extracellular side of the plasma membrane, cell protrusions, and vesicular structures in the cytoplasm (arrows in Figure 1D-F). Flot-1 protein (smaller gold particles), on the contrary, was detected in specific loci at the cytoplasmic side of the plasma

880 Bioconjugate Chem., Vol. 20, No. 5, 2009

Sa¨a¨lik et al.

Figure 1. Localization of CPP-protein complexes in HeLa cells in relation to flotillin-1. For CLSM experiments, cells were incubated for 30 min with 0.5 µM biotinylated transportan (A), TP10 (B), or pTat (C) complexed with 0.15 µM avidin-Texas Red (red). Flotillin-1 was visualized with antiflotillin-1 pAb and AF 488-conjugated secondary Ab (green). A single optical section of cells is presented, which is indicated with a green line on the z-projections. Yellow lines in the xy-projection indicate the location of z-projections, and arrowheads indicate the apical part of cells in z-projections. (D-F) TEM images of cells incubated with complexes of neutravidin-gold (10 nm, arrows) and 1.5 µM biotinylated transportan (D), 3 µM TP10 (E), or pTat (F) for 1 h. Flotillin-1 is visualized with antiflotillin-1 pAb and secondary antibody tagged with nanogold (1.4 nm) particles (arrowheads in D-F). The bar in CLSM images represents 10 µM.

membrane (arrowheads in Figure 1E,F) but also on the membranes of multivesicular endosomes (arrowheads in Figure 1D) in the cytoplasm. Although the complexes of CPP-neutravidin were found close to the location sites of flotillin-1 protein, we could not detect colocalization between Flot-1 and CPP-protein complexes at the plasma membrane by TEM (Figure 1E,F). However, a small population of the intracellular vesicles containing TP/TP10-protein complexes also included Flot-1 protein. The same results were obtained with biotinylated Tat peptide complexed with avidin-Texas Red (CLSM) or neutravidin-gold (TEM) (Figure 1C,F). For transportan and TP10-protein complexes, overlapping structures with Flot-1 in TEM sections were estimated to remain around 14% of the total number of complex-containing vesicles (Table 1). Still, as revealed by quantitative analysis, the colocalization of TP/TP10-protein complexes with Flot-1 on the membrane areas and in the cortical cytoplasm was insignificant, only 1-5%, and the vesicles containing both markers were almost exclusively observed in the perinuclear region. We also performed CLSM experiments using a shorter incubation time, as peptide-protein complexes might pass the flotillin-positive vesicles earlier. However, after 10 min of incubation, very few complexes had internalized, and the colocalization between Flot-1 and TP/TP10-protein complexes was rarely detected on the plasma membrane (data not shown). In addition, we conducted experiments at a lower incubation temperature to assess the initial interaction of complexes with the plasma membrane. Although the peptide-protein complexes were not taken up by cells after 30 min incubation at 4 °C, the complexes interacted with the plasma membrane in a similar manner and colocalization with flotillin-1 was identical to that observed at physiological temperature (Supporting Information Figure 1A-C). CtxB interacted with the plasma membrane at 4 °C less avidly,

but still colocalized with Flot-1 at certain loci, and no intracellular staining was detected (Supporting Information Figure 1D). Transportan and TP10 Avoid the Early Endosomal Pathway Marked by Rab5 and PI3P in the Cellular Delivery of Cargo Protein. In order to assess whether the early endosomal pathway is exploited in peptide-mediated protein transport, we characterized the cellular localization of CPP-avidin complexes in comparison to Rab5 and PI3P, markers of early endosomes. The analogous rare intracellular colocalization of CPP-protein complexes as seen with flotillin-1 was also detected with PI3P (Figure 2A,B) and Rab5 (Figure 2D,E) as observed by CLSM after 30 min incubation. On the contrary, marked overlap was detected between PI3P/Rab5 and transferrin (Figure 2C,F), a marker of clathrin-dependent endocytosis, which passes early endosomes. Though both the CPP-protein complexes and PI3P/Rab5 were confined to punctate structures in cells and on the plasma membrane, they clearly represented different localization areas. Somewhat surprisingly the shorter (10 min) incubation time also revealed even less colocalization between the internalized CPP-protein complexes and Rab5 or PI3P at the plasma membrane, and a small amount of complexes was detected inside cells (Supporting Information Figure 2). Transferrin, on the contrary, entered the cell interior rapidly, already in 10 min (Supporting Information Figure 2). Again, in cases of rare overlap, the focuses stained by anti-Rab5 or antiPI3P antibody were situated in the middle of large conglomerates of CPP-protein complexes. The negligible role of the PI3Pdependent pathway in CPP-mediated protein delivery (Figure 2A,B) correlates well with our unpublished results showing that internalization of TPb/TP10b-avidin complexes was not influenced in HeLa cells where the activity of PI3-kinase was inhibited by wortmannin treatment (data not shown).

Protein Delivery with Transportans

Bioconjugate Chem., Vol. 20, No. 5, 2009 881

Figure 2. Localization of transportan- and TP10-protein complexes in HeLa cells in relation to Rab5 and PI3P. Cells were incubated for 30 min with 0.5 µM biotinylated transportan (A, D), TP10 (B, E) complexed with 0.15 µM avidin-Texas Red, or with AF 594-conjugated transferrin (20 µg/mL, C, F). To visualize PI3P and Rab5, cells were stained with anti-PI3P mAb (A, B) or anti-Rab5 pAb (D, E) and with appropriate AF 488-conjugated secondary Ab (green). A single optical section from the equatorial level of cells is presented. The bar represents 10 µM.

A Fraction of CPP-Protein Complexes Is Targeted to Caveolae. Earlier reports have demonstrated the involvement of caveolae-dependent endocytosis in both CPP internalization and the CPP-mediated cargo delivery (16, 17). Our results obtained by electron microscopy have shown that TP/ TP10-protein complexes enter cells often in small (50-100 nm in diameter) morphologically caveosome-like or noncoated vesicles. Therefore, we looked for the presence of caveolin in these structures by both immunofluorescence and immunoelectron microscopy using antibody against all caveolin (Cav) isoforms. Indeed, a fraction of CPP-protein complexes colocalized with caveolin on the plasma membrane and in cells as observed by confocal microscopy. A higher number of colocalization could be detected on the plasma membrane than in the cytoplasm for all peptide-protein complexes. About 30% of CPP-avidin complexes on the membrane and 20% inside the HeLa cells exhibited similar localization to caveolin by visual estimation of CLSM images (Figure 3A-C). As expected, CtxB known to be also endocytosed via caveolae (27), exhibited a good overlap with caveolin-positive cell areas (Figure 4B). The results of electron microscopy revealed that most of the small vesicles or bigger rosette-like structures had caveolin on the cytosolic side (small gold particles in Figure 3D), and these structures were always located in close proximity to actin fibers. Although the caveosomes were abundant in the cortical cytoplasm, most of them did not contain the TP/TP10-neutravidin complexes (bigger gold particles). However, several of the TP/ TP10-protein, as well as pTat-protein complexes, were found to reside in the Cav-positive vesicles in the perinuclear region (arrows in Figure 3G and 3H). The presence of both TP/ TP10-neutravidin complexes and caveolin was also detected in bigger vesicles (0.4 µm), which had small caveosomes at close proximity or were connected to them (arrowheads in enlarged section of Figure 3E). Among complex-containing endosomal structures, about 20% for TP and 15% for TP10 were also marked by caveolin (Table 1) as revealed by the quantitative analysis of TEM results. The overlap of pTat-protein with

caveolin-positive structures was comparable to that of TP-protein and caveolin, reaching about 20% (data not shown). About onethird of all CPP-protein complex-containing vesicles that did not contain caveolins were found in close proximity to caveosomes as shown in Figure 3F. It might have led to some overestimation of the degree of colocalization when using confocal microscopy (Figure 3A-C). In addition, from the analysis of TEM images we found that the overall number of caveosomes in control cells was about 2-fold lower than in cells incubated with CPP-protein complexes, indicating a triggering of caveolae internalization by CPPs. Down-Regulation of Caveolin-1 Inhibits Transportan/ TP10-Mediated Protein Delivery, but Down-Regulation of Flotillin-1 Has No Effect. In parallel with the colocalization experiments we down-regulated the Flot-1 with siRNA to see whether the lack of cellular Flot-1 interferes with the uptake of peptide-protein complexes. Surprisingly, the used siRNA decreased the Flot-1 concentration only moderately, to 40-50% as estimated by quantification of Western blot results (Figure 5A). Flotillin-1 seemed to be down-regulated differently in the plasma membrane and central cytoplasm; therefore, we visualized the plasma membrane with β1 integrin antibodies and analyzed the amount of Flot-1 at plasma membrane and in cytoplasm separately (see Experimental Procedures). The quantification corroborated that the down-regulation was more efficient at the plasma membrane compared to the cell interior, reaching 75-80% decrease of the Flot-1 signal in plasma membrane areas whereas the concentration of flotillin-1 in intracellular structures was reduced less than 2 times (Figure 5B). The uptake of TP/TP10-protein complexes by Flot-1 down-regulated cells, however, was not influenced (Figure 5C). Analogously to Flot-1, we checked whether depletion of Cav-1 from the plasma membrane influences the TP/TP10-mediated protein delivery. siRNA treatment yielded about 70-80% downregulation of cellular Cav-1 protein expression as quantified by Western blot (Figure 5A). The decrease in Cav-1 concentration inhibited the uptake of TP10-protein complexes by 40-50%

882 Bioconjugate Chem., Vol. 20, No. 5, 2009

Sa¨a¨lik et al.

Figure 3. Localization of CPP-protein complexes in HeLa cells in relation to caveolin. Cells were incubated for 30 min with 0.5 µM biotinylated transportan (A), TP10 (B), or pTat (C) complexed with 0.15 µM avidin-Texas Red (red). To visualize the cellular caveolin, cells were stained with anticaveolin pAb and with appropriate AF 488-conjugated secondary Ab (green). A single optical section from the equatorial level of cells is presented. (D-H) TEM images of cells incubated with complexes of neutravidin-gold (10 nm, arrows) and 1.5 µM biotinylated transportan (D, G), TP10 (E), or pTat (F, H) for 1 h. Caveolin is visualized with anticaveolin pAb and 1.4 nm gold particle-conjugated secondary antibodies. The bar in CLSM images represents 10 µM.

while the TP-mediated transport was slightly less affected (Figure 5D). The coapplication of Flot-1 and Cav-1 siRNA did not lead to an additional inhibitory effect on the protein delivery as compared to Cav-1 knock-down (Figure 5E). In addition to the impaired protein uptake in caveolin-1 down-regulated cells by CLSM, the colocalization between the CPP-protein complexes and the remaining caveolin was markedly decreased for both peptides, most probably due to the lower amount of membranous caveolin (data not shown). However, by visual estimation the amount of membrane-associated complexes in Cav-1 down-regulated cells was on the same level with untreated control cells, indicating that the interaction with plasma membrane is not inhibited by a deficit of Cav-1. On the other hand, the extensive trypsin digestion and trypan blue quenching used in the FACS analysis helped to remove the loosely bound CPP-protein complexes from the cell surface and to diminish the extracellular signal, respectively. These treatments led to lower signal intensity and more correct estimation of the reduction in protein transduction in flow cytometry assays. To completely exclude the induction of caveolin-1-dependent pathways by CPPs, we conducted analogous experiments in Cav-1 knockout mouse embryonic fibroblasts (MEFs). The results obtained by flow cytometry and confocal microscopy corroborated the high impact of Cav-1-dependent pathways on the uptake of avidin induced by peptides. The transduction yield of avidin into Cav-1 knockout MEFs was 60-70% lower than

to wild-type fibroblasts for both TP10 and transportan. In analogy with the HeLa cells, the absence of caveolin-1 interfered with transportan-mediated delivery slightly less than that of TP10 (Figure 6A). Visualization of caveolins with anticaveolin antibody revealed its marked colocalization with TP/TP10-protein complexes in wild type MEFs (Figure 6B,D), which correlates well with the results obtained in HeLa cells. As expected, in Cav-1 knockout MEFs we detected the presence of caveolin-2 isoform at a very low level (28), which did not overlap with peptide-protein complexes (Figure 6C,E), corroborating the substantial role of membranous caveolin-1 in protein delivery with transportans.

DISCUSSION The plasma membrane is a complex cellular organelle, being simultaneously a portal and a barrier for all metabolites. Cellpenetrating peptides are under intense study as carriers, which cross this barrier together with biologically active cargo. On the other hand, their ability to penetrate through membranes also represents an interesting scientific phenomenon per se. The endocytic processes have proven to mediate the CPP/CPP-cargo entry, and the role of pathways that originate from the lipid rafts and/or caveolae is often emphasized (7, 16, 29). Involvement of rafts is usually suggested based on the decreased cellular uptake of CPPs by cells depleted of cholesterol in the plasma membrane after treatment with methyl-β-cyclodextrin. However,

Protein Delivery with Transportans

Figure 4. Localization of cholera toxin B subunit (CtxB) in relation to cellular flotillin-1 and caveolin. After incubation of HeLa cells with Alexa Fluor 594-conjugated CtxB (3 µg/mL, red) for 30 min, the cells were stained with antiflotillin-1 (A) or anticaveolin (B) polyclonal antibodies and with Alexa Fluor 488 conjugated secondary Ab (green). Enlarged sections of indicated areas are depicted on the right column with z-projections. Yellow lines in enlarged sections indicate the location of z-projections, and arrowheads indicate the apical part of cells in z-projections. Scale bar 10 µM.

it does not distinguish between the caveolar and raft-mediated pathways, which have different cellular destinations. Removal of cholesterol also influences other endocytic processes and might lead to changes in cell morphology (30). Moreover, very recently 60% of cellular cholesterol was demonstrated to localize in the inner leaflet of plasma membrane and endocytic recycling compartment in chinese hamster ovary cells (31), suggesting that cholesterol depletion might substantially interfere with the organization of lipids in the cytoplasmic leaflet of membranes and also endocytosis. Continuously accumulating data about the nature and complexity of endocytosis and involvement of its different types in CPP mechanisms prompted us to reassess the initial steps of CPP-mediated protein delivery. We chose flotillin-1 and caveolin as markers for two types of plasma membrane microdomains, and Rab5 and PI3P for characterization of early endosomal vesicles, to elucidate their involvement in TP/TP10-mediated protein delivery into HeLa cells. Cholesterol- and sphingolipid-rich rafts represent highly heterogeneous populations of functionally distinct membrane domains. The availability of caveolae-independent but morphologically similar endocytic pathways involved in the SV40 cell entry was recently demonstrated by two groups in parallel (21, 22). In analogy with CPP-mediated delivery into cells, this type of endocytosis is dependent on plasma membrane cholesterol and dynamin activity. Flotillins, in turn, are shown to be responsible for the assembly of a particular type of rafts, which participate in clathrin- and caveolin-independent endocytosis (24). Coassembly of flotillin-1 and flotillin-2 into microdomains induces formation of membrane invaginations that are morphologically reminiscent of caveolae and considered to be capable of forming intracellular vesicles, which localize in the cortical cytoplasm and look similar to caveosomes but lack caveolin-1 (32). According to our earlier electron microscopy results, the vesicular pathway of TP/TP10-mediated protein uptake also resembled by morphology the noncaveolar, flotillin-1-dependent internalization

Bioconjugate Chem., Vol. 20, No. 5, 2009 883

mode described by Frick and co-workers. Therefore, we assessed the putative involvement of a flotillin-1-dependent internalization pathway in the cellular translocation of TP/TP10-protein complexes. However, somewhat unexpectedly we could not find the colocalization of CPP-protein complexes with flotillin-1 on the plasma membrane or inside cells by immunocytochemical analysis either by confocal or by electron microscopy. In addition, efficient down-regulation of flotillin-1 in the plasma membrane, using siRNA silencing, increased rather than decreased the CPP-protein uptake, corroborating the microscopy results and suggesting that the flotillin-1-dependent pathway has a negligible, if any, role in CPP-mediated protein delivery. The cholera toxin B subunit, used in this study for assessing the activity of the flotillin-dependent pathway, can enter cells via several endosomal pathways (22, 33). Therefore, CtxB could only with great circumspection be considered a marker of the flotillin-dependent endocytic route. Unfortunately, no specific cargo has been characterized for the flotillin-mediated pathway yet, making the explicit evaluation of the role of this mechanism difficult. For the sake of clarity, we also checked the localization of cell-transduced protein in relation to flotillin-2, but in analogy with flotillin-1 no significant overlap was detected (data not shown). On the other hand, based on the experiments with flotillin-1 and -2, we could not exclude the participation of lipid rafts in delivery by CPPs, as the nature of lipid rafts is not unambiguously characterized yet. Moreover, the possibility that complexes indeed show equal or even better uptake in partially flotillin-1-depleted cells suggests that the phospholipids of nonraft areas of the plasma membrane might also play a role in CPP internalization mechanisms (34, 35), independently or in concert with the cell surface proteoglycans. Considering the ability of transportan to modulate the dynamics of the membrane lipids (36), and taking into account that the location and function of cell surface proteoglycans is probably interfered in flotillin1-deficient cells, one can imagine that the changed properties of the plasma membrane could lead to the vesicular uptake mode(s) that is not common to cells with a properly organized membrane. The involvement of another cholesterol-dependent pathway, caveolar endocytosis, has been demonstrated in both the cellular uptake of Tat peptide and pTat-mediated cargo delivery (16, 17). Recent data about caveolae not just merely as a subtype of rafts but as an independent membrane structure (37-39), encouraged us to include it in the study. Our CLSM results demonstrated that approximately 30% of CPP-avidin complexes exhibited similar localization to caveolin, and a higher overlap was observed at the plasma membrane than inside cells. Still, electron microscopy revealed a somewhat lower degree of colocalization (15- 20%) of complexes with caveolin-positive vesicles, especially in the cortical cytoplasm where the main fraction of caveosomes was found. Electron microscopy results revealed that the CPP-neutravidin complexes and caveolin localized often in very close proximity, which might have led to the overestimation of the degree of colocalization by CLSM. However, we cannot exclude that the smaller ratio of colocalization detected in electron microscopy could be caused by different isoelectric points of avidin and neutravidin and by the higher affinity of avidin to negatively charged components of the plasma membrane. Still, after caveolin-1 down-regulation by 60-70%, a substantial inhibition of CPP-mediated protein uptake was detected by FACS analysis, which emphasizes the importance of caveolin-1-containing membrane areas in the internalization process. Moreover, the finding that the amount of caveosomes in cells incubated with CPP-protein complexes is about 2-fold higher than in untreated cells might suggest the induction of caveolae-dependent endocytosis, which contributes most to the TP/TP10-mediated protein delivery among the observed endosomal pathways. The

884 Bioconjugate Chem., Vol. 20, No. 5, 2009

Sa¨a¨lik et al.

Figure 5. (A) Effect of down-regulation of flotillin-1 and caveolin-1 by siRNA on protein cellular delivery by transportan and TP10. (A) Estimation of protein concentration after siRNA treatment by Western blot. Characteristic result of three separate experiments is depicted. Proteins (0.5 µg per lane) were detected with anti-Flot-1 pAb (upper panel), anti-Cav pAb (middle panel), or anti-β-actin mAb as loading control (lower panel). (B) Quantification of the flotillin-1 in the plasma membrane and intracellular structures by the analysis of z-sections of CLSM images of cells treated with nonspecific (gray bars) or Flot-1 specific siRNA (striped bars) and stained for Flot-1 and β1-integrin to visualize the plasma membrane. Results show the percentage of Flot-1-positive pixels of all pixels from the analyzed areas. (C, D) The uptake of complexes by siRNA-treated cells. HeLa cells were treated with nonspecific (gray bars) or specific siRNAs (striped bars) against Flot-1 (C), Cav-1 (D), or both Flot-1 and Cav-1 siRNAs (E) and incubated with complexes of biotinylated TP/TP10 (0.5 µM) and avidin-FITC (0.15 µM) for 30 min prior to FACS analysis. Protein delivery efficiency into cells treated with nonspecific siRNA is represented as 100%. Error bars represent standard deviation, and statistical relevance is characterized with P values.

confocal micrographs reveal that siRNA treatment led to strong reduction of caveolin-positive areas from the plasma membrane and marked loss from intracellular structures. However, the lack of caveolin-1 in plasma membrane inhibits the TP10-induced protein uptake more strongly than that of transportan, and this trend is also observable when the protein delivery by transportans is compared in wild type and caveolin-1 knockout MEFs. CLSM shows rather comparable colocalization between caveolin and avidin complexed with both transportans in normal HeLa cells. Electron microscopy, on the contrary, revealed a slightly higher occupancy of caveolin-coated vesicles by TP-neutravidin than by TP10 complexes, which is in discrepancy with the cellular uptake data in Cav-1 siRNA-treated cells. The results might suggest that although transportan-protein complexes occupy more caveosomes in cells than TP10, it could more flexibly switch between different endosomal pathways. Our earlier results demonstrated that about 10-15% of TP/ TP10-protein complexes localized in transferrin-containing vesicles showing the involvement of clathrin-mediated endocy-

tosis in the uptake of complexes both by fluorescence and by electron microscopy (8). In addition, the hyperosmolar medium known to interfere with the clathrin-mediated endocytosis inhibited the avidin delivery in HeLa cells to the same extent, corroborating the proportion of the clathrin-dependent mechanism in transportan-mediated protein delivery (14). Assuming good concordance of results obtained by three above-mentioned methods and relatively low contribution of the clathrin-dependent pathway in the complexes’ uptake, we did not apply a siRNA-included knock-down for evaluating the role of this pathway. The electron microphotos demonstrated that a substantial amount of cell-transduced protein was confined in macropinosome-like big vesicles (8), implicating that several endocytic pathways were active in CPP-mediated protein transduction. On the basis of this, we suggest that in conditions where the functioning of one entry route is restrained, TP/TP10 target the cargo into other/unimpaired endocytic pathways, although with smaller efficiency and depending on the features of the particular peptide. The further intracellular targeting of

Protein Delivery with Transportans

Bioconjugate Chem., Vol. 20, No. 5, 2009 885

Figure 7. Endocytic pathways used for cellular entry by transportan/ TP10-protein complexes. The scheme includes our earlier data (8). The complexes of peptide and protein (A) can enter cells via macropinocytosis (B), clathrin- (C) or caveolae (D)-mediated endocytosis, and clathrin- and caveolin-independent endocytosis (E). The flotillin-mediated pathway (F) is not utilized. Intracellular trafficking routes are indicated by solid arrows, putative directing is represented by dashed arrows. TGN-trans-Golgi network.

Figure 6. Internalization and localization of transportan/TP10-protein complexes in cells lacking caveolin-1. Wild-type or caveolin-1 knockout MEFs were incubated with complexes of biotinylated TP/ TP10 (0.5 µM) and avidin-FITC (0.15 µM) for 30 min and analyzed by FACS (A) or confocal microscopy (B-E). The protein uptake into wt MEFs is presented as 100% in A. (B-E) Cells were stained with anticaveolin pAb and AF 594 conjugated secondary Ab (red) for visualizing caveolins. The caveolin signal in caveolin-1 knockout cells is very weak and rather diffuse, representing the localization of residual caveolin-2 (C, E). Images shown are merged from all optical slices, and the enlarged sections with z-projections are taken from indicated places and presented in the right column. Arrowheads indicate the apical part of cells in z-projections, and the bar represents 10 µM.

CPP-transduced cargos is even more important than the starting point. Upon long-time incubations, the majority of cargo protein delivered into cells by CPPs accumulates in LAMP2-containing structures in the perinuclear region (8) and is targeted to digestion (3). Therefore, mapping of the early steps of CPPmediated delivery might provide valuable information for finding a way to induce liberation of cargo molecules from the endosomal structures before direction to degradation. The small GTPase Rab5 and PI3P are both typical constituents of early endosomes, and both of them are also reported to have role in macropinocytosis (40). In addition, the lateral transport between early endosomal vesicles and caveosomes has been shown to be Rab5-dependent (38). However, most of the CPP-protein complexes were observed to evade early endosomes since only a negligible fraction of complexes showed colocalization with Rab5 or PI3P after 10 or 30 min. Furthermore, the treatment of cells with wortmannin, a specific inhibitor of phosphoinositide 3-kinase, did not inhibit the CPP-induced protein transport into HeLa cells, which is in line with an earlier study by Payne et al. (41). Although early endosomes are reported to be capable of receiving contents from caveosomes (38), the fraction of TP/ TP10 protein complexes that internalized in caveolin-positive vesicles might avoid early endosomes by using other transport route(s), probably analogous to SV40 (42). Still, the caveolar pathway is considered to lead to the Golgi complex, but only a minor part of TP/TP10-protein complexes is targeted there as we showed earlier (8). Although the CPP-protein complexes seem to avoid being trapped into early endosomal structures after translocation into HeLa cells, upon longer incubation they accumulate in vesicles containing endolysosomal marker protein LAMP2 (8). Recently, it has been reported that at least three endocytic pathways are used for cellular entry by CPPs: macropinocytosis, clathrin-mediated, and caveolar/lipid raftmediated endocytosis (5). Our results on protein cellular delivery are well in line with this study suggesting analogous concurrently acting endocytic uptake mechanisms, and in our system the caveolin-dependent pathway but not the flotillin-1-mediated route is preferentially used. However, the intracellular targeting of complexes seems to avoid canonical trafficking pathways (schematic representation in Figure 7). Additionally, there is a fraction of small vesicles induced by CPPs and mediating the internalization of cargo molecules, which we have not been able to classify using markers of known endocytic pathways, and flotillin-1 did not help with filling this gap.

886 Bioconjugate Chem., Vol. 20, No. 5, 2009

ACKNOWLEDGMENT We thank M. Kure, T. Toivonen, M. Lindman, and A. Strandell for excellent technical assistance in electron microscopy, M. Ehandi for peptide synthesis, S. Ingerpuu for antiβ1-integrin antibody, and H. Ra¨a¨gel for critical reading of manuscript. The work was supported by grants from Estonian Science Foundation (ESF 5588 and 7058), European Community Program QLK3-CT-2002-01989, Swedish Research Council (VR-NT, VR-Med), Academy of Finland, and University of Tartu (PP1TI07902). Supporting Information Available: Localization of CPPavidin complexes (1) in relation to cellular flotillin-1 at low temperature, and (2) compared to cellular Rab5 and PI3P after short incubation time. This material is available free of charge via the Internet at http://pubs.acs.org

LITERATURE CITED ¨ . (2007) Cell-penetrating peptides. In Cell-penetrating (1) Langel, U peptides, pp 1-600, CRC Press, Boca Raton. (2) Ziegler, A., Nervi, P., Du¨rrenberger, M., and Seelig, J. (2005) The cationic cell-penetrating peptide CPP(TAT) derived from the HIV-1 protein TAT is rapidly transported into living fibroblasts: optical, biophysical, and metabolic evidence. Biochemistry 44, 138–48. (3) Fretz, M. M., Penning, N. A., Al-Taei, S., Futaki, S., Takeuchi, T., Nakase, I., Storm, G., and Jones, A. T. (2007) Temperature-, concentration- and cholesterol-dependent translocation of L- and D-octa-arginine across the plasma and nuclear membrane of CD34+ leukaemia cells. Biochem. J. 403, 335–42. (4) Tu¨nnemann, G., Martin, R. M., Haupt, S., Patsch, C., Edenhofer, F., and Cardoso, M. C. (2006) Cargo-dependent mode of uptake and bioavailability of TAT-containing proteins and peptides in living cells. FASEB J. 20, 1775–84. (5) Duchardt, F., Fotin-Mleczek, M., Schwarz, H., Fischer, R., and Brock, R. (2007) A comprehensive model for the cellular uptake of cationic cell-penetrating peptides. Traffic 8, 848–66. (6) Ter-Avetisyan, G., Tu¨nnemann, G., Nowak, D., Nitschke, M., Herrmann, A., Drab, M., and Cardoso, M. C. (2008) Cell entry of arginine-rich peptides is independent of endocytosis. J. Biol. Chem. 284, 3370-8. (7) Wadia, J. S., Stan, R. V., and Dowdy, S. F. (2004) Transducible TAT-HA fusogenic peptide enhances escape of TAT-fusion proteins after lipid raft macropinocytosis. Nat. Med. 10, 310– 315. (8) Padari, K., Sa¨a¨lik, P., Hansen, M., Koppel, K., Raid, R., Langel, ¨ ., and Pooga, M. (2005) Cell transduction pathways of U transportans. Bioconjugate Chem. 16, 1399–1410. ¨. (9) El-Andaloussi, S., Johansson, H. J., Holm, T., and Langel, U (2007) A Novel Cell-penetrating Peptide, M918, for Efficient Delivery of Proteins and Peptide Nucleic Acids. Mol. Ther. 10, 1820–6. (10) Sigismund, S., Woelk, T., Puri, C., Maspero, E., Tacchetti, C., Transidico, P., Di Fiore, P. P., and Polo, S. (2005) Clathrinindependent endocytosis of ubiquitinated cargos. Proc. Natl. Acad. Sci. U.S.A. 102, 2760–5. (11) Di Guglielmo, G. M., Le Roy, C., Goodfellow, A. F., and Wrana, J. L. (2003) Distinct endocytic pathways regulate TGFbeta receptor signalling and turnover. Nat. Cell Biol. 5, 410–21. (12) Lakadamyali, M., Rust, M. J., and Zhuang, X. (2006) Ligands for clathrin-mediated endocytosis are differentially sorted into distinct populations of early endosomes. Cell 124, 997–1009. (13) Richard, J. P., Melikov, K., Vive`s, E., Ramos, C., Verbeure, B., Gait, M. J., Chernomordik, L. V., and Lebleu, B. (2003) Cellpenetrating peptides. A reevaluation of the mechanism of cellular uptake. J. Biol. Chem. 278, 585–90.

Sa¨a¨lik et al. (14) Sa¨a¨lik, P., Elmquist, A., Hansen, M., Padari, K., Saar, K., ¨ ., and Pooga, M. (2004) Protein cargo delivery Viht, K., Langel, U properties of cell-penetrating peptides. A comparative study. Bioconjugate Chem. 15, 1246–53. (15) Parton, R. G., and Simons, K. (2007) The multiple faces of caveolae. Nat. ReV. Mol. Cell Biol. 8, 185–94. (16) Fittipaldi, A., Ferrari, A., Zoppe, M., Arcangeli, C., Pellegrini, V., Beltram, F., and Giacca, M. (2003) Cell membrane lipid rafts mediate caveolar endocytosis of HIV-1 Tat fusion proteins. J. Biol. Chem. 278, 34141–34149. (17) Ferrari, A., Pellegrini, V., Arcangeli, C., Fittipaldi, A., Giacca, M., and Beltram, F. (2003) Caveolae-mediated internalization of extracellular HIV-1 tat fusion proteins visualized in real time. Mol. Ther. 8, 284–294. (18) Norbury, C. C. (2006) Drinking a lot is good for dendritic cells. Immunology 117, 443–51. (19) Nakase, I., Tadokoro, A., Kawabata, N., Takeuchi, T., Katoh, H., Hiramoto, K., Negishi, M., Nomizu, M., Sugiura, Y., and Futaki, S. (2007) Interaction of arginine-rich peptides with membrane-associated proteoglycans is crucial for induction of Actin organization and macropinocytosis. Biochemistry 46, 492– 501. (20) Gerbal-Chaloin, S., Gondeau, C., Aldrian-Herrada, G., Heitz, F., Gauthier-Rouviere, C., and Divita, G. (2007) First step of the cell-penetrating peptide mechanism involves Rac1 GTPasedependent Actin-network remodelling. Biol. Cell 99, 223–38. (21) Damm, E. M., Pelkmans, L., Kartenbeck, J., Mezzacasa, A., Kurzchalia, T., and Helenius, A. (2005) Clathrin- and caveolin1-independent endocytosis: entry of simian virus 40 into cells devoid of caveolae. J. Cell Biol. 168, 477–88. (22) Kirkham, M., Fujita, A., Chadda, R., Nixon, S. J., Kurzchalia, T. V., Sharma, D. K., Pagano, R. E., Hancock, J. F., Mayor, S., and Parton, R. G. (2005) Ultrastructural identification of uncoated caveolin-independent early endocytic vehicles. J. Cell Biol. 168, 465–76. (23) Janich, P., and Corbeil, D. (2007) GM1 and GM3 gangliosides highlight distinct lipid microdomains within the apical domain of epithelial cells. FEBS Lett. 581, 1783–7. (24) Glebov, O. O., Bright, N. A., and Nichols, B. J. (2006) Flotillin-1 defines a clathrin-independent endocytic pathway in mammalian cells. Nat. Cell Biol. 8, 46–54. (25) Rajendran, L., Le Lay, S., and Illges, H. (2007) Raft association and lipid droplet targeting of flotillins are independent of caveolin. Biol. Chem. 388, 307–14. (26) Arai, R., Geffard, M., and Calas, A. (1992) Intensification of labelings of the immunogold silver staining method by gold toning. Brain Res. Bull. 28, 343–5. (27) Nichols, B. J. (2002) A distinct class of endosome mediates clathrin-independent endocytosis to the Golgi complex. Nat. Cell Biol. 4, 374–8. (28) Razani, B., Engelman, J. A., Wang, X. B., Schubert, W., Zhang, X. L., Marks, C. B., Macaluso, F., Russell, R. G., Li, M., Pestell, R. G., Di Vizio, D., Hou, H., Jr., Kneitz, B., Lagaud, G., Christ, G. J., Edelmann, W., and Lisanti, M. P. (2001) Caveolin-1 null mice are viable but show evidence of hyperproliferative and vascular abnormalities. J. Biol. Chem. 276, 38121–38. (29) Foerg, C., Ziegler, U., Fernandez-Carneado, J., Giralt, E., Rennert, R., Beck-Sickinger, A. G., and Merkle, H. P. (2005) Decoding the entry of two novel cell-penetrating peptides in HeLa cells: lipid raft-mediated endocytosis and endosomal escape. Biochemistry 44, 72–81. (30) Rodal, S. K., Skretting, G., Garred, O., Vilhardt, F., van Deurs, B., and Sandvig, K. (1999) Extraction of cholesterol with methylbeta-cyclodextrin perturbs formation of clathrin-coated endocytic vesicles. Mol. Biol. Cell 10, 961–74. (31) Mondal, M., Mesmin, B., Mukherjee, S., and Maxfield, F. R. (2009) Sterols are mainly in the cytoplasmic leaflet of the plasma membrane and the endocytic recycling compartment in CHO cells. Mol. Biol. Cell 20, 581–8.

Protein Delivery with Transportans (32) Frick, M., Schmidt, K., and Nichols, B. J. (2007) Modulation of lateral diffusion in the plasma membrane by protein density. Curr. Biol. 17, 462–7. (33) Torgersen, M. L., Skretting, G., van Deurs, B., and Sandvig, K. (2001) Internalization of cholera toxin by different endocytic mechanisms. J. Cell. Sci. 114, 3737–3747. (34) Lamaziere, A., Burlina, F., Wolf, C., Chassaing, G., Trugnan, G., and Ayala-Sanmartin, J. (2007) Non-metabolic membrane tubulation and permeability induced by bioactive peptides. PLoS ONE 2, e201. (35) Yandek, L. E., Pokorny, A., Flore´n, A., Knoelke, K., Langel, Ü., and Almeida, P. F. (2007) Mechanism of the cell-penetrating peptide transportan 10 permeation of lipid bilayers. Biophys. J. 92, 2434–44. (36) Barany-Wallje, E., Andersson, A., Gra¨slund, A., and Ma¨ler, L. (2006) Dynamics of transportan in bicelles is surface charge dependent. J. Biomol. NMR 35, 137–47. (37) Foster, L. J., De Hoog, C. L., and Mann, M. (2003) Unbiased quantitative proteomics of lipid rafts reveals high specificity for signaling factors. Proc. Natl. Acad. Sci. U.S.A. 100, 5813–8.

Bioconjugate Chem., Vol. 20, No. 5, 2009 887 (38) Pelkmans, L., Burli, T., Zerial, M., and Helenius, A. (2004) Caveolin-stabilized membrane domains as multifunctional transport and sorting devices in endocytic membrane traffic. Cell 118, 767–80. (39) Sprenger, R. R., Fontijn, R. D., van Marle, J., Pannekoek, H., and Horrevoets, A. J. (2006) Spatial segregation of transport and signalling functions between human endothelial caveolae and lipid raft proteomes. Biochem. J. 400, 401–10. (40) Schnatwinkel, C., Christoforidis, S., Lindsay, M. R., Uttenweiler-Joseph, S., Wilm, M., Parton, R. G., and Zerial, M. (2004) The Rab5 effector Rabankyrin-5 regulates and coordinates different endocytic mechanisms. PLoS Biol. 2, E261. (41) Payne, C. K., Jones, S. A., Chen, C., and Zhuang, X. (2007) Internalization and trafficking of cell surface proteoglycans and proteoglycan-binding ligands. Traffic 8, 389–401. (42) Nichols, B. J., Kenworthy, A. K., Polishchuk, R. S., Lodge, R., Roberts, T. H., Hirschberg, K., Phair, R. D., and LippincottSchwartz, J. (2001) Rapid cycling of lipid raft markers between the cell surface and Golgi complex. J. Cell Biol. 153, 529–41. BC800416F