Internalization of novel non-viral vector TAT-streptavidin into human cells

Background The cell-penetrating peptide derived from the Human immunodeficiency virus-1 transactivator protein Tat possesses the capacity to promote the effective uptake of various cargo molecules across the plasma membrane in vitro and in vivo. The objective of this study was to characterize the uptake and delivery mechanisms of a novel streptavidin fusion construct, TAT47–57-streptavidin (TAT-SA, 60 kD). SA represents a potentially useful TAT-fusion partner due to its ability to perform as a versatile intracellular delivery vector for a wide array of biotinylated molecules or cargoes. Results By confocal and immunoelectron microscopy the majority of internalized TAT-SA was shown to accumulate in perinuclear vesicles in both cancer and non-cancer cell lines. The uptake studies in living cells with various fluorescent endocytic markers and inhibiting agents suggested that TAT-SA is internalized into cells efficiently, using both clathrin-mediated endocytosis and lipid-raft-mediated macropinocytosis. When endosomal release of TAT-SA was enhanced through the incorporation of a biotinylated, pH-responsive polymer poly(propylacrylic acid) (PPAA), nuclear localization of TAT-SA and TAT-SA bound to biotin was markedly improved. Additionally, no significant cytotoxicity was detected in the TAT-SA constructs. Conclusion This study demonstrates that TAT-SA-PPAA is a potential non-viral vector to be utilized in protein therapeutics to deliver biotinylated molecules both into cytoplasm and nucleus of human cells.


Background
Due to the limitations of current drug delivery systems, which have been hampered by their inefficiencies in tra-versing the cell membrane, there is a pressing need to develop methods for increasing intracellular delivery of protein-based cargoes. Over the past decade, numerous strategies to overcome the cell membrane barrier have been proposed, including electroporation, microinjection, viral vectors, liposome encapsulation and receptormediated endocytosis. These methods have, however, been plagued by low delivery efficiencies and, to some extent, increased cellular toxicity. Naturally occurring short cell-penetrating peptides (CPPs) derived from viral, insect or mammalian proteins, have attracted considerable interest in the field of drug delivery for their ability to direct cellular uptake through active transport mechanisms. CPPs are oligopeptides, 11-30 amino acid residues in length, that are capable of conferring their apparent translocation activity to proteins and other macromolecular cargo to which they are linked [1]. In recent years, CPPs have been studied extensively, both in vitro and in vivo, for their ability to delivery an array of pharmalogically relevant cargoes, such as antisense oligonucleotides, peptides, proteins, plasmids, liposomes and nanometer-sized particles, with encouraging results [2]. Lately, CPPs have also been used to treat preclinical models of human disease [3,4].
A variety of internalization routes for the TAT [47][48][49][50][51][52][53][54][55][56][57] sequence and TAT-mediated cargoes have been suggested. In the study of Fittipaldi et al. (2003) and Ferrari et al. (2003) Tat11EGFP and GST-Tat-eGFP proteins were reported to internalize into cells via caveolae-mediated endocytosis and transported further to the perinuclear area via an actin cytoskeleton-mediated mechanism [19,20]. Wadia et al. (2004) in turn showed the internalization of the TAT-Cre protein into cells by lipid raftdependent macropinocytosis [21], and Richard et al. (2005) suggested the uptake of the TAT peptide via clathrin-mediated endocytosis [22]. Recently, also Säälik et al. (2004) demonstrated the uptake of biotinylated TAT, detected with FITC-labeled avidin, via both clathrindependent and clathrin-independent endocytosis [23]. Central to the use of TAT, however, is not only its ability to deliver cargo to cells but, importantly, its non-cytotoxicity and stable biological activity over long time periods [5].
Core streptavidin (SA; 125-127 aa) from Streptomyces avidinii has been used in many pharmalogical applications. The exact mechanism of its cellular uptake and intracellular delivery is, however, not well known. The internalization of SA has been suggested to occur via receptor-mediated endocytosis, involving lysine residues [24] and the RYD sequence (Arg-Tyr-Asp) [25]. In vivo, the biodistribution of SA has been shown to exhibit slow clearance from the bloodstream due to accumulation in the kidney [26,27]. The present study was designed to gain insight into the internalization of a novel TATstreptavidin (TAT-SA) construct [10] in human cells. Additionally, the ability of TAT-SA as a transporter of biotin and biotinylated molecules was examined. The subcellular distribution of TAT-SA was altered by biotinylated, pH-responsive polymer poly(propylacrylic acid) (PPAA), which further promoted the endosomal release. These studies provide insights into the mechanism of TAT-SA uptake in cells and may have implications for the optimal use of TAT-SA and PPAA for the intracellular delivery of numerous biotinylated macromolecules.

Characterization of TAT-SA constructs
The structure of TAT-streptavidin (TAT-SA) has been previously described [10] as a tetrameric fusion protein, in which the TAT 47-57 peptide has been attached to the N-terminus of each streptavidin monomer. Biotins or biotinylated molecules are located into binding pockets of SA. In this study, the stability and biotin-binding ability of TAT-SA were analyzed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblot analysis. Both constructs, TAT-SA and Alexa488labeled TAT-SA (TAT-SA-A488), retained their tetrameric (60 kD) conformations in reducing conditions as either apoform or bound to biotin. When samples were preheated to 68°C, a minor proportion of TAT-SA or TAT-SA-A488 were detected as monomeric (15 kD) or dimeric (30 kD) forms. Binding of biotin to TAT-SA, however, stabilized the constructs at 68°C, since only tetrameric proteins were detected (unpublished data). Importantly, reducing conditions and preheatmeant to 37°C did not change the tetrameric conformations of either TAT-SA or TAT-SA bound to biotin (Additional file 1).

Microscopical analysis of TAT-SA and SA uptake
TAT-SA (2 μM) was shown to internalize into human epithelial carcinoma (HeLa), human lung carcinoma (A549) and human lung fibroblast (MRC-5) cell lines at 4 h post transduction ( Fig. 1A-C). In living HeLa cells TAT-SA-A488 was observed to cross the cell membrane rapidly, starting at 5 min post transduction, by confocal microscopy. At later timepoints the majority of TAT-SA-A488 was localized in vesicular compartments within the cytoplasm, and only few cells displayed detectable nuclear accumulation at 4 h post transduction (Fig. 1D-E). Control studies in which cells were first transduced either with TAT-SA-A488 or TAT-SA, fixed at 4 post transduction and immunostained for SA confirmed that both studies in living and fixed cells resulted in a similar subcellular localization. Additionally, in cells transduced with SA-A488 (2 μM) alone, the cellular uptake was reduced ( Fig. 1G-H). The role of TAT during the nuclear import of SA was confirmed by injecting TAT-SA-A488 or SA-A488 (~1.5 × 10 6 molecules/cell) directly into the cytosol of HeLa cells. At 4 h post injection the majority of TAT-SA-A488 was found to accumulate in the nucleoplasm with only minor localization in the cytoplasm. In the nucleus TAT-SA-A488 was distributed evenly throughout the nucleoplasm without accumulation into distinct subnuclear structures (Fig. 1F). In contrast, SA-A488 was distributed randomly throughout the cytoplasm but not in the nucleus (Fig. 1I).
To monitor further the internalization and intracellular transport of TAT-SA and SA, transduced HeLa cells were labeled with an antibody against SA and nanogold-immunolabeling electron microscopy with a silver enhancement technique was used. Intracellular TAT-SA, visualized as small, intensely labeled, grainy spots was localized in large intracellular vesicles near the cell surface or in endocytic-like vesicles adjacent to the nuclear membrane at 4 h p.t ( Fig. 2A). At the same time a small proportion of TAT-SA was observed in the cytosol and the nucleus of cells (Fig. 2B). Interestingly, in the control experiments SA was observed in smaller cytoplasmic vesicles without any nuclear localization than in those displaying TAT-SA at 30 min to 4 h post transduction.

Endocytic entry pathways of TAT-SA
To study the potential role of caveolar endocytosis in the uptake of TAT-SA, HeLa cells were transduced and immunolabeled with caveolin-1 antibody. No colocalization of TAT-SA-A488 and caveolin was, however, apparent at 15-60 min post transduction. In addition, the internalization of TAT-SA-A488 into living HeLa cells was not blocked in the presence of the cholesterol-depleting agent β-methylcyclodextrin, a known inhibitor of caveolae-mediated endocytosis. Importantly, control studies with human hepatoma (HepG2) cells, which do not express caveolins endogenously, demonstrated the efficient internalization of TAT-SA (unpublished data).
In order to characterize clathrin-dependent and clathrinindependent endocytosis in the uptake of TAT-SA, AP180-C plasmid encoding a dominant-negative form of AP180, an inhibitor of clathrin-mediated endocytosis, was used [28]. In HeLa cells overexpressing AP180-C, the internalization of TAT-SA was reduced, but not completely inhibited at 4 h post transduction (Fig. 3A), whereas the uptake of a clathrin-mediated endocytotic marker, Transferrin (Tf; 200 μg/ml), was markedly decreased. In the absence of AP180-C plasmid, immunofluorescence studies revealed slight colocalization of TAT-SA-A488 and an early endosomal marker (rab5) at 15 min post transduction as well as of TAT-SA-A488 and a recycling endosomal marker (rab11) at 30 min post transduction (unpublished data). In addition, the lysosomal marker LAMP-2 showed clear colocalization with TAT-SA-A488 at 2-4 h post transduction (Fig. 3B).
The kinetics of TAT-SA translocation from the cell periphery towards the nuclear periphery at different times was monitored in living HeLa cells in the presence of the fluorescent endocytic markers transferrin and dextran. Partial colocalization of TAT-SA-A488 and TRITC-labeled Transferrin (TRITC-Tf) was observed first at 5 min post transduction in close proximity to the cell membrane and later at 15-30 min post transduction in perinuclear vesicles (Fig. 4A). In order to examine the role of non-clathrinmediated endocytosis, a fluid-phase endosomal marker TRITC-labeled Dextran (TRITC-Dextran; 10 MW, 250 μg/ ml), was used to monitor macropinocytic entry pathway. As shown in Figures  The uptake of TAT-SA was further examined in living HeLa cells by quantitative image analysis (n = [25][26][27][28][29][30] in the presence of different endocytic inhibitors. In comparison to relative fluorescence intensity of untreated, transduced control cells (1.0 ± 0.3; Fig. 5B) a microtubule-disrupting drug, nocodazole, caused only a slight decrease (0.78 ± 0.3) of intracellular TAT-SA-A488. In cells treated with the filamentous F-actin elongation inhibitor cytochalasin D, affecting both macropinocytosis and clathrin-dependent endocytosis, a clear decrease (0.22 ± 0.1) in the amount of cytoplasmic TAT-SA was, however, measured. Furthermore, amiloride, an inhibitor of Na+/H+ exchange required for macropinocytosis, displayed significantly reduced (0.14 ± 0.1) amounts of TAT-SA uptake. Importantly, in control experiments extensive disruption of microtubules or actin filaments was observed in cells treated either with nocodazole or cytochalasin D and amiloride. Additionally, cytochalasin D and amiloride markedly decreased the internalization of TRITC-Dextran in living cells (unpublished data).

The endosomal release of TAT-SA and TAT-SA bound to biotin via the PPAA polymer
The streptavidin-biotin association is one of the strongest known non-covalent interactions in nature. In order to alter the subcellular distribution of TAT-SA, a biotinylated, pH-responsive polymer poly(propylacrylic acid) (PPAA; 4 μM) was complexed to TAT-SA (2 μM) to promote endosomal release. The 3D illustrations of live HeLa The ability of TAT-SA to act as an intracellular delivery vector by transporting biotin and biotinylated molecules into cells was also examined by confocal microscopy. As shown in the 3D illustration, TAT-SA-A488 (2 μM) and Biotin-DY-633 (1-2 μM) accumulated in large intracellular vesicles near the nucleus (Fig. 7A). No nuclear localization of TAT-SA was observed, however. Further experiments demonstrated strong colocalization of the TAT-SA-biotin complex and TRITC-Dextran in vesicles at the nuclear periphery at 2 h and 4 h post transduction, implying that most of the TAT-SA-biotin complexes were unable to escape from the intracellular vesicles. When endosomal releasing polymer PPAA was complexed with TAT-SA bound biotin, however, a major subnuclear accumulation was observed (Fig. 7B, Additional file 4).
Finally, cell viability experiments with MTT cell proliferation assay showed that TAT-SA (2 μM) was not cytotoxic to HeLa cells at 4-72 h post transduction (Table 1). Also no indication of a cytotoxic effect was observed in the presence of TAT-SA-complexed, biotinylated PPAA (4 μM) at 4-24 h post transduction and only a slight decrease in cell viability was detected at 48 and 72 h post transduction (94.0% ± 3.8% survival) compared to untransduced control cells (100.0% ± 3.3% survival).

Discussion
The cell membrane is a barrier to the intracellular delivery of many pharmacologically important biological macromolecules. Several cell penetrating peptides have been shown to possess the ability to direct cellular uptake, including the HIV-1 TAT peptide. So far TAT has been used for directing the intracellular uptake of various biomolecular cargoes. In this study, however, a novel streptavidin fusion protein TAT-SA [10] was examined, which enables the transport of biotinylated cargoes into cells. The internalization of TAT-SA was characterized in both human cancer (HeLa and A549) and non-cancer cell lines (MRC-5) (Fig. 1A-E). The internalization of SA was a relatively slow process in which detectable amounts of SA-A488 were present in the cytoplasm but not the nucleus of HeLa cells at 4 h post transduction (Fig. 1G-H). On the contrary, TAT-SA displayed rapid and efficient internalization in living HeLa cells starting at 5 min post transduc-

Figure 2 Localization of internalized TAT-SA in HeLa cells by immunoelectron microscopy at 4 h post transduction.
Intracellular TAT-SA was detected with pre-embedding immunolabeling technique, in which anti-SA staining was followed by silver-enhanced nanogold and gold-toning treatments. Localization of SA was shown as dark spots (arrow heads) within (A) cytoplasmic vesicular structures and further (B) in the nucleus of cells (nu). Scale bars, 1 μm. tion with distribution into vesicular structures in the cytoplasm after 15 min (Fig. 1A, 1D-E). At later time points, confocal and EM imaging demonstrated that the majority of TAT-SA was localized in the cytoplasmic vesicles with trace amounts in the nucleus (Fig, 1D-E, Fig. 2A and 2B). Previously, numerous possible internalization routes for TAT have been proposed, such as lipid-raftmediated macropinocytosis [21], caveolae-mediated endocytosis [19,20] and clathrin-independent and dependent endocytosis [22,23]. However, the uptake characteristics of the TAT peptide alone and of TAT-conjugated cargoes have been demonstrated to differ significantly [1,4,22]. Furthermore, TAT-mediated internalization process has proposed to be dependent on the properties of the cargo molecule, TAT concentration and cell line [1,29,30]. In our study, streptavidin (60 kD) as a larger partner of the TAT-SA fusion construct (TAT 47-57 -peptide, 11aa) is likely to affect the uptake and intracellular trafficking of the vector. Notably, we show here that direct microinjection of high concentrations of TAT-SA or SA into the cytoplasm resulted in efficient nuclear uptake of TAT-SA but not SA ( Fig. 1F and 1I). This verifies previous findings that upon introduction into the cytosol, the TAT peptide is capable of mediating the nuclear import of its streptavidin fusion partner. Moreover, it is known that positively charged molecules internalize the cells efficiently. Consequently, taken to account that the plain SA is negatively charged (theoretical pI 6.04) and TAT-SA is positively charged (theoretical pI 9.92), TAT-SA internalizes the cells more efficiently than SA. Taken together, these data demonstrate that TAT-SA is efficiently internalized into various human cells but that only a relatively small proportion is further released into the cytoplasm and transported into the nucleus, most likely reflecting the inability of TAT-SA to escape from endocytic vesicles.
In order to use TAT-SA as a vector to deliver biotinylated molecules into cells, the internalization and delivery mechanisms of the construct have to be characterized in vitro and in vivo. In recent years a number of studies have suggested a variety of internalization routes for TAT pep-Immunofluorescence microscopy studies of the cellular distribution of TAT-SA tide and TAT-mediated cargoes. To study this, three major endocytic pathways involving caveolae, lipid rafts and clathrin-coated pits were analyzed using specific endocytic markers, inhibition agents and immunofluorescence labelings of each pathway. Previous studies have suggested the TAT peptide enters cells by temperaturedependent, caveolae-mediated endocytosis [19,20]. However, no evidence for the use of the caveolae route in the internalization of TAT-SA was observed in the present confocal microscopial colocalization studies with a caveolar marker protein. Moreover, treatment with methyl-βcyclodextrin, a cholesterol depletion agent known to inhibit the caveolae route [31] or transduction of the caveolin-deficient HepG2 (unpublished data) and Jurkat T cells [10,21,32,33] did not prevent the entry of TAT-SA-A488 into living cells. Altogether, these data imply that endosomal routes other than caveolae-mediated entry are required for the uptake of TAT-SA.
It has been suggested that the internalization of SA occurs via clathrin-mediated endocytosis [24,25]. Here, the role of clathrin-dependent uptake of TAT-SA was monitored in cells overexpressing the mutant AP180-C protein [34]. AP-180C is required for the efficient assembly of clathrincoated pits by interacting with the clathrin heavy chain through its C-terminal clathrin-binding motifs. Our data indicated that in AP180 overexpressing cells the internalization of TAT-SA was only slightly affected, whereas the uptake of Tf, a marker of clathrin-dependent endocytosis [35], markedly decreased (Fig. 3A). The double-immu- nolabeling studies of clathrin-mediated endocytosis showed only minor colocalization of TAT-SA with early and recycling endosomal markers at early time points (unpublished data). However, at later stages of endocytosis TAT-SA accumulated in lysosomes, the final destination of multiple endocytic pathways (Fig. 3B). Clathrinmediated endocytosis seemed therefore to be only partially involved in the uptake of TAT-SA, the major internalization being mediated by another, more efficient pathway.

Colocalization of TAT-SA with endocytic markers in living cells
Many oligoarginine peptides have been proposed as candidates for cellular internalization via macropinocytosis [30]. Recently, uptake via clathrin-independent endocytosis was also suggested in the internalization of TAT-Cre and biotinylated TAT [21,23,36]. In our study, the intracellular localization of TAT-SA was monitored in living HeLa cells with various fluorescent endocytic markers. Shortly after internalization, only slight colocalization was detected with TAT-SA and Tf, a clathrin-dependent endocytic marker (Fig. 4A), whereas extensive colocalization between TAT-SA and the fluid-phase endosomal and macropinocytic marker dextran was observed (4B, 4C and Additional file 2). Quantitative measurements of fluorescent intensity at 4 h post transduction displayed an approximately 10-fold increase in the colocalization of TAT-SA and dextran as compared to TAT-SA and Tf (Fig.  5A), thus supporting previous observations of the role of macropinocytosis as an entry mechanism for TAT. Additionally, when cells were treated with amiloride, a known inhibitor of macropinocytosis [37], the uptake of TAT-SA-A488 was almost entirely inhibited. In the presence of cytochalasin D, an F-actin-disrupting agent affecting both macropinocytosis and clathrin-mediated endocytosis [34,38], only a minor internalization of TAT-SA-A488 occurred (Fig. 5B). To conclude, we were able to establish macropinocytosis as a major internalization pathway of TAT-SA, as illustrated (Fig. 8). The central interest in this study was to determine the ability of TAT-SA to act as a carrier of biotinylated molecules into nucleus of human cells. Although numerous studies have been conducted on TAT-related protein therapeutics, only a few have attempted to address the intracellular release of macromolecules from endocytic vesicles. Various approaches for endosomal escape have been proposed, such as endosome-disrupting agents (HA2, 4 3 E), acid-sensitive linkers or laser illumination [39]. In a recent study by Wadia et al. (2004) it was shown that an internalized TAT-Cre fusion protein was capable of nuclear uptake and the subsequent induction of reporter gene expression. Most of the TAT-Cre peptides were, however, observed to reside in cytoplasmic vesicles and therefore the endosomal escape was enhanced by incorporation of the HA2 peptide from the influenza virus hemagglutin protein. Albarran et al. (2005) have previously shown the delivery of biologically active alkaline phosphatase (140 kD) and R-phycoerythrin (240 kD) into the cytoplasm of Jurkat T cells by TAT-SA. In present study, it was, however, shown that both TAT-SA and TAT-SA bound to biotin remained enclosed in endocytic vesicles for several hours after internalization (Fig. 4C, 7A, Additional file 3). Therefore, the effect of a biotinylated pH-responsive polymer PPAA [10] on the subcellular distribution of TAT-SA was investigated. In previous studies PPAA has shown to disrupt endosomes at pH 6.5 or below, causing the cytosolic release of cargo molecules [10,[40][41][42][43]. Recently, PPAA has also been shown to enhance the delivery of antibody-targeted conjugates into the cytoplasm [40] and PPAA-containing lipoplexes have improved wound healing in mice [42]. Our data indicated clearly that the nuclear import of TAT-SA-A488 or TAT-SA bound to biotin was considerably increased when combined with biotinylated PPAA (Fig.  6A and 7B, Additional file 4). Quantitative analysis of live cells showed an approximately 2-fold increase in the nuclear localization of TAT-SA-A488-PPAA complexes compared to TAT-SA-A488 alone (Fig. 6B). The endosomal escape of TAT-SA-A488-PPAA may, however, been more extensive, since the detection of fluorescence intensity by confocal microscope is not sensitive enough to observe all the nuclear fusion proteins. Importantly, no significant cytotoxicity of TAT-SA-PPAA was observed by MTT assay even at 72 h post transduction (Table 1). Overall, these data further demonstrate that the incorporation of PPAA with TAT-SA alters the subcellular distribution of the internalized fusion protein, resulting in improved nuclear delivery of its biotin partner (Fig. 8).

Quantification of internalized TAT-SA in living HeLa cells
Recently, several therapeutic applications for TAT-mediated cellular internalization have been developed, including the transport of an inhibitor for human papillomavirus type 16 [44] and for the apoptosis-promoting caspase-3 protein used in HIV-therapy [45], the extension of the cytotoxic activity of herpes simplex virus-1 thymidine kinase for cancer therapy [46], dendritic cellbased immunotherapy [47] and enhancement of viralmediated gene delivery, [48] to name just a few. In this study, the fusion of TAT with SA imparts the versatility and precision of the SA-biotin system and allows the complexation of numerous organic and inorganic cargoes. Additionally, the TAT-SA fusion protein and its use in combination with the endosomal-releasing polymer PPAA demonstrates the potential of this construct for delivering biotinylated molecules into various intracellular compartments, depending on the chemistry of the chosen biotin partner.

Conclusion
Non-viral vector TAT-SA internalizes into human cells via both macropinocytosis and clathrin-dependent endocytosis. The subcellular distribution of TAT-SA is significantly Internalization and intracellular targeting of TAT-SA combined with biotin in HeLa cells altered through the incorporation of a pH-sensitive polymer PPAA. Here, we have characterized of a novel and versatile vector that is capable of delivering an array of biotinylated macromolecular cargoes into cells.

TAT-Streptavidin (TAT-SA) construction and expression
The design and construction of the TAT-SA gene, T7 expression system (pET-21a, Novagen, Inc., Madison, WI) and the isolation, refolding, purification as well as structural and preliminary functional characterization of the TAT-SA fusion protein have previously been reported [10]. TAT-SA and SA was labeled with Alexa-488 according to the protocol for amine-reactive probes (Molecular Probes). The biotin-binding ability and stability of protein constructs were detected by SDS-PAGE and immunoblot analysis. In these experiments part of the samples were first conjugated with biotin at RT for 15 min, and then TAT-SA and biotinylated TAT-SA samples were preheated to 22°C, 37°C or 68°C for 10 min. All samples were predisposed to a reducing agent β-mercaptoethanol for 30 min prior to loading. For live cell studies by a laser scanning confocal microscope (Zeiss LSM 510) the cells were maintained in Scotch chambers [50] in CO 2 -independent medium (Gibco) supplemented with 10% inactivated FCS and penicillin-streptomycin. Prior to detection the objective and sample holder were heated to 37°C. In the imaging, appropriate excitation and emission settings together with the multitracking mode were used to avoid false colocalization.

Transduction of TAT-SA proteins
To quantitate the colocalization of TAT-SA-A488 with TRITC-Tf or TRITC-Dextran, the collected intracellular fluorescence intensity data for each cellular fluorescent marker was processed using 3D LSM and ImageJ programs (Colocalization Finder Plug-in). The influence of different drugs on the cellular internalization of TAT-SA-488 was quantitated by collecting the intracellular fluorescence intensity data from multiple series of drug-treated cells and then processed with the 3D LSM program. Quantitative analysis corresponding to the nuclear import of TAT-SA-488 in the presence of PPAA was performed by summarizing the fluorescence intensity data from multiple series of optical sections of the nuclear area and analyzing with the 3D LSM program. Prior to all the quantitative analyses the data from each channel were corrected by reducing the background signal of untransduced cells.
For immunofluorescence microscopy, the cells were fixed at set time intervals post transduction either with absolute methanol (-20°C) at RT for 6 min or with 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS, pH 7.4) at RT for 20 min. PFA fixed cells were permeabilized with 0.5% Triton X-100. Cells were sequentially immunolabeled with primary and labeled secondary antibodies, embedded with Mowiol-DABCO or ProLong ® Gold antifade reagent with DAPI (Molecular Probes) and subjected to confocal microscopy (Zeiss LSM 510 or Olympus Fluo-View 1000, Olympus Optical Co., Tokyo, Japan).

Microinjection
Hela cells were grown to subconfluency on microgrid coverslips (grid size, 175 nm; Eppendorf, Hamburg, Germany). The injections were performed into the cytoplasm of living cells using a semiautomatic system comprising a Transjector 5246 and Micromanipulator 5171 (Eppendorf) attached to an inverted microscope. Capillaries for injections (Clark Electromedical Instruments, Pangbourne, UK) were prepared with a model P97 capillary puller from Sutter Instruments (Novato, CA). The sample ejection volume (0.1 pl) was measured by injecting radioactive [ 3 H]-biotin (Amersham Biosciences, Little Chalfont, UK) standard. Since the concentration of injected TAT-SA-A488 and SA-A488 was 1.4 mg/ml in PBS, the actual intracellular amount of the injected proteins was estimated to be approximately 1.5 × 10 6 molecules/cell.

Nanogold pre-embedding immunoelectron microscopy
After transduction of TAT-SA (5 μM) and fixing with glutaraldehyde, HeLa cells were washed with phosphate buffer (0.1 M Na 2 HPO 4 pH 7.4) and permeabilized with a saponin buffer (0.01% saponin/0.1% BSA/0.1 M Na 2 HPO 4 ). Cells were then labeled with anti-SA Ab at RT for 1 h, followed by nanogold-conjugated anti-rabbit IgG at RT for 1 h. After appropriate washes with saponin and phosphate buffers, post-fixing (1% glutaraldehyde in 0.1 M phosphate buffer) and quenching (50 mM NH 4 CL in Na 2 HPO 4 ) were performed. Moreover, silver enhancement and gold toning (2% Na-acetate, 0.05% HAuCl 4 , and 0.3% Na-thiosulphate in EM water) were followed by post-fixing (1% osmium tetroxide, 0.1 M phosphate buffer, K 4 Fe(CN) 6 15 mg/ml) at 4°C for 1 h. The samples were dehydrated in ethanol, stained with 2% uranylacetate, and embedded in LX-112 epon. Polymerization of epon was performed over a 24 h period, first at 45°C and then at 60°C, after which the samples were stained with toluidine blue, cut with an ultramicrotom (Reichert-Jung, Ultracut E) and stained again with uranylacetate and lead citrate. Detection was performed by a JEOL JEM-1200EX transmission electron microscope operated at ~60 kV.