Endosomal acidic pH-induced conformational changes of a cytosol-penetrating antibody mediate endosomal escape

Endosomal acidic pH-induced conformational changes of a cytosol-penetrating antibody mediate endosomal escape

Journal of Controlled Release 235 (2016) 165–175 Contents lists available at ScienceDirect Journal of Controlled Release journal homepage: www.elsev...

3MB Sizes 116 Downloads 96 Views

Journal of Controlled Release 235 (2016) 165–175

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

Endosomal acidic pH-induced conformational changes of a cytosol-penetrating antibody mediate endosomal escape Ji-Sun Kim a, Dong-Ki Choi a, Ju-Yeon Shin a, Seung-Min Shin a, Seong-Wook Park a, Hyun-Soo Cho b, Yong-Sung Kim a,⁎ a b

Department of Molecular Science and Technology, Ajou University, Suwon, 16499, Korea Department of Systems Biology, Yonsei University, Seoul 03722, Korea

a r t i c l e

i n f o

Article history: Received 19 March 2016 Received in revised form 13 May 2016 Accepted 31 May 2016 Available online 02 June 2016 Keywords: Endosomal escape Cytosol-penetrating antibody Antibody intracellular trafficking pH-dependent conformational change Membrane pore formation

a b s t r a c t Endosomal escape after endocytosis is a critical step for protein-based agents to exhibit their effects in the cytosol of cells. However, antibodies internalized into cells by endocytosis cannot reach the cytosol due to their inability to escape from endosomes. Here, we report a unique endosomal escape mechanism of the IgG-format TMab4 antibody, which can reach the cytosol of living cells after internalization. Dissociation of TMab4 from its cell surface receptor heparan sulfate proteoglycan by activated heparanase in acidified early endosomes and then local structural changes of the endosomal escape motif of TMab4 in response to the acidified endosomal pH were critical for the formation of membrane pores through which TMab4 escaped into the cytosol. Identification of structural determinants of endosomal escape led us to generate a TMab4 variant with ~3-fold improved endosomal escape efficiency. Our finding of the endosomal escape mechanism of the cytosol-penetrating antibody and its improvement will establish a platform technology that enables a full-length IgG antibody to directly target cytosolic proteins. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The ability of an antibody to reach the cytosol of target mammalian cells from outside of cells is highly desired for diverse purposes of research, diagnostic and therapeutic applications [1–3]. However, like other proteins and peptides, antibodies cannot passively diffuse across the plasma membrane of living cells owing to their large size and hydrophilicity [2,3]. Instead, they can be internalized into cells via receptormediated endocytosis after binding to cell surface receptors. After internalization, the antibody-receptor cargo is first transported to early endosomes (EEs), which serve as sorting stations for the internalized cargo, and subsequently routes to late endosomes (LEs) and finally lysosomes for degradation or can be transported back out of cells by recycling endosomes through the receptor- or neonatal Fc receptor-mediated recycling pathway [4,5]. To reach the cytosol, antibodies need to escape from endosomes along the endocytic pathway; however, they cannot do due to the inability of endosomal escape. Furthermore, the tight association of antibodies with the antigen receptor hampers their recycling or endosomal escape, causing lysosomal degradation, as exploited in antibody-drug conjugate technology [4]. To overcome these limitations of antibody itself, antibody delivery systems into the ⁎ Corresponding author at: Dept. of Molecular Science and Technology, Ajou University, 206 Worldcup-ro, Yeongtong-gu, Suwon 16499, Korea. E-mail address: [email protected] (Y.-S. Kim).

http://dx.doi.org/10.1016/j.jconrel.2016.05.066 0168-3659/© 2016 Elsevier B.V. All rights reserved.

cytosol of living cells, such as polymeric micelles [6] and cell-penetrating and fusogenic peptide-embedded liposomes [7], have been recently developed. Some bacterial toxins and cell-penetrating peptides (CPPs) that are derived from natural proteins or have been designed have the ability to reach the cytosol of cells after endocytosis from outside of cells [2, 3]. Although the detailed endosomal escape mechanisms of most toxins and CPPs are still poorly defined, they are believed to exploit changes in the endosomal environment, such as an acidified pH, a reducing environment and/or activation of some proteases, to trigger endosomal escape during endosomal trafficking [2,8,9]. Importantly, the pH of the endocytic compartments progressively decreases from EEs (pH 5.5– 6.5) to LEs (pH 4.5–5.5) to lysosomes (pH b 5.0) by the vesicular proton-ATPase pump [2,5]. The acidified endosomal pH plays a critical role to make toxins and CPPs undergo the necessary secondary/tertiary structural changes for interactions with the phospholipids of endosomal membranes to cause endosomal escape by membrane pore formation [8,9]. However, their endosomal escape efficiency is extremely low (less than 4% of the internalized molecule pools), with a negligible fraction of the endocytosed cargo reaching the cytosol [10,11], which limits their uses for the cytosolic delivery of biologically active cargos [12–15]. We recently reported the cytosol-penetrating antibody TMab4, a socalled cytotransmab, which in the intact human IgG1 format can access the cytosol of living cells after internalization through clathrin-mediated endocytosis using cell surface-expressed heparan sulfate

166

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

proteoglycan (HSPG) as a receptor [16,17]. However, the endosomal escape mechanism has not been elucidated. Here, we report that TMab4 undergoes conformational changes in response to the acidified endosomal pH in order to cause endosomal membrane pore formation for the endosomal escape of TMab4. Through systematic mutational and functional studies, we identified the structural determinants of TMab4 that respond to the acidic pH for conformational changes of the endosomal escape motif to induce interaction with endosomal membranes, resulting in membrane pore formation. Understanding the endosomal escape mechanism allowed us to generate a TMab4 variant with much improved endosomal release efficiency, thereby efficiently reaching the cytosol from outside of cells. 2. Materials and methods 2.1. Construction of TMab4 variant expression plasmids The mammalian expression plasmid, pcDNA 3.4-LC, encoding the light chain (LC) of TMab4 with hT4 light chain variable domain (VL) and Cκ constant domain sequence (residues 108-214 in EU numbering) was described before [16]. Site-directed mutagenesis for generating hT4 VL variants of TMab4 variants was performed by overlapping PCR using designed oligonucleotides (Macrogen Inc., Korea) [18]. The hT4-03 was prepared by DNA synthesis (Bioneer Inc., Korea). The mutated hT4 variants were subcloned into the NotI/BsiWI site of the pcDNA 3.4-LC vector for LC expression [16]. All of the constructs were confirmed by sequencing (Macrogen Inc.). 2.2. Expression and purification of IgG antibodies The mammalian expression plasmid, pcDNA 3.4-HC, encoding the heavy chain (HC) of TMab4 with heavy chain variable domain (VH) and human IgG1 constant domain sequence (CH1-hinge-CH2-CH3, residues 118–447 in EU numbering) was described before [16]. The plasmids that encode HC and LC of TMab4 and its variants were transiently co-transfected in pairs at an equivalent molar ratio into 0.2–1 L of HEK293F cell cultures in Freestyle 293F media (Invitrogen) following a previously described standard protocol [17,19,20]. Culture supernatants were harvested after 7 d culture at 37 °C in a humidified 8% CO2 incubator by centrifugation and filtration (pore size 0.22 μm, cellulose acetate, Corning, CLS430521). Antibodies were purified from the culture supernatants using a protein-A agarose chromatography column (GE Healthcare) and dialyzed to achieve a final buffer composition of PBS with pH 7.4. Prior to cell treatments, antibodies were sterilized by filtration using a cellulose acetate membrane filter (pore size 0.22 μm, Corning). Antibody concentrations were determined both using a Bicinchoninic Acid (BCA) Kit (Pierce, 23225) and by measuring absorbance at 280 nm.

TAT, as specified in the figure legends. To determine effects of pharmacological inhibitors, HeLa cells were pretreated with wortmannin (0.2 μM), bafilomycin (0.2 μM), and brefeldin A (7 μM) for 30 min and then treated with TMab4 or FITC-TAT. After 2 × washes with PBS, the cells were washed 2 × for 30 s at 25 °C with low-pH glycine buffer (200 mM glycine, 150 mM NaCl, pH 2.5), followed by 2 additional washes with PBS to remove non-internalized and nonspecifically surface-bound antibodies [23,24]. After fixation with 4% paraformaldehyde (PFA) in PBS for 10 min at 25 °C, permeabilization with PERM-buffer (0.1% saponin, 0.1% sodium azide, 1% BSA in PBS) for 10 min at 25 °C, and then blocking with 2% BSA in PBS for 1 h at 25 °C, the internalized cytotransmab was detected with Alexa488- or Alexa555-conjugated goat anti-human IgG antibodies for 2 h at 25 °C. Nucleus was stained with Hoechst 33342 for 5 min in PBS. After mounting the coverslips onto glass slides with Perma Fluor aqueous mounting medium (Thermo Scientific, TA-030-FM), center-focused single z-section images were obtained on a Zeiss LSM710 system with ZEN software (Carl Zeiss). In case of using 63× objective lens in confocal microscope, zoom factor 2 was applied for better resolution. The FITC or Alexa488 fluorescence was quantified using ImageJ software (NIH, USA) [1]. 2.5. Pulse-chase experiments to monitor intracellular trafficking of TMab4 HeLa cells (2 × 104) grown on coverslips in 24-well culture plates were incubated with TMab4 (3 μM) or FITC-TAT (20 μM) for 30 min, quickly washed 3× with PBS, and then incubated at 37 °C for the indicated periods. After washing with PBS, stripping with low pH glycine buffer, fixation, permeabilization, and blocking of the cells, the internalized TMab4 were stained with Alexa488-conjugated goat anti-human antibody, and subcellular endocytotic organelles were stained with anti-Rab5 and anti-Flag for Flag-Rab11, followed by the appropriate FITC- or TRITC-conjugated secondary antibodies. The plasmid encoding Flag-Rab11 was transiently transfected 24 h before the treatment of TMab4. Lysosomes were visualized with LysoTracker® Red DND-99 that was diluted in medium (1 μM) and pretreated for 30 min at 37 °C before the cellular fixation. The subsequent confocal microscopy was performed as described above. To examine effects of heparanase knockdown, control siRNA- and heparanase siRNA-treated HeLa cells were incubated with TMab4 or TAT for 30 min, quickly washed 3 × with PBS, and then incubated at 37 °C for 2 h. After washing with PBS, stripping with low pH glycine buffer, fixation, permeabilization, and blocking of the cells, the internalized TMab4 were stained with Alexa488-conjugated goat anti-human antibody, and subcellular endocytotic organelles were stained with anti-LAMP-1, followed by the appropriate TRITC-conjugated secondary antibody. The subsequent confocal microscopy was performed as described above. 2.6. Cytosolic calcein release assay

2.3. Modeling of TMab4 variable fragment (Fv) Modeling of the three-dimensional structure of TMab4 Fv from the primary amino acid sequence was performed on the web antibody modeling (WAM) algorithm [21]. The WAM algorithm performs homology modeling of VH and VL frameworks (FRs) and complementarity-determining regions (CDRs) by aligning the submitted sequence to the most sequence-homologous FRs and CDRs of the same canonical class, respectively, from the Brookhaven Protein Data Bank of known Ab structures. The modeling was visualized with COOT program [22]. 2.4. Confocal immunofluorescence microscopy Confocal microscopy was performed for detection of internalized cytotransmabs or FITC-TAT in cultured cells, as described before [20, 23,24]. Briefly, cells (2 × 104) grown on 12-mm diameter coverslips in 24-well culture plates were treated with indicated antibodies or FITC-

For tracing endosomal release of cytotransmabs or TAT using calcein [16,25], cells were incubated for 4 h with the indicated concentrations of antibodies or TAT in serum-free media at 37 °C using adalimumab as a non-internalizing negative control. Subsequently, 150 μM of calcein was added for 2 h at 37 °C. Cells were then washed three times with PBS and fixed. Calcein fluorescence images were obtained by confocal microscopy focusing on the center of cells at the same laser intensity for a set of samples. To analyze endosomal-released calcein, we selected the diffusive fluorescence signal area throughout the cytosolic region of cells, while excluding the punctuate fluorescence signal areas, and quantified the fluorescence intensity using ImageJ software. We analyzed more than 20 different cells for each sample and averaged the fluorescence intensity to get the calcein mean fluorescence intensity (MFI). The data were presented as the percentage of calcein MFI in the antibody/peptide-treated cells versus calcein MFI in the corresponding control cells.

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

To examine effects of pharmacological inhibitors, HeLa cells were pretreated with wortmannin (0.2 μM), bafilomycin (0.2 μM), and brefeldin A (7 μM) for 30 min and further treated with TMab4 (3 μM) or TAT (20 μM) for 6 h. The following calcein endosomal release assay was performed as described above. To examine effects of heparanase knockdown, control siRNA- and heparanase siRNA-treated HeLa cells were incubated with TMab4 (3 μM) or TAT (20 μM) for 6 h. The following calcein endosomal release assay was performed as described above. 2.7. Analysis of antibody influx into Ramos cells TMab4 or adalimumab was labelled using a DyLight 488 Antibody Labeling Kit (Thermo Scientific, USA) according to the manufacturer's protocol. Ramos cells were incubated for 2 h at 37 °C with DyLight488-labelled TMab4 (10 μM) or DyLight488-labelled adalimumab (2 μM) along with TMab4 (10 μM). Cells were washed, fixed, stained with Hoechst332 to label nuclei and then examined under a Zeiss LSM710 confocal microscope with ZEN software (Carl Zeiss, USA) [20].

167

otherwise specified. Comparisons of data between two groups were analyzed for statistical significance by a one-tailed, unpaired Student's t-test using Excel (Microsoft). For statistical analysis of multiple comparisons, one-way ANOVA followed by the NewmanKeuls post hoc test was used to determine significance using GraphPad Prism software. A P value of less than 0.05 was considered statistically significant. 2.12. Supplementary materials and methods Details of reagents; cell lines; RNA interference; cell viability of Ramos cells at pH 7.4 and 5.5 buffer conditions; flow cytometry; Annexin-V assay; generation of anti-α-tubulin cytotransmabs; Western blotting; quantitative Western blotting; cytosolic concentration and endosomal escape efficiency of cytotransmabs are detailed in the Supplementary Materials and methods. 3. Results

2.8. Trypan blue uptake assay

3.1. Internalized TMab4 escapes into the cytosol from acidified EEs

Each well of a 24-well plate was treated with 0.01% poly-L-lysine for 15 min at 25 °C. After washing the well, Ramos cells (4 × 105/well) were incubated for 30 min at 37 °C to allow them to attach to the poly-L-lysine-coated 24-well plate. For proteoglycan-deficient CHO-K1 mutant pgsD-677 cells, cells (2 × 105/well) were seeded in 24-well culture plate and cultured overnight. Then, cells were treated for 2 h at 37 °C with the indicated antibodies in pH 7.4 HBSS buffer (HBSS (0.37 g of KCl, 0.06 g of KH2PO4, 0.35 g of NaHCO3, 8 g of NaCl, 0.05 g of Na2HPO4, and 1 g of D-glucose per liter), 50 mM HEPES, pH 7.4) or pH 5.5 HBSS buffer and then further incubated with 5% (v/v) trypan blue dye solution for 1 min. After washing of cells, the images were acquired with a Zeiss AxioCam digital camera and Zeiss AxioVision software (Carl Zeiss). The trypan blue uptake efficiency was calculated as the percentage of trypan blue-stained cells among the incubated cells (n = 400 cells).

To determine the endocytic compartment from which TMab4 escapes to reach the cytosol, HeLa cells were pulsed with TMab4, which was followed by co-staining with distinct endosomal markers, including Rab5 for EEs, Rab11 for recycling endosomes and LysoTracker for LEs/lysosomes [10,26]. As a control, TAT peptide derived from the HIV-1 transactivator protein was employed because it is also internalized through the HSPG receptor [27,28]. During the chasing periods (0, 2 and 6 h immediately after internalization for 30 min), TMab4 had a diffuse cytosolic distribution and was also detected in punctate, endocytic vesicles that dominantly co-localized with Rab5, but neither with Rab11 nor LysoTracker even after internalization for 6 h (Fig. 1A), indicating that TMab4 escapes into the cytosol directly from EEs. Unlike TMab4, TAT rapidly trafficked to EEs as well as to LEs/lysosomes to be largely degraded, releasing a tiny amount of TAT into the cytosol (Fig. 1A), consistent with previous observations that TAT is released from LEs/lysosomes rather than EEs [10, 27,28]. To further delineate the cytosolic release pathways of TMab4, the effects of pharmacological inhibitors of endosomal maturation and acidification processes were examined on the intracellular localization of TMab4 and TMab4-mediated cytosolic release of the calcein tracer in HeLa cells. Calcein, a membrane-impermeable green fluorophore, was employed as a cargo tracer of cytosolic release from endosomes after passive endocytic uptake mediated by co-treated TMab4 or TAT [16]. Pretreatment of HeLa cells with bafilomycin, an inhibitor of the endosomal vesicular proton-ATPase that blocks the acidification of EEs and LEs [29], caused entrapment of TMab4 in endocytic vesicles, thereby blocking its cytosolic localization (Fig. S1) and the endosomal escape of endocytosed calcein into the cytosol (Fig. 1B), demonstrating the essential role of EE acidification in the endosomal escape of TMab4. However, preventing the maturation of EEs into LEs/lysosomes by the phosphoinositide 3-kinase inhibitor wortmannin [30] as well as the trafficking of EEs to the Golgi and endoplasmic reticulum (ER) by the retrotransport inhibitor brefeldin A [31] did not significantly affect the cytosolic access of TMab4 (Fig. S1) and co-treated calcein (Fig. 1B). Wortmannin and bafilomycin, but not brefeldin A, blocked the cytosolic release of TAT and co-treated calcein, leaving TAT and calcein entrapped in endosomes (Fig. 1B and Fig. S1). These results further support the idea that TMab4 is directly released into the cytosol from acidified EEs, whereas TAT escapes into the cytosol later from LEs/lysosomes rather than from EEs [27,28]. Side-by-side comparisons of TMab4 and TAT at several concentrations revealed that TMab4 more efficiently induced the cytosolic release of co-treated calcein (by ~ 1.3–1.8-fold) than TAT (Fig.

2.9. Membrane pore formation recovery assay This assay was performed using Trypan blue uptake assay in Ramos cells or pgsD-677 cells, as described in detail earlier, with the following modifications. Ramos cells (4 × 105) attached to poly-L-lysine coated 24-well plate or pgsD-677 cells (4 × 105) grown in 24-well plate were first treated for 2 h at 37 °C with 10 μM of TMab4 in pH 5.5 HBSS buffer. Then cells were washed carefully to remove TMab4 and then allowed to recover in fresh RPMI medium supplemented with 10% FBS for 2 h then further incubated with 5% (v/v) trypan blue dye solution for 1 min. Images were acquired with a Zeiss AxioCam digital camera and Zeiss AxioVision software (Carl Zeiss). The trypan blue uptake efficiency was calculated in percentage by counting the trypan blue stained cells out of incubated cells (n = 400 cells). 2.10. Co-localization of α-tubulin with anti-α-tubulin cytotransmabs HeLa cells (2 × 104) grown on 12-mm diameter coverslips in 24well culture plates were treated with TuT4 (3 μM) or TuT4-WYW (3 μM) for 6 h at 37 °C. The subsequent confocal microscopy was performed as described above. Cytosolic α-tubulin was labelled with antiα-tubulin antibody (Millipore) for 1 h at 25 °C, and subsequently with TRITC-conjugated goat anti-mouse antibody. 2.11. Statistical analysis Data represent the mean ± SD of triplicate samples from one representative experiment of three independent experiments, unless

168

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

Fig. 1. Cytosolic localization of TMab4 requires EE acidification. A. Pulse-chase intracellular trafficking of internalized TMab4 and FITC-TAT monitored by co-localization with the EE marker Rab5, recycling endosome marker Rab11, and LE/lysosome marker LysoTracker Red, as visualized with the indicated colours by confocal fluorescence microscopy. TMab4 (3 μM) or FITCTAT (20 μM) was pulsed for 30 min and chased for 0, 2 and 6 h in HeLa cells. Blue represents Hoechst332-stained nuclei. B. Cytosolic calcein release assay to assess the effects of wortmannin, bafilomycin and brefeldin A on the endosomal escape of TMab4 (3 μM) and TAT (20 μM) in HeLa cells. The right panel shows the percentage of the mean fluorescence intensity (MFI; arbitrary units) of calcein in cytoplasmic regions of cells relative to that in cells not treated with the inhibitors (control). Error bars, ±SD of 20 cells per sample. ***P b 0.001 versus the control using one-way ANOVA followed by Newman-Keuls post-hoc test. Scale bar, 10 μm.

S2), indicating the superior endosomal escape ability of TMab4 compared with that of TAT.

to dissociate TMab4 and TAT from HSPG for their endosomal escape into the cytosol.

3.2. Dissociation of TMab4 from the cell surface receptor HSPG by heparanase is required for endosomal escape

3.3. TMab4 induces membrane pore formation at an endosomal acidic pH

Like other examples of cell surface receptor catabolism, internalized HSPG eventually routes to lysosomes for degradation [32]. We questioned how TMab4 dissociates from HSPG in EEs for cytosolic release, instead of undergoing lysosomal degradation by further co-trafficking with HSPG. Heparanase, a mammalian endoglycosidase, is tethered to HSPG at the cell surface in an inactive form at a neutral pH and is internalized together with HSPG [33]. In acidified EEs, heparanase becomes active to cleave heparan sulfate glycosaminoglycans from HSPG [34]. Thus, we hypothesized that activated heparanase in acidified EEs cleaves heparin sulfate from HSPGs, leading to dissociation of TMab4 from HSPGs. To test this hypothesis, heparanase was knocked down by siRNA transfection of HeLa cells prior to chasing TMab4 (Fig. 2A). Intriguingly, in heparanase-knockdown HeLa cells, TMab4 was mainly detected in punctuate endosomes with a negligible diffuse cytosolic distribution and was further co-stained with lysosome-associated membrane protein (LAMP)-1 (Fig. 2B). Lysosomal trafficking of TAT was consistently observed, but its cytosolic release was also substantially reduced in heparanase-knockdown HeLa cells (Fig. 2B). Furthermore, knockdown of heparanase significantly diminished both TMab4- and TAT-mediated cytosolic release of calcein (Fig. 2C). These results indicate that activation of heparanase in response to an acidified endosomal pH is critical

The next question is how HSPG-dissociated TMab4 translocates from EEs to the cytosol. To address this question, the membrane translocation activity of TMab4 was assessed using human Burkitt lymphoma Ramos cells as a membrane model of EEs because these cells lack expression of HSPG on the cell surface, which excludes HSPG-mediated binding of TMab4, and the outer leaflet of the plasma membrane has the same overall lipid composition as the inner leaflet of EEs [35]. To simulate TMab4 behaviour within the lumen of acidified EEs, the assay was performed at pH 5.5, or at pH 7.4 for comparison. Noticeably, DyLight488labelled TMab4 entered the cytosol of living Ramos cells only at pH 5.5, but not at pH 7.4 (Fig. 3A). TMab4 facilitated the cellular uptake of co-treated DyLight488-labelled adalimumab (Humira®), a human IgG1 monoclonal antibody (mAb) against tumour necrosis factor, only at the acidic pH. Importantly, cells that had taken up the mAb maintained a spherical morphology without changes in their diameter (Fig. 3A), indicating that TMab4 does not cause cell lysis, but rather induces membrane pores that enable the passage of TMab4 itself and the cotreated adalimumab mAb into Ramos cells. The pH-dependent membrane pore formation ability of TMab4 was further assessed by monitoring cellular uptake of the co-treated trypan blue dye, a tracer of membrane pore formation in viable cells [36]. Buffer- or adalimumab-treated Ramos cells did not allow permeation of

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

169

Fig. 2. HSPG degradation by heparanase in acidified endosomes is essential for the endosomal escape of TMab4. A. Western blots showing the levels of heparanase in HeLa cells, which were transfected with scrambled (control) and heparanase-targeting siRNAs for 72 h. B. Intracellular trafficking of internalized TMab4 and FITC-TAT in heparanase-knockdown HeLa cells was monitored by co-localization with the lysosome marker LAMP-1, as visualized with the indicated colours by confocal fluorescence microscopy. TMab4 (3 μM) or FITC-TAT (20 μM) was pulsed for 30 min and chased for 2 h in HeLa cells transfected with the indicated siRNA for 72 h. Blue represents Hoechst332-stained nuclei. C. Cytosolic calcein release assay to assess the effects of heparanase knockdown on the endosomal escape of TMab4 (3 μM) and TAT (20 μM) in HeLa cells, which was performed after 72 h of the indicated siRNA transfection. The lower panel shows the MFI of calcein in cytoplasmic regions of cells compared with that in the PBS-treated control. Error bars, ±SD of 20 cells per sample. *P b 0.05, ***P b 0.001. In (B, C), scale bar, 10 μm.

trypan blue dye even at pH 5.5 (Fig. 3B). By contrast, TMab4 triggered the dose-dependent cellular uptake of trypan blue dye only at pH 5.5, but not at pH 7.4 (Fig. 3B). The viability of Ramos cells did not differ between pH 5.5 and 7.4 buffer conditions (Fig. S3), excluding the acidic pH buffer effect on the pH-dependent cellular uptake of trypan blue dye. The intact morphology of trypan blue-stained Ramos cells without lysis (Fig. 3B) further supports the idea that entry of TMab4 is mediated by a pore formation mechanism rather than by membrane lysis or collapse [37]. Ramos cells that were first treated with TMab4 at pH 5.5, washed to remove TMab4, allowed to recover for 2 h in fresh media and then exposed to trypan blue dye did not demonstrate trypan blue permeation (Fig. 3C), indicating that TMab4-induced membrane pore formation was transient and reversible [36]. We further investigated the pH-dependent trypan blue uptake in Chinese hamster ovary proteoglycan-deficient CHO-K1 mutant pgsD677 cells, a HSPG-deficient cell line [16], like the Ramos cells. In psgD677 cells, TMab4 triggered the dose-dependent cellular uptake of trypan blue dye only at pH 5.5, but not at pH 7.4 (Fig. S4A) and the TMab4-induced membrane pore formation at pH 5.5 was transient and reversible (Fig. S4B), which is in good agreement with the results of Ramos cells (Fig. 3B). 3.4. TMab4 binds to and destabilizes phospholipid membranes at an endosomal acidic pH To elucidate the pH-dependent membrane pore formation mechanism of TMab4, we first investigated pH-dependent membrane binding of TMab4 to HSPG-deficient Ramos and pgsD-677 cells by flow

cytometric analysis. Noticeably, Alexa488-labelled TMab4, but not Alexa488-labelled adalimumab, bound to Ramos and pgsD-677 cells only at pH 5.5, with negligible binding at pH 7.4 (Fig. 3D, Fig. S4C). The next question was how TMab4 induces formation of membrane pores after its membrane binding at an acidic pH. When membrane pore-forming amphipathic peptides, such as melittin and magainins, bind to phospholipid membranes, they induce local membrane destabilization and thinning, thereby leading to membrane lipid flip-flop and subsequently the formation of membrane toroidal pores [38,39]. We assessed the ability of TMab4 to trigger membrane lipid flip-flop by the annexin-V assay. Annexin-V specifically binds to phosphatidylserine, a component of the inner leaflet of the cell plasma membrane [40]. Thus, it can discriminate between the outer and inner leaflets, allowing the estimation of membrane lipid flip-flop by flow cytometric analysis. Annexin-V bound to TMab4-treated cells only at pH 5.5, but not at pH 7.4, whereas it did not react with adalimumab-treated cells at either pH in both Ramos (Fig. 3E) and pgsD-677 cells (Fig. S4D). Taken together, the above results suggest that, at acidified EEs, TMab4 binds to the endosomal membrane to destabilize it, which eventually leads to the formation of membrane toroidal pores composed of TMab4 and phospholipids for the endosomal escape of TMab4 (Fig. 3F). 3.5. Acidic pH-induced conformational changes of TMab4 trigger endosomal escape The next big question is how TMab4 responds to a mildly acidified endosomal pH in order to trigger endosomal escape. The shift to a

170

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

Fig. 3. TMab4 has the ability to induce membrane pores at an endosomal acidic pH. A. Representative microscopic images showing the plasma membrane permeability of antibodies in Ramos cells incubated with DyLight488-TMab4 (10 μM) or DyLight488-adalimumab (2 μM) together with TMab4 (10 μM) for 2 h at 37 °C and pH 7.4 or 5.5. Green represents DyLight488-labelled antibodies. Blue represents Hoechst332-stained nuclei. Scale bar, 5 μm. B and C. Representative microscopic images showing the uptake of trypan blue dye by Ramos cells co-treated with the indicated concentrations of antibodies for 2 h at 37 °C and pH 7.4 or 5.5. In ‘TMab4 (10 μM) + recovery’ shown in C, Ramos cells were first incubated with TMab4 (10 μM) alone at pH 5.5 for 2 h at 37 °C, washed to remove TMab4, allowed to recover in fresh medium for 2 h and then treated with trypan blue dye for 2 h prior to imaging. In (B, C), the graph show the percentage of trypan blue-stained cells among the incubated cells (n = 400 cells). In (A–C), scale bars, 10 μm. D. Flow cytometric analysis of binding of the indicated antibodies to the plasma membrane of Ramos cells at pH 7.4 or 5.5. E. Flow cytometric analysis of FITC-annexin-V binding to Ramos cells. Cells were incubated with the indicated antibodies (5 μM) for 1 h at 4 °C, washed to remove antibodies and then treated with FITC-annexin-V for 15 min at 25 °C. In (B, C, E), error bars, ±SD. *P b 0.05; **P b 0.01; ***P b 0.001 using one-tailed, unpaired Student's t-test (B, E) or one-way ANOVA followed by Newman-Keuls post-hoc test (C). F. Proposed model of membrane toroidal pores composed of TMab4 and phospholipids for TMab4 endosomal escape.

weakly acidic pH in the lumen of EEs can drive the carboxyl group of the negatively charged Asp and Glu residues to become more protonated and thereby more hydrophobic, facilitating new hydrophobic interactions with other hydrophobic residues [41,42]. This pH-dependent event is called the Tanford transition [43]. To identify such structural determinants of TMab4 that respond to the acidified endosomal pH, we carefully investigated the modelled light chain variable domain (VL) structure of TMab4 because the cytosol-penetrating activity of TMab4 is derived from the domain [16]. We identified AspL1 (where L = light chain with Kabat numbering) in the framework region and MetL95 in the CDR3 of the VL (VL-CDR3), which is 6.5 Å away, but can get closer (within 3 Å) without steric hindrance upon AspL1 protonation through favourable hydrophobic interactions between these two residues (Fig. 4A). The AspL1-MetL95 interaction can induce local structural changes in the MetL95-neighbouring residues of TyrL92, TyrL93 and HisL94 in VL-CDR3, causing the side chains to become upright and exposed from their kinked orientations toward VL frameworks (Fig. 4A). Tyr and His have a high and moderate propensity to interact with phospholipids constituting biological membranes, respectively [44]; therefore, we reasoned that the TyrL92, TyrL93 and HisL94 residues (dubbed YYH) of VL-CDR3 are the endosomal escape motif. To test our hypothesis, we first generated the TMab4-D1A and M95A variants, in which AspL1 and MetL95 were substituted with Ala, respectively (Fig. 4B). Indeed, these two IgG variants lost the ability to

induce uptake of trypan blue dye into Ramos cells even at pH 5.5 (Fig. 4B). By contrast, the TMab4-D1E variant, in which AspL1 was replaced with Glu, and the TMab4-M95 L variant, in which MetL95 was replaced with Leu, maintained the pH-dependent membrane pore formation ability (Fig. 4B). These mutational studies validated our hypothesis that the hydrophobic interaction of the protonatable acidic Asp/Glu residues at L1 with the hydrophobic Met/Leu residues at L95 at an acidified endosomal pH plays a critical role in the formation of membrane pores for the endosomal escape of TMab4.

3.6. The endosomal escape structural motif resides in VL-CDR3 of TMab4 Next, we validated the YYH motif composed of the TyrL92, TyrL93 and HisL94 residues in VL-CDR3 as the endosomal escape motif of TMab4 by Ala scanning (Fig. 4C). TMab4 variants with TyrL92Ala, TyrL93Ala or HisL94Ala substitution significantly lost the membrane pore formation ability for trypan blue dye permeation at pH 5.5, whereas variants with TyrL91Ala and TyrL96Ala mutation maintained this ability (Fig. 4C), excluding TyrL91 and TyrL96 as the critical endosomal escape motif. The TMab4-AAA variant with Ala substitution of the YYH motif completely lost the membrane pore formation ability at pH 5.5 (Fig. 4C), defining the YYH motif in VL-CDR3 as the structural determinant for the endosomal escape of TMab4.

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

171

Fig. 4. Biochemical identification of structural determinants of the TMab4 VL involved in the endosomal acidic pH-induced conformational changes and endosomal escape. A. Homology modeling of TMab4 Fv to highlight the interaction between AspL1 and MetL95 of the VL at an endosomal acidic pH. Protonation of AspL1 at pH 5.5 brings MetL95 toward AspL1, from a distance of 6.5 Å to 3 Å, by hydrophobic interactions, resulting in local structural rearrangement of the L95-neighbouring endosomal escape YYH motif in VL-CDR3 of TMab4 to facilitate interactions with endosomal membranes. Details are described in the text. The image was generated using COOT program. B–D Trypan blue dye uptake assay induced by TMab4 variants with the indicated mutations to identify the interaction pair of AspL1-MetL95 responding to the endosomal acidic pH (B), to determine three residues, TyrL92, TyrL93 and HisL94, in VLCDR3 of TMab4 as the endosomal escape motif (C), and to show the irrelevance of VL-CDR1 for the endosomal escape of TMab4 (D). Ramos cells were treated with trypan blue dye plus the indicated antibodies (10 μM) for 2 h at 37 °C and pH 7.4 or 5.5. The percentage of trypan blue-stained cells among the incubated cells (n = 400 cells) is represented. The sequence of TMab4–03 (D) is shown in Fig. S5A. Error bars, ±SD. **P b 0.01, ***P b 0.001, n.s., not significant versus TMab4-treated cells at pH 5.5 using one-way ANOVA followed by NewmanKeuls post-hoc test.

We previously demonstrated that cellular internalization of TMab4 through electrostatic interactions with HSPG is conferred by VL-CDR1 of TMab4 with the unique cationic patch composed of ArgL27f, ArgL29 and LysL30 [16]. To determine the role of VL-CDR1 in the endosomal escape of TMab4, we replaced both VL-CDR1 and VL-CDR2 with a germline-derived corresponding sequence containing the same number of amino acids from the most homologous framework to the VL of TMab4, generating the TMab4-03 variant (Fig. S5A). As expected, TMab4-03 failed to internalize into HeLa cells (Fig. S5B). However, TMab4-03 exhibited a membrane pore formation ability at pH 5.5 comparable to that of TMab4 in Ramos cells (Fig. 4D), indicating that the two functions of cellular internalization and endosomal escape of TMab4 are independently embedded in VL-CDR1 and VL-CDR3 of TMab4, respectively. 3.7. Rational design of TMab4 variants with improved endosomal escape activity To generate TMab4 with improved endosomal escape activity, we designed a series of TMab4 variants with permutations of Tyr and His residues on the YYH motif by Trp, which exhibits the highest propensity to interact with membrane-constituting phospholipids among amino acids [44,45]. Most of the TMab4 variants showed ~5–7-fold improved trypan blue uptake activity at pH 5.5, but also significantly improved

activity at pH 7.4 in Ramos cells (Fig. 5A). Noticeably, only the TMab4WYW variant exhibited pH-dependent membrane pore formation activity specifically at pH 5.5, but not at pH 7.4, showing ~5-fold enhanced activity (Fig. 5A). Compared with TMab4, TMab4-WYW also induced ~1.5–2-fold higher cytosolic release of co-treated calcein in HeLa cells (Fig. 5B), while showing a similar cellular internalization efficiency (Fig. S6). Proteins can interact with the head and/or tail groups of lipid molecules. To determine which interactions of the endosomal escape motif with either the head or tail group of membrane lipids are important for membrane pore formation, we generated the TMab4-RYR, TMab4IYI and TMab4-GYG variants by replacing Trp of the WYW motif with Arg, Ile and Gly, respectively (Fig. S7A). The propensities for interacting with the head and tail groups of phospholipids are in the order of Arg ≥ Trp N Tyr N Gly ≥ Ile and Trp ≥ Ile N Tyr N Gly N Arg, respectively [1, 44]. When assessed in Ramos cells, TMab4 and TMab4-RYR showed similar membrane-binding activity (Fig. S7B), but ~4–5-fold lower trypan blue uptake efficiency at pH 5.5 compared with TMab4-WYW (Fig. S7C). TMab4IYI and TMab4-GYG showed negligible activities for both membrane binding and trypan blue uptake at pH 5.5, indicating that binding of the endosomal escape motif to the head group of phospholipids is the first event for efficient membrane pore formation. The trypan blue uptake activities of TMab4 variants in Ramos cells related well with their endosomal escape activities for co-treated calcein in HeLa cells (Fig. S7D).

172

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

Fig. 5. Generation of TMab4 variants with improved endosomal escape activity. A. Trypan blue dye uptake assay induced by TMab4 variants with the indicated mutations to determine their endosomal escape ability. Ramos cells were treated with trypan blue dye plus 1 μM of the indicated antibodies for 2 h at 37 °C and pH 7.4 or 5.5. The percentage of trypan bluestained cells among the incubated cells (n = 400 cells) is represented. Error bars, ±SD. B. Cytosolic calcein release assay to assess the endosomal escape ability of TMab4 and TMab4WYW in HeLa cells treated with the indicated concentrations. The right panel shows the MFI of calcein in cytoplasmic regions of cells compared with that in the PBS-treated control. Error bars, ±SD of 20 cells per sample. C. Confocal microscopic analysis of complemented GFP signals in HeLa-SA-GFP1-10 cells, which were treated at 37 °C for 6 h with the indicated concentrations of TMab4-GFP11-SBP2 or TMab4-WYW-GFP11-SBP2. The assay is detailed in Fig. S8. Blue represents Hoechst332-stained nuclei. The right panel shows the MFI of GFP in cytoplasmic regions of cells compared with that in the PBS-treated control. Error bars, ±SD of 20 cells per sample. In (B, C), scale bar, 10 μm. In (A–C), *P b 0.05, **P b 0.01 and ***P b 0.001 versus TMab4-treated cells at each pH using one-way ANOVA followed by Newman-Keuls post-hoc test (A) or one-tailed, unpaired Student's t-test (B, C).

These results indicate that the strong interactions with both the head and tail groups of membrane lipids are critical for the improved endosomal escape motif, as evidenced by the Trp residues in TMab4-WYW. Next, we sought to quantify the cytosolic amount of TMab4-WYW in comparison with TMab4 using the recently developed enhanced splitGFP (green fluorescent protein) complementation assay [17]. Extracellular treatment of HeLa cells expressing streptavidin-GFP1–10 in the cytosol with TMab4-WYW genetically fused to GFP11-SBP2 (streptavidinbinding peptide 2) resulted in complemented GFP fluorescence only when extracellular TMab4-WYW reached the cytosol after its cellular internalization (Fig. S8). TMab4-WYW exhibited complemented GFP fluorescence in the cytosol of HeLa cells in proportion to the concentration, with N 2-fold higher levels than those of TMab4 at each concentration (Fig. 5C). Quantification of the cytosolic amount of TMab4-WYW revealed that the cytosolic concentrations of TMab4-WYW were ~ 3fold higher than those of TMab4, being about 34, 232 and 527 nM after extracellular treatment with 0.1, 0.5 and 1 μM TMab4-WYW for 6 h, respectively (Table S1), despite almost the same amount of TMab4-WYW and TMab4 being internalized by cells, as determined by quantitative Western blotting (Fig. S9, Table S1). When the endosomal escape efficiency was determined by dividing the cytosolic amount by the internalized amount, TMab4-WYW exhibited a much higher endosomal escape efficiency (~13%) than TMab4 (~4.3%) when the initial extracellular concentration was 1 μM (Supplementary Table 1).

3.8. Generation of a cytosolic protein-targeting cytotransmab with improved endosomal escape efficiency To exploit the enhanced endosomal escape ability of the VL of TMab4-WYW, we generated cytotransmabs that can directly target cytosolic microtubules by recognizing their α-tubulin component. HC with the VH originated from anti-α-tubulin 2G4 scFv [46] was paired with the LC of TMab4 and TMab4-WYW to generate TuT4 and TuT4WYW IgG-format cytotransmabs, respectively (Fig. 6A). When TuT4 and TuT4-WYW were incubated at the same concentration with HeLa cells, TuT4-WYW showed much better direct recognition of cytosolic microtubules, with a staining pattern of the distinctive filamentous cytoskeleton structure, than TuT4 after cellular internalization (Fig. 6B). This result can be attributed to the WYW motif in the VL of TuT4WYW, which conferred an ~ 3-fold higher endosomal escape ability than the YYH motif in the VL of TuT4. 4. Discussion With their unique endosomal escape ability, cytotransmabs are a new class of antibodies that reach the cytosol after internalization into living cells without affecting cell viability or proliferation. Here, we elucidated the detailed endosomal escape mechanism by which the internalized TMab4 cytotransmab is released into the cytosol of cells (Fig. 7). We found that the acidified pH of EEs played a critical role in 1)

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

Fig. 6. Generation of a cytosolic protein-targeting cytotransmab with an improved endosomal escape motif. A. Diagram of α-tubulin-targeting cytotransmabs, TuT4 and TuT4-WYW, generated by replacing the VH domain of TMab4 and TMab4-WYW with the α-tubulin-binding 2G4 VH, respectively. B. Co-localization of internalized TuT4 (green) or TuT4-WYW (green) with α-tubulin (red) in HeLa cells. Cells were treated with TuT4 (3 μM) or TuT4-WYW (3 μM) for 6 h at 37 °C and then examined by confocal fluorescence microscopy. Blue represents Hoechst332-stained nuclei. Scale bar, 10 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

activating heparanase to degrade HSPG, resulting in dissociation of TMab4 from HSPG and 2) protonating the carboxyl group of negatively charged residues at L1, such as Asp and Glu, to trigger hydrophobic interactions with hydrophobic residues at L95, such as Met and Leu, which subsequently causes local structural rearrangement of the YYH endosomal escape motif composed of TyrL92, TyrL93 and HisL94 in VL-CDR3 of TMab4 to interact with endosomal membranes. The interactions of the YYH motif with endosomal membranes resulted in membrane destabilization that caused membrane lipid flip-flop, subsequently leading to membrane pore formation for the transport of TMab4 to the cytosol from EEs. To our knowledge, our study is the first to describe the endosomal escape mechanism of an antibody, which has the ability to induce membrane pore formation via the endosomal escape motif in VL-CDR3. Furthermore, although pH-dependent antigen-binding antibodies with His residues in the CDR paratopes have been reported [47], antibodies that undergo pH-dependent interaction between residues in the framework region and CDR to induce local structural rearrangement in the CDR were hitherto not reported. Such pH-dependent conformational changes of protein in response to the endosomal acidic pH are observed in the endosomal escape of bacterial toxins, viral fusogenic peptides and CPPs [2,8,9]. The acidification of endosomal pH can protonate the side chains of Asp, Glu and/or His residues, leading to the necessary secondary structural changes (most commonly transition of a random coil to an α-helix) of the endosomal escape motifs/peptides to mediate membrane interaction for endosomal escape through membrane pore formation [8,37]. However, the acidic pH-induced conformational change of TMab4 is not due to local secondary structural changes, but allosteric tertiary

173

structural changes that expose the endosomal escape YYH motif in VLCDR3 as a result of hydrophobic interactions between AspL1 and MetL95 (Fig. 4). Then, the YYH motif with clustered aromatic side chains seems to intercalate into the endosomal membrane lipid bilayer. We speculate that, as a threshold concentration of embedded TMab4 is reached in a particular area of the membrane, the membrane becomes locally destabilized and thinned, as evidenced by membrane lipid flipflop (Fig. 5), subsequently leading to formation of a pore composed of TMab4 and phospholipids (Fig. 3F), like the toroidal proteolipidic pore models for melittin and magainin peptides [38,48]. The TMab4-induced membrane pore seems large enough to transport TMab4 itself (150 kDa) and the bystanders (trypan blue (0.96 kDa) and adalimumab (150 kDa)) into cells, but is neither significant nor long-lived enough to cause cytotoxicity (Fig. 3). This toroidal pore formation ability of TMab4 is activated only in acidified EEs, but not in the plasma membrane or other membranous organelles with a neutral pH environment. This pH-dependent activity is an important property of TMab4 that explains its negligible toxicity to living cells [16], thus such antibodies hold great potential as research and therapeutic agents. Noticeably, the two separate functions of cellular internalization and endosomal escape of TMab4 were independently embedded in VLCDR1 and VL-CDR3 of TMab4, respectively. As shown with TMab4–03 (Fig. 4, Fig. S5), retention of the VL-CDR3 residues and AspL1 was sufficient to maintain the pH-dependent membrane pore formation activity of the antibody. Since most of human VL germline sequence has Asp or Glu at L1, the endosomal escape activity can be conferred to other conventional antibodies by appropriate grafting the VL-CDR3, particularly the YYH or its improved WYW endosomal escape motif. The essential role of heparanase activity for the endosomal escape of TMab4 and TAT by dissociating them from the receptor HSPG is an important finding (Fig. 2). Although cationic CPPs utilize HSPGs as receptors for cellular uptake, how they dissociate from HSPGs to reach the cytosol has not been clarified. Our results suggest that activation of heparanase in the acidified endosomes functions to make HSPG-sequestered cationic CPP free for the subsequent endosomal escape into the cytosol. Like other endosomal escape proteins and CPPs, the endosomal escape efficiency of the prototype TMab4 was poor (Table S1). However, identification of the YYH endosomal escape motif led us to design the TMab4-WYW variant with Trp residues, which are more lipophilic than Tyr or His. TMab4-WYW showed ~3-fold improved endosomal escape efficiency, leading to higher cytosolic concentrations of TMab4WYW than of TMab4, upon extracellular treatment of living cells (Figs. 5 and 6). As shown in our systematic mutation studies (Fig. 5), the positioning of Trp, rather than the number of Trp residues in the endosomal escape motif, was critical for the acidic pH-induced membrane pore formation activity, reminiscent of the position effect of Trp or Arg on the membrane interaction of Arg-rich CPP [10,49]. However, the exact mechanism remains to be determined.

5. Conclusion Engineering an antibody to have efficient endosomal escape ability after receptor-mediated endocytosis will provide a great opportunity to design an antibody that exhibits an effect in the cytosol and for antibody-based agents that carry a wide variety of biologically active cargos into the cytosol. As shown with TuT4-WYW (Fig. 6), cytosol proteintargeting cytotransmabs can be generated by combining a cytosol-penetrating LC with a HC that targets cytosolic proteins. Our elucidation of the endosomal escape mechanism of this cytotransmab and its improvement will establish a platform technology that enables a fulllength IgG antibody to directly target cytosolic proteins. Such cytosolpenetrating antibodies can be developed as analytical and functional research agents in living cells and as imaging and therapeutic agents that target cytosolic proteins.

174

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175

Fig. 7. Schematic diagram showing the endosomal escape mechanism of TMab4 cytotransmab. 1) TMab4 internalizes into cells using HSPG as the receptor and TMab4-containing vesicles routes to early endosomes. 2) In the acidified EE lumen, heparanase becomes active to cleave heparan sulfate glycosaminoglycans from HSPG, resulting in dissociation of TMab4 from HSPG. 3) TMab4 undergoes endosomal acidic pH-induced conformational changes, thereby binding to the endosomal membrane to destabilize it, which eventually leads to the formation of membrane toroidal pores composed of TMab4 and phospholipid for the endosomal escape of TMab4 into the cytosol. 4) Consequently, endocytosed TMab4 escapes into the cytosol from early endosomes without further trafficking to LEs/lysosomes or the nucleus through Golgi and ER.

Acknowledgements This work was supported by grants from the Pioneer Research Center Program (2014M3C1A3051470) and the Global Frontier Project (2013M3A6A4043874) from the National Research Foundation funded by the Korean government. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jconrel.2016.05.066. References [1] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis, Nat. Methods 9 (2012) 671–675.

[2] S. Guillard, R.R. Minter, R.H. Jackson, Engineering therapeutic proteins for cell entry: the natural approach, Trends Biotechnol. 33 (2015) 163–171. [3] A. Marschall, C. Zhang, A. Frenzel, T. Schirrmann, M. Hust, F. Perez, S. Dubel, Delivery of antibodies to the cytosol: Debunking the myths, mAbs 6 (2014) 943–956. [4] M. Ritchie, L. Tchistiakova, N. Scott, Implications of receptor-mediated endocytosis and intracellular trafficking dynamics in the development of antibody drug conjugates, mAbs 5 (2013) 13–21. [5] G.J. Doherty, H.T. McMahon, Mechanisms of endocytosis, Annu. Rev. Biochem. 78 (2009) 857–902. [6] Y. Lee, T. Ishii, H.J. Kim, N. Nishiyama, Y. Hayakawa, K. Itaka, K. Kataoka, Efficient delivery of bioactive antibodies into the cytoplasm of living cells by charge-conversional polyion complex micelles, Angew. Chem. Int. Ed. Eng. 49 (2010) 2552–2555. [7] Y. Yamada, S.M. Perez, M. Tabata, J. Abe, Y. Yasuzaki, H. Harashima, Efficient and high-speed transduction of an antibody into living cells using a multifunctional nanocarrier system to control intracellular trafficking, J. Pharm. Sci. 104 (2015) 2845–2854. [8] H.K. Shete, R.H. Prabhu, V.B. Patravale, Endosomal escape: a bottleneck in intracellular delivery, J. Nanosci. Nanotechnol. 14 (2014) 460–474. [9] A.K. Varkouhi, M. Scholte, G. Storm, H.J. Haisma, Endosomal escape pathways for delivery of biologicals, J. Control. Release 151 (2011) 220–228.

J.-S. Kim et al. / Journal of Controlled Release 235 (2016) 165–175 [10] J.S. Appelbaum, J.R. LaRochelle, B.A. Smith, D.M. Balkin, J.M. Holub, A. Schepartz, Arginine topology controls escape of minimally cationic proteins from early endosomes to the cytoplasm, Chem. Biol. 19 (2012) 819–830. [11] A. Erazo-Oliveras, N. Muthukrishnan, R. Baker, T.Y. Wang, J.P. Pellois, Improving the endosomal escape of cell-penetrating peptides and their cargos: strategies and challenges, Pharmaceuticals 5 (2012) 1177–1209. [12] M. Li, Y. Tao, Y. Shu, J.R. LaRochelle, A. Steinauer, D. Thompson, A. Schepartz, Z.Y. Chen, D.R. Liu, Discovery and characterization of a peptide that enhances endosomal escape of delivered proteins in vitro and in vivo, J. Am. Chem. Soc. 137 (2015) 14084–14093. [13] X. Liao, A.E. Rabideau, B.L. Pentelute, Delivery of antibody mimics into mammalian cells via anthrax toxin protective antigen, Chembiochem 15 (2014) 2458–2466. [14] W.P. Verdurmen, M. Luginbuhl, A. Honegger, A. Pluckthun, Efficient cell-specific uptake of binding proteins into the cytoplasm through engineered modular transport systems, J. Control. Release 200 (2015) 13–22. [15] N.J. Yang, D.V. Liu, D. Sklaviadis, D.Y. Gui, M.G. Vander Heiden, K.D. Wittrup, Antibody-mediated neutralization of perfringolysin o for intracellular protein delivery, Mol. Pharm. 12 (2015) 1992–2000. [16] D.K. Choi, J. Bae, S.M. Shin, J.Y. Shin, S. Kim, Y.S. Kim, A general strategy for generating intact, full-length IgG antibodies that penetrate into the cytosol of living cells, mAbs 6 (2014) 1402–1414. [17] J.S. Kim, D.K. Choi, S.W. Park, S.M. Shin, J. Bae, D.M. Kim, T.H. Yoo, Y.S. Kim, Quantitative assessment of cellular uptake and cytosolic access of antibody in living cells by an enhanced split GFP complementation assay, Biochem. Biophys. Res. Commun. 467 (2015) 771–777. [18] C.H. Lee, K.J. Park, E.S. Sung, A. Kim, J.D. Choi, J.S. Kim, S.H. Kim, M.H. Kwon, Y.S. Kim, Engineering of a human kringle domain into agonistic and antagonistic binding proteins functioning in vitro and in vivo, Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 9567–9571. [19] T.H. Shin, E.S. Sung, Y.J. Kim, K.S. Kim, S.H. Kim, S.K. Kim, Y.D. Lee, Y.S. Kim, Enhancement of the tumor penetration of monoclonal antibody by fusion of a neuropilintargeting Peptide improves the antitumor efficacy, Mol. Cancer Ther. 13 (2014) 651–661. [20] Y.J. Kim, J. Bae, T.H. Shin, S.H. Kang, M. Jeong, Y. Han, J.H. Park, S.K. Kim, Y.S. Kim, Immunoglobulin Fc-fused, neuropilin-1-specific peptide shows efficient tumor tissue penetration and inhibits tumor growth via anti-angiogenesis, J. Control. Release 216 (2015) 56–68. [21] N.R. Whitelegg, A.R. Rees, WAM: an improved algorithm for modelling antibodies on the WEB, Protein engineering 13 (2000) 819–824. [22] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta crystallographica, Sect. D Biol. Crystallogr. 60 (2004) 2126–2132. [23] J.Y. Jang, J.G. Jeong, H.R. Jun, S.C. Lee, J.S. Kim, Y.S. Kim, M.H. Kwon, A nucleic acid-hydrolyzing antibody penetrates into cells via caveolae-mediated endocytosis, localizes in the cytosol and exhibits cytotoxicity, Cell. Mol. Life Sci. 66 (2009) 1985–1997. [24] A. Kim, T.H. Shin, S.M. Shin, C.D. Pham, D.K. Choi, M.H. Kwon, Y.S. Kim, Cellular internalization mechanism and intracellular trafficking of filamentous M13 phages displaying a cell-penetrating transbody and TAT peptide, PLoS One 7 (2012), e51813. [25] F. Salomone, F. Cardarelli, M. Di Luca, C. Boccardi, R. Nifosi, G. Bardi, L. Di Bari, M. Serresi, F. Beltram, A novel chimeric cell-penetrating peptide with membrane-disruptive properties for efficient endosomal escape, J. Control. Release 163 (2012) 293–303. [26] M. Moisenovich, A. Tonevitsky, N. Maljuchenko, N. Kozlovskaya, I. Agapov, W. Volknandt, J. Bereiter-Hahn, Endosomal ricin transport: involvement of Rab4- and Rab5-positive compartments, Histochem. Cell Biol. 121 (2004) 429–439. [27] A. Erazo-Oliveras, K. Najjar, L. Dayani, T.Y. Wang, G.A. Johnson, J.P. Pellois, Protein delivery into live cells by incubation with an endosomolytic agent, Nat. Methods 11 (2014) 861–867. [28] S.T. Yang, E. Zaitseva, L.V. Chernomordik, K. Melikov, Cell-penetrating peptide induces leaky fusion of liposomes containing late endosome-specific anionic lipid, Biophys. J. 99 (2010) 2525–2533.

175

[29] T. Yoshimori, A. Yamamoto, Y. Moriyama, M. Futai, Y. Tashiro, Bafilomycin A1, a specific inhibitor of vacuolar-type H(+)-ATPase, inhibits acidification and protein degradation in lysosomes of cultured cells, J. Biol. Chem. 266 (1991) 17707–17712. [30] G. Li, C. D'Souza-Schorey, M.A. Barbieri, R.L. Roberts, A. Klippel, L.T. Williams, P.D. Stahl, Evidence for phosphatidylinositol 3-kinase as a regulator of endocytosis via activation of Rab5, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 10207–10211. [31] T. Yoshida, C.C. Chen, M.S. Zhang, H.C. Wu, Disruption of the Golgi apparatus by brefeldin A inhibits the cytotoxicity of ricin, modeccin, and Pseudomonas toxin, Exp. Cell Res. 192 (1991) 389–395. [32] H.C. Christianson, M. Belting, Heparan sulfate proteoglycan as a cell-surface endocytosis receptor, Matrix Biol. 35 (2014) 51–55. [33] S. Gingis-Velitski, A. Zetser, V. Kaplan, O. Ben-Zaken, E. Cohen, F. Levy-Adam, Y. Bashenko, M.Y. Flugelman, I. Vlodavsky, N. Ilan, Heparanase uptake is mediated by cell membrane heparan sulfate proteoglycans, J. Biol. Chem. 279 (2004) 44084–44092. [34] M. Belting, Heparan sulfate proteoglycan as a plasma membrane carrier, Trends Biochem. Sci. 28 (2003) 145–151. [35] C. Bissig, J. Gruenberg, Lipid sorting and multivesicular endosome biogenesis, Cold Spring Harb. Perspect. Biol. 5 (2013) a016816. [36] S.L. Tran, A. Puhar, M. Ngo-Camus, N. Ramarao, Trypan blue dye enters viable cells incubated with the pore-forming toxin HlyII of Bacillus cereus, PLoS One 6 (2011), e22876. [37] Y.J. Lee, G. Johnson, J.P. Pellois, Modeling of the endosomolytic activity of HA2-TAT peptides with red blood cells and ghosts, Biochemistry 49 (2010) 7854–7866. [38] R.J. Gilbert, M. Dalla Serra, C.J. Froelich, M.I. Wallace, G. Anderluh, Membrane pore formation at protein-lipid interfaces, Trends Biochem. Sci. 39 (2014) 510–516. [39] F.Y. Chen, M.T. Lee, H.W. Huang, Evidence for membrane thinning effect as the mechanism for peptide-induced pore formation, Biophys. J. 84 (2003) 3751–3758. [40] G. Koopman, C.P. Reutelingsperger, G.A. Kuijten, R.M. Keehnen, S.T. Pals, M.H. van Oers, Annexin V for flow cytometric detection of phosphatidylserine expression on B cells undergoing apoptosis, Blood 84 (1994) 1415–1420. [41] T. Korte, K. Ludwig, A. Herrmann, ph-dependent hydrophobicity profile of hemagglutinin of influenza virus and its possible relevance in virus fusion, Biosci. Rep. 12 (1992) 397–406. [42] N.V. Di Russo, D.A. Estrin, M.A. Marti, A.E. Roitberg, pH-Dependent conformational changes in proteins and their effect on experimental pK(a)s: the case of Nitrophorin 4, PLoS Comput. Biol. 8 (2012), e1002761. [43] B.Y. Qin, M.C. Bewley, L.K. Creamer, H.M. Baker, E.N. Baker, G.B. Jameson, Structural basis of the Tanford transition of bovine beta-lactoglobulin, Biochemistry 37 (1998) 14014–14023. [44] M. Morita, A.M. Katta, S. Ahmad, T. Mori, Y. Sugita, K. Mizuguchi, Lipid recognition propensities of amino acids in membrane proteins from atomic resolution data, BMC Biophys. 4 (2011) 21. [45] K.A. Browne, E. Blink, V.R. Sutton, C.J. Froelich, D.A. Jans, J.A. Trapani, Cytosolic delivery of granzyme B by bacterial toxins: evidence that endosomal disruption, in addition to transmembrane pore formation, is an important function of perforin, Mol. Cell. Biol. 19 (1999) 8604–8615. [46] L. Guglielmi, V. Denis, N. Vezzio-Vie, N. Bec, P. Dariavach, C. Larroque, P. Martineau, Selection for intrabody solubility in mammalian cells using GFP fusions, Protein Eng. Des. Sel. 24 (2011) 873–881. [47] T. Igawa, S. Ishii, T. Tachibana, A. Maeda, Y. Higuchi, S. Shimaoka, C. Moriyama, T. Watanabe, R. Takubo, Y. Doi, T. Wakabayashi, A. Hayasaka, S. Kadono, T. Miyazaki, K. Haraya, Y. Sekimori, T. Kojima, Y. Nabuchi, Y. Aso, Y. Kawabe, K. Hattori, Antibody recycling by engineered pH-dependent antigen binding improves the duration of antigen neutralization, Nat. Biotechnol. 28 (2010) 1203–1207. [48] K. Matsuzaki, O. Murase, N. Fujii, K. Miyajima, An antimicrobial peptide, magainin 2, induced rapid flip-flop of phospholipids coupled with pore formation and peptide translocation, Biochemistry 35 (1996) 11361–11368. [49] M.L. Jobin, M. Blanchet, S. Henry, S. Chaignepain, C. Manigand, S. Castano, S. Lecomte, F. Burlina, S. Sagan, I.D. Alves, The role of tryptophans on the cellular uptake and membrane interaction of arginine-rich cell penetrating peptides, Biochim. Biophys. Acta 1848 (2015) 593–602.