Dhb Microcystins Discovered in USA Using an Online Concentration LC–MS/MS Platform

Based on current structural and statistical calculations, thousands of microcystins (MCs) can exist; yet, to date, only 246 MCs have been identified and only 12 commercial MC standards are available. Standard mass spectrometry workflows for known and unknown MCs need to be developed and validated for basic and applied harmful algal bloom research to advance. Our investigation focuses on samples taken in the spring of 2018 from an impoundment fed by Oser and Bischoff Reservoirs, Indiana, United States of America (USA). The dominant cyanobacterium found during sampling was Planktothrix agardhii. The goal of our study was to identify and quantify the MCs in the impoundment sample using chemical derivatization and mass spectrometry. Modifying these techniques to use online concentration liquid chromatography tandem mass spectrometry (LC–MS/MS), two untargeted MCs have been identified, [d-Asp3, Dhb7]-MC-LR and tentative [Dhb7]-MC-YR. [Dhb7]-MC-YR is not yet reported in the literature to date, and this was the first reported incidence of Dhb MCs in the United States. Furthermore, it was discovered that the commercially available [d-Asp3]-MC-RR standard was [d-Asp3, Dhb7]-MC-RR. This study highlights a workflow utilizing online concentration LC–MS/MS, high-resolution MS (HRMS), and chemical derivatization to identify isobaric MCs.


Introduction
The World Health Organization (WHO) provides microcystin (MC) drinking and recreational health advisory guidelines. In the last five years, the United States Environmental Protection Agency (US EPA) posted more conservative health advisory guidelines referring to the total class of MCs in drinking and recreational waters. The US EPA set a two-tier drinking water health advisory at 0.3 µg MCs/L for susceptible populations (such as young children) and 1.6 µg MCs/L for the remaining population. The recreational water health advisory is 8 µg MCs/L. Since MCs are hepatotoxins, recent studies suggest that people with non-alcoholic fatty liver disease (NAFLD) would be a susceptible population [1]. With increased incidence of NAFLD due to the rise in obesity and diabetes, the susceptible population to MC toxicity expands across all age groups. In order to determine the concentrations of all MC congeners or prevalent MCs, a robust, sensitive, untargeted analytical method needs to be developed.
MCs are cyclic heptapeptides (Figure 1) produced by over 10 genera of cyanobacteria. The most reported genera in harmful algal bloom (HAB) literature are Microcystis, Dolichospermum, Planktothrix, and Nostoc. In the advent of more sensitive and higher-resolution mass spectrometric techniques, the number of new MCs increased rapidly and is now over 246 MC congeners [2][3][4]. Figure 1 shows that positions 2 (X) and 4 (Z) can vary and are reflected in the MC congener name. MC-LA has leucine and alanine in positions 2 (X) and 4 (Z), respectively. MC has four conserved amino acids: d-methyl aspartic acid (d-MAsp); a unique cyanobacteria amino acid, (2S,3S,4E,6E,8S,9S)-3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid (Adda); d-glutamic acid (d-Glu); and methydehydroalanine (Mdha) (Figure 1). MC congener methylation/demethylation sites occur on these four conserved amino acids, and these substitutions are designated with R as shown in Figure 1. Furthermore, a seventh position substitution was reported with the presence of amino acids serine and methionine. The diversity of MC congeners gives rise to a broad range of chemical and physical properties, suggesting congener-dependent toxicokinetics and toxicodynamics. MCs are cyclic heptapeptides ( Figure 1) produced by over 10 genera of cyanobacteria. The most reported genera in harmful algal bloom (HAB) literature are Microcystis, Dolichospermum, Planktothrix, and Nostoc. In the advent of more sensitive and higher-resolution mass spectrometric techniques, the number of new MCs increased rapidly and is now over 246 MC congeners [2][3][4]. Figure 1 shows that positions 2 (X) and 4 (Z) can vary and are reflected in the MC congener name. MC-LA has leucine and alanine in positions 2 (X) and 4 (Z), respectively. MC has four conserved amino acids: D-methyl aspartic acid (D-MAsp); a unique cyanobacteria amino acid, (2S,3S,4E,6E,8S,9S)-3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid (Adda); D-glutamic acid (D-Glu); and methydehydroalanine (Mdha) (Figure 1). MC congener methylation/demethylation sites occur on these four conserved amino acids, and these substitutions are designated with R as shown in Figure  1. Furthermore, a seventh position substitution was reported with the presence of amino acids serine and methionine. The diversity of MC congeners gives rise to a broad range of chemical and physical properties, suggesting congener-dependent toxicokinetics and toxicodynamics. The two most common methodologies to quantitate MCs are an Adda hapten-based enzymelinked immunosorbent assay (Adda-ELISA) and liquid chromatography tandem mass spectrometry (LC-MS/MS) with triple quadrupole being most commonly used. The Adda-ELISA is an indirect competitive ELISA which is based on the recognition of the Adda group in position 5, and it is now a staple platform to determine the "total MC" concentration by field scientists, ecologists, and environmental scientists because the assay can be performed onsite with minimal overhead. The Adda-ELISA has cross-reactivity with the MC reference standards at 40-150% [5,6], with MC biodegradation products [7], and with MC disinfection byproducts [8][9][10]. While these crossreactivities makes the Adda-ELISA an excellent MC screening tool, its low specificity interferes with quantitation of intact MC among byproducts, making it unsuitable as an analytical tool. LC-MS/MS can only identify and quantify MCs the method has reference standards for. The   The two most common methodologies to quantitate MCs are an Adda hapten-based enzyme-linked immunosorbent assay (Adda-ELISA) and liquid chromatography tandem mass spectrometry (LC-MS/MS) with triple quadrupole being most commonly used. The Adda-ELISA is an indirect competitive ELISA which is based on the recognition of the Adda group in position 5, and it is now a staple platform to determine the "total MC" concentration by field scientists, ecologists, and environmental scientists because the assay can be performed onsite with minimal overhead. The Adda-ELISA has cross-reactivity with the MC reference standards at 40-150% [5,6], with MC biodegradation products [7], and with MC disinfection byproducts [8][9][10]. While these cross-reactivities makes the Adda-ELISA an excellent MC screening tool, its low specificity interferes with quantitation of intact MC among byproducts, making it unsuitable as an analytical tool. LC-MS/MS can only identify and quantify MCs the method has reference standards for. The (Table 3). The weakness in using LC-MS/MS to determine total MC concentration is four-fold: less than 3% of known MCs have commercially available reference standards, new MCs are being discovered more frequently, most commercial reference standards are not certified, and the available standards are not verified as the most prevalent MCs in specific regions or worldwide. Therefore, both methods can commonly over-or underestimate the diversity and concentration of MCs present.
Several methodologies using high-resolution MS (HRMS) and chemical derivatization were developed to identify "untargeted" MCs. HRMS of the precursor and product ions provides chemical formulas derived with a known certainty. LC-HRMS, with accurate mass and automated Data analyses, was used to determine unknown and new MCs [11,12]. Others targeted specific conserved MC MS/MS fragments [13][14][15][16][17], specifically, fragment 135 m/z, [Ph-CH 2 -CH(OCH 3 )], and the MH + minus the neutral 135 m/z fragment that were observed by fast atom bombardment MS using collision-induced dissociation (FABMS/CID) [13] and electrospray ionization/CID (ESI/CID) [18] mass spectrometry. Another approach was chemical derivatization of the position 7 amino acid, Mdha/Dha [14,19]. Mdha and Dha have an α,β unsaturated carbonyl that readily undergoes a Michael's addition in the presences of a good nucleophile such as thiol. In recent literature, several publications used chemical derivatization to discover new MCs [3,20], differentiate isomeric MCs [20], and identify MCs in complex matrices [14]. By using new technologies and understanding the strengths and weakness of each methodology, new platforms and workflows can be created to simplify untargeted MC methods.
Freshwater HABs have an increasingly negative impact on human and environmental health. Primary exposure for humans is through ingestion of drinking and recreational waters. To Date, little resources were allotted for untargeted MC platform development and certified standards of the most prevalent MCs. Since toxicokinetics and toxicodynamics are MC congener-dependent, more studies are needed to determine which MCs are most prevalent and their geographic distribution. In order to execute these studies, robust, sensitive untargeted platforms are needed. The goal of this study was to determine the identity of the untargeted MCs in an impoundment sample from Indiana, United States of America (USA) using a new platform and workflow for untargeted MC analysis.

LC-MS/MS and Adda-ELISA Quantitation
A sample from an impoundment fed by Oser and Bischoff Reservoirs in Indiana was analyzed for total MC concentration using online concentration LC-MS/MS and an Adda-ELISA kit. Online concentration LC-MS/MS is a technique which uses a dual-column system. The first column, the "concentrator" column, concentrates a large volume of the sample. The "concentrator" column is backflushed onto a second column, the "analytical" column, for separation and analysis on an MS system. Previous studies showed a discrepancy in the total MC detected using these two methods [5]. In this study, the LC-MS/MS method used two criteria, chromatographic identification and quantifier to qualifier ion ratios, to verify the identification and quantification of MCs. Chromatographic identification for retention time criteria followed the US EPA method 544, LC-MS/MS method analyses of MCs and nodularin [21]. The peak retention time criterion was calculated by ± 3 standard deviations of the retention time and was calculated to be ± 0.05 min. For the qualifier/quantification transition ratio, a ± 25% criterion was used. Table 1 has the qualifier/quantification transition ratios and retention times for the MCs from the standard and the sample. The retention times observed for both [d-Asp 3 ]-MC-RR and MC-YR in the impoundment sample were outside the peak retention time criteria, at −0.06 min and +0.06 min, respectively. This was not the case for [d-Asp 3 ]-MC-LR with a difference of only 0.01 min (Table 1). Furthermore, the MC-YR and [d-Asp 3 ]-MC-LR in the sample were not within the qualifier/quantification transition ratio range, at 51-85% and 47-79%, respectively. The reported variation in retention times between the calibrants and sample for [d-Asp 3 ]-MC-RR and MC-YR was greater than the peak and the qualifier/quantification transition ratio for MC-YR and [d-Asp 3 ]-MC-LR, as well as outside the ± 25% range. Using a more stringent retention time and qualifier/quantifier ratio, it was concluded that the three major MCs in the sample were not identified and quantified correctly.

Using HRMS to Verify MC Isomers
The impoundment sample was fractionated using high-performance liquid chromatography with a photodiode array detector (HPLC-PDA) into fractions containing the three major MCs and concentrated by N 2 evaporation. Accurate mass by HRMS was used to determine the chemical formula of the three prevalent MCs in the sample and corresponding standards ( Table 2). These results suggested that MCs in the sample had the same chemical formula as the standards of [d-Asp 3 ]-MC-RR, [d-Asp 3 ]-MC-LR, and MC-YR within <2 ppm. As the samples contained the same chemical formula as the standards, but with slight shifts in retention time and changing qualifier/quantification transition ratio, this suggested that the samples MCs were isomers of the standards.

Chemical Derivatization and HRMS Fragmentation to Identify Untargeted MCs
Michael's addition of 2-mercaptoethanol is a commonly used method to determine the identity of Mdha or Dha at position 7 [14,19]. In short, thiol addition indicates Mdha or Dha, while no reaction indicates either Dhb or no α,β unsaturated carbonyl. Dha and Mdhb variants would result in parent masses of one fewer or more methyl group, respectively. The impoundment sample was divided into control and experiment samples and recombined to perform online concentration LC-MS/MS in full scan mode. MC masses, as well as the addition of the 2-mercaptoethanol mass (+78), were targeted for identification. Three targeted MCs and three untargeted MCs were identified ( HRMS with product ion scan was used to determine the structure of the isomers. Fragments identified had <7 ppm agreement with the m/z ion molecular formula used to derive the fragment chemical structure (

Discussion
MCs containing Dhb 7 were observed from the cyanobacteria Planktothirix and Nostoc spp. [22][23][24][25]. The sample in this study was from an impoundment fed from Oser and Bischoff reservoirs in 2018 where the two dominant forms of cyanobacteria present were Planktothrix agardhii and Pseudanabaena limnetica, which are known producers of MCs [26]. Analysis of the sample determined the presence of  [28,29], the [Dhb 7 ] MC congener has not been reported in the US. These studies found that the occurrence of Planktothirx and Oscillatoria spp. was 38% [28], and they were the dominate cyanobacteria contributing to >1 ppb MC in three eco-regions (northern, temperate, and southern plains) [28,29], in which the current sample site was located. Other studies using cultures of Planktothrix sp. showed more diversity in MC congeners to include [d-Asp 3 ] and [d-Asp 3 , Dhb 7 ], containing the interchangeable fourth position MCs to contain amino acids other than arginine [20,30].
The  Table 1). The impoundment sample qualifier/quantification transition ratios were calculated to be 28% and 26% for MC-YR and [d-Asp 3 ]-MC-LR, respectively, which was outside the ±25% window set within the method. The impoundment sample MC-YR and [d-Asp 3 ]-MC-LR peaks visually showed lower peak height for the qualifier ion (213 m/z), suggesting that they have the same chemical formulas but different atom connectivity. The slight shift in chromatography and the monitored product ions suggest that (

1) isomers of MC-YR, [d-Asp 3 ]-MC-LR, and [d-Asp 3 ]-MC-RR exist, and (2) for the 213 m/z product ion present for MC-YR and [d-Asp 3 ]-MC-LR, the isomeric center is likely to be in position 7 as Dhb not
Mdha. The change in the qualifier/quantifier (213 m/z/135 m/z) ratio observation could potentially be a mechanism to determine the presence of Dhb. Several of these trends were also observed by others who determined isomeric MCs [3,24,27]. Therefore, the three major MCs were suspect Mdha/Dhb isomers and were further investigated using an online concentration LC-MS/MS platform for chemical derivatization LC-MS/MS. Results were verified by precursor and product ion HRMS.
In order to advance freshwater HAB research, simple, robust, sensitive MS platforms need to be developed. Untargeted MC analytical platforms center around HRMS because of its ability to determine the chemical formulas of precursor and product ions with certainty [14,31]. Furthermore, chemical derivation with HRMS was used to determine if Mdha/Dha was present in the seventh position. The Dha chemical derivatization method was moved to a new platform, online concentration LC-MS/MS. The online concentrator dilutes 1 mL of sample with 0.1% formic acid water as it is injected onto the concentrator column while being cleaned with 0.1% formic acid in water during the 1-min injection period. This procedure was shown to remove unreacted reagents and buffers from the sample with little to no interferences observed [32]. Since the MCs in the sample were suspected to be Dhb isomers of the Mdha standards, the impoundment and standard samples were modified by chemical derivatization using 2-mercaptoethanol. This reagent reacts via nucleophilic addition with Mdha/Dha versus a minimal reaction with Dhb [14,19]. A 1.5-mL 50:50 mixture of derivatized and unreacted sample was analyzed by online concentration LC-MS full scan. Figure 2 (Figure 4). Each MC fragmentation piece was identified within 7 ppm for each MC, which modeled the pieces of the molecule as a whole (Figure 3 and Supplementary Materials). Since the biosynthetic pathway is well described, the Dhb MC stereochemistries were presumed to be the same as known MCs. With the information provided by these complementary techniques, a new MC, [Dhb 7 ]-MC-YR, was discovered, and the standard [d-Asp 3 ]-MC-RR used for analysis was in fact [d-Asp 3 , Dhb 7 ]-MC-RR. At present, Microcystis is not reported to produce Dhb-containing MCs; however, the manufacturers state that the MC standard assessed during the present study was produced using a Microcystis culture [3,14,19]. This raises the question of whether Microcystis sp. may be able to produce [d-Asp 3 , Dhb 7 ]-MC-RR (and/or other Dhb-containing MCs), and highlights the importance of undertaking characterization work that discriminates between Mdha-and Dhb-containing MCs.

Conclusions
Toxicokinetics and toxicodynamics were shown to be MC congener-dependent; thus, reliable platforms and workflows are needed to determine which  (2) for certified reference standards to be readily available. . Standards were prepared in methanol and diluted in LC-MS grade water. Samples were stored on ice after collection and until arrival to the laboratory where they were then frozen at −4 • C. Samples then went through a series of three freeze-thaw cycles before analysis [31,39,40].

Sample Collection
On 1 May 2018, a grab sample was taken from an impoundment fed by Oser and Bischoff Reservoirs, Indiana, USA. The sample was collected in a 500-mL bottle with no preservatives. The bottle was immediately placed on ice and stored at 4 • C until shipping. Samples were shipped overnight on ice to Lumigen Instrument Center, Wayne State University, Detroit MI within 24 h of collection. The samples underwent three freeze-thaw cycles prior to analyses or concentration. On the Day of analysis, the thawed sample was centrifuged at 10,000 rpm for 10 min at 4 • C, and the supernatant was collected and analyzed.

ELISA
Samples were tested using the Abraxis Adda-ELISA following procedures provided by the kit and as described in Birbeck et al. (2019) [32]. The ELISA kit was stored at 4 • C until used. The plate was read with a SpectraMax M2e microplate reader (Molecular Devices, LLC. San Jose, CA, USA) at 450 nm.

Online Concentration Liquid Chromatography Mass Spectrometry
Samples were analyzed by an online concentration method set up in house using a Thermo Scientific TSQ Altis™ triple quadrupole mass spectrometer (Thermo Scientific, Waltham, MA, USA) with an EQuan MAX Plus™ system. The system was set up as described in Birbeck et al. (2019) [32]. Briefly, 1 mL of sample was injected onto a loading column (Thermo Scientific Hypersil GOLD aQ 2.1 × 20 mm, 12 µm particle size) using an HTC PAL autosampler (CTC Analytics, Zwingen, Switzerland) equipped with a 5-mL sampling syringe. The flow rate was 1.5 mL/min using 0.1% formic acid in water. Analytes were separated on a Thermo Accucore aQ, 50 × 2.1 mm, 2.6 µm particle size column at 35 • C. The analytical separation gradient consisted of 0.1% formic acid in water (mobile phase A), and 0.1% formic acid in acetonitrile (mobile phase B), at a flow rate of 0.5 mL/min. The gradient was set as follows: isocratic from 0-0.7 min at 24% B, then ramped to 26% B from 0.7-1.7 min, and then increased to 50% B from 1.7-5.5 min. The column was washed with 98% B from 5.51-6.5 min at a flow rate 1.0 mL/min, then equilibrated at 24% B from 6.5-8.5 min.
Mass spectrometry analysis was performed using an electrospray ionization source (ESI). The MS source settings were as follows: spray voltage set to 3500 V in positive ion mode, ion transfer tube temperature set to 325 • C, vaporizer temperature at 420 • C, sheath gas pressure at 50 units, auxiliary gas pressure at 10 units, sweep gas at one unit, collision gas at 1.5 mTorr, Q1 resolution set to 0.7 Da, and Q3 resolution set to 1.2 Da. A list of the SRM transitions for each standard can be found in Table 3. Quantitation Data results were accomplished using TraceFinder™ EFS 4.1. Retention time, and quantitative and qualitative ion qualifications were set in TraceFinder™ and monitored to ensure that the proper MC was detected. Furthermore, a second method for full scan Q3 analysis was used. It was identical to the quantitative run for the LC gradient and MS source settings. The scan range was 300-1200 (m/z), the scan rate was set to 1000 (Da/s), and the Q3 resolution was 1.2 full width at half maximum (FWHM).

High-Resolution Mass Spectrometry
The MC isolates were analyzed using nano-LC-HRMS with a Thermo Easy-nano-LC 1200 Orbitrap LTQ-XL (Thermo Scientific, Waltham, MA, USA) by gradient elution through a Thermo peptide map C18 (100 Å, 50 µm × 15 cm column, 2 µm particle size). Mobile phase A was 0.1% formic acid water and mobile phase B was 0.1% formic acid in 80% acetonitrile in water. At time 0, mobile phase B was at 5% and it was increased to 25% at t = 30 min, followed by an increase to 35% at t = 40 min, and an increase to 95% at t = 50 min, where it was held for 10 min. A high-resolution scan was run using the Orbitrap followed by Data-dependent analysis of the [M + 2H] 2+ exact mass for the expected MCs. Target m/z values were fragmented by HCD set to 15 and CID at 35 with an isolation width of 2 m/z. Internal mass locks to background polysiloxanes and phthalates were used to improve mass accuracy. In fragmentation spectra, mass axis correction was determined by shifting to the assumption that the 135.08 peak was the common Adda fragment of 135.0804. The correction factor was applied to all reported masses to improve mass accuracy.

Solid-Phase Extraction
Solid-phase extraction (SPE) was performed on the samples using a Waters 6cc Oasis HLB cartridge (Waters Corporation, Milford, MA, USA). The column was conditioned with 7 mL of LC-MS grade methanol, then equilibrated with 7 mL of LC-MS grade water. Next, 100 mL of sample was brought through the column at <1 mL/min. Then, the column was rinsed with 7 mL of 5% methanol. A second rinse of 7 mL of 100% methanol was completed, and this rinse was collected. The rinses were evaporated under N 2 gas. Samples were stored at −4 • C until ready for analysis.

Fractionation of SPE Concentrated Samples by HPLC-PDA
The concentrated samples were further separated and fractionated using a Waters Alliance 2695 HPLC with a Waters 996 photodiode array detector (Waters Corporation, Milford, MA). The samples were separated using gradient analysis on a Waters Atlantis dC18 OBD prep column (10 × 100 mm, 5 µm particle size). The mobile phases used for the gradient were 20 mM ammonium acetate in water (mobile phase A) and 100% methanol (mobile phase B), run as follows: 60% B from 0-1 min, then ramped to 75% B from 1-10 min, followed by column cleaning at 90% B from 16-18 min, and equilibration at 60% B from 18-21 min. The flow rate was set to 3 mL/min, and the column oven temperature was 35 • C. The fractionated samples were collected every minute and were evaporated using N 2 gas and stored at −4 • C until ready for analysis.

Chemical Derivatization of MCs
To test unknown MC structures, known standards and samples containing MCs were chemically derived with 2-mercaptoethanol. This procedure was used by others to determine unknown MCs, but was modified here to determine structures of known isomers of MCs [14,20,41]. The reaction proceeded as follows: 1 mL of the sample or standard (in water) was added to a vial, and 100 µL of 0.2 M sodium carbonate was added to the vial and vortexed to bring the pH to 9. Next, 5 µL of 2-mercaptoethanol was added to the sample or standard solution and was vortexed and left to sit capped in a hood. The reaction was stopped after 2 h by adding 10 µL of 1% acetic acid to the vial to bring the pH to 6. Control samples and standards were also prepared as above, but the addition of 2-mercaptoethanol was excluded.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6651/11/11/653/ s1: Figure S1: HRMS of the three main MCs found in the impoundment sample and the corresponding isomeric standard.