Advertisement

Analytical and Bioanalytical Chemistry

, Volume 410, Issue 22, pp 5405–5420 | Cite as

Capillary electrophoresis–tandem mass spectrometry for multiclass analysis of polar marine toxins

  • Daniel G. Beach
  • Elliott S. Kerrin
  • Krista Thomas
  • Michael A. Quilliam
  • Pearse McCarron
Paper in Forefront
Part of the following topical collections:
  1. Food Safety Analysis

Abstract

Polar marine toxins are more challenging to analyze by mass spectrometry-based methods than lipophilic marine toxins, which are now routinely measured in shellfish by multiclass reversed-phase liquid chromatography–tandem mass spectrometry (MS/MS) methods. Capillary electrophoresis (CE)–MS/MS is a technique that is well suited for the analysis of polar marine toxins, and has the potential of providing very high resolution separation. Here, we present a CE–MS/MS method developed, with use of a custom-built interface, for the sensitive multiclass analysis of paralytic shellfish toxins, tetrodotoxins, and domoic acid in seafood. A novel, highly acidic background electrolyte (5 M formic acid) was designed to maximize protonation of analytes and to allow a high degree of sample stacking to improve the limits of detection. The method was applied to a wide range of regulated and less common toxin analogues, and exhibited a high degree of selectivity between toxin isomers and matrix interference. The limits of detection in mussel tissue were 0.0052 mg/kg for tetrodotoxins, 0.160 mg/kg for domoic acid, and between 0.0018 and 0.120 mg/kg for paralytic shellfish toxins, all of which showed good linearity. Minimal ionization suppression was observed when the response from neat and mussel-matrix-matched standards was corrected with multiple internal standards. Analysis of shellfish matrix reference materials and spiked samples demonstrated good accuracy and precision. Finally, the method was transferred to a commercial CE–MS/MS system to demonstrate its widespread applicability for use in both R & D and routine regulatory settings. The approach of using a highly acidic background electrolyte is of broad interest, and can be considered generally applicable to simultaneous analysis of other classes of small, polar molecules with differing pKa values.

Graphical abstract

Keywords

Biotoxins Tetrodotoxin Paralytic shellfish poisoning Domoic acid Harmful algal bloom Capillary electrophoresis 

Introduction

Several different classes of natural toxins are known to occur in the marine environment, each produced as secondary metabolites of one or more different species of microorganism. At certain times of the year or under certain environmental conditions, these microorganisms can experience exponential growth, and toxins can accumulate to high enough concentrations to become problematic. Accumulation of toxins in filter-feeding bivalves or other seafood species is a significant public health and economic concern worldwide [1, 2]. Blooms of toxic algae can also create significant environmental problems, leading to multispecies mortality events [3]. In general, increased proliferation of toxin-producing microorganisms in the environment can be linked with global change factors, including warming oceans and nutrient input into the coastal environment [4].

The classes of marine toxins span a broad range of different mechanisms of toxicity and physicochemical properties, but can most broadly be classified as either polar or lipophilic. Polar marine toxins include paralytic shellfish toxins (PSTs), tetrodotoxins (TTXs), and domoic acid (DA). The chemical structures of representative toxins from these three classes are shown in Fig. 1.
Fig. 1

Structures of polar marine toxins analyzed by capillary electrophoresis–mass spectrometry. dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcNEO decarbamoylneosaxitoxin, dcSTX decarbamoylsaxitoxin, NEO neosaxitoxin

DA and the related compound kainic acid (KA) are potent toxins that act as glutamate agonists and were first found in seaweed [5]. DA was first encountered as a shellfish toxin in 1987 during a lethal outbreak of shellfish poisoning in Canada [6, 7]. It was later determined that it was produced by several diatom species in the genus Pseudo-nitzschia [6, 7]. Acute human exposure to DA leads to a condition known as “amnesic shellfish poisoning” [2], but a range of potential effects are also increasingly being associated with long-term, low-level exposure to DA, which may be relevant below the current regulatory limit of 20 mg/kg [8].

PSTs are a class of potent neurotoxins that exist as a mixture of structural analogues of saxitoxin (STX). They occur worldwide, and are produced by dinoflagellates of the genera Alexandrium, Pyrodinium, and Gymnodinium in the marine environment and by some species of cyanobacteria in freshwater [9]. Human exposure to PSTs leads to a condition known as “paralytic shellfish poisoning,” with the current regulatory limit set at 0.8 mg saxitoxin dihydrochloride equivalents per kilogram of shellfish tissue [2, 9]. Chemically, these toxins share a tetrahydropurine skeleton, and can be categorized further on the basis of their charge state in solution resulting from substitutions at four sites as shown in Fig. 1: STXs, with a charge of +2; gonyautoxins (GTXs), with a charge of +1, and neutral N-sulfocarbamoylgonyautoxins (C-toxins).

TTX and several of its structural analogues are found in a wide range of organisms across both the marine and the terrestrial environment, but have most commonly been considered as causative agents in pufferfish poisoning [10]. Unlike PSTs and DA, whose producing organisms have been identified, TTX is still largely of unknown origin. Although several species of macroorganisms are known to contain TTX, it is generally thought that the toxin itself is a bacterial metabolite. A number of reports of sporadic TTX production by several different species of bacteria can be found in the literature, but production has been difficult to replicate under controlled conditions in a laboratory environment [11]. More recently, TTX has been identified in shellfish in Europe, including in the Netherlands and the UK, which have temperate climates [12, 13, 14, 15].

Most marine biotoxins were originally monitored routinely by mouse bioassays, particularly the PSTs [16]. More recently, most regulatory agencies have moved toward analytical chemical methods for toxin detection, including two official methods based on either precolumn or postcolumn oxidation combined with reversed-phase liquid chromatography (RPLC) with fluorescence detection [17, 18]. Robust, routine, multiclass analytical methods based on RPLC coupled with tandem mass spectrometry (MS/MS) are now a reality for several classes of lipophilic shellfish toxins [19, 20]. Many of these multiclass RPLC–MS/MS methods also allow the analysis of DA, but PSTs and TTXs are too polar to be analyzed effectively by RPLC–MS/MS. Separation of PSTs and TTXs by RPLC requires the use of strong ion-pairing agents that are not compatible with or can limit the sensitivity of electrospray ionization (ESI) mass spectrometry (MS) [17, 18]. Alternatively, hydrophilic interaction liquid chromatography (HILIC) has been applied to the separation of PSTs and TTXs coupled with MS detection [21, 22, 23, 24]. For PSTs, either a long separation time [21] or desalting using an activated carbon solid-phase extraction (SPE) cleanup [22] is required, and ionization suppression and method robustness issues exist. For TTXs, the amino acids arginine and hydroxyarginine have been identified as matrix components that can lead to ionization suppression, and therefore require careful method optimization to be separated from TTXs by HILIC [24]. Alternatively, RPLC methods using precolumn or postcolumn derivatization to fluorescent products are widely used for both PSTs and TTXs [17, 18, 25, 26, 27]. These methods are of similar sensitivity to HILIC–MS/MS methods but do not provide the selectivity or structural information of MS/MS and can require multiple chromatographic runs per sample. Compared with TTXs and PSTs, rapid, robust analysis of DA directly from shellfish extracts without the need for sample cleanup is routine by liquid chromatography (LC)–MS or LC with UV detection [28], and even direct MS analysis without sample cleanup or chromatographic separation is feasible [29].

Capillary electrophoresis (CE) is an alternative mode of analytical separation to LC that is ideally suited for the separation of polar, charged analytes. Compared with LC, CE offers higher resolution and generally reduced matrix effects when coupled with ESI-MS. Since it was first introduced in 1987 [30], CE–MS has seen broad application in fields such as metabolomics, proteomics, and DNA sequencing. Its use for the separation of polar marine toxins was appreciated relatively early on in its development [31]. Early reports of CE analysis of marine toxins included the separation of fluorescamine- and dansyl-labeled STX, TTX, and microcystin LR by capillary zone electrophoresis with fluorescence detection [31] and the direct analysis PSTs [32] and DA [33] by CE with UV detection. This work was promising, but generally suffered from high limits of detection (LODs; greater than 5 μM) and significant matrix interference resulting from challenges with UV detection or derivatization of PSTs. Also, the full separation of most PSTs from interference in complex biological matrices has never been achieved.

CE–MS methods for PSTs were developed, showing application to real samples with a method using a 35 mM morpholine buffer (pH 5) and sample stacking by isotachophoresis to achieve a theoretical plate count of more than 200,000 [34, 35, 36]. However, this method requires the use of irreproducible coated capillaries, and the range of PSTs analyzed was limited. Despite these limitations, the method has seen some broader implementation [37, 38]. Similarly, early proof-of-concept work on CE analysis of DA was later fully developed into a quantitative method of analysis using a highly selective combined cation- and anion-exchange double SPE cleanup [39], but the method has not seen wider adoption, probably because of the simplicity and effectiveness of available chromatographic methods for DA.

Our group recently introduced a CE–MS/MS method for analysis of the neurotoxin β-N-methylamino-l-alanine (BMAA) that was successful at separating the analyte from potentially interfering matrix components, including several isomers [40]. Here the utility of CE–MS/MS as a multiclass analytical method for polar marine toxins is demonstrated on the basis of the approach developed for BMAA using a highly acidic background electrolyte (BGE). The method was optimized for qualitative and quantitative analysis of a wide range of PSTs, TTXs, and DA. Ionization suppression and the need for sample cleanup were assessed, and a variety of typical validation parameters, such as selectivity, precision, linearity, and LODs, were measured. Finally, accuracy of the method was evaluated through quantitative analysis of mussel matrix (certified) reference materials and spiked samples.

Experimental

Chemicals and reagents

Acetone and 2-propanol (high-performance LC grade) were from Caledon (Georgetown, ON, Canada). Ammonium acetate was from VWR International (Mississauga, ON, Canada). Acetic acid (CH3COOH) and formic acid (HCOOH) (ACS grade, purity greater than 98%) were from EMD, Gibbstown, NJ, USA). Ammonium hydroxide (ACS grade, 28–30% NH3 w/v), a mixed l-amino acid standard (2.5 mM alanine, arginine, aspartic acid, glutamic acid, glycine, histidine, isoleucine, leucine, lysine hydrochloride, methionine, phenylalanine, proline, serine, threonine, tyrosine, and valine and 1.2 mM l-cysteine in 0.1 M HCl), acetonitrile (MeCN), and methanol (MeOH) (Optima LC–MS grade), were from Fisher Scientific (Ottawa, ON, Canada). Hydrochloric acid (high-performance LC grade, 85–90%), KA, and trichloroacetic acid (reagent grade) were obtained from Sigma-Aldrich (Oakville, ON, Canada). Polycaprolactone was from Plastic World (Toronto, ON, Canada).

Certified reference material (CRM) calibration solutions, including CRM-STX-f, CRM-NEO-d, CRM-dcSTX-b, CRM-GTX1&4-d, CRM-GTX2&3-d, CRM-dcGTX2&3-b, CRM-C1&2-b, CRM-GTX5-c, CRM-GTX6, CRM-dcNEO-d, CRM-DA-g, and CRM-TTX (prerelease), were obtained from National Research Council Canada (NRCC; Halifax, NS, Canada). A mixed standard calibration solution was prepared by our combining STX, decarbamoylsaxitoxin (dcSTX), neosaxitoxin (NEO), decarbamoylgonyautoxin 2 (dcGTX2), decarbamoylgonyautoxin 3, GTX1, GTX2, GTX3, GTX4, TTX, and DA standards gravimetrically. An in-house BMAA-d3 reference material (25 μM in 2 mM HCl) was prepared from material purchased from Abraxis (Warminster, PA, USA) and quantitated by 1H nuclear magnetic resonance spectroscopy [41]. 4,9-Anhydrotetrodotoxin and 4-epitetrodotoxin were made by our heating a 30 μM solution of TTX in 10 mM ammonium acetate, adjusted to pH 4.5 with CH3COOH, for 15 h at 80 C.

Matrix samples

Matrix CRMs CRM-Zero-Mus and CRM-ASP-Mus-d as well as a prerelease CRM for PSTs and DA (NRCC CRM-PSP-Mus) were obtained from NRCC. Two additional PST-positive mussel tissue matrix reference materials were used; these had already been extensively characterized by RPLC with postcolumn oxidation and fluorescence detection and HILIC–MS/MS [21, 42]. Two sea slug (Pleurobranchaea maculata) samples naturally containing TTX and several analogues were generously provided by the Cawthron Institute (Nelson, New Zealand). A dinoflagellate culture (Alexandrium tamarense) isolated from the St Lawrence estuary (Quebec, Canada) was also used. Farmed pufferfish fillets (Takifugu sp.) purchased from a supermarket in Qingdao City (China) were generously provided by Aifeng Li.

For qualitative work, a mussel tissue sample contaminated with TTXs, PSTs, and DA was simulated with a combination of approaches used previously to create laboratory reference materials and CRMs for these toxin classes [21, 24, 42]. This was done by our blending control mussel (Mytilus edulis) tissue (15 g), sea slug contaminated with TTXs (1.9 g), mussel tissue highly contaminated with DA (0.16 g), and A. tamarense that had been heat treated to stabilize its toxin profile (0.15 g). The mixture was vortexed, homogenized with a Polytron P6100 homogenizer, and vortexed again before being aliquoted into plastic storage vials.

Sample preparation

A modified version of the dispersive extraction procedure used previously for LC with postcolumn oxidation and fluorescence detection [26] and HILIC–MS/MS [21] method development and validation was used. Wet tissue homogenate was combined with an equal mass of 0.1 M HCl before homogenization with an OmniPrep tissue homogenizer at 10,000 rpm. Samples were boiled in a water bath for 5 min and allowed to cool to room temperature before centrifugation at 6700g for 10 min. Final volumes were determined gravimetrically as the sum of the extraction solvent and tissue. A 100-μL subsample of the supernatant was vortexed with 300 μL of MeCN to precipitate protein and other insoluble biomolecules, which were removed by spin filtration (0.22-μm Ultrafree-MC GV Durapore polyvinylidene difluoride filter, Merck Millipore, Cork, Ireland) at 2500g for 5 min. Then 250 μL of the filtrate was spiked with 16 μL of 25 μM BMAA-d3.

A matrix-matched calibration curve was prepared to test for method linearity, matrix effects, and use in calibration. Control mussel tissue (CRM-Zero-Mus) was extracted as described, and was spiked with a high-level mixed standard, and a sevenfold serial dilution was done with mussel extract. A 100-μL sample corresponding to each point was precipitated, filtered, and spiked with BMAA-d3 according to the procedure described above.

CE–MS/MS instrumentation and optimized method

CE separations were performed with a G1600A CE system (Agilent Technologies, Mississauga, ON, Canada). Bare fused-silica capillary tubing (50-μm inner diameter, 363-μm outer diameter) was obtained from Polymicro Technologies (Phoenix, AZ, USA) and cut to a length of 100 cm. Capillaries were conditioned by our flushing them with 1 M NaOH solution for 30 min at the start of each sequence. The optimized BGE consisted of 5 M HCOOH in 10% MeCN/H2O (v/v). Hydrodynamic injections were performed for 30 s at 50 mbar, and electrokinetic injections were performed for 45 s for matrix samples and 15 s for standards at a constant current of 50 μA. Between injections, the ESI voltage was switched off and the capillary was conditioned for 2 min with 1 M ammonium hydroxide, 1 min with deionized water, and 4 min with BGE. During runs, the capillary temperature was maintained at 17 °C by Peltier cooling while a positive separation voltage of 30 kV was applied. To prevent damage to the mass spectrometer electronics, the capillary current was kept below 50 μA.

The CE system was coupled to an API 4000 or a QTRAP 5500 mass spectrometer (Sciex, Concord, ON, Canada) in a fashion similar to that in previous reports [40] with use of positive ESI. The sprayer tip consisted of an Agilent CE–ESI-MS sprayer (part number G1607-60001) with a custom housing built with moldable polycaprolactone, and the sprayer tip oriented to point downward at an approximately 25° angle to the orifice plate. A Dino-Lite premier USB camera (Dino Canada, London, ON, Canada) was used to monitor spray stability (Fig. 2).
Fig. 2

Custom capillary electrophoresis (CE)–mass spectrometry interface connecting an Agilent 1600 CE system to a Sciex QTRAP 5500 mass spectrometer (MS)

Sheath liquid, composed of 1:1 MeOH/H2O with 0.1% HCOOH, was delivered via an Agilent 1100 series capillary pump (G1376A) at 4 μL min-1. Stable ESI was achieved with a spray voltage of 4000 V, coaxial nebulizing gas pressure of 4 psi, curtain gas pressure of 40 psi, and an interface heater temperature of 250 °C. Selected reaction monitoring (SRM) parameters are shown in Table 1. SRM transitions for other analytes tested qualitatively during method development are given as figure annotations. Initial work used a method that monitored all SRM transitions at all migration times, whereas the final method used for validation and quantitative work used migration time scheduling with a window of ±5 min around the measured migration time. Other instrument parameters included collision gas pressure of 5 psi, declustering potential of 130 V, collision cell entrance potential of 10 V, collision cell exit potential of 13 V, and unit/low (quadrupole 1/quadrupole 3) resolution.
Table 1

Capillary electrophoresis (CE) migration times measured and selected reaction monitoring parameters used in the final quantitative method used with the QTRAP 5500 mass spectrometer with positive electrospray ionization

Analyte

Migration time (min)a

Relative migration time (relative to BMAA-d3)a

Precursor ion m/z

Product ion m/z

CE (V)

dcSTX

18.6 ± 0.4

0.921 ± 0.003

257

180

30

257

222

25

STX

19.9 ± 0.4

0.982 ± 0.003

300

282

25

300

258

30

BMAA-d3

20.2 ± 0.4

1.000

122

76

17

122

88

20

NEO

20.6 ± 0.4

1.019 ± 0.002

316

220

30

316

298

25

TTX

32.8 ± 0.9

1.62 ± 0.02

320

302

35

320

284

35

dcGTX2

32.9 ± 0.9

1.63 ± 0.02

353

273

25

353

255

25

dcGTX3

33.6 ± 1

1.66 ± 0.02

353

255

25

353

273

25

GTX2

34.5 ± 1

1.70 ± 0.02

396

316

20

396

298

45

GTX3

35.0 ± 0.9

1.75 ± 0.01

396

298

25

396

316

25

GTX1

35.5 ± 1

1.76 ± 0.02

412

332

20

412

314

25

GTX4

36.2 ± 1

1.79 ± 0.03

412

314

25

412

332

20

DA

37.8 ± 0.7

1.87 ± 0.02

312

266

35

312

161

45

BMAA-d3 β-N-methylamino-l-alanine-d3, DA domoic acid, dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcSTX decarbamoylsaxitoxin, GTX1 gonyautoxin 1, GTX2 gonyautoxin 2, GTX3 gonyautoxin 3, GTX4 gonyautoxin 4, NEO neosaxitoxin, STX saxitoxin, TTX tetrodotoxin

aAbsolute and relative migration times represent the average ± combined standard deviation of five matrix and five neat standard injections.

Commercial CE–MS/MS system

An Agilent G7150A CE instrument was coupled to an Agilent 6400 triple-quadrupole mass spectrometer with use of a G1607B sheath flow CE–MS interface kit and a Jet Stream source operated in positive ionization mode with the nozzle voltage turned off. The CE BGE and samples were identical to those used with the other systems but with a sheath flow of 10 μL/min. The source parameters included a gas temperature of 300 C, gas flow of 10 L/min, nebulizer gas pressure of 10 psi, sheath gas temperature of 100 C, a sheath gas flow of 5 psi, and a capillary voltage of 3500 V. SRM parameters included a dwell time of 75 ms and transitions based on previous LC–MS/MS reports for similar MS systems [43, 44, 45], which are shown in detail in Table S1.

Results and discussion

CE–MS/MS method development

A custom sheath flow interface was modified to more effectively couple the CE instrument with the ESI-MS/MS system. This was based on the combination of an Agilent sheath flow ESI sprayer with the XYZ stage and interface ring of a Sciex nanospray source. Two significant improvements were made here over the original design [40], including the use of an improved USB camera to monitor spray stability and the use of a malleable polymer to construct a support for the sprayer at a downward angle of approximately 25 to the orifice plate (Fig. 2). This had the overall advantage of increasing spray stability compared with the on-axis orientation used previously, and also of minimizing contamination of the spray needle when the capillary was flushed. This generation of an interface was first used for method development and optimization with a Sciex API 4000 mass spectrometer and then subsequently transferred to the more sensitive 5500 QTRAP mass spectrometer for sample analysis. The MS/MS parameters for PSTs for both instruments were those previously optimized for HILIC–MS/MS [21], and those for TTX were a combination of literature parameters [24] and additional optimization with available standards.

Of the properties of an analyte that control electrophoretic mobility (size, shape, charge), charge has the biggest impact. For multifunctional organic analytes such as the marine toxins shown in Fig. 1, charge is mediated by solution pH. Above pH ~ 3, an electro-osmotic flow (EOF) is induced in CE when a bare fused-silica capillary is used. This flow is essential for the detection of neutral or negatively charged analytes when a positive field is applied in CE, but since all neutral species are detected together at the migration time of the EOF, this can lead to poor selectivity and ionization suppression in ESI. It was therefore desirable to investigate the impact of pH on CE separation with respect to varying the analyte charge state and EOF.

The charge state distribution of each toxin subclass was calculated between pH 0.2 and pH 12 with available published pKa data and the Henderson–Hasselbalch equation, and is shown in Fig. S1. Three different charge states in solution exist for PSTs, based on the number of hydroxysulfate groups in the structure and the pKa of the guanidinium groups, and further influenced by N-hydroxylation at R1 in Fig. 1 [46, 47, 48]. Below pH ~ 4, all STXs carry a double charge, all GTXs are singly charged, but C-toxins are neutral. Under most basic conditions, PSTs have a mixture of +1, neutral, and -1 charge states. TTX carries a single charge, and is fully protonated below pH ~ 7 [49]. DA, with one secondary amine and three carboxylic acids, has a complex charge state distribution, becoming predominantly protonated only at very low pH (less than 2) [50].

To investigate the migration and separation of marine toxins under a variety of conditions with and without EOF, a mixture of toxin standards was analyzed using several different BGE compositions spanning the pH range between 0.5 and 9. To account for changes in EOF, the migration times were normalized to the earliest migrating analyte, dcSTX. These are shown in Fig. 3 and confirm the charge state distribution calculations for each class from Fig. S1. Notably, neutral C-toxins are detected only when there is EOF present at pH ≥ 4. Similarly, DA was detected only when it retained a partial positive charge at pH < 3. The best resolution of PST isomers from one another occurred at low or high pH, but when trichloroacetic acid was used as the BGE at pH ≤ 0.8, the absolute migration times were extremely long (64 min to more than 120 min) and analyte peaks were wide. The best compromise between toxin resolution, run time, and sensitivity of analysis was observed when a high concentration of HCOOH (5 M) was used as the BGE along with 10% v/v MeCN to reduce the conductivity and keep the current below the system maximum of 50 μA. Under these conditions, all isomeric toxins examined are separated from one another. The only toxins that could not be analyzed under these conditions are the neutral C-toxins. C-toxins are regulated PSTs in shellfish in North America and Europe, but compared with the other PSTs examined, their relative toxicity is extremely low [51]. Since C-toxins and their corresponding GTXs share SRM transitions, it is important to completely resolve C-toxins from GTXs so as not to significantly overestimate toxicity. This can be challenging with rapid HILIC–MS/MS methods, but can be reliably done with longer HILIC separations [21] or, as achieved here, by CE–MS/MS. In cases where measurement of C-toxins is required, a second CE–MS/MS analysis of a sample after base hydrolysis of C-toxins to their corresponding GTXs could be done, as is done in precolumn oxidation workflows currently used in regulatory analysis [18].
Fig. 3

Impact of background electrolyte (BGE) pH on separation of toxins by capillary electrophoresis (CE). The composition of the BGE is shown across the top axis. DA domoic acid, dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcSTX decarbamoylsaxitoxin, GTX2 gonyautoxin 2, GTX3 gonyautoxin 3, NEO neosaxitoxin, STX saxitoxin, TTX tetrodotoxin

Another significant advantage of using a high-concentration (5 M) BGE is that it facilitates the use of field-amplified stacking as a method of online sample preconcentration [52]. In cases where samples with a higher conductivity than the BGE are injected, defocusing can occur as analytes migrate more slowly through the sample solution than through the BGE. Conversely, when the injected sample is of lower conductivity than the BGE, field-amplified stacking occurs because analytes migrate more quickly in the sample plug and preconcentrate at the boundary between the sample plug and the BGE. A 1:1 dilution of shellfish extracts with MeCN had previously been used in HILIC–MS analysis of PSTs as a method of sample cleanup for precipitating proteins and other biological macromolecules that have the potential to plug small particle size LC columns [21]. Here it was found that a higher dilution factor with MeCN (1:3) was more effective at preventing plugging of CE capillaries while also improving sample stacking by decreasing sample conductivity. Under the separation conditions chosen, negligible EOF is observed, requiring use of a sheath flow CE–MS interface. For this purpose, a sheath flow solvent of 4 μL/min 1:1 MeOH/H2O with 0.1% HCOOH was optimal, with small adjustments to the flow rate required to maintain a stable spray.

Activated carbon SPE sample cleanup [22] was investigated to see whether further lowering of sample ionic strength and conductivity could improve stacking or increase resolution as compared with MeCN dilution. The carbon SPE method was applied to a TTX-spiked extract of CRM-PSP-Mus, as published [22], and the MeCN dilution [21] was taken to the same dilution factor as for the SPE method for comparison. No significant qualitative or quantitative difference was observed between the two aliquots, and dilution with MeCN was therefore used for further studies.

The method developed was used for analysis of a wide variety of toxin standards, and for spiked and naturally incurred mussel and pufferfish samples. Figure 4 shows the separation of available standards for PSTs, TTXs, DA, and KA, as well as BMAA-d3, which was of interest as an internal standard because it had been analyzed previously under similar CE–MS conditions [40] and does not occur naturally. Unlike in HILIC, where doubly charged STXs are co-eluted with one another at a retention time similar to that of other doubly charged salts and matrix components [21], by CE they are well resolved from one another. As in HILIC, GTX epimers are all well separated from one another by CE and migrate just after the singly charged TTXs. Finally, DA, which accommodates only a partial positive charge at pH 1.5, has the longest migration time.
Fig. 4

Capillary electrophoresis–tandem mass spectrometry separation of toxin standards. 4,9-anhydroTTX 4,9-anhydrotetrodotoxin, BMAA-d3 β-N-methylamino-l-alanine-d3, DA domoic acid, dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcNEO decarbamoylneosaxitoxin, dcSTX decarbamoylsaxitoxin, 4-epi-TTX 4-epitetrodotoxin, GTX1 gonyautoxin 1, GTX2 gonyautoxin 2, GTX3 gonyautoxin 3, GTX4 gonyautoxin 4, GTX5 gonyautoxin 5, GTX6 gonyautoxin 6, KA kainic acid, NEO neosaxitoxin, STX saxitoxin, TTX tetrodotoxin

Another important consideration in analytical method development is selectivity between analytes and potentially interfering matrix components, which was addressed here in two experiments. First, a control mussel tissue matrix CRM, previously determined to be free of DA and PSTs by a variety of analytical methods, was analyzed by the CE–MS/MS method developed (Fig. S2). No significant matrix interferences were observed in this analysis of the control CRM. Further investigation into possible low-level TTX detection in this control sample will be done. In TTX analysis by HILIC–MS/MS, amino acids in the sample matrix, particularly arginine, have been identified as the principal matrix components leading to ionization suppression [24]. To investigate the potential for co-migration of amino acids and marine toxins, a mixed amino acid standard was also analyzed by the CE–MS/MS method developed, with use of previously published SRM parameters [53]. This separation (Fig. 5) showed excellent resolution of marine toxins from proteinogenic amino acids, with only glycine being detected in the same time range as the doubly charged STXs. All other amino acids migrated either before the doubly charged toxins or between the doubly and singly charged toxins.
Fig. 5

Analysis of a mixed amino acid standard using capillary electrophoresis separation and tandem mass spectrometry parameters published previously [53]. Arrows on the x-axis represent typical migration times of polar marine toxins (see Table 1)

It was of interest to analyze both shellfish and pufferfish matrices, since routine food safety testing is being done worldwide on both types of seafood for many of the marine toxins considered here. PSTs, TTXs, and DA are all known to occur in shellfish, particularly mussels, oysters, and clams [15]. However, it has principally been TTXs and STXs that have been reported in pufferfish [10, 54]. The analysis of a preliminary mussel tissue reference material for PSTs, TTXs, and DA is shown in Fig. 6a. CE–MS analysis of pufferfish tissue spiked with STXs and TTX is shown in Fig. 6b.
Fig. 6

Capillary electrophoresis–tandem mass spectrometry separation of tetrodotoxin (TTX), paralytic shellfish toxins, and domoic acid (DA) in a mussel tissue reference material (a) and toxin-spiked control pufferfish extract (b), each spiked with β-N-methylamino-l-alanine-d3 (BMAA-d3) as an internal standard. dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcSTX decarbamoylsaxitoxin, GTX2 gonyautoxin 2, GTX3 gonyautoxin 3, NEO neosaxitoxin, STX saxitoxin

Both hydrodynamic (pressure) injection and electrokinetic (voltage) injection [31] were investigated and used successfully during method development. In both cases, significantly less sample is injected than in LC (33 nL for hydrodynamic injection estimated with the zeecalc calculator [55]), which could help to limit the consumption of costly standards in routine monitoring. Electrokinetic injection had the advantage of increased sensitivity, fewer instances of capillary plugging from matrix samples, and more consistent migration times between standards and matrix samples. However, because of differential loading between analytes of different electrophoretic mobilities, differences in the response factor were observed between samples and standards when electrokinetic injection was used. This was evident when matrix-matched and neat standards were compared during the evaluation of matrix effects, as shown in Figs. 7 and S3. When a single internal standard (BMAA-d3) is used to normalize analyte signals, as shown in Fig. S3a, significant differences in response are observed between neat and matrix-matched standards, but this can be attributed primarily to differential loading of analytes depending on their electrophoretic mobility during electrokinetic injection, rather than matrix effects in ESI. BMAA-d3 does an excellent job of normalizing the response of the other doubly charged analytes (dcSTX, STX, NEO) but a poor job of normalizing the response of the singly charged analytes, particularly DA. Multiple internal standards were therefore used for the evaluation of matrix effects: BMAA-d3 for doubly charged analytes, TTX for GTXs, and KA for DA (Fig. 7). This approach showed minimal ionization suppression as the result of analyzing polar toxins in a sample matrix. The amount of sample loaded during hydrodynamic injection varies with sample viscosity. This was more straightforward to correct for with a single spiked injection standard. However, the overall sensitivity and reproducibility of this approach were too poor to allow routine use of the method. Therefore, matrix matching was still used in the quantitative analysis done here with electrokinetic injection. Future investigation of multiple internal standards to correct analytes of different mobilities should allow the use of an external standardization approach without matrix matching.
Fig. 7

Evaluation of relative molar response of polar toxins analyzed by capillary electrophoresis–tandem mass spectrometry and matrix effects in mussel tissue matrix. Bar heights represent the relative molar response of each analyte corrected for the response of the internal standard indicated and then normalized to the molar response of decarbamoylgonyautoxin 3 (dcGTX3). Error bars represent the standard deviation of triplicate injections. BMAA-d3 β-N-methylamino-l-alanine-d3, DA domoic acid, dcGTX2 decarbamoylgonyautoxin 2, dcSTX decarbamoylsaxitoxin, GTX1 gonyautoxin 1, GTX2 gonyautoxin 2, GTX3 gonyautoxin 3, GTX4 gonyautoxin 4, KA kainic acid, NEO neosaxitoxin, STX saxitoxin, TTX tetrodotoxin

Quantitative characterization of the CE–MS/MS method

The quantitative capabilities of the method developed were evaluated by our considering a range of typical validation parameters, including selectivity and matrix effects (discussed earlier) as well as linearity, LODs, precision, and accuracy.

Linearity, LODs, and matrix effects were examined by our analyzing neat and matrix-matched standard curves for a mixture of calibration solution CRMs for common PSTs, TTX, and DA. Good linearity (R2 > 0.99) was observed for all analytes for both matrix-matched and neat standards over the calibration range examined (Table 2). Instrument LODs (nM) were evaluated by our calculating the solution concentration at which a peak with signal-to-noise ratio of 3 would be observed with use of signal-to-noise ratios for low-level standards (signal-to-noise ratio approximately 10–20), and are shown in Table 2. These values were converted to method LODs (mg/kg tissue) with use of LODs of matrix-matched standards, sample masses, and dilution factors of sample preparation, and are also shown in Table 2. In general, these LODs for PSTs (0.0018–0.12 mg/kg) were similar to those recently reported for HILIC–MS/MS (0.002–0.02 mg/kg) with the same mass spectrometer [21], despite an extra twofold dilution used during sample preparation for CE–MS. In some cases, particularly the doubly charged STXs, HILIC–MS/MS showed lower LODs (e.g., 0.0085 mg/kg vs 0.120 mg/kg for dcSTX). In other cases, including GTX1, GTX2, and dcGTX2, CE–MS/MS showed lower LODs. It is expected that further improvements in ESI source interfacing could further increase the sensitivity for doubly charged STXs. The LODs make this method suitable for analysis of marine toxins at levels below the current regulatory limits (given in Table 2).
Table 2

Range, linearity, and limits of detection (LODs) of capillary electrophoresis–tandem mass spectrometry for marine toxins in shellfish and the regulatory limits

Analyte

Range investigated (μM)

R2 (matrix)

Instrument LOD (nM)

Method LOD (mg/kg tissue)

Regulatory limit

dcSTX

0.050–3.2

0.994

240

0.12

Total PSTs = 0.8 mg STX equivalents/kg

STX

0.042–2.7

0.992

57

0.034

NEO

0.052–3.3

0.997

59

0.037

dcGTX2

0.074–4.8

0.998

1.2

0.0083

dcGTX3

0.022–1.4

0.996

2.5

0.0018

GTX2

0.088–5.6

0.996

1.2

0.0038

GTX3

0.67–2.1

0.996

2.5

0.0079

GTX1

0.19–5.9

0.998

4.3

0.0035

GTX4

0.029–1.9

0.997

8

0.0064

TTX

0.034–2.2

0.998

8.1

0.0052

0.044 mg/kg

DA

1.1–17

0.998

250

0.16

20 mg/kg

DA domoic acid, dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcSTX decarbamoylsaxitoxin, GTX1 gonyautoxin 1, GTX2 gonyautoxin 2, GTX3 gonyautoxin 3, GTX4 gonyautoxin 4, NEO neosaxitoxin, PST paralytic shellfish toxin, STX saxitoxin, TTX tetrodotoxin

Precision and accuracy of the method developed were assessed by analysis of available matrix CRMs and in-house reference materials for polar marine toxins as well as spiked control samples with use of matrix-matched calibration. These included a CRM for DA (NRCC CRM-ASP-Mus) and a prerelease CRM for PSTs and DA (NRCC CRM-PSP-Mus) that was additionally spiked with TTX. A control pufferfish sample was spiked with toxin calibration solution CRMs to evaluate method performance for TTX in combination with PSTs in this matrix. Figure 8 shows the results of these quantitative runs as compared with reference values (including certified values) and the results of analysis of the same samples using established methods. In general, good agreement was observed along with good precision between triplicate sample injections.
Fig. 8

Quantitation of marine toxins in seafood samples by capillary electrophoresis–tandem mass spectrometry (CE-MS/MS) compared with certified or published values established with other validated methods including a a tetrodotoxin (TTX)-spiked mussel tissue matrix reference material for paralytic shellfish toxins and domoic acid (DA) (NRC CRM-PSP-Mus) [21], b a mussel tissue reference material certified for DA (NRC CRM-ASP-Mus), and c a control pufferfish sample spiked with paralytic shellfish toxins and TTX. DA domoic acid, dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcSTX decarbamoylsaxitoxin, GTX1 gonyautoxin 1, GTX2 gonyautoxin 2, GTX3 gonyautoxin 3, GTX4 gonyautoxin 4, LC-FLD liquid chromatography with fluorescence detection, LC-UV liquid chromatography with UV detection, NEO neosaxitoxin, STX saxitoxin

Method transfer to a commercial CE–MS/MS system

To investigate the feasibility of wider adoption of the CE–MS/MS method developed here, it was transferred to a commercially available CE–MS/MS system that does not rely on the custom-built interface. Because the CE method developed relies on a highly acidic BGE, for which no EOF is observed, a CE–MS interface with sheath flow was required, as opposed to a sheathless interface where the BGE serves as the spray solvent. For this work, the SRM parameters used were similar to those used in previous LC–MS studies that used similar mass spectrometers [43, 44, 45]. For the fragile GTX epimers (GTX1, GTX2, and dcGTX2) significant source fragmentation was observed, so the [M + H − SO3]+ source fragment was preferred as a precursor in SRM, as reported previously [43]. Figure 9 shows analysis of the same PST/TTX/DA mussel tissue reference material as in Fig. 6a, but analyzed with the commercial system. In general, similar sensitivity and resolution of GTX epimer pairs was observed between the two systems. The most notable difference was the increased sensitivity of analysis of the doubly charged PSTs (dcSTX, STX, and NEO), most notably dcSTX, which can be attributed to differences in MS source design between the two mass spectrometers used. It is expected that the method developed will be fully transferrable to the commercial CE–MS/MS system tested with minimal further optimization.
Fig. 9

Analysis of the paralytic shellfish toxin/tetrodotoxin (TTX)/domoic acid (DA) mussel tissue reference material with a commercial capillary electrophoresis–tandem mass spectrometry system. BMAA-d3 β-N-methylamino-l-alanine-d3, dcGTX2 decarbamoylgonyautoxin 2, dcGTX3 decarbamoylgonyautoxin 3, dcSTX decarbamoylsaxitoxin, GTX1 gonyautoxin 1, GTX2 gonyautoxin 2, NEO neosaxitoxin, STX saxitoxin

Conclusions

In conclusion, we have demonstrated the feasibility of using CE–MS/MS for multiclass analysis of polar marine toxins in seafood. This was facilitated by use of a very low pH BGE where EOF is almost entirely suppressed, but nearly all analytes investigated (with the exception of low-toxicity C-toxins) were protonated in solution. The CE separation conditions also afforded a high degree of sample stacking and a very high resolution separation of very similar toxin isomers and potential matrix interferences. The CE system was coupled to an ESI-MS/MS system with a custom-built interface that allowed the sensitive detection of PSTs, TTXs, and DA in mussel and pufferfish tissue, where they are known to co-occur. The LODs were low enough to allow analysis well below regulatory limits for each class of toxin in a single run, along with minimal ionization suppression.

Overall, CE–MS/MS appears to be a promising alternative to existing LC-based methods for the multiclass analysis of polar marine toxins. Therefore further work optimizing the conditions for a commercially available system and applying the method to a larger sample set of real samples is merited. The highly acidic BGE also has significant potential for more general application to other mixed class analysis of polar compounds by MS/MS.

Notes

Acknowledgements

The authors thank Michael Boundy from the Cawthron Institute and Aifeng Li from Ocean University of China for generously providing the sea slug and pufferfish tissue, respectively. Thanks are due to Jean-Francois Roy, Sylvie Larocque, and Dat Phan of Agilent Technologies for arranging access to the commercial capillary electrophoresis–tandem mass spectrometry system and for technical assistance with the related experiments.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no competing interests.

Supplementary material

216_2018_1089_MOESM1_ESM.pdf (420 kb)
ESM 1 (PDF 420 kb)

References

  1. 1.
    Sanseverino I, Conduto D, Pozzoli L, Dobricic S, Lettieri T. Algal bloom and its economic impact. Report no. EUR 27905 EN. Ispra: Joint Research Centre; 2016.Google Scholar
  2. 2.
    Sobel J, Painter J. Illnesses caused by marine toxins. Clin Infect Dis. 2005;41:1290–6.CrossRefPubMedGoogle Scholar
  3. 3.
    Starr M, Lair S, Michaud S, Scarratt M, Quilliam M, Lefaivre D, et al. Multispecies mass mortality of marine fauna linked to a toxic dinoflagellate bloom. PLoS One. 2017;12:e0176299.CrossRefPubMedPubMedCentralGoogle Scholar
  4. 4.
    Wells ML, Trainer VL, Smayda TJ, Karlson BS, Trick CG, Kudela RM, et al. Harmful algal blooms and climate change: Learning from the past and present to forecast the future. Harmful Algae. 2015;49:68–93.CrossRefPubMedPubMedCentralGoogle Scholar
  5. 5.
    Impellizzeri G, Mangiafico S, Oriente G, Piattelli M, Sciuto S, Fattorusso E, et al. Constituents of red algae. I. Amino acids and low-molecular-weight carbohydrates of some marine red algae. Phytochemistry. 1975;14:1549–57.CrossRefGoogle Scholar
  6. 6.
    Bates SS, Bird CJ, de Freitas ASW, Foxall R, Gilgan M, Hanic LA, et al. Pennate diatom Nitzschia pungens as the primary source of domoic acid, a toxin in shellfish from Eastern Prince Edward Island. Canada. Can J Fish Aquat Sci. 1989;46:1203–15.CrossRefGoogle Scholar
  7. 7.
    Quilliam MA, Wright JLC. The amnesic shellfish poisoning mystery. Anal Chem. 1989;61:1053A–60A.CrossRefPubMedGoogle Scholar
  8. 8.
    Lefebvre KA, Kendrick PS, Ladiges W, Hiolski EM, Ferriss BE, Smith DR, et al. Chronic low-level exposure to the common seafood toxin domoic acid causes cognitive deficits in mice. Harmful Algae. 2017;64:20–9.CrossRefPubMedPubMedCentralGoogle Scholar
  9. 9.
    Wiese M, D'Agostino PM, Mihali TK, Moffitt MC, Neilan BA. Neurotoxic alkaloids: saxitoxin and its analogs. Mar Drugs. 2010;8:2185–211.CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Bane V, Lehane M, Dikshit M, O'Riordan A, Furey A. Tetrodotoxin: chemistry, toxicity, source, distribution and detection. Toxins. 2014;6:693–755.CrossRefPubMedPubMedCentralGoogle Scholar
  11. 11.
    Chau R, Kalaitzis JA, Neilan BA. On the origins and biosynthesis of tetrodotoxin. Aquat Toxicol. 2011;104:61–72.CrossRefPubMedGoogle Scholar
  12. 12.
    Andrew D, Turner AP. A Schofield, D N Lees, C Baker-Austin. Detection of the pufferfish toxin tetrodotoxin in European bivalves, England, 2013 to 2014. Euro Surveillance. 2015;20:1–7.Google Scholar
  13. 13.
    Vlamis A, Katikou P, Rodriguez I, Rey V, Alfonso A, Papazachariou A, et al. First detection of tetrodotoxin in Greek shellfish by UPLC-MS/MS potentially linked to the presence of the dinoflagellate Prorocentrum minimum. Toxins. 2015;7:1779–807.CrossRefPubMedPubMedCentralGoogle Scholar
  14. 14.
    Knutsen HK, Alexander J, Barregård L, Bignami M, Brüschweiler B, Ceccatelli S, et al. Risks for public health related to the presence of tetrodotoxin (TTX) and TTX analogues in marine bivalves and gastropods. EFSA J. 2017;15:1–65.Google Scholar
  15. 15.
    Turner AD, Dhanji-Rapkova M, Coates L, Bickerstaff L, Milligan S, O'Neill A, Faulkner D, McEneny H, Baker-Austin C, Lees DN, Algoet M. Detection of tetrodotoxin shellfish poisoning (TSP) toxins and causative factors in bivalve molluscs from the UK. Mar Drugs. 2017;15:277-294.Google Scholar
  16. 16.
    AOAC International. AOAC official method 959.08. Paralytic shellfish poison, biological method. Gaithersburg: AOAC International; 2000.Google Scholar
  17. 17.
    AOAC International. AOAC official method 2011.02. Quantitative determination of paralytic shellfish toxins in mussels, clams, oysters, and scallops using liquid chromatography and post-column oxidation with fluorescence detection. Gaithersburg: AOAC International; 2011.Google Scholar
  18. 18.
    AOAC International. AOAC official method 2005.06, Quantitative determination of paralytic shellfish poisoning toxins in shellfish using pre-chromatographic oxidation and liquid chromatography with fluorescence detection. Gaithersburg: AOAC International; 2005.Google Scholar
  19. 19.
    van den Top HJ, Gerssen A, McCarron P, van Egmond HP. Quantitative determination of marine lipophilic toxins in mussels, oysters and cockles using liquid chromatography-mass spectrometry: inter-laboratory validation study. Food Addit Contam A. 2011;28:1745–57.Google Scholar
  20. 20.
    McCarron P, Wright E, Quilliam MA. Liquid chromatography/mass spectrometry of domoic acid and lipophilic shellfish toxins with selected reaction monitoring and optional confirmation by library searching of product ion spectra. J AOAC Int. 2014;97:316–24.CrossRefPubMedGoogle Scholar
  21. 21.
    Thomas KM, Beach DG, Reeves KL, Gibbs RS, Kerrin ES, McCarron P, et al. Hydrophilic interaction liquid chromatography-tandem mass spectrometry for quantitation of paralytic shellfish toxins: validation and application to reference materials. Anal Bioanal Chem. 2017;409:5675–87.CrossRefPubMedGoogle Scholar
  22. 22.
    Boundy MJ, Selwood AI, Harwood DT, McNabb PS, Turner AD. Development of a sensitive and selective liquid chromatography–mass spectrometry method for high throughput analysis of paralytic shellfish toxins using graphitised carbon solid phase extraction. J Chromatogr A. 2015;1387:1–12.CrossRefPubMedGoogle Scholar
  23. 23.
    Coleman R, Lemire SW, Bragg W, Garrett A, Ojeda-Torres G, Hamelin E, et al. Development and validation of a high-throughput online solid phase extraction - liquid chromatography - tandem mass spectrometry method for the detection of tetrodotoxin in human urine. Toxicon. 2016;119:64–71.CrossRefPubMedGoogle Scholar
  24. 24.
    Turner AD, Boundy MJ, Rapkova MD. Development and single-laboratory validation of a liquid chromatography tandem mass spectrometry method for quantitation of tetrodotoxin in mussels and oysters. J AOAC Int. 2017;100:1469–82.CrossRefPubMedGoogle Scholar
  25. 25.
    Shoji Y, Yotsu-Yamashita M, Miyazawa T, Yasumoto T. Electrospray ionization mass spectrometry of tetrodotoxin and its analogs: liquid chromatography/mass spectrometry, tandem mass spectrometry, and liquid chromatography/tandem mass spectrometry. Anal Biochem. 2001;290:10–7.CrossRefPubMedGoogle Scholar
  26. 26.
    van de Riet J, Gibbs RS, Muggah PM, Rourke WA, JD MN, Quilliam MA, et al. Liquid chromatography post-column oxidation (PCOX) method for the determination of paralytic shellfish toxins in mussels, clams, oysters, and scallops: collaborative study. J AOAC Int. 2011;94:1154–76.PubMedGoogle Scholar
  27. 27.
    Turner AD, Hatfield RG, Rapkova M, Higman W, Algoet M, Suarez-Isla BA, et al. Comparison of AOAC 2005.06 LC official method with other methodologies for the quantitation of paralytic shellfish poisoning toxins in UK shellfish species. Anal Bioanal Chem. 2011;399:1257–70.CrossRefPubMedGoogle Scholar
  28. 28.
    de la Iglesia P, Barber E, Gimenez G, Rodriguez-Velasco ML, Villar-Gonzalez A, Diogene J. High-throughput analysis of amnesic shellfish poisoning toxins in shellfish by ultra-performance rapid resolution LC-MS/MS. J AOAC Int. 2011;94:555–64.PubMedGoogle Scholar
  29. 29.
    Beach DG, Walsh CM, Cantrell P, Rourke W, O'Brien S, Reeves K, et al. Laser ablation electrospray ionization high-resolution mass spectrometry for regulatory screening of domoic acid in shellfish. Rapid Commun Mass Spectrom. 2016;30:2379–87.CrossRefPubMedPubMedCentralGoogle Scholar
  30. 30.
    Olivares JA, Nguyen NT, Yonker CR, Smith RD. On-line mass spectrometric detection for capillary zone electrophoresis. Anal Chem. 1987;59:1232–6.CrossRefGoogle Scholar
  31. 31.
    Smith RD, Udseth HR, Loo JA, Wright BW, Ross GA. Sample introduction and separation in capillary electrophoresis, and combination with mass spectrometric detection. Talanta. 1989;36:161–9.CrossRefPubMedGoogle Scholar
  32. 32.
    Thibault P, Pleasance S, Laycock MV. Analysis of paralytic shellfish poisons by capillary electrophoresis. J Chromatogr. 1991;542:483–501.CrossRefPubMedGoogle Scholar
  33. 33.
    Nguyen A-L, Luong JH, Masson C. Capillary electrophoresis for detection and quantitation of domoic acid in mussels. Anal Lett. 1990;23:1621–34.CrossRefGoogle Scholar
  34. 34.
    Locke SJ, Thibault P. Improvement in detection limits for the determination of paralytic shellfish poisoning toxins in shellfish tissues using capillary electrophoresis/electrospray mass spectrometry and discontinuous buffer systems. Anal Chem. 1994;66:3436–46.CrossRefGoogle Scholar
  35. 35.
    Pleasance S, Ayer SW, Laycock MV, Thibault P. Ionspray mass spectrometry of marine toxins. III. Analysis of paralytic shellfish poisoning toxins by flow-injection analysis, liquid chromatography/mass spectrometry and capillary electrophoresis/mass spectrometry. Rapid Commun Mass Spectrom. 1992;6:14–24.CrossRefPubMedGoogle Scholar
  36. 36.
    Pleasance S, Thibault P, Kelly J. Comparison of liquid-junction and coaxial interfaces for capillary electrophoresis-mass spectrometry with application to compounds of concern to the aquaculture industry. J Chromatogr. 1992;591:325–39.CrossRefPubMedGoogle Scholar
  37. 37.
    Keyon AS, Guijt RM, Gaspar A, Kazarian AA, Nesterenko PN, Bolch CJ, et al. Capillary electrophoresis for the analysis of paralytic shellfish poisoning toxins in shellfish: comparison of detection methods. Electrophoresis. 2014;35:1496–503.CrossRefPubMedGoogle Scholar
  38. 38.
    Gago-Martinez A, Rodriguez-Vazquez JA, Thibault P, Quilliam MA. Simultaneous occurrence of diarrhetic and paralytic shellfish poisoning toxins in Spanish mussels in 1993. Nat Toxins. 1996;4:72–9.CrossRefPubMedGoogle Scholar
  39. 39.
    Zhao J-Y, Thibault P, Quilliam MA. Analysis of domoic acid and isomers in seafood by capillary electrophoresis. Electrophoresis. 1997;18:268–76.CrossRefPubMedGoogle Scholar
  40. 40.
    Kerrin ES, White RL, Quilliam MA. Quantitative determination of the neurotoxin beta-N-methylamino-L-alanine (BMAA) by capillary electrophoresis-tandem mass spectrometry. Anal Bioanal Chem. 2017;409:1481–91.CrossRefPubMedGoogle Scholar
  41. 41.
    Burton IW, Quilliam MA, Walter JA. Quantitative 1H NMR with external standards: use in preparation of calibration solutions for algal toxins and other natural products. Anal Chem. 2005;77:3123–31.CrossRefPubMedGoogle Scholar
  42. 42.
    Reeves K, Thomas K, Quilliam MA. A mussel tissue certified reference material for paralytic shellfish poisoning toxins. In: Henshilwood K, Deegan B, McMahon T, Cusack C, Keaveney S, Silke J, et al., editors. Proceedings of the 5th international conference on molluscan shellfish safety. Galway: Marine Institute; 2006. p. 116-22.Google Scholar
  43. 43.
    Ding L, Qiu J, Li A. Proposed biotransformation pathways for new metabolites of paralytic shellfish toxins based on field and experimental mussel samples. J Agric Food Chem. 2017;65:5494–502.CrossRefPubMedGoogle Scholar
  44. 44.
    Rossi R, Arace O, Buonomo MG, Capozzo D, Castellano V, Imbimbo S, et al. Monitoring the presence of domoic acid in the production areas of bivalve molluscs. Ital J Food Saf. 2016;5:5706.PubMedPubMedCentralGoogle Scholar
  45. 45.
    Saito T, Miura N, Namera A, Ota S, Miyazaki S, Inokuchi S. A rapid sample preparation procedure using MonoSpin CBA and amide columns for tetrodotoxin detection in serum and urine using LC–MS/MS analysis. Chromatographia. 2014;77:687–93.CrossRefGoogle Scholar
  46. 46.
    Shimizu Y, Hsu C, Fallon WE, Oshima Y, Miura I, Nakanishi K. Structure of neosaxitoxin. J Am Chem Soc. 1981;100:6791–3.CrossRefGoogle Scholar
  47. 47.
    Shimizu Y, Hsu C, Genenah A. Structure of saxitoxin in solutions and stereochemistry of dihydrosaxitoxins. J Am Chem Soc. 1981;103:605–9.CrossRefGoogle Scholar
  48. 48.
    Rogers RS, Rapoport H. The pKa's of saxitoxin. J Am Chem Soc. 1980;102:7335–9.CrossRefGoogle Scholar
  49. 49.
    Goto T, Kishi Y, Takahashi S, Hirata Y. Tetrodotoxin. Tetrahedron. 1965;21:2059–88.CrossRefPubMedGoogle Scholar
  50. 50.
    Walter JA, Leek DM, Falk M. NMR study of the protonation of domoic acid. Can J Chem. 1992;70:1156–61.CrossRefGoogle Scholar
  51. 51.
    Selwood AI, Waugh C, Harwood DT, Rhodes LL, Reeve J, Sim J, et al. Acute toxicities of the saxitoxin congeners gonyautoxin 5, gonyautoxin 6, decarbamoyl gonyautoxin 2&3, decarbamoyl neosaxitoxin, C-1&2 and C-3&4 to mice by various routes of administration. Toxins. 2017;9:73–80.CrossRefPubMedCentralGoogle Scholar
  52. 52.
    Osbourn DM, Weiss DJ, Lunte CE. On-line preconcentration methods for capillary electrophoresis. Electrophoresis. 2000;21:2768–79.CrossRefPubMedPubMedCentralGoogle Scholar
  53. 53.
    Beach DG, Kerrin ES, Giddings SD, Quilliam MA, McCarron P. Differential mobility-mass spectrometry double spike isotope dilution study of release of b-methylaminoalanine and proteinogenic amino acids during biological sample hydrolysis. Sci Rep. 2018;8:117.CrossRefPubMedPubMedCentralGoogle Scholar
  54. 54.
    Landsberg JH, Hall S, Johannessen JN, White KD, Conrad SM, Abbott JP, et al. Saxitoxin puffer fish poisoning in the United States, with the first report of Pyrodinium bahamense as the putative toxin source. Environ Health Perspect. 2006;114:1502–7.CrossRefPubMedPubMedCentralGoogle Scholar
  55. 55.
    V. González-Ruiz, N. Drouin, E. Reginato, S. Rudaz, J. Schappler. zeecalc 1.0. https://epgl.unige.ch/labs/fanal/zeecalc. Accessed 5 April, 2018.

Copyright information

© Crown 2018

Authors and Affiliations

  • Daniel G. Beach
    • 1
  • Elliott S. Kerrin
    • 1
  • Krista Thomas
    • 1
  • Michael A. Quilliam
    • 1
  • Pearse McCarron
    • 1
  1. 1.Measurement Science and StandardsNational Research Council CanadaHalifaxCanada

Personalised recommendations