Selection of single chain variable fragment (scFv) antibodies from a hyperimmunized phage display library for the detection of the antibiotic monensin

Journal of Immunological Methods 360 (2010) 103–118

Contents lists available at ScienceDirect

Journal of Immunological Methods j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j i m

Research paper

Selection of single chain variable fragment (scFv) antibodies from a hyperimmunized phage display library for the detection of the antibiotic monensin Shokouh Makvandi-Nejad a, Claudia Sheedy b, Linda Veldhuis a, Gabrielle Richard a, J. Christopher Hall a,⁎ a b

School of Environmental Sciences, University of Guelph, Guelph, ON, Canada N1G 2W1 Agriculture and Agri-Food Canada, Lethbridge Research Center, 5403 1st Avenue South, P.O. Box 3000, Lethbridge, AB, Canada T1J 4B1

a r t i c l e

i n f o

Article history: Received 26 February 2010 Received in revised form 14 June 2010 Accepted 17 June 2010 Available online 2 July 2010 Keywords: Monensin Antibiotics Single chain variable fragment Site-directed mutagenesis Fluorescence polarization

a b s t r a c t Concerns over the occurrence of the veterinary antibiotic monensin (MW 671 Da) in animal food products and water have given rise to the need for a sensitive and rapid detection method. In this study, four monensin-specific single chain variable fragments (scFvs) were isolated from a hyperimmunized phage-displayed library originating from splenocytes of a mouse immunized with monensin conjugated to bovine serum albumin (BSA). The coding sequences of the scFvs were engineered in the order 5′-VL-linker-VH-3′, where the linker encodes for Gly10Ser7Arg. Three rounds of selection were performed against monensin conjugated to chicken ovalbumin (OVA) and keyhole limpet hemocyanin (KLH), alternately. In the third round of selection, two different strategies, which differed in the number of washes and the concentration of the coating conjugates, were used to select for specific binders to monensin. A total of 376 clones from round two and three were screened for their specific binding to monensin conjugates and positive clones were sequenced. It was found that 80% of clones from round three contained a stop codon. After removing the stop codon by sitedirected mutagenesis, ten binders with different amino acid sequences were subcloned into the vector pMED2 for soluble expression in Escherichia coli HB2151. Four of these scFvs bound to free monensin as determined using competitive fluorescence polarization assays (C-FPs). IC50 values ranged from 0.031 and 231 μM. A cross-reactivity assay against salinomycin, lasalocid A, kanamycin and ampicillin revealed that the two best binders were highly specific to monensin. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Monensin (MW 671 Da) belongs to the family of antibiotics known as carboxylic ionophore polyethers. The exterior makeup of this molecule is highly hydrophobic (Fig. 1), thus even salts of monensin have very low solubility in water and relatively high solubility in organic solvents such as acetone,

⁎ Corresponding author. School of Environmental Sciences, University of Guelph, Bovey Building, 50 Stone Rd. E, Guelph, ON, Canada N1G 2W1. Tel.: + 1 519 824 4120x52740; fax: + 1 519 837 0442. E-mail address: [email protected] (J.C. Hall). 0022-1759/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2010.06.015

methanol and chloroform (Pinkerton and Steinrauf, 1970). The antibacterial and antimalarial properties of monensin are attributed to its ability to transport Na+, K+ and protons across the lipid membrane of cells, which affects the net accumulation of protons, and results in a decrease of the intracellular pH (Pressman and Fahim, 1982). Monensin is regularly used as an additive in poultry feed for the prevention and treatment of coccidiosis caused by Eimeria necatrix, E. tenella, E. acervulina, E. brunetti, E. mitis, and E. maxima (Canadian Food Inspection Agency). It is also used in the beef and dairy industries to prevent coccidiosis caused by E. bovis and E. zuernii and as a growth promotant to increase the rate of weight gain, to minimize loss of body tone during lactation in dairy cows, and

104

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

polarization (FP) assay to characterize the binding properties of the isolated scFv antibodies specific to monensin. Some of the problems associated with the isolation of scFvs against hydrophobic haptens are also discussed. 2. Methods and materials 2.1. Synthesis of monensin conjugates

Fig. 1. Chemical structure of monensin (MW 671 Da).

to improve milk production in lactating dairy cows (Canadian Food Inspection Agency). Various dosages of monensin can be toxic to a number of animals. In mice, the oral LD50 of monensin is 44 mg kg−1 (Pinkerton and Steinrauf, 1970). Monensin is also highly toxic to animals such as horses and dogs (Canadian Food Inspection Agency). Due to the frequent administration of monensin as a growth promotant in livestock, this antibiotic was detected in surface waters at seven different agricultural sites in Ontario (Hao et al., 2006). This has given rise to the need for a sensitive assay to detect this antibiotic in water samples to monitor and minimize the possibility of the environmental contamination with monensin. There are two main challenges associated with producing high affinity antibodies and subsequently developing a sensitive assay against a hydrophobic hapten such as monensin. The first challenge lies in their lack of immunogenicity due to their small size. To overcome this problem, haptens have been covalently conjugated to larger molecules such as bovine serum albumin (BSA). However, immunization of an animal with a conjugated hapten could result in production of a heterogeneous pool of antibodies against the target hapten, conjugate, and carrier protein. Since haptens occupy a very small fraction of the immunogen, most of the B cells are triggered by the conjugate and carrier protein and only a few are activated by the haptens (Kramer, 2002). The second challenge with producing antibodies against hydrophobic haptens relies in the difficulties occurring during the conjugation reaction in which hydrophobic haptens may react with hydrophobic domains of carrier proteins. As a result of this reaction, haptens “hide” within the protein structure and thus are not exposed to B cells in the host animal or to antibodies during the selection process (Fasciglione et al., 1996). To develop a sensitive and specific detection assay against monensin, we immunized four mice and two rabbits with either monensin-BSA or monensin-KLH (keyhole limpet hemocyanin) to determine the animal with the best immune response to monensin. The splenocytes of a mouse were utilized to construct a phage-displayed scFv library. Two different selection strategies were applied to isolate monensinspecific scFvs from this library by eliminating non-specific binders. We further describe the development of a fluorescence

Monensin was conjugated to three carrier proteins: BSA, OVA and KLH using the method described by Fleeker (1987). In brief, monensin (150 mg, 0.124 mM) was mixed with N-hydroxysuccinimide (50 mg, 0.124 mM) in 2 mL of methanol. To this solution, dicyclohexylcarbodiimide (39 mg, 0.124 mM) in 1 mL of methanol was added, mixed gently and incubated overnight at 22 °C in the dark. Either BSA (200 mg), OVA (ovalbumin; 200 mg) or KLH (200 mg) in 3 mL of borate buffer (0.1 M, pH 9.0) was added drop-wise to the vial with constant stirring. The reaction mixture was stirred for 2 h at 22 °C. The resulting conjugates were dialyzed against 1X phosphate-buffered saline (PBS; 8 g of NaCl, 0.2 g of KCl, 1.44 g of Na2HPO4, and 0.24 g of KH2PO4 per liter of water, pH 7.5) at 4 °C for 24 h. 2.2. Mice and rabbit immunization Four eight-week-old Balb/c female mice and a pair of New Zealand white female rabbits were immunized with either monensin-BSA or monensin-KLH. Each conjugate was diluted in sterile 1X PBS and mixed with an equal volume of TiterMax Classic Adjuvant (Sigma-Aldrich Chemical Co., St. Louis, MO). For primary immunization, each animal was injected subcutaneously with 100 μg mL−1 of either conjugate. Two weeks after the primary injection, booster immunizations were administered at one week intervals. Each animal was injected with 50 μg mL−1 of a monensin conjugate mixed with an equal volume of Freund's incomplete adjuvant (ca. 100 μL; Sigma-Aldrich Chemical Co.). Sera (ca. 30 μL from mice and 500 μL from rabbits) were collected a week after each immunization and the immune response was monitored by ELISA. The last boost took place three days before sacrificing the animals. The immune response was monitored against monensinOVA and free monensin using indirect and competitive indirect (CI) ELISAs, respectively. For the indirect ELISA, the wells of the first row of microtitre plates were coated with 1:2000 (v/v) dilution of monensin-OVA (1 μg mL−1) in 1X PBS, followed by a serial dilution (1:2 v/v across the plate) with 1X PBS. As a negative control, one row was coated with 5 μg mL−1 of OVA. The coated plates were incubated at 4 °C for 16 h. After washing the wells three times with 1X PBS, wells were blocked with 200 μL of 8% MPBS (8 g of skim milk in 100 mL of 1X PBS) for 2 h at room temperature. Wells were washed as described above and serum was diluted (1:50 and 1:100 v/v for mice and rabbits, respectively) in 1X PBS. After adding the diluted sera to the first column, a serial dilution (1:2 v/v across the plate) in 1X PBS was performed. The sera were allowed to bind to the conjugate by incubation for 1.5 h at room temperature. Wells were washed five times with 1X PBST (0.05% Tween 20 in 1XPBS) and incubated for 1 h with 100 μL of either polyclonal goat anti-mouse or goat anti-

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

rabbit antibody conjugated to horseradish proxidase (1:6000 v/v in 1X PBS; Promega, Madison, WI). The signals were detected by adding TMB peroxidase substrate (Bio-Rad, Hercules, CA). After 30 min, reactions were stopped by adding 1 M H2SO4 (100 uL) and absorbance was measured at 450 nm using a microtiter plate reader (Model 3550-UV, Bio-Rad, Hercules, CA). All final volumes are 100 μL unless otherwise mentioned. To verify the specific immune response to free monensin, a CI-ELISA was performed on the third and fifth bleeds collected from mice and the final bleed collected from rabbits. Wells were coated with monensin-OVA (1:8000 v/v in 1 XPBS) and incubated at 4 °C for 16 h. A dilution of serum (1:2000 v/v) was prepared and its binding to monensin-OVA was inhibited with 50 μL of free monensin (concentration range 0 to 1.7 μM) in 1X PBS. The solution containing free monensin was added to the coated wells and incubated for 1.5 h at room temperature. The remaining steps were performed as described above. All samples were run in triplicates and each assay was repeated three times (on different days) for testing its reproducibility. 2.3. Library construction The spleen of mouse B1 was excised and stored in RNAlater (RNA stabilization reagent; Qiagen, Mississauga, ON). The splenocytes were extracted from 25 mg of spleen tissue using a homogenizer. Total RNA was extracted with an RNeasy RNA extraction kit, according to manufacturer's instructions (Qiagen) and complementary DNA (cDNA) was synthesized using the First Strand cDNA Synthesis Kit (AmershamPharmacia Biotech, Piscataway, NJ). To construct scFv coding sequences, the cDNA product was used as a template for the amplification of the variable heavy (VH) and light (VL) domains of the immunoglobulin genes by PCR. The primers in Appendix A were used for amplification of VH and VL genes. SfiI restriction sites were added to 5′ and 3′ ends of VH and VL, respectively, for cloning into the pMED1 vector (Arbabi Ghahroudi et al., 1997). PCR was performed in a total volume of 100 μL, which contained 0.4 μM of dNTPs, 3.17 μM VH sense primer mixture, 0.75 μM VH reverse primer, 4 units of Taq polymerase (Roche, Mississauga, ON) and 5 μL of cDNA. Vκ and Vλ were amplified as described for VH, except the concentrations of Vκ and Vλ sense primers were 0.625 μM and 0.32 μM, respectively. The PCR conditions were as follow: an initial melt of 94 °C for 4 min, followed by 35 cycles of 94 °C for 30 s, 56 °C for 30 s, and 72 °C for 1 min, and a final extension at 72 °C for 7 min. A total of 10 μg of VH, Vκ and Vλ genes were amplified and purified using a QIAquick Gel Extraction Kit (Qiagen). To add the coding sequence of the linker, the first repeat of the G10S7R linker was added to the 3′ end of the VL primers and the last repeat of the G10S7R was introduced at the 5′ end of VH primers (Appendix A). An overlap PCR was used to assemble the coding sequences of VH and VL using primers in Appendix B. The PCR cocktail was prepared as follows: 0.4 mM dNTP, 0.6 μM of RSF (forward) and RSR (reverse) primers (Appendix B), 40 ng of VH, 110 ng of either Vκ or Vλ, and 4 units of Taq polymerase (Roche) in a total volume of 100 μL. The PCR reaction was performed under the following conditions: an initial melt at 94 °C for 1 min, followed by

105

30 cycles of 94 °C for 15 s, 56 °C for 30 s, 72 °C for 2 min, and a final extension at 72 °C for 10 min. The 850-bp amplicons were gel-purified as described previously. To construct the scFv phage-display library, 7 μg of pMED1 vector and 15 μg of scFv inserts were digested with 100 units of SfiI restriction enzyme (New England Biolabs, Ipswich, MA). The digestion reaction was performed at 50 °C for 24 h and digested products purified using a QIAquick PCR Purification Kit (Qiagen). ScFv (10.8 μg) was ligated into 3.6 μg of digested pMED1 and incubated at 16 °C for 24 h using 10 units of T4 DNA ligase (Promega, Madison, WI) in a final volume of 20 μL. The ligated products were electroporated into E. coli TG1 electrocompetent cells. A total of 70 transformation reactions were performed. Each transformation product was allowed to grow for 1 h at 37 °C with shaking at 220 rpm. After pooling all the transformants, they were grown in LB medium containing 100 μg mL−1 of ampicillin and 1% glucose for 16 h at 37 °C with shaking at 250 rpm. To estimate the size of the library, serial dilutions of 10−1 to 10−5 were prepared and plated on LB agar (10 g Bacto-tryptone, 5 g yeast extract, 10 g NaCl, 15 g of agar in 1 L of water, pH 7.4) plates supplemented with 100 μg mL−1 of ampicillin and 1% glucose. These cultures were incubated at 37 °C for 16 h. Colony PCR was performed to verify the percentage of vectors containing the insert, and twenty randomly chosen clones were sequenced to estimate the diversity of the library (CBS-DNA Sequencing Facilities, Guelph, ON). 2.4. Phage amplification and precipitation Library was grown in 200 mL of 2 × YT (16.0 g of tryptone, 10.0 g of yeast extract, 5.0 g of NaCl, pH 7.4) supplemented with ampicillin (100 μg mL−1) and 1% glucose, and incubated at 37 °C until an OD600 of 0.4 to 0.6 was obtained. The culture was superinfected with 1010 helper phage KM13 at 37 °C for 30 min. Subsequently, the bacteria were harvested by centrifugation and resuspended in 100 mL 2 × YT containing ampicillin (100 μg mL−1), kanamycin (50 μg mL−1) and 0.1% glucose and incubated at 30 °C for 16 h. To harvest the phages, overnight culture was spun at 3300 ×g for 30 min. The phages were precipitated with PEG/NaCl (20% polyethylene glycol 6000 in 2.5 M NaCl), incubated for 1 h on ice, and centrifuged at 3300 × g for 30 min. Precipitated phages were suspended in 1 mL 1X PBS and used directly in the selection procedure described below. 2.5. Selection of specific phage-scFvs A subtractive panning was performed to isolate monensinspecific binders. The selection strategy has been summarized in Table 1. Briefly, in each panning round a Maxisorp microtiter tube (Nunc, Rochester, NY) was blocked with the appropriate blocking buffer (Tube 1) and incubated at 4 °C for 16–20 h. Two other immunotubes were coated with either the corresponding carrier protein (Tube 2) or the coating conjugate (Tube 3) and incubated under the same conditions as Tube 1. Subsequently, precipitated phages (500 μL; described in Section 2.4) in blocking buffer were added to Tube 1 and incubated at 4 °C for 16–20 h. Tubes 2 and 3 were blocked with the blocking buffer and incubated at 4 °C for 16–20 h. Phages were

106

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

Table 1 Summary of the selection strategies to isolate specific anti-monensin scFvs. Subtractive panning was performed using either OVA or KLH as coating antigen followed by either monensin-OVA or monensin-KLH, respectively. Rounds two and three were performed twice (a and b; Methods and materials) by modifying the concentration of the coating conjugates, blocking reagents and number of washes. In round 3, the phage-scFvs were eluted using three increasing concentrations of free monensin (i.e., 0.1 μM, 1 μM and 10 μM). Round

Blocking buffer

Coating antigen

Coating antigen concentration (μg mL−1)

Number of washes

Elution reagent

1 2a 2b 3a 3b

4% milk + 2% OVA + 2% BSA 5% milk + 2% KLH + 2%BSA 4% casein + 2% KLH + 2% BSA 2% casein + 2% OVA + 2% BSA 8% milk + 2% OVA + 2% BSA

OVA and monensin-OVA KLH and monensin-KLH KLH and monensin-KLH OVA and monensin-OVA OVA and monensin-OVA

100 35 35 10 1

10 × PBST 20 × PBST 15 × PBST 25 × PBST 25 × PBST

Trypsin (1 mg mL−1) Trypsin (1 mg mL−1) Trypsin (1 mg mL−1) Monensin Monensin

transferred from Tube 1 to Tube 2 and incubated for 1 h at 25 °C. To select for monensin binders, the phage-scFvs were then transferred from Tube 2 to Tube 3 and allowed to bind to the conjugate for 2 h at 25 °C. The unbound phage-scFvs were removed by washing the tubes with 1X PBST followed by 1X PBS (Table 1). The bound phages were eluted with either trypsin (1 mg mL−1) or with one of three concentrations of free monensin, i.e., 0.1 μM (Elution 1), 1 μM (Elution 2) and 10 μM (Elution 3). Each concentration of free monensin was incubated with phage-scFv at 37 °C for 1 h. The eluted phages were used to infect log phase E. coli TG1 cells (OD600 0.4–0.5) at 37 °C for 30 min. The TG1 cells were spread on 2 × YT agar plates containing 100 μg mL−1 of ampicillin and 0.1% glucose followed by incubating at 30 °C for 16–20 h. Colonies were scraped from the plate and resuspended in 2 × YT broth containing ampicillin and glucose. The phages were grown and precipitated as described in Section 2.4. Precipitated phages were used for the next round of selection or in polyclonal phage ELISA (Section 2.6). The precipitated phages form round 1 of panning were used in both rounds 2a and 2b, which was followed by using the precipitated phages from rounds 2a and 2b in rounds 3a and 3b, respectively. Simultaneously, a second selection was performed to estimate the background titer. The procedure was performed exactly the same as described above except the tubes were coated with the appropriate carrier proteins.

2.6. Screening for specific monensin binders Polyclonal phage ELISA was performed using phages collected from each round of panning. Each well of a microtitre plate was coated with either monensin-BSA, monensin-OVA or monensin-KLH diluted 1:200 v/v in 1 XPBS. Some wells were coated with corresponding carrier proteins as negative controls. After incubating the plates at 4 °C for 16–20 h, wells were blocked with 8% MPBS and incubated as described in Section 2.2. Wells were washed three times with 1X PBS, 109 phage were added into the wells in the presence of 8% MPBS, and incubated for 1 h at room temperature. Each plate was washed 15 times with PBST (0.5% Tween 20) and anti-M13 antibody (1:6000 v/v) (GE Healthcare, Piscataway, NJ) was added into each well prior to being incubated for 1 h at room temperature. After washing the wells 12 times with PBS, the binding signals were detected as described in Section 2.2. In monoclonal phage ELISA, 94 clones were selected from each round of panning with the exception of round one. Random colonies were picked and grown in 96-well microtiter plates containing 200 μL/well of LB medium with 100 μg mL−1

and and and and and

10 × PBS 20 × PBS 15 × PBS 25 × PBS 10 × PBS

ampicillin and 1% glucose. After incubating at 37 °C for 16 h, 2 μL of each culture was transferred into 200 μL of LB containing ampicillin and glucose. These cultures were allowed to grow at 37 °C for 2 h. Each culture was infected with 1010 helper phage and incubated at 37 °C for 1 h. Each plate was centrifuged, and the pellet was resuspended in fresh LB medium containing ampicillin (100 μg mL−1) and kanamycin (50 μg mL−1), prior to being incubated at 30 °C for 16 h. After centrifugation of the cultures at 1800 rpm for 30 min, 50 μL of supernatant was used for ELISA. The coding sequences of the scFvs binding to monensin conjugates were determined (CBS-DNA Sequencing Facilities, Guelph, ON).

2.7. Site-directed mutagenesis To remove the stop codon, primers shown in Table 2 were used in a two-step PCR procedure (Wang and Malcolm, 1999). In the first step, the PCR cocktail contained only the forward or reverse primer. Each reaction mix was prepared as follows: 1XPCR buffer, 5 units of PfuTurbo DNA polymerase (Stratagene, La Jolla, CA), 0.4 mM of dNTP, 0.8 μM of either primers and 50 ng of template. The conditions were 95 °C for 1 min, followed by 10 cycles of 95 °C for 1 min, 68 °C for 10.5 min. In the second step, 25 μL of each of the forward and reverse reaction solution were mixed together followed by adding 2 units of PfuTurbo DNA polymerase. This reaction took place under the following conditions: 95 °C for 1 min, (95 °C for 1 min, 68 °C for 2 min) for 20 cycles, 68 °C for

Table 2 The primers used in site-directed mutagenesis to remove the stop codons, which are represented by an asterisk (*) in Fig. 9. The underlined nucleotides are the modified bases. Primer name

Clone

Oligonucleotide sequence

Red-CF

45-100, 46-10, 15-10

5′-GCT GAC ATT GTG ATG ACC CAG TCT CAA AAA TTC ATG TC-3′ 5′-GAC ATG AAT TTT TGA GAC TGG GTC ATC ACA ATG TCG AGC-3′ 5′-CTA ATG TAG CCT GGT ATC AAC AGA AAC CAG GGC AAT C-3′ 5′-GAT TGC CCT GGT TTC TGT TGA CAG GCT ACA TTA G-3′ 5′-GAG CTC GAT ATT AAG ATG ACC CAG TCT CAA AAA TTC ATG TCC A-3′ 5′-TGG ACA TGA ATT TTT GAG ACT GGG TCA TCT TAA TAT CGA CGT C-3′

Red-CR 4-10-CF

4-10

4-10-CR 5-10-CF 5-10-CR

5-10

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

7 min. The parental sequences were digested with 10 units of Dpn I (Stratagene) for 16 h at 37 °C. After purification of the products, they were transferred into E. coli HB2151 electrocompetent cells. DNA sequencing confirmed the removal of stop codons.

2.8. Soluble expression of anti-monensin scFvs For soluble expression, all the ten positive coding sequences were cloned into the expression vector pMED2 (Source: Dr. M. Arbabi Ghahroudi) and transferred into E. coli HB2151 for soluble expression. Each clone was grown in 5 mL of LB medium containing ampicillin and glucose at 37 °C for 16 h. A fresh LB medium (50 mL) containing ampicillin and glucose was inoculated with 500 μL of the overnight culture and grown at 37 °C until the OD600 reached 0.7–0.8. The cultures were centrifuged at 5000 rpm for 30 min and the pellets resuspended in 50 mL of fresh LB containing 100 μg mL−1 of ampicillin and 1 mM of isopropyl β-D-1thiogalactopyranoside (IPTG). These cultures were incubated at 28 °C for 24 h. To verify the localization of the soluble scFvs, the proteins from whole cell, periplasm, and supernatant were extracted and tested using an ELISA. Each overnight culture was divided into two equal volumes and centrifuged at 3000 rpm for 30 min. The supernatant of each culture was collected to test for the presence of functional scFvs. For extracting protein from the whole cells, one fraction of the cell pellet was resuspended in 1 mL of 1X PBS. The cells were ruptured by sonication and centrifuged at 15,000 rpm for 15 min. The supernatant was collected to be tested for the presence of soluble scFv. To obtain soluble protein from the periplasm, the other fraction of the cell pellet was resuspended in 0.5 mL ice-cold 1X TES buffer (0.2 M Tris–HCl pH 8.0, 0.5 mM EDTA, 0.5 M Sucrose). After adding 0.75 mL of TES (diluted 1:4 in water), cells were vortexed for 30 s and incubated on ice for 1 h. The protein was collected as described previously. An indirect ELISA was performed with solutions containing protein from each fraction (i.e., whole cell, periplasm and supernatant). For this experiment, wells were coated with one of the three monensin conjugates (i.e., monensin-BSA, monensin-OVA or monensin-KLH) and blocked as described for the phage ELISA (Section 2.6). The solution from each fraction was added to the wells and incubated for 2 h at room temperature. ScFv binding was confirmed using a monoclonal mouse anti-penta His (QIAGEN) followed by goat anti-mouse antisera conjugated to horseradish peroxidase (Promega). ELISAs were developed as described previously.

2.9. Anti-monensin scFv expression and purification The clones that were positive for binding to immobilized monensin were grown in 1-L cultures and expressed scFvs were extracted as described previously. Soluble scFvs were purified using HisTrap HP metal affinity chromatography (GE Healthcare, Piscataway, NJ) by following the manufacturer's instructions. The eluted scFvs were dialyzed against 1X PBS at 4 °C for 24 h and tested by immunoblot to verify the presence of the purified scFv antibodies.

107

2.10. Preparation of fluorescent-labeled tracer Monensin was conjugated to fluorescein as described by Furzer et al. (2006). In brief, fluorescein thiocarbamyl ethylenediamine (EDF) was prepared first by adding triethylamine (100 μL) to 10 mL of methanol (solution A). Fluorescein isothiocyanate isoform I (FITC; 30 μmol) and ethylenediamine dihydrochloride (150 μmol) were dissolved in 1 mL and 5 mL of solution A, respectively. These two solutions were mixed and incubated for 16 h at 4 °C. The resulting precipitate was filtered through a glass microfiber filter (Whatman, Piscataway, NJ) and allowed to dry for 18 h at 22 °C in the dark. Monensin (20 μmol), N-hydroxysuccinimide (40 μmol), and dicyclohexylcarbodiimide (40 μmol) were added to 1 mL of dimethylformamide and mixed for 4 h at 22 °C. This solution (250 μL) was added drop-wise to EDF and stored at 4 °C. To purify the product of the conjugation reaction thinlayer chromatography was used. The monensin–fluorescein conjugate solution (50 μL) was spotted onto LK6F silica gel plates (1000 μm, 20 cm × 20 cm; Whatman) along with solutions of monensin (20 nM) and FITC (30 nM), which were used as reference standards. The chromatography plates were developed in methylene chloride:methanol (4:1 v/v). Standards and products were visualized at 366 nm. The spot containing monensin–fluorescein was scraped from the plate, added to 0.5 mL of methanol and allowed to stand overnight at 4 °C. The solution was filtered through glass wool, and the eluant was stored at −20 °C until required. 2.11. Optimal concentration of tracer and scFv To verify the optimal dilution of monensin–fluorescein (tracer), which was used in the fluorescence polarization assay, this conjugate was diluted (1:20 v/v) in borate buffer (BB; 50 mM boric acid, pH 8, adjusted with NaOH, 0.01% NaN3) and added to the first column of a black 96-well microtitre plate followed by serial dilution (1:2 v/v across the plate) in BB. The sample was subjected to an excitation wavelength of 488 nm and the fluorescence measured at a wavelength of 530 nm. The fluorescence value for each dilution was plotted against the tracer dilutions. The dilution (1:2000 v/v) corresponding to the first point on the lower plateau of the plotted curve represents the lowest concentration of monensin–fluoresceine that was accurately measurable. To determine the optimal concentration of either the antimonensin scFvs or rabbit sera, they were serially diluted (1/2, 1/4 …, 1/4096 for scFvs; and 1/10, 1/200…, 1/10,240 for rabbit serum) in BB across the plate and 80 μL of each dilution was mixed with 20 μL of the optimal concentration of tracer. The reaction was incubated for 15 min at room temperature, and the polarization was measured. A standard curve was plotted and the optimal antibody concentration was determined as the concentration at which 50% tracer binding was observed (i.e., midpoint of the curve). 2.12. Inhibition FP assay To prepare standard solutions, monensin was dissolved in methanol (1 mg mL−1). Aqueous standard solutions of free monensin (1 × 10−6, 5 × 10−5 …, 1, and 5 μg mL−1) were

108

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

prepared by dilution of the stock solution with BB buffer, pH 8. For each reaction, 70 μL of the scFv (1 μg mL−1) was mixed with 250 μL of each monensin standard solution and allowed to stand for 1 h at 25 °C. Polarization readings were made after adding the optimal concentration of the tracer into each well and incubating for 15 min. All samples were run in triplicate and each assay was repeated five times on different days for testing its reproducibility. 2.13. Cross-reactivity assay The cross-reactivity assays were performed with four antibiotics: salinomycin, lasalocid A, kanamycin and ampicillin. A C-FP assay for each antibiotic was performed as described in Section 2.12. Percentage of cross-reactivity was calculated according to the following equation: % CR = [IC50 (monensin)/IC50 (analyte)] × 100; where IC50 is the concentration of analyte at which 50% of the antigen binding sites on the antibodies are occupied. All samples were run in triplicate and each assay was repeated three times on different days for testing its reproducibility. 3. Results 3.1. Mice humoral responses to conjugated and free monensin Mice B1 and B2 were immunized with monensin-BSA and mice K1 and K2 were immunized with monensin-KLH. All four mice showed an increase in immune response to monensin-

OVA over time (Fig. 2). In mice B1, K1 and K2, the immune response to monensin-OVA increased by the third bleed, i.e. after second boost (Fig. 2A, C and D), whereas in mouse B2 the immune response was delayed and the serum titer increased only after the fourth bleed, i.e. after third boost (Fig. 2B). Following a checkerboard analysis to determine the optimal concentrations of serum and coating conjugate (monensin-OVA), a CI-ELISA was performed with the sera from the third and fifth bleeds of all four mice (Fig. 3A–D). The sera collected from the third and fifth bleeds of mouse B1 were inhibited by 100% and 70%, respectively, in the presence of 1 μg mL−1 (1.7 mM) of free monensin (Fig. 3A), while sera of both bleeds from mouse K2 showed 55–60% inhibition at the same monensin concentration (Fig. 3D). There was no specific inhibition caused by monensin when sera from mice B2 and K1 were used in the CI-ELISA (Fig. 3B, C). 3.2. Rabbits humoral response to conjugated and free monensin Rabbits BM and KM were immunized with monensin-BSA and monensin-KLH, respectively. The sera collected from both rabbits showed an increase in immune response over time (Fig. 4). Both rabbits responded to immunization after the second boost (i.e., third bleed), with the response from rabbit BM (Fig. 4A) being better than that from rabbit KM (Fig. 4B). Response of the rabbits to the two immunogens was similar after the third and fourth boosts (bleeds four and five, respectively). When CI-ELISAs and C-FP assays were used, the binding of serum from rabbit BM to monensin-OVA and

Fig. 2. Immunoassay results showing immune responses to monensin-OVA in four mice immunized with either monensin-BSA (mice B1 and B2) or monensin-KLH (mice K1 and K2) on days 0, 28, 42 and 70. In mice B1, K1 and K2, the immune response was seen after the second boost (Bleed 3) whereas mouse B2 response was not seen until the third boost (Bleed 4).

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

109

Fig. 3. CI-ELISA showing the specificity and sensitivity towards monensin of the sera collected from the four mice after the second boost (Bleed 3) and fourth boost (Bleed 5) with monensin-BSA for mice B1 and B2 and monensin-KLH for mice K1 and K2. Mouse B1 showed the lowest IC50 value, 0.35 μM.

monensin–fluorescein was inhibited by 54% and 80%, respectively, in the presence of 1 μg mL−1 (1.7 mM) of free monensin (Fig. 5A and B), while sera of rabbit KM did not show any inhibition (data not shown). Fig. 5 also shows that the C-FP assay was more sensitive than the CI-ELISA. 3.3. Selection and screening Since sera from mouse B1 had the highest response to inhibition by soluble monensin (Fig. 3A), the coding sequences of V H, V κ and V λ were isolated from the splenocytes of this mouse by PCR which amplified VH (∼ 450 bp; Fig. 6, Lane 2) and VL (400 bp; Fig. 6; Lanes 3 and 4) coding sequences. To construct scFv genes, VH and VL fragments were joined using overlap PCR, i.e. complementary linker sequences between VH and VL coding sequences acted as PCR primers. The final products of this reaction were scFv coding sequences of ∼850 bp (Fig. 6, Lane 5). A library of 1 × 109 clones with 90% diversity was thereby synthesized; 95% of the clones tested contained inserts. To make this library, 95% of the Vκ–VH and 5% of Vλ–VH solutions were used since these proportions are similar to those of the two respective light chains in the mouse immune system (Goldsby et al., 2002). Three rounds of selection were performed to isolate monensin-specific scFvs from this library. Rounds two and three were repeated twice with minor modifications in blocking buffer, concentrations of coating antigen and the number of washes (Table 1). The titers of the eluted phage

were estimated after each round of selection to determine the enrichment of the binders specific to monensin. The titers of phages displaying scFv from round 1 to rounds 2a and 2b were reduced by 52 and 11 times, respectively (Table 3). In rounds 3a and 3b the bound phages were eluted with three different concentrations of free monensin. The phage titer from round 2a to 3a increased 757 and 221 times when eluted with 0.1 μM and 10 μM monensin, respectively (Table 3). However, in round 3a, the phage titer after elution with 1 μM monensin was lower than background. No enrichment was observed in round 3b with 1 μM monensin. The collected phage-scFvs from each round of panning were checked for specific binding to the three coating conjugates (i.e., monensin-BSA, monensin-KLH and monensin-OVA) (Fig. 7). The binding to each protein (i.e., BSA, KLH and OVA) was subtracted from the binding to the corresponding monensin conjugate. ELISAs against BSAmonensin did not show a significant enrichment until round 2 after which there was no change in binding, except for the decline in round 3b (i.e.,10 μM; Fig. 7). With monensin-KLH there was a significant increase in binding in round 1 and again in round 2, with no further change in round 3. Due to some binding to OVA (data not shown), the detected signals from the ELISA against monensin-OVA were lower compared to the other two conjugates but were statistically significant (Fig. 7). There was no significant increase in binding until round three. There was no difference among the three eluates within rounds 3a and round 3b, with 3b being lower than 3a.

110

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

Fig. 4. Immunoassay results showing response to monensin-OVA of sera from two rabbits, one of which was immunized with monensin-BSA (Rabbit BM) and the other with monensin-KLH (Rabbit KM) on days 0, 28, 42 and 70. Bleed 0 was collected a week before the immunization started. Bleeds 2, 3, 4 and 5 were collected after the first, second, third and forth boosts, respectively. Both rabbits showed an immune response to monensin conjugates after the second boost (Bleed 3).

Fig. 5. CI-ELISA (A) and C-FP (B) showing the specificity and sensitivity to monensin of the serum collected from rabbit BM after the forth boost with monensin-BSA. C-FP was more sensitive than CI-ELISA. Numbers are the average of triplicates. Standard error of the means (SEMs) are shown with bars; when bars are not shown they are smaller than the symbol.

3.4. Soluble scFv expression Based upon these results, monoclonal phage-scFv ELISAs were conducted using clones selected from rounds 2a and 3a. A total of 33 binders specific to all three conjugates were isolated (Fig. 8A, B). However, when phage-scFv clones from rounds 2b and 3b were selected only four binders were isolated (Fig. 8C, D), all of which had little or no binding to monensin-BSA (Fig. 8C, D). Sequencing results showed that ten different clones were selected among these 37 positive clones (Fig. 9). Except for 6-T, 35-T and 39-T, all positive binders from round 2a were also identified in round 3a. These results also show that all the scFvs from round 3a contained stop codon except scFv 1–100. Interestingly, the three scFvs (6-T, 35-T and 39-T) that were obtained in round 2a did not contain stop codons. Stop codons were found either in framework region 1 (FR1) or FR2 (Fig. 9). Both positive scFvs from round 2b were also detected in round 3b. Most of the diversity was observed in the complementarity determining regions (CDRs) of the VH; however, there were some differences in the FRs (Fig. 9). All selected binders contained a Vκ light chain. The families to which each VH and VL belongs are shown in Table 4. To remove the UAG stop codons, site-directed mutagenesis was performed to change TAG to CAG, encoding glutamine, as suggested by Marcus et al. (2006).

After removing the stop codon, the scFvs were expressed in E. coli HB2151 and the proteins from the medium, periplasm and whole cells were each extracted to verify the

Fig. 6. A 2% agarose gel showing VH (450 bp; Lane 2), Vκ and Vλ (400 bp; Lanes 3 and 4) and scFv (850 bp; Lane 5).

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

111

Table 3 An overview of selection results describing the isolation of specific binders to monensin. The phage-scFvs were simultaneously panned against both the carrier proteins (i.e., background) and the monensin conjugates. Bound phage-scFvs were eluted with trypsin in rounds one and two and in round three with three increasing concentrations of free monensin (i.e., 0.1 μM1, 1 μM2 and 10 μM3). ND, not determined, i.e. the titer of background was higher than the phage titer collected from selection against monensin conjugates. Pfu, plaque forming unit. Round

Input (Pfu)

Output (Pfu)

Output (Pfu)1

Output (Pfu)2

Output (Pfu)3

Phage recovery %

1 2a 2b 3a 3b

5.5 × 1015 2.4 × 1014 2.4 × 1014 3.7 × 1013 1.4 × 1013

1.73 × 106 3.3 × 104 1.6 × 105 – –

– – – 2.5 × 107 ND

– – – ND ND

– – – 7.3 × 106 1.0 × 104

3.1 × 10−8 1.3 × 10−8 6.6 × 10−8 2.0 × 10−5–6.7 × 10−5 7.1 × 10−8

location and functionality of each soluble scFv. In spite of removing the stop codons, only half of the soluble scFvs exhibited significant binding to monensin (Fig. 10). These antibodies were purified and used for immunoblotting. Single bands of approximately 30 kDa were detected for each (Fig. 11). 3.5. Binding characteristics of soluble scFvs The five soluble scFvs specific to the monensin conjugates were assessed for their specificity to free monensin by C-FP assay. The concentration of monensin–fluorescein conjugate, blocking conditions, and scFv dilutions were optimized for this assay. The binding of each scFv to monensin–fluorescein was inhibited with different concentration of free monensin (Section 2.12). Milli-Polarization (mP) values were recorded and the standard curves were plotted as percentage inhibition versus log of analyte concentration. The IC50 values were calculated based on the generated equation for each standard curve (Fig. 12). Although scFv T-6 bound to the conjugates (Fig. 10), its binding was not inhibited by monensin (data not shown). As Fig. 12 shows, scFv 1-100 had the lowest IC50 (i.e., 0.031 μM). The cross-reactivities to salinomycin, lasalocid A, kanamycin and ampicillin of the two scFvs with the lowest IC50 values, i.e. clones 1-100 and T-39, were determined. There was no cross-reactivity to lasalocid and salinomycin which are also members of the polyether ionophore family of antibiotics or to kanamycin and ampicillin which belong to the aminoglycoside and beta-lactam families of antibiotics, respectively. 4. Discussion Many environmental contaminants such as antibiotics, pesticides and drugs are hydrophobic small molecules. The development of sensitive detection immunoassays for these molecules is very challenging due to the difficulties associated with their small size, which makes them non-immunogenic, and their hydrophobicity, which often makes it difficult to solubilize them for conjugation to hydrophilic proteins. Furthermore, following successful conjugation, it has been shown that the majority of these antibody fragments are specific to either the conjugate or the derivatized form of the hapten and not the free hapten (de Haard et al., 1999; Dorsam et al., 1997; Little et al., 1999). Despite all these difficulties, a number of antibody fragments have been successfully isolated against many hydrophobic haptens (Yang et al., 2007).

In this study, we have isolated and characterized antimonensin scFvs. To our knowledge there have been only two other antibiotics for which scFvs have been isolated: ampicillin (Burmester et al., 2001) and sulfamethazine (Yang et al., 2007). To construct a hyperimmunized scFv library for selection of specific binders to monensin, four mice were immunized with two different monensin conjugates (i.e., monensin-BSA and monensin-KLH). Both conjugates elicited an immune response in all four mice as well as in two rabbits (Figs. 2 and 4). However, only mice B1 and K2, and rabbit BM showed a specific immune response against the hapten monensin (Fig. 3A and D; Fig. 5A and B). Conversely, mice B2 and K1, and rabbit KM elicited a response against the conjugates and not to monensin. The CI-ELISA also revealed that after the second boost (third bleed) the sensitivity of the assay to monensin did not improve when sera of mouse B1 or K2 were used (Fig. 3A and D). These results indicate that after the second boost the immune response was mainly directed towards either the altered hapten (i.e., altered via conjugation to the protein) or to other antigenic determinants of the carrier proteins. Thus, it is reasonable to assume that after the third immunization (second boost), no further benefit was achieved, i.e., there was no increase in affinity and specificity for monensin with subsequent boosts. Among all six animals (i.e. four mice and two rabbits), mouse B1 showed the best immune response to free monensin. The splenocytes of this mouse were utilized for construction of a hyperimmunized phage display scFv library using universal primer sets (Krebber et al., 1997) to amplify VH and VL genes. Monensin-specific binders were selected from the constructed scFv library. In the first couple of attempts, we selected a large number of non-specific binders (data not shown). We thus chose a subtractive panning procedure to eliminate the binders to polystyrene, blocking agents and/or the carrier proteins. Alternative panning approaches, which differed in the number of washes and the concentration of the coating conjugates, were used to target high affinity binders (Table 1). One of the strategies in this panning procedure included extensive washing steps to favor the selection of high affinity binders. In both rounds 2a and 3a the number of washes was greater than for rounds 2b and 3b (Table 1). In addition, the concentrations of the coating conjugates were decreased from 35 μg mL−1 in rounds 2a and 2b to 10 μg mL−1 in round 3a and 1 μg mL−1 in round 3b. The same elution strategy was used in both panning procedures. Phage-scFvs were eluted using trypsin in the first two rounds of panning to isolate and separate lowand high affinity binders. In the last rounds of panning (i.e., 3a

112

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

Fig. 7. Polyclonal phage ELISAs to determine specific binding of phage-scFvs to monensin-BSA (A), monensin-KLH (B) and monensin-OVA (C) over the three rounds of selection. The background (i.e., wells coated with BSA, KLH or OVA;OD450 b 0.1) was subtracted from the absorbance of the corresponding monensin conjugates. SEMs are shown with bars.

and 3b), phage-scFvs were eluted by incubating with increasing concentrations of free monensin to isolate monensin-specific binders. With this elution strategy, a low concentration of monensin (i.e., 0.1 μM) was used to elute monensin-specific binders with higher affinity followed by elution with higher concentrations of monensin (i.e., 1

and 10 μM) for elution of binders with lower affinities. Moghaddam et al. (2001) showed a dramatic increase in percentage of isolated scFvs that bound to free aflatoxin-B1 when they eluted phage-antibody with free aflatoxin-B1. They also showed that scFvs isolated using this approach had higher affinity. In spite of using free monensin to elute

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

113

Fig. 8. Monoclonal phage ELISAs to determine the binding activity of each phage-scFv isolated from an individual clone after rounds 2a (A), 3a (B), 2b (C) and 3b (D). Ninety four clones from each round of selection (except round one) were randomly screened and their binding to three different conjugates (i.e., monensin-BSA, monensin-KLH and monensin-OVA) was determined. The background (i.e., wells coated with BSA, KLH and OVA, OD450 b 0.1) was subtracted from the absorbance to the respective monensin-protein conjugate. Numbers are the average of triplicates. SEMs are shown with bars; when bars are not shown they are smaller than the symbol. A total of 33 positive clones were isolated from rounds 2a plus 3a, while only four positive clones were identified from rounds 2b plus 3b.

114

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

Fig. 9. Deduced amino acid sequences of the ten different scFvs from the phage ELISA that were specific for binding to at least one of the three monensin conjugates (i.e., monensin-BSA, -KLH, and -OVA). The stop codons are represented by an asterisk (*). All the stop codons were located in either VL-FR1 or VL-FR2. Conserved residues are represented by dots. The deleted residues are represented by a dash (-). FRs and CDRs are determined according to (Kabat et al., 1991).

monensin-specific binders, some non-specific phage-scFvs were also eluted; this was particularly evident when the two highest concentrations of the antibiotic (i.e., 1 μM and 10 μM) were used. Isolation of these non-specific binders may be due in part to the fact that these antibodies were not removed during the washing steps and were instead released during the prolonged duration of the elution with monensin. To overcome this problem, we hypothesize that additional rounds of selection with shorter incubation times for the selective monensin-based elution steps may be necessary. For further characterization of how the panning strategies discussed above affected selection of specific monensin

binders, a polyclonal phage ELISA was also performed on collected phage-scFv from each round to determined if there was specific binding to monensin. Although during the panning procedure the number of bound phage appeared to drop from the first to the second round of panning (Table 2), the polyclonal phage ELISA showed an increase in the binding signals (Fig. 7). This suggests that after the second round of selection there was enrichment of specific binders to the conjugates. Despite high background in round 2a, the polyclonal phage ELISA showed no non-specific binding (very low background) and 16 positive clones were isolated in this round (Fig. 8A). When determined by monoclonal

Table 4 VH, VL Families and J segments to which each of the anti-monensin scFv belong. The sequences were analyzed using the Mouse Immunoglobulin V-BASE2 database (Retter et al., 2005). Clone

VH family

VH gene

JH segment

VL family

Vκ gene

Jκ segment

45-100 46-10 5-10 1-100 4-10 2-10 15-10 39-T 6-T 35-T

VHSM7 VHJ588 VHSM7 VHSM7 VHSM7 VHSM7 VHSM7 VHSM7 VHSM7 VHSM7

HV14-3 HV1-48 HV14-3 HV14-3 HV14-3 HV14-3 HV14-3 HV14-3 HV14-3 HV14-3

JH3 JH3 JH3 JH3 JH3 JH JH3 JH3 JH3 JH3

κV19/28 κV19/28 κV19/28 κV19/28 κV19/28 κV19/28 κV19/28 κV19/28 κV19/28 κV19/28

κV6-15 κV6-15 κV6-15 κV6-15 κV6-15 κV6-15 κV6-15 κV6-15 κV6-15 κ-V6-15

Jκ5 Jκ2 Jκ5 Jκ2 Jκ5 Jκ5 Jκ2 Jκ2 Jκ4 Jκ2

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

115

Fig. 10. Binding of soluble scFvs to monensin-BSA, monensin-KLH and monensin-OVA in an indirect ELISA. Soluble scFvs 1-100, 45-100, 4-10, 6-T and 39-T showed good binding to all three conjugates. The background (i.e., wells coated with BSA, KLH or OVA) was subtracted from the absorbance of each monensin conjugate. Numbers are average of triplicates. SEMs are shown with bars.

phage ELISA, a total of 33 monoclonal scFv-expressing phage from rounds 2a and 3a were found to be positive for binding to the three monensin conjugates (Fig. 8A and B), while only four positive clones were isolated from rounds 2b and 3b (Fig. 8C and D). These four phage-scFvs bound to monensinOVA and monensin-KLH, which were used as coating conjugates in the panning procedure, and not to monensinBSA (Fig. 8C and D), which was not used for panning. This implies that minor changes in the blocking buffers, concentrations of coating antigen, and the number of washing steps have a profound effect on the outcome of selection (i.e., selection of specific binders vs. non-specific binders). The coding sequences of the 37 positive phage-scFv clones were identified. Sequence analysis revealed that these 37 clones consisted of ten different sequences (Fig. 9). These sequences

Fig. 11. Immunoblot analysis of the four soluble anti-monensin scFvs (∼ 30 kDa). Only the four monensin-specific scFvs are shown (6-T is not shown). Lane 1, protein standards; Lane 2, scFv 45-100; Lane 3, scFv 4-10; Lane 4, scFv 1-100; Lane 5, scFv 39-T.

displayed a high level of homology with each other. Due to the small volume of the hapten, there are few epitopes available for antibody selection. Furthermore, many of these epitopes may be simultaneously buried in the antigen binding pocket, thus making only a few epitopes available to the B cells for recruitment of antibodies with affinity for the hapten. Consequently, only a small portion of each antibody is specific to the target hapten and these are highly homologous (Kramer, 2002). Further analysis of the ten clones identified to which VH and VL families the heavy and light chains of these scFvs belong. The light chains of the ten clones isolated belong to the Vκ19/28 family and are joined to one of three different Jκ segments (Table 4). In adult Balb/c mice, it has been shown that 20% of expressed Vκ generally belong to this family (Kaushik and Lim, 1996). Except for clone 46-10, all isolated scFvs contain heavy chains belonging to the VHSM7 family. The heavy chain of clone 46-10 belongs to the VHJ558 family, which has been shown to be closely related to the VHSM7 family (Tutter et al., 1991). The sequence analysis also revealed that except for scFv 1100, all other specific binders which were selected in round 3a contained a stop codon in either VLFR1 or VLFR2 (Fig. 9). ScFvs T-35, T-39 and T6, which were positive binders with no stop codon in round 2a, were not selected in round 3a. It appears the selection pressure exercised in round 3a favored binders with stop codons. Although this phenomenon is not well understood, it has been argued that the presence of a stop codon can be used as a defense mechanism by E. coli against the toxic effect of the scFvs (Marcus et al., 2006). In E. coli TG1, three different codons (i.e., CAG, CAA, or TAG) may translate to glutamine. CAG and CAA codons are translated to glutamine 100% of the time, while TAG is expressed as glutamine only 20% of the time and the other 80% of the time it is translated to termination (Marcus et al., 2006). Therefore, selection of stop codon-containing clones will result in reduction of scFv expression which may be potentially toxic to the bacteria when expressed in high concentration (Marcus et al., 2006). On the other hand, in E. coli HB2151,

116

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

Fig. 12. C-FP assay to determine the affinity of the soluble scFvs based on IC50 values. Different concentrations of free monensin, ranging from 0 to 667 μM were used to inhibit the binding of the scFvs to monensin–fluorescein conjugate. SEMs are shown with bars; when bars are not shown, they are smaller than the symbol.

which is used for soluble expression of the scFvs, TAG is translated only as a stop codon, which results in incomplete translation of a scFv. After replacing the stop codon with the CAG codon, each soluble scFv was expressed and characterized. Although all ten phage-scFvs bound to monensin conjugates, only five of these antibodies showed binding to these antigens in soluble format (Fig. 10). Tout et al. (2001) suggested this variation may be due to differences in the valency and folding of the antigen binding sites of the phage-bound versus the soluble scFv. Another explanation was given by Scott et al. (2008) after showing the amount of expressed scFv on the surface of the phage does not correlate with the concentration of expressed soluble scFv. This group also observed a positive correlation between the total amount of expressed scFv and the amount of pIII-scFv in periplasm but not with the concentration of soluble scFv expression. They explain that once the proteins are translocated across the inner membrane of E. coli, the pIII-scFv fusion stays anchored in the inner membrane via the C-terminal domain of pIII whereas a soluble scFv remains in the oxidative environment of the periplasmic space of E. coli where it is subjected to potential mis-folding and aggregation which results in lowering the concentration of functional scFv. Developing a competitive immunoassay to verify the binding characteristics of a specific antibody to a hapten can be limited by factors such as the affinity of the antibody (i.e., high affinity antibodies are needed to achieve low detection limits in competitive assays) and water solubility

of the target hapten. In the case of monensin, the latter factor (low solubility in H2O) was anticipated to be the main challenge in developing a good immunoassay; however, we show that this factor did not pose a barrier to the development of two IgG-based assays (C-FP and CI-ELISA using rabbit sera, Fig. 5), and a scFv-based C-FP (Fig. 12). Unlike C-FP, the CI-ELISA method is based on the competition between monensin and multiple monensin molecules conjugated to OVA (i.e., monensin-OVA) for a limited number of binding sites (i.e., antibody). Furthermore, in ELISA, the conjugate is bound to microtiter wells (heterogeneous assay) which allows for bivalent binding of IgG to the coating conjugate and may also change the presentation of monensin to the antibody. Conversely, in a C-FP assay, the interaction between the analyte-fluorescein conjugate (1:1 monensin–fluorescein) and the IgG or scFv occurs in an aqueous environment (i.e., a homogeneous assay format), thereby creating a monovalent interaction between the IgG or scFv and the monensin–fluorescein conjugate as well as eliminating any potential negative effects of sorption of the conjugate. Furthermore, based upon our experience, a C-FP assay can be completed in less time than a CI-ELISA and is more amenable to high throughput analysis of environmental samples. Based on this information, we chose the C-FP format over that of the CI-ELISA format (Fig. 12) to create the final quantitative assay. The sensitivity and specificity of the scFvs were determined based on their IC50 values (Fig. 12) using the C-FP assay. ScFvs 1-100 (IC50 = 0.03 μM or 0.018 ng mL−1) and T-39 (IC50 = 0.046 μM or 0.03 ng mL−1) showed very similar

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

sensitivity to free monensin, a result in accordance with the fact that these two scFvs have the highest homology in their amino acid sequences (Fig. 9). In this study, the IC50 values of these two scFvs, as determined by C-FP (Fig. 12), was lower than the calculated IC50 from B1 mouse serum as determined by CI-ELISA (i.e., 16 μM or 9.98 ng mL−1; Fig. 3A). It is possible that the IC50 value of B1 mouse serum may be higher than that of the two scFvs because the FP assay is more sensitive than ELISA. It was unfortunate that the volume of serum collected from mouse B1 was too low to also perform C-FP in order to make a more appropriate comparison of the IC50 values. These scFv binders were also more sensitive than the polyclonal serum (IC50 = 164.9 ng mL−1) and the monoclonal antibody (IC50 = 50 ng mL−1) developed by Crooks et al. (1998) and Watanabe et al. (1998), respectively. Our results are particularly interesting when one considers that our C-FP assay is monovalent in nature, while our IgG-based ELISAs and those of the previous authors are most likely bivalent in nature. Accordingly, the higher affinity of the C-FP assay is not based on avidity as is the case with the IgG-based ELISAs (Crooks et al., 1998 and Watanabe et al., 1998), but rather solely on affinity and thus 1:1 competition between free hapten and the monensin–fluorescein conjugate. Furthermore, although it is difficult to know the exact basis for the higher affinity of the scFvs versus IgGs, we speculate it is due to the separation of parental VL and VH genes and the subsequent recombinations of VL–VH DNA sequences during library construction that result in combinations with improved binding properties for monensin. Cross-reactivity studies were performed with structurally related and unrelated antibiotics to assess the specificity of the two high affinity scFvs (i.e., 1-100 and T-39). There was little or no cross-reactivity indicating high specificity of these two scFvs for monensin. Even salinomycin and lasalocid A, which also belong to carboxylic acid ionophore family, were not bound by either of these two antibodies. Crooks et al. (1998) and Watanabe et al. (1998) also found little or no cross-reactivity of their anti-monensin polyclonal and monoclonal antibodies to structurally related polyether ionophoric antibiotics. In conclusion, several authors (e.g., Eremin and Smith, 2003; Furzer et al., 2006) have shown that C-FP assays can be successfully applied for the detection of a number of small molecules. Thus, our C-FP assay for monensin can be used for constant monitoring of water and food for monensin, and will be less time consuming and more sensitive than previously developed ELISAs for monensin. Furthermore, the C-FP is amenable to high throughput analysis of environmental samples. Finally, due to the monovalent nature of our C-FP, this assay had a much higher affinity, as determined by comparing IC50s, than both the bivalent IgG-based ELISAs developed in this paper and previously developed by other authors to detect monensin. Acknowledgements We gratefully acknowledge the financial support to Dr. J. Christopher Hall from the Canada Research Chairs Program, the Natural Sciences and Engineering Research Council of Canada (NSERC) via the NSERC Networks Program to the Sentinel Bioactive Paper Network, and the Ontario Ministry of Agriculture and Food and Rural Affairs.

117

Appendix A. Universal primers for amplification of VL (A) and VH (B) of mouse IgG genes (Modified from Krebber et al., 1997). Y= C/T; W=A/T; M=A/C; R=A/G; D=A/G/T; S=C/G; K=G/T. The underlined and italic sequences represent the SfiI restriction sites and the linker (G10S7R), respectively

(A) Vκ Forward mix

Oligonucleotide sequences

Vκ-1

5′ GGG CCC AGC CGG CCG AGC TCG AYA TCC AGC TGA CTC AGC C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TTC TCW CCC AGT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TGM TMA CTC AGT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TGY TRA CAC AGT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TRA TGA CMC AGT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTM AGA TRA MCC AGT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTC AGA TGA YDC AGT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TYC AGA TGA CAC AGA C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TTC TCA WCC AGT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG WGC TSA CCC AAT C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTS TRA TGA CCC ART C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYR TTK TGA TGA CCC ARA C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TGA TGA CBC AGK C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TGA TAA CYC AGG A 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TGA TGA CCC AGW T 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTG TGA TGA CAC AAC C 3′ 5′ GGG CCC AGC CGG CCG AGC TCG AYA TTT TGC TGA CTC AGT C 3′

Vκ-2 Vκ-3 Vκ-4 Vκ-5 Vκ-6 Vκ-7 Vκ-8 Vκ-9 Vκ-10 Vκ-11 Vκ-12 Vκ-13 Vκ-14 Vκ-15 Vκ-16 Vκ-17

Vκ reverse mix VκR1

VκR2

VκR3

Vλ Forward Vλ F

Vλ Reverse Vλ R

5′ GGA AGA TCT AGA GGA ACC ACC CCC ACC ACC GCC CGA GCC ACC GCC ACC AGA GGA TTT KAT TTC CAG YTT GGT CCC 3′ 5′GGA AGA TCT AGA GGA ACC ACC CCC ACC ACC GCC CGA GCC ACC GCC ACC AGA GGA TTT TAT TTC CAA CTT TGT CCC 3′ 5′GGA AGA TCT AGA GGA ACC ACC CCC ACC ACC GCC CGA GCC ACC GCC ACC AGA GGA TTT CAG CTC CAG CTT GGT CCC 3′

5′ GGG CCC AGC CGG CCG AGC TCG ATG CTG TTG TGA CTC AGG AAT C 3′

5′GGA AGA TCT AGA GGA ACC CCC ACC ACC GCC CGA GCC ACC GCC ACC AGA GGA GCC TAG GAC AGT CAG TTT GG 3′

118

S. Makvandi-Nejad et al. / Journal of Immunological Methods 360 (2010) 103–118

(B) Forward mix

Oligonucleotide sequences

VH1

5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTR MAG CTT CAG GAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTB CAG CTB CAG CAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG CAG CTG AAG SAS TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTC CAR CTG CAA CAR TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTY CAG CTB CAG CAR TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTY CAR CTG CAG CAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTC CAC GTG AAG CAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG AAS STG GTG GAA TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG AWG YTG GTG GAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG CAG SKG GTG GAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG CAM CTG GTG GAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG AAG CTG ATG GAR TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG CAR CTT GTT GAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTR AAG CTT CTC GAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG AAR STT GAG GAG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTT ACT CTR AAA GWG TST G 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTC CAA CTV CAG CAR CC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG AAC TTG GAA GTG TC 3′ 5′ GGT GGT TCC TCT AGA TCT TCC CTC GAG GTG AAG GTC ATC GAG TC 3′

VH2 VH3 VH4 VH5 VH6 VH7 VH8 VH9 VH10 VH11 VH12 VH13 VH14 VH15 VH16 VH17 VH18 VH19

VH3′ Reverse primers VHR1 VHR2 VHR3

5′ CCT GGC CGG CCT GGC CAC TAG TGA CAG ATG GGG STG TYG TTT TGG C 3′ 5′ CCT GGC CGG CCT GGC CAC TAG TGA CAG ATG GGG CTG TTG TTG T 3′ 5′ CCT GGC CGG CCT GGC CAC TAG TGA CAT TTG GGA AGG ACT GAC TCT C 3′

Appendix B. Oligonucleotide sequences of primers to link VL and VH genes in an overlap PCR modified from Krebber et al. (1997)

Overlap extension primers

Oligonucleotide sequences

RSC-F (Forward)

5′ GAG GAG GAG GAG GAG GAG GCG GGG CCC AGC CGG CCG AGC TC 3′ 5′ GAG GAG GAG GAG GAG GAG CCT GGC CGG CCT GGC CAC TAG TG 3′

RSC-B (Reverse)

References Arbabi Ghahroudi, M., Desmyter, A., Wyns, L., Hamers, R., Muyldermans, S., 1997. Selection and identification of single domain antibody fragments from camel heavy-chain antibodies. FEBS Lett. 414, 521.

Burmester, J., Spinelli, S., Pugliese, L., Krebber, A., Honegger, A., Jung, S., Schimmele, B., Cambillau, C., Plückthun, A., 2001. Selection, characterization and x-ray structure of anti-ampicillin single-chain Fv fragments from phage-displayed murine antibody libraries. J. Mol. Biol. 309, 671. Canadian Food Inspection Agency. http://www.inspection.gc.ca/english/ anima/feebet/mib/mib57e.shtml Crooks, S.R., Fodey, T.L., Gilmore, G.R., Elliott, C.T., 1998. Rapid screening for monensin residues in poultry plasma by a dry reagent dissociation enhanced lanthanide fluoroimmunoassay. Analyst 123, 2493. de Haard, H.J., van Neer, N., Reurs, A., Hufton, S.E., Roovers, R.C., Henderikx, P., de Bruine, A.P., Arends, J.W., Hoogenboom, H.R., 1999. A large non-immunized human Fab fragment phage library that permits rapid isolation and kinetic analysis of high affinity antibodies. J. Biol. Chem. 274, 18218. Dorsam, H., Rohrbach, P., Kurschner, T., Kipriyanov, S., Renner, S., Braunagel, M., Welschof, M., Little, M., 1997. Antibodies to steroids from a small human naive IgM library. FEBS Lett. 414, 7. Eremin, S.A., Smith, D.S., 2003. Fluorescence polarization immunoassays for pesticides. Comb. Chem. High Throughput Screen. 6, 257. Fasciglione, G.F., Marini, S., Bannister, J.V., Giardina, B., 1996. Hapten-carrier interactions and their role in the production of monoclonal antibodies against hydrophobic haptens. Hybridoma 15, 1. Fleeker, J., 1987. Two enzyme immunoassays to screen for 2, 4-dichlorophenoxyacetic acid in water. J. Assoc. Off. Anal. Chem. 70, 874. Furzer, G.S., Veldhuis, L., Hall, J.C., 2006. Development and comparison of three diagnostic immunoassay formats for the detection of azoxystrobin. J. Agric. Food Chem. 54, 688. Goldsby, R.A., Kindt, T.J., Osborne, B.A., Kusby, J., 2002. Immunology, 5th ed. W.H. Freeman and Company, New York. Hao, C., Lissemore, L., Nguyen, B., Kleywegt, S., Yang, P., Solomon, K., 2006. Determination of pharmaceuticals in environmental waters by liquid chromatography/electrospray ionization/tandem mass spectrometry. Anal. Bioanal. Chem. 384, 505. Kabat, E., Wu, T., Rei-Miller, M., Perry, H., Gottesman, K., Foeller, C., 1991. Sequences of Proteins of Immunological Interest. Department of Health and Human Service, Public Service, National Institute of Health, Washington, DC, U.S. Kaushik, A., Lim, W., 1996. The primary antibody repertoire of normal, immunodeficient and autoimmune mice is characterized by differences in V gene expression. Res. Immunol. 147, 9. Kramer, K., 2002. Synthesis of a group-selective antibody library against haptens. J. Immunol. Methods 266, 209. Krebber, A., Bornhauser, S., Burmester, J., Honegger, A., Willuda, J., Bosshard, H.R., Plückthun, A., 1997. Reliable cloning of functional antibody variable domains from hybridomas and spleen cell repertoires employing a reengineered phage display system. J. Immunol. Methods 201, 35. Little, M., Welschof, M., Braunagel, M., Hermes, I., Christ, C., Keller, A., Rohrbach, P., Kurschner, T., Schmidt, S., Kleist, C., Terness, P., 1999. Generation of a large complex antibody library from multiple donors. J. Immunol. Methods 231, 3. Marcus, W.D., Lindsay, S.M., Sierks, M.R., 2006. Identification and repair of positive binding antibodies containing randomly generated amber codons from synthetic phage display libraries. Biotechnol. Prog. 22, 919. Moghaddam, A., Lobersli, I., Gebhardt, K., Braunagel, M., Marvik, O.J., 2001. Selection and characterisation of recombinant single-chain antibodies to the hapten Aflatoxin-B1 from naive recombinant antibody libraries. J. Immunol. Methods 254, 169. Pinkerton, M., Steinrauf, L.K., 1970. Molecular structure of monovalent metal cation complexes of monensin. J. Mol. Biol. 49, 533. Pressman, B.C., Fahim, M., 1982. Pharmacology and toxicology of the monovalent carboxylic ionophores. Annu. Rev. Pharmacol. Toxicol. 22, 465. Retter, I., Althaus, H.H., Munch, R., Muller, W., 2005. VBASE2, an integrative V gene database. Nucleic Acids Res. 33, 671. Scott, N., Reynolds, C.B., Wright, M.J., Qazi, O., Fairweather, N., Deonarain, M.P., 2008. Single-chain Fv phage display propensity exhibits strong positive correlation with overall expression levels. BMC Biotechnol. 8, 97. Tout, N.L., Yau, K.Y., Trevors, J.T., Lee, H., Hall, J.C., 2001. Synthesis of ligandspecific phage-display ScFv against the herbicide picloram by direct cloning from hyperimmunized mouse. J. Agric. Food Chem. 49, 3628. Tutter, A., Brodeur, P., Shlomchik, M., Riblet, R., 1991. Structure, map position, and evolution of two newly diverged mouse Ig VH gene families. J. Immunol. 147, 3215. Wang, W., Malcolm, B.A., 1999. Two-stage PCR protocol allowing introduction of multiple mutations, deletions and insertions using QuikChange Site-Directed Mutagenesis. Biotechniques 26, 680. Watanabe, H., Satake, A., Matsumoto, M., Kido, Y., Tsuji, A., Ito, K., Maeda, M., 1998. Monoclonal-based enzyme-linked immunosorbent assay and immunochromatographic rapid assay for monensin. Analyst 123, 2573. Yang, Z.Y., Shim, W.B., Kim, M.G., Lee, K.H., Kim, K.S., Kim, K.Y., Kim, C.H., Ha, S.D., Chung, D.H., 2007. Production and characterization of monoclonal and recombinant antibodies against antimicrobial sulfamethazine. J. Microbiol. Biotechnol. 17, 571.