This article was downloaded by: [University of Guelph]On: 13 March 2013, At: 02:01Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Journal of Environmental Science and Health, Part B:Pesticides, Food Contaminants, and Agricultural WastesPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/lesb20

Synthesis of novel hapten and production of genericmonoclonal antibody for immunoassay of penicillinsresidues in milkSai N. Jiao a b , Ping Wang c , Guo X. Zhao a , Huui C. Zhang a b , Jing Liu a b & Jian P. Wang ab

a College of Veterinary Medicine, Agricultural University of Hebei, Baoding Hebei, Chinab Hebei Engineering and Technology Research Center of Veterinary Biological Products,Agricultural University of Hebei, Baoding Hebei, Chinac Hebei Institute of Veterinary Drug Control, Shijiazhuang Hebei, ChinaVersion of record first published: 01 Mar 2013.

To cite this article: Sai N. Jiao , Ping Wang , Guo X. Zhao , Huui C. Zhang , Jing Liu & Jian P. Wang (2013): Synthesis ofnovel hapten and production of generic monoclonal antibody for immunoassay of penicillins residues in milk, Journal ofEnvironmental Science and Health, Part B: Pesticides, Food Contaminants, and Agricultural Wastes, 48:6, 486-494

To link to this article: http://dx.doi.org/10.1080/03601234.2013.761908

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.

Journal of Environmental Science and Health, Part B (2013) 48, 486–494Copyright C© Taylor & Francis Group, LLCISSN: 0360-1234 (Print); 1532-4109 (Online)DOI: 10.1080/03601234.2013.761908

Synthesis of novel hapten and production of genericmonoclonal antibody for immunoassay of penicillinsresidues in milk

SAI N. JIAO1,2, PING WANG3, GUO X. ZHAO1, HUUI C. ZHANG1,2, JING LIU1,2 and JIAN P. WANG1,2

1College of Veterinary Medicine, Agricultural University of Hebei, Baoding Hebei, China2Hebei Engineering and Technology Research Center of Veterinary Biological Products, Agricultural University of Hebei,Baoding Hebei, China3Hebei Institute of Veterinary Drug Control, Shijiazhuang Hebei, China

The objective of this study was to produce a generic monoclonal antibody for determination of penicillins residues in milk. Thecompound 6-aminopenicillanic acid was used as the template to synthesize two novel generic haptens that were used to produce themonoclonal antibodies. The obtained monoclonal antibodies simultaneously recognized 11 penicillin drugs (amoxicillin, ampicillin,penicillin G, penicillin V, sulbenicillin, carbencillin, methicillin, cloxacillin, dicloxacillin, oxacillin, and nafcillin). After evaluationof different reagent combinations, a heterologous indirect competitive enzyme immunoassay was developed to multi-determine the11 drugs in milk. The crossreactivities to the 11 drugs were in a range of 16%–117% and the limits of detection were in a range of0.7–9.3 ng/mL depending on the drug. The recoveries from the fortified blank milk were in a range of 77.6%–99.4% with coefficientsof variation lower than 13.5%. This method could be used as a rapid screen tool for routine monitoring the residues of the 11 penicillindrugs in animal derived foods.

Keywords: Penicillins, 6-aminopenicillanic acid, monoclonal antibody, heterologous enzyme immunoassay, milk.

Introduction

Penicillin drugs (PCs) are usually used to treat variousbacterial-induced diseases in animals, e.g. mastitis in dairycow. The eleven common PCs are shown in Figure 1. How-ever, the broad use of PCs in animals may lead to theirresidues in animal-derived foods that could cause allergicreactions to the consumers [1] and accelerate the spreadingof antimicrobial resistance.[2,3] For protection of consumerhealth, the European Union has laid down different max-imum residue limits (MRLs) for some PCs in milk: peni-cillin G (PCG), ampicillin (APC) and amoxicillin (AOC),4 ng/mL; oxacillin (OXC), cloxacillin (CLC), dicloxacillin(DCLC) and nafcillin (NFC), 30 ng/mL.[4] Therefore, itis very important to monitor the PCs residues in animalderived foods.

By now, there have been many articles reporting highperformance liquid chromatography (HPLC),[5–7] liquidchromatography-mass spectrometry/mass spectrome-

Address correspondence to Jian Ping Wang, College of Veteri-nary Medicine, Agricultural University of Hebei, Baoding Hebei,China; E-mail: wjpcau@yahoo.com.cnReceived August 17, 2012.

try,[8–11] biosensor,[12,13] chemiluminescence technique,[14]

receptor-based method [15,16] and immunoassay [17–28] fordetermination of the residues of PCs or PCs metabolitesin animal derived foods. Among these methods, enzymelinked immunosorbent assay (ELISA) is a method capableof screening large number of samples in a single test.The core detection reagent of an ELISA method is theantibody. In the previously reported ELISA methods,the antibodies of AOC and APC were usually used asthe detection reagent.[19–26] Due to the different molecularstructures of these PCs (Fig. 1), the antibodies of AOC andAPC showed various crossreactivities to other analogs.

In two previous reports, the antibodies of hydrolyzed PCswere produced that only recognized the relative metabolitesbut did not recognize the parent PCs.[17,18] Cliquet et al.produced a polyclonal antibody against APC that cross re-acted with APC, AOC, PCG, OXC, CLC, and DCLC.[19]

Strasser et al. produced a polyclonal antibody against APCthat cross reacted with APC, PCG, AOC, OXC, CLC,DCLC, and NFC.[20] Samsonova et al. produced a poly-clonal antibody against APC that cross reacted with APC,PCG, azlocillin, piperacillin, and carbenicillin.[21] Usleberet al. produced a polyclonal antibody against APC thatwas able to recognize 15 PCs.[22] Richard et al. produced a

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

Penicillins residue in milk 487

NO

S CH3

CH3

COOH

H2N NH

O

NH2

NO

S CH3

CH3

COOHHO

NH

O

NH2

NO

S CH3

CH3

COOH

6-aminopenicillanic acid (6-APA) Amoxicillin (AOC) Ampicillin (APC)

NH

O NO

S CH3

CH3

COOHO

OH2C NH

NO

S CH3

CH3

COOH

NH

O

SO3Na

NO

S CH3

CH3

COOH

Penicillin G (PCG) Penicillin V (PCV) Sulbenicillin (SBC)

NH

O

COOH

NO

S CH3

CH3

COOHO

NH

NO

S CH3

CH3

COOH

OCH3

OCH3

NH

NO

S CH3

CH3

COOH

O

NO CH3

Cl

Carbencillin (CBC) Methicillin (MTC) Cloxacillin (CLC)

NH

NO

S CH3

CH3

COOH

O

NO CH3

Cl

Cl

NH

NO

S CH3

CH3

COOH

O

NO CH3 O

NH

NO

S CH3

CH3

COOHOEt

Dicloxacillin (DCLC) Oxacillin (OXC) Nafcillin (NFC)

Fig. 1. Molecular structures of 6-aminopenicillanic acid and 11 common penicillin drugs.

monoclonal antibody against APC that was able to recog-nize 13 PCs.[23] Cliquet et al. also produced a monoclonalantibody against APC that cross reacted with PCG, OXC,DCLC and carbenicillin (CBC).[24] Yeh et al. produced apolyclonal antibody against AOC that cross reacted withAOC, APC, PC, OXC and CLC.[25] Recently, a polyclonalantibody against AOC was produced in our lab that crossreacted with AOC, PCG, APC, CBC, penicillin V (PCV)and sulbenicillin (SBC).[26] There were other two reportsin which the antibodies of 6-aminopenicillanic acid [27] andAPC [28] were incorporated in fluorescent immunoassay forthe detection of PCs residues.

For development of a generic ELISA, the commonlyused method is to employ an antibody capable of recogniz-ing all the analytes of interest. This conception has beenused to develop the generic ELISA for sulfonamides,[29]

organophosphorus pesticides,[30] nitrofurans [31] and Sudandyes.[32] As shown in Figure 1, the core structure of thesePCs was 6-aminopenicillanic acid (6-APA). The specificantibody toward 6-APA should recognize all these PCs. Inthe previous report, 6-APA was directly coupled to key-hole limpet hemocyanin to prepare an immunogen and theobtained anti-APA polyclonal antibody recognized sevenPCs, but the crossreactivities were various: PCV (145%),PCG (100%), AOC (50%), APC (47%), OXC (12%), CLC(3.9%) and DCLC (1.2%).[27]

As shown in Figure 1, APA part is only a part of thePCs molecules. Therefore, the anti-APA antibody showedvarious crossreactivities to the PCs.[27] In order to produce

a broad specific antibody capable of recognizing all thesePCs, two novel generic haptens were synthesized in thepresent study with 6-APA as the template. Then the ob-tained monoclonal antibody was used to develop an ELISAfor multi-determination of PCs residues in milk. The detailsof the experiments are described below.

Materials and methods

Reagents and chemicals

The standards of 6-aminopenicillanic acid (APA), peni-cillin G potassium salt (PCG), penicillin V potassium salt(PCV), amoxicillin anhydrous (AOC), ampicillin trihydratesodium salt (APC), oxacillin sodium salt monohydrate(OXC), cloxacillin sodium salt monohydrate (CLC), di-cloxacillin sodium salt monohydrate (DCLC), methicillin(MTC), nafcillin sodium salt monohydrate (NFC), bovineserum albumin (BSA), ovalbumin (OA) and Freund’s adju-vants were all purchased from Sigma (St. Louis, MO, USA).Carbencillin disodium salt (CBC) was obtained from San-gon Biotech Co., Ltd. (Shanghai, China). Sulbenicillin(SBC) was obtained from China Institute of VeterinaryDrug Control (Beijing, China). The compound 3,3′,5,5′-tetramethylbenzidine (TMB) was purchased from Serva(Heidelberg, Germany). The p-hydroxybenzaldehyde, p-phenylenediamine and other chemical reagents were allfrom Beijing Chemical Company (Beijing, China).

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

488 Jiao et al.

glutaraldehyde

protein

H2N NH2 + +

glutaraldehyde

N

S

H2N

CH3

CH3

COOHO

H2N N CH(CH2)3CH

N

S

N

CH3

CH3

COOHO

protein N CH(CH2)3CH N N CH(CH2)3CH

N

S

N

CH3

CH3

COOHO

hapten PAPA

immunogen 1

hapten

immunogen 2

protein

HO CHO

N

O

S CH3

CH3

COOH

N

S

H2N

CH3

CH3

COOHO

HO HC N

+

carbonyldiimidazo

N

O

S CH3

CH3

COOH

O HC N

HAPA

protein

protein

N

S CH3

CH3

COOHO

O CH

NH2

C

O

NHcoating antigen 3

Fig. 2. Synthetic processes of the generic haptens and the conjugates.

Standard stock solutions of these PCs were preparedwith ultrapure water (100 µg/mL). These solutions werestored in the dark at 4◦C to be stable for two weeks.Working solutions with series concentrations (0.5, 1, 2,5, 10, 20, 50, 100 ng/mL) were prepared by diluting thestock solutions with PBS immediately before use. PBS (pH7.2) was prepared by dissolving 0.2 g KH2PO4, 0.2 g KCl,1.15 g Na2HPO4, and 8.0 g NaCl in 1000 mL deionizedwater. Washing buffer (PBST) was PBS containing 0.05%Tween. Coating buffer was carbonate buffer (0.1 M, pH9.6). Substrate buffer was 0.1 M citrate (pH 5.5). Thesubstrate system was prepared by adding 200 µL 1% (w/v)TMB in DMSO and 64 µL 0.75% (w/v) H2O2 into 20 mLsubstrate buffer.

Synthesis of hapten 1 PAPA

Hapten 1 was synthesized as shown in Figure 2. Briefly,216 mg of 6-APA (1 mmol) and 100 mg of glutaraldehyde(1 mmol) were added into 8 mL of 1,4-dioxane in a roundbottom flask. The solution was heated and refluxed for30 min to obtain a brown solution. Then 108 mg of p-phenylenediamine (1 mmol) was added and the mixturewas refluxed until some solid deposit appeared on the innerwall of the flask. After the temperature was cooled down toroom temperature, the mixture was filtered under vacuum.Finally, the deposit was washed with 50 mL of water anddried to yield hapten 1 PAPA (melting point, 226◦C; IR(KBr) Vmax 3579, 3500, 3050, 2929, 2825, 1772, 1646, 1338,1116, 810–720, 669 cm−1).

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

Penicillins residue in milk 489

Preparation of conjugates 1

Hapten 1 PAPA was coupled to carrier protein by usingglutaraldehyde method (Fig. 2). About 5 mL of DMF con-taining 42 mg of PAPA was added dropwise into 2 mL ofPBS containing 70 mg of BSA or 100 mg of OA. Then100 µL of 25% glutaraldehyde was added and the solutionwas stirred for 4 h at room temperature to prepare im-munogen 1 (PAPA-BSA) or coating antigen 1 (PAPA-OA).The conjugates were dialyzed against PBS for 3 days at4◦C and the dialysates were stored at −20◦C until use. Thecoupling ratios were determined according to the previous2,4,6-trinitrobenzene sulfonic acid method.[33]

Synthesis of hapten 2 HAPA

Hapten 2 was synthesized as shown in Figure 2.Briefly, 216 mg of 6-APA (1 mmol) and 120 mg of p-hydroxybenzaldehyde (1 mmol) were added into 8 mL of1,4-dioxane in a round bottom flask. The solution washeated and refluxed for 30 min to obtain some lemon yel-low sediment. After the temperature was cooled down toroom temperature, the mixture was filtered under vacuum.Finally, the deposit was washed with 50 mL of water anddried to yield hapten 2 HAPA (melting point, 220◦C; IR(KBr) Vmax 3648, 3110, 3000, 2983, 1772, 1623, 1580–1410,1338, 1114, 889, 763, 669 cm−1).

Preparation of conjugates 2

Hapten 2 HAPA was coupled to carrier protein by usingN,N′-carbonyldiimidazo method (Fig. 2). Briefly, 31 mgof HAPA and 17 mg of N,N’-carbonyldiimidazo were dis-solved in 5 mL of acetone to be stirred for 4 hours. Thenthe solution was added dropwise into a PBS solution con-taining 65 mg of BSA or 100 mg of OA. The mixture wasstirred for 6 hours to prepare immunogen 2 (HAPA-BSA)or coating antigen 2 (HAPA-OA). The conjugates were dia-lyzed and the coupling ratios were determined as describedabove.

Preparation of coating antigen 3

Amoxicillin (AOC) was used to prepare coating anti-gen 3 by coupling AOC to OA (AOC-OA, Fig. 2). Thepreparation process was carried out as the procedure forcoating antigen 2 by using 21 mg AOC, 8.5 mg N,N’-carbonyldiimidazo and 110 mg OA.

Production of monoclonal antibodies

The two immunogens (PAPA-BSA and HAPA-BSA) wereall used to produce the monoclonal antibodies. The ex-periments were performed at Animal Experiment Centerof College of Veterinary Medicine, Agricultural Univer-sity of Hebei according to the Regulation Guideline for

Experimental Animals issued by the Ministry of Scienceand Technology of China. Ten 8-week-old female BALB/cmice were randomly divided into two groups. The micein group 1 (numbered R1 to R5) were immunized withPAPA-BSA and the mice in group 2 (numbered R6 to R10)were immunized with HAPA-BSA. The mice were immu-nized subcutaneously with an emulsion of the immunogen(50 µg per animal, calculated as protein) in Freund’s com-plete adjuvant on the dorsal region. Thereafter the micewere boosted 8 times at 2-week intervals. Through the 8boosters, the sera were collected and the antibody titerswere monitored. The spleen from the mouse with the high-est titer in each group was removed and the splenocyteswere fused with SP2/O myeloma cells. The splenocytefused myeloma cells were cultured in 96-well plates andthe positive hybridomas were screened by using the indi-rect competitive ELISA described below with AOC as thecompetitor. The hybridomas producing the specific mono-clonal antibody to AOC were sub-cloned twice by limitingdilution method and the sub-cloned hybridomas were col-lected, centrifuged and frozen in liquid nitrogen. The as-cites from hybridoma-induced mice were purified by usingthe saturated ammonium sulfate precipitation method toobtain the monoclonal antibody.

Development of the ELISA

The optimal dilutions of coating antigen and antibody weredetermined by using the checkerboard procedure, in whichthe well with an absorbance of 1.0 was defined as the op-timal dilutions of the coating antigen and the antibody.After that, each well of a microtiter plate was coated with100 µL of coating antigen, incubated overnight at 4◦C,and then blocked with 1% fetal calf serum. The plate waswashed three times with PBST. Then 50 µL of the optimalantibody dilution and 50 µL of serial dilutions of AOCstandard were added to the wells (in triplicate) for incu-bation for 1 h at 37◦C. After washes as described above,100 µL of horseradish peroxidase-labeled goat anti-mouseIgG was added. The plate was incubated for 30 min at 37◦C.After washes, 100 µL of TMB substrate system was addedand the plate was incubated for 15 min at 37◦C. Finally,the reaction was stopped by adding 50 µL of 2 M H2SO4,and the absorbance was measured at 450 nm to obtain theabsorbance (B) of each well.

In the present study, the optimal monoclonal antibodyand the three coating antigens were arranged into ho-mologous and heterologous formats to optimize the coat-ing antigen (Table 1). Other PCs shown in Figure 1 andseveral other classes of drugs (cefalexin, cefquinome, flor-fenicol) were determined as described above. The half ofinhibition concentrations (IC50) and the limits of detection(LOD) were defined as the concentrations showing 50% and10% of inhibition, respectively. The competitive inhibitorycurves were developed by plotting the concentrations(Log C) verse the B/B0 values (mean absorbance for the

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

490 Jiao et al.

Table 1. IC50, LODs, and CRs of different reagent combinations to the 11 PCs.

AOC APC PCG PCV CBC SBC MTC CLC DCLC OXC NFC

combination 1 (PAPA-OA and antibody R4M5)IC50

a 6.2 6.7 7.6 11.9 29.5 36.5 47.6 82.7 182 95.3 86.1CR b 100 93 81 52 21 17 13 7.5 3.4 6.5 7.2

combination 2 (HAPA-OA and antibody R7M9)IC50 8.3 8.6 11.2 11.5 19.3 21.9 19.8 43.6 59.3 36.1 92.1CR 100 96 74 73 43 38 42 19 14 23 9

combination 3 (PAPA-OA and antibody R7M9)IC50 5.8 5.0 6.4 6.7 8.7 9.2 10.2 18.7 17.5 14.9 36.2CR 100 117 91 86 67 63 57 31 33 39 16LODa 0.9 0.7 1.2 2.0 3.7 4.3 4.5 6.3 8.0 6.6 9.3

combination 4 (AOC-OA and antibody R7M9)IC50 6.5 6.8 8.2 8.0 13.9 13.3 14.4 32.5 34.2 24.1 59.1CR 100 96 79 81 47 49 45 20 19 27 11

aThe unit of IC50 and LOD is ng/mL.b The CR is expressed as %.

standards divided by the zero standards). The crossreactiv-ity (CR) among these competitors was calculated accordingEquation 1.

CR (%) = 100 × IC50 AOC/IC50 competitor (1)

Sample preparation

The extraction of PCs from milk was according to a previ-ous report.[5] A volume of 5 mL milk and 5 mL acetonitrilewere added into a centrifuge tube and the mixture was vor-texed vigorously for 5 min. Then, the tube was centrifugedat 1800 rpm for 5 min. The supernatant was transferred toa clean tube and the left milk sample was re-extracted with5 mL acetonitrile. The acetonitrile phases were combinedand evaporated to dryness under vacuum on a water bath at50◦C. Finally, the dry residue was reconstituted in 5 mL ofPBS and filtered with a 0.22 µm Millipore filter for ELISAanalysis.

Some blank raw milk samples from the controlled healthcows were used to assess the matrix influence and themethod recovery. The 11 PCs were fortified into the blankmilk respectively at levels of 4–60 ng/mL to determine theaccuracy (intra- and inter-assay recovery) and the precision(coefficient of variation, CV). Furthermore, 48 unknownmilk samples (28 raw milk samples and 20 packaged milksamples) obtained from several farms and markets in Chinawere analyzed by the developed method.

Results and discussion

Haptens and immunogens

In the previous reports for immunoassay of PCs, parentAOC and APC were usually used as the haptens that weredirectly coupled to carrier protein via the amidogen groupin their molecules to prepare the immunogens.[19–26,28] In

these cases, the individual molecule of AOC or APC waspresented to the animal immune system, so the resultantantibodies showed various CRs to other PCs. In a previousreport, 6-aminopenicillanic acid (6-APA) was used as thehapten that was coupled to carrier via the amidogen groupin its molecule to prepare the immunogen.[27] The resultantpolyclonal antibody was able to recognize seven PCs, butthe CRs varied (1.2%–145%). As shown in Figure 1, thecore structure of these PCs is 6-APA. For production ofa broad specific antibody toward these PCs, the first thingis to design and synthesize a generic hapten containingthe APA structure. Therefore, two novel haptens based on6-APA were synthesized in the present study in order toproduce a monoclonal antibody capable of recognizing allthese PCs.

Hapten 1 PAPA was synthesized by coupling 6-APAto p-phenylenediamine with glutaraldehyde as the linker(Fig. 2). The melting point of PAPA (226◦C) was differ-ent from that of 6-APA (208◦C) and p-phenylenediamine(138◦C), indicating a new compound was obtained. Theinfrared spectroscopy (IR) data showed that PAPA con-tained a benzene ring, a – (CH2)n – chain and a C N bondbesides the main chemical groups of 6-APA. Furthermore,the UV spectrum of PAPA was a typical UV scan diagram(data not shown), indicating a benzene ring-contained com-pound was obtained (6-APA has no UV absorption). Thesedata could be used to deduce a conclusion that the haptenPAPA was synthesized.

Hapten 2 HAPA was synthesized by coupling 6-APA to p-hydroxybenzaldehyde (Fig. 2). The meltingpoint of HAPA (220◦C) was different from that ofp-hydroxybenzaldehyde (112◦C) and 6-APA (208◦C),indicating a new compound was obtained. The IR datashowed that HAPA contained a benzene ring, a hydroxyland a C N bond besides the main chemical groupsof 6-APA. Furthermore, the UV spectrum of HAPAbeing a typical UV scan diagram indicated a benzene

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

Penicillins residue in milk 491

ring-contained compound was obtained (data not shown).These data could be used to deduce a conclusion that thehapten HAPA was synthesized.

The two haptens all contained the APA part and theirmolecular structures were larger and more complex thanthat of 6-APA, so the antibodies from the two haptensshould be better than the previous anti-APA antibody.[27]

Then, PAPA and HAPA were coupled to BSA respectivelyto prepare two immunogens. In the two immunogens, thecentral group of these PCs (APA part) was presented to theimmune system, i.e. acting as an epitope of the immunogen,so the antibodies from the two immungens should recog-nize the 11 PCs simultaneously. The molecules of 6-APA,PAPA and HAPA all contain a β-lactam ring that is unsta-ble under acid or alkaline condition. Therefore, the reac-tion environments during hapten synthesis and immunogenpreparation were all controlled at neutral condition in orderto keep the β-lactam ring intact.

Antibody performances

During the experiments, one of these PCs, AOC, was usedas the competitor to screen the positive hybridomas. Eighthybridomas producing specific antibody to AOC were ob-tained: three strains from PAPA-BSA and five strains fromHAPA-BSA. Their specificities and sensitivities to otherPCs were determined by the indirect competitive ELISA.Results showed antibody R4M5 from PAPA-BSA and an-tibody R7M9 from HAPA-BSA showed the best perfor-mance in each group, so the two antibodies were used forthe subsequent experiments. Their IC50 and CRs to thesePCs are shown in Table 1 (combination 1 and 2). Duringthe experiments, the PCs standards were prepared with car-bonate buffer (0.1 M, pH 9.6) for ELISA analysis. Resultsshowed the antibodies showed no crossreactivity to the car-bonate buffer treated PCs, because the β-lactam ring wasbroken under alkaline condition.

As shown in combination 1 and 2, both of the two an-tibodies recognized the 11 PCs simultaneously with IC50in the range of 8.3–182 ng/mL and CRs in the range of3.4%–100%. As expected, the two antibodies all showed noCR to other classes of drugs. However, the CRs of anti-body R7M9 to these analytes (9%–100%) were higher thanantibody R4M5 (3.4%–100%). This was because of the dif-ferent spacer arms in the two immunogens and the differentmolecular structures of the two haptens.

Franek et al. showed an antibody from an immunogenwith a short spacer arm showed low ability to discrimi-nate the structural changes, whereas an antibody from animmunogen with a long spacer arm showed high abilityto recognize the structural elements.[34] In immunogen 1PAPA-BSA, there was a long spacer arm between APApart and BSA (Fig. 2), so antibody R4M5 showed high dis-crimination ability for these PCs; thereby resulting in lowCRs. In immunogen 2 HAPA-BSA, there was a short spacerbetween APA part and BSA (Fig. 2), so antibody R7M9

showed low discrimination ability for these PCs; therebyresulting in high CRs. Furthermore, the general molecularstructure of HAPA was similar to these PCs whereas PAPAwas different from these PCs (Fig. 1, Fig. 2), so antibodyR7M9 showed higher CRs to these PCs than R4M5. Afterthorough consideration, antibody R7M9 was selected forthe subsequent experiments.

As shown in combination 2, the CRs to the 11 analytesranged from 9% to 100%. This was because of their differ-ent molecular structures. As shown in Figure 1, the generalstructures of AOC, APC, PCG, PCV, SBC and CBC aresimilar to that of HAPA (benzene ring-short chain-APApart). Therefore, the CRs to them were higher than 38%.In the molecules of CLC, DCLC, OXC and NFC, there is asubstituted benzene ring, a naphthalene ring, or a complexchain between APA part and benzene ring. These struc-tures could not be recognized by the antibody, so the CRsto them were in the range of 9%–23%. MTC also con-tains a substituted benzene ring (Fig. 1), but the CR to itwas high (42%), which was not understood. In general, theperformances of the monoclonal antibody were better thanthe previous anti-APC monoclonal antibodies [23,24] and theanti-APA antibody.[27] This is the first paper reporting theproduction of a broad specific monoclonal antibody forPCs with 6-APA as the hapten template.

Heterologous ELISA

Heterology in the coating antigen has been commonly usedto improve the sensitivity of an immunoassay.[29–32,34–36]

For a fixed quantity of antibody, the better assay sensitivitywas obtained when the antibody affinity for the analytewas higher than that for the coating antigen.[35] Recently,Xu et al. showed the use of a partial structure of the targetmolecule as the coating hapten was a good strategy.[36]

Therefore, antibody R7M9 and the three coating antigens(PAPA-OA, HAPA-OA and AOC-OA) were arrangedinto three combinations to optimize the coating antigen(Table 1).

The performances of antibody R7M9 in homologousformat (combination 2) were described above. As shownin Table 1, antibody R7M9 showed better performancesin heterologous format (combination 3 and 4) withCRs in the range of 11%–117% and IC50 in the range of5.0–59.1 ng/mL. In addition, the CRs differences amongthese competitors in heterologous format were lowerthan that in homologous format. In heterologous format,antibody R7M9 showed low affinity for the coating antigen(PAPA-OA in combination 3 or AOC-OA in combination4), thereby improving the antibody binding for the com-petitors. This was the advantage of heterologous ELISA.

Between the two heterologous combinations, theperformances when using PAPA-OA (CRs of 16%–117%and IC50 of 5.0–36.2 ng/mL) were better than thatwhen using AOC-OA (CRs of 11%–100% and IC50 of6.5–59.1 ng/mL). There were two possible reasons. First,

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

492 Jiao et al.

001011

0

20

40

60

80

100

B/B

0 (%

)

Log C (ng/mL)

OXC PCG AOC

Fig. 3. Competitive inhibitory curves for the standards of AOC,PCG, OXC.

the coating hapten in AOC-OA (AOC) was similar toHAPA, but that in PAPA-OA (PAPA) was different fromHAPA (Fig. 2). This meant antibody R7M9 showed lowaffinity for PAPA-OA, so the performance when usingPAPA-OA was better. Second, PAPA-OA was equivalentto linking APA part and BSA with an extremely longspacer (Fig. 2). This meant the actual coating haptenin PAPA-OA (APA part) was only a part of the coatinghapten in AOC-OA (AOC). These results were consistentwith the previous reports.[35,36] In addition, the extremelylong linker in PAPA-OA may also influence the antibodybinding. Therefore, antibody R7M9 and coating antigenPAPA-OA were used for the subsequent experiments. TheLODs for the 11 drugs were in the range of 0.7–9.3 ng/mL(combination 3). The standard competitive inhibitorycurves for five representative PCs are shown in Figure 3(AOC, PCG, and OXC) and Figure 4 (APC and MTC) withanalytes concentrations in the range of 0.5–100 ng/mL.

Sample extraction and recovery

In a previous paper, acetonitrile was used to extract eightPCs from milk with high recoveries (>82.9%),[5] so ace-tonitrile was also used as extraction solvent in this study.For assessment of the matrix effect, APC and MTC wereprepared with the extracts of blank milk to develop thematrix matched competitive curves. As shown in Figure 4,their matrix matched competitive curves were similar tothat of their standards in PBS, revealing the matrix influ-ence was minimal. Then, the 11 drugs were fortified intoblank milk respectively to evaluate the accuracy and theprecision. The fortification levels for AOC, APC, PCG,PCV, CLC, DCLC, OXC and NFC were at and higherthan their respective MRLs. Due to lack of official MRLsfor sulbenicillin (SBC), methicillin (MTC), and carben-

0010110

20

40

60

80

100

B/B

0 (%

)

Log C (ng/mL)

APC matrix matched APC MTC matrix mathced MTC

Fig. 4. Competitive inhibitory curves for the standards of APC,MTC and their matrix matched standards.

cillin (CBC), their fortification concentrations were selectedat two medium levels (10 and 50 ng/mL). As shown inTable 2, the intra-assay recoveries ranged from 81.7% to99.4% with CVs lower than 10.3% and the inter-assay re-coveries ranged from 77.6% to 98.0% with CVs lower than13.5%.

Table 2. Recoveries of the 11 penicillin drugs from fortified blankmilk (n = 6).

intra-assay inter-assay

added recovery CV recovery CVanalyte (ng/mL) (%) (%) (%) (%)

AOC 4 92.6 7.5 86.7 13.520 93.7 5.2 97.4 6.3

APC 4 96.3 6.7 90.2 9.420 99.4 5.6 89.0 6.7

PCG 4 92.6 8.2 89.0 11.220 94.6 6.7 78.2 5.4

PCV 4 92.6 10.3 80.3 8.120 89.7 5.5 89.6 5.9

CBC 10 92.4 8.6 86.4 9.250 84.2 4.3 98.0 6.7

SBC 10 94.0 9.2 89.6 12.350 96.7 6.7 91.8 7.3

MTC 10 95.3 8.6 93.0 9.950 97.8 5.3 90.1 7.7

CLC 30 90.5 5.4 78.6 7.190 94.6 3.9 84.5 8.2

DCLC 30 86.1 6.7 83.7 6.890 81.7 4.1 94.2 5.4

OXC 30 93.2 5.0 89.4 7.390 94.6 3.0 77.6 9.4

NFC 30 84.3 6.8 81.5 8.090 97.4 4.9 93.5 8.4

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

Penicillins residue in milk 493

The 48 unknown milk samples were also analyzed bythe ELISA. Three raw milk samples and one packagedmilk sample were determined as positive samples, andthe residue levels calculated as AOC were 11, 19, 33 and6 ng/mL. This means one sample containing any of the 11PCs can be determined as positive, but the result can only beexpressed as AOC equivalent and the specific analyte can-not be verified due to the antibody’s high crossreactivity.Therefore, the positive ELISA results need to be confirmedwith an instrumental method, e.g. HPLC or LC-MS. Suchan instrumental method for simultaneous determination ofthe 11 PCs remains to be studied.

Conclusion

The residues of penicillin drugs in milk are dangerousto the consumers. Therefore, it is important to develop amulti-analytes analytical method to monitor their residues.This study first synthesized a generic hapten based on 6-aminopenicillanic acid and produced a broad specific mon-oclonal antibody capable of recognizing 11 penicillin drugs.Then, a heterologous ELISA method was developed tomulti-determine the 11 penicillin drugs in milk. From theanalysis of fortified blank samples and unknown samples,this method could be used as a rapid and simple tool forroutine monitoring the residues of PCs in milk, thoughthe results need to be confirmed with other instrumentalmethod.

Acknowledgment

The authors of this work are grateful for the finan-cial support of Hebei Scientific and Technology Project(12220402D), Hebei Natural Science Foundation (C2011204021) and Shijiazhuang Scientific and TechnologicalProject (11150252A). The authors Jiao and Wang con-tributed equally to this work.

References

[1] Sullivan, T.J.; Wedner, H.J.; Shatz, G.S.; Yecies, L.D.; Parker, C.W.Skin testing to detect penicillin allergy. J. Allergy Clin. Immu. 1981,68, 171–180.

[2] Jacoby, G.A. β-Lactamase nomenclature. Antimicrob. Agents Ch.2006, 50, 1123–1129.

[3] Roesch, M.; Perreten, V.; Doherr, M.G.; Schaeren, W.; Schal-libaum, M.; Blum, J.W. Comparison of antibiotic resistance of udderpathogens in dairy cows kept on organic and on conventional farms.J. Dairy Sci. 2006, 89, 989–997.

[4] European Communities. Official Journal of the European Commu-nities, Commission Regulation (EEC) NO 2377/90, 1990, L 224,1–8.

[5] Bailon-Perez, M.I.; Garcıa-Campana, A.M.; del Olmo Iruela, M.;Cruces-Blanco, C. Trace determination of 10 β-lactam antibioticsin environmental and food samples by capillary liquid chromatog-raphy. J. Chromatogr. A 2009, 1216, 8355–8361.

[6] Hsieh, S.H.; Huang, H.Y.; Lee, S. Determination of eight penicillinantibiotics in pharmaceuticals, milk and porcine tissues by nano-liquid chromatography. J. Chromatogr. A 2009, 1216, 7186–7194.

[7] Kukusamude, C.; Santalad, A.; Boonchiangma, S.; Burakham, R.;Srijaranai, S.; Chailapakul, O. Mixed micelle-cloud point extrac-tion for the analysis of penicillin residues in bovine milk by highperformance liquid chromatography. Talanta 2010, 81, 486–492.

[8] Van Holthoon, F.; Mulder, P.P.; van Bennekom, E.O.; Heskamp,H.; Zuidema, T.; van Rhijn, H.J. Quantitative analysis of peni-cillins in porcine tissues, milk and animal feed using derivatisa-tion with piperidine and stable isotope dilution liquid chromatog-raphy tandem mass spectrometry. Anal. Bioanal. Chem. 2010, 396,3027–3040.

[9] Gaugain-Juhel, M.; Delepine, B.; Gautier, S.; Fourmond, M.P.;Gaudin, V.; Hurtaud-Pessel, D.; Verdon, E.; Sanders, P. Validationof a liquid chromatography-tandem mass spectrometry screeningmethod to monitor 58 antibiotics in milk: a qualitative approach.Food Addit. Contam. A 2009, 26, 1459–1471.

[10] Kantiani, L.; Farre, M.; Sibum, M.; Postigo, C.; Lopez de Alda,M.; Barcelo, D. Fully automated analysis of beta-lactams in bovinemilk by online solid phase extraction-liquid chromatography-electrospray-tandem mass spectrometry. Anal. Chem. 2009, 81,4285–4295.

[11] Zhang, X.; Chen, L.; Xu, Y.; Wang, H.; Zeng, Q.; Zhao, Q.; Ren, N.;Ding, L. Determination of β-lactam antibiotics in milk based onmagnetic molecularly imprinted polymer extraction coupled withliquid chromatography-tandem mass spectrometry. J. Chromatogr.B 2010, 878, 3421–3426.

[12] Gustavsson, E.; Bjurling, P.; Ase, S.O. Biosensor analysis of peni-cillin G in milk based on the inhibition of carboxy peptidase activity.Anal. Chim. Acta. 2002, 468, 153–159.

[13] Song, K.; Jeong, E.; Jeon, W.; Cho, M.; Ban, C. Aptasensorfor ampicillin using gold nanoparticle based dual fluorescence-colorimetric methods. Anal. Bioanal. Chem. 2012, 402, 2153–2161.

[14] Li, Y.; Lu, J. Chemiluminescence flow-injection analysis of beta-lactam antibiotics using the luminol-permanganate reaction. Lu-minescence 2006, 21, 251–255.

[15] Abouzied, M.; Sarzynski, M.; Walsh, A.; Wood, H.; Mozola, M.Validation study of a receptor-based lateral flow assay for detectionof beta-lactam antibiotics in milk. J. AOAC Int. 2009, 92, 959–974.

[16] Janine, L.; Michael, P. Development of a receptor-based microplateassay for the detection of beta-lactam antibiotics in different foodmatrices. Anal. Chim. Acta. 2007, 586, 296–303.

[17] Grubelnik, A.; Padeste, C.; Tiefenauer, L. Highly sensitive enzymeimmunoassay for the detection of beta-lactam antibiotics. FoodAgric. Immunol. 2001, 13, 161–169.

[18] Zhang, Y.; Jiang, Y.; Wang, S. Development of an enzyme-linkedimmunosorbent assay to detect benzylpenicilloic acid, a degradationproduct of penicillin G in adulterated milk. J. Agric. Food Chem.2010, 58, 8171–8175.

[19] Cliquet, P.; Goddeeris, B.M.; Okerman, L.; Cox, E. Production ofpenicillin-specific polyclonal antibodies for a group-specific screen-ing ELISA. Food Agric. Immunol. 2007, 18, 237–252.

[20] Strasser, A.; Usleber, E.; Schneider, E.; Dietrich, R.; Burk, C.;Martlbauer, E. Improved enzyme immunoassay for group-specificdetermination of penicillins in milk. Food Agric. Immunol. 2003,15, 135–143.

[21] Samsonova, Zh.; Shchelokova, O.; Ivanova, N.; Rubtsova, M.;Egorov, A. Enzyme linked immunosorbent assay of ampicillin inmilk. Appl. Biochem. Micro. 2005, 41, 589–595.

[22] Usleber, E.; Litz, S.; Martlbauer, E. Production and characteriza-tion of group specific antibodies against penicillin antibiotics. FoodAgric. Immunol. 1998, 10, 317–324.

[23] Richard, D.; Usleber, E.; Martlbauer, E. The potential of mono-clonal antibodies against ampicillin for the preparation of a multi-immunoaffinity chromatography for penicillins. Analyst 1998, 123,2749–2754.

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3

494 Jiao et al.

[24] Cliquet, P.; Cox, E.; Van Dorpe, C.; Schacht, E.; Goddeeris, B.M.Generation of class-selective monoclonal antibodies against thepenicillin group. J. Agric. Food Chem. 2001, 49, 3349–3355.

[25] Yeh, L.C.; Lee, W.M.; Koh, B.W.; Chan, J.P.; Liu, C.H.; Kao, J.P.;Chou, C.C. Development of amoxicillin enzyme-linked immunosor-bent assay and measurements of tissue amoxicillin concentrationsin a pigeon microdialysis model. Poult. Sci. 2008, 87, 577–587.

[26] Jiao, S.N.; Liu, J.; Zhang, Y.F.; Zhao, G.X.; Wang, J.P. Preparationof a bi-hapten antigen and the broad-specific antibody for simulta-neous immunoassay of penicillins and tetracyclines in milk. FoodAgric. Immunol. 2012, 23, 273–287.

[27] Benito-Pena, E.; Moreno-Bondi, M.C.; Orellana, G.; Maquieira,A.; Amerongen, A.V. Development of a novel and automated fluo-rescent immunoassay for the analysis of beta-lactam antibiotics. J.Agric. Food Chem. 2005, 53, 6635–6642.

[28] Bacigalupo, M.A.; Meroni, G.; Secundo, F.; Lelli, R. Time-resolvedfluoroimmunoassay for quantitative determination of ampicillin incow milk samples with different fat contents. Talanta 2008, 77,126–130.

[29] Franek, M.; Diblikova, I.; Cernoch, I.; Vass, M.; Hruska, K. Broad-specificity immunoassays for sulfonamide detection: immunochem-ical strategy of generic antibodies and competitors. Anal. Chem.2006, 78, 1559–1567.

[30] Xu, Z.L.; Xie, G.M.; Li, Y.X.; Wang, B.F.; Beier, R.C.; Lei, H.T.;Wang, H.; Shen, Y.D.; Sun, Y.M. Production and characteriza-tion of a broad-specificity polyclonal antibody for O,O-diethyl

organophosphorus pesticides and a quantitative structure-activityrelationship study of antibody recognition. Anal. Chim. Acta. 2009,647, 90–96.

[31] Li, J.; Liu, J.; Zhang, H.C.; Li, H.; Wang, J.P. Broad-specificity indi-rect competitive ELISA for determination of nitrofurans in animalfeeds. Anal. Chim. Acta. 2010, 678, 1–6.

[32] Liu, J.; Zhang, H.; Zhang, D.; Gao, F.; Wang, J. Production of themonoclonal antibody against Sudan 2 for immunoassay of Sudandyes in egg. Anal. Biochem. 2012, 423, 246–252.

[33] Sashidhar, R.B.; Capoor, A.K.; Ramana, D. Quantitation of aminogroups using amino acids as reference standards by trinitrobenzenesulfonic acid: A simple spectrophotometric method for the estima-tion of hapten to carrier protein ratio. J. Immunol. Method. 1994,167, 121–127.

[34] Franek, M.; Zeravik, J.; Eremin, S.A.; Yakovleva, J.; Badea, M.;Danet, A.; Nistor, C.; Ocio, N.; Emneus, J. Antibody based meth-ods for surfactant screening. J. Fresenius Anal. Chem. 2001, 371,456–466.

[35] Goodrow, M.H.; Hammock, B.D. Hapten design for compound-selective antibodies: ELISAS for environmentally delecterious smallmolecules. Anal. Chim. Acta. 1998, 376, 83–91.

[36] Xu, Z.L.; Xie, G.M.; Li, Y.X.; Wang, B.F.; Beier, R.C.; Lei, H.T.;Wang, H.; Shen, Y.D.; Sun, Y.M. Broad-specificity immunoassay forO,O-diethyl organophosphorus pesticides: application of molecularmodeling to improve assay sensitivity and study antibody recogni-tion. Anal. Chem. 2010, 82, 9314–9321.

Dow

nloa

ded

by [

Uni

vers

ity o

f G

uelp

h] a

t 02:

01 1

3 M

arch

201

3