A Quantum Dot-Based Immunoassay for Screening of Tetracyclinesin Bovine MuscleJenifer García-Fernandez, Laura Trapiella-Alfonso, Jose M. Costa-Fernandez, Rosario Pereiro,and Alfredo Sanz-Medel*

Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, 33006 Oviedo, Spain

*S Supporting Information

ABSTRACT: A reliable and robust direct screening methodology based on a quantum dot (QD) fluorescent immunoassay hasbeen developed to detect trace levels of different antibiotic species from the family of the tetracyclines (e.g., oxytetracycline,tetracycline, chlortetracycline, and doxycycline) in contaminated bovine muscle tissues. First, the synthesis and characterizationof a new immunoprobe (oxytetracycline-bovine serum albumin-QD) has been carried out for its further application in thedevelopment of a competitive fluorescent QD-immunoassay. The developed fluoroimmunoassay provides sensitive and binary“yes/no” responses being appropriate for the screening of this family of antibiotics above or below a preset concentrationthreshold. The detection limit achieved with this strategy, 1 μg/L in aqueous media and 10 μg/kg in bovine muscle samples, is10-fold lower than the maximum level concentration allowed by International Legislation in muscle tissue, enabling suitable andefficient screening of the antibiotics.

KEYWORDS: quantum dots, immunoassay, screening, tetracyclines, fluorescence

■ INTRODUCTION

Tetracyclines (TCCs) are antibiotics with a broad antibacterialspectrum and bacteriostatic activity. Moreover, their low cost,availability, easy administration, and efficiency for the treatmentof bacterial diseases make them the most frequently used inhuman health, animal husbandry, and some agricultural areas.1,2

Oxytetracycline (Oxy), tetracycline (TC), chlortetracycline(Chlor), and doxycycline (Doxy) are the most representativecompounds (see Figure 1) among all the members of the TCCfamily. However, they can be considered a risk to human healthnot only due to their increasing abuse for therapeutic use butalso because they can be employed as fraudulent promoters ofaccelerated growth in food-producing animals. Residues ofthese TCCs remain even in cooked animal tissues and could

have harmful effects on human consumers, including allergicreactions, liver damage, yellowing of teeth, and gastrointestinaldisturbance. In addition, long-term exposure to such speciescan lead to an increased drug-resistance of microbial strains.3

Therefore, in order to preserve safety and food control,different worldwide governmental organizations (e.g., FAO/WHO,4 US FDA,5 and Japanese Ministry of Health, Welfare,and Labour6) have established maximum residue limits (MRLs)allowed to be present in different foodstuff. Particularly, one ofthe most complete and restrictive legislations has beenelaborated by the European Commission, which has setMRLs for TCCs in bovine muscle (100 μg/kg), kidney (600μg/kg), and liver (300 μg/kg).7

Several methods of TCC residues analyses in food have beendeveloped. Some of them are based on microbiological assaysas they are easy to perform. Unfortunately such methods aretime-consuming (usually require 2−3 days for microbe growth)and lack the needed specificity and sensitivity.8−11 Capillaryelectrophoresis has been evaluated also as a possiblealternative,12 but it presents important drawbacks includinglaborious pretreatment steps and difficulty to determine lowconcentrations of those antibiotics. Due to their high sensitivityand specificity, chromatographic techniques have been alsoevaluated13 (especially HPLC-based methods). However, suchmethodologies require high-cost instrumentation and samplesusually need extensive and time-consuming processing(extraction and purification) before the analysis.14 Therefore,the development of sensitive, rapid, selective, and cost-effective

Received: November 12, 2013Revised: January 17, 2014Accepted: January 20, 2014Published: January 20, 2014

Figure 1. Chemical structure of TCCs.

Article

pubs.acs.org/JAFC

© 2014 American Chemical Society 1733 dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−1740

methods for TCC residue detection and screening in routineassays is still a research demand.Immunochemical assays could be an alternative to meet all

those requirements, as detection of low concentrations of TCCresidues in many samples in a short time and in some casesavoiding long and tedious sample pretreatment steps may bepossible. In fact, several approaches of competitive ELISAformat assays have been already developed for the analysis ofTCCs, being applied in most of the cases to milksamples.1,15−17 Assuming that a majority of samples analyzedmay not be contaminated by any of the TCCs, samplescreening methods providing a reliable “yes/no” binaryresponse related to a preset concentration threshold are ofincreasing interest in food control. Screening systems givemainly qualitative or semiquantitative information as they arefocused on identifying and selecting a group of samples thatcontain one or more analytes above a preset concentration orthreshold level.18,19 In this context, the high selectivity andsensitivity typically provided by immunological methods areparticularly suitable for the development of such screeningstrategies.The use of quantum dots (QDs) as fluorescent labels in

bioassays has been increasing due to their advantageousfeatures, including excellent photostability and chemicalstability, long fluorescence lifetimes, intense and narrowemission spectra, and size- and composition-tunable proper-ties.20 Furthermore, the possibility of attaching QDs tobiomolecules keeping their functionality has allowed develop-ment of different formats of immunoassays such as micro-arrays21 or Western blotting22 as well as luminescent sensorsfor the analysis of health-risk environmental pollutants (e.g.,based on molecular imprinted polymers23). However, thedevelopment of fluorescent immunoassay can report anadditional benefit over the ELISA immunoassays, that is, savingtime and costs in avoiding the final step of enzymatic reaction.In this context, a QD-based immunoassay for direct

screening of four TCCs in bovine muscle tissue is heredescribed. The fluorescent probe consists of QDs bioconju-gated to oxytetracycline−BSA conjugate (Oxy−BSA−QD),prepared following the carbodiimide chemistry. Due to thesmall size of the TCCs (haptens with only one site for bindingto the antibody), the immunoassay design is restricted to acompetitive format. So, competition for antibody binding sitesbetween free antigen of the sample (unlabeled TCC) and thebioconjugates (Oxy−BSA−QD) is established.

■ MATERIALS AND METHODSReagents, materials, and instrumentation used through the work areavailable in the Supporting Information.Characterization of Oxy−BSA conjugates. The synthesis and

purification of Oxy−BSA conjugates are based on a commercial kit,and a detailed description about the procedure is collected in theSupporting Information. Once Oxy−BSA conjugates are obtained, anexhaustive characterization step is mandatory. A Bradford test24 wasused to determine the BSA concentration in the Oxy−BSA conjugate.The test correlates the ratio between absorbance of the samplesmeasured at 465 and 595 nm (595/465) with the BSA (protein)concentration. According to a linear regression, BSA concentration inthe conjugate was then calculated.Molecular weight and stoichiometry of the conjugate were evaluated

by MALDI-MS. A solution of 5 mg/mL of sinapinic acid in 30% ofacetonitrile, containing 0.1% of trifluoroacetic acid, was prepared. Anamount of 1−10 pmol of protein in the sample was needed, being 2.42pmol in this case. As the sample has to be ideally clean of salts and the

Oxy−BSA conjugate was stored in a 28 mM Na3PO4, 300 mM NaCl,and 33 mM sorbitol buffer (PBS), a previous cleanup step was carriedout (sample desalting by Zip Tips). A solution of 1 mg/mL of BSAstandard in water was used as reference. The MALDI plate was loadedwith standards, matrix, and/or samples, and when the spots had beendried (at least after 30 min), the MALDI-MS analysis was performed.

Finally, following a strategy previously developed in our lab25 basedon the combination of the information provided by the Bradford testand the MALDI analysis, the final concentration of Oxy in theconjugate can be calculated. Such information is necessary to furtheroptimize the bioconjugation of Oxy−BSA to QDs.

Synthesis and Characterization of the Immunoprobe: Oxy−BSA−QD Bioconjugate. The fluorescent labels used for labeling theOxy−BSA conjugate giving rise to the tracer of the immunoassay arethe CdSe/ZnS QDs. A brief description about their synthesis andsolubilization are presented in the Supporting Information. Thebioconjugation strategy used to synthesize the fluorescent tracer wasbased on a coupling procedure between −COOH groups from outerpolymeric layers of QDs and −NH2 groups from the BSA of the Oxy−BSA conjugate owing to the catalytic activity of EDC.26 After stirringof the mixture of reagents during 2 h at room temperature, apurification step is needed to separate the bioconjugate from the restof undesirable products (e.g., excess of reagents). The purification stepis based on differences in molecular weights between the Oxy−BSAconjugate (71.2 kDa) and the Oxy−BSA−QD bioconjugate (>150kDa). So, it can be carried out by ultrafiltration (UHF) using a 100kDa membrane filter. Experimental conditions used for UHF were3000 g, 5 min, making 3 washes with 100 mM PBS buffer at pH 7.4 at4 °C.

The characterization of this bioconjugate consists on the evaluationof the fluorescent properties of QDs after the reaction and the stabilityin time of the tracer. These studies were carried out using fluorescentmeasurements.

Competitive Immunoassay Protocol. In a conventionalimmunoassay analysis the following experimental procedure wasperformed. First, the microtiter plate is coated with 100 μL/well of a 1μg/mL antibody solution and is incubated at 37 °C for 2 h. In this stepthe antibody is immobilized on the microtiter plate wells by absorptionover the surface of the support of polystyrene. Antibody solution isthen removed and a blocking step is performed by adding 200 μL/wellof 3% casein solution in water, in order to minimize further unspecificbindings. The microtiter plate is then incubated overnight at 4 °C. In afurther step, the microtiter plate is washed three times using a washingsolution (200 μL/well of 10 mM PBS pH 7.4 + 0.05% Tween20) toremove the excess of reagents. Then, the competition is established bythe addition of a mixture of the standard (or the sample) and a knownamount of the tracer (Oxy−BSA−QD) in a total volume of 150 μL/well. The reaction is incubated at 37 °C for 2 h. After thecorresponding washing step, the fluorescence emission of the QDsfrom the Oxy−BSA−QD recognized by the antibody is measured.

Cross Reactivity Calculation. Cross reactivity (CR) is generallydefined as the necessary amount of mass or concentration ofinterference able to produce an equal signal as when the analyte isassayed to provoke a signal inhibition of 50%.27 Therefore, in this workCR rates, in terms of percentage (%), were calculated according to theexpression (eq 1)

= ×CR [(IC (analyte)/IC (interference))] 10050 50 (1)

where IC50 is the necessary concentration of species (analyte orinterference) to induce a signal inhibition of 50%.

Screening Theory and Determination of the UnreliabilityRegion. A screening system provides mainly qualitative or semi-quantitative information. In this type of analysis it is not possible toexpress uncertainty in the same way as in quantitative analysis. For thatreason it is necessary to define new concepts that characterizescreening methods. Thus, “reliability” of a screening test is defined asthe “proportion of correct yes/no responses from a large number oftests carried out on ‘n’ aliquots of the same sample”. On the otherhand, the term “uncertainty” should be replaced by “unreliability”, thatis the concentration range of the analyte around the limit where false

Journal of Agricultural and Food Chemistry Article

dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−17401734

positives, false negatives, and dubious results are obtained.18,28 Inorder to determine the unreliability region (UR) in the screeningprocedures, the cutoff level must be defined. The UR is finallycalculated from the probability−concentration curve established forthe screening test. In this work the limits defining the UR weredetermined as the levels of false positives (probability that a negativesample is classified as positive, also called α error) and false negatives(probability that a positive sample is classified as negative, also called βerror).29

In this work, to obtain the probability vs concentration graph, 20assays (random samples containing variable concentrations of theindividual TCC but ensuring the same total antibiotic concentration)were carried out for each point of the curve, and the relativeproportion (in terms of percentage) of “yes/no” answers wasdetermined. In all cases, the binary “yes/no” response was obtainedfrom the analytical signal (fluorescent intensity measured at the finalstep of the immunoassay) and was classified as “negative” or “positive”after its comparison with the fluorescence intensity signal registeredfor the corresponding cutoff concentration (I0).

Analysis of Bovine Muscle Samples. Bovine muscle samplesfrom noncontaminated cows were extracted following the Oka et al.procedure,30 blending successively the meat samples (5 g) with threesuccessive portions of 20, 20, and 10 mL of 0.1 M Na2EDTA-McIlvaine buffer pH 4.0, centrifuging the mixture (at 850g for 5 min),and collecting and combining the supernatants by decanting. As nodetectable levels of TCC residues were found in the meat samples, 5 gof chopped meat was spiked with the appropriate amount of TCCmixtures (see Table S1 in the Supporting Information for details of theconcentrations of each TCC) and carefully stirred manually during 5min. The samples were kept at 4 °C for 1 h. Afterward, the McIlvainebuffer was added and the Oka et al. procedure30 was then followed.However, additional ultrafiltration steps were needed to remove thecolor that remains in the extract, which can interfere in thefluorescence measurements. For that reason, ultrafiltration withAmicon membrane filter of 10 kDa followed by ultrafiltration withAmicon membrane filter of 3 kDa (centrifuged at 12000g for 15 min inboth cases) was performed, collecting the filtered solution. Two-hundred microliter aliquots of such solution were then used in the

Figure 2. Work-flow diagram of the TCC screening in bovine muscle samples by fluorescent competitive QD-based immunoassay.

Figure 3. MALDI-MS characterization of the synthesized Oxy−BSA conjugate. The mass spectra were obtained using sinapic acid as matrix. BSAstandard (black line) and Oxy−BSA conjugate (gray line) mass spectra are represented, showing clearly the mass shift between both (ΔM).

Journal of Agricultural and Food Chemistry Article

dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−17401735

immunoassays. A work-flow diagram of the complete process for theanalysis of TCCs in bovine muscle samples is shown in Figure 2.

■ RESULTS AND DISCUSSIONCharacterization of the Oxy−BSA Conjugate. As it is

not possible to purchase commercial Oxy−BSA conjugate, it isnecessary to synthesize it as described in Materials andMethods. A complete characterization of synthesized Oxy−BSA conjugate was performed following the strategy developedby Trapiella-Alfonso et al.25 The molecular weight of theconjugate, the stoichiometry, and the concentration of Oxy andBSA in the conjugate are obtained.Based on results obtained from a Bradford assay, BSA

concentration was calculated to 335 ± 9 μg/mL. Molecularweight and stoichiometry were investigated by MALDI-MS.Results from MALDI experiments were used to calculate themolecular weight of the conjugate, being 71223 ± 99 Da (n =15). Comparing the mass spectrum of BSA standard with theOxy−BSA spectrum (see Figure 3), a mass shift (ΔM) isnoticed, and this can only be attributed to the incorporation ofthe Oxy in the BSA structure.If Oxy molecular weight is considered as 460.434 Da,

dividing the ΔM observed by the Oxy molecular weight, thenumber of Oxy per BSA can be estimated. For this case, thefound value was close to 10, leading to a ratio of Oxy:BSA of10:1. Finally, applying eq 2 the final concentration of Oxy inthe conjugate can be estimated.25

=C C10 (MW /MW )Oxy BSA Oxy BSA (2)

where COxy is the concentration of oxytetracycline in theconjugate, 10 is the Oxy:BSA ratio, CBSA is the concentration ofBSA in the conjugate obtained by the Bradford test, and MWOxyand MWBSA are the molecular weights of oxytetracycline andBSA respectively.Synthesis and Characterization of Oxy−BSA−QD

Bioconjugate. The molar ratios used to carry out thebioconjugation reaction were optimized by varying the ratiosQDs:Oxy−BSA from 1:1 to 3:1 (always with an excess of EDCin a ratio of 1500 mol per mol of biomolecule to ensure thatbioconjugation is completed). Optimum molar ratio valueswere observed for 2:1:1500 (QDs:Oxy−BSA:EDC) attendingto the highest fluorescence intensity signals measured,indicating a high efficiency on the labeling of the Oxy−BSAconjugate with the nanoparticles. According to Trapiella-Alfonso et al.31 the concentration of Oxy and BSA in thetracer and the bioconjugation yield can be calculated.In order to characterize the purified bioconjugate,

fluorescence measurements were performed to check thatQD emission is not interfered with the bioconjugation reaction.As can be seen in Figure 4, the Oxy−BSA−QD bioconjugatedoes not modify significantly QD emission wavelength orfluorescence intensity; therefore, it is possible to develop afluorescent immunoassay with this tracer.Finally, the stability of the immunoprobe (utility valid period

of the tracer for immunoassays) was evaluated attending to thestability of the fluorescence emission and the ability of theantibody to recognize the tracer. It was observed that 10 dayswas established as the maximum period where the Oxy−BSA−QD is stable and suitable for immunosensing, being stored at 4°C. In the Supporting Information, Figure S1 shows data froma stability study. It is observed that the fluorescence emission ofthe immunoprobe is reduced more than 50% after 10 days,limiting its utility in the immunoassay.

Development of the Competitive Fluorescent Immu-noassay and Analytical Performance Characteristics.The immunoassay format consists of a competitive fluorescentimmunoassay where nanoparticles are used as antigen labels. Inthis format, sample antigen and tracer (Oxy−BSA−QD)compete for the limited binding sites of the immobilizedantibody.First, the concentration of anti-Oxy antibody used for coating

the well-plate and immunoprobe concentration added in thecompetition step were optimized for the development of theimmunoassay. Optimum conditions (referring to the highestfluorescence intensity, the lowest signal deviation, and the bestinhibition curve) were achieved with 1 μg/mL of the antibodysolution and 10 μg/mL of the tracer (Oxy−BSA−QD)solution, ensuring a 1:1 volume ratio between the sample andthe immunoprobe involved in the competition (see Table 1).

The analytical performance characteristics are assessed by theanalysis of a series of Oxy standard solutions at differentconcentration levels, using the optimized QD-based immuno-assay. The corresponding inhibition curve was fitted using afour-parameter equation with SOFTmax Pro software (seeFigure 5). The principal parameters (see Table 2) are definedby different inhibitory concentration (IC) values calculatedfrom the inhibition curve. Thus, a linear relationship wasachieved ranging from 0.26 to 3.53 μg/L (IC20−IC80) enablingOxy trace determination in bovine muscle. The detection limit(IC10) was 0.12 μg/L of Oxy and the sensitivity of the assay(IC50) was 0.96 μg/L of Oxy in aqueous media.

Figure 4. QDs (gray line) and Oxy−BSA−QD bioconjugate (blackline) fluorescence emission spectra in buffer 10 mM PBS (pH = 7.4).

Table 1. Optimized Experimental Conditions of theImmunoassay Based on QDs

step/condition optimized value

coating 100 μL/well Aboxy (1 μg/mL)blocking 200 μL/well 3% casein[Oxy] in the bioconjugate 10 μg/mLsample:bioconjugate ratio 1:1

Journal of Agricultural and Food Chemistry Article

dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−17401736

The here developed competitive QD-based fluoroimmuno-assay shows some attractive features that make it advantageousover the established conventional immunoassays for theanalysis of TCCs in bovine meat (in general competitiveELISA formats). Such advantages include a lower detectionlimit (<1.5 μg/kg), less time consumption (avoiding somesteps required in the enzyme-linked immunosorbent assay),and low-cost assays (due to the smaller number of reagentsused).1,15,17,32 In addition, comparing with other analysistechniques, such as chromatographic methods, immunochem-ical techniques have the advantage that up to 96 samples can beanalyzed within 2 h (in the best of the cases up to 24 samplescan be analyzed by LC−MS33).Cross Reactivity Test. A study of the cross reactivity was

evaluated in order to test the performance of the antibody anti-Oxy rabbit IgG reaction with the other selected members of theTCC family under scrutiny (TC, Chlor, and Doxy). Theselection of these TCCs was based on EU CommissionRegulation7 on pharmacologically active substances and theirclassification regarding maximum residue limits in foodstuffs ofanimal origin.Different inhibition curves (corresponding to the different

selected TCCs) were registered, employing the same optimizedconditions as those used in the case of the Oxy determinations,and deviations from the standard curve were investigated. IC50values for each analyte were determined and the cross reactivityrates were calculated applying the eq 1 given in Materials andMethods. Results obtained are collected in Table 3.Considering the high values observed for the four analytes, itis confirmed that anti-Oxy could be an appropriate antibody forthe design of a screening test able to detect the presence of anyof the four selected TCCs.Additivity Studies. The immunoassay based screening

system under development aims at recognizing the presence of

total TCCs in the samples over a preset concentration level(e.g., the maximum residue level authorized for such family ofantibiotics). Therefore, fluorescence signals from the differentTCCs that could be present in the sample should be expectedas additive.To evaluate whether the analytical signals from the TCCs are

additive, sets of aqueous samples containing only a single typeof tetracyclines, as well as binary, ternary, and quaternarymixtures of the TCCs under study, were prepared and analyzedby the QD-based immunoassay. In all cases solutions contained2 μg/L of TCC total concentration. The samples were analyzedaccording to the general procedure described above, and resultsare shown in Figure 6. A good agreement between experimentalconcentrations extracted from immunoassay fluorescencemeasurements and expected concentrations were observed.This is a proof of the analytical potential of theimmunochemical method proposed for the development ofthe intended screening system of TCCs.

Characterization of the Screening System for TCCDetection. In order to characterize the here-developedscreening test, and thus obtain the UR region, a screeningcurve for TCC detection was designed and characterized inaqueous media. For this purpose, a probability versusconcentration graph was elaborated selecting a cutoff level of1 μg/L total concentration of TCCs, according to themaximum sensitivity reached by the immunoassay method-ology (IC50 factor extracted from the inhibition curve was 0.96μg/L). The limits defining the UR of our method weredetermined by calculating the concentrations of TCCs thatproduce a probability of 5% to obtain a false positive and aprobability of having a 95% of positive response (α, β errors) asit is explained in Materials and Methods. According to thesepatterns, the experimental unreliability region obtained was inthe range between 0.56 and 1.25 μg/L total TCC concentration(see Figure 7).This means that the presence of TCCs in aqueous media can

be confirmed reliably by the proposed screening system forconcentrations higher than 1.25 μg/L and, on the other hand,samples giving total concentration of TCCs below 0.56 μg/Lcan be reliably considered as noncontaminated.

Application to Bovine Muscle Samples. The applic-ability of this QD-based immunochemical method for thescreening of TCC residues in contaminated bovine meatsamples was finally evaluated. As no TCCs were detected byimmunoassay in the scrutinized samples, they were spiked withknown amounts of TCCs and then screened following thegeneral procedure above-described.In order to get a validation of the used methodology for the

extraction of analytes from bovine muscle, a high-performanceliquid chromatography (HPLC) with spectrophotometricdetection for the determination of TCCs in bovine tissueswas carried out. Bovine meat samples were extracted and thenanalyzed following the Cinquina et al. procedure.34 Theanalyses were performed using a mobile phase of 0.01 Moxalic acid:acetonitrile:methanol (60:25:15 in volume) on a

Figure 5. Fluorescent inhibition curve obtained for the analysis ofTCCs in bovine muscle samples using oxytetracycline labeled withQDs as tracer in the competitive immunoassay. All the points that plotthe curve exhibit a RSD < 5%.

Table 2. Analytical Characteristics of the Inhibition Curvefor the Analysis of TCCs in Bovine Muscle

analytical parameter obtained value

limit of detection (IC10) 0.12 μg/Llimit of quantification (IC20) 0.26 μg/Llinear range (IC20−IC80) 0.26−3.53 μg/Lsensitivity (IC50) 0.96 μg/LR2 0.999

Table 3. Cross Reactivity Test Using as Antibody RabbitAnti-Oxytetracycline IgG

antigens assayed (TCC family) CR (%)

tetracycline 89chlortetracycline 90doxytetracycline 82

Journal of Agricultural and Food Chemistry Article

dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−17401737

reversed-phase C8 column. The recoveries were obtained byspiking bovine muscle samples at three different concentrations(50, 100, and 150 μg/kg) of the TCC quaternary mixture andthen by analyzing the samples with reference to the blank.Good recoveries (average recoveries were higher than 95%)were observed in all cases, validating the selected samplepretreatment strategy here used.A 1 μg/L total concentration of TCCs was set out as cutoff

level considering the analyzed bovine muscle sample amounttaken and the dilution carried out eventually before analysis,corresponding to 10 μg/kg in the bovine muscle. Resultsobtained for the screening in the bovine muscle samples usingthe proposed QD-based immunoassay are collected in Table 4.Results were classified as positive or negative, depending ontheir values in comparison to the cutoff one. Good agreementbetween expected and experimental results was obtained. Thisconfirms that the immunoassay developed is suitable for TCCscreening of bovine meat samples. Moreover, the here-developed screening methodology may provide rigorous andreliable analytical results in the case that legislation on these

contaminants turns stricter in the future due to the lowestdetection limit (10-fold lower than the current MRL given byEuropean Legislation) and the selectivity achieved by thismethod for the TCC family compared to those developed up tonow for similar purposes.18,33 In addition, compared with otherTCC screening luminescent approaches recently proposed18

Figure 6. Additive study. In the graphic are represented the fluorescent signal intensities from different TCC mixtures (bars in black) compared withthe expected signal for the concentration assayed (2 μg/L; bars in gray). As can be seen the results are in agreement.

Figure 7. Experimental probability−concentration graph for the screening of TCCs in aqueous solution for a cutoff concentration value of 1 μg/L.

Table 4. Screening of Bovine Muscle Samples Spiked withTCCs

[TCCs] μg/kg screening result expected screening result

0a negative negative3 negative negative4 negative negative9 uncertain negative10 cutoff cutoff11 uncertain positive18 positive positive19 positive positive

aBlank.

Journal of Agricultural and Food Chemistry Article

dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−17401738

and HPLC methodologies,14 the QD-based immunoassaypresents some benefits including the possibility of analyzing ahigh number of samples simultaneously in a short time (up to96 samples/plate in 2 h), as well as the nonexistence of someproblems related to phosphorescent labels (e.g., deoxygena-tion). The drawback of the proposed methodology is the needto carry out a sample preparation step in order to extract theanalyte from the bovine muscle and make the sample colorless(filtration step) because the coloration can interfere in the finalfluorescence emission of QDs. Nevertheless, the designedprocedure is simple and up 10 samples can be prepared in 30min.

■ ASSOCIATED CONTENT*S Supporting InformationAdditional details of materials and methods including reagentsand materials; instrumentation; synthesis and solubilization ofCdSe/ZnS quantum dots; and preparation of Oxy−BSAconjugates. This material is available free of charge via theInternet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*E-mail: asm@uniovi.es. Tel: +34 985 103474.

FundingThe authors are grateful for financial support from the SpanishMinistry of Education and Science (CTQ2010-16636 andMAT2010-20921-C02-01) and the European FEDER program.Finally, L.T.-A. acknowledges the FPU program (ref: AP2008-01504) from the Spanish Education Ministry for thepredoctoral fellowship.

NotesThe authors declare no competing financial interest.

■ REFERENCES(1) Pastor-Navarro, N.; Maquieira, A.; Puchades, R. Review onimmunoanalytical determination of tetracycline and sulfonamideresidues in edible products. Anal. Bioanal. Chem. 2009, 395, 907−920.(2) Griffin, M. O.; Fricovsky, E.; Ceballos, G.; Villarreal, F.Tetracyclines: a pleitropic family of compounds with promisingtherapeutic properties. Review of the literature. Am. J. Physiol. 2010,299, C539−C548.(3) Jeon, M.; Kim, J.; Paeng, K.-J.; Park, S.-W.; Paeng, I. Biotin−avidin mediated competitive enzyme-linked immunosorbent assay todetect residues of tetracyclines in milk. Microchem. J. 2008, 88, 26−31.(4) FAO/WHO Food Standards Codex Alimentarius (2007).Veterinary drug residues in food. Maximum residue limits. URL(http://www.codexalimentarius.net/mrls/vetdrugs/jsp/vetd_q-e.jsp)(accessed 25 September 2013).(5) US Food and Drug Administration (1975). Titled 21 − Food anddrugs. Chapter I −Food and Drugs Administration Department ofHealth and Human Services, Subchapter E − Animal drugs, feed andrelated products. Part 556. Tolerances for residues of new animaldrugs in food. URL (http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/cfrsearch.cfm?fr=556.500) (accessed 25 September2013).(6) Japanese Ministry of Health Welfare and Labor (2008). Foodsafety information. URL (http://www.mhlw.go.jp/english/topics/foodsafety/index.html) (accessed 25 September 2013).(7) Commission Regulation (EU) No 37/2010 of 22 December2009.(8) Kurittu, J.; Lonnberg, S.; Virta, M.; Karp, M. A Group-SpecificMicrobiological Test for the Detection of Tetracycline Residues inRaw Milk. J. Agric. Food Chem. 2000, 48, 3372−3377.

(9) Virolainen, N. E.; Pikkemaat, M. G.; Alexander Elferink, J. W.;Karp, M. T. Rapid Detection of Tetracyclines and Their 4-EpimerDerivatives from Poultry Meat with Bioluminescent BiosensorBacteria. J. Agric. Food Chem. 2008, 56, 11065−11070.(10) Kanda, M.; Kusano, T.; Osonai, T.; Ushiyama, K.; Takeba, K.;Sakamoto, M.; Hayashi, H.; Igusa, K.; Ibe, A.; Nagayama, T. Rapiddetermination of residues of 4 tetracyclines in meat by microbiologicalscreening, HPLC and LC/MS/MS. J. Food Hyg. Soc. Jpn. 2008, 49,37−44.(11) Goudin, V.; Hedou, C.; Rault, A.; Verdon, E. Validation of a fiveplate test, the STAR protocol, for the screening of antibiotic residuesin muscle from different animal species according to EuropeanDecision 2002/657/EC. Food Addit. Contam., Part A 2010, 27, 935−952.(12) Miranda, J. M.; Rodríguez, J. A.; Galan-Vidal, C. A.Simultaneous determination of tetracyclines in poultry muscle bycapillary zone electrophoresis. J. Chromatogr., A 2009, 1216, 3366−3371.(13) Onal, A. Overview on liquid chromatographic analysis oftetracycline residues in food matrices. Food Chem. 2011, 127, 197−203.(14) Rodríguez, C.; Granados, F.; Sanchez, J.; García, F. A novelGreen chemistry method for non-aqueous extraction and HPLCdetection of first-, second-, and third-generation tetracyclines, 4-epitetracycline, and tylosin in aninal feeds. J. Agric. Food Chem. 2012,60, 7121−7128.(15) Pastor-Navarro, N.; Morais, S.; Maquieira, A.; Puchades, R.Synthesis of haptens and development of a sensitive immunoassay fortetracycline residues: Application to honey samples. Anal. Chim. Acta2007, 594, 211−218.(16) Adrian, J.; Fernandez, F.; Sanchez-Baeza, F.; Marco, M. P.Preparation of antibodies and development of an enzyme-linkedimmunosorbent assay (ELISA) for the determination of doxycyclineantibiotic in milk samples. J. Agric. Food Chem. 2012, 60, 3837−3846.(17) Zhang, Y. L.; Lu, S. X.; Liu, W.; Zhao, C. B.; Xi, R. M.Preparation of anti-tetracycline antibodies and development of anindirect heterologous competitive enzyme-linked immunosorbentassay to detect residues of tetracycline in milk. J. Agric. Food Chem.2007, 55, 211−218.(18) Traviesa-Alvarez, J. M.; Costa-Fernandez, J. M.; Pereiro, R.;Sanz-Medel, A. Direct screening of tetracyclines in water and bovinemilk using room temperature phosphorescence detection. Anal. Chim.Acta 2007, 589, 51−58.(19) Valcarcel, M.; Cardenas, S. Current and future screeningsystems. Anal. Bioanal. Chem. 2005, 381, 81−83.(20) Chan, W. C. W.; Maxwell, D. J.; Gao, X.; Bailey, R. E.; Han, M.;Nie, S. Luminescent quantum dots for multiplexed biological detectionand imaging. Curr. Opin. Biotechnol. 2002, 13, 40−46.(21) Nichkova, M.; Dosev, D.; Davies, A. E.; Gee, S. J.; Kennedy, I.M.; Hammock, B. D. Quantum Dots as Reporters in MultiplexedImmunoassays for Biomarkers of Exposure to Agrochemicals. Anal.Lett. 2007, 40, 1423−1433.(22) Xu, H.; Xu, J.; Wang, X.; Wu, D.; Chen, Z.; Georgia, W.;Andrew, Y. Quantum dot-based, quantitative, and multiplexed assayfor tissue staining. ACS Appl. Mater. Interfaces 2013, 5, 2901−2907.(23) Xie, M-X; Chen, Y.; Wang, D.; Yin, Y.; Wang, L.; Wang, X.Quantum dots capped with dummy molecularly imprinted film asluminescent sensor for determination of tetrabromobisphenol A inwater and soils. J. Agric. Food Chem. 2012, 60, 10472−10479.(24) Bradford, M. A rapid and sensitive method for the quantitationof microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248−254.(25) Trapiella-Alfonso, L.; Costa-Fernandez, J. M.; Pereiro, R.; Sanz-Medel, A. Synthesis and characterization of hapten-quantum dotsbioconjugates: application to development of a melamine fluorescentimmunoassay. Talanta 2013, 106, 243−248.(26) Sun, B.; Xie, W.; Yi, G.; Chen, D.; Zhou, Y.; Cheng, J.Microminiaturized immunoassays using quantum dots as fluorescent

Journal of Agricultural and Food Chemistry Article

dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−17401739

label by laser confocal scanning fluorescence detection. J. Immunol.Methods 2001, 249, 85−89.(27) Mikkelsen, S. R.; Corto n, E. Chapter 6: Quantitativeimmunoassays with labels. Bioanalytical Chemistry; John Wiley &Sons, Inc.: Hoboken, NJ, 2004.(28) Simonet, B. M.; Ríos, A.; Valcarcel, M. Unreliability of screeningmethods. Anal. Chim. Acta 2004, 516, 67−74.(29) Trullols, E.; Ruisanchez, I.; Rius, F. X. Validation of qualitativeanalytical methods. TrAC, Trends Anal. Chem. 2004, 23, 137−145.(30) Oka, H.; Ikai, Y.; Ito, Y.; Hayakawa, J.; Harada, K.; Suzuki, M.;Odani, H.; Maeda, K. Improvement of chemical analysis of antibiotics:XXIII. Identification of residual tetracyclines in bovine tissues byelectrospray high-performance liquid chromatography−tandem massspectrometry. J. Chromatogr., B 1997, 693, 337−344.(31) Trapiella-Alfonso, L.; Costa-Fernandez, J. M.; Pereiro, R.; Sanz-Medel, A. Development of a quantum dot-based fluorescentimmunoassay for progesterone determination in bovine milk. Biosens.Bioelectron. 2011, 26, 4753−4759.(32) Gao, F.; Zhao, G. X.; Zhang, H. C.; Wang, P.; Wang, J. P.Production of monoclonal antibody against doxycycline for immuno-assay of seven tetracyclines in bovine muscle and milk. J. Environ. Sci.Health, Part B 2013, 48, 92−100.(33) Granelli, K.; Elgerud, C.; Lundstrom, A.; Ohlsson, A.; Sjoberg,P. Rapid multi-residue analysis of antibiotics in muscle by liquidchromatography−tandem mass spectrometry. Anal. Chim. Acta 2009,637, 87−91.(34) Cinquina, A. L.; Longo, F.; Anastasi, G.; Giannetti, L.; Cozzani,R. Validation of a high-performance liquid chromatography methodfor the determination of oxytetracycline, tetracycline, chlortetracyclineand doxycycline in bovine milk and muscle. J. Chromatogr., A 2003,987, 227−233.

Journal of Agricultural and Food Chemistry Article

dx.doi.org/10.1021/jf500118x | J. Agric. Food Chem. 2014, 62, 1733−17401740