*3.2. Enzyme-Linked Immunosorbent Assay (ELISA)*

According to the literatures, ELISA is the most popular and most frequently used technique for mycotoxin analysis especially aflatoxins [56]. Among different types of ELISA,

the direct competitive ELISA is commonly used in mycotoxin detection. In recent years, only a few studies have been focused on mycotoxin detection using ELISA method. However, there are a large number of commercial ELISA kits produced by different companies worldwide.

Traditional ELISA uses antibody as recognition element and HRP-catalyzed TMB to generate color as a signal reporter. Although ELISA has been recognized as an excellent and accurate method for mycotoxin analysis, but the procedure is somewhat time-consuming (incubation time of approximately 1–2 h), uneconomical, unsuitable for field testing due to the need for specialist plate readers, and unreliable due to the similarity of the structure of mycotoxins, which causes false positive results [57]. Therefore, many efforts have been made to improve the shortcomings mentioned. One attempt is to improve the colorimetric signal. Conventional colorimetric signal using HRP and TMB is not suitable for naked-eye detection in deprived areas with limited resources because a plate reader is required to distinguish the tonality of analytes with similar concentrations. Recently, colorimetric ELISA has gained considerable attention due to its simple readout without specialist devices. Acid–base indicators are ideal signal reporters for naked-eye distinction because most of them provide a significant contrasting color at their titration end points under a narrow pH range. Several enzymes including alkaline phosphatase, urease and penicillinase have been used in ELISA to change the pH through catalyzing the related specific substrate to produce hydrogen or hydroxide ions [58]. In this regard, Xinog et al. developed a direct competitive colorimetric ELISA using glucose oxidase (GOx) as an alternative to HRP for glucose oxidation into gluconic acid and H2O<sup>2</sup> (Figure 4a) [58]. The pH indicator bromocresol purple (BCP), which was highly sensitive to pH variation, was applied as signal output. BCP indicator showed a vivid color change from yellow to grayish purple in the presence of 100 pg·mL−<sup>1</sup> AFB1. Therefore, the cutoff limit was determined to be 100 pg·mL−<sup>1</sup> by the naked eye. The developed GOx-based colorimetric ELISA exhibited high sensitivity and excellent selectivity with IC50 value at 66.27 pg·mL−<sup>1</sup> , which was approximately 10-fold lower than that of traditional HRP-based ELISA. However, long incubation time and multi-step washing were still the major limitations of the ELISA method. The proposed assay was applied for AFB1 determination in corn samples with acceptable accuracy and precision.

Among colorimetric ELISA methods, plasmonic ELISA is another attempt with simple readout format suitable for on-site detection. Gold nanoparticles (AuNPs) are good candidate as colorimetric indicator in plasmonic ELISA due to high molar extinction coefficient and localized surface plasmon resonance (LSPR) characteristics. The LSPR of AuNPs is related to their size, shape, composition, and agglomerate mood [59]. The LSPR variation of AuNPs generates a significant color change that is easily observable by the naked eyes. Based on differences in producing LSPR mechanism, plasmonic ELISA is classified into four types that employ the aggregation, etching, controlled growth kinetics, and AuNPs metallization. Among them, enzyme-induced silver metallization on the AuNPs surface can produce a remarkable LSPR, and provide a multicolor change in the solution [60]. Several enzymes GOx, alkaline phosphatase and β-galactosidase have been used to catalyze their substrates and produce reducing agents such as H2O2, ascorbic acid, and p-aminophenol which can reduce the silver ions on the AuNPs surface. In this regard, Pei et al. developed a colorimetric plasmonic ELISA for OTA detection based on the urease-induced metallization of gold nanoflowers (AuNFs) [60]. OTA-labeled urease was employed as competing antigen to hydrolyze urea into ammonia. In the presence of ammonia, silver ions were reduced by the formyl group from glucose to produce a silver shell around AuNFs resulted in the solution color change from blue to brownish red (Figure 4b). The plasmonic ELISA exhibited high sensitivity with a cutoff limit of 40 pg·mL−<sup>1</sup> and LOD at 8.205 pg·mL−<sup>1</sup> (19-folds lower than those of HRP-based ELISA). The proposed procedure provided a highly selective and sensitive, simple, robust, and high-throughput screening method for the quantitative determination of OTA in food and feed samples. However, it suffered from long incubation time.

**Figure 4.** (**a**) Colorimetric GOx-based ELISA using acid-base indicator bromocresol purple **Figure 4.** (**a**) Colorimetric GOx-based ELISA using acid-base indicator bromocresol purple (BCP) for AFB1 detection; (**b**) Plasmonic ELISA based on the urease-induced metallization of gold nanoflowers for OTA detection; (**c**) DLS-ELISA method associated with H2O<sup>2</sup> -mediated tyramine signal amplification system for AFB1 detection. Reproduced from [58,60,61], respectively, with permission.

−1 In another plasmonic ELISA, aggregation-induced color change of AuNPs, as a main strategy to regulate the plasmonic signal, was employed for the ultrasensitive detection of AFB1 using dynamic light scattering (DLS) signal instead of absorbance (Figure 4c) [61]. In the developed DLS-ELISA, GOx-AFB1 was used as competing antigen because the GOx can effectively convert glucose to H2O2. Then, the produced H2O<sup>2</sup> converted into hydroxyl radical in the presence of HRP to induce AuNPs aggregation. Indeed, H2O2-mediated TYR was used as signal amplification system. The DLS-ELISA exhibited a LOD as low as 0.12 pg·mL <sup>−</sup><sup>1</sup> which was about 153- and 385-folds lower than those of conventional plasmonic and colorimetric ELISA, respectively. The ultrahigh sensitivity is attributed to the high sensitivity of light-scattering intensity to particle size changes. The DLS-ELISA was employed for AFB1 detection in corn samples with good reliability and precision.

Mukherjee et al. compared aptamer-based enzyme linked apta-sorbent assay (ELASA) with antibody-based ELISA and its potential to replace antibodies in usual immunoassay formats either as capture probe or detection probe without affecting the sensitivity [62]. The ELASA was based on the principle of target capture by aptamer where, OTA specific aptamer was used for toxin detection. Then, anti-OTA IgG primary antibody and anti-rabbit secondary antibody labeled with alkaline phosphatase (ALP) where added as detection agents. The colorimetric signal was produced under addition of para-nitrophenyl phosphate (p-NPP) as substrate. The LOD was obtained 0.84 pg·mL−<sup>1</sup> . The developed ELASA exhibited a similar sensitivity to the conventional antibody-based ELISA with a LOD of 1.13 pg mL−<sup>1</sup> . However, the OTA aptamer showed about 40% cross-reactivity with aflatoxins. By selecting aptamer with a low percentage of cross-reactivity, ELASA can be a good alternative to the conventional ELISA. The proposed ELASA was used for OTA detection in groundnut and coffee bean.

Another innovation in improving the ELISA characteristics is to replace nanomaterialbased enzyme mimics (nanozymes) as artificial enzymes with natural enzymes. Nanozymes exhibit excellent properties such as easy synthesis, high stability, low cost, and design flexibility. Different kinds of nanomaterials—including noble metal nanoparticles (e.g., AuNP and AgNPs), graphene oxide, magnetic iron oxide, etc.—have been used in sensing methods. Xu et al. proposed a nanozyme-linked immunosorbent assay using metal–organic frameworks (MOFs) for AFB1 detection [63]. MOF with peroxidase-like activity was replaced with HRP for antibody labeling and catalyzing TMB to generate colorimetric signal. The MOF-ELISA system increased the accuracy of detection and inhibited false positive problems in the detection method, indicating that MOFs exhibited better catalytic activity and more stability than HRP. The LOD was obtained 0.009 ng·mL−<sup>1</sup> which was 20-folds lower than those of HRP-based ELISA. The proposed ELISA was employed for AFB1 detection in peanut milk and soymilk.

Representative examples of recent developed ELISA methods for the detection of mycotoxins are reported in Table 3.

*Toxins* **2021**, *13*, 13

