Determination of Microcystins in Fish Tissue by ELISA and MALDI-TOF MS Using a Highly Specific Single Domain Antibody

Abstract

The development of simple, reliable, and cost-effective methods is critically important to study the spatial and temporal variation of microcystins (MCs) in the food chain. Nanobodies (Nbs), antigen binding fragments from camelid antibodies, present valuable features for analytical applications. Their small antigen binding site offers a focused recognition of small analytes, reducing spurious cross-reactivity and matrix effects. A high affinity and broad cross-reactivity anti-MCs-Nb, from a llama antibody library, was validated in enzyme linked immunosorbent assay (ELISA), and bound to magnetic particles with an internal standard for pre-concentration in quantitative-matrix-assisted laser desorption ionization-time of flight mass spectrometry (Nb-QMALDI MS). Both methods are easy and fast; ELISA provides a global result, while Nb-QMALDI MS allows for the quantification of individual congeners and showed excellent performance in the fish muscle extracts. The ELISA assay range was 1.8-29 ng/g and for Nb-QMALDI, it was 0.29-29 ng/g fish ww. Fifty-five fish from a MC-containing dam were analyzed by both methods. The correlation ELISA/sum of the MC congeners by Nb-QMALDI-MS was very high (r Spearman = 0.9645, p < 0.0001). Using ROC curves, ELISA cut-off limits were defined to accurately predict the sum of MCs by Nb-QMALDI-MS (100% sensitivity; ≥89% specificity). Both methods were shown to be simple and efficient for screening MCs in fish muscle to prioritize samples for confirmatory methods.

Keywords: ELISA; MALDI-TOF; fish; microcystins; nanobody.

Figure 1

Calibration curve of the Nb-ELISA (n = 4, r² = 0.999). % A/A₀ = % of absorbance of the zero concentration. The concentration range was 0.01–10 µg/L MC-LR and the calculated IC₅₀ (half maximal inhibitory concentration) = 0.71 µg/L (95% confidence interval = 0.63–0.81 µg/L).

Figure 2

MALDI-TOF spectra of water samples from Palmar Dam in both campaigns. Peaks with m/z corresponding MCs are labeled. (A) February 2017, m/z = 995.6 (MC-LR); (B) March 2021 m/z = 995.6, 1023.5, and 1037.6 (which correspond to the single protonated ion of MC-LR; [D-Leu¹, Dha⁷] MC-LR; and [D-Leu¹] MC-LR).

Figure 3

Boxplots for the MC analysis in fish, classified by method and campaign (n = 55). (A) All data classified by method (Nb-ELISA and Nb-QMALDI); (B) Nb-ELISA and sum of congeners by Nb-QMALDI MS in ng/g fish wet weight, classified by campaign.

Figure 4

Nb-QMALDI MS of the fish samples. Typical spectra of fish samples from Palmar in both campaigns analyzed. The m/z = 1123.6 corresponds to the internal standard used for quantification. (A) Unexposed fish (LMB); (B) February 2017 sample with MCs: m/z = 995.5 and 1116.6 (which correspond to the single protonated ion of MC-LR and MC-LR-Cys); (C) March 2021 sample with MCs: m/z = 995.6, 1023.6, 1037.6, and 1068.5 (which correspond to MC-LR, [D-Leu¹, Dha⁷] MC-LR, [D-Leu¹] MC-LR, and MC-WR, respectively).

Figure 5

Correlation between the methods of MC analysis in fish: ELISA and the sum of congeners by Nb-QMALDI MS for each campaign (2017 n = 26 and 2021 n = 29). Ln (concentration ng/g fish) is the variables x and y for normalization.

Figure 6

ROC curves for the Nb-ELISA as a test for the prediction of three threshold-values of the sum of all variants by Nb-QMALDI MS. AUC = Area under the curve.