Microcystins (MCs) are secondary metabolites generated by cyanobacteria, which can be present in drinking water sources during harmful algal blooms (HABs). MCs are potent liver toxins that inhibit the function of protein phosphatases 1 and 2A (PP1 and PP2A) by binding to the enzymes’ active sites. Large MC doses lead to acute liver failure, but prolonged exposure to low levels of MCs may be more prevalent and pernicious. The effects of such exposure in humans are not well understood and are generally extrapolated from animal models. Further complicating their study, over 275 MC congeners have been discovered, though many remain unidentified. This dissertation advances MC research by describing novel extraction and data analysis methods coupled with liquid chromatography-mass spectrometry (LC-MS) for quantitative and qualitative analyses of MCs.
A method was developed to extract and quantify MCs from mouse liver with limits of quantification (LOQs) lower than previously reported. MCs were extracted from 40-mg liver samples using 85:15 (v:v) CH3CN:H2O containing 200 mM ZnSO4 and 1% formic acid. Solid-phase extraction with a C18 cartridge was used for sample cleanup. MCs were detected and quantified using LC-orbitrap-MS with simultaneous MS/MS detection of the 135.08 m/z fragment from the conserved Adda amino acid for structural confirmation. The method was used to extract six MCs (MC-LR, MC-RR, MC-YR, MC-LA, MC-LF, and MC-LW) and the MC-LR cysteine adduct (MC-LR-Cys), which can be created from MC-LR in vivo by the glutathione detoxification pathway, from spiked liver tissue. Matrix-matched internal standard calibration curves were constructed for each MC (R2 ≥ 0.993), with LOQs between 0.25 ng per g of liver tissue (ng/g) and 0.75 ng/g for MC-LR, MC-RR, MC-YR, MC-LA, and MC-LR-Cys, and 2.5 ng/g for MC-LF and MC-LW. The protocol was applied to extract and quantify MC-LR and MC-LR-Cys from the liver of mice that had been gavaged with 50 μg or 100 μg of MC-LR per kg bodyweight and were euthanized 2 h, 4 h, or 48 h after final gavage. C57Bl/6J (wild type, control) and Leprdb/J (sample) mice were used as a model to study non-alcoholic fatty liver disease (NAFLD). The Leprdb/J mice were relatively inefficient in metabolizing MC-LR into MC-LR-Cys, which is an important defense mechanism against MC-LR exposure. Trends were also observed as a function of MC-LR gavage amount and time between final MC-LR gavage and euthanasia/organ harvest.
In the second project, putative identification of novel MCs was achieved in water originating from HAB of 2020 using ultra-high-performance (UHP) LC-high resolution (HR) MS and a new bottom-up sequencing strategy. Maumee River water samples were collected during a HAB and analyzed with simultaneous HRMS and MS/MS. Unidentified ions with characteristic MC fragments (135 and 213 m/z) were recognized as possible novel MC congeners. An innovative workflow, involving separation and detection of MC ions and their fragments by ultrahigh pressure (UPLC)-orbitrap-MS, MS/MS and MS/MS/MS, was developed for putative identification of those ions. Two codes were written using the Python programming language. The first predicts molecular composition from exact mass measurements by generating potential structures using 59 amino acids that have been found in known MCs. The second assigns fragment ions after multistage MS (MS/MS and MSn) for structural confirmation. Peptide fragmentation is well-characterized, but existing software for rapid fragment ion identification is not suitable for MCs due to their cyclic structure and unique amino acids. The developed workflow was used for putative identification of two unknown congeners found in Lake Erie: MC-HarR and MC-E(OMe)R. The method was also used to putatively identify eight MCs for which standards are not available, and was able to distinguish between isomeric MCs such as [Asp3] MC-YR and [Dha7] MC-YR.