Article Cite This: Anal. Chem. 2018, 90, 2716−2724
pubs.acs.org/ac
Four Hapten Spacer Sites Modulating Class Specificity: Nondirectional Multianalyte Immunoassay for 31 β‑Agonists and Analogues Lanteng Wang,†,∇ Wenmeng Jiang,†,∇ Xing Shen,† Xiangmei Li,† Xin-an Huang,‡ Zhenlin Xu,† Yuanming Sun,† Shun-Wan Chan,§ Lingwen Zeng,∥ Sergei Alexandrovich Eremin,⊥,# and Hongtao Lei*,† †
Guangdong Provincial Key Laboratory of Food Quality and Safety, South China Agricultural University, Guangzhou 510642, China Tropical Medicine Institute & South China Chinese Medicine Collaborative Innovation Center, Guangzhou University of Chinese Medicine, Guangzhou 510405, China § Faculty of Science & Technology, Technology & Higher Education Institute of Hong Kong, Hong Kong, China ∥ South China Institute for Stem Cell Biology and Regenerative Medicine, Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou 510530, China ⊥ Faculty of Chemistry, M.V. Lomonosov Moscow State University, 119991 Moscow, Russia # A.N. Bach Institute of Biochemistry, Research Center of Biotechnology of the Russian Academy of Sciences, 119071 Moscow, Russia ‡
S Supporting Information *
ABSTRACT: Immunoassay methods are important for monitoring β-agonists illegally used for reducing animal fat deposition in livestock. However, there is no simultaneous screening surveillance immunoassay for detecting various βagonist chemicals that are possibly present in food. In this study, through the use of an R-(−)-salbutamol derivative as the immunizing hapten, an antibody recognizing 31 β-agonists and analogues was generated for the first time. Three-dimensional quantitative structure−activity relationship (3D QSAR) revealed that strong steric and hydrophobic fields around the hapten spacer near C-2, as well as a chirality at C-1′, dominantly modulated the class specificity of the raised antibody. However, a hapten spacer linked at C-2′ or C-1 would lead to a narrow specificity, and the spacer charge at C-6 could affect the raised antibody specificity spectrum. A class specificity competitive indirect enzyme-linked immunosorbent assay (ciELISA) was established with an ideal recovery ranging from 81.8 to 118.3% based on the obtained antibody. With a good agreement to the HPLC/MS method, the proposed ciELISA was confirmed to be reliable for the rapid surveillance screening assay of β-agonists in urine. This investigation will contribute to the rational design and control of the immunoassay specificity. β-Agonists are a class of therapeutic drugs commonly used in the treatments for acute symptoms of asthma, owing to their bronchodilator activities, e.g., salbutamol, clenbuterol, terbutaline, salmeterol, etc.1 An overdose of β-agonists could lead to symptoms of nausea, dizziness, and palpitation and even cause death.2 However, some of these compounds were sometimes found to be illegally used as bronchodilating agents, because they can reduce carcass fat content and then result in a muscle hypertrophy.3 Salbutamol, clenbuterol, brombuterol, and ractopamine were well-known as “lean meat agents” in China; these agents resulted in serious public health accidents.4 In 2009 and 2011, a number of β-agonist poisoning incidents occurred in China, and these incidents caused sickness to 70 people in Guangzhou and 91 people in Changsha, respectively.5 In the European Union (EU) and China, the illegal use of β© 2018 American Chemical Society
agonists as a growth promoter in animal husbandry is forbidden.6 Antibody-based immunoassays are effective tools for either quantitative or qualitative detection of chemical residues in foods and environment. They do not require complicated equipment and are capable of analyzing a quantity of samples simultaneously. Most of the immunoassay methods, such as enzyme-linked immunosorbent assay (ELISA), DNA labeled immunoprobes, immunochromatographic tests, and so on, had focused on the detection of single β-agonist7,8 or a limited number of β-agonists.9−11 However, numerous β-agonists have Received: November 13, 2017 Accepted: January 21, 2018 Published: January 21, 2018 2716
DOI: 10.1021/acs.analchem.7b04684 Anal. Chem. 2018, 90, 2716−2724
Article
Analytical Chemistry
Figure 1. Structures of β-agonists, analogues, and haptens.
comparative molecular field analysis (CoMFA) and a comparative molecular similarity indices analysis (CoMSIA) were then used for the investigation of the three-dimensional quantitative structure−activity relationship (3D QSAR) of the antibody and β-agonists. The effect of hapten spacer sites on the class specificity were reasonably elaborated and predicted.
been used or might be used in animal products.12 There is an urgent need to develop an immunoassay for the simultaneous detection of a series of β-agonists and analogues instead of one assay for each specific target. However, there is not yet a multianalyte immunoassay for a nondirectional screening purpose to detect a wide enough variety of β-agonists and possibly occurring analogues in animal food. This is at least ascribed to the lack of a class specificity antibody and understanding of the class recognition structure−activity relationship, thus leading to the difficult artificial control of the resultant antibody specificity.13 Salbutamol possesses two enantiomers, R-(−)-salbutamol and S-(+)-salbutamol. The R-(−)-isomer is approximately 80 times more potent than the S-(+)-isomer regarding the therapeutically bronchodilating effects, and pure R-(−)-salbutamol can reduce side effects.14,15 In our previous study, the raised antibody against Rac-salbutamol demonstrated an unexpected cross-reactivity (CR, 447.3 and 255.8%) to brombuterol and clenbuterol.16 This suggested that salbutamol would possibly be a hapten candidate to produce a class specificity antibody to β-agonists. In this study, to develop a class specific immunoassay and better investigate the recognition between the antibody and βagonists, the pure isomer R-(−)-salbutamol, instead of Racsalbutamol, was selected as immunizing hapten to raise the antibody. It was interestingly found that 31 β-agonists and analogues could be well-recognized by this resultant antibody. Based on the cross-reactivity and the structure analysis, a
■
EXPERIMENTAL SECTION Reagents and Animals. Rac-salbutamol, R-(−)-salbutamol, and S-(+)-salbutamol were obtained from Yunhui Trade Co., Ltd., (Guangzhou, China). 3-Dehydroxy salbutamol (3DSAL), pirbuterol, zilpaterol, octopamine, colterol, bitolterol, indacaterol, olodaterol, fenspiride, reproterol, t-butylnorsynephrine, bromoclenbuterol, benzaldehyde,5-[2-[(1,1dimethylethyl)amino]-1-hydroxyethyl]-2-hydroxy (DAH), salbutamol methyl ether, salbutamon, hydroxymethyl clenbuterol (HYD), clenhexerol, 5-hydroxy salbutamol, 1-(3,5-dimethoxyphenyl)-2-(isopropylamino)ethanol (DIM), cimbuterol, and procaterol were purchased from TRC (Toronto, Canada). Salmeterol, vilanterol, mabuterol, mapenterol, carbuterol, ritodrine, orciprenaline, clenbuterol, clenpenterol, clenproperol, and bambuterol were purchased from Dr. Ehrenstorfer (Augsburg, Germany). Penbutolol, brombuterol, terbutaline, tulobuterol, ractopamine, clorprenaline, hexoprenaline, and benzenemethanol,4-amino-α-[[(1,1-dimethylethyl)amino]methyl] (BEN) were purchase from Witega (Berlin, Germany). Alprenolol, pindolol, cimaterol, arformoterol, propafenone and 2717
DOI: 10.1021/acs.analchem.7b04684 Anal. Chem. 2018, 90, 2716−2724
Article
Analytical Chemistry
energy minimization. Calculated oil/water partition coefficient (ClogP) values were computed by the CLOGP program. The detailed methods are described in the Supporting Information. Sample Analysis. In order to evaluate the accuracy and precision of the immunoassay, the negative swine urine samples gifted by Dr. Yu Wang from the Guangzhou Institute for Food Control, China, were spiked with salbutamol, clenbuterol, brombuterol, and cimbuterol at three concentrations (1, 5, and 10 ng mL−1), respectively, and then centrifuged at 27 000g for 15 min. The supernatants were collected as spiked samples for the ciELISA analysis. The method confirmation experiments were carried out by comparing ciELISA with HPLC/MS. The HPLC/MS analysis of spiked urine samples was conducted following the procedure of the China National Standard (GB/T 22286−2008).
propranolol were purchased from Aladdin (Shanghai, China). Epinephrine, norepinephrine, and isoproterenol were purchased from Kewei Co., Ltd., (Shanghai, China). R-(−)-2tert-Butylamino-1-phenylethanol (R-(−)-TER), S-(+)-2-tertbutylamino-1-phenylethanol (S-(+)-TER), N,N-dimethylformamide (DMF), isobutyl chloroformate, bovine serum albumin (BSA), ovalbumin (OVA), and complete and incomplete Freund’s adjuvants were purchased from Sigma (St. Louis, MO, U.S.A.). Phosphate-buffered saline with 0.1% Tween-20 (PBST, 0.01 mol L−1, pH 7.4) was used as the working buffer for aqueous standard solutions of analytes. All other chemicals and organic solvents that were analytical grade or better were obtained from a local chemical supplier (Yunhui Trade Co., Ltd., Guangzhou, China). New Zealand white rabbits that were 2−3 months old (about 2 kg) were purchased from Guangdong Experimental Animal Center. Apparatus. The ultraviolet−visible (UV−vis) spectrum was recorded on a UV-4000 spectrophotometer (Hitachi, Japan). The electronic circular dichroism (ECD) spectrum was obtained on a Chirascan Circular Dichroism Spectrometer (Applied Photophysics, U.K.). The purified antibody concentration was determined by a NanoDrop 2000c spectrophotometer (Thermo Scientific, U.S.A.). ELISA plates were washed in an MK2 microplate washer (Thermo Scientific, U.S.A.). The ELISA absorbance was measured at a wavelength of 450 nm with a Multiskan MK3 microplate reader (Thermo Scientific, U.S.A.). The nuclear magnetic resonance (NMR) spectrum was obtained with a DRX-400 NMR spectrometer (Bruker, Germany). The chromatography was manipulated on the HPLC/MS system (LC-30-API5500, Shimadzu, Japan), and an Inertsil ODS-SP HPLC column (C18, 4.6 × 150 mm, 5 μm, GL Science, Japan) was used. Hapten Synthesis. Succinic anhydride (77 mg, 0.77 mmol) was added to a stirred mixture of R-(−)-salbutamol (220 mg, 0.77 mmol) in dry ethanol (40 mL) under nitrogen at room temperature. The solution was stirred at room temperature for 12 h and filtered to obtain the hapten R-(−)-salbutamol (R(−)-sal-hapten, 39 mg, 11.5%). 1H NMR (DMSO, 400 MHz): δ (ppm) 1.12 (9H, s), 2.38−2.41 (2H, m), 2.53−2.56 (2H, m), 2.76−2.78 (2H, m), 4.61−4.62 (1H, m), 5.05 (2H, d, J = 2.0 Hz), 6.77 (1H, d, J = 8.0 Hz), 7.10 (1H, dd, J = 8.4, 2.0 Hz), 7.27 (1H, d, J = 2.0 Hz). MS m/z (+ESI): 340.1 [M + H]+. Antibody Production. The hapten−protein conjugate preparation and antibody production methods are described in the Supporting Information. The antibodies from rabbits were purified with caprylic-acid-saturated ammonium sulfate precipitation, and the purity was confirmed by sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE).17 ciELISA Procedure. The competitive indirect enzymelinked immunosorbent assay (ciELISA) for β-agonists was conducted according to the previously reported procedure.18 The ciELISA calibration curves were fitted with a fourparameter logistic function.19 IC50 is the concentration of the analyte that resulted in 50% inhibition. The limit of detection (LOD) was defined as the concentration of analyte that inhibited 10% binding (IC10).20 The linear range was defined as the lower and upper concentration that provided 20−80% inhibition.21 The antibody specificity was measured with the cross-reactivity (CR) of the structurally related β-agonists (Figure 1).22 Molecular Modeling. The molecular modeling was conducted by the SYBYL-X 2.1.1 program package.23 Tripos force field and Gasteiger−Huckel charges were used in the
■
RESULTS AND DISCUSSION Absolute Configuration of Hapten. The experimental ECD spectrum of the R-(−)-hapten in acetonitrile is shown in Figure S1. The energy, oscillation, and rotational strength of the R-(−)-isomer structure were calculated by Gaussian 1624 using time-dependent density functional theory (TD-DFT). The calculated ECD spectrum was computed at the B3LYP/6311+G(2d,p) level using the B3LYP/6-311+G(2d,p) optimized geometries. It was found that the experimental spectrum was similar to the calculated spectrum, and both spectra exhibited strong positive Cotton effects near 203 nm and negative Cotton effects near 192 nm (Figure S1). Therefore, the hapten absolute configuration could be confirmed as the R-(−)-isomer.25 Immunoreagent Preparation. UV−vis spectra of the hapten−protein conjugate, carrier protein (BSA, OVA,) and R(−)-sal-hapten (Figure S2) implied that the coupling of the hapten and carrier protein was successful.26 SDS−PAGE (Figure S3) showed that the purified R-(−)-salbutamol antibody exhibited a heavy chain at about 50 kDa and a light chain at about 25 kDa. This indicated that the antibody purity was ideal for further investigation.27 The purified antibody concentration determined using an A280 absorbance measurement was 3.8 mg mL−1.28 ciELISA. Optimal concentrations of coating antigen and antibody were critical to the sensitivity of ciELISA.29,30 It was found that the combination of the coating antigen at 12.5 ng mL−1 together with the antibody at 0.1 μg mL−1 (1/32 000 dilution) exhibited a maximum absorbance around 1.0 and an optimal sensitivity.31 The established ciELISA calibration curve for R-(−)-salbutamol exhibited an IC50 of 0.5 ng mL−1, a concentration range from 0.1 to 40.9 ng mL−1, and an LOD of 0.04 ng mL−1 (Figure 2). Specificity. In the United States, only ractopamine and zilpaterol were approved as feed additives for cattle and swine by the FDA, and other compounds are banned in animal products.32 However, all β-agonists have been banned for animal products in the European Union (EU) and China. Ractopamine, clenbuterol, salbutamol, zilpaterol, cimaterol, terbutaline, and tulobuterol were officially banned and are most often found being illegally used in China.5,32 For purpose of investigating the impact of hapten structure on the antibody specificity, 54 compounds were used for the evaluation of crossreactivity. These compounds included the most common βagonists and their analogues, e.g., clenbuterol, ractopamine, zilpaterol, clenhexerol, and 3-dehydroxy salbutamol. Moreover, some potentially abused compounds in animal products, e.g., brombuterol, bambuterol, penbuterol, clorprenaline, mabuterol, 2718
DOI: 10.1021/acs.analchem.7b04684 Anal. Chem. 2018, 90, 2716−2724
Article
Analytical Chemistry
two 3D QSAR methods, CoMFA and CoMSIA, were conducted for the structure−activity analysis. Generation of 3D QSAR Models. The interactions between the antibody and β-agonists were analyzed by CoMFA and CoMSIA models, which simulated the binding process from electrostatic, steric, and hydrophobic fields, respectively (Table S1). In the CoMFA model, the q2 and r2 values were 0.560 and 0.972. The q2 and r2 of the CoMSIA model were 0.530 and 0.978. All the parameters could prove the simulation result was reliable. The scatter plots of the predicted value versus experimental activity are shown in Figure S6, and the values of training set and test set molecules are shown in Table S2. Since the activities could be well-predicted by the two models, this confirmed that the accuracy of the models was satisfactory. The contour maps reflected the effects of different groups or structures on the molecular binding affinity in various force fields. In the CoMFA model, from the view of the steric field (Figure 3a), green contours meant that bulk steric groups around this area could improve the biding affinity, and yellow contours would exert a reverse effect. From the view of the electrostatic field (Figure 3c), blue contours meant that positive charges were favorable to a high affinity, and red contours would exert a reverse effect. In the CoMSIA model, from the view of the hydrophobic field, magenta contours represented that an increase in hydrophobicity was favorable to a high affinity, and white contours had a reverse effect (Figure 3e). C-2′ Effect. The tertamyl at C-2′ in clenpenterol (39) was surrounded by a large block of green and magenta contours (Figure 3a,f), and this suggested that a bulky or hydrophobic substituent could be favored for the antibody biding. HYD (37) showed low cross-reactivity (CR, 0.3%); this could be ascribed to its unfavorable hydrophilicity of the hydroxyl at C-2′ (ClogP, 1.392). The steric field of the tertamyl on clenpenterol (39) was larger than that of the tert-butyl on clenbuterol (36) or the isopropyl on clenproperol (40) (Figure 1), and the hydrophobicity (ClogP, 2.922) of the tertamyl on clenpenterol was also higher than that of the tert-butyl on clenbuterol (ClogP, 2.393) and the isopropyl on clenproperol (ClogP, 1.994). Thus, the cross-reactivity difference (12.8% for clenperterol, 2.9% for clenbuterol, and ND for clenproperol) was ascribed to the difference of both the steric and hydrophobic fields at C-2′ of these compounds. Similarly, it was also ascribed to the steric and hydrophobic field effects that six groups of molecules demonstrate a cross-reactivity difference. (I) t-butylnorsynephrine (16) (CR, 571.4%) and octopamine (17) (CR, ND); (II) cimbuterol (47) (CR, 4.4%) and cimaterol (48) (CR, 0.1%); (III) mapenterol (43) (CR, 1.9%) and mabuterol (44) (CR, ND); (IV) terbutaline (11) (CR, 8.3%) and orciprenaline (12) (CR < 0.1%); (V) colterol (7) (CR, 0.2%) and isoproterenol (8) (CR, ND); (VI) tulobuterol (45) (CR, 1.7%) and clorprenaline (46) (CR, ND). The yellow contour near C-2′ implied that a huge steric field on the tertamyl would decrease the cross-reactivity. Thus, it was reasonable that those compounds with bulky rings at C-2′ (e.g., arformoterol (29) (CR < 0.1%), olodaterol (31), indacaterol (32), fenspiride (20), reproterol (22), ractopamine (24), and ritodrine (25) (CR, ND)) demonstrated very low crossreactivity and even no recognition. CoMFA analysis revealed that a stronger hydrophobic field near C-2′ could increase cross-reactivity, but a bigger steric field near C-2′ could decrease cross-reactivity conversely. Clenhexerol (38) contained a bigger and stronger hydrophobic tertiary
Figure 2. ciELISA calibration curve for R-(−)-salbutamol.
and cimaterol, were also included in the 54 compounds.33,34 The cross-reactivity to R-(−)-salbutamol was set as the reference cross-reactivity (100%); it was found that there were 31 compounds that could be recognized by the antibody against R-(−)-salbutamol (Table 1). Comparing the salbutamol hapten structures reported previously,4,30,35,36 it could be found that there were other four salbutamol derivatives with different spacers for the antibody generation, herein named sal-hapten-1,30 sal-hapten2,4 sal-hapten-3,35 and sal-hapten-436 (Figure 1 and Table 1). Both sal-hapten-1 and sal-hapten-2 had a spacer at a same site (C-6), and their spacer difference was one benzene ring on the sal-hapten-2 spacer. However, the antibody against sal-hapten-1 exhibited a poorer recognition to terbutaline and clenbuterol (CR < 0.1%), but the antibody against sal-hapten-2 showed a better recognition (CR, 10 and 107%). It could be inferred that the supplemental benzene ring at the C-6 spacer on sal-hapten2 contributed to the better antibody binding affinity, but both resultant antibodies were still a narrow specific (not class specificity) due to the limited recognition numbers. Different from the same spacer sites of sal-hapten-1 and salhapten-2, the spacer sites of sal-hapten-3 and sal-hapten-4 were derived at C-2′ and C-1, respectively. Sal-hapten-3 possessed two carbons spacer similar to that of sal-hapten-1, and salhapten-4 had a longer spacer. It could be found that both resultant antibodies raised from sal-hapten-3 and sal-hapten-4 recognized only salbutamol, which is still a similar narrow specificity. This suggested that the hapten with a spacer at site C-2′ and C-1 might be not suitable for a class specificity generation. In the present investigation, the R-(−)-sal-hapten spacer was derived at C-2, different from other three hapten spacer sites, C-2′, C-1, and C-6, of the four previously reported sal-hapten 1−4, and the obtained antibody resulted from R-(−)-sal-hapten could recognize 31 compounds within the tested 54 β-agonists and analogues (Table 1). This was the most class specificity to β-agonist compounds reported so far. Based on the comparison of hapten structures and specificities above, it could be found that these four hapten spacer sites, C-2′, C-1, C-6, and C-2, could significantly differ the specificity of the resultant antibody to β-agonist compounds. 3D QSAR Analysis. In order to better understand the effect of hapten spacer sites on the antibody specificity to β-agonists, 2719
DOI: 10.1021/acs.analchem.7b04684 Anal. Chem. 2018, 90, 2716−2724
Article
Analytical Chemistry Table 1. Specificity Comparison of Antibodies Raised from Variously Derived Salbutamol Hapten R-(−)-sal-hapten (this investigation) molecules 3D-SAL t-butylnorsynephrine R-(−)-salbutamol carbuterol Rac-salbutamol DAH clenpenterol clenhexerol terbutaline salbutamol methyl ether R-(−)-TER cimbuterol bromoclenbuterol brombuterol clenbuterol bitolterol mapenterol S-(+)-salbutamol tulobuterol pirbuterol DIM pindolol S-(+)-TER HYD colterol 5-hydroxy salbutamol cimaterol salmeterol penbutolol orciprenaline propafenone arformoterol alprenolol ractopamine mabuterol clorprenaline olodaterol indacaterol propranolol zilpaterol fenspiride epinephrine clenproperol procaterol bambuterol norepinephrine isoproterenol hexoprenaline reproterol vilanterol octopamine ritodrine salbutamon BEN
h-130a
−1
IC50 (nmol L ) 0.07 0.3 1.6 4.4 5.1 6.8 12.5 16.5 19.3 27.3 33.5 36.4 38.9 50.5 54.5 60.5 84.8 91.7 94.0 116.3 123.4 301.3 436.7 551.7 643.6 905.7 1178.5 1258.4 1464.7 2494.8 3111.2 3449.4 6334.5 NDc ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND
h-24a
h-335a
h-436a
100
