Translocation Mechanisms of Cell-Penetrating Polymers Identifiedby Induced Proton DynamicsTatsuro Goda,*,†,‡ Yuki Imaizumi,† Hiroaki Hatano,† Akira Matsumoto,†,§ Kazuhiko Ishihara,∥,⊥

and Yuji Miyahara†

†Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University (TMDU), 2-3-10 Kanda-Surugadai, Chiyoda,Tokyo 101-0062, Japan‡Nano Innovation Institute, Inner Mongolia University for Nationalities, No. 22 HuoLinHe Street, Tongliao, Inner Mongolia028000, P. R. China§Kanagawa Institute of Industrial Science and Technology (KISTEC), 705-1 Shimoimaizumi, Ebina, Kanagawa 243-0435, Japan∥Department of Materials Engineering, School of Engineering, and ⊥Department of Bioengineering, School of Engineering, TheUniversity of Tokyo, 7-3-1 Hongo, Bunkyo, Tokyo 113-8656, Japan

*S Supporting Information

ABSTRACT: Unlike the majority of nanomaterials designed for cellular uptake viaendocytic pathways, some of the functional nanoparticles and nanospheres directly enterthe cytoplasm without overt biomembrane injuries. Previously, we have shown that awater-soluble nanoaggregate composed of amphiphilic random copolymer of 2-methacryloyloxyethyl phosphorylcholine (MPC) and n-butyl methacrylate (BMA),poly(MPC-random-BMA) (PMB), passes live cell membranes in an endocytosis-freemanner. Yet, details in its translocation mechanism remain elusive due to the lack ofproper analytical methods. To understand this phenomenon experimentally, weelaborated the original pH perturbation assay that is extremely sensitive to the poreformation on cell membranes. The ultimate sensitivity originates from the detection of the smallest indicator H+ (H3O

+) passedthrough the molecularly sized transmembrane pores upon challenge by exogenous reagents. We revealed that water-solublePMB at the 30 mol % MPC unit (i.e., PMB30W) penetrated into the cytosol of model mammalian cells without any protonleaks, in contrast to conventional cell-penetrating peptides, TAT and R8 as well as the surfactant, Triton X-100. While exposureof PMB30W permeabilized cytoplasmic lactate dehydrogenase out of the cells, indicating the alteration of cell membranepolarity by partitioning of amphiphilic PMB30W into the lipid bilayers. Nevertheless, the biomembrane alterations by PMB30Wdid not exhibit cytotoxicity. In summary, elucidating translocation mechanisms by proton dynamics will guide the design ofnanomaterials with controlled permeabilization to cell membranes for bioengineering applications.

■ INTRODUCTION

Nanomaterials have acquired malleable features on cellularuptake, allowing them to be applied to the drug delivery system(DDS), subcellular imaging, cell theranostics, and cellreprogramming.1−3 It is appreciated that their element,composition, size, shape, charge, polarity, and surfacechemistry are the determinant factors for cell−nanomaterialinteractions.4,5 Nanomaterials manifest a bimodal pattern ofinternalization, that is, endocytic and nonendocytic fashions.6,7

Endocytosis or pinocytosis requires cells to engulf a nanoma-terial−protein (e.g., opsonin) complex in an energy-dependentmanner.8,9 The endocytic cellular uptakes restrict intracellulardistribution in the endosomal compartment. The encapsulatedmolecules cannot perform their expected biological activities.Therefore, escaping from the lysosomal vesicle is essential forefficient cytoplasmic delivery of a drug conjugate andsuccessful nanomedicine. Meanwhile, nonendocytic internal-ization frequently occurs by nanomaterials with cationic oramphiphilic features by evading barrier functions of lipidbilayer biomembranes.3,7 Cationic peptides such as trans-

a c t i v a t i n g t r a n s c r i p t i o n a l a c t i v a t o r ( T AT :GRKKRRQRRRPQ) and octa-arginine (R8: RRRRRRRR)are recognized as cell-penetrating peptides (CPPs) because oftheir permeability to the lipid bilayer cell membrane.10

Chemically derivatized metallic nanoparticles (nanoplexes)can directly enter or deliver their cargo molecules into thecytoplasm.11−13 Bioengineered proteins with abundant positiveor negative net charges (supercharged proteins) are capable ofpenetrating into mammalian cells.14 These nonendocyticinternalizations are liberated from endosomal entrapmentsthat may prevail in the efficiency of the current DDS. However,direct penetration has a concern about the cytotoxic or biocidaleffects by temporary breaking down of the biomembranebarriers.15−17

Synthetic macromolecules with nonendocytic cell penetra-tion have emerged as a new class of nanomaterials for cell

Received: March 22, 2019Revised: May 5, 2019Published: May 16, 2019

Article

pubs.acs.org/LangmuirCite This: Langmuir 2019, 35, 8167−8173

© 2019 American Chemical Society 8167 DOI: 10.1021/acs.langmuir.9b00856Langmuir 2019, 35, 8167−8173

Dow

nloa

ded

via

JIL

IN U

NIV

on

Apr

il 22

, 202

0 at

11:

26:4

1 (U

TC

).Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

manipulations (Figure 1a).16 Alkyl-modified poly(acrylicacid)s enhanced the permeability to liposomal membranes

by controlled disruptions of lipid bilayers.18 A polyethyleneglycol (PEG)-based polymeric surfactant passed through theintestinal cell membrane via nonendocytic internalizations.19

We have discovered direct penetration of the aggregate of awater-soluble and amphiphilic random copolymer composedof 2-methacryloyloxyethyl phosphorylcholine (MPC) and n-butyl methacrylate (BMA), poly(MPC-random-BMA) (PMB),

into live cells without cytotoxicity.20,21 The polymericderivative of phospholipids originally developed for nonfoulingbiomaterials can deliver anticancer drugs or nucleic acids intothe cytosol efficiently, both in vitro and in vivo.22−27 Similar toPMB, a methacrylate copolymer bearing zwitterionic sulfobe-taine and methoxy PEGs in the side chains translocated intocells via nonendocytic pathways.28 It is said that directpenetration of a macromolecule is mediated either by makingtransient holes or by fusing with cell membranes.16 Passivediffusion of net-neutral PMB depended on its amphiphilicity sothat the absence of the BMA unit abolished nonendocyticinternalization.20

Acquisition of experimental evidence on the internalizationmechanism for nanomaterials is crucial toward efficient andsafe therapeutics. However, it is challenging to find a suitablemethod for identifying the types of injuries to thebiomembranes. Traditionally, patch-clamp assays are used tomeasure transmembrane electrical resistance as a signature ofbiomembrane integrity, but the distinction of pore formationfrom membrane disorder is not possible merely from theresistive index.17 In addition, it requires skilled techniques tomanipulate a microelectrode without disrupting the biomem-brane. The inherent low-throughput also limits the applicationsfor drug screening. Meanwhile, an optical assay that measuresendogenous or preloaded probes leaked out of the cells is astandard technique for determining biomembrane injuries.3

Their sensitivity to minute pores depends on the dimension ofa reporter molecule. In addition, most of the indicators shouldbe preloaded intracellularly before the assays. Furthermore,some indicators permeate through altered biomembranes bypartitioning or fusion mechanisms, thereby obscuring thetranslocation mechanism by pore formation.We have recently established a new technique for specifically

detecting biomembrane pores (Figure 1b).29,30 Prompted bythe nature that H+ (H3O

+) cannot permeate through the lipidbilayers but has access to molecularly sized pores due to thesmallest size ever (the hydrodynamics radius: RH ≈ 0.33 nm),3

the assay measures changes in the H+ permeability through themembranes of live cells cultured on a pH-sensing ion-sensitivefield-effect transistor (ISFET). To focus on the H+

permeabilization by excluding gradual pH changes caused bycell metabolism and transporter activities, pH was temporallyperturbed by 1.4−1.7 unit by simply exposing to a weak acidNH4

+ in a buffered solution using a superfused fluidicsystem.31,32 The transient pH changes at the point of NH4

+

Figure 1. Understanding translocation mechanism of amphiphilicpolymers by the weak acid equilibrium-induced proton dynamics. (a)Aggregate of amphiphilic copolymer PMB30W directly translocatesinto the cytosol in live mammalian cells either by forming small poresor fusing with the biomembranes. (b) Schematic illustration of anNH4Cl-induced pH perturbation assay for model HepG2 cellscultured on a gate dielectric of the Nernstian pH-sensing ion-sensitivefield-effect transistor (ISFET, −55.4 ± 1.8 mV/pH). Two isotonicbuffers with/without NH4Cl were alternately subjected to the cellsusing a superfused fluidic system that caused the pH perturbation(ΔV) as a result of the temporary imbalanced NH4

+/NH3 equilibriumin the cell microenvironment. A decrease in ΔV is specifically causedby the permeabilization of H+ and NH4

+ as the smallest indicatorsever through the pores on damaged biomembranes. The ultimatesensitivity and selectivity to biomembrane pores highlight trans-location mechanisms of PMPB30W and other nanomaterials.

Figure 2. Demonstration of direct penetration of rhoPMB30W by time-lapse confocal laser scanning fluorescence imaging of HepG2 cell spheroidsin a serum-free medium. Upper panels: 1 mg/mL rhoPMB30W, lower panels: 1 mg/mL rhoPMPC (negative control). Scale bar = 50 μm.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b00856Langmuir 2019, 35, 8167−8173

8168

addition/removal originated from the inherent ion imperme-ability of the lipid bilayers. Namely, the degree of pHperturbation attenuated specifically by pore-forming activitiesof external reagents without interference by cell metabolismand reagent. In fact, we have succeeded in identifying smallpores that were otherwise invisible using a standard lactatedehydrogenase (LDH, RH ≈ 4.2 nm) leakage assay.29 TheISFET-based active pH sensing has features in sensitivity,specificity, simplicity, stability, reproducibility, wide applic-ability to cell types, and amenability to high-throughputanalysis.30,32 Moreover, the comparison of the ISFET assaywith existing cytotoxicity assays enabled to classify the type ofbiomembrane injuries and the cause of cell deaths.33

Therefore, the aim of the study was to inspect how PMBand CPPs translocate into the cytoplasm by employing theISFET assay in combination with the LDH release andproliferation assays.

■ RESULTS AND DISCUSSION

Characterization of PMB30W. The molar ratio of MPC/BMA in the PMB copolymer with 30 mol % MPC unit in thefeed (PMB30W) was 30.3/69.7 as determined by 1H NMR(Figure S1). The molecular weights of PMB30W and thehomopolymer of MPC lacking the BMA unit (PMPC) wereMn = 4.2 × 103 with Mw = 9.0 × 103 and Mn = 8.3 × 103 withMw = 1.3 × 104, respectively, by gel permeation chromatog-raphy. PMB30W formed a net-neutral aggregate in aqueousmedia with the mean diameter of 11.2 ± 0.2 nm and the meanzeta potential of −4.1 ± 0.9 mV (Figure S2). The meanhydrodynamic size and zeta potential of PMPC were 23.5 ±0.6 nm and 0.8 ± 1.0 mV, respectively.Cytoplasmic Entry of PMB30W. First, we confirmed the

rapid internalization of PMB30W by time-lapse confocalmicroscopy (Figure 2). Live human Caucasian hepatocytecarcinoma (HepG2) cells in spheroids were stained by

rhodamine B-tagged PMB30W (rhoPMB30W) within 5 minfollowing the exposure. Our previous study confirmed thatrhoPMB30W translocated under the conditions of cell energydepletion or at 4 °C.20 The nonendocytic internalizationoccurred at the molar ratio of MPC/BMA ranged from 30/70to 80/20 without limitations by the molecular weight oraggregate size. Moreover, rhoPMB30W was enriched inmitochondria due to the affinity between rhodamine B andmitochondria by bypassing endosomal entrapment. The rapidentry indicates the translocation via passive transport acrossthe biomembranes rather than by active cellular uptakes. Bycontrast, the cytosol was not stained by rhodamine B-taggedPMPC (rhoPMPC), which is the homopolymer of MPClacking BMA. The results indicate the necessity ofamphiphilicity for the nonendocytic internalization.

ISFET and LDH Assays Probing the InternalizationMechanism of PMB30W. Nonendocytic internalization ofPMB30W suggests that the mechanism is either by poreformation or membrane fusion. To gain direct evidence forpore formations, we performed the ISFET-based pHperturbation assay (Figures 1b and 3). The degree of pHperturbation was determined based on the signal before (ΔV0)and after one, two, and three 1 min exposures (ΔVi) of areagent to give the normalized ISFET signal as [(ΔV0 − ΔVi)/ΔV0]. The ISFET signal remained unchanged for membrane-permeable PMB30W and a transfection reagent Lipofectamineas well as membrane-impermeable PMPC and PEG. On theother hand, the degree of pH perturbation increased by theconcentration and the number of exposures for TAT and R8.The pH perturbation disappeared (ΔVi ≈ 0 mV) when cellstotally lysed by treating 1 mg/mL Triton X-100 (TX-100) as aPEG-based nonionic surfactant. Because the degree of pHperturbation is ultimately sensitive to small pores due to thesmallest H+ and NH4

+ indicators with RH of ∼0.33 nm (Figure1), the results show no such pore formation by PMB30W

Figure 3. Left: Time course of the gate potential that indicates local pH at the cells/ISFET interface during the NH4Cl-induced pH perturbationassay with repeated exposures of a panel of reagents to cells by superfusion. Right: Changes in the degree of pH perturbation [(ΔV0 − ΔVi)/ΔV0]following one, two, and three 1 min exposures of each reagent at varying concentrations to cells. Mean ± standard deviation (SD, n = 3). Thearrows indicate the reagent and its concentration for the corresponding time-series data.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b00856Langmuir 2019, 35, 8167−8173

8169

during internalization. Together, we confirmed that exposuresof cationic TAT and R8 (0.1 mg/mL) formed small poresaccessible for H+ and NH4

+. These minute pores were notdetectable using a conventional electrochemical impedancespectroscopy (Figure S3).To further examine any alterations of the biomembrane by

PMB30W, we performed the LDH release assay for HepG2cells (Figure 4a). Membrane-permeable PMB30W causedLDH leakage from the cytosol in a concentration-dependentmanner. This is in contrast to membrane-impermeable PMPC.Lipofectamine at 1/10 dilution, 0.1−1 mg/mL TX-100, and0.01−1 mg/mL PEG (Mw of 15−25k) also exhibited LDHsignals, whereas cultures with TAT and R8 induced no LDHleakages.To classify the type of biomembrane injuries, the ISFET

signals were compared with the LDH results in a scatteredplot.29 By setting thresholds, the plots were categorized intofour regimes: ISFET−/LDH− (1), ISFET+/LDH− (2),ISFET−/LDH+ (3), and ISFET+/LDH+ (4) (Figure 4b,c).Among them, the ISFET+/LDH− (2) regime for 0.1 mg/mLTAT and R8 indicates the formation of molecularly sized poresand which is permeable for H+ and NH4

+ (RH ≈ 0.33 nm) andnot for LDH (RH ≈ 4.2 nm). In contrast, ISFET−/LDH+ (3)for 0.01−10 mg/mL PMB30W, Lipofectamine at 1/18 and 1/10 dilutions, 0.1 mg/mL TX-100, and 0.01−1 mg/mL PEG isinterpreted as chemically induced structural disorders orpolarity alterations of biomembranes, thereby permeatingLDH enzyme as an amphiphilic protein while maintainingbarrier functions of cell membranes against small ions.ISFET−/LDH− (1) for PMPC represents the intact plasmamembrane by blocking the leaks of both ion and LDH.ISFET+/LDH+ (4) for 1 mg/mL TX-100 stands for large-scalepore formation or membrane lysis. Identification of the type ofmembrane injuries in the regimes (2) and (3) was exclusive ofthe ISFET assay. In our previous study, ISFET+/LDH− (2)

was assigned to cationic reagents poly(ethylene imine) (PEI,Mw of 25k and 750k) and stearyl trimethyl ammonium chloride(STAC) for a short period of incubation (<15 min). Thisaccords with the present results on TAT and R8 with richpositive charges (Figure 4b). For PEI and STAC, the plotsshifted to ISFET+/LDH+ (4) when the incubation extended to>30 min, indicating that the membrane damage expanded overtime that permeated LDH. Cationic molecules are known toform transient pores by pulling the polar headgroup toward thehydrophobic core in the bilayer.34−37

PMB30W has an n-butyl group and is less hydrophobic thanthe fatty acid groups in phospholipids. Therefore, the polarityof the cell membrane is altered when PMB30W hydrophobi-cally interacts with the core of the lipid bilayer plasmamembranes. As with PMB30W, we have previously observedthe ISFET−/LDH+ (3) signals for nonionic surfactants Tween20, Tween 80, TX-100, and an anionic surfactant, sodiumdodecyl sulfate. These detergents are known to alter the cellmembrane polarity and permeabilize drugs and LDH throughthe biomembrane with a minimum cytotoxic effect.38−44 Weenvisage that the LDH leakage is induced by the association ofthese amphiphilic reagents on the bilayer structure. Thepartitioning of amphiphilic molecules into biomembranes has athree-stage equilibrium.39,41,43 (A) At the low detergent/lipidratio, a nonmicellar detergent is inserted into biomembranes;(B) at the middle ratio, lipid-saturated micelles coexist withdetergent-saturated biomembranes; and (C) at the high ratio,detergent-saturated biomembranes lyse. The ISFET−/LDH+

condition falls into the stage (A) or (B). Permeability isexpressed as the product of the diffusion coefficient andsolubility coefficient. The solubility coefficient of the alteredcell membranes differs depending on the solute. Hydrated ionswere more hydrophilic than proteins, so they did not dissolvein cell membranes. Therefore, it can be explained that protons

Figure 4. A comparison between the ISFET and LDH assays to identify the type of biomembrane injuries. (a) Amount of LDH released duringincubation with a panel of reagents for 15 min. (b) Scatter plot between the ISFET and LDH assays. Data points identify the two signals for thesamples at set concentrations with the mean ± SD (n = 5). Dashed lines represent the thresholds defined at 0.1. Correlation coefficient: r = 0.77.(c) Schematic illustrations showing the types of biomembrane injuries based on the assignment of the four regimes in the correlation diagram.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b00856Langmuir 2019, 35, 8167−8173

8170

and ammonium ions could not permeate the alteredmembranes.Translocation by amphiphilic (macro)molecules has been

studied by simulation in the literature.45−48 According to them,passive translocation occurs at a point of balanced hydro-philicity/hydrophobicity, where hydrophobic units are insertedinto the core of the phospholipid bilayer by the Brownianmotion while compensating the energy required for disorder-ing the bilayer structure. The description explains our resultson PMB30W. The in silico studies also predict biomembranepermeabilization to solvents and small solutes following theinvasion of amphiphilic polymers.47 This accords with theenhanced LDH leakage for PMB30W (Figure 4), but not withthe maintained barrier properties for H+ (H3O

+) and NH4+

(Figure 3). Namely, the PMB30W-induced permeabilizationdepends on the solute types and the bilayer composition. Infact, the hemolysis assay indicates no leakage of hemoglobinfrom erythrocytes by PMB30W (Figure S4).Cytotoxicity Assay Analyzing the Consequence of

Biomembrane/PMB30W Interplay. Finally, the cell pro-liferation (WST-8) assay was performed to examine the causalrelationship between the biomembrane alterations byPMB30W and acute cytotoxicity. In agreement with theliterature,20,25 PMB30W, TAT, R8, and PEG did not causecytotoxicity (Figure 5a). Lipofectamine and TX-100 werecytotoxic in a dose-dependent manner.The ISFET and WST-8 results were again analyzed in the

bivariate chart with threshold lines (Figure 5b,c).29,33 Thescattered plots for TAT and R8 at 0.1 mg/mL were categorizedinto ISFET+/WST-8−. This indicates the maintained cellularviability due to homeostatic recovery from pore formations bythe cationic CPPs. Lipofectamine at 1/18 and 1/10 dilutionsand TX-100 at 0.01−0.1 mg/mL fell into ISFET−/WST-8+ byaltering biomembranes or damaging organelles, not by formingpores (Figure 4b). Cytotoxicity of Lipofectamine is associatedwith a polycationic component as well as the structure of thehydrophobic domain in the lipids.49 Yet, a Lipofectamine/polynucleotide polyplex for transfection is less cytotoxic due tothe cancellation of net-positive charges by anionic nucleicacids.49 Amphiphilic TX-100 at 1 mg/mL irreversibly lysedbiomembranes followed by cell death (ISFET+/WST-8+).

Trends were similar between the ISFET/WST-8 and LDH/WST-8 comparisons (Figure S5).It is intriguing how PMB30W invasion to biomembranes

prevents the permeation of small ions during the LDH release,thereby supporting the homeostatic maintenance. Our interestlies in the effect of the chemical structure of phospholipid-mimetic PMB30W on the support to ionic impermeability inaltered biomembranes.21,25 A further investigation is needed toelucidate the molecular design of amphiphilic polymers for safeand efficient translocation.

■ CONCLUSIONSTranslocation mechanisms of PMB30W were examined invitro using our original ISFET-based pH perturbation assaythat can specifically detect molecularly sized transmembranepores by the leakage of H+ and NH4

+ as the smallest indicators.The ISFET assay confirmed no pore-forming activity ofPMB30W during passive translocation. Meanwhile, PMB30Wpermeabilized cytoplasmic LDH, indicating alterations inbiomembrane integrity by the association of hydrophobicBMA on the bilayer structure. Therefore, we concluded thatPMB30W entered cells by the amphiphilicity-inducedmembrane fusion, not by pore formation. The fusion alsooccurred by amphiphilic Lipofectamine and TX-100, butPMB30W only preserved the cell proliferation. The com-parative study between the ISFET and LDH assays identifiedthe pore formation by cationic TAT and R8. Scanning minutepores by proton dynamics will help develop safe and efficientpolymeric nanocarriers for gene transfection and cell therapy.

■ METHODSMaterials. PMB30W, PMPC, rhoPMB30W, and rhoPMPC were

synthesized by free radical polymerization as previously reported.20 Inbrief, MPC (2.66 g, 9.00 mmol) and BMA (2.99 g, 21.0 mmol) werereacted in degassed ethanol (30 mL) in the presence of an initiatorperbutyl-ND (70 wt % in hydrocarbons, 0.367 g, 1.05 mmol) for 12 hat 60 °C. For the synthesis of rhoPMB30W and rhoPMPC,methacryloxyethyl thiocarbonyl rhodamine B (20.5 mg, 30.0 μmol,Polysciences, PA) was added to the monomer solution. Thecopolymer was purified by reprecipitation in diethylether/chloroform(7/3 v/v) and by dialysis against water through regenerative cellulosemembrane (MWCO: 3.5k). TAT and R8 peptides were ordered

Figure 5. Comparison between the ISFET and WST-8 assays to identify the cause of cell death. (a) Normalized WST-8 signals after incubatingwith a panel of reagents for 6 h. (b) Scatter plot between the ISFET and WST-8 assays. Data points identify the two signals of each reagent at setconcentrations with the mean ± SD (n = 5). Dashed lines represent the threshold set at 0.1. Correlation coefficient: r = 0.49. (c) Flow chartshowing the different cytotoxic activities identified by the bivariate analysis.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b00856Langmuir 2019, 35, 8167−8173

8171

through Medical & Biological Laboratories Co., Ltd. (Nagoya, Japan).Lipofectamine, which is a mixture of 3-dioleoyloxy-2-(6-carbox-yspermyl)-propylamid and dioleoyl phosphatidylethanolamine (3/1w/w), was obtained from ThermoFisher Scientific Inc. (Waltham,MA). The HepG2 cell line was obtained from DS Pharma Biomedical(Osaka, Japan). All other reagents were purchased from thecommercial sources and used as received.Cell Culture. HepG2 cells were seeded at 1.3 × 104 cells/cm2 on a

tissue culture polystyrene dish in Dulbecco’s modified Eagle medium(DMEM, high glucose) with 10% fetal bovine serum (FBS) and 0.1mg/mL penicillin/streptomycin and were incubated in 5% CO2 at 37°C. The cells were passaged at subconfluent conditions by treatingwith 0.25% trypsin/EDTA.Confocal Microscopy. Time-lapse confocal images were taken for

HepG2 cells cultured on a poly-L-lysine-treated glass bottom dish(Matsunami Glass Ind. Ltd., Osaka, Japan) in DMEM containing 1mg/mL rhoPMB30W or rhoPMPC by using a Nikon Eclipse Ti-Einverted microscope with a confocal laser scanning system (Tokyo,Japan). A 60× water immersion objective lens (1.2 NA, Nikon) wasused. The rhodamine B-tagged polymers were excited with the 561nm line of a diode laser (Nikon Lu-N4), and emission was detectedusing a standard PMT detector (Nikon C2-DU3) with a 570−613 nmband-pass filter.ISFET-Based pH Perturbation Assay. HepG2 cells were

cultured in the medium on a poly(L-lysine)-coated gate dielectric (a40 nm-thick Ta2O5/140 nm-thick Si3N4/125 nm-thick SiO2 layer) ofan open gate n-channel depletion type ISFET (width/length = 340μm/10 μm, ISFETCOM, Saitama, Japan) overnight before the assay.For inducing pH perturbations, the isotonic buffers (1 mmol/L [bis-tris-propane], 140 mmol/L [NaCl], 4 mmol/L [KCl], and 1 mmol/L[MgCl2]; pH 7.4) containing 10 mmol/L [NH4Cl] or 20 mmol/L[sucrose] were alternately exposed to the cells for 1 min each, using afluidic system at 37 °C. The inlet tube was placed ∼80 μm away fromthe pH-sensing area of ISFET for instant exchanges of the solution(<1 s). ISFET was operated as a source−drain follower forpotentiometric readout.29,50,51 The drain−source voltage, drain−source current, and direct current bias potential against the Ag/AgClpellet (Warner Instruments, Hamden, CT) reference electrode were0.5 V, 0.5 mA, and 0 V, respectively. We recorded the time course ofgate potential and converted it into pH changes by the Nernstianresponse. The mean ± SD of our ISFET response was −55.4 ± 1.8mV/pH (n = 3).LDH and WST-8 Assays. HepG2 cells (5.0 × 103 cells) were

incubated in DMEM with 10% FBS in a 96-well microtiter plate for24 h before the assays. Thereafter, the cells were subjected to areagent in DMEM for 15 min (LDH) or 6 h (WST-8) in 5% CO2 at37 °C. Then, the cell-free supernatant was analyzed using commercialLDH-Cytotoxic Test or Cell Counting Kit-8 reagents (Fujifilm WakoPure Chemical Corp., Osaka, Japan) following the manufacturers’instructions. We confirmed that PMB30W and PMPC did notinterfere with the reactions in the LDH and WST-8 assays. UVabsorbance (A) was measured using the Infinite M200 microplatereader (Tecan Corp., Mannedorf, Switzerland). Cells treated with 1.0mg/mL TX-100 or DMEM were used as positive and negativecontrols. The degree of LDH leaked was determined as [LDH] =(A560 − A560,n.c.)/(A560,p.c. − A560,n.c.). The degree of cell death wasdetermined at the set endpoint as [WST-8] = (A450 − A450,n.c.)/(A450,p.c. − A450,n.c.).

■ ASSOCIATED CONTENT

*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.lang-muir.9b00856.

1H NMR charts for PMB30W and PMPC; dynamic lightscattering results; electrochemical impedance spectros-copy for reagents-exposed cells; hemolysis assay; and

bivariate analysis between the LDH and WST-8 assays(PDF)

■ AUTHOR INFORMATIONCorresponding Author*E-mail: goda.bsr@tmd.ac.jp.ORCIDTatsuro Goda: 0000-0003-2688-8186Kazuhiko Ishihara: 0000-0002-4196-5488Yuji Miyahara: 0000-0003-2703-0958Author ContributionsT.G. conceived the research and wrote the paper. T.G. and K.I.synthesized the polymers. Y.I. and H.H. conducted theexperiments. A.M. and Y.M. supervised the project.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe appreciate financial support in part from a Grant-in-Aid forScientific Research on Innovative Areas “NanomedicineMolecular Science” (#26107705) from MEXT of Japan, theNakatani Foundation of Electronic Measuring TechnologyAdvancement, and the Tateisi Science and TechnologyFoundation.

■ REFERENCES(1) Fu, A.; Tang, R.; Hardie, J.; Farkas, M. E.; Rotello, V. M.Promises and pitfalls of intracellular delivery of proteins. BioconjugateChem. 2014, 25, 1602−1608.(2) Sun, T.; Zhang, Y. S.; Pang, B.; Hyun, D. C.; Yang, M. X.; Xia, Y.N. Engineered nanoparticles for drug delivery in cancer therapy.Angew. Chem., Int. Ed. 2014, 53, 12320−12364.(3) Stewart, M. P.; Langer, R.; Jensen, K. F. Intracellular delivery bymembrane disruption: mechanisms, strategies, and concepts. Chem.Rev. 2018, 118, 7409−7531.(4) Nel, A. E.; Madler, L.; Velegol, D.; Xia, T.; Hoek, E. M. V.;Somasundaran, P.; Klaessig, F.; Castranova, V.; Thompson, M.Understanding biophysicochemical interactions at the nano-biointerface. Nat. Mater. 2009, 8, 543−557.(5) Elsabahy, M.; Wooley, K. L. Design of polymeric nanoparticlesfor biomedical delivery applications. Chem. Soc. Rev. 2012, 41, 2545−2561.(6) Iversen, T.-G.; Skotland, T.; Sandvig, K. Endocytosis andintracellular transport of nanoparticles: Present knowledge and needfor future studies. Nano Today 2011, 6, 176−185.(7) Du, S.; Liew, S. S.; Li, L.; Yao, S. Q. Bypassing endocytosis:direct cytosolic delivery of proteins. J. Am. Chem. Soc. 2018, 140,15986−15996.(8) Hillaireau, H.; Couvreur, P. Nanocarriers’ entry into the cell:relevance to drug delivery. Cell. Mol. Life Sci. 2009, 66, 2873−2896.(9) Sahay, G.; Alakhova, D. Y.; Kabanov, A. V. Endocytosis ofnanomedicines. J. Controlled Release 2010, 145, 182−195.(10) Heitz, F.; Morris, M. C.; Divita, G. Twenty years of cell-penetrating peptides: from molecular mechanisms to therapeutics. Br.J. Pharmacol. 2009, 157, 195−206.(11) Rosi, N. L.; Giljohann, D. A.; Thaxton, C. S.; Lytton-Jean, A. K.R.; Han, M. S.; Mirkin, C. A. Oligonucleotide-modified goldnanoparticles for intracellular gene regulation. Science 2006, 312,1027−1030.(12) Verma, A.; Uzun, O.; Hu, Y.; Hu, Y.; Han, H.-S.; Watson, N.;Chen, S.; Irvine, D. J.; Stellacci, F. Surface-structure-regulated cell-membrane penetration by monolayer-protected nanoparticles. Nat.Mater. 2008, 7, 588−595.(13) Ghosh, P.; Yang, X.; Arvizo, R.; Zhu, Z.-J.; Agasti, S. S.; Mo, Z.;Rotello, V. M. Intracellular delivery of a membrane-impermeable

Langmuir Article

DOI: 10.1021/acs.langmuir.9b00856Langmuir 2019, 35, 8167−8173

8172

enzyme in active form using functionalized gold nanoparticles. J. Am.Chem. Soc. 2010, 132, 2642−2645.(14) Cronican, J. J.; Beier, K. T.; Davis, T. N.; Tseng, J.-C.; Li, W.;Thompson, D. B.; Shih, A. F.; May, E. M.; Cepko, C. L.; Kung, A. L.;Zhou, Q.; Liu, D. R. A class of human proteins that deliver functionalproteins into mammalian cells in vitro and in vivo. Chem. Biol. 2011,18, 833−838.(15) Vega-Villa, K. R.; Takemoto, J. K.; Yanez, J. A.; Remsberg, C.M.; Forrest, M. L.; Davies, N. M. Clinical toxicities of nanocarriersystems. Adv. Drug Delivery Rev. 2008, 60, 929−938.(16) Marie, E.; Sagan, S.; Cribier, S.; Tribet, C. Amphiphilicmacromolecules on cell membranes: from protective layers tocontrolled permeabilization. J. Membr. Biol. 2014, 247, 861−881.(17) de Planque, M. R. R.; Aghdaei, S.; Roose, T.; Morgan, H.Electrophysiological characterization of membrane disruption bynanoparticles. ACS Nano 2011, 5, 3599−3606.(18) Vial, F.; Rabhi, S.; Tribet, C. Association of octyl-modifiedpoly(acrylic acid) onto unilamellar vesicles of lipids and kinetics ofvesicle disruption. Langmuir 2005, 21, 853−862.(19) Mathot, F.; Schanck, A.; Van Bambeke, F.; Arien, A.; Noppe,M.; Brewster, M.; Preat, V. Passive diffusion of polymeric surfactantsacross lipid bilayers. J. Controlled Release 2007, 120, 79−87.(20) Goda, T.; Goto, Y.; Ishihara, K. Cell-penetrating macro-molecules: Direct penetration of amphipathic phospholipid polymersacross plasma membrane of living cells. Biomaterials 2010, 31, 2380−2387.(21) Goda, T.; Ishihara, K.; Miyahara, Y. Critical update on 2-methacryloyloxyethyl phosphorylcholine (MPC) polymer science. J.Appl. Polym. Sci. 2015, 132, 41766.(22) Ukawa, M.; Akita, H.; Masuda, T.; Hayashi, Y.; Konno, T.;Ishihara, K.; Harashima, H. 2-Methacryloyloxyethyl phosphorylcho-line polymer (MPC)-coating improves the transfection activity ofGALA-modified lipid nanoparticles by assisting the cellular uptakeand intracellular dissociation of plasmid DNA in primary hepatocytes.Biomaterials 2010, 31, 6355−6362.(23) Lin, X.; Konno, T.; Ishihara, K. Cell-membrane-permeable andcytocompatible phospholipid polymer nanoprobes conjugated withmolecular beacons. Biomacromolecules 2014, 15, 150−157.(24) Tamura, K.; Kikuchi, E.; Konno, T.; Ishihara, K.; Matsumoto,K.; Miyajima, A.; Oya, M. Therapeutic effect of intravesicaladministration of paclitaxel solubilized with poly(2-methacryloylox-yethyl phosphorylcholine-co-n-butyl methacrylate) in an orthotopicbladder cancer model. BMC Cancer 2015, 15, 317.(25) Ishihara, K.; Mu, M.; Konno, T. Water-soluble and amphiphilicphospholipid copolymers having 2-methacryloyloxyethyl phosphor-ylcholine units for the solubilization of bioactive compounds. J.Biomater. Sci., Polym. Ed. 2018, 29, 844−862.(26) Ishihara, K. Blood-compatible surfaces with phosphorylcholine-based polymers for cardiovascular medical devices. Langmuir 2019,35, 1778−1787.(27) Ishihara, K. Revolutionary advances in 2-methacryloyloxyethylphosphorylcholine polymers as biomaterials. J. Biomed. Mater. Res.,Part A 2019, 107A, 933−943.(28) Morimoto, N.; Wakamura, M.; Muramatsu, K.; Toita, S.;Nakayama, M.; Shoji, W.; Suzuki, M.; Winnik, F. M. Membranetranslocation and organelle-selective delivery steered by polymericzwitterionic nanospheres. Biomacromolecules 2016, 17, 1523−1535.(29) Imaizumi, Y.; Goda, T.; Schaffhauser, D. F.; Okada, J.-i.;Matsumoto, A.; Miyahara, Y. Proton-sensing transistor systems fordetecting ion leakage from plasma membranes under chemical stimuli.Acta Biomater. 2017, 50, 502−509.(30) Imaizumi, Y.; Goda, T.; Toya, Y.; Matsumoto, A.; Miyahara, Y.Oleyl group-functionalized insulating gate transistors for measuringextracellular pH of floating cells. Sci. Technol. Adv. Mater. 2016, 17,337−345.(31) Schaffhauser, D.; Fine, M.; Tabata, M.; Goda, T.; Miyahara, Y.Measurement of rapid amiloride-dependent pH changes at the cellsurface using a proton-sensitive field-effect transistor. Biosensors 2016,6, 11.

(32) Hatano, H.; Goda, T.; Matsumoto, A.; Miyahara, Y. Inducedproton perturbation for sensitive and selective detection of tightjunction breakdown. Anal. Chem. 2019, 91, 3525−3532.(33) Imaizumi, Y.; Goda, T.; Matsumoto, A.; Miyahara, Y.Identification of types of membrane injuries and cell death usingwhole cell-based proton-sensitive field-effect transistor systems.Analyst 2017, 142, 3451−3458.(34) Hong, S.; Leroueil, P. R.; Janus, E. K.; Peters, J. L.; Kober, M.-M.; Islam, M. T.; Orr, B. G.; Baker, J. R.; Banaszak Holl, M. M.Interaction of polycationic polymers with supported lipid bilayers andcells: Nanoscale hole formation and enhanced membrane perme-ability. Bioconjugate Chem. 2006, 17, 728−734.(35) Sikor, M.; Sabin, J.; Keyvanloo, A.; Schneider, M. F.; Thewalt, J.L.; Bailey, A. E.; Frisken, B. J. Interaction of a charged polymer withzwitterionic lipid vesicles. Langmuir 2010, 26, 4095−4102.(36) Choudhury, C. K.; Kumar, A.; Roy, S. Characterization ofConformation and Interaction of Gene Delivery Vector Polyethyle-nimine with Phospholipid Bilayer at Different Protonation State.Biomacromolecules 2013, 14, 3759−3768.(37) Kwolek, U.; Jamroz, D.; Janiczek, M.; Nowakowska, M.;Wydro, P.; Kepczynski, M. Interactions of polyethylenimines withzwitterionic and anionic lipid membranes. Langmuir 2016, 32, 5004−5018.(38) Stavrovskaya, A. A.; Potapova, T. V.; Rosenblat, V. A.;Serpinskaya, A. S. The effect of non-ionic detergent tween 80 oncolcemid-resistant transformed mouse cells in vitro. Int. J. Cancer1975, 15, 665−672.(39) Lasch, J. Interaction of detergents with lipid vesicles. Biochim.Biophys. Acta, Biomembr. 1995, 1241, 269−292.(40) Kragh-Hansen, U.; le Maire, M.; Møller, J. V. The mechanismof detergent solubilization of liposomes and protein-containingmembranes. Biophys. J. 1998, 75, 2932−2946.(41) Heerklotz, H.; Seelig, J. Correlation of membrane/waterpartition coefficients of detergents with the critical micelleconcentration. Biophys. J. 2000, 78, 2435−2440.(42) Heerklotz, H. Interactions of surfactants with lipid membranes.Q. Rev. Biophys. 2008, 41, 205−264.(43) Lichtenberg, D.; Ahyayauch, H.; Alonso, A.; Goni, F. M.Detergent solubilization of lipid bilayers: a balance of driving forces.Trends Biochem. Sci. 2013, 38, 85−93.(44) Vaidyanathan, S.; Orr, B. G.; Banaszak Holl, M. M. Detergentinduction of HEK 293A cell membrane permeability measured underquiescent and superfusion conditions using whole cell patch clamp. J.Phys. Chem. B 2014, 118, 2112−2123.(45) Werner, M.; Sommer, J.-U.; Baulin, V. A. Homo-polymers withbalanced hydrophobicity translocate through lipid bilayers andenhance local solvent permeability. Soft Matter 2012, 8, 11714−11722.(46) Sommer, J.-U.; Werner, M.; Baulin, V. A. Critical adsorptioncontrols translocation of polymer chains through lipid bilayers andpermeation of solvent. Europhys. Lett. 2012, 98, 18003.(47) Werner, M.; Sommer, J.-U. Translocation and inducedpermeability of random amphiphilic copolymers interacting withlipid bilayer membranes. Biomacromolecules 2015, 16, 125−135.(48) Werner, M.; Bathmann, J.; Baulin, V. A.; Sommer, J.-U.Thermal tunneling of homopolymers through amphiphilic mem-branes. ACS Macro Lett. 2017, 6, 247−251.(49) Zhi, D.; Zhang, S.; Wang, B.; Zhao, Y.; Yang, B.; Yu, S.Transfection efficiency of cationic lipids with different hydrophobicdomains in gene delivery. Bioconjugate Chem. 2010, 21, 563−577.(50) Casans, S.; Ramirez, D.; Navarro, A. E. Circuit ProvidesConstant Current for ISFETs/MEMFETs. EDN Network, 2000, Vol.45 (22), pp 164−176.(51) Nakazato, K. An integrated ISFET sensor array. Sensors 2009,9, 8831−8851.

Langmuir Article

DOI: 10.1021/acs.langmuir.9b00856Langmuir 2019, 35, 8167−8173

8173