Ultrasensitive ELISA Using Enzyme-Loaded Nanospherical Brushes asLabelsZhenyuan Qu,† Hong Xu,*,† Ping Xu,† Kaimin Chen,§ Rong Mu,† Jianping Fu,‡ and Hongchen Gu*,†

†State Key Laboratory of Oncogenes and Related Genes, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai200030, P. R. China§Chemistry and Chemical Engineering, Shanghai University of Engineering Science, Shanghai 201620, China‡Department of Mechanical Engineering, University of Michigan, Ann Arbor, Michigan 48109, United States

*S Supporting Information

ABSTRACT: Improving the detection sensitivity of enzyme-linked immunosorbent assay (ELISA) is of utmost importancefor meeting the demand of early disease diagnosis. Herein wereport an ultrasensitive ELISA system using horseradishperoxidase (HRP)-loaded nanospherical poly(acrylic acid)brushes (SPAABs) as labels. HRP was covalently immobilizedin SPAABs with high capacity and activity via an efficient“chemical conjugation after electrostatic entrapment” (CCEE)process, thus endowing SPAABs with high amplificationcapability as labels. The periphery of SPAAB-HRP was furtherutilized to bind a layer of antibody with high density forefficient capture of analytes owing to the three-dimensionalarchitecture of SPAABs. Using human chorionic gonadotro-phin (hCG) as a model analyte, the SPAAB-amplified system drastically boosted the detection limit of ELISA to 0.012 mIUmL−1, a 267-fold improvement as compared to conventional ELISA systems.

Enzyme-linked immunosorbent assay (ELISA) has becomethe gold standard for laboratorial and clinical analysis

owing to its simplicity, low cost, and easy operation andinstrumentation.1−3 However, one limitation of conventionalELISA using enzyme-antibody conjugates as labels is itsrelatively low sensitivity, which has become a bottleneck formeeting the ever-growing demand of early disease diagnosisbased on low-abundant biomarkers. To solve this problem,efforts have been made in developing more sensitivesubstrates,4 introducing biotin−streptavidin systems,5 andperforming ELISA in conjuction with polymerase chainreaction (immuno-PCR),6 etc. In recent years, the advent ofa variety of innovated immunoassay techniques has promotedthe frontier of detection sensitivity to an unprecedented level(∼fM or even lower).7−10 Representative techniques includebiobar-code assay,7 Erenna immunoassay,8 Digital ELISA,9 andplasmonic ELISA,10 to name a few.Among various strategies to achieve higher sensitivity of

ELISA, use of enzyme-loaded particles as labels presents asimple and promising method, where multiple enzymes areimmobilized on the surface of a single particle to improveELISA detection signal and thus enhance sensitivity.11,12

Currently, the most common way to prepare particle labels isthrough covalent immobilization of enzyme on particles via avariety of well-established conjugation methods.13 Many typesof particles including liposome,14 gold nanoparticles,15

polymeric particles,16 micrometer-sized magnetic particles,17

silica nanoparticles,12 and mesoporous silica nanoparticles18,19

have been reported as good candidates for this purpose.However, the current particle systems share several commonlimitations including a relatively low enzyme loading capacity14

and a significant loss of enzyme activity during theimmobilization process.20

To address these issues, herein we report the development ofan ultrasensitive immunoassay system by using sphericalpoly(acrylic acid) (PAA) brushes (SPAABs) as labels. SPAABshave many superior properties over conventional particles asenzyme carriers: the three-dimensional, flexible, and soft PAAbrushes can serve as an ideal scaffold for enzyme loading whilepreserving their biological activities.21−23 However, use ofenzyme-loaded SPAABs as labels in ELISA still presents asignificant challenge due to enzyme binding stability in SPAABsin a biological medium condition. In the present work, weleveraged a “chemical conjugation after electrostatic entrap-ment” (CCEE) method24 for covalent immobilization ofproteins in SPAABs to demonstrate for the first time thatenzyme-loaded SPAABs can be used as ultrasensitive andchemically stable labels in ELISA. Horseradish peroxidase(HRP), the most frequently used enzyme in ELISA, wasemployed. Human chorionic gonadotrophin (hCG), a bio-

Received: July 9, 2014Accepted: September 5, 2014Published: September 8, 2014

Letter

pubs.acs.org/ac

© 2014 American Chemical Society 9367 dx.doi.org/10.1021/ac502522b | Anal. Chem. 2014, 86, 9367−9371

marker for early pregnancy and trophoblastic tumor, was usedas the model analyte.The SPAAB-amplified ELISA immunoassay procedure was

illustrated in Scheme 1: (i) SPAAB-HRP complex was prepared

by covalent immobilization of HRP in SPAABs via the CCEEprocess (Scheme 1a); (ii) SPAAB-HRP was decorated with acorona of antibody (anti-β-hCG antibody) via the N-hydroxysuccinimide/N-(3-dimethylaminopropyl)-N′-ethyl-car-bodiimidehydrochloride (NHS/EDC) process to obtainSPAAB-HRP-Ab with dual functionalities of recognizingantigen and generating signals (Scheme 1b); (iii) ELISAdetection of hCG was achieved by a sandwich assay usingantibody (anti-α-hCG antibody) functionalized magnetic beadsas solid-phase substrates and SPAAB-HRP-Ab as labels(Scheme 1c). In the presence of analyte, an immunocomplexwould form and could be further separated by an externalmagnetic field. The concentration of analyte was quantifiedusing the signal amplified by HRP immobilized in SPAABs,which catalyzed its 3,3′,5,5′-tetramethylbenzidine (TMB)/H2O2 substrate to produce yellow products with the maximumabsorbance at 450 nm.The SPAABs used in this work was composed of an 80 nm

silica core surrounded by densely grafted, long-stretching PAAchains synthesized via surface-initiated RAFT polymerization(SI-RAFT)25 (see Table S1 in the Supporting Information forcharacterization information). The SI-RAFT process allows aprecise tailoring of PAA grafting density and lengths with high

uniformity.24,25 The CCEE process for covalent immobilizationof HRP (Scheme 1a) consisted of two sequential steps: tandemelectrostatic entrapment (step 1) and chemical conjugation(step 2). While step 1 leveraged the unique “Donnan effect”26

for SPAABs to achieve a heightened enzyme binding capacity,step 2 effectively turned labile electrostatic interaction intostable covalent binding, allowing resuspension of SPAAB-HRPinto the biological medium or buffer used in immunoassays.The pH of buffer in step 1 and the EDC dosage in step 2 wasoptimized to achieve the high binding capacity of HRP (FigureS1 in the Supporting Information). Figure 1a shows a TEMimage of SPAAB-HRP, displaying a clear core−shell structurewith a uniform size distribution. The 80 nm silica core wassurrounded by the PAA corona with an average dry thickness of44 nm.Successful immobilization of HRP into SPAABs could be

intuitively visualized by eye (Figure 1b). After immobilization,SPAABs possessed a characteristic brownish color of HRP.Immobilized HRP can be separated by centrifugation, leavingthe supernatant colorless. For further verification, the SPAAB-HRP complex was subjected to UV−visible spectrometry. Asshown in Figure 1c, the characteristic absorption of HRP wasclearly seen in SPAAB-HRP, demonstrating their successfulimmobilization in SPAABs. The binding capacity was estimatedto be about 677 μg mg−1 by subtracting the backgroundabsorption of SPAABs (which was fitted by an exponentialfunction27) and comparing with the absorption of free HRP at403 nm. Furthermore, no leaking of HRP was observed afterSPAAB-HRP was redispersed in PBS (as judged from A403 inthe supernatant). Noteworthily, excellent disparity of SPAABswas maintained after immobilization of HRP, as SPAAB-HRPcould be stored in a dispersion state for weeks without notableprecipitation. This property of HRP-loaded SPAABs isextremely important for their practical applications as labelsin immunoassays to obtain reproducible results. The strongelectrostatic repulsion and steric stabilization effect of PAAbrushes and the simple CCEE immobilization process mighthave both contributed to this superb dispersity of SPAAB-HRP.For comparison, conventional carboxylated silica nanoparticles(SiO2−COOH) with a similar size (90 nm) were synthesizedand conjugated with HRP via the NHS/EDC process atoptimal conditions. The binding capacity of SiO2−COOH wasestimated to be 14 μg mg−1 by the depletion method, much lessthan achieved using SPAABs.The catalytic property of SPAAB-HRP was measured using

TMB/H2O2 as a substrate. As shown in Figure 2a, activities ofsingle immobilized HRP relative to free HRP was estimatedfrom the initial slope of catalysis kinetics. This activity couldalso be determined by the end-point method that was adoptedin the subsequent immunoassay system (Figure 2b, see theSupporting Information for experimental conditions), where alinear relationship was observed between A450 and HRPconcentration and activities of HRP were calculated fromslopes of linear regression fitting. Both methods led to the sameconclusion that relative activities of immobilized HRP ascompared to free HRP (As) were ∼67% and 5.0% for SPAAB-HRP and SiO2−COOH-HRP, respectively. The remarkableenhancement in enzyme activity clearly demonstrated theadvantage of SPAABs as efficient enzyme carriers. The retainedhigh activity of immobilized HRP was consistent with previousreports,21 where enzymes are immobilized in SPAABs viaelectrostatic adsorption. Together, our data supported that theadditional EDC conjugation step in the CCEE method did not

Scheme 1. (a) Covalent Immobilization of HRP IntoSPAABs by CCEE Process, (b) Conjugation of Antibodyonto SPAAB-HRP via NHS/EDC Process, and (c) SPAABsAmplified Sandwich ELISAa

aImmunocomplex forms between SPAAB-HRP-Ab and immuno-magnetic beads in the presence of analytes and can be separated by anexternal magnetic field. The signal is generated from the yellowproduct of TMB/H2O2 substrate under the catalysis of immobilizedHRP.

Analytical Chemistry Letter

dx.doi.org/10.1021/ac502522b | Anal. Chem. 2014, 86, 9367−93719368

significantly affect activities of enzymes immobilized inSPAABs. The activity of immobilized enzymes in SPAABscould remain unchanged over a long period of time and keptthe same in a biological sample (Figure S2 in the SupportingInformation).The catalytic efficiency of SPAAB-HRP was characterized by

the number of effective HRP on a single particle, which wascalculated using the following equation:

=N NAeff s (1)

σ=NM

Mparticle

HRP (2)

where As described the enzymatic activity of a singleimmobilized HRP relative to a free HRP, N was the numberof HRP immobilized on a single particle, σ (mg HRP per mgparticle) was the enzyme binding capacity, and Mparticle andMHRP were molecular weights of particle and HRP (44kD),respectively. Compared with conventional particles, SPAABsachieved an improvement of 48-fold in binding capacity forHRP and 13.4-fold in maintaining their enzyme activity, whichtogether resulted in an almost 3-order of magnitude enhance-ment in catalysis efficiency (Table 1).After loading with HRP, the carboxyl groups remaining on

SPAABs could be utilized for antibody conjugation via theNHS/EDC process (Scheme 1b). The influence of antibody

Figure 1. (a) TEM image of SPAAB-HRP, the enlarged image clearly shows the core−shell structure of SPAAB-HRP. (b) States of SPAABs beforeand after immobilization of HRP. The color change of particles and their response to centrifugation indicate the success of HRP immobilization. FreeHRP cannot be separated by centrifugation. (c) UV−visible spectra of free HRP, SPAABs, and SPAAB-HRP.

Figure 2. Activity of immobilized HRP relative to free HRP measured by (a) catalysis kinetics and (b) end-point method using TMB/H2O2 assubstrates. The results were normalized to concentration of HRP for three forms of HRP.

Table 1. Comparison of SPAABs and SiO2−COOH as Carrier for HRP Immobilization

particles immobilization method Mparticlea σ (μg mg−1) Nb As (%) Neff

SPAABs-HRP CCEE 6.9 × 108a 677 10 600 67.3 7100SiO2−COOH-HRP NHS/EDC 5.4 × 108a 14 172 5.0 8.6

aSee the Supporting Information for the calculation method of particle molecular weight. bN was calculated by eq 2 using 44 kDa as the molecularweight of HRP.

Analytical Chemistry Letter

dx.doi.org/10.1021/ac502522b | Anal. Chem. 2014, 86, 9367−93719369

binding capacity and dosage of SPAAB-HRP-Ab on biosensingsignal detection was examined (Figure S3a in the SupportingInformation). A higher signal-to-background ratio (S/B) wasachieved by SPAAB-HRP-Ab with high antibody bindingcapacity (SPAAB-HRP-AbH). This observation is consistentwith the recent report that particles with a higher antibodycoverage possess a greater association constant with antigens28

due to multiple attachment. A higher antibody concentration insolution for SPAAB-HRP-Ab might also contribute to theenhanced signal by accelerating the immunocomplex formation.The optimal dosage of SPAAB-HRP-Ab was set as 1 μg, wherethe highest S/B was reached (Figure S3a in the SupportingInformation). In the same way, the concentration of anti-β-hCG-Ab-HRP was optimized to be 1 μg mg−1 (Figure S3b inthe Supporting Information). After conjugation of antibody,HRP-loaded SPAABs retained 60% of their catalysis activity(Figure S4 in the Supporting Information). Thus, one SPAAB-HRP-AbH carried ∼600 anti-β-hCG antibodies and 4200effective HRPs (see the Supporting Information for calculationmethods).ELISA detection of hCG using SPAAB-HRP-AbH as labels

was performed and compared with conventional assays usinganti-β-hCG-Ab-HRP as labels. Both systems were done at theirrespective optimal conditions. As shown in Figure 3, bothsystems exhibited excellent linearity between A450 and hCGconcentration for the analyte concentration ranges tested.Judging from the slopes of linear regression fitting for the twosystems, we conclude that optical signal from SPAAB-HRP-AbHwas 400-fold improved as compared to conventional assaysusing anti-β-hCG-Ab-HRP as labels (The dramatic amplifica-

tion of signal is visually evident, see Figure S5 in the SupportingInformation). On the other hand, the background of SPAAB-amplified system (A450 at hCG = 0) was slightly higher than theconventional system, which was reflected by the intercepts oftwo linear regression equations. As a result, the limit ofdetection (LOD), defined by the signal at zero analyteconcentration plus 3 standard deviations, was 0.012 mIUmL−1 (corresponding to 1.3 pg mL−1) for SPAAB-amplifiedsystem and 3.2 mIU mL−1 (corresponding to 0.35 ng mL−1) forconventional assay. By virtue of the ultrasensitive SPAAB-HRP-Ab labels, the detection sensitivity of ELISA was improved by267-fold, much greater than similar immunoassay systemsreported in the literature using conventional silica nanoparticlesof similar sizes.29−31The LOD of the present SPAAB-amplifiedimmunoassay system, which was in the low pg mL−1 range, wasmore sensitive or comparable to other similar particle-amplifiedimmunoassay systems.17−19,32−34 Compared with even moresensitive immunoassay systems developed so far,7−10 thepresent technique improved the sensitivity in a simple waywith no additional operation or specialized equipment needed.The SPAAB-amplified system could also be used for thedetection of biological samples (e.g., hCG in fetal bovineserum) with the sensitivity improved by about 200-fold (FigureS6 in the Supporting Information), exhibiting potentialapplicability in detecting clinical samples.The drastic improvement of LOD of the SPAAB-amplified

system clearly resulted from the efficient loading of HRP inSPAABs with their high activity properly preserved. In addition,the high density of antibody (134 μg per mg SPAABs,corresponding to ∼357 ng cm−2) immobilized on SPAAB-HRPalso contributed to an enhanced sensitivity. All these improve-ments were attributable to the unique 3D architecture ofSPAABs and our rational design of protein immobilizationscheme: the massive inner space binding sites were reserved forhigh-capacity HRP immobilization whose substrates were smallmolecules that could diffuse rather freely, while the peripheryregion was decorated with an antibody corona to effectivelycapture analytes of interest (as we show in Scheme 1). Such aspatial covalent protein immobilization scheme in SPAABs hasbeen previously studied by us,24 enabling a rational design andconvenient control of protein distribution in SPAABs withtandem CCEE and NHS/EDC processes. In contrast, inpreparation of conventional particles labels, competition ofantibody and enzyme molecule for binding sites on the particleis a common issue, which inevitably affects detection sensitivityby either sacrificing the efficiency of capturing analytes orlowering the amplification factor.30

In summary, we have developed an ultrasensitive immuno-assay system using HRP-loaded SPAABs as labels. The highcapacity and high activity for covalent immobilization of HRPin SPAABs endows SPAAB-HRP with remarkable signalamplification capability. We envision that the ultrasensitiveSPAAB-HRP can be used as universal amplification labels inbiosensing and molecular diagnostics.

■ ASSOCIATED CONTENT

*S Supporting InformationExperimental and additional figures (Figure S1−S6) andinformation as noted in the text. This material is availablefree of charge via the Internet at http://pubs.acs.org.

Figure 3. ELISA detection of hCG using TMB/H2O2 as substrates:(a) conventional assay using β-hCG-HRP as labels and (b) amplifiedsystem using SPAAB-HRP-Ab as labels (n = 3). The insets show themagnified results of low hCG concentration and have the samecoordinate with the original graphs.

Analytical Chemistry Letter

dx.doi.org/10.1021/ac502522b | Anal. Chem. 2014, 86, 9367−93719370

■ AUTHOR INFORMATION

Corresponding Authors*E-mail: xuhong@sjtu.edu.cn.*E-mail: hcgu@sjtu.edu.cn.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe acknowledge financial support for this work from the UM-SJTU Collaboration on Biomedical Technologies (to J.F. andH.G., Grant No. 12X120010007), 863 High Tech Program(Grants 2013AA032203, 2012AA020103), SJTU funding(Grant YG2013MS29), Shmec Project (Grant 14ZZ023), andthe U.S. National Science Foundation (Grant ECCS 1231826and Grant CBET 1263889 to J.F.).

■ REFERENCES(1) Lequin, R. M. Clin. Chem. 2005, 51, 2415−2418.(2) Wu, A. H. B. Clin. Chim. Acta 2006, 369, 119−124.(3) Crowther, J. R.; Walker, J. M. The ELISA Guidebook, 2nd ed.;Springer: New York, 2009; Vol. 149.(4) Frey, A.; Meckelein, B.; Externest, D.; Schmidt, M. A. J. Immunol.Methods 2000, 233, 47−56.(5) Gould, E.; Buckley, A.; Cammack, N. J. Virol. Methods 1985, 11,41−48.(6) Malou, N.; Raoult, D. Trends Microbiol. 2011, 19, 295−302.(7) Nam, J.-M.; Thaxton, C. S.; Mirkin, C. A. Science 2003, 301,1884−1886.(8) Todd, J.; Freese, B.; Lu, A.; Held, D.; Morey, J.; Livingston, R.;Goix, P. Clin. Chem. 2007, 53, 1990−1995.(9) Rissin, D. M.; Kan, C. W.; Campbell, T. G.; Howes, S. C.;Fournier, D. R.; Song, L.; Piech, T.; Patel, P. P.; Chang, L.; Rivnak, A.J.; Ferrell, E. P.; Randall, J. D.; Provuncher, G. K.; Walt, D. R.; Duffy,D. C. Nat. Biotechnol. 2010, 28, 595−599.(10) de La Rica, R.; Stevens, M. M. Nat. Nanotechnol. 2012, 7, 821−824.(11) Tekin, H. C.; Gijs, M. A. M. Lab Chip 2013, 13, 4711−4739.(12) Knopp, D.; Tang, D.; Niessner, R. Anal. Chim. Acta 2009, 647,14−30.(13) Rusmini, F.; Zhong, Z.; Feijen, J. Biomacromolecules 2007, 8,1775−1789.(14) Singh, A. K.; Kilpatrick, P. K.; Carbonell, R. G. Biotechnol. Prog.1995, 11, 333−341.(15) Xiao-Yan, Y.; Ying-Shu, G.; Sai, B.; Shu-Sheng, Z. Biosens.Bioelectron. 2009, 24, 2707−2711.(16) Dhawan, S. Peptides 2002, 23, 2099−2110.(17) Mani, V.; Chikkaveeraiah, B. V.; Patel, V.; Gutkind, J. S.;Rusling, J. F. ACS Nano 2009, 3, 585−594.(18) Yang, M.; Li, H.; Javadi, A.; Gong, S. Biomaterials 2010, 31,3281−3286.(19) Chen, L.; Zhang, Z.; Zhang, P.; Zhang, X.; Fu, A. Sens. Actuators,B 2011, 155, 557−561.(20) Rodrigues, R. C.; Ortiz, C.; Berenguer-Murcia, A.; Torres, R.;Fernandez-Lafuente, R. Chem. Soc. Rev. 2013, 42, 6290−6307.(21) Haupt, B.; Neumann, T.; Wittemann, A.; Ballauff, M.Biomacromolecules 2005, 6, 948−955.(22) Kudina, O.; Zakharchenko, A.; Trotsenko, O.; Tokarev, A.;Ionov, L.; Stoychev, G.; Puretskiy, N.; Pryor, S. W.; Voronov, A.;Minko, S. Angew. Chem., Int. Ed. 2014, 53, 483−487.(23) Kudina, O.; Zakharchenko, A.; Trotsenko, O.; Tokarev, A.;Ionov, L.; Stoychev, G.; Puretskiy, N.; Pryor, S. W.; Voronov, A.;Minko, S. Angew. Chem. 2014, 126, 493−497.(24) Qu, Z.; Chen, K.; Gu, H.; Xu, H. Bioconjugate Chem. 2014, 25,370−378.(25) Qu, Z.; Hu, F.; Chen, K.; Duan, Z.; Gu, H.; Xu, H. J. ColloidInterface Sci. 2013, 398, 82−87.

(26) Wittemann, A.; Haupt, B.; Ballauff, M. Phys. Chem. Chem. Phys.2003, 5, 1671−1677.(27) Duan, Z.; Qu, Z.; Hu, F.; Yang, Y.; Chen, G.; Xu, H. Appl. Surf.Sci. 2014, 300, 104−110.(28) Mani, V.; Wasalathanthri, D. P.; Joshi, A. A.; Kumar, C. V.;Rusling, J. F. Anal. Chem. 2012, 84, 10485−10491.(29) Nilsson, K. G. J. Immunol. Methods 1989, 122, 273−277.(30) Ke, R.; Yang, W.; Xia, X.; Xu, Y.; Li, Q. Anal. Biochem. 2010,406, 8−13.(31) Wu, Y.; Chen, C.; Liu, S. Anal. Chem. 2009, 81, 1600−1607.(32) Tang, D.; Ren, J. Anal. Chem. 2008, 80, 8064−8070.(33) Chen, L.; Chen, C.; Li, R.; Li, Y.; Liu, S. Chem. Commun. 2009,2670−2672.(34) Qian, J.; Zhang, C.; Cao, X.; Liu, S. Anal. Chem. 2010, 82,6422−6429.

Analytical Chemistry Letter

dx.doi.org/10.1021/ac502522b | Anal. Chem. 2014, 86, 9367−93719371