Polyhistidine facilitates direct membrane translocation of cell-penetrating peptides into cells

The bovine lactoferricin L6 (RRWQWR) has been previously identified as a novel cell-penetrating peptide (CPP) that is able to efficiently internalize into human cells. L6 interacts with quantum dots (QDs) noncovalently to generate stable L6/QD complexes that enter cells by endocytosis. In this study, we demonstrate a modified L6 (HL6; CHHHHHRRWQWRHHHHHC), in which short polyhistidine peptides are introduced into both flanks of L6, has enhanced cell-penetrating ability in human bronchoalveolar carcinoma A549 cells. The mechanism of cellular uptake of HL6/QD complexes is primarily direct membrane translocation rather than endocytosis. Dimethyl sulfoxide (DMSO), but not pyrenebutyrate (PB), ethanol, oleic acid, or 1,2-benzisothiazol-3(2 H)-one (BIT), slightly enhances HL6-mediated protein transduction efficiency. Neither HL6 nor HL6/QD complexes are cytotoxic to A549 or HeLa cells. These results indicate that HL6 could be a more efficient drug carrier than L6 for biomedical as well as biotechnological applications, and that the function of polyhistidine peptides is critical to CPP-mediated protein transduction.

transient pore formation or membrane destabilization 2 . The direct membrane translocation pathway may be a more effective route to convey CPP-derived pharmaceutics and therapeutics, insofar as it does not entail endosomal entrapment 14 .
Administration concentrations of arginine-rich CPPs employed during protein transduction played a vital part in identifying two internalization pathways employed by CPPs 15 . For instance, it was found in one study that endocytosis is the main pathway at low CPP concentrations, while direct membrane translocation is noticeable at higher CPP concentrations than the threshold of 5 μM 15 . Arginine-rich CPPs dynamically alter local membrane structure during direct membrane translocation 15 , which originates from nucleation zones (NZs) 16 . These spatially restricted NZ regions on the plasma membrane provide potent platforms for CPP internalization 16 . Enzymatic digestion with heparinases strongly suppresses direct membrane translocation of arginine-rich CPPs, indicating a role of heparan sulfates on cell surface in internalization through NZs 15,16 .
In our previous study, a bovine lactoferricin L6 CPP 17 has been demonstrated to noncovalently interact with carboxyl-functionalized quantum dots (QDs) and form stable L6/QD complexes that can enter cells by endocytosis 18 . The aim of the present study was to assess whether modification of L6 with additional penta-histidine and mono-cysteine residues on both sides would influence its 1) ability to form noncovalent complexes with QDs, 2) transduction efficiency, and 3) mechanism of cellular uptake.  12 , and FITC-nonCPP (RPPGFSPFR, bradykinin) 19 peptides were purchased from Genomics (Taipei, Taiwan). Carboxyl-functionalized and water-soluble QDs (CdSe/ZnS-PEG-COOH-520) with a maximal emission wavelength of 523 nm were purchased from Mesolight (Zhejiang, China).

Gel retardation assay.
To characterize noncovalent interactions between CPPs and QDs, a series of concentrations of HL6 ranging from 0 to 2.5 μM were incubated with 0.1 μM of QDs at 37 °C for 1 h. Complexes formed at molar ratios of 0 (QDs alone), 5, 10, 15, 20, and 25 were analyzed by electrophoresis on a 1% SeaKem Gold agarose gel (Lonza Group, Basel, Switzerland) in 0.5× TBE buffer (44.5 mM of Tris-borate and 1 mM of EDTA, pH 8.3) at 50 V for 30 min 20 . ChemiDoc XRS+ gel imaging system (Bio-Rad, Hercules, CA, USA) was used to capture and analyze gel images 17 . The relevant shift-percentage of a QD band shift in the gel is defined as a reciprocal mobility ratio normalized to scale in which the minimal molar ratio of 0 was 0%, and the maximal ratio of 25 was 100%. Noncovalent protein transduction. A549 cells were washed with phosphate buffered saline (PBS) and changed to serum-free RPMI 1640 medium for protein transduction experiments 21 . The presence of 10% serum in the medium as the solvent caused a decrease in the uptake of CPPs or their complexes/conjugates. For cellular internalization, cells were seeded at a density of about 1 × 10 5 per 35-mm petri dish and then treated with serum-free medium (as a control), 80 μM of FITC isomer I (Sigma-Aldrich, St. Louis, MO, USA), FITC-nonCPP, FITC-L6, FITC-HL6, or FITC-HR9 at 37 °C for either 1 h or various time durations. The fluorescent tracker Hoechst 33342 was used to stain the nuclei of cells in blue, according to the manufacturer's instructions (Thermo Fisher).
To analyze protein transduction of noncovalent CPP/QD complexes, 6 μM of HL6 peptides were premixed with 0.3 μM of QDs at 37 °C for 1 h. After complex formation, the mixtures were incubated with A549 cells at 37 °C for various durations 21 . Cells were washed with PBS to remove non-transduced material from cell surface.

Fluorescent microscopy.
Bright-field and fluorescent images were captured using an AE31 inverted Epi-fluorescence microscope (Motic, Hong Kong, China) equipped with an IS1000 eyepiece (Tucsen, Fujian, China) 17 . Excitation filter modules of green and blue fluorescent protein (GFP, BFP) channels were set at 480/30 and 350/50 nm, respectively. Emission filter modules of GFP and BFP channels were set at 535/40 and 460/50 nm, respectively.
Flow cytometric analysis. Cells were seeded at a density of about 2 × 10 5 per well in a 24-well plate.
Following the protein transduction, cells were washed with PBS and harvested by adding 250 μl/well of 2.9 mM ethylenediaminetetraacetic acid (EDTA, pH 6.14) 22 . Samples were analyzed using a Cell Lab Quanta SC MPL flow cytometer (Beckman Coulter, Fullerton, CA, USA) possessed with 488 nm laser at a speed of 20 μl/min. Digital measurements were analyzed using Cell Lab Quanta SC MPL software version 1.0 (Beckman Coulter). Results are shown as the percentage of either the total cell population displaying green fluorescence or internalization efficiency.
Cytotoxicity assay. To assess the cytotoxicity of HL6, QDs, and HL6/QD complexes, both human A549 and HeLa cells were treated at 37 °C for 30 min with 6 μM of HL6 alone, 0.3 μM of QDs alone, or HL6/QD complexes (prepared at a ratio of 20 at 37 °C for 1 h). One hundred % DMSO and serum-free medium were used to treat cells for 30 min as positive and negative controls, respectively. HeLa cells were pretreated with 4 °C for 30 min or heparinases I, II, and III for 6 h. Cells were then treated with HL6/QD complexes (prepared at a ratio of 20 at 37 °C for 1 h) in the absence or presence of various endocytic inhibitors at 37 °C for 30 min. After treatment, cells were washed with PBS and then incubated in complete culture medium at 37 °C for 24 h. Eighty μl of serum-free medium and 20 μl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich; 5 mg/ml in PBS) solution were added to each well. After incubation at 37 °C for 30 min, soluble formazan crystals converted only by metabolically living cells were dissolved in 300 μl of DMSO. The 24-well plates were then read using a SpectraMax M2 microplate reader with SoftMax Pro 6.3 microplate analysis software (Molecular Devices, Sunnyvale, CA, USA) at 540 nm wavelength. statistical analysis. Data are shown as mean ± standard deviation (SD) from more than three independent experiments performed in triplicates for each treatment group. Statistical comparisons were carried out by One-way ANOVA, using SigmaPlot software version 12.5 (Systat Software, San Jose, CA, USA) at levels of statistical significance when the P-value was less than 0.05 (*, α) or 0.01 (**, αα).

Results
Cellular internalization of HL6. HL6 internalization by human A549 cells was characterized by treating cells with FITC-CPPs and other vehicles at 37 °C for 1 h, followed by staining with Hoechst 33342. Fluorescent microscopy demonstrates green fluorescence in the cells treated with either FITC-L6 or FITC-HL6 (Fig. 1). Cells treated with serum-free medium, FITC, and FITC-nonCPP did not display green fluorescence, indicating that HL6 when comprised of L6 and polyhistidine is an effective CPP.
A time course analysis of cellular internalization of HL6. The kinetics of HL6 internalization were evaluated using cells treated with FITC-HL6, FITC-L6 as an endocytic control (unpublished observations), or FITC-HR9 as a direct translocation control 12 for various time durations, and then analyzed using a flow cytometer. FITC-L6 and control groups showed little green fluorescence at 0 min (Fig. 2). Green fluorescence was present in FITC-HL6 and FITC-HR9 groups within 5 min. Cellular internalization of FITC-HL6 and FITC-HR9 was consistently higher than FITC-L6 or control. At 60 min, uptake in the FITC-HL6 group was 46.7 times higher than in the FITC-L6 group. The fraction of cells showing uptake was higher with FITC-HR9-treated cells than FITC-HL6 and FITC-L6 at all time points. Endocytosis involves formation of endosomes that requires at least 5-15 min to form 10,11 , while HR9 has been proven to enter cells by direct membrane translocation 12 . Collectively, these results indicate that HL6 enters cells by direct membrane translocation just as HR9, but with a relatively lower efficiency.

HL6 CPPs and QDs in vitro.
The gel retardation assay was performed to determine whether HL6 CPPs form stable, noncovalent complexes with QDs. HL6 and QDs were incubated at molar ratios varying from 0 to 25. QDs mobility was found to be inversely related to the concentration of HL6 being incubated (Fig. 3A, Supplementary  Fig. S1). Using relative shift analysis, ratio-dependent interactions of HL6/QD complexes were maximized at ratios above 15 (Fig. 3B). Accordingly, a molar ratio of 20 was selected and applied in subsequent experiments. Intracellular delivery of QDs using CPPs. The QD delivery capabilities of HL6 CPPs were assessed using 20 molar HL6/QD complexes incubated with A549 cells, followed by staining with Hoechst 33342. Fluorescent microscopy detected green fluorescence in the cells treated with HL6/QD complexes, but not in control cells or in cells treated with QDs alone (Fig. 4A). Quantitative fluorescent intensity was determined after protein transduction, using cells treated with 20 molar HL6/QD complexes and then analyzed using flow cytometry. Cellular uptake of QDs mediated by HL6 started within 5 min and continued to show high protein transduction over time (Fig. 4B). These results indicate that HL6 CPPs are capable of quickly delivering QDs into cells.
Molecular mechanism of cellular uptake of HL6/QD complexes. To reveal the mechanism of HL6-mediated cellular delivery of QDs, a series of inhibition studies were conducted. Flow cytometry was used www.nature.com/scientificreports www.nature.com/scientificreports/ to analyze cells treated by the HL6/QD complexes and pharmacological modulators, low temperature, or heparinases. Internalization of HL6/QD complexes was not affected by treatments with various endocytic inhibitors, including EIPA, CytD, FIL, NCO, and incubation at 4 °C for 5 min (Fig. 5A). In contrast, the transduction efficiency of HL6/QD complexes was dramatically decreased by treatment with heparinases. According to these findings, HL6/QD complexes are mainly delivered into cells using direct membrane translocation.
The transduction efficiency of HL6/QD complexes was decreased by treatment with pharmacological modulators EIPA, CytD, or NCO at 37 °C for 30 min, or incubated at 4 °C for 30 min (Fig. 5B). This suggests that macropinocytosis and clathrin-dependent endocytosis become involved in cellular uptake of HL6/QD complexes after extended periods of exposure. Together, these results indicate that direct membrane translocation is initially evident as the main uptake mechanism for HL6/QD complexes during cellular internalization. Subsequently, endocytosis contributes to the cellular uptake of HL6/QD complexes.
The HL6/QDs complex direct translocation was evaluated using several chemical enhancers of membrane transduction and analyzed using flow cytometry. Cells were treated with HL6/QDs complexes in PB, DMSO, EtOH, OA, or BIT. Only DMSO treatment among all tested chemical enhancers had a significant positive effect on cellular internalization of HL6/QD complexes (Fig. 5C). www.nature.com/scientificreports www.nature.com/scientificreports/ Cytotoxicity of HL6-mediated cellular uptake of QDs. Cell viability after HL6-directed QD delivery was analyzed using the MTT assay. Cells were treated with HL6 alone, QDs alone, or HL6/QD complexes for 30 min, washed with PBS, incubated in complete culture medium at 37 °C for 24 h, and then analyzed using the MTT assay. No cytotoxicity was detected in A549 (Fig. 6A) or HeLa cells (Fig. 6B) for any of the treatments. Additionally, HeLa cells were not damaged by HL6/QD complexes in the presence of a series of endocytic inhibitors, low temperature, or heparinases (Fig. 6C).  www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
This study demonstrates the novel HL6 peptides, which have flanking penta-histidine and mono-cysteine residues on the well-characterized cell-penetrating peptide L6, are able to deliver noncovalently complexed QDs into cells by direct membrane translocation. Cell viability assay with both A549 and HeLa cells indicates that none of the protein transduction components (HL6, QDs, and HL6/QD complexes) used in this study are cytotoxic. In our previous report, we demonstrated that L6/QD complexes enter cells by endocytosis 18  Flow cytometry was used to analyze the cells after treatment for 5 min. Significant differences were determined when the P-value was less than 0.01 (**). Data are shown as mean ± SD from twelve independent experiments for each treatment group. (B) Effect of endocytic modulators on cellular internalization of HL6/QD complexes for 30 min. HL6/QD complexes (prepared at a ratio of 20 at 37 °C for 1 h) were used to treat A549 cells in the presence or absence of pharmacological inhibitors (including EIPA, CytD, FIL, and NCO) and low temperature treatment (4 °C) for 30 min. Flow cytometry was used to analyze the cells after treatment for 30 min. Significant differences were determined when the P-value was less than 0.05 (*). Data are shown as mean ± SD from eight independent experiments for each treatment group. (C) Effect of chemical enhancers of membrane transduction on HL6-mediated cellular uptake of QDs. Flow cytometry was used to analyze the cells treated with HL6/QDs complexes in the presence or absence of chemical enhancers PB, DMSO, EtOH, OA, or BIT. Significant differences were determined when the P-value was less than 0.05 (*). Data are shown as mean ± SD from eight independent experiments for each treatment group.
www.nature.com/scientificreports www.nature.com/scientificreports/ studies demonstrate that the addition of polyhistidine peptides to an endocytic CPP allows internalization by direct membrane translocation. It is likely that HL6/QD complexes can enter cells by both direct membrane translocation and endocytosis simultaneously. , and heparinases (pretreated with heparinases I, II, and III for 6 h) at 37 °C for 30 min, or incubation at 4 °C (pretreated at 4 °C for 30 min) for 30 min. Cells were washed with PBS after treatment and then grown in complete culture medium at 37 °C for 24 h. The MTT assay was applied to assess cell viability. Significant differences compared with the negative control were determined when the P-value was less than 0.01 (**). Data are shown as mean ± SD from twelve independent experiments for each treatment group.
www.nature.com/scientificreports www.nature.com/scientificreports/ Endocytosis and direct membrane translocation contribute to cellular internalization of CPPs 2,6,15,29,30 . Recently, Futaki's team has demonstrated that CPP concentration can play an important role in determining the mode of cellular internalization 15 . In general, the majority of studies investigating cellular uptake mechanisms of CPPs have utilized CPP concentrations ranging from 0.5 to 10 μM 31 . Direct membrane translocation mediated by CPPs tends to occur at relatively high concentrations of CPPs, while endocytosis is a dominant pathway at lower CPP concentrations. For instance, endocytosis is the uptake mechanism of 2 μM of fluorescently labeled R8 in leukemia cells 31 , while direct membrane translocation occurs at concentrations above 5 μM 15,31 .
GAGs are biopolymers with greatly negative charges across membranes of all animal cells. GAGs contain heparan sulfate and chondroitin sulfate, and are long polysaccharides made of alternating disaccharide units 32 . Heparinase acts on the extracellular surface to degrade GAGs but keeps the integrity of the plasma membrane intact 33 . The detailed molecular actions of GAGs, especially heparan sulfate proteoglycans, in cellular internalization of CPPs are still a matter of controversy [33][34][35] . It was previously postulated that direct membrane translocation originates from spatially restricted NZ regions on plasma membranes, with heparan sulfates on cell surface being necessary for internalization at these NZ sites 16 . Recently, loosening of lipid packing in the membrane bilayer was proposed to be the key factor governing the membrane translocation of CPPs 36 . Arginine-rich CPPs are able to interact with negatively charged constituents of plasma membranes, such as heparan sulfate proteoglycans and glycosylated lipids 15,36 . This interaction may induce dynamic alternations in membrane structure, including particle formation, local membrane inversion and pore formation, resulting in direct membrane translocation of CPPs into cells 13,15,36 . As revealed by flow cytometry, the uptake of HL6/QD complexes was reduced by 63.2% following heparinase digestion (Fig. 5A). These results are consistent with previous reports that enzymatic digestion by heparinases led to great suppression of direct membrane translocation of arginine-rich CPPs 15,16 .
The positive charges of cationic CPPs, most rich in arginines or lysines, are an important functional characteristic required for cell entry 2,37 . Surprisingly, polyhistidine peptides, such as a hexadeca-histidine (H16) peptide, were recently reported as novel CPPs in human HepG2 and other cells 37 as well as in plant and yeast cells 38 . Cellular uptake of H16 peptides is mainly the result of macropinocytosis, and most H16 peptides localize in lysosomes and the Golgi apparatus 37 . Endosomolytic agents, such as hemagglutinin-2 (HA2) 39 , have been engaged to conquer endosomal entrapment and to enhance protein transduction efficiency. A modified PTD of transactivator of transcription (Tat) containing additional polyhistidine and cysteine residues 40 and LAH4 peptides containing four histidines 41 possessed endosomolytic properties. Similarly, the modified LKH-stEK CPP derived from the replacement of Lys residues of stEK CPP with His moieties was found to facilitate endosomal escape of short interfering RNAs (siRNAs) and promote >90% gene silencing 42 . Recently, polyhistidine-arginine (H6R6) peptide was demonstrated to show higher cell-penetrating efficiency and greater endosomal escape capacity compared to unmodified chitosan nanoparticles in vitro 43 . It has commonly been assumed that electrostatic interactions shared by both anionic cell membrane surfaces and cationic CPPs play a vital part in cellular internalization 15 . It should be noted that histidine possesses an imidazole ring with a pKa around 6.5, and thus is neutral without electric charge under physiological conditions. Nevertheless, the imidazole ring of histidine starts to be protonated in acidic tumor microenvironments 44,45 . Therefore, possible protonation of polyhistidine residues in acidic tumor microenvironments may facilitate cellular internalization of histidine-rich HL6. Additionally, we previously found that the nona-arginine CPP modified with additional polyhistidine and cysteine residues (becoming HR9) entered cells by direct membrane translocation, while the unmodified nona-arginine CPP entered cells by endocytosis 12 . Consistently with the previous study, HL6 with additional polyhistidine residues was shown to internalize into cells by direct membrane translocation, whereas the unmodified L6 CPP entered cells by endocytosis 18 . Collectively, our results indicate that polyhistidine peptides can play a determinant role in the mechanism of CPP-mediated protein transduction.

Conclusion
A novel HL6 (CHHHHHRRWQWRHHHHHC) CPP containing polyhistidine peptides flanking the L6 CPP can generate stable complexes with QDs at an optimal molar ratio of 20 and then deliver QDs into cells. Mechanistic studies revealed that direct membrane translocation is a principal route of internalization for HL6/QD complexes. Cytotoxicity studies proved HL6, QDs, and HL6/QD complexes were not harmful to A549 and HeLa cells. Overall, HL6 CPP may be an efficient and biocompatible carrier of nanoparticles or therapeutic cargos in biomedical applications and theranostics.