4.3.2. Total Adda MCs/NODs Oxidation, Extraction and Analysis (MMPB)

MMPB oxidations and extractions were conducted as previously described [40]. Briefly, samples were oxidized as 100 mg (liver & kidney; wet weight), 50 mg (hair; dry weight), 500–1500 μL (blood), 0.2 μL–500 μL (bile, urine) and 10 mg (vomit; wet weight) subsets. All samples were paired with at least one pre-oxidation matrix spike of certified reference material (CRM) MC-LR (National Research Council Canada; Halifax, NS, Canada) for quantification. Spikes ranged from 5 ng mL−<sup>1</sup> (or ng g<sup>−</sup>1) to 50,000 ng mL−<sup>1</sup> and were added to approximately double native peak areas (positive samples) for quantification (Table S2). Subsets of liver, hair, kidney, vomit and blood were oxidized by adding 5 mL of oxidant (0.2 M K2CO3, 0.1 M KMnO4 and 0.1 M NaIO4), while urine and bile samples were oxidized with 2.5 mL oxidant. Reactions were stopped after 2 h with the addition of 40% sodium bisulfite (w/v) until solutions turned white/opaque. The liver, kidney, hair, and blood oxidized aliquots were cooled (10 min; −20 ◦C), centrifuged (1500 *g*; 10 min) and supernatants retained. The pellets were rinsed with DI (2 mL) by vortex mixing followed by centrifugation (1500× *g*; 10 min). Extracts were loaded onto preconditioned (MeOH→DI) 200 mg Strata X SPE (Phenomenex, Torrance, CA, USA), rinsed (DI; 3x) and eluted (5 mL; 90% acetonitrile). Extracts were blown to dryness (N2, 40 ◦C) and samples (with exception to liver and blood) reconstituted in DI (0.5–1.0 mL). The liver and blood were reconstituted in 0.1 M HCl (1.5 mL) and loaded onto simplified liquid extraction (SLE) columns (12cc; Phenomenex, Torrance, CA, USA), allowed to sit (10 min) and eluted with 2x–5mL ethyl acetate. Elutions from SLE were blown to dryness (N2, 40 ◦C) and reconstituted in DI (0.5–1.0 mL). All samples were filtered using 0.2 μm polyvinylidene difluoride (PVDF) syringe filters (Millipore Sigma, St. Louis MO, USA).

A Thermo Surveyor high performance liquid chromatography (HPLC) system coupled to a TSQ Quantum Access MAX Triple quadrupole mass spectrometer system coupled with a Kinetex C18 column (2.6 μm; 100 Å; 150 × 2.1 mm; Phenomenex; Torrance, CA, USA) were utilized as described previously [36]. The [M−H]<sup>−</sup> ion of MMPB (*m*/*z* 207) was fragmented and *m*/*z* 131 (CE = 12%) was monitored. A five-point standard curve (0.5–100 ng/mL of oxidized MC-LR) was used to interpolate spike returns. Quantification of samples was conducted using individually prepared pre-oxidation spikes of MC-LR (standard addition). The instrument detection limit coupled with spike responses were used to determine the method detection limits (MDLs; Table 4).



4.3.3. Free Adda MCs/NODs Extraction and Analyses

Samples with total MCs >15 ppb (ng g−1; ng mL−1) via MMPB analysis and representative negative control samples (unexposed individuals) were extracted and analyzed for free MCs, unless exhausted. Subsets of liver (100 mg), kidney (100 mg), blood/serum (500 μL), and hair (50 mg) were suspended in 5 mL extractant (75% acidified acetonitrile in 100 mM acetic acid) and sonicated (bath, 25 min). The samples were cooled (10 min; −20 ◦C), centrifuged (1500× *g*; 10 min) and supernatants retained. The pellets were resuspended in extractant (1 mL), cooled (10 min; −20 ◦C) and re-centrifuged. The pooled supernatants (including previously homogenized vomitus) were diluted (70 mL DI). Samples of urine (ranging from 10 μL–500 μL) and bile (10 μL) were diluted (10 mL DI) prior to SPE. Preconditioned Oasis HLB (200 mg; Waters Corporation, Milford, MA) or Strata X were loaded with diluted sample, rinsed with DI (2×; 5 mL and 3 mL) and eluted (5 mL; 90% acetonitrile). Solutions were blown to dryness (N2; 60 ◦C) and reconstituted in DI at sample concentrations within range of the calibration curves for each analysis used. All samples were filtered using 0.2 μm PVDF prior to analysis.

Final liver and kidney extracts were washed 1:1 (*v*/*v*) using hexane (ACS grade; Fisher Scientific, Waltham, MA, USA). Hexane was added to sample, vortex mixed, centrifuged (1500× *g*; 5 min) and the hexane layer discarded. Analysis on these extracts by ELISA was conducted both prior to and after hexane washing.

An MCs/NODs Adda ELISA (Abraxis; Warminster, PA) was used as previously described [40]. Dilutions were prepared using DI to achieve absorbance values within the range of the standard curve (0.15–4.0 ng mL−1) and all solutions were analyzed in duplicate. The minimum method detection limit (MDL) was 15 ppb (ng g−1/ng mL<sup>−</sup>1) based on the 100-fold dilution factor (DF) used and assay sensitivity (0.15 ng mL<sup>−</sup>1).

Extracts prepared for ELISA were post-spiked with the internal standard (IS) *d*7-MC-LR. The vomitus sample extracts were spiked with a second IS (*d5*-MC-LF). Both IS were acquired from Abraxis Kits (Warminster, PA, USA). A targeted analysis for MC-LR was used on all extracts using a Thermo Surveyor HPLC system coupled with an LTQ XL Linear Ion Trap Mass Spectrometer. Separation was achieved using the same column used in MMPB analysis (Kinetex C18 column) with mobile phase C (2 mM formic acid and 3.6 mM ammonium formate in deionized water) and D (95% acetonitrile (*v*/*v*) in 2 mM formic acid and 3.6 mM ammonium formate). Acetonitrile (Optima LC/MS), water (HPLC), ammonium formate (ACS grade), and formic acid (98%) were from Thermo Fisher Scientific (Waltham, MA, USA). The gradient (0.2 mL min-1) was as follows: solvent C 70–30% over 5 min, 30–70% C over 2 min, and held at 70% C for 3 min. Each chromatographic run was 10 min and 20 μL full loop injections were employed. The analysis of MC-LR (*m*/*z* 995.5→375.0, 553.4, 599.4, 866.6) was calibrated using a seven-point IS curve (0.5–100 ng mL<sup>−</sup>1) with *<sup>d</sup>*7-MC-LR (*m*/*<sup>z</sup>* 1002.5→599.5).

*Toxins* **2019**, *11*, 456

The analysis of the vomit extracts was conducted using targeted LC-MS/MS (Table S1) and LC-MS scans (*m*/*z* 400–1800 (-ve and +ve)) with a PDA set to λ 238 nm. The gradient (0.2 mL min-1) was as follows: solvent C held at 70% for 11 min, 70–30% C over 9 min, 30–70% C over 2 min, and held at 70% C for 5 min. XCalibur v 2.2 (Thermo Fisher Scientific, Waltham, MA, USA) was utilized for data processing. Standards used to calibrate the method included CRMs (NOD-R, MC-LR, MC-RR, [Dha7]MC-LR) and the RM (MC-RY) from the National Research Council Canada (Halifax, NS, Canada), RMs (MC-WR, [Asp3]MC-RR, [Asp3]MC-LR, MC-HtyR, MC-LF, MC-LW, MC-HilR) from Enzo Biochem (Farmingdale, NY, USA) and RMs ([Leu1]MC-LR, MC-YR, MC-LA, MC-LY) from GreenWater Laboratories (Palatka, FL, USA). Additional RMs ([ADMAdda5]MC-LR, [ADMAdda5]MC-LHar) were produced in-house through the extraction and purification of a strain of *Nostoc* sp. 152 generously supplied by Kaarina Sivonen (University of Helsinki) using methods previously described [38]. The desmethyl RMs [DMAdda5]MC-LR and [DMAdda5]MC-LHar were prepared by hydrolysis of the [ADMAdda5]MCs [57]. Quantification of targeted MCs was conducted using the previously reported IS approach [36].

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6651/11/8/456/s1, Figure S1: A general representation of the heptapeptide microcystin and the formation of the MMPB molecule following oxidative cleavage of the Adda side chain; Figure S2: Significant gross lesions observed during autopsy of deceased dog (C-SP); Table S1: Multiple reaction monitoring (MRM) transitions for targeted analysis of MCs (19 variants), NOD-R; Table S2: Matrix spikes (of MC-LR) associated with returns reported in Table 3.

**Author Contributions:** Conceptualization, A.F. and S.B.; methodology, A.F. and S.B.; validation, A.F. and M.A.; formal analysis, A.F.; investigation, B.G., A.M.; resources, A.F., M.A., S.F., B.G., A.M.; writing—original draft preparation, A.F.; writing—review and editing, A.F., S.B., N.M.

**Funding:** The authors declare that GreenWater Laboratories received funding for MMPB analysis of one liver sample and ELISA analysis of the vomitus sample from Pet Emergency of Martin County (Stuart FL). All other research received no external funding.

**Acknowledgments:** The authors thank Kamil Cieslik for his assistance in laboratory (MC) testing and the pet owners for their assistance throughout this study.

**Conflicts of Interest:** The authors declare no conflict of interest
