ORIGINAL PAPER Screening for multiple classes of marine biotoxins by liquid chromatography – high-resolution mass spectrometry Pearl Blay & Joseph P. M. Hui & James Chang & Jeremy E. Melanson Received: 17 November 2010 / Revised: 2 February 2011 / Accepted: 3 February 2011 / Published online: 24 February 2011 # Crown copyright in right of Canada 2011 Abstract Marine biotoxins pose a significant food safety risk when bioaccumulated in shellfish, and adequate testing for biotoxins in shellfish is required to ensure public safety and long-term viability of commercial shellfish markets. This report describes the use of a benchtop Orbitrap system for liquid chromatography – mass spectrometry (LC-MS) screening of multiple classes of biotoxins commonly found in shellfish. Lipophilic toxins such as dinophysistoxins, pectenotoxins, and azaspiracids were separated by reversed phase LC in less than 7 min prior to MS data acquisition at 2 Hz with alternating positive and negative scans. This approach resulted in mass accuracy for analytes detected in positive mode (gymnodimine, 13-desmethyl spirolide C, pectenotoxin-2, and azaspiracid-1, -2, and -3) of less than 1 ppm, while those analytes detected in negative mode (yessotoxin, okadaic acid, and dinophysistoxin-1 and -2) exhibited mass errors between 2 and 4 ppm. Hydrophilic toxins such as domoic acid, saxitoxin, and gonyautoxins were separated by hydrophilic interaction LC (HILIC) in less than 4 min, and MS data was collected at 1 Hz in positive mode, yielding mass accuracy of less than 1 ppm error at a resolving power of 100,000 for the analytes studied ( m/z 300 – 500). Data were processed by extracting 5 ppm mass windows centered around the calculated masses of the analytes. Limits of detection (LOD) for the lipophilic toxins ranged from 0.041 to 0.10 μ g/L (parts per billion) for the positive ions, 1.6 – 5.1 μ g/L for those detected in negative mode, while the domoic acid and paralytic shellfish toxins yielded LODs ranging from 3.4 to 14 μ g/L. Toxins were detected in mussel tissue extracts free of interference in all cases. Keywords Marine biotoxins . Paralytic shellfish toxins . High-resolution mass spectrometry . Accurate mass screening . LC-MS . HILIC Introduction Marine biotoxins are produced by naturally occurring microalgae such as Alexandrium tamarense , Dinophysis acuminata , and Azadinium spinosum [1]. Under certain environmental conditions, algal populations can increase significantly to form a harmful algal bloom (HAB). During the incidence of a HAB, marine biotoxins pose a significant food safety risk when bioaccumulated in shellfish. There- fore, adequate testing for biotoxins in shellfish is required to ensure public safety and long-term viability of commer- cial shellfish markets. Since the 1980s, official testing in many countries has been carried out with the mouse bioassay [2]. In addition to ethical concerns surrounding animal usage, the mouse bioassay suffers from poor precision and does not have sufficient sensitivity to detect certain lipophilic toxins, such as the okadaic acid group, at the current European Union regulatory limits [3]. There- fore, there are increasing efforts worldwide to develop and validate instrumental methods that could replace the mouse bioassay [4]. Electronic supplementary material The online version of this article (doi:10.1007/s00216-011-4772-2) contains supplementary material, which is available to authorized users. P. Blay : J. P. M. Hui : J. E. Melanson ( * ) National Research Council of Canada, Institute for Marine Biosciences, 1411 Oxford Street, Halifax, NS B3H 3Z1, Canada e-mail: jeremy.melanson@nrc-cnrc.gc.ca J. Chang Thermo Fisher Scientific, 355 River Oaks Parkway, San Jose, CA 95134, USA Anal Bioanal Chem (2011) 400:577 – 585 DOI 10.1007/s00216-011-4772-2 Marine toxins can be divided into two general classi- fications based on their water solubility. The lipophilic toxins include the dinophysistoxins, azaspiracids, pecteno- toxins, and yessotoxins. The lipophilic toxins are extremely structurally diverse and thus do not contain a common UV chromophore or reactive functional group for fluorescence derivatization. Therefore, liquid chromatography – mass spectrometry (LC-MS) is the method of choice for their analyses, and several methods have been reported [5 – 7], one of which has undergone a successful single-laboratory validation [8]. Chemical structures of the lipophilic toxins can be found in Fig. S1 (Electronic Supplementary Material). The hydrophilic toxins include domoic acid and the paralytic shellfish-poisoning toxins such as saxitoxin, neosaxitoxin, and gonyautoxins. Domoic acid has conven- tionally been analyzed by LC-UV [9], but specific LC-MS methods have also been reported [10]. Despite the existence of an AOAC accredited method for the paralytic shellfish- poisoning toxins based on pre-column oxidation with fluorescence detection [11, 12], method development continues for the paralytic shellfish poisoning (PSP) toxins using post-column oxidation and fluorescence detection [13, 14] and also LC-MS [15 – 17]. Therefore, general acceptance of one particular analytical technique for domoic acid and the PSP toxins has yet to be established. Chemical structures of domoic acid and the paralytic shellfish toxins are shown in Fig. S2 (Electronic Supple- mentary Material). In general, LC-MS analysis of biotoxins is performed on triple-quadrupole MS operating in multiple reaction mon- itoring (MRM) mode [5 – 7]. This approach is well suited for quantification due to its inherent selectivity and high sensitivity. However, the relatively low resolution of quadrupoles (typically unit resolution) renders the tech- nique prone to interference from ions of similar mass in complex samples [18]. In addition, significant method development time is generally required, as MRM transi- tions are typically optimized for each analyte. Finally, due to the targeted nature of MRM, only known toxins specified in the method will be detected. Therefore, new or modified biotoxins could remain undetected indefinitely, even at high abundance. To overcome many of the issues associated with targeted analysis, LC coupled to high-resolution MS has been successfully implemented for screening and quantification in food safety [19, 20], pharmaceutical [21, 22], and environmental [23] applications. The lower cost, higher mass accuracy, and ease of use of modern quadrupole time- of-flight (QTOF) and Orbitrap-based mass spectrometers have made high-resolution systems viable alternatives to triple-quadrupole systems for routine analysis. After full- spectrum data acquisition, specificity is typically achieved by extracting narrow mass windows (i.e., 2 – 5 ppm) centered around a list of target analytes. Using this approach, it has been demonstrated that a resolving power of 50,000 or greater was required for correct mass assign- ments of 151 pesticides, veterinary drugs, mycotoxins, and plant toxins, in both honey and animal feed matrices [20]. This non-targeted approach provides high-resolution data over the entire chromatographic separation, allowing detection of new or unknown compounds in addition to those of interest. Furthermore, the technique generally requires less method development time than MRM analy- ses, as settings are not tuned for individual analytes. This report describes the use of a single high-resolution MS platform for screening multiple classes of biotoxins commonly found in shellfish. Lipophilic toxins such as dinophysistoxins, pectenotoxins, and azaspiracids were analyzed by a reversed phase LC method, while hydrophilic toxins such as domoic acid and PSP toxins were separated by hydrophilic interaction LC (HILIC). The methods were optimized using mixtures of biotoxin standards, and then applied to various shellfish tissue extracts. Experimental Chemicals and materials Certified calibration solutions and mussel tissue reference materials were purchased from the NRC Certified Refer- ence Materials Program (Halifax, NS, Canada). Certified calibration solutions were used for the following biotoxins: domoic acid (DA), N -sulfocarbamoyl toxins (C1 and C2), gonyautoxins (GTX1, GTX2, GTX3, GTX4, and GTX5), decarbamoylgonyautoxins (dcGTX2 and dcGTX3), neo- saxitoxin (NEO), saxitoxin (STX), gymnodimine (GYM), 13-desmethyl spirolide C (SPX1), yessotoxin (YTX), okadaic acid (OA), pectenotoxin-2 (PTX2), and azaspiracid 1 (AZA1). Calibration solutions for azaspiracid 2 (AZA2) and azaspiracid 3 (AZA3) were not certified but have preliminary values assigned [24], as have the calibration s o l u t i o n s f o r d i n o p h y s i s t o x i n - 1 ( D T X 1 ) a n d dinophysistoxin-2 (DTX2). Concentrations of the calibra- tion solutions used are listed in Table S1 (Electronic Supplementary Material). A mussel tissue material contain- ing the hydrophilic toxins DA, GTX1, GTX2, GTX3, GTX4, dcGTX2, dcGTX3, STX, and NEO was employed as a test sample. Similarly, a freeze-dried mussel tissue reference material (NRC RM-FDMT) contained prelimi- nary certified levels of DA, OA, DTX1, DTX2, YTX, PTX2, AZA1, AZA2, AZA3, and SPX1 [25] was used as a test sample for the lipophilic toxins method. HPLC grade acetonitrile and formic acid (98%) were purchased from EMD chemicals (Gibbstown, NJ, USA). Distilled-in-glass grade methanol was acquired from 578 P. Blay et al. Caledon Laboratories (Georgetown, ON, Canada), and ammonium formate ( ≥ 99.0%) was from Fluka (St. Louis, MI, USA). Reconstitution and extraction of lipophilic toxins from freeze-dried mussel tissue Reconstitution of the freeze-dried tissue consisted of adding 1.65 mL of deionized water to 0.35 g of freeze-dried tissue, followed by vortex mixing for 30 s and sonication for 1 min in an ultrasonic bath. For extraction of lipophilic toxins, 2 g of reconstituted tissue was homogenized with 5.5 mL of methanol using a multi-tube vortex mixer (Model DVX- 2500, VWR International, Radnor, PA, USA) at 2,500 rpm for 3 min. The sample was then centrifuged at 2,300× g for 15 min, and the supernatant was decanted into a 25-mL volumetric flask. The pellet was extracted three times further using the same procedure, and the supernatants from each step were combined. The final volume was made up to 25 mL with methanol. Approximately 0.5 mL of this solution was filtered through a 0.45- μ m spin-filter (Millipore, Billerica, MA, USA) prior to analysis. Extraction of domoic acid and PSP toxins from mussel tissue Approximately 0.5 g of mussel tissue was transferred to a microcentrifuge tube into which 0.5 mL of 0.1 M HCl was added. The tube was vortexed, adjusted to pH 3 with 1 M ammonium hydroxide to minimize toxin conversion, and was boiled in a hot water bath for 5 min [14]. It was cooled down to room temperature and was centrifuged at 5,000 rpm for 10 min, after which, the supernatant was carefully transferred to a 3-mL volumetric flask. A second aliquot of 0.5 mL of 0.1 M HCL was added to the same microcentrifuge tube, the extraction procedure was repeated, and the resulting supernatant was combined with the first one. The volumetric flask was filled up to 3 mL with acetonitrile. The supernatant was filtered through a 0.22- μ m spin filter (Millipore) prior to the LC-MS analysis. LC-MS instrumentation LC-MS analysis was carried out on an Accela ™ High Speed LC coupled to an Exactive ™ mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA), equipped with an Orbitrap mass analyzer and a heated electrospray ionization probe (HESI-II). The instrument was mass calibrated daily for positive and negative modes, and the capillary, tube lens, and skimmer voltages were optimized daily, using the automated script within the ExactiveTune 1.1 software in both cases. For positive mode, mass calibration was performed with a mixture consisting of caffeine, MRFA tetrapeptide, and Ultramark 1621, while the negative mode calibration was performed with sodium dodecyl sulfate, sodium taurocholate, and Ultramark 1621. All analyses were performed using the “ balanced ” auto- matic gain control (AGC) setting with a 50-ms maximum inject time across a mass range from m/z 100 to 2,000. Data acquisition was carried out using Xcalibur 2.1 (Thermo Fisher Scientific). LC-MS method for lipophilic toxins Lipophilic toxins were separated on a 2.1×100-mm, 1.9- μ m particle Hypersil Gold C18 column (Thermo Fisher Scientific), at a flow rate of 400 μ L/min and using 3- μ L injections. Mobile phases were prepared from a stock solution of 1% formic acid solution in water with the pH adjusted to 3.0 using concentrated ammonium hydroxide. This stock solution was then diluted tenfold with water (A) or acetonitrile (B), resulting in 0.1% formic acid in water for mobile phase A and 0.1% formic acid in 90% acetonitrile ( v / v ) for B. Analytes were eluted with a linear gradient from 20% to 100% B from 0 to 5 min, held for 2 min, before returning to the initial conditions of 20% B. Optimal ion source and interface conditions consisted of a spray voltage of 3 kV in positive mode or − 2.7 kV in negative mode, sheath gas flow of 50, auxiliary gas flow rate of 10, capillary temperature of 360 °C, and a heater temperature of 250 °C. Alternating positive and negative polarity scans were acquired at a scan rate 2 Hz (50,000 resolution instrument specification) for an overall cycle time greater than 1 s. It should be noted that this scan rate generated resolution at full-width-half-maximum (FWHM) of roughly 25,000 in the mass range of the lipophilic toxins (i.e., m / z 800). LC-MS method for domoic acid and PSP toxins Separation of DA and PSP toxins was carried out on a Waters Acquity BEH Amide column (2.1 × 100 mm, 1.7 μ m, Milford, MA, USA) operating at 30 °C, at a flow rate of 500 μ L/min, using 3- μ L injections. The solvent system comprised of 2 mM ammonium formate pH 3.5 (A) and 0.1% formic acid in acetonitrile (B). Analytes were eluted with a linear gradient from 75% to 55% B in 4 min, held for 1 min, before returning to the initial conditions of 75% B. Optimal ion source and interface conditions consisted of a spray voltage of 3 kV, sheath flow of 55, auxiliary gas flow rate of 15, capillary temperature of 150 °C, and a heater temperature of 200 °C. Positive polarity scans were acquired at 1 Hz (100,000 resolution instrument specifica- tion) for a cycle time of roughly 1 s. Screening for multiple classes of marine biotoxins by LC-MS 579 Results and discussion High-resolution LC-MS method for lipophilic toxins Lipophilic toxins were separated by reversed phase chro- matography coupled to the Exactive ™ mass spectrometer. As shown in Fig. 1, ten lipophilic toxin standards were baseline separated in just under 6 min, and the data shown represents 5 ppm extracted mass chromatograms centered around the calculated masses of the target analytes. As OA, DTX1, DTX2, and YTX ionize significantly better in negative mode, alternative positive and negative polarity scans were acquired to achieve maximum signal for all analytes. To maintain a sufficient number of data points across chromatographic peaks, data was collected at a scan rate of 2 Hz for an overall cycle time greater than 1 s, yielding resolution of roughly 25,000 at m/z 800. This scan rate generated much lower resolution than that possible with the mass spectrometer, but was selected as a reasonable compromise between selectivity and quantitative performance. The ability to rapidly scan both positive and negative polarities allows data collection in a true non- targeted fashion and permits independent optimization of the LC method without consideration of the retention time of positive and negative analytes. In contrast, some LC-MS methods on triple-quadrupole mass spectrometers employ extreme mobile phase pH conditions (i.e., pH 11) to group positive and negative ionizing analytes into different time regions of the chromatogram to enhance the duty cycle of the MRM method and maintain sufficient sensitivity [7]. Operating at pH 11 limits LC column selection to only a handful of choices, and many laboratories prefer to avoid the use of such high pH for routine analysis. Listed in Table 1 are mass measurement data and limits of detection (LOD) for the lipophilic toxins obtained with a mixture of toxin standards. Based on triplicate injections, Table 1 lists averages of the measured masses with their corresponding standard deviations. The errors listed in Table 1 represent averages of the absolute values of the individual mass errors [26], thereby eliminating positive and negative errors canceling each other. In general, accurate masses were measured within 1 ppm error for analytes detected in positive mode, while those detected in 25x10 6 20 15 10 5 0 Intensity 7 6 5 4 3 2 1 0 Time (min) GYM SPX YTX OA DTX2 PTX2 AZA3 DTX1 AZA1 AZA2 Fig. 1 Reversed phase LC-MS chromatograms of ten lipophilic biotoxin standards acquired with alternating positive (GYM, SPX, PTX2, and AZA-1, -2, and -3) and negative (YTX, OA, and DTX-1 and -2) scans at 2 Hz. Data shown represents 5 ppm mass windows centered around the analyte mass Table 1 Mass measurement data and LODs for the lipophilic toxin class Toxin ID T ret (min) Chemical formula Ion detected Calculated m / z Measured m / z Measured std. dev. Error (ppm) LOD ( μ g/L) GYM 2.80 C 32 H 45 NO 4 [M+H] + 508.34214 508.34196 0.00025 0.47 0.041 SPX1 3.27 C 42 H 61 NO 7 [M+H] + 692.45208 692.45196 0.00045 0.52 0.054 YTX 4.63 C 55 H 82 O 21 S 2 [M − H] − 1,141.47063 1,141.47421 0.00032 3.1 5.1 OA 4.81 C 44 H 68 O 13 [M − H] − 803.45762 803.46053 0.00046 3.6 2.8 DTX2 5.04 C 44 H 68 O 13 [M − H] − 803.45762 803.46064 0.00041 3.8 1.6 PTX2 5.19 C 47 H 70 O 14 [M+NH 4 ] + 876.51038 876.50971 0.00063 0.78 0.10 AZA3 5.45 C 46 H 69 NO 12 [M+H] + 828.48925 828.48848 0.00031 0.93 0.062 DTX1 5.59 C 45 H 70 O 13 [M − H] − 817.47326 817.47712 0.00025 4.7 2.0 AZA1 5.78 C 47 H 71 NO 12 [M+H] + 842.50490 842.50427 0.00036 0.75 0.052 AZA2 5.96 C 48 H 73 NO 12 [M+H] + 856.52055 856.51969 0.00035 1.0 0.064 580 P. Blay et al. negative mode yielded errors between 3 and 5 ppm. It should be noted that accurate masses were achieved using external calibration exclusively, without any mass correc- tion on an internal standard or a background ion. The larger error associated with negative ions is consistent with most mass spectrometers, presumably as ion optics are typically optimized for the more commonly detected positive ions. Similarly, LODs ranged from 0.041 to 0.099 μ g/L (parts 2.0x10 7 1.5 1.0 0.5 0.0 Intensity 7 6 5 4 3 2 1 0 Time (min) B 3.0x10 2.5 2.0 1.5 1.0 0.5 0.0 Intensity D 1.5x10 7 1.0 0.5 0.0 Intensity 846.0 845.0 844.0 843.0 842.0 m/z 842.50488 844.51111 843.50848 845.51520 AZA1 3x10 5 2 1 0 Intensity 80 7.0 806.0 805.0 804.0 803.0 802.0 m/z 803.46 143 804.46533 805 .46844 DTX2 1.6x10 9 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 Intensity 7 6 5 4 3 2 1 0 TIC (+) 3.0x10 2.5 2.0 1.5 1.0 0.5 0.0 Intensity 7 6 5 4 3 2 1 Time (min) C TIC (-) SPX PTX2 (5X) AZA3 AZA1 AZA2 5 YTX OA DTX2 DTX1 1.5x10 7 1.0 0.5 0.0 Intensity 846.0 845.0 844.0 843.0 842.0 m/z 842.50488 844.51111 843.50848 845.51520 AZA1 3x10 5 2 1 0 Intensity 80 7.0 806.0 805.0 804.0 803.0 802.0 m/z 803.46 143 804.46533 805 .46844 DTX2 7 6 5 4 3 2 1 0 Time (min) A TIC (+) 8 7 6 5 4 3 2 1 7 6 5 4 3 2 1 Time (min) 7 6 5 4 3 2 1 TIC (-) Fig. 2 Analysis of freeze-dried mussel tissue reference material containing various lipophilic toxins. a TIC of positive mode scans. b LC-MS chromatogram at 5 ppm of biotoxins detected in positive mode; inset , representative mass spectrum of AZA1. c TIC of negative mode scans. d LC-MS chromatogram at 5 ppm of analytes detected in negative mode; inset , representative mass spectrum of DTX2. Positive and negative ions were detected simultaneously via polarity switching but are shown in different traces for clarity due to the much larger response for positive ions Screening for multiple classes of marine biotoxins by LC-MS 581 per billion) for the positive ions, while those detected in negative mode are distinctly higher at 1.6 – 5.1 μ g/L. The utility of the screening method for lipophilic toxins was evaluated by analyzing a pre-release freeze- dried mussel tissue reference material containing prelim- inary certified levels of various lipophilic toxins (120 – 700 μ g/kg). As naturally contaminated mussel tissue rarely contains more than one or two different biotoxins, this reference material was chosen as a test sample as it offered a broader range of toxins and was able to test the entire method simultaneously. In addition, this material also offered a reasonable test of the sensitivity required for screening applications, as toxin levels were very near or in some cases lower than the current EU regulatory limits [25]. Shown in Fig. 2 are the high-resolution LC- MS chromatograms for the mussel tissue reference material. Although both positive and negative ions were detected simultaneously via polarity switching, they are shown in different traces for clarity due to the much larger response for positive ions. Figure 2a,c represent the total ion chromatograms (TIC) in positive and negative mode, respectively. Although the TICs do not reveal an over- whelming number of peaks, it should be noted that the scales of the TIC traces are two and three orders magnitude larger than their respective positive and negative analyte traces. Therefore, the biotoxins are extremely minor components in a very complex matrix. As shown in Fig. 2b, specificity was demonstrated by the minimal background peaks detected in the 5-ppm mass windows associated with SPX1, PTX2, AZA3, AZA1, and AZA2. Figure 2b (inset) shows the corresponding mass spectrum of AZA1 from a single scan at retention time 5.76 min. The monoisotopic peak at m / z 842.50488 falls within 0.2 ppm of the calculated mass, and the resolution was measured at 28,100, demonstrating no impact of the complex matrix of analytical performance. Similarly, Fig. 2d displays the LC-MS chromatograms for YTX, OA, DTX2, and DTX1 detected in negative mode. Shown in Fig. 2d (inset) is the corresponding mass spectrum of DTX2 for a single scan at retention time 5.05 min. The monoisotopic peak detected at m / z 803.46143 is within 4.7 ppm of the calculated mass, and the resolution was measured at 26,400. It should be noted that domoic acid was also detected in this mussel tissue reference material, but yielded relatively poor peak shape on the particular column employed, such that LODs were significantly 3.5x10 6 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Intensity 5 4 3 2 1 0 Time (min) Domoic acid C1 C2 GTX2 396.09321 GTX1 GTX4 dcGTX2 dcGTX3 GTX5 STX NEO GTX2-SO 3 312.14416 493.07657 GTX3 412.08812 353.08739 380.09829 300.14148 316.13639 Fig. 3 Hydrophilic interaction liquid chromatography (HILIC-MS) chromatograms at 5 ppm of hydrophilic toxin standards consisting of domoic acid and paralytic shellfish toxins Table 2 Mass measurement data and LODs for domoic acid and the paralytic shellfish toxins Toxin ID T ret (min) Chemical formula Ion detected Calculated m / z Measured m / z Measured std. dev. Error (ppm) LOD ( μ g/L) DA 0.81 C 15 H 21 NO 6 [M+H] + 312.14416 312.14422 0.000042 0.18 3.4 C1 2.01 C 10 H 17 N 7 O 11 S 2 [M+NH 4 ] + 493.07657 493.07688 0.000165 0.64 14 C2 2.26 C 10 H 17 N 7 O 11 S 2 [M+NH 4 ] + 493.07657 493.07693 0.000144 0.73 11 GTX2 2.35 C 10 H 17 N 7 O 8 S [M+H] + 396.09321 396.09326 0.000095 0.21 7.5 GTX1 2.41 C 10 H 17 N 7 O 9 S [M+H] + 412.08812 412.08823 0.000117 0.29 9.9 dcGTX2 2.46 C 9 H 16 N 6 O 7 S [M+H] + 353.08739 353.08735 0.000113 0.21 7.2 GTX3 2.60 C 10 H 17 N 7 O 8 S [M+H] + 396.09321 396.09321 0.000124 0.24 4.5 GTX4 2.67 C 10 H 17 N 7 O 9 S [M+H] + 412.08812 412.08821 0.000070 0.23 6.7 dcGTX3 2.72 C 9 H 16 N 6 O 7 S [M+H] + 353.08739 353.08733 0.000147 0.30 4.0 GTX5 3.04 C 10 H 17 N 7 O 7 S [M+H] + 380.09829 380.09824 0.000093 0.18 6.9 STX 3.75 C 10 H 17 N 7 O 4 [M+H] + 300.14148 300.14123 0.000040 0.84 6.0 NEO 3.86 C 10 H 17 N 7 O 5 [M+H] + 316.13639 316.13619 0.000049 0.64 8.1 582 P. Blay et al. higher than that obtained in HILIC mode (see below). In addition, various other toxin analogs were also detected (data not shown), simply by expanding on the target list of analyte masses during data processing. High-resolution LC-MS method for domoic acid and paralytic shellfish poisoning toxins As domoic acid and the PSP toxins exhibit little or virtually no retention on reversed phase columns, alternate separa- tion modes are typically employed for their analysis. The use of ion-pair reagents in the mobile phase has been successful at separation of PSP toxins [11, 13], but these reagents are typically not compatible with MS detection. HILIC coupled to MS had been reported as a viable alternative for analyzing PSP toxins [15 – 17]. HILIC separation of PSPs generally requires an additional mech- anism for selectivity beyond conventional bare silica particles, such as amide or zwitterionic functionality. Until recently, a commercial column with such additional HILIC selectivity has not been available to take advantage of modern high-pressure LC pumps, and separations have been limited to 30 – 45 min. Shown in Fig. 3 is the LC-MS chromatogram of domoic acid and 11 PSP toxin standards using an Acquity BEH- Amide column in HILIC mode. Baseline separation could not be achieved over the 4-min run for the gonyautoxins (GTX2-3 and GTX1-4), so the mass chromatograms are displayed offset for clarity. However, baseline separation of all structural isomers was achieved (i.e., GTX1 and GTX4), so all co-eluting analytes differed in mass. Data was acquired in positive mode only as domoic acid, and all known PSP toxins ionize very effectively in positive mode. The scan rate employed was 1 Hz, and this yielded a sufficient number of data points across chromatographic peaks given the broader peaks generally observed in HILIC mode. This scanning speed has an instrument setting of 100,000, and this was near the actual resolution observed in most cases for the analytes in the m/z 300 – 400 range. Other than the scan rate, the only significant difference between the MS method employed for the PSP toxins was the milder source conditions (i.e., heater temperature of 200 °C, capillary temperature of 150 °C). Above these temper- atures, significant in-source fragmentation occurred. As listed in Table 2, accurate masses for the mixture of toxin standards were measured within 1 ppm error, based on averages of the absolute values of mass errors from three replicates. Linear calibration was achieved over three orders of magnitude, peak area RSDs ranged from 5% to 10%, and LODs ranged from 3.4 to 14 μ g/L. 2.5x10 4 2.0 1.5 1.0 0.5 0.0 Intensity 300.35 300.30 300.25 300.20 300.15 300.10 300.05 300.00 m/z STX 300.14139 300.16171 300.20154 Fig. 5 Mass spectrum of mussel tissue containing hydrophilic toxins at retention time 3.60 min, showing the monoisotopic peak of STX at m / z 300.14139 and potential interference peaks at m / z 300.16171 and 300.20154 2.5x10 6 2.0 1.5 1.0 0.5 0.0 Intensity 6 5 4 3 2 1 0 Time (min) GTX2 GTX3 GTX1 GTX4 dcGTX2 dcGTX3 STX NEO GTX2-SO 3 Domoic acid 312.14416 (reduced 100X) TIC (reduced 1000X) 396.09321 412.08812 353.08739 300.14148 316.13639 Fig. 4 HILIC-MS analysis of preliminary mussel tissue reference material containing domoic acid and paralytic shellfish poisoning toxins Screening for multiple classes of marine biotoxins by LC-MS 583 To demonstrate the applicability of the method to screening shellfish tissues, a preliminary reference material containing domoic acid, GTX1, GTX2, GTX3, GTX4, dcGTX2, dcGTX3, STX, and NEO was analyzed, as shown in Fig. 4. The top trace of Fig. 4 depicts the TIC, while the other traces represent 5 ppm mass window centered around the analyte masses. As with the lipophilic toxins, all analytes were detected void of any interference from the complex mussel matrix. Due to the consistently high mass accuracies of the hydrophilic toxins (all below 1 ppm), mass windows as narrow as 2 ppm could be extracted without any loss of signal. This could be advantageous for even more complex samples, but 5 ppm was sufficient to filter out most interference from the mussel tissue matrix, as shown in Fig. 4. Figure 5 demonstrates the greater need for high resolution at the lower mass range of the hydrophilic toxins, whereby interference peaks m / z 300.16171 and 300.20154 are clearly resolved from the monoisotopic peak of STX at m / z 300.14139. With a mass difference of 68 ppm between the STX peak and peak at m / z 300.16171, this interference is easily removed with the 5-ppm mass filter. Conclusions A single LC-MS platform has been developed for non- targeted screening of two major classes of biotoxins commonly found in shellfish. Although two different modes of separation were employed for the two classes, the mass spectrometer acquisition methods differ only in a few ionization sources conditions, scan rate, and the polarity switching for the lipophilic toxin class. This minimal MS method development time could be attractive to laboratories that monitor these two classes of toxins, or potentially others, as MS settings are not tuned to individual analytes. Although the results described above were limited to a relatively small subset of biotoxins, extending the approach to other toxins or toxin analogues is simply a matter of expanding on the target list of analyte masses during data processing. Similarly, archived data can be subjected to retrospective analysis to determine the onset of new toxins or toxin analogues once discovered. However, accurate mass alone is typi- cally not sufficient for the structural elucidation of unknown compounds, and other techniques such as tandem MS or nuclear magnetic resonance spectroscopy are generally required. Acknowledgements The authors wish to thank Dr. Pearse McCarron for reviewing this manuscript and for providing extracts of the reference material containing lipophilic toxins. Loan of the Exactive ™ mass spectrometer from Thermo Fisher Scientific is also gratefully acknowl- edged. This is NRC publication number 51791. References 1. Moestrup Ø, Akselman R, Cronberg G, Elbraechter M, Fraga S, Halim Y, Hansen G, Hoppenrath M, Larsen J, Lundholm N, Nguyen LN, Zingone A (2009) IOC-UNESCO taxonomic reference list of harmful micro algae. Available online at http:// www.marinespecies.org/HAB. Accessed on 25 Oct 2010 2. Commission Regulation (EC) No. 2074/2005 of 5 December 2005 laying down implementing measures for certain products under Regulation (EC) No. 853/2004 of the European Parliament and of the Council and for the organisation of official controls under Regulation (EC) No. 854/2004 of the European Parliament and of the Council and Regulation (EC) No. 882/2004 of the European Parliament and of the Council, derogating from Regulation (EC) No. 852/2004 of the European Parliament and of the Council and amending Regulations (EC) No. 853/2004 and (EC) No. 854/2004 (2005). Official Journal of the European Union L 338/48 3. (2009) Scientific Opinion of the Panel on Contaminants in the Food Chain on a request from the European Commission on Marine Biotoxins in Shellfish — summary on regulated marine biotoxins. The EFSA Journal 1306:1 – 23 4. Hess P (2010) Requirements for screening and confirmatory methods for the detection and quantification of marine biotoxins in end-product and official control. Anal Bioanal Chem 397:1683 – 1694 5. McNabb P, Selwood AI, Holland PT (2005) Multiresidue method for determination of algal toxins in shellfish: single-laboratory validation and interlaboratory study. J AOAC Int 88:761 – 772 6. Fux E, McMillan D, Bire R, Hess P (2007) Development of an ultra-performance liquid chromatography-mass spectrometry method for the detection of lipophilic marine toxins. J Chromatogr A 1157:273 – 280 7. Gerssen A, Mulder PPJ, McElhinney MA, de Boer J (2009) Liquid chromatography-tandem mass spectrometry method for the detection of marine lipophilic toxins under alkaline conditions. J Chromatogr A 1216:1421 – 1430 8. Gerssen A, Van Olst EHW, Mulder PPJ, De Boer J (2010) In- house validation of a liquid chromatography tandem mass spectrometry method for the analysis of lipophilic marine toxins in shellfish using matrix-matched calibration. Anal Bioanal Chem 397:3079 – 3088 9. Quilliam MA, Xie M, Hardstaff WR (1995) Rapid extraction and cleanup for liquid chromatographic determination of domoic acid in unsalted seafood. J AOAC Int 78:543 – 546 10. Mafra LL Jr, Leger C, Bates SS, Quilliam MA (2009) Analysis of trace levels of domoic acid in seawater and plankton by liquid chromatography without derivatization, using UV or mass spectrometry detection. J Chromatogr A 1216:6003 – 6011 11. Lawrence JF, Niedzwiadek B (2001) Quantitative determination of paralytic shellfish poisoning toxins in shellfish by using prechromatographic oxidation and liquid chromatography with fluorescence detection. J AOAC Int 84:1099 – 1108 12. Lawrence JF, Niedzwiadek B, Menard C (2005) Quantitative determination of paralytic shellfish poisoning toxins in shellfish using prechromatographic oxidation and liquid chromatography with fluorescence detection: collaborative study. J AOAC Int 88:1714 – 1732 13. Rourke WA, Murphy CJ, Pitcher G, Van De Riet JM, Burns BG, Thomas KM, Quilliam MA (2008) Rapid postcolumn methodology for determination of paralytic shellfish toxins in shellfish tissue. J AOAC Int 91:589 – 597 14. Van De Riet JM, Gibbs RS, Chou FW, Muggah PM, Rourke WA, Burns G, Thomas K, Quilliam MA (2009) Liquid chromatographic post-column oxidation method for analysis of paralytic shellfish 584 P. Blay et al. toxins in mussels, clams, scallops, and oysters: single-laboratory validation. J AOAC Int 92:1690 – 1704 15. Dell'Aversano C, Hess P, Quilliam MA (2005) Hydrophilic interaction liquid chromatography-mass spectrometry for the analysis of paralytic shellfish poisoning (PSP) toxins. J Chromatogr A 1081:190 – 201 16. Turrell E, Stobo L, Lacaze JP, Piletsky S, Piletska E (2008) Optimization of hydrophilic interaction liquid chromatography/ mass spectrometry and development of solid-phase extraction for the determination of paralytic shellfish poisoning toxins. J AOAC Int 91:1372 – 1386 17. Diener M, Erler K, Christian B, Luckas B (2007) Application of a new zwitterionic hydrophilic interaction chromatography column for determination of paralytic shellfish poisoning toxins. J Separation Sci 30:1821 – 1826 18. Schurmann A, Dvorak V, Cruzer C, Butcher P, Kaufmann A (2009) False-positive liquid chromatography/tandem mass spec- trometric confirmation of sebuthylazine residues using the identification points system according to EU directive 2002/657/ EC due to a biogenic insecticide in tarragon. Rapid Comm Mass Spectrom 23:1196 – 1200 19. Grimalt S, Sancho JV, Pozo OJ, Hernandez F (2010) Quantifica- tion, confirmation and screening capability of UHPLC coupled to triple quadrupole and hybrid quadrupole time-of-flight mass spectrometry in pesticide residue analysis. J Mass Spectrom 45:421 – 436 20. Kellmann M, Muenster H, Zomer P, Mol H (2009) Full scan MS in comprehensive qualitative and quantitative residue analysis in food and feed matrices: how much resolving power is required? J Am Soc Mass Spectrom 20:1464 – 1476 21. Zhang NR, Yu S, Tiller P, Yeh S, Mahan E, Emary WB (2009) Quantitation of small molecules using high-resolution accurate mass spectrometers — a different approach for analysis of biological samples. Rapid Commun Mass Spectrom 23:1085 – 1094 22. Bateman KP, Kellmann M, Muenster H, Papp R, Taylor L (2009) Quantitative-qualitative data acquisition using a Benchtop Orbitrap mass spectrometer. J Am Soc Mass Spectrom 20:1441 – 1450 23. Krauss M, Singer H, Hollender J (2010) LC – high resolution MS in environmental analysis: from target screening to the identification of unknowns. Anal Bioanal Chem 397:943 – 951 24. Perez RA, Rehmann N, Crain S, LeBlanc P, Craft C, MacKinnon S, Reeves K, Burton IW, Walter JA, Hess P, Quilliam MA, Melanson JE (2010) The preparation of certified calibration solutions for azaspiracid-1,-2, and-3, potent marine biotoxins found in shellfish. Anal Bioanal Chem 398:2243 – 2252 25. McCarron P, Emteborg H, Nulty C, Rundberget T, Loader JI, Miles CO, Emons H, Quilliam MA, Hess P (2011) A mussel tissue certified reference material for multiple toxins: part 1: design and preparation. Anal Bioanal Chem doi:10.1007/s00216-011-4786-9 26. Brenton AG, Godfrey AR (2010) Accurate mass measurement: terminology and treatment of data. J Am Soc Mass Spectrom 21:1821 – 1835 Screening for multiple classes of marine biotoxins by LC-MS 585