Short communication

A nanosensor based on quantum-dot haptens for rapid, on-siteimmunoassay of cyanotoxin in environmental water

Long Feng a,c,n, Anna Zhu b,c, Hongchen Wang a,c, Hanchang Shi b,c

a School of Environment and Natural Resources, Renmin University of China, Haidian, Beijing 100872, Chinab Research Institute of Chemical Defense, Beijing 102205, Chinac School of Environment, Tsinghua University, Haidian, Beijing 100084, China

a r t i c l e i n f o

Article history:Received 1 July 2013Received in revised form1 September 2013Accepted 9 September 2013Available online 25 September 2013

Keywords:NanosensorQuantum dotCyanotoxinFluorescence resonance energy transferOptofluidic

a b s t r a c t

A nanoprobe based on quantum-dot (QD) haptens was synthesized by conjugating carboxyl quantumdots with aminoethyl-microcystin (MC)–leucine–arginine (LR). A two-alkyl group was introduced tosupply a spacer between the QD nanoprobe and anti-MC–LR antibody to reduce the steric hindrances ofimmunoreaction. The sensor system based on a portable optofluidic platform exhibited a liner range of0.10–4.0 mg/L for MC–LR with a detection limit of 0.03 mg/L. The proposed sensor has potentialapplication in the rapid, on-site detection of MC–LR in real water samples.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

The frequency of occurrence of cyanobacterial blooms hasbecome a global concern because of the resulting liberation ofcyanobacterial toxins and the deterioration of water quality(Ye et al., 2009). Microcystins (MCs), a group of cyclic heptapeptidehepatotoxins consisting of a seven-amino-acid peptide ring, havebeen proven to be highly toxic to vertebrates (Campo and Ouahid,2010). Many reported cases of animal poisoning, human diseases,and even death are attributed to MC exposure through drinkingand surface waters (Amé et al., 2010). MC–leucine–arginine (MC–LR; MW¼995.2) is one of the most frequently detected and toxiccyanotoxins (Amé et al., 2010; Campo and Ouahid, 2010; Dennisand Bao, 2008; Herranz et al., 2012). As a result, the World HealthOrganization (WHO) has set a guideline value of 1.0 mg/L for MC–LR in drinking water to minimize the risk to the public (WorldHealth Organization, 2003). A number of detection methods,including high-performance liquid chromatography, invertebratebioassays, and protein phosphatase inhibition assays, are available.However, these techniques generally require long analysis times andmay be strongly affected by matrix effects (Herranz et al., 2012).Recently, immunosensors based on mono-/polyclonal antibodies have

been developed as screening techniques because of their specificity,sensitivity, ease of use and rapidity. However, the sensitivity of animmunoassay strongly depends on the affinity of specific antibodiesand the sensitivity of the detection method.

Fluorescence detection techniques have the advantages ofversatility, high sensitivity, and the simplicity of signal detection.In particular, fluorescence resonance energy transfer (FRET)-basednanoscale biosensors are widely used detection systems that rely onthe fluorescence response (Jaiswal et al., 2003). In this technique,one of the optimum donors is quantum dots (QDs) because of theirunique optical properties, including high quantum yield, photo-stability, narrow emission spectrum, and broad absorption. Fluor-escent dyes are generally used as acceptors. QD-FRET basedbiosensors have been widely used in immunoassays, clinical/diag-nostic assays, and biomolecular binding assays. Morevoer, QDs-antibody (Ab) conjugates are the most developed and most widelyused detection bioprobes for QD integration in bioanalyses (Dennisand Bao, 2008). For example, a QD/Ab probe was used for theimmunological recognition of MC–LR as well as for electrochemicaltransduction (Yu et al., 2010). QD responses were amplified andconverted to electrochemical signals by measuring the release ofcadmium ions from the QDs. However, this method is complex andrequires a long analysis time (45 h). More importantly, when QDsare conjugated with antibodies by covalent or noncovalent methods,antibody activity loss occurs. Moreover, the number of antibodiesper QD and their orientation and position relative to the QDs aredifficult to control (Medintz et al., 2004). In addition, these bioprobes

Contents lists available at ScienceDirect

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

Biosensors and Bioelectronics

0956-5663/$ - see front matter & 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.bios.2013.09.018

n Corresponding author at: School of Environment and Natural Resources,Renmin University of China, Beijing 100872, China. Tel.: þ86 10 62796935;fax: þ86 10 62771472.

E-mail addresses: longf04@ruc.edu.cn, longf04@mails.thu.edu.cn (L. Feng).

Biosensors and Bioelectronics 53 (2014) 1–4

have to be freshly prepared because antibodies require cryopreser-vation but QDs cannot be frozen. Therefore, challenges remain forthe fabrication of new types of QD-based nanosensors for MC–LRdetection with high sensitivity and high stability.

In the present study, a hapten-coupled QD nanoprobe protocolwas developed for the rapid and sensitive detection of MC–LR inreal water samples. QD-hapten nanoprobes were prepared byconjugating carboxyl QDs with aminoethyl-MC–LR (H2N-etMC–LR),which is responsible for the immunological recognition of the anti-MC–LR antibody and for optical transduction. The nanosensor'ssensitivity, specificity, stability, and resistance to background inter-ferents were determined.

2. Experiments

2.1. Materials and chemicals

Bovine serum albumin (BSA), ovalbumin (OVA), and 1-ethyl-3(-3dimethylamino-propyl)carbodiimide hydrochloride (EDC) werepurchased from Sigma-Aldrich (Steinheim, Germany). MC–LR,MC–YR, MC–RR, MC–LW, and MC–LF were obtained from Alexis(Läufelfingen, Switzerland). All other reagents, unless otherwisespecified, were supplied by the Beijing Chemical Agents; theywere also analytical grade and used without further purification.Distilled deionized water was used throughout the investigation.About 1 mg/mL MC–LR stock solutions were prepared in 0.01 mol/L phosphate-buffered saline (PBS) and stored at 4 1C. Standardconcentrations of the analyte were prepared from the stocksolution by serial dilutions in 0.01 mol/L PBS.

Monoclonal anti-MC–LR-MAb (MC-LR-MAb, reference no. 8C10)was produced by our research group and labeled by Cy5.5 aspreviously described (Long et al., 2008).The estimated number ofCy5.5 molecules attached to each antibody determined was2.4 according to producer's guide.

2.2. Preparation of QD nanoprobe

First, aminoethyl-MC–LR was synthesized by the reaction of2-mercaptoethylamine with the seventh amino acid residue (dehy-droAla) of MC–LR (Fig. 1A). This residue introduces a primary aminogroup to MC–LR and allows its easy conjugation to the QDs. Thereason for choosing the seventh amino acid residue is its location,which is the farthest from both of the variable amino acid residuesand from 3-amino-9-methoxy-10-phenyl-2,6,8-trimethyl-deca-4(E)6(E)-dienoic acid (Adda). The latter plays an important role in toxicityand is recognized by most of the currently available mono-/poly-clonal antibodies. A two-alkyl group was added to supply a suitableand biologically compatible spacer between the QD nanoprobe andthe anti-MC–LR antibody to reduce steric hindrances for immunor-eaction (Moorhead et al., 1994). QD-hapten nanoprobes wereprepared by conjugating carboxyl QDs with H2N-etMC–LR (Fig. 1B),which is responsible for the immunological recognition of the anti-MC–LR antibody and for optical transduction.

2.3. Sensing mechanism

Fig. 1C shows the competitive immunoassay mechanism for theQD-FRET-based MC–LR detection. During one detection cycle, differ-ent concentrations of MC–LR solutions, the fluorescence-labeledantibodies, and a fixed concentration of the QD nanoprobe weremixed and incubated for 5 min. During incubation, the free MC–LR insolution and the antigens immobilized onto the QD nanoprobe surfacesimultaneously and competitively bound with the fluorescence-labeled antibodies. Once the equilibrium state of this competitiveantigen-antibody reaction was reached, the mixture was introduced

into the optofluidic channel for detection. The fluorescence intensitywas inversely proportional to the MC–LR concentration in thesamples.

2.4. Instruments: Portable optofluidic platform

A portable optofluidic platformwas developed to achieve rapid,on-site MC–LR detection (Fig. S3). In this system, the 405 nm,20 mW pulse diode laser with a pigtail was used as the excitedlight source, whereas a single multifiber optic coupler was used forthe transmission of the excitation light and the collection andtransmission of fluorescence. The emission signal of the QDnanoprobe was collected by the multimode optical fiber, filteredby a bandpass filter, and then detected by photodiodes through adigital lock-in amplifier that was interfaced to a minicomputer.

3. Results and discussions

3.1. Characterization of QD-hapten nanoprobe

The QD-hapten nanoprobe was prepared by conjugatingcarboxyl QDs with H2N-etMC–LR conjugate, which is regarded as

MC–LR, , QD, , Cy5.5-labeled antibody.

EDC/NHS

405nm excitation

625 nm emission

Cy5.5 emission

H2N-etMC-LR

Fig. 1. (A) Synthesis of aminoethyl-MC–LR using the reaction of 2-mercaptoethy-lamine with the seventh amino acid residue of MC–LR; (B) Preparation of the QDnanoprobe by conjugating amine of BSA to carboxyl-coated Qdot 605 through theEDC/sulfo-NHS chemistry; and (C) Sensing mechanism of MC–LR detection basedon indirect competitive immunoassay.

L. Feng et al. / Biosensors and Bioelectronics 53 (2014) 1–42

the biorecognition unit of fluorescence-labeled anti-MC–LR anti-body as well as for optical transduction. TEM images show that theQD-MC–LR nanoprobe was uniform in size. Moreover, only theinorganic particles were directly visualized at approximately10 nm�5 nm (Fig. S2). When the anti-MC–LR antibody was addedto the QD nanoprobe solution, an additional halo around the QDsappeared after negative staining of the antibody bound to the QDsurface (Fig. S2). This result is consistent with the conjugation ofproteins to the particles and indicates the successful covalentimmobilization of MC–LR on the QD surface using the EDC/NHScoupling strategy.

3.2. FRET efficiency

The FRET efficiency (E) is a key parameter of FRET-basedsensors and strongly depends on the spectral overlap of the donorand acceptor absorption spectra, as well as on the relativeorientation of the donor emission dipole moment, the separationdistance between donors and acceptors, and the acceptor absorp-tion dipole moment (Medintz et al., 2003a). In this study, Cy5.5-labeled anti-MC–LR antibodies were used as biorecognitionmolecules of MC–LR as well as acceptor. Fig. S4 shows that theenergy transfer efficiency increased with increased fluorescence-labeled antibody concentration. When the ratio of antibodies tothe QD nanobioprobe was 420, E was estimated at 452.3%. Thishigh FRET may be due to several factors. First, by using the Försterdipole-dipole interaction formalism, the energy transfer efficiencyE for a single donor FRET with multiple identical acceptors can beexpressed as (Medintz et al., 2003b),

E¼ nR60

r6þnR60

ð1Þ

where n is the average number of acceptor molecules interactingwith one single donor, and R0 is the Förster radius of the FRET pairin a single donor-single acceptor situation, in which E¼50% and r isthe apparent donor–acceptor distance. In the QD-hapten sensingsystem, the high number of hapten molecules attached to each QDnanoprobe (the estimated number of MC–LR is approximately 12)enable several fluorescence-labeled antibodies to simultaneouslybind to the QD surface. Several acceptor dyes strongly interact witha single QD donor, which improves the FRET efficiency according toEq. (1). This result was confirmed in the FRET data collected fromexperiments that used increasing ratios of dye-labeled antibody perQD. In these data, the measured efficiencies increased as expectedwhen the acceptor-to-donor ratios increased in each conjugate (Fig.S4). Second, the Förster energy transfer is generally most efficientwhen the distance between donors and acceptors is in the 20–60 Årange. By assuming that the H2N-etMC–LR and the antibodies aresequentially and perpendicularly attached to the QD surface and byaccounting for the increased QD size (9–16 nm) due to the antibody(6–10 nm), the center-to-center separation distances between theQD donor and the antibodies in our system exceeded 6 nm.However, the QD nanoprobe is unsymmetrical (Fig. S2) and theshort diameter is approximately 5 nm. Meanwhile, the Cy5.5dye-to-antibody ratio is approximately 2.4, and these fluorescencemolecules easily bound to the antibodies in various randomorientations. Therefore, some dyes are in close proximity to theQD surface. In addition, a two-alkyl group spacer was introducedbetween the QD-haptens nanoprobe and the anti-MC–LR antibody,which may have prevented steric hindrance and improved theantigen-antibody binding reaction.

3.3. Dose–response curves

Fig. 2A shows the temporal fluorescence signal during a typical testcycle for MC–LR detection using the portable optofluidic platform,

including the introduction of a mixture of the fluorescence-labeledantibody, QD nanoprobe, and sample as well as a washing step. Thedecreased fluorescence emission intensity is linear with the MC–LRconcentration in the 0.10 mg/L to 4.0 mg/L range (Fig. 2B), with arelative standard deviation as a percentage (%RSD) ranging from1.73% to 6.82%. Based on the dose–response curve and on a signal-to-noise ratio of 7.5, the detection limit was determined as approxi-mately 0.03 mg/L using the three-times standard deviation rule (WorldHealth Organization 2003). The detection limit is sufficient to monitorMC–LR changes from basal levels and is lower than the WHOstandards. The detection limit obtained is also comparable to thoseof SPR detection (0.07 mg/L) (Herranz et al., 2010), electrochemicalimmunoassay (0.099 mg/L) (Yu et al., 2010), and bead-based competi-tive fluorescence immunoassay (0.03 mg/L) (Yu et al., 2011). However,compared with these sensors, the proposed nanosensor is significantlysimpler and faster. These attributes result from the dynamic bindinginteraction between the antibodies and the QD nanoprobe. Moreover,without any loss in sensitivity, the dimensions of the optofluidicdevice ensures a total volume (approximately 35 mL) that is lower thanthat of conventional ELISA, which requires at least 100 mL in micro-wells. This small volume requirement significantly reduces the assaycost, given that some reagents are highly expensive.

0 100 200 300 400 500 600 7000

10

20

30

40

50

60

70

80

8.0 µg/L MC-LR

4.0 µg/L MC-LR

2.0 µg/L MC-LR

1.0 µg/L MC-LR

1 ×PBS

0.5 µg/L MC-LR

0.1 µg/L MC-LR

0 µg/L MC-LR

Fluo

resc

ence

Sig

nal (

a.u.

)

Time (s)

0.0

10

20

30

40

50

60

70

0.1 110

20

30

40

50

60

Fluo

resc

ence

Sig

nal (

a.u.

)

Log (MC-LR concentration, µg/L)

10.01.0

Fluo

resc

ence

sig

nal (

a.u.

)

MC-LR concentration (µg/L)0.1

Fig. 2. Competitive immunoassay of MC–LR. (A) Signal traces observed as the mixtureof 5 nM fluorescence labeled antibody, 1.5 nMQD immunoprobe andMC–LR solutionsof various concentrations flow over the optofluidic channel. (B) Dose–response curveof MC–LR. Inset indicates the dynamic range, showing the linear relationship betweenthe MC–LR concentration and fluorescence intensity.

L. Feng et al. / Biosensors and Bioelectronics 53 (2014) 1–4 3

3.4. Specificity of nanosensor

The specificity of the QD-FRET-based immunoassay was evaluatedusing a wide range of compounds structurally related to MC–LR.As previously shown using ic-ELISA (Sheng et al., 2007), MC8C10antibody exhibits high cross-reactivity for MC-LR-related compounds(e.g., MC–YR, MC–RR, MC–LW, MC–LF, and Nodularin). Moreover, thecross-reactivities of the different interferents are considerably smalland fall below 12% (Table S1). This selectivity must be due to theantibody specificity, which is a vital factor for rapid and on-sitedetermination of water samples without separation.

3.5. Spiked water samples assay

The matrix susceptibility caused by component variability inseveral environmental water samples is a source of potentialinterference and should thus be considered to verify the accuracyand applicability of the QD-FRET-based nanosensor. In the presentstudy, the percentage recovery (% recovery) of three water sampleswas estimated by spiking with low (0.3 μg/L), moderate (1.0 μg/L),and high (3.0 μg/L) concentrations of standard MC-LR. These con-centrations were designed to fall and be evenly spaced within thedynamic range of the assay (Table S2). All samples were measuredunspiked to ensure noncontamination (o0.03 μg/L). The percentrecoveries of the MC–LR added to the water samples at differentconcentrations ranged from 80% to 115%. All coefficients of variation(CVs) were acceptable and did not exceed 9% (n¼3). These dataconfirm that the proposed QD nanoprobe is considerably resistantto the nonspecific binding of dissolved components in water andare thus applicable to the matrix-independent and direct measure-ment of MC–LR.

4. Conclusion

A rapid and simple analytical system set up for cyanotoxindetection was successfully achieved by the effective integration ofan optofluidic platform with a QD-FRET-based competitive immu-noassay. Compared with traditional techniques, this system providesseveral advantages. First, by using assembled functional haptensthat are conjugated to QD surfaces as recognition elements, theQD-hapten nanoprobe is more stable in complex environmentalsamples. Moreover, the binding properties of immobilized biomole-cules are not compromised when the probes are prepared byimmobilizing haptens onto the QD surface. Second, a two-alkylgroup was introduced to supply a spacer between the QDs and thehaptens to prevent steric hindrance and maintain the high activity ofthe QD nanoprobe for its antibody. Third, the FRET efficiency ishigher because of the increased number of acceptor dyes bound toone QD surface, which results in the high sensitivity of the QD-FRET

assay. Fourth, proposed QD nanoprobe can be stored for over threemonths, with significantly low deterioration (o15%) in performance.More importantly, through the integration of a portable optofluidicplatform, rapid, on-site cyanotoxin detection with enhanced sensi-tivity, reduced analysis time, lower reagent volumes, and simplifiedmanipulation can be achieved. This will pave the way for a vitalroutine analysis that satisfies the high demand for safe drinkingwater sources and thus ensure human health.

Acknowdgement

This research was financially supported by the National NaturalScience Foundation of China (21077063, 21277173), the NationalInstrument Major Project of China (2012YQ3011105), and the BasicResearch funds in Renmin University of China from the CentralGovernment (13XNLJ01).

Appendix A. Supplementary material

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.bios.2013.09.018.

References

Amé, M.V., Galanti, L.N., Menone, M.L., Gerpe, M.S., Moreno, V.J., Wunderlin, D.A.,2010. Harmful Algae 9, 66–73.

Campo, F.F., Ouahid, Y., 2010. Environmental Pollution 158, 2906–2914.Dennis, A.M., Bao, G., 2008. Nano Letters 8, 1439–1445.Herranz, S., Marazuela, M.D., Moreno-Bondi, M.C., 2012. Biosensors and Bioelec-

tronics 33, 50–55.Herranz, S., Bocková, M., Marazuela, M.D., Homola, J., Moreno-Bondi, M.C., 2010.

Analytical and Bioanalytical Chemistry 398, 2625–2634.Jaiswal, J.K., Mattoussi, H., Mauro, J.M., Simon, S.M., 2003. Nature Biotechnology 21,

47–51.Long, F., He, M., Shi, H.C., Zhu, A.N., 2008. Biosensors and Bioelectronics 23,

952–958.Medintz, I.L., Konnert, J.H., Clapp, A.R.I., Twigg, M.E., Mattousssi, H., Mauro, J.M.,

Deschamps, J.R., 2004. Proceedings of the National Academy of Sciences 121,9612–9617.

Medintz, I.L., Clapp, A.R., Mattoussi, H., Goldman, E.R., Fisher, B., Mauro, J.M., 2003a.Nature Materials 2, 630–638.

Medintz, I.L., Goldman, E.R., Lassman, M.E., Mauro, J.M.A., 2003b. BioconjugateChemistry 14, 909–918.

Moorhead, G., Mackintosh, R.W., Morrice, N., Gallagher, T., Mackintosh, C., 1994.FEBS Letters 356, 46–50.

Sheng, J.W., He, M., Shi, H.C., 2007. Analytica Chimica Acta 603, 111–118.World Health Organization, 2003. Guidelines for Drinking Water Quality, third ed.

World Health Organization, Geneva, Switzerland.Ye, W., Liu, X., Tan, J., Li, D., Yang, H., 2009. Harmful Algae 8, 637–644.Yu, H.W., Kim, L.S., Nguyen, G.H., Kim, I.S., 2010. Analytical and Bioanalytical

Chemistry 398, 2625–2634.Yu, H.W., Jang, A., Kim, L.H., Kim, S.J., Kim, I.S., 2011. Environmental Science and

Technology 45, 7804–7811.

L. Feng et al. / Biosensors and Bioelectronics 53 (2014) 1–44