S

Utn

XKQ

a

ARR1AA

KCE4D

1

oiZfstaWcbaMclomvD

0d

Biosensors and Bioelectronics 24 (2009) 2707–2711

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journa l homepage: www.e lsev ier .com/ locate /b ios

hort communication

ltrasensitive enhanced chemiluminescence enzyme immunoassay forhe determination of �-fetoprotein amplified by double-codified goldanoparticles labels

iao-Yan Yang, Ying-Shu Guo, Sai Bi, Shu-Sheng Zhang ∗

ey Laboratory of Eco-chemical Engineering, Ministry of Education, College of Chemistry and Molecular Engineering, Qingdao University of Science and Technology,ingdao 266042, China

r t i c l e i n f o

rticle history:eceived 8 September 2008eceived in revised form2 November 2008ccepted 1 December 2008

a b s t r a c t

A novel enhanced chemiluminescent (CL) immunoassay for ultrasensitive determination of �-fetoprotein(AFP) was reported. The method made full use of 4-(4′-iodo)phenylphenol (IPP) as a new potential signalenhancer and double-codified gold nanoparticles (DC-AuNPs) labels modified with horseradish peroxi-dase (HRP)-conjugated anti-AFP used for further signal amplification. This protocol involved a sandwich

vailable online 11 December 2008

eywords:hemiluminescent immunoassaynhancer-(4′-iodo)phenylphenol

format, in which the antigen in the sample was first captured by the immobilized primary antibody onthe surface of magnetic beads, and then recognized by the second antibody labeled with DC-AuNPs. Thecombination of the remarkable sensitivity of the enhanced CL method and the use of AuNPs as an anti-AFP–HRP carrier for the enzymatic signal amplification, provided a linear response range of AFP from0.008 to 0.3 ng mL−1 with an extremely low detection limit of 5 pg mL−1, much lower than those achievedby the classical enzyme-linked immunosorbent assay (ELISA). This new system can be easily extended to

tion

ouble-codified gold nanoparticles a variety of immunodetec

. Introduction

Gold nanoparticles (AuNPs) have been widely used in numer-us bioassays (Rosi and Mirkin, 2005; Gómez–Hens et al., 2008)nvolving luminescent (Fan et al., 2005; Ao et al., 2006; Chen andu, 2007) or electrochemical detection (Selvaraju et al., 2008). Theunctional gold nanoparticles that are linked to biological moleculesuch as proteins, enzymes, and nucleic acids, provide interestingools for several biological systems (Ackerson et al., 2006; Stoeva etl., 2006; You et al., 2006; Rosi et al., 2006; Sapsford et al., 2006;ang et al., 2008; Pavlov et al., 2004; Tang et al., 2008). AuNPs

an provide a biocompatible microenvironment for HRP, which haseen the most widely used for immunoassay as carriers, and greatlymplify enzymatic signal (Cui et al., 2008; Li et al., 2008). Recently,erkoci et al. reported a novel double-codified nanolabel AuNPs

onsisting of gold nanoparticles conjugated to a peroxidase (HRP)-abeled anti-human IgG antibody (DC-AuNPs) for the determinationf IgG by using either a spectrophotometric or an electrochemical

ethod (Ambrosi et al., 2007). These detection methods have pro-

ided a versatile, simple tool for the protein detection as well asNA analysis.

∗ Corresponding author. Tel.: +86 532 84022750; fax: +86 532 84022750.E-mail address: shushzhang@126.com (S.-S. Zhang).

956-5663/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.bios.2008.12.009

as well as DNA analysis.© 2008 Elsevier B.V. All rights reserved.

Because of the high sensitivity, rapidity of reaction, simpleinstrumentation, and wide dynamic range, chemiluminescence(CL) assay has been used as an attractive analytical method indifferent fields, such as biotechnology, pharmacology, molecu-lar biology, and clinical environmental chemistries (Elbaz et al.,2008; Zhou et al., 2006; Konry et al., 2005; Kuroda et al., 2000).However, the sensitivity of CL immunoassay needs to be furtherimproved to meet the demand in clinical disease diagnosis andtreatment. An alternative approach for signal amplification in CLimmunoassay is to use an efficient enhancer or increase the numberof labels conjugated to each antibody. Recently, many improve-ments including enhanced sensitivity and reduced detection timehave been made in CL immunoassay (Previte et al., 2006; Ahn etal., 2007). For example, p-iodophenol (PIP) is the most popularenhancer of luminol–H2O2-horseradish peroxidase (HRP) CL reac-tion and applied to a wide variety of immunoassays and nucleic acidhybridization assays (Zhou et al., 2006). However, the applicationof DC-AuNPs in enhanced CL enzyme immunoassay has not beenreported, to the best of our knowledge.

Herein, we developed a novel ultrasensitive enhanced CLenzyme immunoassay based on the combination role with 4-(4′-

iodo)phenylphenol (IPP) as a new enhancer and double-codifiedgold nanoparticles (DC–AuNPs) labels for further signal ampli-fication. AFP, an oncofetal protein which was widely used as atumor marker for diagnosis and management of hepatocellularcarcinoma, was chosen to prove the novel CL immunoassay as a

2708 X.-Y. Yang et al. / Biosensors and Bioelectronics 24 (2009) 2707–2711

F munom MBs).d RP. (f)A

t(maaMttstmswd

2

2

An3R&(ot(N

ig. 1. Schematic protocol of the developed sandwich-type CL enzyme-linked imagnetic beads (MBs). (b) activation of a carboxylate moiety on the surface of (

ifferent concentrations of the human AFP antigen. (e) incubation with anti-AFP–HFP/AFP/anti-AFP–HRP. (h) enhanced CL detection of MB/anti-AFP/AFP/DC-AuNPs.

ypical model. The principle of CL detection with magnetic beadsMBs) as bimolecular immobilizing carriers to separate target

olecules is depicted in Fig. 1. The new CL immunoassay involvedbinding event between carboxylic acid coated MBs and anti-AFP,nd the formation of sandwich immunocomplexes between theBs and anti-human-AFP–HRP or the DC-AuNPs. The concentra-

ion of AFP was monitored based on the HRP label activity towardhe oxidation of luminol, which was quantified by CL method. Aignificant amplification for the detection of AFP was obtained byhe use of a new efficient enhancer and gold nanoparticles as a

ulti-AFP–HRP carrier, which therefore amplified the enzymaticignal. In view of these intrinsic advantages, this method haside potential applications in immunoassay and pathogenicetection.

. Experimental

.1. Reagents

An anti-�-fetoprotein (AFP) antibody and a HRP-labeledFP were purchased from Biocell. Co. Ltd. (China). Paramag-etic microbeads coated with carboxyl groups (particle size:.0–4.0 �m) were purchased from Tianjin BaseLine ChroTechesearch Centre (China). Luminol was purchased from ABCR GmbH

Co. (Germany). p-Iodophenol (PIP), 4-(4′-iodo)phenylphenol

IPP) were purchased from Merry (China). Horseradish per-xidase (∼250 U mg−1), hydrogen peroxide (H2O2), hydrogenetrachloroaurate(III) tetrhydrate (HAuCl4·4H2O), 1-ethyl-3-3-dimethylaminopropyl)carbodiimide hydrochloride (EDC),-hydroxysuccinimide (NHS), trisodium citrate, imidazole, and

assay by using magnetic beads: (a) introduction of carboxyl groups-coated para-(c) incubation with the primary anti-human AFP antibody. (d) incubation with

incubation with gold-labeled anti-AFP–HRP. (g) enhanced CL detection of MB/anti-

Tween-20 were purchased from Sigma–Aldrich. Analytical reagentgrade chemicals and deionized, doubly distilled water were usedthroughout.

2.2. Apparatus

CL measurements were performed with a BPCL Ultra WeakLuminescence analyzer (Biophysics Institute of Chinese Academyof Science, China). UV–vis spectra were carried out on a Cary50 UV–vis–NIR spectrophotometer (Varian). Transmission electronmicroscopy (TEM) image was taken with a JEOL JSM-6700F instru-ment (HITACHI). The optical densities of ELISA were carried out ona DG 3022A enzyme-linked immunoassay analyzer. The CL spectrawere measured on a model FL 4500 spectrofluorometer (HITACHI)with the excitation light source being turned off.

2.3. Preparation of the DC-AuNPs

A colloidal solution of 3 nm-diameter gold nanoparticles wassynthesized by the hydroborate reduction method (Brown et al.,1996), while colloidal solutions of 15-, 30-, 50-, and 60 nm-diametergold nanoparticles were synthesized by the citrate reductionmethod with a slight modification (Frens, 1973). The size andshape of the synthesized gold nanoparticles were characterizedby TEM and UV–vis spectrum. TEM images revealed that the

average diameters of the gold nanoparticles were about 3.0, 15,30, 50, 60 nm, respectively (data not shown except for 30 nm,Fig. S1).

The DC-AuNPs were prepared by the following published pro-cedure with a slight modification. (Ambrosi et al., 2007). The

Bioel

adlawr4acaD

2

a2p(0tgaiitCc

aawaM

0AiFptm

2

p0(Lwt

3

3

eriFs83

X.-Y. Yang et al. / Biosensors and

nti-AFP–HRP (10% more than the minimum amount, which wasetermined using CL method, Fig. S2) was added to the 30 nm col-

oidal Au suspension (Fig. S3). The mixture was stirred for 15 minnd then centrifuged at 45000 × g for 30 min. The clear supernatantas carefully removed, and the precipitated gold conjugates were

esuspended in 1 mL of 0.01 M PBS (pH 7.4) buffer and stored at◦C. Furthermore, the amount of HRP molecules in anti–AFP–HRPdsorbed onto AuNPs was determined by CL method based on theatalysis of HRP to the luminol–H2O2 system. The ratio of AuNPsnd HRP was estimated to be approximately 1/9 (Supplementaryata, Fig. S4).

.4. Preparation of the sandwich-type immunocomplexes

The anti-AFP immobilized magnetic beads were preparedccording to the references with a slight modification (Zhang et al.,005). Briefly, a 60 �L (Fig. S7) of a magnetic microbeads slurry waslaced in a test tube and washed with a 0.1 M imidazol–HCl bufferpH 7.0) three times. A 100 �L of 0.2 M NHS solution and 100 �L of.8 M EDC solution were added to the tube and the resultant mix-ure was incubated at 37 ◦C for 1 h to activate the carboxylic acidroups on the microbeads. The beads were then washed three timesnd 200 �L of anti-AFP antibody solution was added, followed byncubation at 37 ◦C for 24 h. The washed beads were resuspendedn 100 �L of 0.01 M PBS buffer. The number of anti-AFP loaded onhe MB could be quantitatively calculated based on sandwich-typeL enzyme-linked immunoassay as shown in Fig. S8. The averageoverage of an MB was calculated (∼11700 anti-AFP/MB).

200 �L of PBS–0.1% (w/v) bovine serum albumin (BSA) wasdded to the above test tube. The resultant mixture was incubatedt 37 ◦C for 1 h to block remaining active surface of MBs. Afterashing, 200 �L of AFP antigen at different concentrations was

dded and incubated at 37 ◦C for 2 h to form the immunocomplex:B/anti-AFP/AFP (Fig. 1(d)).The washed MB/anti-AFP/AFP was resuspended in the 200 �L

.01 M PBS–0.02% (w/v) Tween-20 and incubated with the anti-FP–HRP (or DC-AuNPs) at 37 ◦C for 2 h, forming the sandwich-type

mmunocomplex: MB/anti-AFP/AFP/anti-AFP–HRP as shown inig. 1(e) (or MB/anti-AFP/AFP/DC-AuNPs as shown in Fig. 1(f)). Therecipitate was then separated and washed with 0.01 M PBS. Finally,he precipitate was resuspended in 200 �L PBS for the following CL

easurements.

.5. CL immunoassay for the determination of AFP

In a typical experiment, 50 �L of sandwich-type immunocom-lexes was transferred into a CL test tube containing 200 �L of.1 mM luminol solution and 100 �L IPP solution containing 0.002%w/v) Tween-20. After vortex mixing, the test tube was placed in theuminescence analyzer, 200 �L of 3.0 mM H2O2 aqueous solutionas then injected and the CL produced was measured for 180 s. The

otal CL intensity was defined as the area under the CL decay curve.

. Results and discussion

.1. Optimization of the enhanced CL reaction conditions

In this work, a novel enhancer, 4-(4′-iodo)phenylphenol wasmployed in the luminol–H2O2–HRP CL system. The optimized CL

eaction conditions and the evaluation of its enhancing capabilitiesn the luminol–H2O2–HRP CL system were described. As shown inig. S9, the favorable conditions for the luminol–H2O2–HRP–IPP CLystem were 0.1 mM luminol solution (0.1 M Tris–HCl buffer, pH.1), 7.0 × 10−5 M IPP solution containing 0.002% (w/v) Tween-20,.0 mM H2O2 aqueous solution.

ectronics 24 (2009) 2707–2711 2709

3.2. Mechanism discussion

As for the kinetics of light emission, Fig. 2(A) shows CL emissionfor luminol–H2O2–HRP–IPP and luminol–H2O2–HRP–PIP (Kurodaet al., 2000) under the optimum conditions. The results showedthat IPP would be a more potent enhancer than PIP, which was themost widely used enhancer, with respect to the sensitivity for thedetermination of HRP. To demonstrate it, a comparison of calibra-tion curves of HRP with PIP and IPP is shown in Fig. 2(B) and (C),respectively. Under the optimized conditions, the dynamic range ofHRP by using PIP as an enhancer was 4.0 × 10−10–5.0 × 10−8 g mL−1

with the detection limit of 100 pg mL−1, whereas the assay sen-sitivity by using IPP as an enhancer could be further increased to9 pg mL−1 with the dynamic range of 5.0 × 10−11–4.0 × 10−8 g mL−1.The results demonstrated that this new enhancer had an advantageover 4-iodophenol in CL determination of HRP. The CL spectrum forluminol–H2O2–HRP in the absence and presence of IPP is acquiredas shown in Fig. 2(D). It clearly indicated that the maximum emis-sions for these two cases was almost the same value of 460 nm, areasonable region for luminol CL, and revealed that the luminophorfor the CL system was still the excited-state 3-aminophthalateanions (3-APA*). Therefore, the addition of IPP did not lead to thegeneration of new luminophor for this CL system. Otherwise, it wasalso suggested that the luminol–H2O2–HRP–IPP CL system did notinvolve an energy transfer process. The enhancement effect of IPPmight depend on its phenol moiety, which accelerates the CL reac-tion like other phenol derivatives such as 4-iodophenol (Easton etal., 1996).

The precise mechanism of the HRP-catalyzed CL oxidation ofluminol in the presence of a p-phenol derivative has not yet been100% proved. However, the main theory involved HRP interme-diates formation and phenoxyl radical generation. The latter wasassumed to affect the CL intensity via (i) the rate of the enzymeturnover and (ii) the electron transfer between radicals and lumi-nol (Easton et al., 1996). Similarly, we suggested that the electronicproperties (i.e. extent of resonance effect) of the substituents atthe 4-position of the phenol played a critical role on radical stabi-lization and therefore on CL intensity enhancement. For instance,substituents at the 4-position of the phenol including an aromaticring provide a resonance stabilization of the phenoxyl radicalsthrough �-delocalization. Also, electron donating groups have asimilar effect on O–H bond dissociation energy, and therefore stabi-lize phenoxyl radicals. The presence of an iodo-substituted phenylgroup at the p-position of the phenol was considered to showa more stabilizing effect for the enhancer radical than a solelyiodo group, which was in agreement with the theoretical calcu-lations (see Supplementary Data), and therefore the superiorityof IPP against PIP on the CL enhancement effect could be easilyexplained.

3.3. CL measurements of AFP

Under the proposed experimental conditions, a new enhancedCL immunoassay based on the HRP-related signal was for thefirst time carried out for the detection of the DC-AuNPs-basedimmunocomplex, and the results were compared with the anal-ysis of the MB/anti-AFP/AFP/anti-AFP–HRP immunocomplex. Thecorresponding standard calibration plots of human AFP for bothimmunocomplexes are shown in Fig. 3 (A) and (B), respectively.Fig. 3(A) illustrated that the use of a novel luminol signal enhancerIPP affected assay performances dramatically, such as detection

limit and the range of the calibration curve. The corresponding cal-ibration plot of relative CL intensity versus the concentration ofhuman AFP was linear over the range from 0.5 to 5.0 ng mL−1, andthe detection limit was 0.5 ng mL−1, which was lower than that inthe ELISA.

2710 X.-Y. Yang et al. / Biosensors and Bioelectronics 24 (2009) 2707–2711

Fig. 2. (A) CL kinetic profiles of luminol–H2O2–HRP–IPP and luminol–H2O2–HRP–PIP. (B) Calibration curve of CL intensity versus the concentration of HRP usingl 1.0 × 1s sing l1 easuro

atotwlte

Fros

uminol–H2O2–HRP–PIP system. Inset: CL intensity versus HRP from 4.0 × 10−10 tourements. (C) Calibration curve of CL intensity versus the concentration of HRP u.0 × 10−9 g mL−1. The error bars represent the relative standard deviation of three mf 5.0 × 10−6 g mL−1 (PBS, pH 7.4) was used for the experiments.

Furthermore, we employed the DC-AuNPs as labels in thebove enhanced CL immunoassay to further improve the detec-ion sensitivity. As shown in Fig. 3(B), a linear relationship wasbtained between the amplified CL intensity and the concentra-ion of the human AFP in the range from 0.008 to 0.3 ng mL−1,

ith a detection limit of 5 pg mL−1. It can be observed that the

imit of detection using DC-AuNPs was about 100 times lowerhan that obtained using the HRP-labeled anti-human AFP. Thenhancement of the sensitivity achieved using DC-AuNPs could

ig. 3. (A) Calibration curves of CL intensity versus the concentration of human AFP recordange from 0.5 to 5.0 ng mL−1. The error bars represent the relative standard deviation of tf human AFP recorded using DC-AuNPs as labels. The inset indicates that the curve is lintandard deviation of three measurements.

0−8 g mL−1. The error bars represent the relative standard deviation of three mea-uminol–H2O2–HRP–IPP system. Inset: CL intensity versus HRP from 5.0 × 10−11 toements (D) CL spectra for luminol–H2O2–HRP (a) with IPP and (b) without PIP. HRP

be attributed to a higher number of HRP capped on the AuNPs(∼9 HRP/one AuNP). The reproducibility of the preparation of theMB/anti-AFP/AFP/anti-AFP–HRP immunocomplex was discussed inthe Supplementary Data, which showed acceptable preparationreproducibility. Table S1 listed the detection limits of AFP using

different nanoparticles-based techniques. Compared with thosesensitive techniques listed in Table S1, our proposed method hasbeen the most sensitive tool for the determination of AFP, to thebest of our knowledge.

ed using anti-AFP–HRP as labels. The inset indicates that the curve is linear over thehree measurements. (B) Calibration curves of CL intensity versus the concentrationear over the range from 0.008 to 0.3 ng mL−1. The error bars represent the relative

Bioel

gshotwecs

3

ccddtbIbat(it

4

ip4saAaAdmbts

X.-Y. Yang et al. / Biosensors and

Comparative studies between IPP and PIP were further investi-ated by MB/anti-AFP/AFP/DC-AuNPs sandwich-type immunoas-ay. The results demonstrated that IPP could provide a muchigher signal than PIP did, also the CL intensity could not bebserved with use of PIP at the concentration range from 0.005o 0.3 ng mL−1 of AFP, while it did in the case of IPP (Fig. S11). Soe could clearly make a conclusion that the new luminol signal

nhancer produces a stronger response of the CL over PIP and itould be used to determine target molecules in a CL immunoas-ay.

.4. Detection of AFP in the clinical serum samples

The feasibility of the immunoassay system for clinical appli-ations was investigated by analyzing several real samples, inomparison with the ELISA method. These serum samples wereiluted to different concentrations with a PBS of pH 7.4. Table S2escribed the correlation between the partial results obtained byhe proposed CL system and ELISA method. The calibration curveetween the proposed method and ELISA is shown in Fig. S12.

t obviously indicated that there was no significant differenceetween the results given by two methods. When the givenmounts of analysts were added into the clinical serum samples,he results showed satisfied recoveries in the range of 96.2–103.0%Table S3). The developed versatile immunoassay may provide annteresting alternative tool for detection protein in clinical labora-ory.

. Conclusions

A versatile and sensitive enhanced sandwich-type CLmmunoassay has been developed in our work. In the pro-osed procedure, a new highly potent signal enhancer,-(4′-iodo)phenylphenol, and a double-codified label con-isting of gold nanoparticles conjugated to an HRP-labelednti-human AFP antibody, have been used to detect humanFP as a model protein. The enormous signal enhancementssociated with the use of an efficient enhancer and DC-uNPs labels provided the basis for the ultrasensitive CL

etection of AFP with a very low detection limit (5 pg mL−1),uch lower than the classical enzyme-linked immunosor-

ent assays. This proposed CL system could be easily extendedo other protein detection schemes as well as in DNA analy-is.

ectronics 24 (2009) 2707–2711 2711

Acknowledgements

The authors thank the National Nature Science Foundation ofChina (no. 20775038), and the National High-tech R&D Program(863 program, no. 2007AA09Z113).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.bios.2008.12.009.

References

Ackerson, C.J., Jadzinsky, P.D., Jensen, G.J., Kornberg, R.D., 2006. J. Am. Chem. Soc. 128,2635–2640.

Ahn, K.C., Lohstroh, P., Gee, S.J., Gee, N.A., Lasley, B., Hammock, B.D., 2007. Anal. Chem.79, 8883–8890.

Ambrosi, A., Castaneda, M.T., Killard, A.J., Smyth, M.R., Alegret, S., Merkoci, A., 2007.Anal. Chem. 79, 5232–5240.

Ao, L.M., Gao, F., Pan, B.F., He, R., Cui, D.X., 2006. Anal. Chem. 78, 1104–1106.Brown, K.R., Fox, A.P., Natan, M.J., 1996. J. Am. Chem. Soc. 118, 1154–1157.Chen, Z.F., Zu, Y.B., 2007. Langmuir 23, 11387–11390.Cui, R.J., Huang, H.P., Yin, Z.Z., Gao, D., Zhu, J.J., 2008. Biosens. Bioelectron. 23,

1666–1673.Easton, P.M., Simmonds, A.C., Rakishev, A., Egorov, A.M., Candeias, L.P., 1996. J. Am.

Chem. Soc. 118, 6619–6624.Elbaz, J., Shlyahovsky, B., Willner, I., 2008. Chem. Commun. 13, 1569–1571.Fan, A.P., Lau, C., Lu, J.Z., 2005. Anal. Chem. 77, 3238–3242.Frens, G., 1973. Nat. Phys. Sci. 241, 20–22.Gómez–Hens, A., Fernández-Romero, J.M., Aguilar–Caballos, M.P., 2008. Trends Anal.

Chem. 27, 394–406.Konry, T., Novoa, A., Shemer-Avni, Y., Hanuka, N., Cosnier, S., Lepellec, A., Marks, R.S.,

2005. Anal. Chem. 77, 1771–1779.Kuroda, N., Shimoda, R., Wada, M., Nakashima, K., 2000. Anal. Chim. Acta 403,

131–136.Li, X.M., Yang, X.Y., Zhang, S.S., 2008. Trends Anal. Chem. 27, 543–553.Pavlov, V., Xiao, Y., Shlyahovsky, B., Willner, I., 2004. J. Am. Chem. Soc. 126,

11768–11769.Previte, M.J.R., Aslan, K., Malyn, S.N., Geddes, C.D., 2006. Anal. Chem. 78, 8020–8027.Rosi, N.L., Giljohann, D.A.C., Thaxton, S., Lytton-Jean, A.K.R., Han, M.S., Mirkin, C.A.,

2006. Science 312, 1027–1030.Rosi, N.L., Mirkin, C.A., 2005. Chem. Rev. 105, 1547–1562.Sapsford, K.E., Berti, L., Medintz, I.L., 2006. Angew. Chem. Int. Ed. 45, 4562–4588.Selvaraju, T., Dasa, J., Hanb, S.W., Yang, H., 2008. Biosens. Bioelectron. 23, 932–938.Stoeva, S.I., Lee, J.S., Smith, J.E., Rosen, S.T., Mirkin, C.A., 2006. J. Am. Chem. Soc. 128,

8378–8379.Tang, D., Yuan, R., Chai, Y., 2008. Anal. Chem. 80, 1582–1588.Wang, Y., Li, D., Ren, W., Liu, Z., Dong, S., Wang, E., 2008. Chem. Commun. 22,

2520–2522.You, C.C., Agasti, S.S., De, M., Knapp, M.J., Rotello, V.M., 2006. J. Am. Chem. Soc. 128,

14612–14618.Zhang, R.Q., Hirakawa, K., Seto, D., Soh, N., Nakano, K., Masadome, T., Nagata, K.,

Sakamoto, K., Imato, T., 2005. Talanta 68, 231–238.Zhou, Y., Zhang, Y., Lau, C., Lu, J., 2006. Anal. Chem. 78, 5920–5924.