Development of a New Monoclonal Antibody against Brevetoxins in Oyster Samples Based on the Indirect Competitive Enzyme-Linked Immunosorbent Assay

The consumption of shellfish contaminated with brevetoxins, a family of ladder-frame polyether toxins formed during blooms of the marine dinoflagellate Karenia brevis, can cause neurotoxic poisoning, leading to gastroenteritis and neurotoxic effects. To rapidly monitor brevetoxin levels in oysters, we generated a broad-spectrum antibody against brevetoxin 2 (PbTx-2), 1 (PbTx-1), and 3 (PbTx-3) and developed a rapid indirect competitive enzyme-linked immunosorbent assay (icELISA). PbTx-2 was reacted with carboxymethoxylamine hemihydrochloride (CMO) to generate a PbTx-2-CMO hapten and reacted with succinic anhydride (HS) to generate the PbTx-2-HS hapten. These haptens were conjugated to keyhole limpet hemocyanin (KLH) and bovine serum albumin (BSA) to prepare immunogen and coating antigen reagents, respectively, using the active ester method. After immunization and cell fusion, a broad-spectrum monoclonal antibody (mAb) termed mAb 1D3 was prepared. The 50% inhibitory concentration (IC50) values of the icELISA for PbTx-2, PbTx-1, and PbTx-3 were 60.71, 52.61, and 51.83 μg/kg, respectively. Based on the broad-spectrum mAb 1D3, an icELISA was developed to determine brevetoxin levels. Using this approach, the limit of detection (LOD) for brevetoxin was 124.22 μg/kg and recoveries ranged between 89.08% and 115.00%, with a coefficient of variation below 4.25% in oyster samples. These results suggest that our icELISA is a useful tool for the rapid monitoring of brevetoxins in oyster samples.


Introduction
Marine biotoxins have negative effects on the seafood industry. Typically, they are transferred along food chains and affect other organisms, including humans. Different types of poisoning are induced by marine biotoxins, e.g., puffer fish poisoning, paralytic shellfish poisoning, scombroid fish poisoning, diarrhetic shellfish poisoning, neurotoxic shellfish poisoning, ciguatera fish poisoning, and amnesic shellfish poisoning [1]. Brevetoxins belonging to the neurotoxic shellfish poisoning group are produced by the Florida red tide organism Karenia brevis and are divided in two groups: (1) those derived from the brevetoxin A backbone (PbTx-1, PbTx-7, and PbTx-10) and (2) those from brevetoxin B (PbTx-2, PbTx-3, PbTx-5, PbTx-6, PbTx-9, PbTx-11, and PbTx-12) [2]. PbTx-1 is the most potent, while PbTx-2 is the most highly produced brevetoxin ( Figure 1) [3]. K. brevis blooms occur most years in brevis blooms occur most years in the Gulf of Mexico, killing high numbers of fish a marine mammals, including sea turtles and aquatic birds, and generate economic loss of USD 2-24 million [4]. Physiologically, brevetoxins appear to activate voltage-sensiti sodium channels causing sodium influx and nerve membrane depolarization resulting respiratory distress [5]. Thus, to protect human health and avoid food poisoning brevetoxins, a rapid, sensitive, and specific assay for brevetoxins detection is required. Several analytical methods have been established to detect brevetoxins, includi liquid chromatography-tandem mass spectrometry (LC-MS/MS) and receptor/antibod based immunoassays. For example, Shin et al. (2018) developed an LC-MS/MS proto for the PbTx-1, PbTx-2, and PbTx-3 brevetoxins with a limit of quantification (LOQ) of μg/kg for each toxin [6]. Similarly, Wunschel et al. (2018) developed an electrospray L MS/MS system for the same brevetoxins with an LOQ of 2.5 μg/kg for each toxin [7]. Do et al. (2018) established a high-resolution LC-MS system to detect 14 brevetoxins, w LOQs of 312 and 324 μg/kg for PbTx-2 in mussel and oyster, respectively [8]. Howev these analytical methods are often time-consuming, expensive, and involve complex sa ple preparation. Therefore, new rapid brevetoxin screening/detection methods are quired. McCall et al. (2012) developed a competitive binding assay based on rat brain sy aptosomes as receptors and BODIPY ® -PbTx-2 as competitive fluorescent probes brevetoxin analogs [9]. In addition, Murata et al. (2019) developed a chemiluminesce receptor binding assay based on rat brain synaptosomes and acridinium-PbTx-2, with detection limit value of 1.4 amol [10]. However, the synaptosomes were unstable and quired storage at -80°C, and the assay was a time-consuming process.
Indirect competitive enzyme-linked immunosorbent assays (icELISAs) and late flow immunoassays (LFAs) are antibody-based and are frequently used as screeni methods for small molecules due to their rapid turnaround, low costs, and high sensit ity. Recently, Zhou et al. (2010) prepared the monoclonal antibody (mAb) 2C4 using t PbTx-2-HS hapten; it exhibited IC50 values of 6.40, 6.57, and 5.31 μg/kg against PbTx PbTx-1, and PbTx-3, respectively, with the accompanying icELISA having a limit of d tection (LOD) of 0.60 ng/well [11]. Zhou et al. (2009) developed an LFA based on a coll dal gold probe for the rapid detection of brevetoxins in fish product samples, with a visu detection limit of 20 μg/kg [12]. Lai et al. (2016) developed a novel colorimetric immun assay for PbTx-2 using an enzyme-controlled Fenton-based reaction and a 3,3',5,5'-tet methylbenzidine-based visual colored system, with an LOD of 0.08 ng/kg [13]. As Several analytical methods have been established to detect brevetoxins, including liquid chromatography-tandem mass spectrometry (LC-MS/MS) and receptor/antibodybased immunoassays. For example, Shin et al. (2018) developed an LC-MS/MS protocol for the PbTx-1, PbTx-2, and PbTx-3 brevetoxins with a limit of quantification (LOQ) of 25 µg/kg for each toxin [6]. Similarly, Wunschel et al. (2018) developed an electrospray LC-MS/MS system for the same brevetoxins with an LOQ of 2.5 µg/kg for each toxin [7]. Dom et al. (2018) established a high-resolution LC-MS system to detect 14 brevetoxins, with LOQs of 312 and 324 µg/kg for PbTx-2 in mussel and oyster, respectively [8]. However, these analytical methods are often time-consuming, expensive, and involve complex sample preparation. Therefore, new rapid brevetoxin screening/detection methods are required. McCall et al. (2012) developed a competitive binding assay based on rat brain synaptosomes as receptors and BODIPY ® -PbTx-2 as competitive fluorescent probes for brevetoxin analogs [9]. In addition, Murata et al. (2019) developed a chemiluminescent receptor binding assay based on rat brain synaptosomes and acridinium-PbTx-2, with a detection limit value of 1.4 amol [10]. However, the synaptosomes were unstable and required storage at -80 • C, and the assay was a time-consuming process.
Indirect competitive enzyme-linked immunosorbent assays (icELISAs) and lateral flow immunoassays (LFAs) are antibody-based and are frequently used as screening methods for small molecules due to their rapid turnaround, low costs, and high sensitivity. Recently, Zhou et al. (2010) prepared the monoclonal antibody (mAb) 2C4 using the PbTx-2-HS hapten; it exhibited IC 50 values of 6.40, 6.57, and 5.31 µg/kg against PbTx-2, PbTx-1, and PbTx-3, respectively, with the accompanying icELISA having a limit of detection (LOD) of 0.60 ng/well [11]. Zhou et al. (2009) developed an LFA based on a colloidal gold probe for the rapid detection of brevetoxins in fish product samples, with a visual detection limit of 20 µg/kg [12]. Lai et al. (2016) developed a novel colorimetric immunoassay for PbTx-2 using an enzyme-controlled Fenton-based reaction and a 3,3',5,5'-tetramethylbenzidinebased visual colored system, with an LOD of 0.08 ng/kg [13]. As is known, the preparation of broad-spectrum antibodies is a key step for developing an immunoassay. Hapten design is an important feature when preparing antibodies against target compounds. In the literature, PbTx-2 always reacted with succinic anhydride (HS) which introduced active sites as hapten [11,14]. In principle, aldehyde groups from PbTx-2 may be conjugated with a spacer arm, such as carboxymethoxylamine hemihydrochloride (CMO) and/or aminobenzoic acid. In addition, molecular modeling has become a powerful tool in guiding and improving hapten design strategies [15][16][17][18][19]. Therefore, we explored and developed novel haptens using molecular modeling to prepare mAbs against brevetoxins. We then developed an icELISA method to detect brevetoxins in oyster samples.
Eight-week-old female BALB/c mice were provided by Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China) and raised under strictly controlled conditions. The mice were manipulated according to the China Agricultural University regulations concerning the protection of animals used for scientific purposes (2010-SYXK-0037).

Preparation of PbTx-2-Protein Conjugates
PbTx-2 was reacted with CMO to insert a carboxyl group to facilitate coupling to a carrier protein, as previously described but with modifications ( Figure 2A) [21,22]. Briefly, 2 mg PbTx-2 in 5 mL pyridine was reacted with 2 mg CMO at 90 • C for 6 h, after which the reaction mixture was evaporated to dryness under nitrogen gas. Then, the residue was dissolved in 5 mL 0.1 mol/L NaHCO 3 and extracted by 5 mL ethyl acetate. The aqueous phase was adjusted to pH 3 with 0.05 mol/L HCl and extracted three times in 5 mL ethyl acetate. The organic phase was dried under N 2 at 40 • C to generate the PbTx-2-CMO hapten. The PbTx-2-HS hapten was similarly synthesized by reacting PbTx-2 with HS as described ( Figure 2B). mg N,N'-dicyclohexylcarbodiimide and 1.5 mg N-hydroxysuccinimide and reacted for 12 h at room temperature. After centrifugation at 8000× g for 10 min, the supernatant of each active hapten solution was divided into two and added drop-wise to 4 mL PBS containing 5 mg KLH and 10 mL PBS containing 10 mg BSA, respectively. Reaction mixtures were stirred at 4 °C for 12 h, and the conjugates of PbTx-2-CMO-KLH/BSA and PbTx-2-HS-KLH/BSA were dialyzed in PBS for 48 h. Antigens were then stored at −20 °C.

Preparation of mAbs
Animal immunization procedures were as follows: twelve female BALB/c mice were immunized by subcutaneous injection. The immunogens PbTx-2-CMO-KLH and PbTx-2-HS-KLH (100 μg) were emulsified in Freund's complete adjuvant for the first immunizations. After this, the mice were boosted with the same immunogen doses in Freund's incomplete adjuvant every 3 weeks. Then, 7-10 days after the last immunization, serum was collected from the caudal vein. Antibody titers were analyzed by ELISA and specificity was characterized by icELISA. Animals with the highest inhibition ratios were sacrificed for fusion studies [23]. The inhibition ratio was calculated as follows: Then, the PbTx-2-CMO and PbTx-2-HS haptens were conjugated to KLH (immunogen) and BSA (coating antigen), respectively, via the active ester method [23]. Briefly, the haptens were respectively redissolved in 0.5 mL N,N-dimethylformamide containing 2 mg N,N'dicyclohexylcarbodiimide and 1.5 mg N-hydroxysuccinimide and reacted for 12 h at room temperature. After centrifugation at 8000× g for 10 min, the supernatant of each active hapten solution was divided into two and added drop-wise to 4 mL PBS containing 5 mg KLH and 10 mL PBS containing 10 mg BSA, respectively. Reaction mixtures were stirred at 4 • C for 12 h, and the conjugates of PbTx-2-CMO-KLH/BSA and PbTx-2-HS-KLH/BSA were dialyzed in PBS for 48 h. Antigens were then stored at −20 • C.

Preparation of mAbs
Animal immunization procedures were as follows: twelve female BALB/c mice were immunized by subcutaneous injection. The immunogens PbTx-2-CMO-KLH and PbTx-2-HS-KLH (100 µg) were emulsified in Freund's complete adjuvant for the first immunizations. After this, the mice were boosted with the same immunogen doses in Freund's incomplete adjuvant every 3 weeks. Then, 7-10 days after the last immunization, serum was collected from the caudal vein. Antibody titers were analyzed by ELISA and specificity was characterized by icELISA. Animals with the highest inhibition ratios were sacrificed for fusion studies [23]. The inhibition ratio was calculated as follows: where B 0 and B correspond to the absorbance value of wells without a standard and the absorbance value of wells with x µg/kg PbTx-2 standard, respectively. Spleen cells from animals with the highest inhibition ratios were collected and fused with Sp2/0 myeloma cells using PEG1450 at a 10:1 ratio. Fusion cells were cultured in hypoxanthine aminopterin thymidine medium for 7 days. Supernatants from fusion cultures were also assayed for the titer and the inhibition ratios using ELISA and icELISA. Positive hybridomas were subcloned three times using the limiting dilution method and injected into BALB/c mice to produce ascites [23].

ELISA and icELISA Protocols
ELISA was conducted as previously described [24]. Briefly, (1) 100 µL of coating antigen PbTx-2-CMO-BSA (or PbTx-2-HS-BSA) diluted in coating buffer at 1 µg/mL was added to the wells of a 96-well plate and incubated at 4 • C for 10-12 h. (2) The coating antigen was then discarded, and the plate was washed three times in wash buffer (250 µL/well).
(3) Then, 200 µL of 2% skimmed milk powder in assay buffer was added per well and incubated at 37 • C for 1 h to reduce unspecific binding. (4) After this, 50 µL 10 mM PBS and 50 µL anti-PbTx-2 mAb diluted in assay buffer were added sequentially to wells and incubated at 37 • C for 30 min. (5) After further washing, 100 µL/well peroxidaseconjugated goat anti-mouse IgG (diluted 1:5000) was added and incubated at 37 • C for 30 min. (6) The plate was washed five times, 100 µL/well substrate solution was added to the reaction, and the plate was incubated at 37 • C for 15 min. (7) The reaction was stopped with 50 µL 2 M H 2 SO 4 , and the absorbance was detected at 450 nm on a Multiskan MK3 microplate reader.
For the icELISA procedure, the 50 µL 10 mM PBS was replaced by 50 µL series concentration of PbTx-2 standard solution at step 4.

Sample Preparation
Blank oyster samples from a local supermarket were confirmed using the HPLC-MS/MS method [6]. The blank oyster samples were spiked with 200, 400, and 800 µg/kg. Then, brevetoxins were extracted from the oysters based on a previously described method with some modifications [6]. Briefly, 5 g homogenized sample was extracted in 5 mL 80% methanol by vortexing for 5 min and then ultrasonication at 60 • C for 10 min. After centrifugation at 2400× g for 10 min, the supernatant was diluted 15-fold in assay buffer to eliminate matrix interference. The LOD value in oyster samples was calculated based on 20 blank samples, accepting no false positive rates, with an average value plus triple standard deviation (SD), then multiplied by 15 (the diluted ratio) [23]. The recovery was calculated as the following equation [23]: Recovery (%) = (measured value (µg/kg)/spiked values (µg/kg)) × 100% Foods 2021, 10, 2398 6 of 11

Hapten Design, Synthesis, and Conjugate Preparation
Hapten design is key to generating excellent antibody performances against small molecules [25]. Generally, haptens should mimic the target molecule in terms of size, shape, electronic properties, and insert functional groups such as, carboxyl, amino, and sulfhydryl groups for carrier protein coupling [16,17]. In principle, there are two ways to synthesize PbTx-2 haptens. First, the hydroxy group of PbTx-2 is reacted with HS, which introduces active sites to construct the PbTx-2-HS hapten ( Figure 2A) [11,14]. Moreover, the PbTx-2 aldehyde group is an active site which potentially reacts with an amino group with the objective of obtaining a probe or hapten.  [10]. Thus, PbTx-2 was reacted with CMO to prepare the hapten of PbTx-2-CMO ( Figure 2B).
To further design the optimal hapten, PbTx-2 was selected as the template molecule to align the haptens of PbTx-2-CMO and PbTx-2-HS based on their lowest energy conformation. As shown in Figure 3A, the PbTx-2-CMO and PbTx-2-HS haptens were exposed on the left side of the structure to animal immunity. In addition, the introduction of spacer arms at the O57 or O61 position of PbTx-2 barely influenced the atom partial charges feature of PbTx-2 ( Figure 3B). However, the introduction of HS hapten caused the spacer arm to form a certain angle with the parent structure of PbTx-2 ( Figure 3A). To further explain the difference between PbTx-2-CMO, PbTx-2-HS, and the target compound PbTx-2 in terms of conformation and electron distribution, the ESP displayed on the van der Waals surfaces of global lowest energy conformation for PbTx-2, PbTx-2-CMO, and PbTx-2-HS is shown ( Figure 3C). The PbTx-2-CMO conformation had the most similar structure to the target, PbTx-2 ( Figure 3C, points A and B). The spacer HS arm of PbTx-2-HS formed a specific spatial conformation with the parent nucleus structure ( Figure 3C, point C), non-conducive to the production of high-affinity antibodies against the target. Thus, PbTx-2-CMO was the ideal hapten to be used for antibody production against PbTx-2. To further verify the quality of haptens, PbTx-2-CMO and PbTx-2-HS were synthesized and conjugated with the carrier protein as complete antigens.
PbTx It is accepted that the coupling ratio is an important factor affecting generated antibodies; higher coupling ratios could induce higher antibody titers [17]. In the literature, appropriate coupling ratios ranged from 3 to 15 [16,17,21]. The coupling ratios of PbTx-2-CMO-KLH/BSA and PbTx-2-HS-KLH/BSA were relatively low due to the lower reaction molar ratio (hapten to carrier protein). However, the conjugation of PbTx-2-CMO- KLH/BSA and PbTx-2-HS-KLH/BSA was successful according to the UV-visible absorption spectra (Figure 4). Therefore, these antigens were used for immunization studies. The maximum absorbance peaks between conjugates and PbTx-2 had shifted, thus indicating the successful synthesis of complete antigens. The calculated molar ratio of hapten to carrier protein was 1.5 for PbTx-2-CMO-KLH, 0.9 for PbTx-2-HS-KLH, 1.2 for PbTx-2-CMO-BSA, and 0.6 for PbTx-2-HS-BSA. It is accepted that the coupling ratio is an important factor affecting generated antibodies; higher coupling ratios could induce higher antibody titers [17]. In the literature, appropriate coupling ratios ranged from 3 to 15 [16,17,21]. The coupling ratios of PbTx-2-CMO-KLH/BSA and PbTx-2-HS-KLH/BSA were relatively low due to the lower reaction molar ratio (hapten to carrier protein). However, the conjugation of PbTx-2-CMO-KLH/BSA and PbTx-2-HS-KLH/BSA was successful according to the UV-visible absorption spectra (Figure 4). Therefore, these antigens were used for immunization studies.

Antibody Production and Characterization
The two immunogens PbTx-2-CMO-KLH and PbTx-2-HS-KLH were used to generate antibodies against PbTx-2. After a third immunization, serum was collected and char-

Antibody Production and Characterization
The two immunogens PbTx-2-CMO-KLH and PbTx-2-HS-KLH were used to generate antibodies against PbTx-2. After a third immunization, serum was collected and characterized using ELISA and icELISA (Table 1). Mice immunized with both complete antigens produced antiserum against PbTX-2; the immune response to PbTx-2-CMO-KLH was superior to that to PbTx-2-HS-KLH from an inhibition rate perspective. However, the PbTx-2-HS-KLH antibody titer was higher than that for PbTx-2-CMO-KLH, possibly suggesting the spacer HS arm of the PbTx-2-HS molecule that formed based on molecular modeling ( Figure 3A,C), a specific spatial conformation with the parent nucleus structure conduced to the production of high-titer antibodies. Additionally, all the antiserum titers from mice were low due to the lower coupling ratios of the hapten to the carrier protein. Finally, mouse No. 1 (PbTx-2-CMO-KLH) and mouse No. 5 (PbTx-2-HS-KLH) were sacrificed for fusion studies as they exhibited a higher inhibition ratio and antiserum titer. After cell fusion, the 1D3 cell line from mouse No.1 (PbTx-2-CMO-KLH) and the cell lines 6D8 and 9E8 from mouse No. 5 (PbTx-2-HS-KLH) were identified as secreting antibodies against PbTx-2. Thus, all three cell lines were used for antibody production, and ELISAs and icELISAs were used to characterize mAbs ( Table 2). All mAbs recognized PbTx-2, PbTx-1, and PbTx-3. Additionally, the IC 50 value for the 1D3 mAb was lower than that for the 6D8 and 9G8 mAbs, consistent with the molecular modeling data (Figure 3). Those results also indicated that PbTx-2-CMO was the best hapten for brevetoxin production, and the PbTx-2-HS-BSA coating antigen improved 1D3 mAb sensitivity. a The concentration of coating antigen PbTx-2-HS-BSA was 0.20 µg/mL, and the antibody titer was 1:1 × 10 5 . b The concentration of coating antigen PbTx-2-CMO-BSA was 1.00 µg/mL, and the antibody titer was 1: 1 × 10 4 . c The concentration of coating antigen PbTx-2-HS-BSA was 0.20 µg/mL, and the antibody titer was 1: 3 × 10 4 .
Next, PbTx-2-HS-BSA and 1D3 mAb were used to establish a standard curve in buffer assay. As shown (Figure 5), the curves that were based on the PbTx-2, PbTx-1 and PbTx-3 gave IC 50 values of 60.71 µg/kg, 52.61 µg/kg and 51.83 µg/kg, respectively. The mAb 1D3 exhibited a CR = 100% toward PbTx-2, CR = 115.40% toward PbTx-1, and CR = 117.13% toward PbTx-3. The LOD in assay buffer was 6.11 µg/kg, and the linearity range (IC 10 -IC 90 ) was between 7.46 and 127.00 µg/kg. In addition, the mAb did not exhibit a measurable CR with other marine biotoxins, including domoic acid, microcystins, nodularin, neosaxitoxin, and tetrodotoxin. These data indicated that mAb 1D3 was a broad spectrum homogeneous antibody for brevetoxins.. The sensitivity of the mAb 1D3 icELISA was lower than that for the mAb 2C4 (IC 50 value of 5.3 µg/L towards PbTx-2) [11]. In this study, the lower coupling ratios of PbTx-2-CMO to the carrier protein was because of the lower feed ratios (2 mg PbTx-2 reacted with 2 mg CMO), resulting in lower antibody titer immunized by PbTx-2-CMO-KLH. Ultimately, this affected the performance parameters of the mAb 1D3. More sensitive antibodies can be obtained if the ratios of haptens (PbTx-2-CMO) to carrier proteins are improved. The recoveries for PbTx-2, PbTx-1, and PbTx-3 from spiked oyster samples at three dose levels (200, 400, and 800 μg/kg) are shown in Table 3 and were 91.21-108.33%, 91.04-115.00%, and 89.08-113.17%, respectively. The LOD in the oyster sample was calculated at 124.22 μg/kg. The sensitivity of this icELISA was higher than the high-resolution LC-MS sensitivity (LOQ of 324 μg/kg for PbTx-2 in oyster samples) [8] and similar to that of the LFA based on colloidal gold probe (visual detection limit of 20 μg/kg in fish product samples) by Zhou et al. [12]. 89.08 ± 1.12%

Conclusions
Using molecular modeling and experimental analyses, the PbTx-2-CMO hapten produced acceptable antibody characteristics against brevetoxins. The IC50 values against The recoveries for PbTx-2, PbTx-1, and PbTx-3 from spiked oyster samples at three dose levels (200, 400, and 800 µg/kg) are shown in Table 3 and were 91.21-108.33%, 91.04-115.00%, and 89.08-113.17%, respectively. The LOD in the oyster sample was calculated at 124.22 µg/kg. The sensitivity of this icELISA was higher than the high-resolution LC-MS sensitivity (LOQ of 324 µg/kg for PbTx-2 in oyster samples) [8] and similar to that of the LFA based on colloidal gold probe (visual detection limit of 20 µg/kg in fish product samples) by Zhou et al. [12].