lable at ScienceDirect

Analytica Chimica Acta xxx (2017) 1e8

Contents lists avai

Analytica Chimica Acta

journal homepage: www.elsevier .com/locate/aca

Genetically functionalized ferritin nanoparticles with a high-affinityprotein binder for immunoassay and imaging

Jong-won Kim a, Woosung Heu b, Sukyo Jeong b, Hak-Sung Kim b, *

a Graduate School of Nanoscience and Technology, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 305-701, South Koreab Department of Biological Sciences, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 305-701, South Korea

h i g h l i g h t s

* Corresponding author.E-mail address: hskim76@kaist.ac.kr (H.-S. Kim).

http://dx.doi.org/10.1016/j.aca.2017.07.0600003-2670/© 2017 Elsevier B.V. All rights reserved.

Please cite this article in press as: J.-w. Kimimmunoassay and imaging, Analytica Chimi

g r a p h i c a l a b s t r a c t

� Human IgG-specific repebody wasgenetically fused to N-terminus offerritin subunit.

� The RbF-NPs exhibited 1000-foldhigher binding affinity for humanIgG than free repebody due tomultivalency.

� The Cy3-laleled RbF-NPs generatedstronger fluorescent signals thanCy3-labeled free repebody in immu-noassays and imaging.

� The RbF-NPs can be effectively usedin various types of immunoassaysand imaging.

a r t i c l e i n f o

Article history:Received 6 May 2017Received in revised form26 July 2017Accepted 28 July 2017Available online xxx

Keywords:RepebodyFerritinSelf-assemblyMultivalencyImmunoassayCell imaging

a b s t r a c t

Molecular detection of target molecules with high sensitivity and specificity is of great significance in bioand medical sciences. Here, we present genetically functionalized ferritin nanoparticles with a high-affinity protein binder, and their utility as a signal generator in a variety of immunoassays and imag-ing. As a high-affinity protein binder, human IgG-specific repebody, which is composed of LRR (Leucine-rich repeat) modules, was used. The repebody was genetically fused to the N-terminal heavy-chainferritin, and the resulting subunits were self-assembled to the repebody-ferritin nanoparticlescomposed of 24 subunits. The repebody-ferritin nanoparticles were shown to have a three-order ofmagnitude higher binding affinity toward human IgG than free repebody mainly owing to a decreaseddissociation rate constant. The repebody-ferritin nanoparticles were conjugated with fluorescent dyes,and the resulting nanoparticles were used for western blotting, cell imaging, and flow cytometricanalysis. The dye-labeled repebody-ferritin nanoparticles were shown to generate about 3-fold strongerfluorescent signals in immunoassays than monovalent repebody. The repebody-functionalized ferritinnanoparticles can be effectively used for sensitive and specific immunoassays and imaging in many areas.

© 2017 Elsevier B.V. All rights reserved.

1. Introduction

Self-assembled protein nanoparticles, such as virus-like

, et al., Genetically functionca Acta (2017), http://dx.doi.

particles, ferritins, or heat shock proteins, hold great promise asnatural nano-carriers with utility in biomedical and biomaterialsciences. Such protein-based nanoparticles are biocompatible andeasy to make a genetic or chemical modification. Among thesenanoparticles, ferritin has attracted considerable attention as pro-tein nano-cages for biotechnological and nanomedical applications

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060

J.-w. Kim et al. / Analytica Chimica Acta xxx (2017) 1e82

owing to its unique structural and biochemical features [1e3].Ferritin (450 kDa) is a hollow globular protein composed of 24subunits that self-assemble into a hollow cage-like structure withinternal and external diameters of about 8 and 12 nm, respectively[4]. Ferritin is present in every cell type, and its primary role in-volves intracellular iron-storage and transport in a non-toxic formand iron-transport in most living organisms [5]. Recently,numerous approaches have been attempted to use ferritin as op-tical imaging, MRI, and drug delivery systems for the diagnosis andtreatment of tumors [6e10].

Detection and analysis of a target molecule with high sensitivityand specificity is of great significance in biosciences and medicalsciences. Many advances have been made in the development ofdetection systems with desirable performance in terms of sensi-tivity, specificity, and reliability. In particular, various immunoassayformats based on antigen-antibody interaction have been of pri-mary choice in analysis of diverse biomolecules, finding wide ap-plications in many areas [11]. To increase the sensitivity ofimmunoassays, chemical conjugation of either enzymes or nano-particles to antibodies has been used [12e16]. However, chemicalconjugation of signal generators to antibodies can produce unsta-ble, heterogeneous, and randomly cross-linkedmolecules, resultingin a loss of binding affinity and activity. In addition, it is difficult tocontrol the conjugation stoichiometry. In an effort to overcomesuch drawbacks, alternative binders have been recently attemptedfor immunoassays, including aptamers, Fc binding domain andsingle chain antibody (scFv) [12,17e19]. Nonetheless, the devel-opment of a new immunoassay format with high sensitivity re-mains a challenge.

We previously developed a repebody scaffold composed of LRR(Leucine-rich repeat) modules [20]. The repebody was shown to beexpressed at high level in E. coli and stable against heat, pH andproteolytic digestion. Repebodies for various targets could easily beselected through a phage display, and their utility and potential as aprotein binder with high specificity have been demonstrated[21e23]. Herein, we present genetically functionalized ferritinnanoparticles with a human IgG-specific repebody and their utilityfor immunoassay and imaging with high sensitivity and specificity.The repebody was fused to a ferritin subunit, and the resultingsubunits were self-assembled to ferritin nanoparticles with highhomogeneity and stability. Biophysical property of each ferritinnanoparticles was investigated in terms of size and multivalency.We demonstrate the utility and potential of fluorescent dye-labeledrepebody-ferritin nanoparticles in immunoassay and imaging.Details are reported herein.

2. Materials and methods

2.1. Gene expression and protein purification

A repebody specifically binding to human IgG was previouslydeveloped [21], and used as a targeting moiety in this work. Anti-human IgG repebody (rF4) was genetically fused to the N-termi-nus of heavy chain ferritin subunit [24] using a (GSS)4 linker. Theconstructed gene contained a 6x-His tag at the C-terminal of therepebody for affinity purification. The gene coding for therepebody-fused ferritin subunit was cloned into a pET21a vector(Invitrogen) between the Nde I and Xho I restriction enzyme sites.The vector was transformed into BL21 (DE3) E. coli cells. Thetransformed cells were grown in a LB medium at 37 �C until theoptical density at 600 nm reached 0.5, and IPTGwas added at a finalconcentration of 0.5mM for induction. Cells were further incubatedat 18 �C for 20 h and harvested by centrifugation at 4000g. Collectedcells were suspended in a lysis buffer (50 mM NaH2PO4, 300 mMNaCl, and 10 mM imidazole, at pH 8.0) and disrupted using

Please cite this article in press as: J.-w. Kim, et al., Genetically functionimmunoassay and imaging, Analytica Chimica Acta (2017), http://dx.doi.

sonication. Following centrifugation at 16000 rpm for 50 min, thesupernatant was collected and purified using affinity chromatog-raphy through Ni-NTA Superflow (Qiagen). The solution wasapplied to the resin-packed column, followed by a washing with abuffer (50 mM NaH2PO4, 300 mM NaCl, and 50 mM imidazole, atpH 8.0) until no protein was detected by a Bradford assay. Theprotein was eluted using an elution buffer (50 mM NaH2PO4,300 mM NaCl, and 250 mM imidazole, at pH 8.0). The collectedproteins were further purified through gel permeation chroma-tography (Superdex 200, GE Healthcare) using a phosphate-buffered saline (PBS, pH 7.4). The concentrations of repebody-ferritin nanoparticles and free repebody were determined bymeasuring the absorbance at 280 nm.

2.2. Transmission electron microscopy (TEM)

Repebody-ferritin nanoparticles and native ferritinwere stainedwith phosphotungstic acid for TEM images. Briefly, the sample wasprepared by dropping ferritin solution on formvar/carbon grid, andthe solution was evaporated under ambient condition. Phospho-tungstic acid (2%, pH 7.4) was dropped on the grid, stained for1 min, and completely removed using a filter paper. Images wereobtained using a 200 kV field-emission transmission electron mi-croscope (JEM-2100F, JEOL LTD., Japan) following further drying thegrid under ambient condition.

2.3. Measurement of hydrodynamic size

Repebody-ferritin nanoparticles and native ferritins werediluted in 1 � PBS buffer solution (pH 7.4) and were subjected toanalysis of dynamic light scattering (DLS) to determine the hy-drodynamic size using a Zetasizer nano zs (Malvern).

2.4. Binding kinetics analysis by surface plasmon resonance (SPR)

The association and dissociation kinetics of repebody-ferritinnanoparticles and free repebody were determined using a SPR(BIACORE 3000, Biacore AB, Sweden). Human IgG (Sigma, USA) wasimmobilized on a sensor chip CM5 through EDC/NHS conjugationchemistry. A mixture of 0.4 M 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC) and 0.1 M N-hydrox-ysulfosuccinimide (sulfo-NHS) was used to activate the chip sur-face, and 10 mg mL�1 human IgG in sodium acetate buffer (10 mM,pH 4.0) was injected for immobilization onto the chip. Ethanol-amine (1.0 M, pH 8.0) was used to block excessive activated func-tional groups. Repebody-ferritin nanoparticles and free repebodywere passed through at various concentrations, and sensorgramswere fitted into a 1:1 Langmuir binding model using the softwareprovided by the supplier to estimate the association and dissocia-tion rate constants. The dissociation constant was determined bydividing the dissociation rate constant by the association rateconstant.

2.5. Cell culture

MDA-MB-468 (human breast cancer), A431 (human epidermoidcarcinoma), MCF-7 (human breast cancer), and SK-Br-3 (humanbreast cancer) cells were maintained in adequate medium. A431cells were cultured in DMEM medium. RPMI 1640 medium wasused for cultivation of MDA-MB-468, MCF-7 and SK-Br-3 cells. Cellswere grown in media supplemented with 10% fetal bovine serum(GE Healthcare Hyclone, UK) in a 5% CO2 incubator (MCO-5AC,Sanyo, Japan) at 37 �C.

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060

J.-w. Kim et al. / Analytica Chimica Acta xxx (2017) 1e8 3

2.6. Labeling with fluorescent dyes

Repebody-ferritin nanoparticles and free repebody were labeledwith Cy3 fluorescent dyes. Ferritin nanoparticles or repebodieswere incubated with 30-fold molar excess of Cy3-NHS (Lumiprobe)in 1 � PBS buffer (pH 7.4) for 2 h at room temperature. Followingthe reaction, remaining non-reactive Cy3-NHS was removed usingdesalting column (PD-10, GE Healthcare). The resulting repbody-ferritin nanoparticles and repebodies were concentrated by usinga 10 K centrifugal filter (Merck Millipore, Germany) and used forwestern blotting, cell imaging, and flow cytometric analysis.

2.7. Western blot analysis using repebody-ferritin nanoparticles

For western blot analysis, we chose EGFR and HER2 as targets.20 mg of cell lysates was separated by 6% SDS-PAGE gel and trans-ferred to nitrocellulose membrane (Bio-rad). The membrane wasblocked with PBST (0.05% Tween-20, pH 7.4) containing 1% BSA for1 h at room temperature. Cetuximab (Merck) and Trastuzumab(Roche) were used as primary antibodies. Each of 3 mg mL�1 pri-mary antibodies was incubated with the membrane at 4 �C over-night, and free antibodies were removed by washing three timeswith PBST at room temperature. The membrane was incubatedwith Cy3-labeled repebodies and repebody-ferritins (8 mg mL�1

and 3 mg mL�1, respectively) to visualize target proteins as a sec-ondary antibody for 1 h at room temperature, and finally washedthree times with PBST. Cy3 signal was acquired using ChemiDocimaging system (Bio-rad). Fluorescent intensity was measured us-ing ImageJ software (https://imagej.net).

2.8. Cell imaging

MDA-MB-468 and SK-Br-3 were used as EGFR- and HER2-positive cell lines, respectively. Cells were grown in 8-chamberculture slide (SPL) for 2 days. After incubation, the cells werewashed with Dulbecco's phosphate-buffered saline (DPBS), andeach of 2 mg mL�1 Cetuximab or Trastuzumab in a serum-freemedium was treated for 1 h at 4 �C, and washed three times withDPBS. Cy3-labeled repebody-ferritin nanoparticles (2 mg mL�1) orrepebodies (5 mg mL�1) were added to a serum-free medium, fol-lowed by incubation for 1 h at 4 �C, and washed three times withDPBS. Cells were fixed with 4% paraformaldehyde for 10 min andwashed three times with DPBS, followed by DAPI staining. Thefluorescent images were obtained using a confocal laser micro-scopy (LSM 710, Carl Zeiss, Germany).

2.9. Flow cytometry

MDA-MB-468 and SK-br-3 (1 � 106 cells) were harvested andresuspended in 1 mL cold PBS containing 2% BSA. The 100 ml of cellsuspension was mixed with each of 1 mg mL�1 Cetuximab orTrastuzumab for 30 min on ice, washed three times with 2% BSA/PBS and stained by 1 mg mL�1 Cy3-labeled repebody-ferritinnanoparticles or repebodies. After incubation for 30 min on ice,stained cells were washed and analyzed using LSR Fortessa (BDBiosciences, USA). Experimental data were analyzed using FACS-Diva software (BD Biosciences).

2.10. Performance and stability test

To test the performance of repebody-ferritin nanoparticles withrespect to pH, immunofluorescence assays were carried out usingCy3-labeled repebody-ferritin nanoparticles and Cy3-labeledrepebody at different pH conditions. Buffers used were: 10 mMsodium acetate buffer (pH 4.0, 4.3, 5.3), 10 mM PIPES (pH 6.5), PBS

Please cite this article in press as: J.-w. Kim, et al., Genetically functionimmunoassay and imaging, Analytica Chimica Acta (2017), http://dx.doi.

(pH 7.4), and 50 mM sodium bicarbonate buffer (pH 8.5). For im-munoassays, 10 mg mL�1 of human IgG was coated onto a 96-wellplate (Nunc) at 4 �C overnight, followed by blocking with eachbuffer solution containing 0.05% Tween 20 and 1% BSA for 1 h atroom temperature. After washing, 10 mg mL�1 of Cy3-labeledrepebody-ferritin nanoparticles or Cy3-labeled repebody whichhad been diluted in each of blocking buffer were added to eachwell. Following incubation for 1 h at room temperature andwashing three times with each buffer containing 0.05% Tween 20,the fluorescent intensity wasmeasured using an Infinite N200 platereader (Tecan) at an excitation wavelength of 540 nm and anemission wavelength of 575 nm, respectively. To assess the storagestability, Cy3-labeled repebody-ferritin nanoparticles and Cy3-labeled repebody were incubated in PBS (pH 7.4) for 28 days, andwere subjected to fluorescence immunoassay using the samemethod as described above.

3. Results and discussion

3.1. Construction and characterization of repebody-ferritinnanoparticles (RbF-NPs)

We previously selected a repebody specifically binding to Fc ofhuman IgG [21], and employed it as a targeting moiety in this work.To develop multivalent repebody-ferritin nanoparticles (RbF-NPs)as a signal generator for immunoassays, human IgG-specific repe-body was fused to the N-terminus of a heavy chain ferritin subunitthrough a GS linker. It was expected that genetically-fused repe-bodies are exposed outward of ferritin nanoparticles. We analyzedthe resulting RbF-NPs by gel permeation chromatography and SDS/PAGE analysis (Fig. S1). Genetic fusion of the repebody with aferritin subunit was verified through an increase in the size ofsubunit and assembly of the RbF-NP. We next checked the shapeand size of the RbF-NPs using transmission electron microscopy(TEM) and dynamic light scattering (DLS) (Fig. 1). TEM image ofnegatively stained RbF-NPs showed a homogenous morphologyand spherical cage-like structure as native ferritin nanoparticles(Fig. 1A and B), indicating a negligible effect of genetically fusedrepebody on the self-assembly of 24 subunits of ferritin. Hydro-dynamic diameter of the RbF-NPs was estimated to be 15.1 nmthrough DLS (Fig. 1C), whereas the size of native ferritin nano-particles was around 11.6 nm, and an increase in the size seems tobe due to genetically fused repebody. This result implies that agenetic fusion of the repebody to the N-terminus of ferritin subunithad a negligible effect on the formation of cage-like ferritin throughself-assembly as soluble protein nanoparticles with high stability.

A distinct feature of our approach lies in simple and straight-forward self-assembly of the repebody-ferritin nanoparticles (RbF-NPs) through genetic fusion. It was reported that fusion of otherprotein binder to a ferritin subunit often results in the formation ofinsoluble aggregates in E. coli cells [9,25]. Repebodies specific forvarious targets could be easily developed through a phage displayand modular engineering, and they were expressed at high level inE. coli, showing high stability [20,21]. The RbF-NPs were shown tobe expressed as a soluble form, and their expression level reachedabout 35 mg per 1 L culture. Based on the result, a variety of RbF-NPs with different targeting moiety seems to be obtained by asimple process without additional purification and assembly steps.

3.2. Multivalency of repebody-ferritin nanoparticles

It has been known that a multivalent system usually confers asignificantly enhanced binding affinity toward a target throughbinding avidity and thereby leads to high sensitivity and specificityin immunoassays [26,27]. In this work, RbF-NPs were observed to

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060

Fig. 1. Characterization of RbF-NPs. TEM images of (A) native ferritin nanoparticles. (B) RbF-NPs stained with 1% phosphotungstic acid. (C) Size distribution of dynamic lightscattering.

J.-w. Kim et al. / Analytica Chimica Acta xxx (2017) 1e84

self-assemble from repebody-fused ferritin subunits, and each RbF-NP is likely to have 24 repebodymolecules on its surface. To check ifthe RbF-NPs display the multivalency, we determined the bindingaffinity of the RbF-NPs for human IgG using surface plasmonicresonance (SPR). Human IgG was immobilized on the surface ofCM-5 sensor chip, and different concentrations of the RbF-NPswere injected into a channel of chip, and the dissociation con-stant KD was determined. As a result, the KD value of the RbF-NPsfor human IgG was estimated to be 615 pM, whereas free repe-body showed a KD value of 620 nM (Fig. 2A and B). It is interestingto note that the RbF-NPs exhibited a three-order of magnitudehigher binding affinity than free repebody, indicating a distinctmultivalent feature of the RbF-NPs. As shown in Table 1, the asso-ciation rate constant of the RbF-NPs was shown to be around 3-foldhigher than that of free repebody, whereas the RbF-NPs had a340-fold lower dissociation rate constant, resulting in a significantincrease in the binding affinity. Previous studies also revealed thatenhanced binding affinity of multivalent system is linked to thedissociation rate constant rather than the association rate constant[28,29], supporting our result. It is likely that a significant increasein the binding affinity of the RbF-NPs for human IgG resulted fromtheir multivalent interaction with human IgG.

Fig. 2. SPR sensorgrams for binding of human IgG to (A) RbF-NPs and (B) freerepebody.

3.3. Immunoassay using repebody-ferritin nanoparticles

To use RbF-NPs for immunoassay, RbF-NPs were labeled withCy3. For comparison, monovalent repebody was also labeled withCy3 using the same method as RbF-NPs, and the resulting dye-conjugated repebody was used for immunoassay. The dye to pro-tein ratios (molar basis) for the RbF-NPs and free repebody wereestimated to be 18 and 2, respectively. The RbF-NPs have a numberof lysine residues on their convex surface, and these residues can bedecorated with diverse functional molecules. We chose epidermalgrowth factor receptor (EGFR) and HER2 as model targets. EGFRand HER2 are widely used biomarkers for cancer diagnosis andtreatment [30e32]. For immunoassays, Cetuximab and

Please cite this article in press as: J.-w. Kim, et al., Genetically functionimmunoassay and imaging, Analytica Chimica Acta (2017), http://dx.doi.

Trastuzumab were employed to detect EGFR or HER2 from analytesas primary antibodies, respectively, followed by addition of

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060

Table 1Kinetic parameters for association and dissociation of RbF-NPs and free repebodyagainst human IgG.

ka [L mol�1 s�1] kd [s�1] KA [mol�1 L] KD [mol L�1]

RbF-NPs þ hIgG 4.78 � 105 2.94 � 10�4 1.63 � 109 6.15 � 10�10

Repebody þ hIgG 1.56 � 105 0.10 1.61 � 106 6.21 � 10�7

J.-w. Kim et al. / Analytica Chimica Acta xxx (2017) 1e8 5

Cy3-labeled RbF-NPs and repebody. We determined the optimalconcentrations of the Cy3-labeled RbF-NPs and repebody whichgave rise to highest signals from respective target proteins withminimal noise signals.

For western blot analysis, we used cell lysates from EGFR-positive MDA-MB-468, A431 and HER2-positive SK-Br-3 cells. Asshown in Fig. 3A and B, the use of the Cy3-labeled RbF-NPsgenerated intensive and clear fluorescent signals compared to theCy3-labeled repebody even though a lower amount of the RbF-NPswas used. To directly compare the performance of the RbF-NPs withmonovalent repebody, we measured the fluorescent intensity fromthe corresponding protein band in western blot analysis when thesame mass concentration of the Cy3-labeled RbF-NPs and freerepebody was used. As a result, the RbF-NPs showed about three-fold higher fluorescent intensity than free repebody for EGFR andHER2 (Fig. 3C). Considering molecular weights of RbF-NPs(1250 kDa) and free repebody (29 kDa), it is noteworthy that themolar concentration of the RbF-NPs was around 43 times lowerthan free repebody when the equal mass concentration was used.This result seems to be due to a large number of labeled dyes andhigh binding affinity of the RbF-NPs compared to free repebody,demonstrating the utility of RbF-NPs as a signal generator forimmunoassays.

We tested applicability of RbF-NPs to imaging of cell surfacetargets. EGFR-positiveMDA-MB-468 cells and HER2-positive SK-Br-3 cells were treated with Cetuximab and Trastuzumab as primaryantibodies, respectively, followed by addition of Cy3-labeled RbF-NPs. As shown in Fig. 4A and B, the Cy3-lableled RbF-NPs exhibi-ted a strong fluorescent signal compared to Cy3-labeled repebodythrough specific binding to human IgGs in images. Measurement of

Fig. 3. Western blot analysis using (A) Cy3-labeled RbF-NPs and (B) Cy3-labeled repebodynitrocellulose membrane was incubated with 3 mg mL�1 primary antibody, Cetuximab or Traof Cy3-labeled repebody for staining the target bands. Molecular weights of EGFR and HER2the corresponding protein bands in western blot analysis when the equal mass concentrati

Please cite this article in press as: J.-w. Kim, et al., Genetically functionimmunoassay and imaging, Analytica Chimica Acta (2017), http://dx.doi.

the fluorescent intensity revealed that the RbF-NPs generatedabout three-fold higher intensity than free repebody for imaging ofEGFR and HER2 when the equal mass concentration of the RbF-NPsand repebody was used (Fig. 4C). In the absence of primary humanIgGs, on the other hand, Cy3-labeled repebody and RbF-NPs dis-played a negligible signal, which indicates that repebodies on RbF-NPs retained their binding capacity toward human IgGs (Fig. S2).Based on the result, it is plausible that RbF-NPs can be effectivelyused for immunofluorescent imaging.

3.4. Flow cytometric analysis using repebody-ferritin nanoparticles

We investigated if RbF-NPs can be used for flow cytometricanalysis. EGFR-expressing MDA-MB-468 and HER2-expressing SK-Br-3 cells were treated with respective EGFR-targeting and HER2-targeting human IgGs as primary antibodies, followed by incuba-tion with the same amount of Cy3-labeled RbF-NPs or Cy3-labeledrepebody, and cells were analyzed by flow cytometry. As a result,the use of Cy3-labeled RbF-NPs gave rise to much stronger fluo-rescent signals for both cell lines, than free repebody even thoughthe amounts of Cy3-labeled RbF-NPs and Cy3-labeled repebodywere the same (Fig. 5A and B). Regardless of primary antibodies,such as Cetuximab and Trastuzumab, the RbF-NPs showed astronger signal than free repebody. This result indicates that therepebody (F4) specifically bound to Fc fragment of human IgG, andthe signal intensity was mainly dependent on the binding ability ofprimary antibodies. The difference of signal intensity between theRbF-NPs and free repebody in MDA-MB-468 cells was smaller thanthat in SK-Br-3 cells. It is likely that the signal of the RbF-NPs fordetection of EGFR was already saturated at 1 mg mL�1 of RbF-NPs(Fig. 5A). Taken together, the RbF-NPs can be used for flowcytometry analysis of cells.

3.5. Performance and stability of repebody-ferritin nanoparticles

Since the binding affinity of human IgG-specific repebody wasknown to decrease at acidic pH [21], the performance of RbF-NPswould also be affected by pH. We evaluated the pH-dependent

for detection of EGFR and HER2. In western blot analysis, 20 mg of cell lysates on thestuzumab, followed by incubation with 3 mg mL�1 of Cy3-labeled RbF-NPs or 8 mg mL�1

are 175 kDa and 185 kDa, respectively. (C) Quantification of fluorescence intensity fromon (3 mg mL�1) of Cy3-labeled RbF-NPs or Cy3-labeled repebody was used.

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060

Fig. 4. Immunofluorescent images of cell surface receptors using Cy3-labeled RbF-NPs. (A) Image of MDA-MB-468 cells expressing EGFR. (B) Image of SK-Br-3 cells expressing HER2.Cells were treated with 2 mg mL�1 Cetuximab or Trastuzumab as a primary antibody, followed by incubation with 2 mg mL�1 of Cy3-labeled RbF-NPs or 5 mg mL�1 of Cy3-labeledrepebody. DAPI (blue) was used to stain the nucleus in merged images. Scale bar ¼ 50 mm. (C) Quantification of fluorescence intensity from the immunofluorescent images when thesame mass concentration (2 mg mL�1) of Cy3-labeled RbF-NPs or Cy3-labeled repebody was used. (For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

J.-w. Kim et al. / Analytica Chimica Acta xxx (2017) 1e86

performance of RbF-NPs by direct immunofluorescence assayusing the same amount of Cy3-labeled RbF-NPs and Cy3-labeledrepebody. As expected, the fluorescent intensities of the RbF-NPs and free repebody were shown to decrease with thedecreasing pH, which is coincident with the previous result(Fig. S3). In contrast, the RbF-NPs exhibited much stronger signalsthan free repebody at each pH, which implies that the RbF-NPsretained multivalency and binding characteristics of the repe-body even at acidic pH condition.

To evaluate storage stability, the Cy3-labeled RbF-NPs and Cy3-labeled repebody were incubated in PBS buffer (pH 7.4) at 4 �C indark for 28 days, and subjected to fluorescence immunoassay usingthe same method as described above. As shown in Fig. S4, no sig-nificant change in the fluorescence intensity of the RbF-NPs wasobserved even after 28 days. In addition, the RbF-NPs exhibitedhigher intensity than free repebody. Based on the result, it is likelythat the RbF-NPs maintained their structural integrity and conse-quently target-binding capability during a long-term storage.

4. Conclusion

We have demonstrated that RbF-NPs can be effectively used forsensitive immunoassays and imaging of cells. The repebody-fusedferritin subunits were shown to self-assemble to cage-like ferritinnanoparticles with high stability and homogeneity. It is noteworthy

Please cite this article in press as: J.-w. Kim, et al., Genetically functionimmunoassay and imaging, Analytica Chimica Acta (2017), http://dx.doi.

that the RbF-NPs showed a significantly enhanced binding affinitytoward a target compared to free repebody through multivalentinteraction with a target, namely binding avidity. Furthermore,single RbF-NP has a number of conjugated fluorescent dyes, andconsequently the dye-labeled RbF-NPs exhibited a strong fluores-cent signal compared to monovalent repebody in immunoassays.Analytical methods with high sensitivity and specificity find wideapplications in biosciences and medical sciences. It was reportedthat magneto-ferritin, which possesses a nanoparticle core of ironoxide, shows a catalytic activity like peroxidase [33]. Integration ofa repebody with such magneto-ferritin is anticipated to enablemore sensitive immunoassays through catalytic activity, instead ofconjugation with fluorescent dyes. Based on the results, it is likelythat the present assay format based on the RbF-NPs can be effec-tively used in various types of immunoassays. Ferritin nano-particles have attracted considerable interest as drug deliveryvehicle, even though their use in immunoassays was recentlyattempted. Ferritin nanoparticles were functionalized with varioustargeting moieties mainly through chemical conjugation to aminoacids on ferritin, and the resulting ferritin nanoparticles were usedfor drug delivery and diagnosis [7,34,35]. The present approachallowed straightforward and simple production of RbF-NPs withhigh target specificity and affinity, finding wide applications to notonly immunoassays, but also diagnosis and drug delivery.

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060

Fig. 5. Flow cytometric analysis of the cells expressing EGFR and HER2 using dye-labeled RbF-NPs. (A) EGFR-positive MDA-MB-468 cells. (B) HER2-expressing SK-Br-3 cells. Cells were incubated with 1 mg mL�1 Cetuximab or Trastuzumab as a primaryantibody, respectively, followed by treatment with 1 mg mL�1 of Cy3-labeled RbF-NPsor Cy3-labeled repebody, and subjected to analysis using a flow cytometry. Histogramsfor EGFR and HER2 expression levels were plotted using 1 mg mL�1 of Cy3-labeled RbF-NPs (red) and Cy3-labeled repebody (green). Negative controls were measured usingCy3-labeled RbF-NPs (dark gray) and Cy3-labeled repebody (light gray) in the absenceof primary antibodies. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

J.-w. Kim et al. / Analytica Chimica Acta xxx (2017) 1e8 7

Acknowledgments

This research was supported byMid-Career Researcher Program(NRF-2014R1A2A1A01004198), Global Research Laboratory Pro-gram (NRF-2015K1A1A2033346) and Bio & Medical TechnologyDevelopment Program (NRF-2017M3A9F5031419) through theNational Research Foundation of Korea funded by the Ministry ofScience, ICT and Future Planning.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.aca.2017.07.060.

References

[1] P. Rucker, F.M. Torti, S.V. Torti, Recombinant ferritin: modulation of subunitstoichiometry in bacterial expression systems, Protein Eng. 10 (1997)967e973.

Please cite this article in press as: J.-w. Kim, et al., Genetically functionimmunoassay and imaging, Analytica Chimica Acta (2017), http://dx.doi.

[2] M. Kim, Y. Rho, K.S. Jin, B. Ahn, S. Jung, H. Kim, M. Ree, pH-dependentstructures of ferritin and apoferritin in solution: disassembly and reassembly,Biomacromolecules 12 (2011) 1629e1640.

[3] S. Stefanini, S. Cavallo, C.Q. Wang, P. Tataseo, P. Vecchini, A. Giartosio,E. Chiancone, Thermal stability of horse spleen apoferritin and human re-combinant H apoferritin, Arch. Biochem. Biophys. 325 (1996) 58e64.

[4] J. Tatur, W.R. Hagen, P.M. Matias, Crystal structure of the ferritin from thehyperthermophilic archaeal anaerobe Pyrococcus furiosus, J. Biol. Inorg. Chem.12 (2007) 615e630.

[5] H.N. Munro, M.C. Linder, Ferritin - structure, biosynthesis, and role in iron-metabolism, Physiol. Rev. 58 (1978) 317e396.

[6] H.J. Kang, Y.J. Kang, Y.M. Lee, H.H. Shin, S.J. Chung, S. Kang, Developing anantibody-binding protein cage as a molecular recognition drug modularnanoplatform, Biomaterials 33 (2012) 5423e5430.

[7] Z. Zhen, W. Tang, H. Chen, X. Lin, T. Todd, G. Wang, T. Cowger, X. Chen, J. Xie,RGD-modified apoferritin nanoparticles for efficient drug delivery to tumors,ACS Nano 7 (2013) 4830e4837.

[8] M.P. Hwang, J.W. Lee, K.E. Lee, K.H. Lee, Think modular: a simple apoferritin-based platform for the multifaceted detection of pancreatic cancer, ACS Nano7 (2013) 8167e8174.

[9] X. Li, L. Qiu, P. Zhu, X. Tao, T. Imanaka, J. Zhao, Y. Huang, Y. Tu, X. Cao,Epidermal growth factor-ferritin H-chain protein nanoparticles for tumoractive targeting, Small 8 (2012) 2505e2514.

[10] K.C. Kwon, H.K. Ko, J. Lee, E.J. Lee, K. Kim, J. Lee, Enhanced in vivo tumordetection by active tumor cell targeting using multiple tumor receptor-binding peptides presented on genetically engineered human ferritin nano-particles, Small 12 (2016) 4241e4253.

[11] J. Wu, Z.F. Fu, F. Yan, H.X. Ju, Biomedical and clinical applications of immu-noassays and immunosensors for tumor markers, Trac-Trend Anal. Chem. 26(2007) 679e688.

[12] M. Iijima, T. Matsuzaki, N. Yoshimoto, T. Niimi, K. Tanizawa, S. Kuroda, Fluo-rophore-labeled nanocapsules displaying IgG Fc-binding domains for thesimultaneous detection of multiple antigens, Biomaterials 32 (2011)9011e9020.

[13] Y. Zhao, Y. Zheng, C. Zhao, J. You, F. Qu, Hollow PDA-Au nanoparticles-enabledsignal amplification for sensitive nonenzymatic colorimetric immunode-tection of carbohydrate antigen 125, Biosens. Bioelectron. 71 (2015) 200e206.

[14] D. Tang, B. Su, J. Tang, J. Ren, G. Chen, Nanoparticle-based sandwich electro-chemical immunoassay for carbohydrate antigen 125 with signal enhance-ment using enzyme-coated nanometer-sized enzyme-doped silica beads,Anal. Chem. 82 (2010) 1527e1534.

[15] Y.L. Wang, H.M. Ma, X.D. Wang, X.H. Pang, D. Wu, B. Du, Q. Wei, Novel signalamplification strategy for ultrasensitive sandwich-type electrochemicalimmunosensor employing Pd-Fe3O4-GS as the matrix and SiO2 as the label,Biosens. Bioelectron. 74 (2015) 59e65.

[16] K.C. Han, E.G. Yang, D.R. Ahn, A highly sensitive, multiplex immunoassay usinggold nanoparticle-enhanced signal amplification, Chem. Commun. 48 (2012)5895e5897.

[17] J. Wang, A. Munir, Z. Li, H.S. Zhou, Aptamer-Au NPs conjugates-accumulatedmethylene blue for the sensitive electrochemical immunoassay of protein,Talanta 81 (2010) 63e67.

[18] Y. Liu, et al., Single chain fragment variable recombinant antibody function-alized gold nanoparticles for a highly sensitive colorimetric immunoassay,Biosens. Bioelectron. 24 (2009) 2853e2857.

[19] J. Wang, A. Munir, Z. Li, H.S. Zhou, Aptamer-Au NPs conjugates-enhanced SPRsensing for the ultrasensitive sandwich immunoassay, Biosens. Bioelectron. 25(2009) 124e129.

[20] S.C. Lee, K. Park, J. Han, J.J. Lee, H.J. Kim, S. Hong, W. Heu, Y.J. Kim, J.S. Ha,S.G. Lee, H.K. Cheong, Y.H. Jeon, D. Kim, H.S. Kim, Design of a binding scaffoldbased on variable lymphocyte receptors of jawless vertebrates by moduleengineering, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 3299e3304.

[21] W. Heu, J.M. Choi, J.J. Lee, S. Jeong, H.S. Kim, Protein binder for affinity puri-fication of human immunoglobulin antibodies, Anal. Chem. 86 (2014)6019e6025.

[22] J.J. Lee, H.J. Choi, M. Yun, Y. Kang, J.E. Jung, Y. Ryu, T.Y. Kim, Y.J. Cha, H.S. Cho,J.J. Min, C.W. Chung, H.S. Kim, Enzymatic prenylation and oxime ligation forthe synthesis of stable and homogeneous protein-drug conjugates for tar-geted therapy, Angew. Chem. Int. Ed. 54 (2015) 12020e12024.

[23] S. Jeong, W. Heu, J.W. Kim, H.S. Kim, Protein binders specific for immuno-globulin g from different species for immunoassays and multiplex imaging,Anal. Chem. 88 (2016) 11938e11945.

[24] M. Dhar, V. Chauthaiwale, J.G. Joshi, Sequence of a cDNA encoding the ferritinH-chain from an 11-week-old human fetal brain, Gene 126 (1993) 275e278.

[25] P.K. Dehal, C.F. Livingston, C.G. Dunn, R. Buick, R. Luxton, D.J. Pritchard,Magnetizable antibody-like proteins, Biotechnol. J. 5 (2010) 596e604.

[26] J.H. Myung, K.A. Gajjar, J. Saric, D.T. Eddington, S. Hong, Dendrimer-mediatedmultivalent binding for the enhanced capture of tumor cells, Angew. Chem.Int. Ed. 50 (2011) 11769e11772.

[27] X. Zhu, L. Wang, R. Liu, B. Flutter, S. Li, J. Ding, H. Tao, C. Liu, M. Sun, B. Gao,COMBODY: one-domain antibody multimer with improved avidity, Immunol.Cell Biol. 88 (2010) 667e675.

[28] S. Hong, P.R. Leroueil, I.J. Majoros, B.G. Orr, J.R. Baker Jr., M.M. Banaszak Holl,The binding avidity of a nanoparticle-based multivalent targeted drug de-livery platform, Chem. Biol. 14 (2007) 107e115.

[29] X. Montet, M. Funovics, K. Montet-Abou, R. Weissleder, L. Josephson,

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060

J.-w. Kim et al. / Analytica Chimica Acta xxx (2017) 1e88

Multivalent effects of RGD peptides obtained by nanoparticle display, J. Med.Chem. 49 (2006) 6087e6093.

[30] F. Ciardiello, G. Tortora, Epidermal growth factor receptor (EGFR) as a target incancer therapy: understanding the role of receptor expression and othermolecular determinants that could influence the response to anti-EGFR drugs,Eur. J. Cancer 39 (2003) 1348e1354.

[31] C. Gravalos, A. Jimeno, HER2 in gastric cancer: a new prognostic factor and anovel therapeutic target, Ann. Oncol. 19 (2008) 1523e1529.

[32] H.J. Burstein, The distinctive nature of HER2-positive breast cancers, New

Please cite this article in press as: J.-w. Kim, et al., Genetically functionimmunoassay and imaging, Analytica Chimica Acta (2017), http://dx.doi.

Engl. J. Med. 353 (2005) 1652e1654.[33] K. Fan, C. Cao, Y. Pan, D. Lu, D. Yang, J. Feng, L. Song, M. Liang, X. Yan, Mag-

netoferritin nanoparticles for targeting and visualizing tumour tissues, Nat.Nanotechnol. 7 (2012) 459e464.

[34] K. Li, Z.P. Zhang, M. Luo, X. Yu, Y. Han, H.P. Wei, Z.Q. Cui, X.E. Zhang, Multi-functional ferritin cage nanostructures for fluorescence and MR imaging oftumor cells, Nanoscale 4 (2012) 188e193.

[35] Z. Zhen, W. Tang, T. Todd, J. Xie, Ferritins as nanoplatforms for imaging anddrug delivery, Expert Opin. Drug Deliv. 11 (2014) 1913e1922.

alized ferritin nanoparticles with a high-affinity protein binder fororg/10.1016/j.aca.2017.07.060