Mechanism of Uptake of C105Y, a Novel Cell-penetrating Peptide*

C105Y, a synthetic peptide (CSIPPEVKFNKPFVYLI) based on the amino acid sequence corresponding to residues 359-374 of α1-antitrypsin, enhances gene expression from DNA nanoparticles. To investigate how this enhancement occurs, C105Y was fluorescently labeled to study its uptake and intracellular trafficking. When human hepatoma cells (HuH7) were incubated with fluorescently labeled C105Y for as little as 3 min, C105Y displayed nuclear and cytoplasmic staining with enrichment of fluorescent signal in the nucleus and nucleolus. Uptake and nucleolar localization were observed with the short sequence PFVYLI, but not with SIPPEVKFNK, and the D-isomer was readily taken up into cells but not into the nucleus. We found that the C105Y peptide is routed to the nucleolus very rapidly in an energy-dependent fashion, whereas membrane translocation and nuclear localization are energy-independent. When we tested the involvement of known endocytosis pathways in uptake and trafficking of this peptide, we demonstrated that C105Y peptide is internalized by a clathrin- and caveolin-independent pathway, although lipid raft-mediated endocytosis may play a role in peptide intracellular trafficking. Efficient energy-independent cell entry with rapid nuclear localization probably accounts for enhancement of gene expression from inclusion of C105Y into DNA nanoparticles.

for their ability to induce translocation, a third class of CPPs has been identified, which needs further characterization (2). Although CPPs are derived from a variety of sources, they have some common features. Some peptides are ␣-helical in structure, and some contain arginine or lysine rich motifs, whereas others have a hydrophobic core sequence. It appears that these motifs are important in mediating the translocation of CPPs across cell membranes. For example, the hydrophilic arginine residues within TAT appear to mediate its cell penetrating properties, whereas the hydrophobic core of penetratin is necessary to mediate its entry into cells (2,8).
Proteins, peptides, and small molecules can enter cells through a variety of mechanisms. For example, small molecules such as amino acids, sugars, and ions can traverse the plasma membrane through pumps or channels. Macromolecules, however, must be transported into the cell by endocytosis. Endocytosis occurs by multiple mechanisms that fall into two distinct categories. Phagocytosis involves the uptake of large particles and is usually restricted to specialized mammalian cells, such as macrophages and neutrophiles. Pinocytosis occurs in all cell types and involves the uptake of fluids and solutes. Pinocytosis is mediated by four identified mechanisms: macropinocytosis, clathrinmediated endocytosis, caveolae-mediated endocytosis, and clathrin-, caveolae-, and dynamin-independent endocytosis (9). It is not fully understood how CPPs enter cells, but the process seems to differ from peptide to peptide. In general, cells rapidly internalize most CPPs in an energy-independent manner, and their uptake is unaffected by most inhibitors of classical endocytic pathways. Furthermore, it does not appear that these peptides are internalized by a receptor-mediated process, because identical transduction properties are observed for their D-isomers (1). These data suggest that many CPPs enter cells by a nonclassical endocytic pathway. Recently published data suggest that some CPPs enter cells by fluid phase macropinocytosis, a specialized form of endocytosis that is independent of caveolae, clathrin, and dynamin (10,11). Consequently, it appears that some peptides may use a distinct mechanism of translocation or that they have the ability to enter cells through alternative pathways (12). Such a pathway has been suggested for penetratin, which enters cells not by an endocytic pathway but by translocation across lipid bilayers (13).
CPPs have tremendous potential in biotechnology because of their ability to increase uptake and/or nuclear targeting of biologically active proteins or molecules by cells. CPPs may be able to deliver a wide variety of therapeutic compounds. A variety of cargoes have been successfully linked to the protein transduction domains of CPPs, including small and intermediate molecules, peptides, full-length proteins and liposomes (14). To achieve this potential, it will be important to understand how CPPs cross a biological membrane and localize to specific intracellular compartments.
Work in our laboratory has focused on developing efficient ways to deliver DNA into cells for gene therapy. We have used poly-L-lysine (polyK) condensed DNA complexes to effectively deliver DNA to cells both in vitro and in vivo. To target and increase cellular uptake of these complexes, we covalently attached a small peptide ligand, called C105Y, to polyK before DNA condensation. This peptide was initially identified as an amino acid sequence within the C-terminal tail of ␣1-antitrypsin (␣1-AT) that could mediate binding to the serpin enzyme complex receptor (sec-R). Presumably, when ␣1-AT binds to elastase, it undergoes conformational change, exposing this peptide sequence, which is then available to its receptor to promote cellular uptake of the ␣1-AT⅐elastase complex (15). This receptor was characterized mainly by biochemical means as a cell surface binding site on human hepatoma (HepG2) cells and blood monocytes and, as of yet, has not been cloned. The synthetic peptide ligand C105Y, based on amino acids 346 -374 of ␣1-AT, competed with the natural ligand for binding to the sec-R and provided gene transfer into cells when it was conjugated polyK⅐DNA complexes (16,17). These particles, which average 25 nm in diameter by electron microscope, can provide gene transfer of the cystic fibrosis transmembrane regulator gene in vivo to airway epithelial cells from cystic fibrosis transmembrane regulator-deficient mice. This gene transfer is sufficient to partially correct the chloride transport defect and to reverse the down-regulation of NOS-2 that is also a characteristic of cystic fibrosis mouse nasal epithelium (18). Other studies in mice show significant enhancement of gene transfer in the lung, liver, and spleen by inclusion of C105Y in the complex (19). In vitro studies show that the enhancement depends on the density of ligand substitution as well as chain length of the poly-L-lysine. In these studies C105Y increased both the intensity and the duration of reporter gene expression (20,21).
To determine why conjugating C105Y peptide to polyK condensed DNA complexes increased the efficacy of gene transfer, we investigated the mechanism by which this peptide enters cells. C105Y by itself could penetrate the cell membrane rapidly, enter the cytoplasm, and localize to the nucleus and nucleolus of live HuH7 cells. In this study we describe the mechanism by which C105Y enters cells and travels to the nucleus. Our data suggest that C105Y does not undergo receptor binding but rather is a novel cell-penetrating peptide.
Peptide Synthesis and Labeling-Fluorescein-labeled peptides and unlabeled C105Y were synthesized by solid phase method, purified, and subjected to high pressure liquid chromatography and matrix-assisted laser desorption ionization time-of-flight mass spectroscopy by BIO-WORLD (Dublin, OH) and the Cleveland Clinic Foundation (Cleveland, OH), respectively. The peptides were dissolved in 5% Me 2 SO and PBS, and the stocks were periodically tested by matrix-assisted laser desorption ionization for peptide integrity. To attach AF-594 to unlabeled C105Y, peptide was dissolved in 0.1 M sodium bicarbonate buffer, pH 8.3, and incubated with AF-594 labeling reagent for 1 h at room temperature. The conjugate was separated from unreacted labeling reagent using Sephadex G-25 columns.
Cells and Cell Culture-Human Hepatoma cells (HuH7) were maintained as previously described in RPMI medium (17) with 10% fetal calf serum (FCS). HuH7 cells were previously shown to bind the sec-R ligand C105Y (17). Fresh medium was added every second day. C6 and C6-Cav-1 glioma cells were provided by Bryan L. Roth (Case Western Reserve University) and were maintained in RPMI medium containing 10% FCS alone or 10% FCS and 0.6 mg/ml G418, respectively (22). 3T3 cells were maintained in Dulbecco's modified Eagle's medium containing 10% FCS. Primary HTE cells were isolated from human tracheas, grown on filters in an air liquid interface, and maintained in medium containing Dulbecco's modified Eagle's medium and Ham's F-12 medium in a 1:1 ratio substituted with 2% Ultrose G.
Stable Cell Line Generation-Stable cell lines HuH7-GFP and HuH7-⌬Eps15 expressing EGFP or EGFP-DIII, respectively, were generated by transfecting 70% confluent HuH7 cells grown on standard six-well tissue culture plates with 1.5 g/well EGFP or EGFP-DIII DNA combined with FuGENE. DNA plasmids encoding EGFP or EGFP-DIII were given to us by Dr. Benmerah from (Faculté Necker-Enfants Malades) (23). Forty-eight hours after transfection, the cells were washed and supplemented with media containing G418 (0.65 mg/ml) as the selectable marker. The surviving colonies were examined for GFP expression by fluorescence microscopy and grown to confluence before further passages.
Peptide Uptake and Colocalization Studies-For experiments on peptide uptake, HuH7, HuH7-GFP, HuH7-⌬Eps15, 3T3, HTE, C6, and C6-Cav-1 cells were grown to 50 -70% confluence on chambered coverglass slides prior to incubation with peptide. To study intracellular peptide localization 1 M FL-C105Y, PFVYLI, SCRAM, ⌬FVYLI, or D-C105Y were incubated with HuH7 cells or 5 min in PBS with Ca 2ϩ / Mg 2ϩ at room temperature. The cells were fixed with 2% paraformaldehyde and analyzed by fluorescence microscopy. To study the effect of low temperature on the cellular uptake of C105Y, HuH7 cells were preincubated at 4°C for 1 h in serum-free medium. The cells were then coincubated with AF-594-C105Y and AF-488-transferrin for 30 min at 4°C. The cell were then washed with PBS and either fixed in 2% paraformaldehyde or further incubated in serum-free medium for 30 min at 37°C. For colocalization studies, the cells were preincubated with fluorescent live cell markers in serum-free medium at 37°C for 15 min followed by addition of fluorescently labeled peptide for 5 min in the same medium at room temperature. Subsequently the cells were fixed and analyzed by fluorescence microscopy.
Fluorescence Microscopy-For analysis of the subcellular localization of C105Y in live cells, HuH7 cells were cultured on coverslips and mounted into a diffusion chamber. FITC-C105Y in PBS was continuously perfused into the chamber for 1 min followed by PBS. The cells were imaged by wide field fluorescent microscopy over a period of 20 min starting with peptide addition. At 20 min, time lapse acquisition was stopped, and a z-stack was acquired to determine peptide location.
Alternatively, the subcellular localization of C105Y was evaluated in fixed cells. The cells to be processed for immunofluorescence were washed with PBS and fixed for 5 min at room temperature with 2% paraformaldehyde in PBS. All of the samples were examined in dual channels using a Zeiss Axiovert 200 utilizing wide field microscopy with Autoquant deconvolution software for analysis, because of the superior signal to noise ratio of wide field microscopy images in comparison with laser scanning confocal microscopy (24). Acquired z1-sections were processed using Hygens confocal deconvolution software.
Antibody Staining-FITC-C105Y (1 M) was added to four-well glass slides containing HuH7 cells plated 2 days prior. After 5 min of incubation of cells with FITC-C105Y peptide at 21°C, the cells were washed and fixed with 2% paraformaldehyde. The cells were then permeabilized with 0.1% Nonidet P-40 for 30 min before blocking with 1% bovine serum albumin in PBS and the addition of primary antibody in 1% bovine serum albumin for 1 h. Following this, the cells were washed and incubated with Texas Red-labeled secondary antibody.
Western Blot Analysis-For Western blot analysis, 20 g of total cell lysate prepared from HuH7, C6, and C6-Cav-1 cells were separated by SDS-PAGE and transferred to polyvinylidene difluoride membrane (Bio-Rad). The primary antibodies were amplified by HRP-conjugated secondary antibodies and detected with Pico Enhanced Chemiluminescence (Pierce).

RESULTS
The peptides used in this study are shown in Table 1. C105Y is a fluorescently labeled, 17-amino acid sequence based on residues 346 -374 of sec-R. PFVYLI is a synthetic peptide that represents the C-terminal portion of C105Y and includes the proposed binding sequence to ␣1-AT. SCRAM is a peptide with the scrambled binding sequence of C105Y (ILVFY). ⌬FVYLI is a C-terminal truncation mutant designed to eliminate the proposed binding sequence of C105Y. D-C105Y is the D-isomer of C105Y, which was synthesized with D-amino acids to determine the specific structural requirements of C105Y uptake.
C105Y Peptide Localizes to the Nucleus and Nucleolus in Fixed HuH7 Cells and Other Cell Types-To determine whether C105Y peptide alone can enter different cell types, the peptide was fluorescently labeled and added to cultured human hepatoma (HuH7), mouse fibroblast (3T3), and primary HTE cells. Surprisingly, after only 5 min of incubation, C105Y peptide ( Fig. 1, A-C) was not only present in the cytoplasm but showed enrichment of fluorescent signal in the nucleus and nucleolus. This pattern was more pronounced in HuH7 cells compared with 3T3 and HTE cells. To locate C105Y in the nucleolus, we stained HuH7 cells with an antibody against the nucleolar protein nucleolin (Fig. 1E), which is normally present in the nucleolus and sometimes in the cytoplasm of cells. The orange color in the merged image (Fig. 1F) indicates that C105Y colocalizes with the nucleolin present in the nucleolus. These data, taken together, suggest that C105Y can enter the nucleus and nucleolus in several different cell types very rapidly.
C105Y Peptide Is Rapidly Internalized by Live HuH7 Cells-The data in fixed cells suggest that C105Y can rapidly enter cells, a characteristic of most CPPs. However, recent evidence suggests that cell fixation, even under mild conditions, can lead to an artificial uptake or redistribution of these peptides (25). To confirm that C105Y can rapidly enter cells and that our previous observation was not due to an artifact of fixation, FL-C105Y was added to live HuH7 cells. The cells were visualized by live cell fluorescent microscopy, and images were taken every 20 s for the first 5 min and every 2 min thereafter (Fig. 2). These time lapse images revealed that C105Y peptide is internalized by HuH7 cells within a few seconds and reaches the nucleus and nucleolus as rapidly as 3 min.
Therefore, the ability of C105Y to penetrate cells and localize to the nucleus and nucleolus is not an artifact of fixation, because C105Y can enter living cells.
Sequence-dependent Uptake of C105Y Peptide-Although cells internalize C105Y peptide, it was unknown which part of the C105Y sequence was responsible for its internalization and nuclear localization. Previous work by others indicates that a pentapeptide sequence in the C-terminal portion of ␣1-AT (FVFLM) was necessary to bind the putative sec-R (15). A synthetic analog of this peptide with sequence FVYLI could block binding and internalization of ␣1-AT complexes. When the amino acid sequence of these synthetic peptides were altered by mutating specific residues or scrambling amino acid sequences, they were unable to block the binding and internalization of ␣1-AT complexes. These data suggested that the binding of ␣1-AT complexes to the sec-R was sequence-specific (16). Therefore, we hypothesized that because the synthetic peptide PFVYLI could bind the putative sec-R and block the internalization of ␣1-AT complexes, it would also be able to enter cells similarly to C105Y. Additionally, we wanted to determine whether altering this putative binding sequence would abolish the ability of C105Y to enter cells. Consequently, we engineered a scrambled mutant (SCRAM) and a truncated peptide (⌬FVYLI) that lacks the proposed binding sequence to determine whether these peptides could be internalized and trafficked to the nucleus. Amino acid sequences for the above mentioned peptide can be found in Table 1.
The previously reported pentapeptide-binding sequence PFVYLI (Fig. 3A) and the scrambled mutant SCRAM (Fig. 3D) were internalized by HuH7 cells and displayed a staining pattern identical to C105Y (Fig.  1A). However, the peptide lacking the pentapeptide-binding sequence (⌬FVYLI) was not taken up by HuH7 cells (Fig. 3G). Therefore, it appears that the C-terminal amino acids of C105Y (PFVYLI) alone can mediate cellular uptake and nuclear targeting of C105Y. Interestingly neither cellular uptake nor nuclear localization of the scrambled mutant is impaired, which suggests that the order of amino acids within the C-terminal region does not influence cell penetration and nuclear localization. This is in contrast to prior data for putative sec-R ligands and suggests that this binding site may not be accessed in this model or at least does not account for most of the uptake.

TABLE 1 Primary structures of peptides used in this study
All of the peptides were synthesized as peptide amides and fluorescently labeled at the N terminus.  JANUARY 13, 2006 • VOLUME 281 • NUMBER 2

JOURNAL OF BIOLOGICAL CHEMISTRY 1235
Uptake of the C105Y D-isomer-To further test whether C105Y accessed HuH7 cells via a receptor mediated pathway or as a cell-penetrating peptide, we synthesized the D-isomer of C105Y. Previously published data suggest that D-isomers of the TAT CPP were taken up as efficiently as the native peptide (26), indicating that TAT entered cells in a receptor-independent fashion. FL-D-C105Y accumulated in the cytoplasm of HuH7 cells as outlined by FM4 -64 (Fig. 4B), a marker for endocytic vesicles, but did not enter the nucleus or nucleolus (Fig. 4A). However, although cytoplasmic accumulation of D-C105Y occurs, its pattern differs from the distribution pattern of C105Y, SCRAM, and PFVYLI in the cytoplasm. These data indicate that no specific spatial orientation of the amino acids comprising the hydrophobic C terminus is needed for cell penetration. This provides further evidence that cellular uptake is not a receptor-mediated event. However, nuclear and nucleolar targeting appears to be stereo-selective.
C105Y Nucleolar Localization, but Not Internalization, Is an Energydependent Process-We wanted to elucidate the mechanism by which C105Y enters cells and determine whether uptake is energy-dependent.
Receptor-mediated endocytosis through clathrin-coated pits is an energy-dependent process that can be inhibited by incubating cells at 4°C. To determine whether the import of C105Y into cells is energydependent, FL-C105Y peptide was incubated with HuH7 cells at 4°C. As a control, AF-488-transferrin, which is internalized through clath-

. C105Y rapidly internalizes into live HuH7 cells.
HuH7 cells were cultured on coverslips and mounted into diffusion chambers. FITClabeled C105Y in PBS was continuously perfused into the chamber followed by PBS. The cells were imaged by wide field microscopy over a period of 20 min starting with peptide addition. At 20 min, time lapse acquisition was stopped, and a z-stack was acquired to determine peptide location. rin-coated pits, an energy-dependent endocytic process, was added to cells as well.
At 4°C, HuH7 cells internalized FL-C105Y into the cytoplasm and nucleus. However, no FL-C105Y was observed in the nucleolus at 4°C (Fig. 5A). In contrast, endocytosis of transferrin was completely inhibited at 4°C, and the protein remained bound to its receptor on the cell surface (Fig. 5B). Therefore, it appears that C105Y enters cells and even the nucleus in an energy-independent process. To determine whether peptide localization to the nucleolus was energy-dependent, HuH7 cells were first incubated with FL-C105Y at 4°C; excess peptide in the medium was removed and replaced with fresh media to be incubated at 37°C. After incubation at 37°C, C105Y was present in the nucleolus (Fig. 5D), suggesting that nucleolar uptake is energy-dependent. Similarly, transferrin was endocytosed by cells at 37°C and localized within the cytoplasm. (Fig. 5E). Therefore, it appears that C105Y does not enter cellsbyanenergy-dependentprocess.However,C105Yrequiresanenergydependent process to enter the nucleolus.
C105Y Peptide Is Internalized by a Clathrin-independent Pathway-Endocytosis can occur through a variety of mechanisms, each providing a distinct pathway for the selective uptake of molecules by cells. These distinct pathways include clathrin-mediated, caveolae/lipid raft-mediated, and clathrin-and caveolae-independent endocytosis. The classically described route for entry of proteins into cells is via clathrin-coated pits, whereby proteins bind to specific receptors on the cell surface, and these ligand⅐receptor complexes cluster and are internalized in an energy-dependent manner by clathrin-coated membrane vesicles (27). To examine the role of clathrin-coated pits in C105Y uptake, stable HuH7-GFP and HuH7-⌬Eps15 cell lines were created expressing either EGFP or EGFP-DIII, respectively (Fig. 6, B and E). The EGFP-DIII plasmid construct was created by Benmerah et al. (23) and encodes a partial Eps15 sequence that is normally required for the early steps of clathrindependent endocytosis. The EGFP-DIII construct has a deletion in the C-terminal domain of Eps15, which contains AP-2-binding regions important for targeting AP-2 to the clathrin-coated pits. It was shown that transfection of this construct into HeLa cells (23) and HuH7 cells (data not shown) inhibited transferrin uptake and thus interfered with clathrin-coated pit formation. When fluorescently labeled C105Y was added to either of these stable cell lines, it was rapidly internalized and routed to the nucleus and nucleolus (Fig. 6, A and D), suggesting that clathrin-coated pits are not involved in C105Y uptake.
C105Y Internalization Occurs Independent of Caveoli-Cholesterolrich domains on the plasma membrane, generally referred to as "lipid rafts," are small structures that freely diffuse on the cell surface. These lipid rafts are subdivided into two distinct categories: caveolin-containing lipid rafts, called caveoli, and clathrin-and caveolin-independent lipid rafts. Caveoli are flask-shaped, noncoated invaginations on the plasma membrane, which contain many diverse signaling molecules and membrane transporters. Clathrin-and caveolin-independent lipid rafts are similar to caveolae in structure but lack the protein caveolin. Additionally, macropinocytosis is a rapid lipid raft-dependent and receptor-independent form of endocytosis. Each of these aforementioned structures can enter cells through an endocytic mechanism. Human hepatoma cells express very low levels of caveolin-1 (Cav-1) and Cav-2 yet rapidly take up C105Y, suggesting a role of noncaveolar raft domains in C105Y uptake and/or trafficking (28). To confirm this, HuH7 cell lysates were immunoblotted with anti-caveolin-1 and anticaveolin-2 antibodies and compared with cell lysates from C6 glioma cells (Fig. 7). Additionally, we tested C6-Cav-1 glioma cells, which exhibit reduced but not absent levels of Cav-1 and Cav-2 compared with wild type C6 cells. The C6-Cav-1 knockdown cells line was created Bhatnagar et al. (22) to stably knock down Cav-1 expression utilizing a DNA vector-based small interfering RNA. When FL-C105Y was added to C6 and C6-Cav-1 cells, it was internalized rapidly, localizing predominantly to the nuclei and nucleoli of these cells (Fig. 7, A and B), with an identical staining pattern to HuH7 cells (Fig. 1). These data, taken together, indicate that C105Y uptake and trafficking occur independent of caveolae.
Endocytosis Is Not Involved in Uptake of C105Y-To investigate the possible involvement of different endocytic pathways in the cellular uptake of C105Y peptide, we tested the effect of chlorpromazine (a known inhibitor of clathrin-mediated endocytosis), methyl-␤-cyclodextrin (which depletes membrane cholesterol, inhibiting pathways dependent on lipid rafts), cytochalasin D (an inhibitor of F-actin elongation required for macropinocytosis and caveolar endocytosis), and nocodazole (which inhibits microtubule formation). Among all these tested drugs only methyl-␤-cyclodextrin (M␤CD) had an inhibitory effect on C105Y localization to the nucleolus (Fig. 8D). For all other drugs C105Y peptide (Fig. 8, A-C) exhibited an identical staining pattern to the control (data not shown). Control experiments were performed to address the effects of these drugs with transferrin, a marker of clathrin-mediated endocytosis. As expected, transferrin uptake was inhibited by chlorpromazine, cytochalasin D, and nocodazole (Fig. 8, E-G), but not M␤CD (Fig. 8H). Taken together, these results confirm that endocytic pathways are not involved in C105Y peptide uptake but may involve lipid rafts or other endocytic pathways that are inhibited by M␤CD for nucleolar targeting.

Intracellular Trafficking of C105Y Occurs by an Endocytic
Mechanism-Pinocytosis (including macropinocytosis), the uptake of fluid and solutes by cells, can be measured by the intracellular accumulation of tracer molecules. This process is conserved in all eukaryotic cells and is required for diverse cellular functions. We wanted to determine whether the intracellular trafficking of C105Y might be facilitated through endocytic vesicles or lipid rafts, as suggested by the effect of M␤CD on C105Y nucleolar localization. Consequently, HuH7 cells were coincubated with C105Y (Fig. 9A) and a fluid phase marker of endocytosis: FM1-43X (Fig. 9B). The merged image (Fig. 9C) shows that there is overlap between the cytoplasmic structures stained by C105Y and the cytoplasmic membrane-bound vesicles stained by the FM1-43X in the cytoplasm. Furthermore, colocalization of C105Y with FM1-43X was also apparent in C6 and C6-Cav-1 cells (Fig. 7, E and F). This suggests that C105Y peptide may utilize endocytic vesicles stained by FM1-43X to traffic inside the cell.

DISCUSSION
In this study we report that C105Y is a novel cell-penetrating peptide, based on its ability to enter the cytoplasm, nucleus, and nucleolus of live cells very rapidly. Although the mechanism of cell penetration by CPPs is not fully understood, in general these peptides enter cells rapidly in a receptor-independent fashion. Most of the previously identified CPPs share one of several motifs: either multiple positive charges, hydrophobicity, or ␣-helical structure. Many CPPs, such as TAT peptide, penetrate cell membranes via positively charged residues, such as arginines, but others, such as penetratin, translocate into cells via its central hydrophobic core (26,29). It appears that C105Y enters cells through its hydrophobic C-terminal sequence (PFVYLI), because this six-amino acid sequence itself can enter cells and localize to the nucleus in an identical fashion to full-length C105Y, and the peptide lacking this sequence (⌬FVYLI) fails to internalize into cells. Furthermore, the order of the amino acids present in the hydrophobic core is not critical, because the scrambled C105Y mutant PILVFY behaves similarly to the authentic C105Y and PFVYLI. These data further support the notion that uptake of C105Y is not via a specific receptor, depending only on the biochemical properties of their sequences and not the order of their amino acid residues.
Earlier studies on cellular translocation of CPPs suggested that none of the classical receptor-, transporter-, or endocytosis-mediated processes are involved in their uptake. Instead it was proposed that the uptake of CPPs might occur through either direct membrane penetration or through the formation of inverted micelles (30). However, other recent evidence suggests that some endocytic process can be involved in the uptake of CPPs. For example the TAT-derived peptide is first endocytosed into a vesicular compartment before release into the cytoplasm (8). Several endocytic pathways exist that may mediate the internalization and intracellular trafficking of CPPs. These include endocytosis via clathrin-coated pits or endocytosis via lipid rafts. C105Y uptake and internalization does not occur through known endocytic pathways. It was particularly important to investigate clathrin-mediated endocytosis for C105Y, because it was originally regarded as a high affinity ligand for such a receptor, the sec-R. Low temperature, chlorpromazine, cytochalasin D, nocodazole, stable cell lines expressing EGFP-DIII, or use of the D-isomer did not interfere with C105Y peptide uptake, indicating no participation of receptor-and clathrin-mediated endocytosis. Furthermore, C105Y peptide can enter cells with or without caveolin-1 and FIGURE 6. C105Y can enter stable cell lines that have impaired clathrin-mediated endocytosis. Stable HuH7-GFP and HuH7-⌬Eps15 cells expressing either EGFP (B) or EGFP-DIII (E), a mutant that has been shown to interfere with clathrin-coated pit formation, were incubated with AF-596-labeled C105Y (A and D) for 5 min before fixation and analysis by fluorescence microscopy with deconvolution software. The merged images of GFP expression (green) and C105Y (red) are shown C and F. caveolin-2, so caveolin-mediated endocytosis is not involved. The aforementioned data are summarized in Table 2.
However, C105Y peptide did colocalize with intracellular vesicles stained by FM1-43X, a fluid phase endocytic marker. This suggests that these membrane-bound vesicles might facilitate intracellular C105Y peptide trafficking. However, the mechanism of membrane translocation of C105Y needs to be further investigated and might occur similarly to penetratin, which enters cells via a nonendocytic and receptor-and transporter-independent pathway. Instead penetratin, which needs the hydrophobic residue tryptophan for internalization, has been shown to cross pure lipid bilayers without pore formation (8,13).
CPPs may be useful reagents for delivering specific cargoes to cells. Already peptides, proteins, plasmids, antisense mRNA, and small interfering RNA have been transported into cells by CPPs. C105Y is no exception. Our previous data show that C105Y mediates cellular uptake of molecular cargoes. C105Y covalently attached to polyK was used to condense plasmid DNA by electrostatic interactions. In vitro these C105Y-polyK-coupled DNA complexes, gave up to 100-fold increase in gene expression compared with polyK⅐DNA complexes that did not contain C105Y peptide. These C105Y⅐polyK⅐DNA complexes also increased gene expression in vivo. In many instances expression from nonpeptide containing complexes is not detectable, but strong expression is observed from C105Y containing nanoparticles (18,19). The results obtained by Wu et al. (31) and Patel et al. (32), which used minor variants on the C105Y sequence, but both containing the FVYLI core sequence to target AAV viruses and other gene delivery systems, respectively, show that the addition of this ligand markedly increases transfection efficiency of virus particles into airway cells and enhanced gene expression when linked to other gene delivery systems. Our data now suggest that enhanced gene expression from sec-R-targeted complexes result from the ability of C105Y to enter cells via a pathway independent of clathrin and caveolae and its ability to be transported to the nucleus and nucleolus by nondegradative pathways.
Interestingly, C105Y localizes to the nucleolus, a plurifunctional organelle responsible for ribosome biogenesis, RNA processing, viral replication, and tumor suppression. Thus C105Y is a good candidate to deliver potentially therapeutic cargoes (e.g. antivirals and tumor suppressors) to this site (33,34). Because the pathway by which nontargeted complexes enter cells is not well understood, it is also possible that

TABLE 2
Summary of data on C105Y uptake and intracellular localization C105Y enhances gene expression simply by facilitating nuclear entry. These results suggest a potential role for using C105Y to deliver other therapeutic cargoes to cells.
The nucleolar targeting of C105Y is likely sequence-dependent. A penta-peptide with the sequence VPMLK, derived from the Ku70 protein, has similar amino acid composition to PFVYLI. This peptide, VPMLK, can penetrate cells, can localize to the nucleus, and is one of the shortest CPPs described to date (35,36). The Lamond laboratory, in an effort to characterize proteins within the nucleolus, determined that some short peptide motifs show specific enrichment in the nucleolar proteome. For example, the protein phosphatase-1-binding motif, (L/R)VXP, is highly enriched in the nucleolar proteome compared with the total nuclear proteome (33). This nucleolar binding motif is very similar to the PFVYLI sequence found within C105Y and the VPMLK sequence derived from the Ku70 protein. Therefore, it may not be coincidence that these two peptides localize to the nucleus and/or nucleolus.
In conclusion, we identified C105Y to be a novel cell-penetrating peptide that is rapidly internalized by live cells by an energy-independent process via caveolin-and clathrin-independent lipid rafts. Moreover, it traffics rapidly to the nucleus and nucleolus. These properties probably account for the ability of this peptide to increase gene transfer and gene expression both in vitro and in vivo, when conjugated to polyK-DNA complexes or included in the capsid of AAV. Consequently, C105Y may be a useful carrier of other molecular cargoes as well.