Prior to disposal, all the containers (e.g., centrifuge tubes and flasks) for phage incubation were immersed in 10% bleach solution overnight, followed by autoclaving. All the disposable items (pipet tips and tubes) in contact with phages were autoclaved before being discarded. TBBPA and its analogues were discarded as hazardous waste according to campus policies.

The synthesis of haptens T1–T6 (Figure 1) and the conjugation of haptens to carrier proteins were described in the previous report.¹⁶ TBBPA standard was purchased from TCI Co. Ltd. (Tokyo, Japan). TBBPA derivatives and other BFR analogues were purchased from AccuStandard (New Haven, CT). The anti-HA tag antibody (horseradish peroxidase (HRP) conjugate) (ab1265) and goat antirabbit IgG (HRP conjugate) (ab97051) were purchased from Abcam (Cambridge, MA). HisPur Ni-NTA resin, B-PER, Halt protease inhibitor cocktail, and Nunc MaxiSorp flat-bottom 96-well plates were purchased from Thermo Fisher Scientific Inc. (Rockford, IL, U.S.A.). Incomplete Freund’s adjuvant, thyroglobulin, bovine serum albumin (BSA), 3,3′,5,5′-tetramethylbenzidine (TMB), polyethylene glycol 8000 (PEG 8000), isopropyl-β-d-thiogalactopyranoside (IPTG) were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, U.S.A.). Mouse anti-M13 phage mAb-HRP was from GE Healthcare (Piscataway, NJ, U.S.A.). Rabbit anti-alpaca PAb was produced in the laboratory. The phagemid vector pComb3X was a gift from Dr. Carlos F. Barbas (The Scripps Research Institute, La Jolla, CA, U.S.A.). Electrocompetent E. coli ER2738 cells were acquired from Lucigen Corporation (Middleton, WI). M13KO7 helper phage and SfiI were purchased from New England Biolabs (Ipswich, MA).

A four-year old castrated male alpaca was immunized subcutaneously with T5-thyroglobulin six times biweekly. The VHH phage display library was constructed with the blood lymphocytes collected after the sixth injection using the method described previously.³⁰ Briefly, the total mRNA was transcribed to complementary DNA, and VHH fragments were amplified by polymerase chain reaction (PCR). VHH IgG variable domains were ligated into the plasmid pComb3X using restriction sites SfiI. Then the ligated material was electroporated into electrocompetent cells E. coli ER 2738.

One well of a microtiter plate was coated overnight with 100 μL of T5-BSA (10 μg mL^–1) at 4 °C, and an additional four wells with 100 μL of 3% BSA in coating buffer. The plate was blocked with 3% skim milk in PBS (0.01 mol L^–1 phosphate, 0.137 mol L^–1 NaCl, 3 mmol L^–1 KCl, pH 7.4) for 1 h at ambient temperature. A 100-μL aliquot of phage-display VHH library was added into the first well with 5% methanol (MeOH) and incubated for 2 h with gentle shaking at ambient temperature. After washing 10 times with PBST (0.05% Tween-20 in PBS), this well was eluted with 100 μL of TBBPA (1000 ng mL^–1) in PBS containing 5% MeOH for 1 h at ambient temperature with shaking. The eluent was transferred in equal aliquots to the next four BSA-coated wells to remove VHH phage that binds nonspecifically. Then the eluent was collected for the determination of phage titer and phage amplification. The phage eluent was amplified with addition of the M13KO7 helper phage (1 × 10¹² cfu mL^–1) for the next round of panning, which was described by Barbas et al.³⁷ The entire panning process was repeated three times, except the concentrations of coating antigen and TBBPA to elute the VHH phage were decreased gradually. The concentrations of T5-BSA for the second, third, and fourth panning were 5, 2.5, and 1 μg mL^–1, respectively. Meanwhile, the concentrations of TBBPA were decreased to 200, 40, and 10 ng mL^–1, respectively. After four rounds of panning, several phage clones were tested for their binding ability with TBBPA by a competitive phage ELISA,³⁰ and the optimal one was selected for the remaining studies.

Similarly, the heterologous coating antigens T1-BSA and T3-BSA were separately coated to isolate VHHs from the VHH library.

The cloned plasmids pComb3X containing the anti-TBBPA VHHs were extracted from ER2738 and heat shock transformed to Top 10F′ cells. A 1-mL aliquot of overnight culture was diluted in 100 mL of Super Broth with 50 μg mL^–1 carbenicillin and 2 mL of 1 M MgCl₂. After shaking for 8 h, the culture was induced with 1 mM IPTG and incubated in a shaker at 37 °C overnight. The culture was centrifuged, and the cell pellet was lysed with B-PER lysis buffer (4 mL g^–1 pellet) containing protease inhibitors at ambient temperature for 10 min. The cell lysate supernatants were collected by centrifugation at 13000 × g for 10 min, followed by purifying on a 1-mL Ni-NTA resin column. The column was equilibrated and washed with 40 mM imidazole (dissolved in 10 mM PBS, pH 7.4). The VHH was eluted with 150 mM imidazole, and the purified VHH was stored at −20 °C after dialysing with three buffer changes with PBS at 4 °C.

The optimal concentrations of coating antigens and VHH dilutions were selected by a checkerboard titration. A 100-μL solution of T3-BSA (2 μg mL^–1) was coated on a 96-well microtiter plate at 4 °C overnight. The plate was blocked with 3% skim milk in PBS for 1 h at ambient temperature the next day. A series dilution of TBBPA (50 μL/well, 10% MeOH in PBS) was added, followed by the addition of 50 μL of VHH in PBST (0.02 μg μL^–1). After incubation at room temperature for 1 h, the plate was washed five times with PBST, and then 100 μL of goat anti-HA tag IgG-HRP (diluted at 1:25000 with PBST) was added. After another incubation step and washing step, 100 μL of TMB solution (400 μL of 0.6% TMB and 100 μL of 1% H₂O₂ dilution in 25 mL citrate buffer, pH 5.5) was added into the plate, and the reaction was stopped 10 min later by the addition of 50 μL of 2 M H₂SO₄. The absorbance was read at 450 nm on a microtiter plate reader (Molecular Devices, Sunnyvale, CA). The IC₅₀, an indicator of the assay sensitivity, and the LOD, set as the IC₁₀, were obtained from a four-parameter logistic equation from SigmaPlot 10.0.

Different concentrations of MeOH or dimethyl sulfoxide (DMSO) (5, 10, 20, and 40%) in PBST were used to dilute the TBBPA to optimize the assay performance.

VHH was diluted with PBST at different pHs (4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, and 11.0). Correspondingly, the TBBPA was dissolved in the PBST with pH ranging from 4.0 to 11.0. Then calibration curves of ELISA were generated at different pH values.

For the thermal stability study, VHH was incubated at 90 °C for 10, 20, 30, 60, and 90 min, followed by cooling to room temperature. The binding of the VHH to the coating antigen T3-BSA was subsequently evaluated by ELISA.

The selectivity of the VHH ELISA was evaluated by determining the cross-reactivity (CR) with several TBBPA derivatives and other analogues. For these studies CR was calculated as

Soil samples were collected from the plow layer (0–10 cm) at the campus of the University of California at Davis. The soils were dried and sieved through a 100-mesh screen and kept in a sealed container until use. Two grams of soil was weighed and spiked with 10, 100, and 1000 ng g^–1 TBBPA. Ethyl acetate (15 mL) was added, and the sample was sonicated for 20 min. The extraction process was repeated twice, and the extracts were combined and centrifuged for 15 min at 3000 × g. The supernatant was evaporated and redissolved with 2 mL of MeOH. The blank fetal bovine serum (FBS) was bought from Gemini Bio Products (Calabasas, CA, U.S.A.). TBBPA was spiked into 1.5 mL of FBS with a final concentration of 10, 100, and 1000 ng mL^–1. A solid phase extraction (SPE) HLB cartridge (6 mL, Oasis, Waters) was conditioned with ethyl acetate and MeOH, and equilibrated with H₂O. After loading the supernatants, the cartridge was washed with 4 mL of H₂O (2% formic acid and 5% MeOH) and eluted with 2 mL of MeOH followed by 2 mL of ethyl acetate. The samples were divided for instrumental and ELISA analysis.

The VHH ELISA was validated with LC–MS/MS (Waters/Micromass, Manchester, UK), which was carried out on a Supelco Discover C18 column (5 cm × 2.1 mm, 5 μm, Sigma). Water (solution A) and acetonitrile containing 0.1% (v/v) acetic acid (solution B) were used as mobile phase with a flow rate of 0.3 mL min^–1. The volume of sample injection was 10 μL, running time was 5 min. The mass spectrometry was performed in a negative ESI mode, m/z 417 and 420 being the two daughter ions for quantitative tracing of the target.

The phage display VHH library was panned in both homologous format with T5-BSA and heterologous format with T1-BSA or T3-BSA. Both the concentrations of coating antigens and TBBPA were gradually decreased in an attempt to capture high-affinity binders. For all the selected phage clones showing binding to the coating antigens T1-BSA, T3-BSA, and T5-BSA, the 50% inhibition values of TBBPA in the competitive phage ELISA varied in a range of 1.0–10 ng mL^–1, <1.0 ng mL^–1, and >10 ng mL^–1, respectively. Based on the sensitivities, the strongest binders, clones T1–4, T3–15, and T5–10, were chosen from selections using templates T1-BSA, T3-BSA, and T5-BSA, respectively. Since the structure of T3 is similar to the immunizing hapten T5, all positive clones had better binding ability with T3-BSA and T5-BSA than with other coating antigens (T1-, T2-, T4-, and T6-BSA) (Table 1). T5–10 recognized most of the coating antigens except T4-BSA which lacks the bromines while T3–15 only recognized T3-BSA and T5-BSA (Table 1). The binding activity of T5–10 is similar to that of the PAb previously produced with immunogen T5-KLH.¹⁶ Hapten T6 with the fragment of TBBPA was only recognized by T5–10. Since T1–4 was isolated from T1-BSA, this clone showed moderate binding ability with both T1-BSA and T2-BSA that varied in linker length.

The three VHHs showed different affinities to TBBPA, because the sensitivities of VHH ELISAs (IC₅₀) varied in a range of 0.41–7.5 ng mL^–1 (Table 1). Clone T5–10 from the homologous VHH panning system showed less sensitivity than clones T1–4 and T3–15 from the heterologous VHH panning system. It is possible that VHH T5–10 has higher binding ability to the linker of coating antigens than VHH T1–4 and VHH T3–15 and this nonspecific binding is difficult to inhibit by free TBBPA. VHH T5–10 showed better selectivity for TBBPA in the homologous assay with T5-BSA than in the heterologous assays with T1-, T2-, T3-, and T6-BSA (Table 1), probably more nonspecific binding presented in the heterologous format than in the homologous one. VHH T1–4 may have less affinity to both coating antigen T1-BSA and TBBPA because analogues T1 and T5 were different in spatial configuration and linker length. Hapten T3 only being two carbons shorter in the linker than hapten T5 would allow VHHs to recognize TBBPA well with negligible recognition of the linker, and the most sensitivity was observed in the combination of VHH T3–15 and T3-BSA (Table 1). Consequently, the most sensitive assays were primarily determined by the heterologous antigen selection, rather than coating antigens competitor optimization. Therefore, T3-BSA was used as the coating antigen in both panning and testing systems for the remaining studies.

The titer of the input phage with T3-BSA selection in each cycle was 10¹³ cfu mL^–1, while the titer of output phage gradually increased from 9.6 × 10⁷ to 1.3 × 10⁹ cfu mL^–1 (Table S-1 in Supporting Information [SI]), showing an enrichment of the specific phages. After the fourth round of panning, 16 clones were selected, and all showed inhibition by TBBPA in the phage ELISA. The DNA was isolated, and sequencing revealed five groups of amino acid sequences. The VHH diversity³⁸ partly resulted from the extension of CDR1 and CDR3, antigen binding loops, frequency of mutation, and the paratope diversity.²⁴ Although there are only 12 amino acids or fewer in the CDR3 of T1–4, T3–15, and T5–10 in our study (Figure S-1 in SI) compared to the average 16 amino acid length observed in llama VHH,³⁹ the VHH showed a complex repertoire and high inhibition by TBBPA. Showing the highest sensitivity in the positive phage ELISAs, the phagemid vector containing VHH T3–15 was transferred to top 10 F′ cells, and the expressed VHH was purified with a Ni-NTA affinity column. The size and purity of VHH T3–15 were verified on a 15% SDS-PAGE gel with one major band at MW 17 kD.

TBBPA is a highly lipophilic (log K_ow = 4.5) compound and a water-miscible cosolvent is necessary in the assay buffer for analysis. MeOH and DMSO were reported to be efficient cosolvents for analyte–antibody reactions⁴⁰ and the effects of these solvents on the VHH T3–15 assay performance were studied (Figure S-2 in SI). The increase of MeOH concentration from 5% to 40% made the IC₅₀ vary between 0.42 and 1.33 ng mL^–1 (Figure S-2A in SI) and the maximal signal A₀ declined from 0.94 to 0.76. The most sensitive assay (IC₅₀ = 0.42 ng mL^–1) was observed at 10% MeOH with a reasonable A₀ (0.9) and the least sensitive (IC₅₀ = 1.33 ng mL^–1) was at 40% MeOH, with a lower A₀ (0.76). These data suggest that a high concentration of MeOH may interfere with the VHH binding with both the coating antigen and the target. The A₀ of the assay declined from 1.22 to 1.07 with the increasing of DMSO. The sensitivities in 5–40% of DMSO were in a range of 1.66–2.41 ng mL^–1 (Figure S-2B in SI), lower than those in MeOH. Thus, 10% MeOH, not only compatible with the VHH affinity but also showing efficient solubility of TBBPA, was chosen to optimize competitive VHH ELISAs.

The optimal concentrations of coating antigen T3-BSA and VHH T3–15 were determined by a checkerboard titration. Figure 2 is a typical calibration curve of competitive VHH ELISA for TBBPA under optimized conditions (10% MeOH, pH 7.4). The assay has a linear range (IC₂₀–IC₈₀) of 0.06–2.53 ng mL^–1, with an IC₅₀ value of 0.40 ng mL^–1 and an IC₁₀ value of 0.02 ng mL^–1. Compared to the heterologous competitive ELISA with alpaca antiserum (IC₅₀ = 500 ng mL^–1), the sensitivity was increased around 1000-fold in the VHH ELISA. In contrast with ELISAs based on PAbs¹⁶ and mAbs,¹⁸ the sensitivities were increased approximately 2-fold and 9-fold, respectively. Instead of using the solubilized VHH in the ELISA, displaying the VHH on phage particles showed a similar sensitivity (IC₅₀ = 0.46 ng mL^–1) to VHH ELISA, but a narrower linear range (0.19–1.07 ng mL^–1) than the latter.

The specificity of the VHH T3–15 ELISA was evaluated by comparing the IC₅₀ of TBBPA with those of several TBBPA derivatives, including 2,2′,6,6′-tetrabromobisphenol A diallyl ether (TBBPA-bAE) and tetrabromobisphenol A bis(2-hydroxyethyl) ether (TBBPA-bOHEE)s or other commonly used BFRs: hexabromocyclododecane (HBCD), PBDEs (BDE-47, BDE-99, BDE-100, BDE-153, BDE-154, and BDE-183), hydroxylated BDE-47 metabolites (5-OH-BDE-47 and 6-OH-BDE-47), 1,2-bis(pentabromodiphenyl) ethane (DBDPE), and bisphenol A (BPA) (Figure S-3 in SI). As shown in Table 2, the VHH assay is very selective for TBBPA because low cross-reactivities with the other tested compounds were observed (<0.1%). The high selectivity of the assay is useful to detect TBBPA which always coexists with its nonbrominated metabolite BPA as well as HBCDs and PBDEs in environmental matrices.^10,13,14

It is reported that thermal stability is one of the most extraordinary features of VHH.⁴¹ In this study, the VHH–antigen binding signal retained about 80% and 20% of the unheated control signal after heating at 90 °C for 10 and 90 min, respectively (Figure S-4A in SI). It was reported that conventional mAbs lost all binding activities within 5 min at 100 °C.⁴¹ The thermal stability of VHH could be explained by its extreme plasticity allowing refolding from the denatured state.⁴² Another reason for this property is cysteines forming in this case a disulfide bond, thus increasing thermal and conformational stabilities.⁴³ The thermal stability of VHH makes it preferable to conventional antibodies for on-site analysis.

The effect of the assay buffer pH on the VHH ELISA was evaluated (Figure S-4B in SI). The VHH ELISA was more susceptible to low pH (≤6.0) than to high pH (7.4–11.0), because low A₀ values were shown at pH 4.0 and 5.0 (0.2 and 0.3, respectively) and a very high IC₅₀ (8.2 ng mL^–1) was shown at pH 6.0, which may partly cause the denaturing of VHH. The protonated state of the protein at low pH might contribute to protein unfolding, or change surface charge or the antibody paratope. The isoelectric point (pI) of VHH T3–15 was estimated to be 9.56 according to the protein sequence and ExPASy (http://web.expasy.org/compute_pi/). When the pH of assay buffer was close to the pI of VHH T3–15, some antibodies may have precipitated; thus, the A₀ declined somewhat at pH 7.4–11.0 but the IC₅₀ varied in a narrow range of 0.48–0.98 ng mL^–1. The optimal pH of assay buffer was 7.4, but this assay was able to be performed at pH ranging from 8.0 to 11.0, with little sensitivity loss.

Matrix effect, inevitable in sample analysis, may cause some false positive results in the format used and can be eliminated with the dilution of the extract. However, dilution generally leads to a loss of assay sensitivity. For soil and FBS extracts, a 100-fold dilution and a 10-fold dilution with PBS were needed to minimize the matrix effect, respectively.

Extraction of TBBPA from soil by sonication was a simple, effective and time-saving method.⁶ MeOH, dichloromethane (DCM), ethyl acetate, and DCM/acetone (1:1, v/v) were initially tested. Ethyl acetate was proven to be an ideal extraction solvent for TBBPA from soil due to the acceptable recovery (Table 3). A reversed-phase solid phase extraction cartridge was used to remove the hydrophilic contaminants in FBS and extract the TBBPA from FBS with reasonable recoveries (Table 3).

The VHH ELISA for the spiked samples was validated by comparing the results of ELISA with those of an instrumental method LC–MS/MS. The recoveries of TBBPA from soil and FBS determined by VHH ELISA were in a range of 102.0–110.7% and 90.3–95.6%, respectively, and by LC–MS/MS in a range of 92.6–96.6% and 90.8–92.5%, respectively (Table 3). Recoveries in a range of 75–125% are usually acceptable for spiked samples.⁴⁴ Both methods showed good recoveries and correlated well with each other (Figure S-5 in SI). The VHH ELISA was demonstrated to be a valid method to detect TBBPA in soil and serum.