Molecular Characterization of Monoclonal Antibodies against

May 2, 2012 - ... against Aflatoxins: A Possible Explanation for the Highest Sensitivity .... Yoon , Yoon-Jee Heo , Daesang Lee , Sung-Gun Kim , Jin-H...
7 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Molecular Characterization of Monoclonal Antibodies against Aflatoxins: A Possible Explanation for the Highest Sensitivity Xin Li,†,‡,§ Peiwu Li,*,†,‡,§,⊥,¶ Qi Zhang,*,†,§ Yuanyuan Li,†,‡,§ Wen Zhang,‡,⊥,¶ and Xiaoxia Ding†,‡,⊥,¶ †

Oil Crops Research Institute of the Chinese Academy of Agricultural Sciences, Wuhan, 430062, P.R. China Key Laboratory of Biology and Genetic Improvement of Oil Crops, Ministry of Agriculture, Wuhan, 430062, P.R. China § Key Laboratory of Detection for Mycotoxins, Ministry of Agriculture, Wuhan, 430062, P.R. China ⊥ Laboratory of Risk Assessment for Oilseeds Products (Wuhan), Ministry of Agriculture, Wuhan, 430062, P.R. China ¶ Quality Inspection and Test Center for Oilseeds and Products, Ministry of Agriculture, Wuhan, 430062, P.R. China ‡

ABSTRACT: We screened and established seven hybridoma cell lines that secrete anti-aflatoxin monoclonal antibodies with different sensitivities. Among these antibodies, 1C11 exhibited the highest sensitivity against all four major kinds of aflatoxins (AFB1, AFB2, AFG1, and AFG2) (IC50 0.0012−0.018 ng mL−1 in the enzyme linked immunosorbent assay (ELISA) system, visual limit of detection of 0.03−0.25 ng mL−1). To better understand the interactions between these antibodies and aflatoxins, as well as to guide their potential sensitivity improvement in recombinant antibodies, we used multiple sequence alignment and molecular modeling combined with molecular docking to clarify the molecular mechanism of the highest sensitivity of 1C11 against aflatoxins. Our results show that hydrogen bond and hydrophobic interaction formed by Ser-H49 and Phe-H103 in the antibody with the hapten played the most important roles in determining the binding affinity. Further experiments performed on antibody mutants, designed on the basis of the computational models, supported the prediction of the interaction mode between the antibody and the hapten. Although the factors that influence antibody sensitivity are highly interdependent, our experimental and modeling studies clearly demonstrate how structural differences influence the binding properties of antibodies against the target hapten with different sensitivities. flatoxins, a group of highly toxic secondary metabolites produced by Aspergillus f lavus and A. parasiticus, contaminate a vast array of foods and agricultural commodities, particularly maize, groundnuts, oilseeds, and tree nuts in tropical and subtropical regions of the world. These metabolites are strong hepatotoxins that are internationally recognized as carcinogens.1 To date, numerous techniques have been developed to detect aflatoxins in foods [e.g., thin-layer chromatography, high-performance liquid chromatography (HPLC), overpressured-layer chromatography, immunoaffinity chromatography (IAC), and near-infrared spectroscopy].2 Immunochemical analyses are widely employed to screen food contaminants because of their rapidity, selectivity, and sensitivity. Most of these methods (e.g., enzyme linked immunosorbent assay (ELISA) and IAC-HPLC) require highquality antibodies.3,4 Over the past 10 years, a wide range of monoclonal antibodies (mAbs) against aflatoxins has been described,5−8 differing in terms of their hapten binding affinities. High-affinity binding between the hapten and the antibody is very important for the sensitivity in antigen detecting assays. With the advances in antibody technologies, recombinant antibodies have been constructed for the detection of aflatoxins.9,10 The recombinant antibody technology could result in low-cost antibodies and facile manipulation of antibody affinity and specificity.11 In

A

© 2012 American Chemical Society

addition, antibody affinity could be improved in vitro by introducing mutations at specific positions in the wild-type antibody (original antibody). A combined strategy involving multiple sequence alignment, homology modeling, and molecular docking was recently developed to predict protein function.12 This strategy can also be used to identify intermolecular interactions responsible for the binding of mAbs to target antigens,13 which will be advantageous to provide clues to improve the binding ability of recombinant antibodies through genetic engineering. Our laboratory has an extensive library of hybridoma cells that secrete mAbs against aflatoxins.2,7,14 Among these cells, one clone (1C11) exhibited the highest sensitivity to all four major kinds of aflatoxins.14 In the present study, generic mAbs against aflatoxin with different sensitivities were selected, and the molecular characterization and binding properties of these mAbs were investigated. The variable region (V) genes encoding these seven mAbs were cloned and sequenced. Homology modeling and molecule docking were further performed to study the antibody−antigen interactions. Our Received: October 17, 2011 Accepted: May 2, 2012 Published: May 2, 2012 5229

dx.doi.org/10.1021/ac202747u | Anal. Chem. 2012, 84, 5229−5235

Analytical Chemistry

Article

aflatoxin. The plates were rewashed with PBST and IgG-HRP (diluted 1/5000 in MPBST, 100 μL well−1), added, and then incubated for 45 min at 37 °C. After six washes, color development was carried out by adding freshly prepared substrate solution [9.5 mL of phosphate-citrate buffer (pH 5.0), 0.5 mL of 2 mg mL−1 3,3′,5,5′-tetramethylbenzidine (TMB), and 32 μL of 3% (w/v) urea hydrogen peroxide] and incubating at 37 °C for 15 min. The enzyme catalyzed reaction was stopped by addition of 50 μL of 2 mol L−1 H2SO4 to each well. Absorption values at 450 nm were measured with a microplate reader. The data obtained were used to create a plot of an inhibition curve, expressed as binding ratio (% B/B0) versus aflatoxin concentration.15 The curve was plotted by measuring each standard in five replicates and fitting them to a four-parameter logistic equation.16 The immunoglobulin isotype of each mAb was identified using a mouse antibody ISO2-1 kit. Cloning of the Variable Region Genes of the mAbs. Total RNA was isolated from hybridoma cells (107 mL−1) using TRIzol reagent. The first-strand cDNA synthesis was performed using RNA with a Superscript Reverse Transcriptase Kit and an oligo(dT)15 primer according to the manufacturer’s protocol (Invitrogen). Mouse immunoglobulin variable heavy chain (VH) and light chain (VL) genes for each mAb were amplified using the primers in the RPAS Mouse scFv Module. The following polymerase chain reaction (PCR) condition was used to amplify the VH gene: 30 cycles at 94 °C for 30 s, 65 °C for 30 s, and 72 °C for 40 s, followed by a final extension at 72 °C for 7 min. The following PCR condition was used to amplify the VL gene: 25 cycles at 94 °C for 30 s, 65 °C for 1 min, and 72 °C for 1 min, followed by a final extension at 72 °C for 7 min. The PCR products of the VH and VL genes were separately cloned into the pMD18-T vector and transformed into E. coli DH5α. Three to five clones were chosen for DNA sequencing in the forward and reverse directions. Variable Region Genes Sequences Analysis. The DNA sequences of the cloned heavy-chain variable-region genes were analyzed for homology with germ-line variable (VH), diversity (DH), and joining (JH), while the cloned light-chain variableregion genes were analyzed for homology with germ-line variable (VL) and joining (JK). To identify putative complementarity-determining region (CDR) loops for the variable region genes reported in the present study, the derived sequences were compared using the Web-based program IGMT/V-QUEST (http://www.imgt.org/IMGT_vquest/).17 All sequences were aligned using the AlignX function in Vector NTI Suite 9 from Informax 200318 and were deposited in GenBank (accession numbers JN109176−JN109182 and JN558712). Modeling the Fv Regions of mAbs. Molecular modeling of the three-dimensional (3-D) structures of the Fv regions of the mAbs 1C11 and 4F12 was performed using MODELER modeling software.19 BLASTP searches in the PDB protein sequence were performed using the VH and VL sequences of antibodies. The 1C11 and 4F12 structures were generated using MODELER based on structural alignment of the templates. Structures from PDB with the following accession codes were used as templates in the modeling: 1aj7, 1c5c, 1jnn, 1kel, 1q72, 1ub5, 1um4, 1wcb, 2cgr, 2g2r, 3fct, 1ar1, and 32c2. For validation and quality assessment, residue-averaged discrete optimized protein energy (DOPE) scores were used to evaluate local regions with high, unfavorable scores that tend to correspond to errors. The models were then subjected to

goal was to determine the molecular mechanisms of the highest-sensitivity of mAb 1C11 against aflatoxins to help us search for “hot spot” residues to generate anti-aflatoxin antibodies with higher sensitivity.

■

MATERIALS AND METHODS Murine Hybridoma Cell Lines. The hybridoma cell lines used in this study were selected from our previous studies.2,14 Hybridoma cultures were grown in complete medium [RPMI1640 medium supplemented with 20% (v/v) fetal calf serum, antibiotics (100 U mL−1 penicillin, 100 μg mL−1 streptomycin), and N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES, free acid, 283.3 g L−1)]. Materials. RPMI-1640 medium containing L-glutamine and HEPES (free acid, 283.3 g l−1) were obtained from HyClone (Logan, UT, USA). Fetal calf serum, penicillin (10 000 U mL−1), and streptomycin (10 000 μg mL−1) were obtained from Gibco (Grand Island, NY, USA). AFB1, AFB2, AFG1, AFG2, and AFM1 standard solutions (3 μg mL−1 in acetonitrile/water, 98:2) were from Supelco (Bellefonte, PA, USA). Aflatoxin B1-BSA (AFB1-BSA) conjugate and goat antimouse immunoglobulin horseradish peroxidase (IgG-HRP) were purchased from Sigma (St. Louis, MO, USA). Cell culture flasks were obtained from Wuxi NEST Biotechnology. Co., Ltd. (Shanghai, China). Microtiter plates (96-well ELISA) were obtained from Costar Inc. (Cambridge, MA, USA). Taq DNA polymerase, DL2000 DNA marker, and pMD18-T vector were purchased from Takara Biotechnology (Dalian, China). TRIzol reagent and oligo(dT)15 primer were obtained from Invitrogen (Beijing, China). Superscript reverse transcriptase kits were purchased from Promega (Madison, WI, USA). Kits for gel extraction and plasmid purification were acquired from Tiangen Biotech Co. (Beijing, China). Heavy chain and light chain primers for the RPAS Mouse scFv Module, pCANTAB 5E vector, anti-M13 monoclonal antibody/horseradish peroxidase conjugate (anti-M13/HRP), M13KO7 helper phage, and E. coli TG1 were purchased from GE Healthcare Bio-Sciences (Shanghai, China). In-Fusion Advantage PCR Cloning Kit was purchased from Clontech (Boston, MA, USA). QuikChange Lightning site-directed mutagenesis kit was obtained from Stratagene (La Jolla, CA, USA). Six-well bacterial growth plates were purchased from Nunc (Roskilde, Denmark). All other reagents were of analytical grade. Evaluation of the mAbs against Aflatoxins. The characterization and application of the seven mAbs have been previously described.2,14 Indirect competitive ELISA was used to evaluate the sensitivity of each mAb and cross-reactivity (CR) with AFB1, AFB2, AFG1, AFG2, and AFM1. Briefly, 96well Costar plates were coated with 100 μL of AFB1−BSA conjugate (2 μg mL−1) in 0.025 M carbonate buffer (pH 9.6) at 4 °C overnight, then washed three times with phosphatebuffered saline (PBS) containing 0.05% Tween 20 (v/v) (PBST), and blocked with 4% Marvel milk powder in PBST (MPBST). The MPBST-blocked blank wells were used as a negative control. Subsequently, 50 μL of AFB1, AFB2, AFG1, and AFG2 at different working concentrations [the aflatoxin standard solution was evaporated to dryness and redissolved in methanol to a final concentration of 1 μg mL−1, respectively, and then diluting further with 0.01 M PBS (pH 7.4) to a series of concentrations] was mixed with the same volume of different mAbs dilutions (which gave an absorbance close to 1.0) and incubated for 1 h at 37 °C. Each competition reaction was carried out in duplicate with at least eight concentrations of 5230

dx.doi.org/10.1021/ac202747u | Anal. Chem. 2012, 84, 5229−5235

Analytical Chemistry

Article

were added to each well of a 96-well plate as the first antibody; for the secondary antibody reaction, horseradish peroxidase (HRP)-conjugated anti-M13 monoclonal antibody diluted at 1/ 5000 in MPBST was added.

energy minimization using assisted model building with energy refinement (AMBER 11).20 Force field ff99SB was used.21,22 We used the steepest descent method for the first 1000 steps and then switched to the conjugate gradient method until the convergence. Molecular Docking of mAbs with Haptens. The AFB1 structures were built from its known structure. The AFB1 ligand was manually docked into the binding site based on several antibody−hapten complex structures to maximize hydrophobic interactions between the antibody and the hapten. The complex structures were minimized using AMBER 11,20 and force field ff99SB was applied. Minimization at each stage was performed using 200 steps of steepest descent and 1500 steps of conjugate gradient algorithms.23 The minimization was mainly to release the bad contact between the antibody and hapten. The antibody−ligand complexes generated were subjected to further analysis. Site-Directed Mutagenesis Experiments. Construction of Phage scFv Antibodies of 4F12 and 10G4. The cloned VH and VL DNA fragments of 4F12 and 10G4, respectively, were assembled into a full-length scFv DNA fragment comprising a single polypeptide using a linker−primer mix encoding (Gly4Ser)3 peptide. The assembled fragment was reamplified with the primers containing ends of 15 bases homologous to the phage-display vector pCANTAB-5E. The full-length scFv DNA fragments were gel-purified and cloned into the vector pCANTAB-5E using In-Fusion Advantage PCR Cloning Kit according to the manufacturer’s instructions (Clontech). The ligation product was transformed into E. coli TG1. All transformed cells were selected on the 2× YTAG plate (2× YT medium containing 2% glucose and 100 μg mL −1 ampicillin). Site-Directed Mutagenesis of scFv Antibody. Mutations of Ser-H49 in H-CDR2 and Phe-H103 in H-CDR3 of both 4F12 and 10G4 were generated using the QuikChange Lightning Site-Directed Mutagenesis Kit (Stratagene). Briefly, we used the constructed scFv plasmid DNA as the template and primers with specific base-pair alterations. Mutations were made by PCR with Pfu DNA polymerase for replication fidelity. The PCR product was treated with DpnI endonuclease to digest the parental DNA template and then was transformed into E. coli XL10-Gold ultracompetent cells (Stratagene). The DNA sequences of the mutants were confirmed by DNA sequencing. Evaluation of the Constructed scFv Antibodies. All of the wild type and mutants of both 4F12 and 10G4 scFv antibodies were expressed and displayed at the tip of the M13 phage, and their sensitivity against AFB1 was assayed by phage ELISA. Briefly, E. coli single colonies from different plates consisting of each different recombinant scFv phagemid were inoculated into 200 μL 2× YTAG medium (100 μg mL−1 ampicillin and 2% glucose) in 1.5 mL Eppendorf tubes and incubated in a shaker overnight at 30 °C. Small inocula (50 μL) were transferred to a new 24-well bacterial growth plate containing 1.5 mL 2× YTAG medium containing 109 plaque forming unit (pfu) of M13KO7 helper phage and cultured at 37 °C in a shaker at 250 rpm. After a 2 h incubation, the 24-well bacterial growth plate was centrifuged at 1300g for 20 min and the supernatant was carefully removed. Bacterial pellets were resuspended in 1.5 mL 2× YTAK(100 μg mL−1 ampicillin, 50 μg mL−1 kanamycin) and cultured overnight at 37 °C. The next day, the bacterial culture plate was centrifuged and supernatants were collected. Indirect competitive phage ELISA was performed as described above, except that 50 μL supernatants of the scFv phage lysates

■

RESULTS AND DISCUSSION Evaluation of the mAbs against Aflatoxins. The isotype and sensitivity data for the seven mAbs are shown in Table 1.

Table 1. Sensitivity of the Seven Selected mAbs with Five Major Aflatoxins IC50 in ELISAa (ng mL−1) mAb

isotype

AFB1

AFB2

AFG1

AFG2

AFM1

1C11 4F12 4F3 1D3 10G4 10C9 1E10

IgG2a IgG2a IgG1 IgG1 IgG1 IgG1 IgG2b

0.0012 0.09 0.29 0.41 0.73 2.09 2.42

0.0013 0.10 0.17 0.77 0.54 2.23 -b

0.0022 0.10 0.14 0.36 0.47 2.19 5.44

0.018 0.42 0.50 2.53 1.46 3.21 -

0.013 0.20 0.27 1.25 1.44 2.95 -

a

The IC50 values for selected mAbs against aflatoxins were estimated by indirect competitive ELISA.2,14 bNot determined.

The antibodies exhibited different sensitivity against AFB1 and good cross-reactivity. The IC50 values against AFB1 ranged from 0.0012 to 2.42 ng mL−1. The seven mAbs could be broadly classified into three distinct groups. 1C11 alone was in group 1 because it displayed the exceptionally highest sensitivity compared with the other mAbs. Group 2 was composed of 4F12, 4F3, 1D3, and 10G4, whose sensitivities against aflatoxin were approximately 100-fold lower than that of 1C11. Group 3 was composed of the remaining mAbs (10C9 and 1E10), whose sensitivities against AFB1 were approximately 1000-fold lower than that of 1C11. Cloning and Analysis of Variable Region Genes of mAbs against Aflatoxins. V-Region Gene Cloning. VH and VL genes of these seven antibodies were amplified by PCR using primers specific for the mouse variable region of each chain. A product of the expected size was obtained for each clone, with a single band ranging in size from 309 to 356 bp for VH and 332 bp for VL. The derived sequences were confirmed by sequence analysis of three to five clones in the forward and reverse directions. The results obtained from each clone were in good agreement. The sequences were further analyzed to find the gene families of these mAbs. The results indicated that the V-region repertoire was restricted to a few relatively highly related VL and VH gene families. All cloned VL genes were from the same gene family (IGKV3), and all VH sequences were from the same germ-line gene family (IGHV5). These sequence similarities are likely correlated with similar antibody-binding specificity. Analysis of the Deduced Amino Acid Sequences. According to the deduced amino acid sequences, all seven VL genes cloned were identical. In contrast, VH sequences differed from the others. These results indicated that the differences in the sensitivity of mAbs against aflatoxin may result from the diversities in VH sequences. Figure 1 shows the alignment of the deduced amino acid sequences of the cloned VH sequences. In our studies, we consider the VH gene of mAb 1C11 as nonmutated (100% identical to their putative germ-line genes) because it displayed the highest sensitivity against aflatoxin 5231

dx.doi.org/10.1021/ac202747u | Anal. Chem. 2012, 84, 5229−5235

Analytical Chemistry

Article

Figure 1. Multiple alignment of the deduced amino acid sequences of the variable regions of the heavy (VH) chains of mAbs against aflatoxins. The alignment was performed using the AlignX function in Vector NTI Suite 9 from Informax 2003. Deleted residues are represented by a dash.

(IC50 = 0.0012 ng mL−1). Its percentage identity with other genes is listed in Table 2 as calculated from pairwise AlignX

structure of the variable fragment (Fv) domains of 1C11 and 4F12 using MODELER based on the determined amino acid sequences of both mAb chains. Visual inspection revealed that the homology model was in good agreement with the characteristic immunoglobulin fold adopted by antibody Fv regions, being composed exclusively of antiparallel β-sheets connected by loops, including those that form the CDRs.24,25 As shown in Figure 2, the differences between the structures of

Table 2. Comparison of Assigned Germline Gene and Amino Acid Sequences with the Respective Expressed Sequences of mAbs against Aflatoxin VH mAb

gene

amino acid

1C11 4F12 4F3 1D3 10G4 10C9 1E10

100a 85.3 89.5 86.1 85.3 70.4 69.2

100 76.9 79.4 76.7 72.4 54.7 54.7

a

Percentage identity, as calculated using the pairwise AlignX function in Vector NTI Suite 9 from Informax 2003. The sequence comparison does not include residues from the primers used to amplify and clone the gene sequences.

alignments. Our results show that the percentage of VH region identity was correlated with the sensitivity against aflatoxin. As shown in Table 2, the sensitivities of mAbs in group 2 (4F12, 4F3, 1D3, and 10G4) against AFB1 were all approximately 100fold lower than that of 1C11; the amino acid sequence of the VH region exhibited high homology compared with 1C11 (>72.4%). The sensitivities of group 3 mAbs (10C9 and 1E10) against AFB1 were 1000-fold lower than that of 1C11, and the amino acid sequence of the VH region exhibited a low homology with 1C11 (54.7%). Hence, the sensitivity diversity of antibodies against aflatoxin may have been related to the differences in sequences of the VH domain. As shown in Figure 1, the sequences of 10C9 and 1E10 (group 3) had the lowest similarities with 1C11, which probably contributed significantly to their lowest sensitivities among those of the seven antibodies against AFB1. However, the reasons for the difference in sensitivity of antibodies in groups 1 and 2 still needed to be identified, so homology modeling and molecular docking analysis were performed to identify the factors affecting the sensitivity of these mAbs. Molecular Docking of 1C11 and 4F12 with the Hapten AFB1. Homology Modeling and Molecular Docking. In the present study, we created homology models for the 3-D

Figure 2. Comparison of the molecular docking models for the 1C11 and 4F12 antibodies generated using AutoDock Vina. (A) Molecular docking model of 1C11 with AFB1 and close-up views of the antibody and hapten interactions. (B) Molecular docking model of 4F12 with AFB1 and close-up views of the antibody and hapten interactions. Amino acid residues involved in interactions with the ligand are displayed as capped sticks. The backbones of the light and heavy chains are shown in red and blue, respectively. Residues numbering 1− 116 stand for VH domain while residues numbering 117−224 stand for VL domain. Hydrogen bonds are shown as green lines. Trp-H101 (Trp-H102 for 4F12) forms a parallel π-stacking interaction with the benzofuranyl ring of AFB1. 5232

dx.doi.org/10.1021/ac202747u | Anal. Chem. 2012, 84, 5229−5235

Analytical Chemistry

Article

the complexes are subtle and the encoded binding pockets have very similar shapes. Both the light (red) and heavy (blue) chains contribute to the antigen-binding site. The binding pocket, a deep cavity, is highly complementary to the hapten AFB1 (Figure 3).

The backbone torsion angles of 193 residues (91.5%) resided in favored regions.Thirteen residues (6.2%) were in allowed regions, and five residues (2.4%) were found in the outlier regions of the plot. Of the five outliers, only Trp102 was predicted to form π interaction with AFB1 according to the docking models. Ramachandran analysis showed that, in the 1C11 structure, the backbone torsion angles of 198 residues (93.8%) resided in favored regions. Ten residues (4.8%) were in allowed regions and three residues (2.4%) were found in the outlier regions of the plot. Of the three outliers, none of them are expected to be involved in interactions with AFB1, as predicted by the docking model. Possible Explanation for the Highest Sensitivity of 1C11. The close-up views of the antibody−hapten complex showed that the hapten binding site was mainly surrounded by amino acids from the VH domain of each mAb (residues numbering 1−116 stand for the VH domain while residues numbering 117−224 represent the VL domain). The results revealed that the heavy chain contributed most contacts to the hapten AFB1, supporting our hypothesis that the differences in the sensitivity of antibodies against aflatoxin were mainly dependent on sequences of the VH domains. This conclusion was consistent with the previous study by Davies et al. showing that residues in all three hypervariable regions in the VH domain affect antigen binding.26 The interaction of antibodies with AFB1 in the docking model was characterized to define the molecular basis for the observed difference in sensitivity between 1C11 and 4F12. As shown in Figure 2, interactions common to these two complexes included hydrogen bond interactions between AspL217 of CDR L3 and the methoxyl group of AFB1, parallel πstacking of the benzofuranyl ring of AFB1 with the aromatic side chain of Trp-H101 of CDR H3 (Trp-H102 for 4F12), and additional hydrophobic interactions involvingTyr-H57 (TyrH58 for 4F12). Differences in binding affinity can be explained by the number of hydrogen bonds formed between the hapten AFB1 and different antibodies. The mAb 1C11, which had the highest sensitivity against AFB1 (IC50 = 0.0012 ng mL−1), apparently formed one more hydrogen bond than 4F12 in the molecular docking model. In 1C11, the Ser-H49 residue in H-CDR2 formed a hydrogen bond with AFB1, while the Thr-H49 residue at the same position in 4F12 formed a hydrogen bond with Ser-H34 (Figure 2). Table 3 shows the deduced amino acid sequences of the CDR loops of the VH domains of the mAbs mentioned above. As shown in the table, the residue at

Figure 3. The binding pocket of molecular docking model with AFB1: (A) 1C11 and (B)4F12. The figure was generated in UCSF chimera.27

Ramachandran analysis using the MolProbity Web site showed that the 4F12 structure, the backbone torsion angles of 193 residues (91.5%) resided in favored regions. Thirteen residues (6.2%) were in allowed regions and five residues (2.4%) were found in the outlier regions of the plot.

Table 3. Affinities of Selected mAbs to AFB1 and Amino Acid Sequences of the CDR Loops in the VH Domain mAb 1C11 4F12 4F3 1D3 10G4 10C9 1E10

IC50 (AFB1)a 0.0012 0.086 0.286 0.411 0.730 2.092 2.420

H-CDR1 sequenceb

H-CDR2 sequenceb

H-CDR3 sequenceb

S I-SGGG-YSYYPDSVKG TISSGGGKNYYPDSMKG ---ST-GGSLTYYPDSVKG TI-SGGGGFTYYPDSVKG TISSGGGKNYYPDSMKG YISS--GSSTLHYADTVKG YISS--GSSTLHYADTVKG

ASHDYAWSFGV VRHNYRWTMDY ARHDFHWSFDV ARHRGTNYYFDY VRHNYRWTMDY NDc ND

d

GFTFSNYG GFSLSSYA GFTFSSYA GFTFSSYG GFSLSSYA GFTFSSFG GFTFSSFG

a Determined by indirect competitive ELISA, as described previously.2,14 bH-CDR1, H-CDR2, and H-CDR3 are defined according to IGMT/VQUEST. Nonconserved amino acids are underlined. cNot determined. dThe underlined residues in 1C11 are Ser-H49 and Phe-H103, which make important contacts with AFB1 in the docking model. The underlined residues in 4F12 are Thr-H49 and Met-H104, which make different contacts with AFB1 in the docking model, compared with 1C11. The underlined residues in other sequences are at the same positions as Ser-H49 and PheH103 in 1C11.

5233

dx.doi.org/10.1021/ac202747u | Anal. Chem. 2012, 84, 5229−5235

Analytical Chemistry

Article

experimental data were qualitatively consistent with the computational predictions, indicating that the side chains of Ser-H49 and Phe-H103 played important roles on high-affinity binding interactions while hydrogen bonding and hydrophobic interactions have the most important effect on aflatoxin antibody-binding affinity.

the H49 position of H-CDR2 is nonconserved. According to the mAbs classified as described above, the residue on this position of group 1 (1C11) is Ser, compared with the residue on the same position of group 2, which is Thr. The loss of hydrogen bonding at the antibody active site may result in the loss of binding affinity because even a single hydrogen bond can significantly improve interactions between the ligand and the receptor. Thus, the different amounts of hydrogen bonds formed may be the main factor influencing the sensitivity of antibodies against aflatoxin, while Ser-H49 in H-CDR2 of 1C11 may be a key amino acid that contributes to additional hydrogen bond formation. Due to the large hydrophobic fragment of aflatoxin, hydrophobic interactions might be another important factor to increase the binding affinity between the receptor and the ligand. As shown in the molecular docking models, the PheH103 residue in 1C11, which presents a large nonpolar surface to AFB1, improved the packing interactions between the ligand and the active site. At the same position in 4F12 is Met-H104, whose hydrophobic interaction with the hapten is much weaker than that of Phe-H103. As shown in Table 3, the difference in sensitivity between 1D3 and 10G4 might be partially due to the presence of Phe-H104 in H-CDR3 of 1D3. These findings imply that hydrophobic interactions contribute to the high sensitivity of mAbs against aflatoxin. Corroboration of the Predictions. To confirm the computational prediction, site-directed mutagenesis was performed at H49 and H104. According to the discussion above, single mutation (ThrH49Ser) and double mutation (ThrH49Ser, Met H104Phe) of 4F12 and 10G4 were constructed. For comparison, wild-type 4F12 and 10G4 were expressed and their sensitivities against AFB1 were assayed through phage ELISA at the same time under the same experimental conditions. All of the IC50 results are listed in Table 4. As

■

CONCLUSIONS Several anti-aflatoxin mAbs with different sensitivities are available in our laboratory. To investigate the molecular mechanisms that affect antibody sensitivity, we modeled the complex structure of antibodies and AFB1 by homology modeling and molecular docking. Although the factors that influence sensitivity are highly interdependent, making it difficult to quantify them individually, our experimental and modeling studies clarified how structural differences influenced the binding properties of antibodies against the target hapten with different sensitivities. After analyzing binding interactions such as hydrogen bonding, π-stacking, and hydrophobic interactions, we concluded that hydrogen bonding and hydrophobic interactions played important roles in mediating aflatoxin and antibody interactions. Furthermore, the amino acids Ser-H49 and Phe-H103, involved in the interactions with the hapten AFB1, appear to be the most important amino acids to warrant the high-affinity binding of aflatoxin with the antibodies. The data obtained from the site-directed mutagenesis of the predicted “hot spot” residues showed that all four mutants increase the sensitivity against AFB1 compared to the wild type. Our results demonstrate that multiple sequence alignment combined with homology modeling and molecular docking can be used to provide clues into the interactions between antibody and hapten, which will guide future modifications of antibodies against aflatoxin through genetic engineering.

■

Table 4. Effect of Mutations on Sensitivity against AFB1 of scFv 4F12 and 10G4 mutation

IC50 against AFB1 (ng mL−1)

4F12 (wild-type) 4F12 (ThrH49Ser) 4F12 (ThrH49Ser, Met H104Phe) 10G4 (wild-type) 10G4 (ThrH49Ser) 10G4 (ThrH49Ser, Met H104Phe)

10 1.0 0.5 0.59 0.15 0.03

AUTHOR INFORMATION

Corresponding Author

*Tel.: +86 27 86812943. Fax: +86 27 86812862. E-mail: [email protected] (P.L.); [email protected] (Q.Z.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS X.L. and P.L. contributed equally to this work. We thank Dr. Xiang Mao, Professor of Nanjing Agricultural University, for kindly helping us to finish the molecular docking research. This work was supported by the National Natural Science Foundation of China (31171702), the Project of National Science & Technology Pillar Plan (2012BAB19B09, 2012BAK08B03, 2011BAD02D02, 2010BAD01B07), the Special Fund for Agro-scientific Research in the Public Interest (201203094, 201203037), the earmarked fund for China Agriculture research system (CARS-13), the Key Project of the Ministry of Agriculture (2011-G5, 2010-G1), Special Foundation of President of the Chinese Agricultural Academy of Sciences (2012ZL042,1610172010003).

shown in the table, the mutant of 4F12 (ThrH49Ser) increased the sensitivity against AFB1 by about 10-fold compared to the wild type, while the mutant of 10G4 (ThrH49Ser) increased the sensitivity against AFB1 by about 4-fold compared to the wild type. The double mutants (ThrH49Ser, Met H104Phe) of both 4F12 and 10G4 increased the sensitivity against AFB1 by about 20-fold compared to the wild type. The sensitivity against AFB1 of the constructed 1C11 scfv was also assayed through phage ELISA. The IC50 of 1C11 scfv was 0.02 ng mL−1, almost the same as with the mutant 10G4 (ThrH49Ser, Met H104Phe). The result indicates that the mutated amino acids, Ser and Phe, were key amino acid influencing binding. However, the IC50 of the mutant 4F12 (ThrH49Ser, Met H104Phe) did not restore sensitivity to that of 1C11, perhaps owing to other unknown factors which resulted in tiny differences in the 3-D structures of the Fv regions between mutant 4F12 (ThrH49Ser, Met H104Phe) and 1C11. Thus, the

■

REFERENCES

(1) Kensler, T. W.; Roebuck, B. D.; Wogan, G. N.; Groopman, J. D. Toxicol. Sci. 2011, 120, 28−48. (2) Li, P.; Zhang, Q.; Zhang, W.; Zhang, J.; Chen, X.; Jiang, J.; Xie, L.; Zhang, D. Food Chem. 2009, 115, 313−317.

5234

dx.doi.org/10.1021/ac202747u | Anal. Chem. 2012, 84, 5229−5235

Analytical Chemistry

Article

(3) Fisher, A. C.; DeLisa, M. P. J. Mol. Biol. 2009, 385, 299−311. (4) Lauer, B.; Ottleben, I.; Jacobsen, H. J.; Reinard, T. J. Agric. Food Chem. 2005, 53, 899−904. (5) Kolosova, A. Y.; Shim, W.-B.; Yang, Z.-Y.; Eremin, S.A..; Chung, D.-H. Anal. Bioanal. Chem. 2006, 384, 286−294. (6) Cervino, C.; Weber, E.; Knopp, D.; Niessner, R. J. Immunol. Methods 2008, 329, 184−193. (7) Zhang, J.; Li, P.; Zhang, W.; Zhang, Q.; Ding, X.; Chen, X.; Wu, W.; Zhang, D. Hybridoma 2009, 28, 67−70. (8) Min, W.-K.; Kweon, D.-H.; Park, K.; Park, Y.-C.; Seo, J.-H. Food Chem. 2011, 126, 1316−1323. (9) Moghaddam, A.; Løbersli, I.; Gebhardt, K.; Braunagel, M.; Marvik, O. J. J. Immunol. Methods 2001, 254, 169−181. (10) Daly, S.; Dillon, P.; Manning, B.; Dunne, L.; Killard, A.; Kennedy, R. Food Agric. Immunol. 2002, 14, 255−274. (11) Yang, L.; Ding, H.; Gu, Z.; Zhao, J.; Chen, H.; Tian, F.; Chen, Y.; Zhang, H.; Chen, W. J. Agric. Food Chem. 2009, 57, 8927−8932. (12) Pierri, C. L.; Parisi, G.; Porcelli, V. Biochim. Biophys. Acta 2010, 1804, 1695−1712. (13) Lape, M.; Paula, S.; Ball, W. J., Jr. Eur. J. Med. Chem. 2010, 45, 2291−2298. (14) Zhang, D.; Li, P.; Zhang, Q.; Zhang, W.; Huang, Y.; Ding, X.; Jiang, J. Anal. Chim. Acta 2009, 636, 63−69. (15) Anfossi, L.; Calderara, M.; Baggiani, C.; Giovannoli, C.; Arletti, E.; Gianfranco, G. J. Agric. Food Chem. 2008, 56, 1852−1857. (16) Ngundi, M. M.; Shriver-Lake, L. C.; Moore, M. H.; Lassman, M. E.; Ligler, F. S.; Taitt, C. R. Anal. Chem. 2005, 77, 148−154. (17) Ho, M.; Segre, M. Biochim. Biophys. Acta 2001, 1521, 135−140. (18) Elansary, O. M. H.; Adamec, L.; Storchova, H. Aquat. Bot. 2010, 92, 214−220. (19) Eswar, N.; Eramian, D.; Webb, B.; Shen, M.-Y.; Sali, A. Methods Mol. Biol. 2008, 1, 145−159. (20) Case, D. A.; Darden, T. A.; Cheatham, T. E.; et al. AMBER 11; University of California: San Francisco, 2010. (21) Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.; Simmerling, C. Proteins 2006, 65, 712−725. (22) Long, D.; Li, D.-W.; Walter, K. F. A.; Griesinger, C.; Bruschweiler, R. Biophys. J. 2011, 101, 910−915. (23) Aparoy, P.; Leela, T.; Reddy, R. N.; Reddanna, P. J. Mol. Graphics Modell. 2009, 27 (6), 744−750. (24) Bork, P.; Holm, L.; Sander, C. J. Mol. Biol. 1994, 242, 309−320. (25) Schiffer, M.; Girling, R. L.; Ely, K. R.; Edmundson, A. B. Biochemistry 1973, 12, 4620−4631. (26) Davies, J.; Riechmann, L. Immunotechnology 1996, 2, 169−179. (27) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. J. Comput. Chem. 2004, 25 (13), 1605−1612.

5235

dx.doi.org/10.1021/ac202747u | Anal. Chem. 2012, 84, 5229−5235