Protein delivery into live cells by incubation with an endosomolytic agent

Abstract

We report that a tetramethylrhodamine-labeled dimer of the cell-penetrating peptide TAT, dfTAT, penetrates live cells by escaping from endosomes with high efficiency. By mediating endosomal leakage, dfTAT also delivers proteins into cultured cells after a simple co-incubation procedure. We achieved cytosolic delivery in several cell lines and primary cells and observed that only a relatively small amount of material remained trapped inside endosomes. Delivery did not require a binding interaction between dfTAT and a protein, multiple molecules could be delivered simultaneously, and delivery could be repeated. dfTAT-mediated delivery did not noticeably affect cell viability, cell proliferation or gene expression. dfTAT-based intracellular delivery should be useful for cell-based assays, cellular imaging applications and the ex vivo manipulation of cells.

Figure 1

Cytosolic delivery of dfTAT in live cells is efficient and exceeds its monomeric counterparts. (a) Cellular localization of acfTAT and dfTAT assessed by fluorescence microscopy. Cells were incubated for 1 h with either acfTAT (20 μM) or dfTAT (5 μM), washed, and imaged with a 100x objective. Monochrome images represent the emission of TMR at 560 nm. Scale bars, 10 μm. (b) Comparison of the cytosolic delivery efficiency of acfTAT, fTAT, and dfTAT. Cells were incubated with acfTAT, fTAT, and dfTAT (1-20 μM) for 1 h. The number of cells with detectable cytosolic and nuclear fluorescence distribution in microcopy images was counted and divided by the total number of cells present (%) (1,000 cells/experiment). (c) dfTAT overall uptake in HeLa cells as a function of the concentration of dfTAT present in the incubation media. Cells were incubated with dfTAT (1-10 μM) for 1h and relative uptake was assessed quantitatively by measuring the bulk fluorescence of cell lysates (300,000 cells/experiment). The data shown in b and c represent the mean of triplicate experiments and the corresponding standard deviations.

Figure 2

dfTAT penetrates the cytosol by escaping from the endocytic pathway. (a) Assay showing the effect of endocytosis inhibitors on the cellular distribution of dfTAT. HeLa cells were pre-treated with each inhibitor for 20 min, washed, and incubated with 5 μM dfTAT and inhibitor for 1h. The percentage of cells displaying a cytosolic fluorescence distribution was quantified as in Fig. 1. Insert: image showing a punctate distribution of dfTAT in the presence of bafilomycin (1,000 cells/experiment, experiments were performed in triplicates, mean and corresponding standard deviations represented). Scale bar, 10 μm. (b) Pulse-chase experiment showing the progressive cytosolic penetration of dfTAT. HeLa cells were incubated with dfTAT (5μM) for 5 min, washed and imaged with a 20X objective (TMR fluorescence images are represented as inverted monochrome). The imaging intervals and corresponding percentages of cells with a cytosolic signal are represented. (c) Microscopy images showing that dfTAT causes the cytosolic release of molecules trapped inside endosomes. Hela cells were incubated with 5 μM DEAC-K9 for 1 h and washed. Cells were subsequently incubated with 5 μM dfTAT for 1 h. Images are represented as inverted monochromes. Scale bars, 10 μm. (d) Assay establishing the endosomolytic efficiency of dfTAT. HeLa cells expressing SNAP-H2B were incubated with 5 μM dfTAT and 5 μM SNAP-Surface 488. Representative fluorescence images of SNAP-Surface are shown (dfTAT is in insert). The SNAP-Surface 488 signal present in the nucleus is indicated as a % of the total signal. Scale bars, 10 μm.

Figure 3

dfTAT-mediated delivery does not significantly affect cell proliferation and transcription. (a) Proliferation assay. HeLa, Neuro-2a and HDF cells were incubated with 5 μM dfTAT for 1 h or left untreated. Proliferation was assessed using a MTT assay (150,000 cells/experiment, experiments in triplicates, mean and standard deviations represented). (b) Microscopy imaging showing that cells containing cytosolic dfTAT divide. HeLa cells were incubated with 5 μM dfTAT for 1 h, washed and imaged in a time-lapse experiment. Scale bars, 10 μm (c) Whole-genome microarray analysis of HDF cells treated with dfTAT. HDF cells were treated with 5 μM dfTAT for 1h and transcriptome analysis was performed immediately, 1h, or 24h after dfTAT treatment. The plot displays microarray intensity values of treated vs. untreated (same incubation steps but without peptide) samples. The red lines represent 2-fold intensity change cut-offs and transcripts up or down-regulated above these cut-offs are circled for clarity. (d) Microscopy imaging showing that dfTAT-mediated endosomal escape can be repeated. HeLa cells were co-incubated with dfTAT (5 μM) and DEAC-K9 (5 μM) for 1 h (step1) (images not shown). After washing, dfTAT (5 μM) and SNAP-Surface 488 (5 μM) were co-incubated in the absence (top panel) or presence (bottom panel) of bafilomycin (200 nM) (step 3). Scale bars, 10 μm.

Figure 4

Delivery of intact and functional proteins using co-incubation with dfTAT. (a) Microscopy imaging showing that dfTAT delivers EGFP into the cytosol of live cells. HeLa, Neuro-2a and HDF cells were co-incubated with EGFP (10 μM) and dfTAT (5 μM) for 1 h. Scale bars, 100X objective: 10 μm, 20X objective: 100 μm. (b) Assay showing that dfTAT improves the delivery of TAT-Cre. HeLa cells transfected with a plasmid containing egfp upstream of a loxP-STOP-loxP sequence were incubated for 1 h with either fTAT (5 μM) or dfTAT (5 μM) in the presence of TAT-Cre (1 μM). EGFP⁺ cells that result from successful TAT-Cre delivery were visualized and counted by microscopy (confirmed by flow cytometry). Scale bars, 100 μM. (c) Microscopy imaging showing that dfTAT mediates the delivery of an antibody. HeLa cells were co-incubated with FITC-anti-ATP5A (20 μg/mL) and dfTAT (5 μM) for 1 h. FITC-anti-ATP5A (green) is delivered in the cytosol of cells and stains tubular mitochondria (more clearly visible in zoomed-in image). Absence of SYTOX Blue staining indicates that the cells imaged do not have a compromised plasma membrane. Black arrows point to tubular mitochondria. Scale bars, 100X objective: 10 μm, zoomed-in image: 2 μm.

Figure 5

dfTAT-mediated delivery improves the delivery and transcriptional output of a transcription factor. (a) Assay showing that dfTAT mediated delivery of HoxB4 and TAT-HoxB4 improves the expression of a luciferase reporter. NIH 3T3 transfected with a HoxB4-dependent luciferase reporter were incubated for 1.5 h with either HoxB4 or TAT-HoxB4 (200 nM) in presence or absence of dfTAT (3 μM). TAT-mCherry (200 nM) and dfTAT (3 μM) serve as negative controls (400,000 cells/experiment, experiments in duplicate). (b) The amount of DEAC-K9 delivered in the cytosol and nucleus of live cells can be titrated. HeLa cells were incubated with dfTAT (5 μM) and increasing amounts of DEAC-K9 (1, 2.5, 5, 10, 20 μM). The fluorescence intensity of cells displaying cytosolic release was assessed by microscopy (representative 20X images are shown using a pseudocolored colorscale: blue=lowest intensity, red=highest intensity) and by measuring to the bulk fluorescence of cell lysates (microscope: 1,000 cells/experiment, fluorometer: 300,000 cells/experiment; experiments in triplicates, average and standard deviations represented). Scale bars, 10 μm (c) Assay showing that the induction of luciferase expression by dfTAT-mediated delivery of HoxB4 can be titrated. NIH 3T3 cells were co-incubated with HoxB4 (25-200 nM) and dfTAT (3 μM) and luciferase induction was measured as described in a (400,000 cells/experiment, experiments in duplicate).