Ho

HSa

Wb

a

ARRAA

KMHAE

1

ii

a3mwddiCn

(

0d

Analytica Chimica Acta 665 (2010) 84–90

Contents lists available at ScienceDirect

Analytica Chimica Acta

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

apten synthesis and antibody production for the developmentf a melamine immunoassay

ongtao Leia,1, Yudong Shena,1, Lijun Songa, Jinyi Yanga, Olivier P. Chevallierb,imon A. Haugheyb, Hong Wanga, Yuanming Suna,∗, Christopher T. Elliottb,∗∗

The Key Laboratory of Food Safety of Guangdong Province/Institute of Food Quality and Safety, South China Agricultural University,ushan Road, Tianhe District, Guangzhou 510642, Guangdong, PR China

Institute of Agri-Food and Land Use, Queen’s University Belfast, Belfast BT9 5AG, Northern Ireland, United Kingdom

r t i c l e i n f o

rticle history:eceived 27 November 2009eceived in revised form 28 February 2010ccepted 7 March 2010vailable online 15 March 2010

eywords:elamineaptenntibodynzyme-linked immunosorbent assay

a b s t r a c t

The incorporation of melamine into food products is banned but its misuse has been widely reportedin both animal feeds and food. The development of a rapid screening immunoassay for monitoring ofthe substance is an urgent requirement. Two haptens of melamine were synthesized by introducingspacer arms of different lengths and structures on the triazine ring of the analyte molecular structure. 6-Aminocaproic acid and 3-mercaptopropionic acid were reacted with 2-chloro-4,6-diamino-1,3,5-triazine(CAAT) to produce hapten 1 [3-(4,6-diamino-1,6-dihydro-1,3,5-triazin-2-ylamino) hexanoic acid] andhapten 2 [3-(4,6-diamino-1,6-dihydro-1,3,5-triazin-2-ylthio) propanoic acid], respectively. The molecu-lar structures of the two haptens were identified by 1H nuclear magnetic resonance spectrometry, massspectrometry and infrared spectrometry. An immunogen was prepared by coupling hapten 1 to bovineserum albumin (BSA). Two plate coating antigens were prepared by coupling both haptens to egg oval-

bumin (OVA). A competitive indirect enzyme-linked immunosorbent assay (ciELISA) was developed toevaluate homogeneous and heterogeneous assay formats. The results showed that polyclonal antibod-ies with high titers were obtained, and the heterogeneous immunoassay format demonstrated a betterperformance with an IC50 of 70.6 ng mL−1, a LOD of 2.6 ng mL−1 and a LOQ of 7.6 ng mL−1. Except for cyro-mazine, no obvious cross-reactivity to common compounds was found. The data showed that the haptensynthesis was successful and the resultant antisera could be used in an immunoassay for the rapid and

bann

sensitive detection of this

. Introduction

Melamine, with an IUPAC name of 1,3,5-triazine-2,4,6-triamine,s an organic base. In its chainlike “polymerized” form, melamines an industrial chemical and has been used for decades in the

Abbreviations: CAAT, 2-chloro-4,6-diamino-1,3,5-triazine; BSA, bovine serumlbumin; DCC, dicyclohexylcarbodiimide; NHS, N-hydroxysuccinimide; TMB,,3′ ,5,5′-tetramethylbenzidine; DMF, dimethylformamide; TLC, thin-layer chro-atography; PBS, phosphate-buffered saline; PBST, phosphate-buffered salineith TW-20; HRMS, high-resolution mass spectrometry; hapten 1, 3-(4,6-iamino-1,6-dihydro-1,3,5-triazin-2-ylamino) propanoic acid; hapten 2, 3-(4,6-iamino-1,6-dihydro-1,3,5-triazin-2-ylthio) propanoic acid; ELISA, enzyme-linked

mmunosorbent assay; iELISA, indirect ELISA; ciELISA, competitive indirect ELISA;R, cross-reactivity; UV, ultraviolet spectrometry; IR, infrared spectrometry; NMR,uclear magnetic resonance.∗ Corresponding author. Tel.: +86 2085280270; fax: +86 2085288282.

∗∗ Corresponding author. Tel.: +44 2890976549; fax: +44 2890976513.E-mail addresses: ymsun@scau.edu.cn (Y. Sun), chris.elliott@qub.ac.uk

C.T. Elliott).1 Equal contributors.

003-2670/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.aca.2010.03.007

ed chemical.© 2010 Elsevier B.V. All rights reserved.

manufacturing of dishes, plastic resins, flame-retardant fibers,components of paper and paperboard and industrial coatings [1,2].Food has had limited exposure to melamine through a number ofthese contact materials. The Food and Drug Administration (FDA)states that the yield of chloroform soluble extracts should notexceed 0.5 mg in.−2 (77.5 �g cm−2) of food contact surface (21 CFR1777.1460). The maximum permitted concentration for melaminein food has been set at 2.5 mg kg−1 by the European Commission[3,4]. Melamine is not approved to be added to foods or feeds, noris it permitted to be used as a fertilizer anywhere in the world. Anyfuture approval is highly unlikely as ingestion of melamine maylead to reproductive damage, bladder or kidney stones, which inturn may lead to bladder carcinogenesis [5].

However, in 2007 residues of melamine were detected in petfood, animal feed, wheat gluten, and other protein-based food

commodities, and subsequently pet illness and death were widelyreported [1]. In the following year, several companies and individ-uals were implicated in a scandal involving milk and infant formulawhich had been adulterated with melamine, leading to kidneystones and renal failure [6–8]. Thus, the discovery of illegally added

H. Lei et al. / Analytica Chimica Acta 665 (2010) 84–90 85

ute of

md

dmwCqGltaoort

ietoiAaac

btaieeo

2

2

u(CfM(taMCAf

Fig. 1. Synthesis ro

elamine to food products resulted in the urgent requirement toetect melamine in food and feed.

Liquid chromatographic determination of melamine waseveloped for beverage samples in 1987 [9]. Ion-pair liquid chro-atography coupled to electrospray tandem mass spectrometryas used to determine residue of melamine in chard samples using-18 chromatography, the limit of detection (LOD) and limit ofuantification (LOQ) were 0.01 and 0.1 mg kg−1, respectively [10].as chromatography–mass spectrometry and ultra-performance

iquid chromatography–tandem mass spectrometry were also usedo detect melamine, the claimed sensitivity of this procedure was 10nd 5 �g kg−1, respectively [11]. However, quantitative detectionf melamine is currently limited to instrumental methods, mostf which share a number of important drawbacks: they usuallyequire costly apparatus, their sample throughput is limited andhey are not suitable for on-site analysis.

Immunoassay technology has been increasingly used for screen-ng food contaminants due to the sensitivity, selectivity, timefficiency, cost-effective and portability of the procedures. Overhe past 20 years the development of immunochemical meth-ds and their potential applications, especially the enzyme-linkedmmunosorbent assay (ELISA), has grown significantly [12,13].lthough several commercial ELISA kits to melamine have becomevailable in recent times [2], no studies on the hapten synthesis,ntibody production, and development of immunoassays to thisompound have been published to our knowledge.

The aim of the present research is to prepare polyclonal anti-ody against melamine for immunoassay development. To achievehis aim, haptens containing two different linkers were synthesizednd one was used as an immunogen and both were used as coat-ng antigen. The recognition performance of antibody obtained wasvaluated with some compounds based on their structural differ-nce and cross-reactivity data. Also, the specificity and sensitivityf the homogeneous and heterogeneous ELISA were compared.

. Experiments

.1. Reagents and chemicals

General reagents and organic solvents were of analytical gradenless specified otherwise. 2-Chloro-4,6-diamino-1,3,5-triazineCAAT) was obtained from Shanghai Chaoyan Biotechnologyo., Ltd. (Shanghai, China). 6-Aminocaproic acid was bought

rom Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).elamine, bovine serum albumin (BSA), dicyclohexylcarbodiimide

DCC), N-hydroxysuccinimide (NHS), ovalbumin (OVA), 3,3′,5,5′-etramethylbenzidine (TMB), complete and incomplete Freund’s

djuvants were purchased from Sigma (St. Louis, MO, USA). 3-ercaptopropionic acid was bought from Alfa Aesar (Tianjin,

hina). Cyanuric chloride and cyanuric acid were obtained fromccela Chembio Co., Ltd. (Shanghai, China). Atrazine was a gift

rom Shandong Zhongke Qiaochang Chemical Co., Ltd. (Shandong,

melamine haptens.

China). Cyromazine was bought from Changzhou Zhineng Ani-mal Pharmaceutical Co., Ltd. (Changzhou, China). HRP-conjugatedgoat anti-rabbit IgG was obtained from Boster Biotech Co., Ltd.(Wuhan, China). Thin-layer chromatography (TLC) was performedon 200 mesh, 2.5 mm precoated silica gel GF254 on glass sheetsobtained from Qingdao Haiyang Chemical Co., Ltd. (Qingdao,China). Polystyrene ELISA plates were obtained from Jincanhua Co.,Ltd. (Shenzhen, China).

Buffers used in this study were prepared as follows: 50 mMcarbonate buffer (pH 9.6) for coating plates, 10 mM PBST solutionphosphate buffer saline (PBS, pH 7.4, containing 0.1% Tween-20)was used for washing plates, 0.1 M citrate and sodium phosphatebuffer (pH 5.4) for substrate buffer, and 2 M H2SO4 as the stoppingreagent. TMB solution was prepared by addition of 10 mL substratebuffer and 150 �L of 15 mg mL−1 TMB in dimethylformamide (DMF)and 2.5 �L of 6% (w/v) H2O2.

2.2. Instrumentation

Ultraviolet spectrometry (UV) were recorded on a UV-3010spectrophotometer (Hitachi, Japan). High-resolution mass spec-trometry (HRMS) analyses were performed using a MAT95XPhigh-resolution mass spectrometer (Thermo, USA). Nuclear mag-netic resonance (NMR) spectra were obtained with the DRX-600NMR spectrometers (Bruker, Germany–Switzerland). Infraredspectrometry (IR) was performed using Nicdet Avatar 360 (Thermo,USA). Melting point was determined on apparatus (Gallenkamp,UK). ELISA plates were washed using a Multiskan MK-2 microplatewasher (Thermo Labsystems, USA). Absorbance was measured ata wavelength of 450 nm using a Multiskan MK3 microplate reader(Thermo Labsystems, USA).

2.3. Hapten synthesis and characterization

2.3.1. Synthesis of 6-(4,6-diamino-1,3,5-triazin-2-ylamino)hexanoic acid (hapten 1, Fig. 1)

CAAT (1.21 g, 8.1 mmol) was dissolved in 150 mL of absoluteethanol, and 6-aminocaproic acid (1.26 g, 9.4 mmol) and KOH(1.64 g, 24.9 mmol) in 10 mL of absolute ethanol were added drop-wise, refluxed at 70 ◦C for 24 h, monitored by TLC. The reactionmixture was filtered, the filtrate was cooled to obtain a whitesolid. The solid was dissolved in 75 mL of 5% sodium carbonateand extracted with 60 mL of methylene chloride (3× 20 mL). Theaqueous phase was acidified to pH 3 with 6 N HCl, concentratedand cooled, 0.128 g of hapten 1 was obtained as a white solid in6.6% yield. m.p. 185.0–186.0 ◦C, TLC Rf = 0.5 (CHCl3:MeOH = 60:20).HRMS (EI) m/z calculated for C9H16N2O2 [M] 240.1335, found

240.1329. 1H NMR (DMSO-d6, 600 MHz) ı: 1.32–1.25 (m, 2H),1.56–1.45 (m, 4H), 2.21 (t, J = 7.35, 7.35 Hz, 2H), 3.25 (dd, J = 12.97,6.65 Hz, 2H); 7.85–8.18 (br, 4H). 13C NMR (MeOH-d4, 150 MHz) ı:177.4 (C-1), 41.6 (C-6), 34.8 (C-2), 30.09 (C-5), 27.2 (C-4), 25.6 (C-3).IR (KBr) �max (cm−1): 3492 (s), 3409 (s), 3331 (s), 3159 (s), 2937 (s),

86 H. Lei et al. / Analytica Chimica Acta 665 (2010) 84–90

tructu

212

2p

e(wmldt1

J(m(18

2

jtj(iTrsd7d

2

n

Fig. 2. Synthesis route and s

860 (w), 1642 (vs), 1553 (s), 1473 (s), 1442 (s), 1371 (s), 1309 (w),165 (w), 811 (w). The maximum wavelength of UV absorption was41 nm.

.3.2. Synthesis of 3-(4,6-diamino-1,3,5-triazin-2-ylthio)ropanoic acid (hapten 2, Fig. 1)

CAAT (1.21 g, 8.1 mmol) was suspended in 150 mL of absolutethanol, and 3-mercaptopropionic acid (1.0 g, 9.4 mmol) and KOH1.64 g, 24.9 mmol) in 10 mL of absolute ethanol were added drop-ise, refluxed at 78 ◦C for 28 h, monitored by TLC. The reactionixture was filtered, the filtrate cake was washed with cool abso-

ute ethanol. The white solid was dissolved in 10 mL of cooledeionized water (0 ◦C) and acidified to pH 6 with 6 N HCl. 0.74 g hap-en 2 was obtained as a white solid in 42.4% yield. m.p. 218–221 ◦C;H NMR (600 MHz, DMSO-d6) ı: 2.64 (t, J = 6.98 Hz, 2H), 3.14 (t,= 7.02 Hz, 2H), 6.9 (br, 4H). 13C NMR (DMSO-d6, 150 MHz) ı:178.3C-1); 173.0 (C-4); 165.3 (C-5,6); 34.3 (C-2); 24.2 (C-3). MS (APCI)/z: 216 [M]; MS2 (APCI) m/z: 198 [M−OH], 144 [M−CH2COOH]; IR

KBr) �max (cm−1): 3422 (vs), 3157 (s), 2924 (s), 2854 (w), 1711 (s),644 (vs), 1531 (s), 1403 (s), 1308 (w), 1252 (w), 1189 (w), 921 (w),03 (w). The maximum wavelength in UV absorption is 275 nm.

.4. Preparation of hapten–protein conjugates (Fig. 2)

Hapten 1 was coupled to BSA to be used as an immunogen (con-ugate 1), and both haptens 1 and 2 were coupled with OVA usinghe active ester method to obtain two plate coating antigens (con-ugates 2 and 3). Briefly, 20.0 mg (100 �mol) of hapten 1 or 21.0 mg100 �mol) of hapten 2, NHS 17.0 mg (150 �mmol) were dissolvedn 1 mL of DMF, followed by addition of DCC 31.0 mg (150 �mmol).he activation reaction was carried out at 4 ◦C overnight with stir-ing. The white dicyclohexylurea precipitate was removed fromolution by centrifugation. The supernatant (900 �L) was addedropwise to BSA (113 mg) or OVA (75 mg) in 8 mL PBS (10 mM, pH.4). The conjugate mixture was stirred at 4 ◦C for 12 h and thenialyzed against 10 mM PBS (pH 7.4).

.5. Production of polyclonal antibodies

Three New Zealand rabbits weighing 1.5–2.0 kg were immu-ized 5 times using conjugate 1 at intervals of 14 days by the

res of melamine conjugates.

Guangdong Medical Laboratory Animal Center. The rabbits wereblood sampled to detect the presence of antibodies to melamineusing an indirect ELISA on the eighth day after each immunization,starting 40 days after the first injection. Serum was divided intoaliquots (1 mL) and stored at −20 ◦C until use.

2.6. ELISA development

In the indirect ELISA (iELISA), all incubations were performed at37 ◦C except for the coating antigen. Flat-bottom polystyrene ELISAplates were coated with hapten–OVA (1 �g mL−1, 100 �L well−1)in carbonate buffer (pH 9.6) over night at 4 ◦C. The wells werewashed 5 times with PBST solution, and then blocked with 5%skim milk in PBS buffer (200 �L well−1) for 1 h. After washing 5times with PBST solution, the wells were incubated with 100 �L ofdiluted antibody in PBST for 1 h and washed 5 times with PBST solu-tion. HRP-conjugated goat anti-rabbit IgG diluted 1:9000 in PBSTwas added (100 �L well−1). After incubation for 1 h and washing5 times with PBST solution, TMB solution was added to the wells(100 �L well−1) and incubated for 15 min. The reaction was stoppedby addition of 50 �L well−1 of 2 M H2SO4, and the absorbance wasrecorded at 450 nm.

For the competitive indirect ELISA (ciELISA), the procedurewas identical except for the addition of a competition step afterthe blocking, which was involved with adding 50 �L well−1 ofmelamine standards dissolved in PBS containing 0.8% methanol,followed by 50 �L well−1 of appropriate concentrations of antiseradiluted with PBST. The concentrations of antibodies and coatingantigens had been optimized by checkerboard titration and com-petitive curves were then obtained by plotting the normalizedsignal B/B0 against the logarithm of analyte concentration, whereB0 is the signal without analyte and B is the signal of each concen-tration of analyte.

A four-parameter logistic equation (OriginPro 7.5 for Windows)was used to fit the sigmoidal curve according to the following for-mula (1):

Y = A − D

[1 + (x/C)B] + D(1)

where A is the asymptotic maximum (maximum absorbance inabsence of analyte, Amax), B is the curve slope at the inflexion point,

H. Lei et al. / Analytica Chimica Acta 665 (2010) 84–90 87

Table 1Parameters comparison of homogeneous and heterogeneous ciELISA.

Coating antigen Rabbit 1 serum Rabbit 2 serum

(ng m

.0

.0

Ccaw

2

cmt

%

3

3

tacttrawctrrg3rm

ewaciattwdvd

ds

3

dT

sensitivity achievable. The results have been presented in Fig. 3 andTable 1. For rabbit 1 serum, the IC50 with heterogeneous format wasfound to be 758.8 ng mL−1 (6.02 nM) compared to 1346.6 ng mL−1

(10.68 nM) with the homogeneous format. The IC50 based on rabbit2 serum was found to be substantially better than that of rabbit 1,

Titer IC50 (ng mL−1) LOD (ng mL−1) LOQ

Conjugate 2 4 × 105 1346.6 37.6 153Conjugate 3 2 × 105 758.8 32.2 90

is the x value at the inflexion point (corresponding to analyteoncentration giving 50% inhibition of Amax, IC50), and D is thesymptotic minimum (background signal). 10% inhibition valueas defined as LOD; 20% inhibition value was defined as LOQ.

.7. Specificity of antibody

The specificity of the antibody was evaluated by performingompetitive assays using several compounds structurally related toelamine as competitors, and the obtained IC50 values were used

o calculate cross-reactivity using the formula:

CR = IC50 (melamine)IC50 (cross reactant)

× 100 (2)

. Results

.1. Hapten synthesis

Hapten design is a critical factor for the successful prepara-ion of highly specific antibodies against low molecular weightntigens [13–15]. A suitable hapten and its subsequent proteinonjugate for immunization should be designed as a mimic tohe target molecule. If melamine itself was utilized as reactanto derive a hapten, due to its three amino groups on the triazineing having the same reactivity, the derivation product would inll likelihood be complicated in that one, two or all three aminoould be substituted when using glutaraldehyde or halogenated

arboxylic acid such as sodium chloro- or bromo- acetic acid orheir esters. For this reason CAAT was selected as the derivationeactant due to the active chloride group being capable of reactingelatively easily with the aliphatic primary amino or the sulphydrylroup of a spacer arm reagent such as 6-aminocaproic acid or-mercaptopropionic acid, respectively. Catalyzed with KOH, theesultant product possessed the antigenic epitopes most likely toimic that of melamine (Fig. 1).In some cases, immunogens prepared without a spacer arm are

asier to produce but result in an assay with poor sensitivity and/oreak recognition of the segment of the target molecule near the

ttachment site on the carrier protein [12–15]. Spacer groups whichontain 4–6 carbons have been shown to be optimal in obtain-ng high quality antibodies [15,16]. In the present study, a spacerrm containing 6 carbons was designed for immunogen prepara-ion. This was selected to expose a specific region of the hapteno the animals’ immune system. Furthermore, a three carbon armith a heterogeneous sulphur atom instead of a nitrogen atom wasesigned for hapten 2. The strategy employed was selected to pre-ent any antibody binding to the spacer arm, which may have aetrimental effect in the assay.

From the data generated by 1H NMR and MS and IR, it wasetermined that the desired haptens, 1 and 2, were successfullyynthesized.

.2. Serum titer determination

Three rabbits were immunized with conjugate 1, however oneied during the immunization period (cause of death unknown).he other two rabbits produced sera that resulted in antibody

L−1) Titer IC50 (ng mL−1) LOD (ng mL−1) LOQ (ng mL−1)

8 × 104 708.8 42.8 95.82 × 104 70.6 2.6 7.6

titers >80,000 (Table 1) as determined by indirect ELISA, with titerbeing defined as the time of the serum dilution that results in anabsorbance value that is about 1.0 OD units.

3.3. Comparison of homogeneous and heterogeneous ciELISA

A heterogeneous approach in the current context refers to thehapten used for immunogen production being different from thatused for the coating antigen. Heterogeneous formats can oftenresult in antibodies having a higher affinity towards the analytein comparison to the coating antigen or tracer hapten [14–17].Thus the sensitivity achievable using this format as opposed to ahomologous one can improve sensitivity substantially.

In the present study, conjugates 2 and 3 were used as homoge-neous and heterogeneous coating antigens in order to compare the

Fig. 3. Comparison of homogeneous and heterogeneous ciELISA. Data representedin mean ± SD (standard deviation) and n = 3. (a) Standard curve of rabbit 1 serum;(b) standard curve of rabbit 2 serum.

88 H. Lei et al. / Analytica Chimica Acta 665 (2010) 84–90

Table 2Cross-reactivity of antiserum to related compounds.

Compounds Structure Conjugate 2 Conjugate 3

IC50 (nmol mL−1) Cross-reactivity (%) IC50 (nmol mL−1) Cross-reactivity (%)

Melamine 5.62 100 0.56 100.0

Cyromazine 2.1 267.6 0.30 186.7

Hapten 1 0.34 1652.9 0.26 215.4

Hapten 2 13.36 42.4 14.08 4.0

CAAT 218.44 2.6 181.34 0.3

Cyanuric acid ND �0.01 ND �0.01

Atrazine ND �0.01 ND �0.01

Cyanuric chloride ND �0.01 ND �0.01

N gistic

iwT2shlfTi[

f

D presented infinite IC50 values and could not be fitted with the four-parameter lo

.e. in the region of 10 times more sensitive. Similar improvementsith both LOD and LOQ measurements were also found (Table 1).

he working range between melamine concentrations giving0–80% inhibition were 0.76–41.54 nM (95.8–6744.2 ng mL−1) forerum 1 and 0.06–5.8 nM (7.6–731.4 ng mL−1) for serum 2 in theeterogeneous assay format. The US FDA have published several

iquid chromatography methods coupled with mass spectrometeror melamine analysis, the claimed LOQ was 10 �g kg−1 [18–20].herefore, the LOQ determined in the presented heterogeneous

mmunoassay could meet the requirements of both the US FDA18–20] and European Commission [2,3].

Due to its better sensitivity, rabbit 2 serum was selected forurther specificity investigation.

equation.

3.4. Specificity of antibody

To determine which structural features of the molecules areimportant to antibody recognition, the cross-reactivity (CR) againsta range of compounds structurally related to melamine was tested.Both conjugates 2 and 3 were employed as homogeneous and het-erogeneous coating antigens to compare the specificity differencesobserved (Table 2). Amongst all the related compound tested inthe present study, the cross-reactivity of hapten 1 was highest in

both homogeneous and heterogeneous format assays, followed bycyromazine, melamine, hapten 2, CAAT, atrazine, cyanuric acid andcyanuric chloride which showed less than 0.01% cross-reactivity tothe developed antibody.

himica

shwcmmhapns

htlta

3

ghmttccottmatiatism

obattechpntotUjbiaaten

d

& D Program of China during the 11th Five-Year Plan Period(No. 2006BAD27B02-05) and China Guangdong Provincial Science

H. Lei et al. / Analytica C

Since the immunogen was prepared using hapten 1, it was rea-onable to expect that the developed antibody would recognizeapten 1 with a high affinity. Although the arm of cyromazineith a large ring structure was different from that of hapten 1,

yromazine displayed more %CR than melamine in both assay for-ats. However, removing the spacer arm to change hapten 1 intoelamine resulted in the %CR decreasing from 1652.9 to 100 in the

omogeneous assay, and from 215.4 to 100 in the heterogeneousssay (Table 2). This data strongly indicated that the spacer armlayed an important role in hapten–antigen binding. Similar phe-omenon has been reported in other immunoassay developmenttudies [21–24].

In the homogeneous format, the 267.6% CR of cyromazine wasigher than the 186.7% CR in the heterogeneous format. Similarly,he %CRs of hapten 2 and CAAT in the heterogeneous assay wereower than those in homogeneous assay. These findings indicatedhat the assay format employed can affect not only sensitivity butlso specificity.

.5. Discussion

If the cyclopropyl group of cyromazine and amidocaproic acidroup of hapten 1 being regarded as two different spacer arms ofaptens, and comparing the structures and cross-reactivity data ofelamine, cyromazine and haptens, it was interesting to find that

he compound without the spacer arm, melamine, had lower %CRhan that with spacer arms, hapten 1 and cyromazine. Althoughyromazine contains one hydrocarbon ring structure which canause steric hindrance to hapten–antibody binding [25,26], the %CRf cyromazine was still higher than melamine but lower than hap-en 1 with a straight hydrocarbon spacer arm. Thus, it is possiblehat the steric hindrance reduces antibody recognition of cyro-

azine compared to hapten 1; and that even though a spacerrm was just a hydrocarbon straight chain, it could take part inhe hapten–antibody interaction. Also, when the hapten in themmunogen contains a spacer arm, even if the arm of recognizednalyte contains a ring structure with the steric hindrance effect,he analyte with ring could possibly be recognized by antibod-es easier than that without a spacer arm. Consequently, the armtructure of analyte could be used to adjust the assay perfor-ance.However, hapten 2 contained one spacer arm, mercaptopropi-

nic acid group, which had 3 carbons less than the arm of hapten 1,ut the %CR of hapten 2 was lower than not only cyromazine withring arm, but also melamine without a spacer arm. It seemed that

here was another factor affecting the hapten–antibody recogni-ion stronger than the spacer arm effect. We think that electronicffects may contributed to the lower %CR of hapten 2. In someases, electronic features were also responsible for governingapten–antibody recognition [26]. In the present study, the sul-hur atom in hapten 2 contained two lone-pair electrons, while theitrogen atom in the spacer arm of hapten 1 and cyromazine con-ained only one single lone-pair electrons. The additional lone-pairf electrons on the sulphur atom resulted in a stronger conjuga-ive effect with the triazine ring, which had been verified by theV data. The UV information of the molecule can indicate the con-

ugative structures, and extending conjugative effects will result inathochromic and hyperchromic shifts in absorption [27]. The max-

mum absorbance wavelength of hapten 1 was found to be 241 nm,s opposed to 275 nm for hapten 2, which are typical bathochromicnd hyperchromic shifts. This data indicated that hapten 2 con-

ained a stronger conjugative structure. Therefore, the electronicffect observed is likely to have played a critical role in the recog-ition of the melamine antibody to different compounds.

Moreover, the chlorine atom on CAAT had strong electron with-rawing effects on the triazine ring, which is another obviously

Acta 665 (2010) 84–90 89

effect different to the conjugative effects of the sulphur atom onhapten 2 and the nitrogen atom of hapten 1 or cyromazine. Whenonly one amino group of melamine was substituted by a chlo-rine atom, the cross-reactivity decreased dramatically from 100%to 2.6% in homogeneous assay or 0.31% in heterogeneous assay,respectively (Table 2). It is possible that the electron withdrawingeffect influenced significantly the binding of the hapten and anti-body. Also, comparison of the %CR of hapten 2 and CAAT showedthat the %CR of the latter was much lower than the former, whichsuggested that the electronic withdrawing effect of the substituentatom in the arm might have inversely affected the hapten–antibodybinding more than the conjugative effect of the sulphur atom inhapten 2. Similarly, comparing the spacer arm steric effect (basedon melamine, cyromazine, hapten 1) and electronic effect (basedon melamine, hapten 2, CAAT), it seemed that the latter resultedin more influence on the recognition of hapten–antibody than theformer.

Furthermore, atrazine can be regarded as a derived CAAT whosetwo free amino groups are blocked with one ethyl and one iso-propyl group, respectively. Then, removing the two alkyl blockinggroups of atrazine to expose both naked amino groups, the cross-reactivity was observed to increase from≤0.01% of atrazine to 2.57%or 0.3% of CAAT in homogeneous and heterogeneous assay, respec-tively (Table 2). Similarly, both cyanuric acid and cyanuric chloride– which contain a triazine ring but no naked amino groups – couldnot be recognized by the antibody either. The result suggestedthat both free amino groups were necessary epitopes for antibodyrecognition.

Cyromazine is a fly-killing pesticide frequently used for prevent-ing housefly and Fannia canicularis in livestock and Liriomyza sativaeand leafminer on vegetable and flowers [1,10]. It is known thatmelamine is the main metabolite of cyromazine [1,10]. To lower theunexpected cross-reactivity from cyromazine, one hapten with anon-hydrocarbon spacer arm, e.g. 6-hydrazinyl-1,3,5-triazine-2,4-diamine, could possibly be used to prepare an immunogen whichcould hopefully raise antibodies recognizing cyromazine weaklierbecause the hydrazine arm of hapten and cyclopropyl group ofcyromazine are quite distinct.

4. Conclusion

Two haptens with different spacer arms were synthesized andwere used successfully for antibody production. Two conjugateswere used to compare and evaluate the antibody performance,the result showed that sensitivity and specificity of antibody weremuch better in the heterogeneous format assay than in homoge-neous one. Based on the comparison of structure and biologicalactivity data, spacer arm effects and electronic effects appear to betwo important factors with regard to the antibody binding to thehaptens. With further development work the antibody produced inthe current study could be applied to a range of immunochemicalplatforms to deliver a highly rapid and sensitive screening proce-dure for melamine in foods and feeds.

Acknowledgements

This work was supported by National Key Technologies R

and Technology Department (2007A020100006-10, zgzhzd0808,2009B011300005).

We are grateful to Dr Yingju Liu, College of Science, South ChinaAgricultural University, for conducting infrared spectrometry andmelting point test.

9 imica

R

[

[

[[

[

[

[[

[

[

[l

[

[

[

0 H. Lei et al. / Analytica Ch

eferences

[1] W.C. Andersen, S.B. Turnipseed, C.M. Karbiwnyk, S.B. Clark, M.R. Madson, C.M.Gieseker, R.A. Miller, N.G. Rummel, R. Reimschuessel, J. Agric. Food Chem. 56(2008) 4340–4347.

[2] E.A.E. Garber, J. Food Prot. 71 (2008) 590–594.[3] European Commission Decision 2008/921/EC, amending Decision

2008/798/EC.[4] European Commission Decision 2008/798/EC, imposing special conditions gov-

erning the import of products containing milk or milk products originating inor consigned from China, and repealing Commission Decision 2008/757/EC.

[5] H. Ogasawara, K. Imaida, H. Ishiwata, K. Toyoda, T. Kawanishi, C. Uneyama, S.Hayashi, M. Takahash, Y. Hayashi, Carcinogenesis 16 (1995) 2773–2777.

[6] C.A. Brown, K.S. Jeong, R.H. Poppenga, B. Puschner, D.M. Miller, A.E. Ellis,K. Kang, S. Sum, A.M. Cistola, S.A. Brown, J. Vet. Diagn. Invest. 19 (2007)525–531.

[7] N. Yan, L. Zhou, Z. Zhu, X. Chen, J. Agric. Food Chem. 57 (2009) 807–811.[8] S. Ehling, S. Tefera, I.P. Ho, Food Addit. Contam. 24 (2007) 1319–1325.[9] H. Ishiwata, I.T. Yamazaki, K. Yoshihira, J. Assoc. Off. Anal. Chem. 70 (1987)

457–460.10] J.V. Sancho, M. Ibanez, S. Grimalt, O.J. Pozo, F. Hernandez, Anal. Chim. Acta 530

(2005) 237–243.11] X. Xia, S. Ding, X. Li, X. Gong, S. Zhang, H. Jiang, J. Li, J. She, Anal. Chim. Acta 651

(2009) 196–200.12] L.G. Bachas, M.E. Meyerhoff, Anal. Biochem. 156 (1986) 223–238.13] C.P. Chan, Y.C. Cheung, R. Renneberg, M. Seydack, Adv. Biochem. Eng. Biotech-

nol. 109 (2008) 123–154.

[

[[

[

Acta 665 (2010) 84–90

14] C. Roucairol, S. Azoulay, M.C. Nevers, C. Créminon, J. Grassi, A. Burger, D. Duval,Anal. Chim. Acta 589 (2007) 142–149.

15] Y.J. Kim, Y.A. Cho, H.S. Lee, Y.T. Lee, S.J. Gee, B.D. Hammock, Anal. Chim. Acta475 (2003) 85–96.

16] M.P. Marco, S.J. Gee, B.D. Hammock, Trends Anal. Chem. 14 (1995) 415–425.17] M. Paknejad, M. Javad Rasaee, J. Mohammadnejad, M. Pouramir, M. Rajabibazl,

M. Kakhki, J. Toxicol. Sci. 33 (2008) 565–573.18] W.C. Andersen, S.B. Turnipseed, C.M. Karbiwnyk, M.R. Madson, Lab. Inf.

Bull. 23 (2007), Available at http://www.fda.gov/Food/ScienceResearch/LaboratoryMethods/DrugChemicalResiduesMethodology/ucm071796.htm.

19] S. Turnipseed, C. Casey, C. Nochetto, D. Heller, Lab. Inf. Bull. 24 (2008),Available at http://www.fda.gov/Food/ScienceResearch/LaboratoryMethods/DrugChemicalResiduesMethodology/ucm071637.htm.

20] M. Smoker, A.J. Krynitsky, Lab. Inf. Bull. 24 (2008), Available athttp://www.fda.gov/Food/ScienceResearch/LaboratoryMethods/Drug.ChemicaResiduesMethodology/ucm071673.htm.

21] J.V. Mercader, C. Suárez-Pantaleón, C. Agulló, A. Abad-Somovilla, A. Abad-Fuentes, J. Agric. Food Chem. 56 (2008) 7682–7690.

22] E.M. Brun, M. Garcés-García, R. Puchades, A. Maquieira, J. Immunol. Methods295 (2004) 21–35.

23] N. Lee, D.P. McAdam, J.H. Skerritt, J. Agric. Food Chem. 46 (1998) 520–534.

24] Q. Zhang, L. Wang, K.C. Ahn, Q. Sun, B. Hu, J. Wang, F. Liu, Anal. Chim. Acta 596

(2007) 303–311.25] D.W. Brown, Y.T. Kim, G.W. Siskind, J. Immunol. Methods 116 (1989) 45–51.26] M.T. Muldoon, C.K. Holtzapple, S.S. Deshpande, R.C. Beier, L.H. Stanker, J. Agric.

Food Chem. 48 (2000) 537–544.27] V. Promarak, S. Ruchirawat, Tetrahedron 63 (2007) 1602–1609.