visual electrofluorochromic detection of cancer cell

9
Visual Electrouorochromic Detection of Cancer Cell Surface Glycoprotein on a Closed Bipolar Electrode Chip Zhaoyan Tian, Li Mi, Yafeng Wu,* ,Fengying Shao, Mingqiang Zou, δ Zhenxian Zhou, and Songqin Liu* ,Key Laboratory of Environmental Medicine Engineering, Ministry of Education, Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, School of Chemistry and Chemical Engineering, Southeast University, Nanjing, 211189, China δ Chinese Academy of Inspection and Quarantine (CAIQ), No. A3, Gaobeidian Road, Chaoyang District, Beijing 100123, China Nanjing Second Hospital, No. 121, Jiangjiayuan, Gulou District, Nanjing, Jiangsu, China * S Supporting Information ABSTRACT: This work reports an electrouorochromic strategy on the basis of electric eld control of uorescent signal generation on bipolar electrodes (BPEs) for visualizing cancer cell surface glycoprotein (mucin 1). The device included two separate cells: anodic sensing cell and cathodic reporting cell, which were connected by a screen-printing electrode patterned on poly(ethylene tereph- thalate) (PET) membrane. In the sensing cell, anti-MUC1 antibody immobilized on a chitosan-multiwalled carbon nanotube (CS- MWCNT)-modied anodic BPE channel was used for capturing mucin-1 (MUC1) or MCF-7 cancer cells. Then ferrocene (Fc)- labeled mucin 1 aptamers were introduced through hybridization. Under an applied voltage, the ferrocene was oxidized and the electroactive molecules of 1,4-benzoquinone (BQ) in the cathodic reporting cell were reduced according to electroneutrality. This produced a strongly basic 1,4-benzoquinone anion radical (BQ - ), which turned on the uorescence of pH-responsive uorescent molecules of (2-(2-(4-hydroxystyryl)-6-methyl-4H- pyran-4-ylidene)malononitrile) (SPM) coexisting in the cathode reporting cell for both spectrophotometric detection and imaging. This strategy allowed sensitive detection of MUC1 at a concentration down to 10 fM and was capable of detecting a minimum of three MCF-7 cells. Furthermore, the amount of MUC1 on MCF-7 cells was calculated to be 6.02 × 10 4 molecules/ cell. Our strategy also had the advantages of high temporal and spatial resolution, short response time, and high luminous contrast and is of great signicance for human health and the promotion of life science development. M ucin-1 (MUC1) is a glycoprotein that has been proved to participate in the metastasis and invasion of multiple tumor types. 1-3 In normal cells, the expression level of MUC1 is very low, whereas its upregulation is associated with the invasion, proliferation, and survival of tumor cells by reducing cell-cell adhesion and cell-extracellular matrix adhesion. 4,5 Therefore, sensitive and specic detection of MUC1 on the cancer cell surface is critical. Currently, the reported methods for MUC1 detection mainly include enzyme-linked immuno- sorbent assay (ELISA), 6 uorescence, 7 electrochemilumines- cence (ECL) 8,9 and electrochemical techniques. 10 However, ELISA and electrochemical detection cannot achieve signal visualization, uorescent labeling methods might cause damage to the cells themselves, and ECL methods cannot obtain high spatial and temporal resolution imaging. The combination of electrical and uorescent signals combines both the sensitivity of the electrochemistry and the visualization of uorescence imaging, enabling optical information to be switchable by the electric eld in situ. Electrouorochromism (EFC) has attracted a lot of attention, in which uorescence signals are monitored by an electrochemical redox process, 11-13 and it has great potential application in optical displays, 14,15 information encryption, 16 and information communications. 17 Electrouorochromic molecules can convert electrical signals to visual uorescence signals with high sensitivity. 18 Generally, electrouorochromic small molecules are privileged chemical structures, consisting of a uorophore bound to an electroactive group by a spacer. 19 An innovative method was developed to separate the uorophore and the electroactive group and make use of proton transfer to control emission. 20,21 Recently, Zhangs group developed an imaging tool, called uorescence-enabled electrochemical microscopy(FEEM), where a redox reaction is coupled to a uorogenic reporter reaction through a closed bipolar electrode (BPE). 22-24 The electrochemical reaction of interest takes place at one pole of the bipolar electrode, whereas a uorescent species is generated and imaged at the Received: April 10, 2019 Accepted: May 28, 2019 Published: May 28, 2019 Article pubs.acs.org/ac Cite This: Anal. Chem. 2019, 91, 7902-7910 © 2019 American Chemical Society 7902 DOI: 10.1021/acs.analchem.9b01760 Anal. Chem. 2019, 91, 7902-7910 Downloaded by SOUTHEAST UNIV at 05:10:19:597 on June 22, 2019 from https://pubs.acs.org/doi/10.1021/acs.analchem.9b01760.

Upload: others

Post on 18-Jan-2022

3 views

Category:

Documents


0 download

TRANSCRIPT

Visual Electrofluorochromic Detection of Cancer Cell SurfaceGlycoprotein on a Closed Bipolar Electrode ChipZhaoyan Tian,† Li Mi,† Yafeng Wu,*,† Fengying Shao,† Mingqiang Zou,δ Zhenxian Zhou,‡

and Songqin Liu*,†

†Key Laboratory of Environmental Medicine Engineering, Ministry of Education, Jiangsu Engineering Laboratory of SmartCarbon-Rich Materials and Device, School of Chemistry and Chemical Engineering, Southeast University, Nanjing, 211189, ChinaδChinese Academy of Inspection and Quarantine (CAIQ), No. A3, Gaobeidian Road, Chaoyang District, Beijing 100123, China‡Nanjing Second Hospital, No. 121, Jiangjiayuan, Gulou District, Nanjing, Jiangsu, China

*S Supporting Information

ABSTRACT: This work reports an electrofluorochromic strategy onthe basis of electric field control of fluorescent signal generation onbipolar electrodes (BPEs) for visualizing cancer cell surfaceglycoprotein (mucin 1). The device included two separate cells:anodic sensing cell and cathodic reporting cell, which were connectedby a screen-printing electrode patterned on poly(ethylene tereph-thalate) (PET) membrane. In the sensing cell, anti-MUC1 antibodyimmobilized on a chitosan-multiwalled carbon nanotube (CS-MWCNT)-modified anodic BPE channel was used for capturingmucin-1 (MUC1) or MCF-7 cancer cells. Then ferrocene (Fc)-labeled mucin 1 aptamers were introduced through hybridization.Under an applied voltage, the ferrocene was oxidized and theelectroactive molecules of 1,4-benzoquinone (BQ) in the cathodicreporting cell were reduced according to electroneutrality. This produced a strongly basic 1,4-benzoquinone anion radical(BQ•−), which turned on the fluorescence of pH-responsive fluorescent molecules of (2-(2-(4-hydroxystyryl)-6-methyl-4H-pyran-4-ylidene)malononitrile) (SPM) coexisting in the cathode reporting cell for both spectrophotometric detection andimaging. This strategy allowed sensitive detection of MUC1 at a concentration down to 10 fM and was capable of detecting aminimum of three MCF-7 cells. Furthermore, the amount of MUC1 on MCF-7 cells was calculated to be 6.02 × 104 molecules/cell. Our strategy also had the advantages of high temporal and spatial resolution, short response time, and high luminouscontrast and is of great significance for human health and the promotion of life science development.

Mucin-1 (MUC1) is a glycoprotein that has been provedto participate in the metastasis and invasion of multiple

tumor types.1−3 In normal cells, the expression level of MUC1is very low, whereas its upregulation is associated with theinvasion, proliferation, and survival of tumor cells by reducingcell−cell adhesion and cell−extracellular matrix adhesion.4,5

Therefore, sensitive and specific detection of MUC1 on thecancer cell surface is critical. Currently, the reported methodsfor MUC1 detection mainly include enzyme-linked immuno-sorbent assay (ELISA),6 fluorescence,7 electrochemilumines-cence (ECL)8,9 and electrochemical techniques.10 However,ELISA and electrochemical detection cannot achieve signalvisualization, fluorescent labeling methods might cause damageto the cells themselves, and ECL methods cannot obtain highspatial and temporal resolution imaging. The combination ofelectrical and fluorescent signals combines both the sensitivityof the electrochemistry and the visualization of fluorescenceimaging, enabling optical information to be switchable by theelectric field in situ.Electrofluorochromism (EFC) has attracted a lot of

attention, in which fluorescence signals are monitored by an

electrochemical redox process,11−13 and it has great potentialapplication in optical displays,14,15 information encryption,16

and information communications.17 Electrofluorochromicmolecules can convert electrical signals to visual fluorescencesignals with high sensitivity.18 Generally, electrofluorochromicsmall molecules are privileged chemical structures, consistingof a fluorophore bound to an electroactive group by a spacer.19

An innovative method was developed to separate thefluorophore and the electroactive group and make use ofproton transfer to control emission.20,21 Recently, Zhang’sgroup developed an imaging tool, called “fluorescence-enabledelectrochemical microscopy” (FEEM), where a redox reactionis coupled to a fluorogenic reporter reaction through a closedbipolar electrode (BPE).22−24 The electrochemical reaction ofinterest takes place at one pole of the bipolar electrode,whereas a fluorescent species is generated and imaged at the

Received: April 10, 2019Accepted: May 28, 2019Published: May 28, 2019

Article

pubs.acs.org/acCite This: Anal. Chem. 2019, 91, 7902−7910

© 2019 American Chemical Society 7902 DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

Dow

nloa

ded

by S

OU

TH

EA

ST U

NIV

at 0

5:10

:19:

597

on J

une

22, 2

019

from

http

s://p

ubs.

acs.

org/

doi/1

0.10

21/a

cs.a

nalc

hem

.9b0

1760

.

other pole of the electrode, which avoids mutual interferencebetween the two poles of the bipolar electrode located in twoseparate compartments.25−27 BPE is an electronic conductorimpregnated in electrolyte without direct electrical contact to apotential source, which facilitates miniaturization of suchdevices.28,29 When a certain external voltage is applied via twodriving electrodes, a linear gradient of potential dropped whichvaried across the length of the bipolar electrode.30−33 Whenthe interface potential is sufficient, faradaic reactions aretriggered at the two poles of the BPE. Because the amount ofelectrons gained or lost at both ends of the BPE is equal, thereduction of the electroactive material in the cathode of BPEand the oxidation in the anode are quantitatively related.34−36

In this paper, a homemade closed BPE device combinedwith EFC was designed to achieve sensitive monitoring ofMUC1 on MCF-7 cancer cells. The device included twoseparate cells: anodic sensing cell and cathode reporting cell,which were connected by a screen-printing electrode patternedon poly(ethylene terephthalate) (PET) membrane with goodelectrical insulation and mechanical stability. Chitosan-multi-walled carbon nanotubes (CS-MWCNTs) were modified onthe anodic BPE channel to enhance conductivity and surfacearea, and anti-MUC1 antibody (Ab1), MUC1 solution, orMCF-7 cancer cells were sequentially captured. Fc-labeledMUC1 aptamers for signal recognition were then introducedthrough hybridization for detection. Electroactive molecule1,4-benzoquinone (BQ) and pH-responsive fluorescentmolecule (2-(2-(4-hydroxystyryl)-6-methyl-4H-pyran-4-ylidene)malononitrile) (SPM) were added to the cathodereporting cell. In the original state, SPM has no fluorescence,but while a voltage of 3.5 V was applied, an electrochemicalredox process occurred, which caused the fluorescence of SPMto change from “OFF” to “ON”, which was recorded by bothspectrophotometry and imaging. The fluorescence intensitystrongly depended on the MUC1 expression level. Aconcentration down to 10 fM and a minimum of threeMCF-7 cells could be detected. With high visual sensitivity, ourwork opens a new door for the development of convenient,low-cost, sensitive, and portable detection methods onbiochips.

■ EXPERIMENTAL SECTIONMaterials and Reagents. Carboxylic group-functionalized

multiwalled carbon nanotubes (MWCNTs, diameter: 10−20nm, length: <2 μm, special surface area >200 m2/g) werepurchased from Nanjing Nano Pioneer Technology Co., Ltd.Polydimethylsiloxane (PDMS) and a curing agent wereobtained from Dow Corning (Midland, MI). Chitosan, (2,6-dimethyl-4H-pyran-4-ylidene)malononitrile (98%), and 4-hydroxybenzaldehyde (98%) were purchased from Aladdin(Shanghai, China) and used without further purification. 1,4-Benzoquinone (BQ, 99%) and tetrabutylammonium hexa-fluorophosphate (TBAPF6, 99%) were obtained from J&KScientific Ltd. (Beijing, China). Acetic acid (HAc) andglutaraldehyde (GA, 50 wt % in water) were purchased fromAlfa Aesar (Shanghai, China). Bovine serum albumin (BSA)was received from Beijing Solarbio Science & Technology Co.,Ltd. (Beijing, China). Mucin 1 (MUC1) and MUC1antibodies were purchased from Abcam (Shanghai, China).Fc-labeled MUC1 aptamer with a sequence of 5′-Fc-GCAGTTGATCCTTTGGATACCCTGG-3′ was purchasedfrom Shanghai Sangon Biotechnology Co., Ltd. (Shanghai,China). Human breast cancer cell lines MCF-7 cells were

received from Shanghai Institute for Biological Sciences ofChinese Academy of Science (Shanghai, China). RPMI 1640culture media and fetal bovine serum were purchased fromHyclone (Shanghai, China). Fluorescein diacetate (FDA) andpropidium iodide (PI) were purchased from KeyGEN Biotech(Nanjing, China). MCF-7 cells were cultured in RPMI 1640medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 U/mL streptomycin at 37 °C in ahumidified atmosphere of 5% CO2. A 0.1 M phosphate buffersolution (PBS, pH 7.4) containing 2.7 mM potassium chloride,136.8 mM sodium chloride, 1.5 mM KH2PO4, and 9.7 mMNa2HPO4 was used for cell culture. All other chemicals were ofanalytical grade. Milli-Q ultrapure water (18.2 MΩ cm) wasused throughout the experiments.

Instruments. The fluorescence emission spectra wererecorded by a fluorescence spectrometer with a slit width of3 nm (Fluoromax-4, Horiba Jobin Yvon, Japan). Confocalfluorescence imaging was achieved by confocal fluorescencemicroscopy (FV3000, Olympus Corporation, Japan). CHI750C electrochemical workstation (Shanghai Chenhua, China)with screen-printed electrode (SPE) as working electrode,saturated calomel electrode as reference electrode, andplatinum electrode as counter electrode were applied forelectrochemical measurements. 1H and 13C NMR spectra werecollected on a Bruker AV600 (600 MHz) spectrometer withCDCl3 as the solvent and tetramethylsilane (TMS) as aninternal standard.

Synthesis of (2-(2-(4-Hydroxystyryl)-6-methyl-4H-pyran-4-ylidene)malononitrile) (SPM). SPM was synthe-sized according to a previous report with slight modification.37

Briefly, (2,6-dimethyl-4H-pyran-4-ylidene)malononitrile (1.03g, 6 mmol) and 4-hydroxybenzaldehyde (0.37 g, 3 mmol) weremixed in 30 mL of toluene. Then 0.5 mL of piperidine and0.25 mL of acetic acid were added and stirred continuouslyunder Ar atmosphere. The reaction was allowed to continuefor 8 h under continuous refluxing. After removal of the solventunder vacuum, purifying the residue by column chromatog-raphy using dichloromethane/ethyl acetate (30:1), andrecrystallizing with ethyl acetate and n-hexane, 0.46 g of ayellow solid in a yield of 56% was thus obtained. 1H NMR(CDCl3, 600 MHz, TMS), δ (ppm): 2.45 (s, 3H, CH3), 6.67(s, 1H), 6.81−6.83 (t, 3H), 7.12 (d, J = 18.0 Hz, 1H), 7.46 (d,J = 18.0 Hz, 1H), 7.56 (d, J = 6 Hz, 2H), 10.06 (s, 1H); 13CNMR (CDCl3, 100 MHz), δ (ppm): 19.86, 55.52, 106.17,106.41, 115.90, 116.07, 116.44, 126.44, 130.48, 138.33, 157.21,160.22, 160.96, 164.43; TOF-SIMS: m/z calcd. forC17H12N2O2 276.1, found: 276.1 (Figures S1−3).

Device Fabrication. The bipolar electrode system wasprepared by printing carbon ink on poly(ethylene tereph-thalate) (PET) membrane. The closed BPE system consistedof three separate parts with a gap of 1.0 mm. The intermediatepart containing five separate electrodes (12.0 mm long × 0.5mm wide, lined with 0.5 mm gap) was used as the BPE, whiletwo side parts of the same size (10.0 mm long × 4.3 mm wide)were used as electrodes for voltage application. Forconstruction of the device, a mixture of PDMS monomerand curing agent at a ratio of 10:1 (after being degassed in avacuum chamber) was gently transferred into a rectangularmold and cured in an oven heated at 80 °C for 1 h. Aftercooling at room temperature, the cured PDMS was thendetached from the mold and cut into small slices (14.0 mm ×24.0 mm), and two square holes (6.0 mm) punched with thegap of 4.0 mm served as reservoirs and one 1.0 mm pore as

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7903

Administrator
高亮

sample inlet at the edge of each hole. After that, the PETmembrane with bipolar electrode arrays was integrated withthe above PDMS blocks using PDMS monomer and curingagent at a ratio of 20:1 and then heated to 80 °C for 1 h.For the anode interface modification, a CS-MWCNT

suspension was first prepared according to the literature.38,39

Briefly, a 3 mg mL−1 carboxylic MWCNT solution wasdispersed in 0.1% CS solution (in 0.1 M HAc) with sonicationto obtain a black homogeneous CS-MWCNT suspension.After 20 μL of CS-MWCNT composites was modified on theBPE anode, 2.5% GA was introduced and incubated for 3 h.After thoroughly washing with PBS (pH 7.4), 30 μL of 10 μgmL−1 anti-MUC1 antibody (Ab1) was added and incubated at4 °C for 12 h, followed by washing with PBS to remove thephysically adsorbed antibodies. Finally, 1 wt % BSA solutionwas added for 40 min to block the nonspecific adsorption onthe interface. Scanning electron microscopy demonstrated theformation of the BPE array on substrates (Figure S4).Fluorescence Detection of MUC1 in Solution and on

MCF-7 Cells on the Bipolar Electrode Chip. The Ab1-functionalized sensing compartment was incubated with 30 μLof different concentrations of MUC1 for 50 min, followed bywashing with PBS to remove the nonspecific adsorbedproteins. Then Fc-labeled MUC1 aptamer was introducedinto the sensing compartment for 50 min and subsequentlyrinsed extensively with deionized water to remove theunreacted aptamer. For the detection of MCF-7 cells, theimmunoassays were similar except that MUC1 was replaced byMCF-7 cells. Another compartment was filled with 0.1 MTBAPF6 acetonitrile solution containing 0.1 mM SPM and 1

mM BQ. For fluorescence spectroscopy, the device was set upin a homemade plexiglass support, where the fabricated chipwas fixed in a black box and the bipolar electrode cell andexcitation light source were at the same height (Figure S5). ADC power supply was connected to two electrodes of the as-prepared device to drive the electrochemical reaction, wherethe electrochemical products could be in-site monitored byfluorescence spectroscopy.

■ RESULTS AND DISCUSSION

Detection Principle. The strategy for detection of proteinson the cell surface using the closed BPE array to electricallycouple two separate redox reactions together with fluorescencemonitoring of the electrochemical products is illustrated inScheme 1A,B. In the sensing compartment, Fc-labeled MUC1aptamer attached to the sensing surface of the BPE arraythrough sandwiched MUC1 recognition reactions. When avoltage was applied between the two driving electrodes, anelectrochemical oxidation of the captured Fc occurred at onepole of the bipolar electrode in the sensing compartment,which triggered an opposite electrochemical process occurringat the other pole of the bipolar electrode in the detectioncompartment due to charge neutrality (Scheme 1C).36 Thecorresponding opposite electrochemical process was usuallydesigned as a fluorogenic redox reaction, which generated ahighly fluorescent product and was detected by fluorescencespectroscopy. Herein, the designed fluorogenic redox reactionwas the reduction of 1,4-benzoquinone (BQ), which generateda radical anion (BQ•−) with strong alkalinity that was able toswitch some pH-sensitive fluorescent material, such as SPM.37

Scheme 1. (A) Sensing Principle for the Detection of MUC1 on MCF-7 Cells. (B) Structure of the Designed Chip. (C) BasicPrinciple of the Closed Bipolar Electrode Chip

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7904

Therefore, an electrochemical process was electronicallycoupled to a fluorescence reaction that converted electro-chemical signals into fluorescent signals. The fluorescenceintensity at the detection compartment depended on theamount of Fc-labeled MUC1 aptamer brought to BPE,providing an approach for MUC1 detection. With this concept,the MUC1 expression on tumor cells was also evaluated,having advantages of high sensitivity, short response time, andhigh luminous contrast.To prove the fluorogenic redox reaction, the fluorescence

properties of SPM were systematically investigated and wereshown in Figure S6. SPM in acetonitrile solution displayed afaint green fluorescence with an emission peak at 507 nmunder excitation at 365 nm and negligible fluorescence underexcitation at 467 nm. In the presence of sodium tert-butoxide(t-BuONa), however, SPM displayed red fluorescence with aremarkable emission peak at 630 nm when excited at 467 nmdue to the formation of SPM−. This confirmed that SPM itselfhas almost no fluorescence under excitation at 467 nm, but itcould be switched to a fluorescence substrate by interactionwith an alkaline substance.In this work, the fluorescence of SPM could be “turned on”

by the alkaline substrate of BQ•−, which was generated by theelectrochemical reduction of BQ. Two separate electricallycoupled redox reactions occurred at the sensing and detectioncompartments, respectively, assisted by the closed BPE arrays.The electroactivity of Fc-aptamer on the sensing interface wasdemonstrated by a three-electrode system, where a screen-printed electrode (SPE) was used as a working electrode, asaturated calomel electrode was used as the referenceelectrode, and a platinum electrode was used as the counterelectrode. The cyclic voltammograms of the Fc-aptamer-coatedelectrode showed a pair of well-defined redox peaks located at0.12 V and −0.055 V, which corresponded to the reversibleredox of Fc/Fc+,40,41 confirming the effective direct electrontransfer between electrode and Fc attached to the sensingsurface through a sandwiched recognition reaction (FigureS7A). On the other hand, the cyclic voltammograms of SPE inacetonitrile solution containing BQ displayed a couple of redoxpeaks located at 0.11 V and −0.37 V (Figure S7B). Thisconfirmed an effective electrochemically driven redox reactionbetween BQ and BQ•−, and quinone anion radical could begenerated by the reduction of BQ. The CVs in the absence ofBQ demonstrated that oxygen reduction begins at around −0.5V (vs SCE) which was much more negative than that of BQ.

Moreover, the mixture of BQ and SPM in acetonitrile showedthe same electrochemical activity as BQ itself, indicating thatSPM was electrochemically stable at the potential window forBQ redox reactions. All these results demonstrated thesuccessful electron transfer between electrode and theelectroactive Fc or BQ. We reason that an oxidation reactionoccurred for the Fc-labeled MUC1 aptamer through asandwiched MUC1 recognition reaction in the sensingcompartment, while a reduction reaction of BQ occurred inthe detection compartment at a certain applied voltage. Asshown in Figure S8A, after CS-MWCNT composites weremodified on the BPE anode easily, the current intensity almostquadrupled compared with the bare BPE because of its largespecific surface area and good conductivity. The modificationof Ab1 sequentially at the sensing interface resulted in adecrease in current intensity at an applied voltage of 3.5 V.When MUC1 protein was coupled to the sensing interfacethrough immunoreaction, the current intensity was furtherdecreased due to the insulation of the protein shell. However,when Fc-labeled MUC1 aptamer attached to the sensinginterface, the current increased obviously, indicating that thepresence of Fc at the sensing interface increased the electrontransfer rate on BPE due to the successful oxidation of Fc. Onthe other hand, the addition of BQ in the detectioncompartment also led to the increase in the cathodic current(Figure S8B), confirming that the reduction of BQ alsofacilitated the electron transfer rate of BPE.As a proof of principle, the corresponding fluorescence

spectroscopy and visual confocal fluorescence imaging wasconducted (Figure 1). After coating with Fc-aptamer with asandwich recognition reaction, a distinct fluorescence peak at630 nm and a red fluorescence band can be observed at theBPE arrays in the detection compartment (curve a and imaginga in Figure 1A,B). Conversely, no fluorescence signal could beseen when the BPE arrays in the sensing compartment werepresented in PBS solution (curve b and imaging b), BPE arrayswithout modification with capture antibodies (curve c andimaging c) reacted with MUC1 and Fc-labeled aptamer, BPEarrays modified with capture antibodies but in absence oftarget mucin (curve d and imaging d) or reacted with MUC1but without attachment of Fc-labeled MUC1 aptamer (curve eand imaging e). Also, no fluorescence signal was observed atBPE arrays in the detection compartment when the detectionsolution was only TBAPF6/CH3CN (curve f and imaging f),TBAPF6/CH3CN containing BQ without SPM (curve g and

Figure 1. Fluorescence intensity at 630 nm (λex = 467 nm) (insert: the corresponding fluorescence spectra) (A) and visual confocal fluorescenceimaging (B) under different conditions. (a−e) The condition in the cathode reservoir was the same (1 mM BQ and 0.1 mM SPM in acetonitrilesolution containing 0.1 M TBAPF6). For the anode reservoir of BPE: modified 10 μg mL−1 MUC1 antibody, 15 nM MUC1, and 100 nM Fc-labeled MUC1 aptamer (a), only PBS (b), in the absence of Ab1 (c), in the absence of MUC1 (d), in the absence of Fc labeled aptamer (e). (f−h)The condition in the anode reservoir was the same. For the cathode reservoir of BPE: the solution was only TBAPF6/CH3CN (f), TBAPF6/CH3CN containing BQ without SPM (g), or TBAPF6/CH3CN containing SPM without BQ (h).

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7905

imaging g), or TBAPF6/CH3CN containing SPM without BQ(curve h and imaging h). All these results demonstrated thatstrong fluorescence signals attributed to the reduction of BQ toa basic anion radical in the site under the influence of anelectric field, which regulates the fluorescence of pH-sensitiveSPM molecules. The applied voltage of 3.5 V allowed theredox reactions of Fc and BQ to occur in the sensing anddetection compartments, respectively, which triggered thefluorogenic redox reaction.Detection of MUC1 with the Fluorogenic Redox

Reaction. Prior to MUC1 detection, the experimentalconditions including the applied voltage, Fc-labeled MUC1aptamer concentration, mucin incubation time, and electrolysistime were optimized. First, the applied voltage on BPE arrays

was crucial for the fluorogenic redox reaction. The reactions ofpoles of BPE are illustrated as follows:

− → = +− + Eanode reaction: Fc e Fc 0.12 V0

+ → = −− •− Ecathode reaction: BQ e BQ 0.37 V0

Therefore, the potential difference between the anode andcathode should be higher than 0.49 V to drive the reactions ofthe two poles of BPE. Considering the distance between thedrive electrodes, the length of the bipolar electrode, and thehigh solution resistance, the fluorescence results under drivingvoltages varying from 1.0 to 4.5 V were evaluated. As shown inFigure S9A, no fluorescence signal was observed when theapplied voltage was less than 1.5 V, indicating that the appliedvoltage was insufficient to initiate the redox reaction. The

Figure 2. Change in (A) fluorescence intensity and (B) confocal fluorescence imaging with electrolytic time in durations of 5, 10, 20, 40, 60, 80,and 100 s, respectively. The applied driving voltage was 3.5 V. The error bars represent the standard deviation of three replicate measurements.

Figure 3. Fluorescence emission spectra with the increasing concentration of MUC1 (insert: the variation trend) (A), the correspondingcalibration curve for MUC1 detection (B), the corresponding confocal fluorescence imaging (C), and fluorescence gray value extracted fromfluorescence imaging (D) with different concentrations of MUC1 from 0.3 pM to 15 nM under a voltage of 3.5 V.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7906

fluorescence intensity increased significantly with an increasein voltage from 1.5 to 3.5 V, which could be attributed to thesuccessful fluorogenic redox reaction. When the voltageexceeded 3.5 V, the fluorescence intensity in the detectioncompartment slightly decreased, most likely because benzo-quinone was significantly consumed at the cathode. Therefore,3.5 V was selected as the driving voltage in this work.Second, the fluorescence intensity in the detection compart-

ment generally depended on the amount of Fc-labeled MUC1aptamer being captured on the sensing interface. The amountof MUC1 in human serum is 0.7−39.8 kU L−1;42 therefore, inthis case, the MUC1 concentration used for coupling with theantibody in the sensing interface was selected as 10 μg mL−1.As the aptamer concentration increased from 0 to 200 nM, thefluorescence signal increased and further addition led to aplateau at 100 nM (Figure S9B).Third, the incubation time for MUC1 coupling and Fc-

labeled aptamer attachment was also optimized. Figure S9C,Dshows that the fluorescence signals were almost linearlyenhanced with an increase in incubation time. In thisexperiment, fluorescence intensity reached a maximum at 50min of MUC1 and Fc-labeled aptamer incubation time in theanodic pole of BPE and maintained a relatively stable value.Therefore, 50 min was chosen for the optimal reaction time.Last, electrolysis time was also investigated. Figure 2A,B

shows the fluorescence spectra and confocal imaging when avoltage of 3.5 V was applied at the driven electrodes for variousperiods. It was observed that the fluorescence intensityincreased with the electrolysis time and reached a maximumat 60 s. The confocal images showed that the red fluorescenceappeared gradually at the BPE arrays in the detectioncompartment. This was most likely because the electrochemi-cally generated BQ•− species reacted quickly with SPM to formfluorescent SPM− before diffusion into the solution.24 Thefluorescence emission was extremely bright, even visible to thenaked eye, and filled the entire compartment due to diffusionof the fluorescent SPM− species. Thereafter, the fluorescenceintensity slightly decreased with the extended electrolysis time,which was caused by the self-quenching phenomenon of thefluorescent molecules.43,44 As a result, an electrolysis time of60 s was selected as the optimized conditions for furtherinvestigation.Under the optimal conditions, the fluorescent signal

increased gradually as the concentration of MUC1 increasedfrom 0.3 pM to 15 nM (Figure 3). After that, fluorescenceintensity reached a plateau with the further increase in MUC1concentration (Figure 3A). The fluorescence intensity wasproportional to the logarithmic value of the MUC1concentration ranging from 0.3 pM to 15 nM (Figure 3B).The linear regression equation could be expressed as I =194593.7 log(C/pM) + 116568.3 with a correlation coefficientR2 of 0.994 (n = 9) and the limit of detection of 10 fM (S/N =3). As the fluorescence image shows in Figure 3C,D, a fittingequation of G = 18.73 log(C/pM) + 11.34 (correlationcoefficient R2 = 0.996) was obtained with Image J software,where G was the fluorescence gray value corresponding to acertain concentration of the sensing analyte. The LODobtained from the gray value was maintained in goodagreement with the fluorescence spectra.If MUC1 was determined directly with the electrochemical

method without coupling to the fluorogenic redox reaction(Supporting Information), a calibration range from 4 pM to 15nM with a linear correlation equation of I = 1.06 log(C/nM) −

0.63 was obtained (Figure S10). The detection limit wascalculated to be 1.36 pM. This indicated that the fluorescencedetection methods based on the fluorogenic redox reaction wasmore sensitive than that of the electrochemical method. Othercomparisons of MUC1 detection methods are listed in TableS1. This biochip had a wider detection range and lowerdetection limits, indicating successful electrochemicallycoupled fluorescent detection.To study the specificity of the biochip, an equal amount of

interfering protein, i.e., BSA, AFP, CA19-9, cTn, or MyO, wasdetermined (Figure S11). Only MUC1 caused significantenhancement of fluorescence intensity, whereas the signals ofother interfering proteins were almost the same as that of theblank. This revealed that the biochip was specific to MUC1.The reproducibility (inter- and intra-assays) of this biochip

was examined. As shown in Figure S12A, the intra-assayprecision was explored by testing the same chip five times andthe calculated relative standard deviation (RSD) was 4.48%.However, the interassay precision was examined by monitoringthe fluorescence intensity of five biochips fabricatedindependently and the RSD value was 6.90%. Therefore, thebiochip has relatively good stability performance. Additionally,the long-term storage stability was also studied by storing thebiochip at 4 °C in a refrigerator when not in use. The resultsindicate that the biochip retained 97.1% of its initial responseafter 2 weeks of storage, and its response decreased to 90.8%after 4 weeks (Figure S12B).To evaluate the accuracy of the proposed biochip for real

sample detection, human serum samples from healthy donorsat Nanjing Drum Tower Hospital were collected and diluted10-fold before use. A recovery study was carried out via astandard addition method, and different concentrations ofMUC1 were added into the serum samples (Table 1). The

recoveries of the human serum samples ranged from 94.44% to106.67% with the RSD less than 7.51%. This indicated that theproposed biochips were reliable for MUC1 detection in realsamples.

Detection of MUC1 Overexpressed on MCF-7 Cells.The anti-MUC1 antibody, different concentrations of MCF-7cells, and Fc-labeled MUC1 aptamers were modifiedsuccessively at the anode of the bipolar electrode. Fluorescenceintensity increased along with the increase of MCF-7 cells, asmore Fc-labeled MUC1 aptamers were captured by MCF-7cells. Fluorescence spectra were recorded under electro-chemical control (Figure 4A). The fluorescence peak increasedwith the increasing concentration of MCF-7 cells from 100 to1.0 × 106 cells mL−1 and reached a plateau corresponding toan excess of cells. As illustrated in Figure 4B, fluorescenceintensity was proportional to the logarithm of the concen-tration of MCF-7 cells in the range of 100 to 1.0 × 106 cells

Table 1. Recovery Results of MUC1 in the Serum Sampleswith the Biochipa

MUC1 addition found recovery (%) RSD (%)

1.00 pM 0.96 pM 96.00 5.472.50 pM 2.60 pM 104.00 7.5125.00 pM 23.61 pM 94.44 2.380.30 nM 0.32 nM 106.67 6.434.00 nM 4.00 nM 100.25 2.32

aDifferent concentrations of MUC1 were added to the diluted realhuman serum samples.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7907

mL−1. The standard curve equation was Y = 94613.2 log X −133464.5 (cells mL−1) with a detection limit of three cells (thevolume of the added sample was 30 μL) according to S/N = 3,in which Y was the fluorescence intensity and X was the cellnumber. The amount of MUC1 on MCF-7 cells was calculatedto be 6.02 × 104 molecules/cell. The gray value extracted fromFigure 4C corresponding to different numbers of cells is shownin Figure 4D. The percentage for LOD obtained from grayvalues was calculated to be less than 8.3% compared to thelinearity obtained from the fluorescence spectrum. To evaluate

the potential clinical application of the device, we comparedthe fluorescence intensity obtained from MCF-7 cells in PBSand serum samples and there was no obvious difference(Figure S13).On the anodic channel surface, three MCF-7 cells (stained

by fluorescein diacetate (FDA)) were clearly observed withfluorescence confocal microscopy (Figure 5). After Fc-aptamerwas introduced on the cell surface, the cathode cell was filledwith 0.1 M TBAPF6 acetonitrile solution containing 0.1 mMSPM and 1 mM BQ, and a weak fluorescent signal was

Figure 4. Variation trend of fluorescence intensity (insert: the corresponding fluorescence spectra) after incubation of different concentrations ofcells (A), a plot of fluorescence intensity versus the logarithm of the number of cells (B), the confocal fluorescence imaging with the number ofcaptured MCF-7 cells from 100 to 1 × 106 cells/mL (C), and the corresponding fluorescence gray value (D).

Figure 5. Confocal fluorescence images of MCF-7 cells (stained by FDA) in the anode of BPE and SPM in the cathode of BPE with an appliedvoltage of 3.5 V and electrolysis time of 60 s. Fc-aptamer was introduced on the MCF-7 cell surface in the anode cell, and the cathode cell was filledwith 0.1 M TBAPF6 acetonitrile solution containing 0.1 mM SPM and 1 mM BQ.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7908

exhibited when a voltage of 3.5 V was applied for 60 s. We alsostudied the influence of the applied voltage and electrolysistime on the MCF-7 cell viability with fluorescein diacetate(FDA) and propidium iodide (PI), which are for viable celland dead cell staining, respectively.45 Live/dead (green/red)cells were visualized by FDA/PI staining (Figure S14). Onlygreen cells were observed, confirming that cell viability was notaffected with an applied voltage of 3.5 V and electrolysis timeof 60 s. This biochip has potential as a feasible tool forquantitative single-cell based bioassay and heterogeneousanalysis of individual cells.To further evaluate the specificity of the biochip for MCF-7

cell detection, several types of interfering cells with low MUC1expression including liver cancer cells (HepG2), lungcarcinoma cell line (A-549), and a mixture of them wereexamined (Figure S15). As expected, a negligible fluorescenceintensity was observed for the interfering cells and a clearlyenhanced peak only in the presence of MCF-7 cells, which wasclear evidence of excellent specificity of the biochip.

■ CONCLUSIONSIn summary, a novel electrofluorochromic strategy forvisualizing cancer cell surface glycoprotein (MUC1) on thebasis of electric field control of fluorescent signal generation onBPEs was developed, which has both the sensitivity ofelectrochemistry and the visualization of fluorescence imaging.The sensing platform revealed good sensitivity and selectivityfor MUC1 by redox electroactive materials including Fc-labeled aptamers and 1,4-benzoquinone. MUC1 at aconcentration down to 10 fM and a minimum of threeMCF-7 cells were successfully detected. Furthermore, theamount of MUC1 on each MCF-7 cell was calculated to be6.02 × 104 molecules/cell. The proposed chip also possessedthe advantages of simple operation, minor sample consump-tion, good reproducibility, and acceptable accuracy. Our workprovides a new platform for detecting biomarker expression oncancer cells based on the combination of BPEs andelectrofluorochromic strategy and thus provides a powerfultool for disease diagnostics and clinical analysis.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.anal-chem.9b01760.

1H NMR, 13C NMR, and TOF-MS spectra of (2-(2-(4-hydroxystyryl)-6-methyl-4H -pyran-4-yl idene)-malononitrile), the description and figures of SEMcharacterization, picture of the real device, emissionspectra of SPM and SPM + t-BuONa under differentexcitation wavelengths, cyclic voltammograms andcurrent−potential curves of the electrode under differentconditions, the optimization of the operation parame-ters, DPV curves and linear calibration plots of differentconcentrations of MUC1 tested by the electrochemicalmethod, the specificity, reproducibility, and stability ofthe biochip, the comparison of the fluorescenceintensities obtained from MCF-7 cells in PBS andserum samples, confocal fluorescence images of MCF-7cells (stained by FDA/PI) in the anode after applying avoltage of 3.5 V for 60 s, comparison of differentmethods for the detection of MUC1 (PDF)

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: [email protected].*E-mail: [email protected] Wu: 0000-0003-0549-5420Songqin Liu: 0000-0002-4686-5291NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the National Key R&D Programof China (no. 2017YFF0108606), National Natural ScienceFoundation of China (nos. 21627806, 21635004, 21705018),and the Fundamental Research Funds for the CentralUniversities (2242017K3DN11).

■ REFERENCES(1) Hui, J. J.; Bao, L.; Li, S. Q.; Zhang, Y.; Feng, Y. M.; Ding, L.; Ju,H. X. Angew. Chem., Int. Ed. 2017, 56, 8139−8143.(2) He, Y.; Lin, Y.; Tang, H. W.; Pang, D. W. Nanoscale 2012, 4,2054−2059.(3) Li, Z.; Mao, G. B.; Du, M. Y.; Tian, S. B.; Niu, L. Q.; Ji, X. H.;He, Z. K. Microchim. Acta 2019, 186, 233.(4) Matsuda, A.; Kuno, A.; Nakagawa, T.; Ikehara, Y.; Irimura, T.;Yamamoto, M.; Nakanuma, Y.; Miyoshi, E.; Nakamori, S.; Nakanishi,H.; Viwatthanasittiphong, C.; Srivatanakul, P.; Miwa, M.; Shoda, J.;Narimatsu, H. Anal. Chem. 2015, 87, 7274−7281.(5) Zhang, Y. A.; Guo, S.; Huang, H. Y.; Mao, G. B.; Ji, X. H.; He, Z.K. Anal. Chim. Acta 2018, 1035, 154−160.(6) Ferreira, C. S. M.; Papamichael, K.; Guilbault, G.; Schwarzacher,T.; Gariepy, J.; Missailidis, S. Anal. Bioanal. Chem. 2008, 390, 1039−1050.(7) Kim, J.; Lee, E.; Kang, Y. Y.; Mok, H. Chem. Commun. 2015, 51,9038−9041.(8) Li, S. K.; Chen, A. Y.; Niu, X. X.; Liu, Z. T.; Du, M.; Chai, Y. Q.;Yuan, R.; Zhuo, Y. Chem. Commun. 2017, 53, 9624−9627.(9) Jiang, X. Y.; Wang, H. J.; Wang, H. J.; Zhuo, Y.; Yuan, R.; Chai,Y. Q. Anal. Chem. 2017, 89, 4280−4286.(10) Hu, R.; Wen, W.; Wang, Q. L.; Xiong, H. Y.; Zhang, X. H.; Gu,H. S.; Wang, S. F. Biosens. Bioelectron. 2014, 53, 384−389.(11) Miomandre, F.; Meallet-Renault, R.; Vachon, J. J.; Pansu, R. B.;Audebert, P. Chem. Commun. 2008, 1913−1915.(12) Maglione, M. S.; Casado-Montenegro, J.; Fritz, E. C.; Crivillers,N.; Ravoo, B. J.; Rovira, C.; Mas-Torrent, M. Chem. Commun. 2018,54, 3038−3041.(13) Sun, J. W.; Chen, Y. N.; Liang, Z. Q. Adv. Funct. Mater. 2016,26, 2783−2799.(14) Kim, Y.; Do, J.; Kim, E.; Clavier, G.; Galmiche, L.; Audebert, P.J. Electroanal. Chem. 2009, 632, 201−205.(15) Kuo, C. P.; Chang, C. L.; Hu, C. W.; Chuang, C. N.; Ho, K. C.;Leung, M. K. ACS Appl. Mater. Interfaces 2014, 6, 17402−17409.(16) Zhu, X.; Liu, R.; Li, Y.; Huang, H.; Wang, Q.; Wang, D.; Zhu,X.; Liu, S.; Zhu, H. Chem. Commun. 2014, 50, 12951−12954.(17) Raymo, F. M. Adv. Mater. 2002, 14, 401−414.(18) Lu, J.; Fan, Y. S.; Howard, M. D.; Vaughan, J. C.; Zhang, B. J.Am. Chem. Soc. 2017, 139, 2964−2971.(19) Al-Kutubi, H.; Zafarani, H. R.; Rassaei, L.; Mathwig, K. Eur.Polym. J. 2016, 83, 478−498.(20) Li, Z. X.; Zhou, Y. Q.; Peng, L. Q.; Yan, D. P.; Wei, M. Chem.Commun. 2017, 53, 8862−8865.(21) Doneux, T.; Bouffier, L.; Goudeau, B.; Arbault, S. Anal. Chem.2016, 88, 6292−6300.(22) Oja, S. M.; Zhang, B. Anal. Chem. 2014, 86, 12299−12307.(23) Oja, S. M.; Guerrette, J. P.; David, M. R.; Zhang, B. Anal. Chem.2014, 86, 6040−6048.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7909

(24) Guerrette, J. P.; Percival, S. J.; Zhang, B. J. Am. Chem. Soc.2013, 135, 855−861.(25) Jansod, S.; Cuartero, M.; Cherubini, T.; Bakker, E. Anal. Chem.2018, 90, 6376−6379.(26) Shi, H. W.; Zhao, W.; Liu, Z.; Liu, X. C.; Xu, J. J.; Chen, H. Y.Anal. Chem. 2016, 88, 8795−8801.(27) Wu, M. S.; Liu, Z.; Shi, H. W.; Chen, H. Y.; Xu, J. J. Anal.Chem. 2015, 87, 530−537.(28) Baek, S.; Kwon, S. R.; Yeon, S. Y.; Yoon, S. H.; Kang, C. M.;Han, S. H.; Lee, D.; Chung, T. D. Anal. Chem. 2018, 90, 4749−4755.(29) Zhai, Q. F.; Fan, D. Q.; Zhang, X. W.; Li, J.; Wang, E. K. NPGAsia Mater. 2017, 9, No. e421.(30) Khoshfetrat, S. M.; Ranjbari, M.; Shayan, M.; Mehrgardi, M. A.;Kiani, A. Anal. Chem. 2015, 87, 8123−8131.(31) Eßmann, V.; Zhao, F.; Hartmann, V.; Nowaczyk, M. M.;Schuhmann, W.; Conzuelo, F. Anal. Chem. 2017, 89, 7160−7165.(32) Loget, G.; Li, G.; Fabre, B. Chem. Commun. 2015, 51, 11115−11118.(33) Hao, R.; Fan, Y. S.; Han, C.; Zhang, B. Anal. Chem. 2017, 89,12652−12658.(34) Essmann, V.; Barwe, S.; Masa, J.; Schuhmann, W. Anal. Chem.2016, 88, 8835−8840.(35) Essmann, V.; Santana Santos, C.; Tarnev, T.; Bertotti, M.;Schuhmann, W. Anal. Chem. 2018, 90, 6267−6274.(36) Zhang, H. R.; Wang, Y. Z.; Zhao, W.; Xu, J. J.; Chen, H. Y.Anal. Chem. 2016, 88, 2884−2890.(37) Wang, X. J.; Li, W.; Li, W. R.; Gu, C.; Zheng, H. Z.; Wang, Y.Y.; Zhang, Y. M.; Li, M. J.; Zhang, S. X.-A. Chem. Commun. 2017, 53,11209−11212.(38) Zhang, X.; Ding, S. N. Biosens. Bioelectron. 2017, 94, 47−55.(39) Feng, Q. M.; Pan, J. B.; Zhang, H. R.; Xu, J. J.; Chen, H. Y.Chem. Commun. 2014, 50, 10949−10951.(40) Zhao, F. J.; Bai, Y.; Zeng, R. S.; Cao, L. L.; Zhu, J. M.; Han, G.C.; Chen, Z. C. Electroanalysis 2018, 30, 2774−2780.(41) Balaji Viswanath, K.; Krithiga, N.; Jayachitra, A.; Sheik Mideen,A. K.; Amali, A. J.; Vasantha, V. S. ACS Omega 2018, 3, 17010−17022.(42) Hasanzadeh, M.; Shadjou, N.; de la Guardia, M. TrAC, TrendsAnal. Chem. 2017, 91, 67−76.(43) Audebert, P.; Miomandre, F. Chem. Sci. 2013, 4, 575−584.(44) Belzile, M. N.; Godin, R.; Durantini, A. M.; Cosa, G. J. Am.Chem. Soc. 2016, 138, 16388−16397.(45) Sun, Y. N.; Song, W. H.; Sun, X. H.; Zhang, S. S. ACS Appl.Mater. Interfaces 2018, 10, 31054−31060.

Analytical Chemistry Article

DOI: 10.1021/acs.analchem.9b01760Anal. Chem. 2019, 91, 7902−7910

7910