**2. Results**

#### *2.1. Monoclonal Antibody Production*

Mouse mAbs to AMAs were generated using the immunogen PERI-AMA-KLH [20]. Following the screening of the fusion plates, there were 14 positive cultures (optical density > 0.7), of which 12 cultures exhibited substantial signal reduction (optical density decreased by 0.5 or greater) in the presence of 100 ng mL−<sup>1</sup> α-AMA in cELISA (Figure 2). Only two (9C12 and 9G3) of these grew stably, and were cloned multiple times until every well of the cell culture plate with cell growth elicited a positive indirect ELISA response to the coating antigen, a periodate-oxidized form of α-AMA conjugated to bovine serum albumin (PERI-AMA-BSA). The resulting mAbs were AMA9G3 (American Type Culture Collection Accession number PTA-125922) and AMA9C12 (American Type Culture Collection Accession number PTA-125923). Both mAbs were isotype IgG1-possessing kappa light chains.

**Figure 2.** Hybridoma clone supernatants screened by indirect enzyme-linked immunosorbent assay (ELISA) (black bars) and by indirect competitive ELISA (gray bars). The cELISAs were completed using 100 ng mL−<sup>1</sup> of α-amanitin as the competing analyte.

#### *2.2. Cross-Reactivity and Sensitivity*

In order to determine how e ffective the assay would be in selectively detecting AMAs, a panel of cyclic peptides and smaller chemicals was tested. These included the bicyclic heptapeptides known as phallotoxins (phalloidin and phallacidin) also produced by *A. phalloides*, chemical toxins (psilocybin, muscimol, and ibotenic acid) produced by other mushrooms, and cyclic peptides (nodularin and microcystin-LR) produced by cyanobacteria. Of these analytes tested, AMA9G3 was competitively inhibited by all of the AMAs, α-AMA, β-AMA, and γ-AMA, while AMA9C12 was only competitively inhibited by α-AMA and γ-AMA (Table 1 and Figure 3). Both mAbs did not bind to any of the other compounds tested.


**Table 1.** Cross-reactivity (%) of compounds found in associated mushrooms or structurally related compounds.

**Figure 3.** Standard cELISA inhibition curves for both monoclonal antibodies AMA9G3 (solid line) and AMA9C12 (dashed line) against toxins α-amanitin (circles), β-amanitin (squares), and γ-amanitin (triangles). R<sup>2</sup> > 0.96 for the linear portion (including a minimum of three points) for every curve.

The standard curves for the binding of both mAbs to the α-AMA, β-AMA, and γ-AMA toxins are shown in Figure 3. There is no reduction in signal response for AMA9C12 when tested against different concentrations of β-AMA, whereas, for AMA9G3, all three of the toxins did competitively inhibit at higher concentrations. The steepness of the curve indicated a small dynamic range for the assay because of the dramatic signal change produced by very small changes in toxin concentration, but simple sample dilutions could be performed to achieve a signal that fits within this range.

While both mAbs exhibited competitive inhibition from α-AMA and γ-AMA, AMA9G3 exhibited slightly higher sensitivity, with an IC50 of 1.57 ng mL−<sup>1</sup> for α-AMA (Table 1 and Figure 3). For AMA9G3, the working range of detection (estimated as IC30 to IC80) of α-AMA is 0.7–3.1 ng mL−1, of γ-AMA is 0.5–2.4 ng mL−1, and of β-AMA is 3.6–129.1 ng mL−1. A conservative estimate for the limit of detection for α-AMA or γ-AMA with the AMA9G3 assay is 1 ng mL−1, accounting for the large (30%) variation in signal at low to no concentrations of toxin. Because of the propensity for samples (mushroom extracts) to contain all three AMAs, AMA9G3 was selected for use in the cELISAs for the extraction studies.

#### *2.3. Kinetic Measurements*

For all three antibodies (two mAbs from this study and one rabbit pAb #58 from our previous work), a final concentration of 10 nM was used in both the Equilibrium and Kinetics Injection studies. Table 2 shows the affinity (*K*d) and kinetic parameter (*k*on and *k*off) values obtained for each antibody tested against α-AMA as the free ligand. *K*d values of 10−<sup>10</sup> M and lower indicate that these antibodies

exhibit a high affinity for their target analyte (α-AMA). Given the similar affinity characteristics between the mAbs and the pAb, the major advantage of the mAbs over the pAb is their ability to produce a continuous supply of the same protein.


**Table 2.** Affinities (*K*d) and kinetic parameters (*k*on and *k*off) for antibodies binding to α-amanitin measured by KinExA.

#### *2.4. Mushroom Extraction*

For the purposes of exploring the feasibility of performing simplified and rapid extractions, five different extraction solutions were tested. The commonly employed extraction using methanol and dilute acid was compared to extractions with more innocuous reagents, such as phosphates, tris, and Tween-20. The extraction solutions were tested on three different mushroom species. Both *A. phalloides* and *A. ocreata* are known to contain AMAs, while *A. gemmata* is known to not contain AMAs.

AMAs were detected by this mAb-based cELISA in both of the extracts for the species known to contain AMAs (*A. phalloides* and *A. ocreata*) (Figure 4a,b). Detection is indicated by 100% inhibition at dilutions up to and including a 9000-fold dilution of the extract (Figure 4a,b). With increasing dilutions, it would be expected that the amount of inhibition would decrease as the AMA concentration decreases in a dose-dependent manner. Indeed, at dilutions greater than 9000-fold, decreased inhibition was observed for all extraction conditions.

**Figure 4.** *Cont.*

**Figure 4.** Representative cELISA inhibition profiles illustrating the amount of inhibition from the tested extraction conditions at varying dilutions of mushroom extracts obtained from (**a**) *Amanita phalloides*, (**b**) *A. ocreata*, and (**c**) *A. gemmata*. The extraction conditions were as follows: (1) MeOH: methanol: water: HCl for 1 h; (2) H2O: diH2O, 1 min, (3) PB: phosphate buffer, 1 min, (4) PBT: PB with Tween-20, 1 min, and (5) TBST: tris-bu ffered saline with Tween-20, 1 min.

There appeared to be very little di fference between the di fferent extraction conditions. The methanol extractions had a slightly higher inhibition at 27,000-fold and 81,000-fold dilutions for *A. phalloides* and the 27,000-fold dilution for *A. ocreata*. This may sugges<sup>t</sup> that the methanol extractions extracted more AMAs than the aqueous extractions, since more inhibition equates to more toxin. However, the methanol extractions were carried out for 1 h, while the aqueous bu ffers were only carried out for 1 min. So, we cannot conclude whether time or composition contributed to the slight di fferences in extraction e fficiency. However, because the di fferences were slight, we conclude that these aqueous bu ffers, with only a 1 min shaking step, were highly e ffective methods for AMA extraction.

With competitive-type assays, the signal intensity is the greatest when the free toxin to be detected is the lowest. In this case, low toxin levels are seen at the highest dilutions, and high signal intensities exhibit the most variation. With the data plotted as percent inhibition, higher variation was seen at higher dilutions (Figure 4). Thus, the background was estimated to be around 20–30% inhibition. Extracts from the non-toxin containing mushroom *A. gemmata* (Figure 4c) were intentionally tested at the more concentrated (e.g., 1- and 3-fold) dilutions, since no toxin was expected to be detected. The strong distinction between samples that exhibit 100% inhibition (i.e., with AMAs) and those without toxin demonstrated the ability of this mAb-based cELISA to selectively and sensitively detect AMAs from mushroom extracts.
