LC/MS/MS Analysis of Cylindrospermopsin and Anatoxin-A in Drinking Water Using US EPA Method 545

LC/MS/MS Analysis of Cylindrospermopsin and Anatoxin-a in Drinking Water Using US EPA Method 545

Application Note

Environmental

Authors Abstract

Ralph Hindle and Don Noot Analysis of drinking water samples according to EPA Method 545 on the Agilent Vogon Laboratory Services 1290 Infinity LC System coupled to an Agilent 6460 Triple Quadrupole LC/MS Cochrane, AB System results in precision and accuracy values that are well within the method Canada requirements and Lowest Calculated Minimum Reporting Limits (LCMRLs) that meet those cited in the method. Introduction Experimental Cyanotoxins are produced by cyanobacteria (blue-green Reagents and materials algae), which can form expansive blooms in lakes and oceans Anatoxin-a, cylindrospermopsin, and the internal standards under the right conditions. These toxins can reach concentra - were supplied by the EPA in Cincinnati, OH. An Agilent Polaris tions that can poison and even kill animals and humans. Two C18-Ether, 3 × 150 mm, 3 µm column (p/n A2021150X030) of the cyantoxins produced by blue-green algae are alkaloids was used for the HPLC separations. known as cylindrospermopsin and anatoxin-a.

Cylindrospermopsin is toxic to liver and kidney tissue, and Internal standard CASRN a Catalog no. toxic blooms producing cylindrospermopsin are usually found Uracil-d , neat material 24897-55-0 C/D/N Isotopes D-5135 in tropical, subtropical, and arid zone waters. 4 L-phenylalanine-d 5, neat 56253-90-8 Cambridge Isotopes Labs Anatoxin-a is a neurotoxin that produces very rapid toxic material DLM -1258-1 effects, including loss of coordination, twitching, convulsions, and rapid death by respiratory paralysis. Because of the dan - Instruments gers they present to recreational and drinking waters, these The system was set up using Agilent 1200 Infinity Series LC cyanotoxins are the subject of monitoring and regulation modules coupled to an Agilent 6460A Triple Quadrupole efforts in several countries, including the US. LC/MS System, using electrospray positive ionization with The Safe Drinking Water Act (SDWA) requires the US EPA to Agilent Jet Stream technology. The LC system used the publish a list of unregulated contaminants that are known or Agilent 1290 Infinity Binary Pump (G4220A), Agilent 1290 expected to occur in public water systems in the U.S., known Infinity Autosampler (G4226A), and Agilent 1290 Infinity as the Contaminant Candidate List (CCL). Cyanotoxins appear Column Compartment (G1316C). The LC/MS run conditions on the three drinking water CCLs, and the EPA has focused on are shown in Table 1. three cyanotoxins for further research activities: microcystins, anatoxin-a, and cylindrospermopsin [1]. The EPA uses the Table 1. HPLC and MS Conditions Unregulated Contaminant Monitoring Rule (UCMR) program to HPLC collect data on these cyanotoxins, and EPA Method 545 has Analytical column Agilent Polaris C18-Ether, 3 × 150 mm, 3 µm been developed to measure anatoxin-a and cylindrospermopsin column (p/n A2021150X030) levels in drinking water for that purpose [2] . Column temperature 30 °C Injection volume 20 µL This application note describes the use of the Agilent 1200 Mobile phase A) 0.15 % acetic acid in water Infinity Series LC and the Agilent 6460A Triple Quadrupole B) 0.15 % acetic acid in methanol LC/MS System with Agilent Jet Stream technology to meet Flow rate 0.3 mL/min the stringent quality control requirements of Method 545. This analysis platform provided Lowest Calculated Minimum Gradient Time (min) Mobile phase (% B) 03 Reporting Limits (LCMRLs) that met or exceeded EPA require - 1 20 ments. In addition, recoveries and precision were well within 5 50 the requirements of the method. Only 20-µL injections were 6 60 required to satisfy the sensitivity requirements of the method, 8 60 rather than the 50-µL injections recommended in the EPA Post time 6 minutes method, enabling a reduction in maintenance requirements Run time 15 minutes, injection to injection for the MS source. MS Acquisition parameters ESI mode, positive ionization, MRM Sheath gas temperature 350 °C Sheath gas flow rate 12 L/min Drying gas temperature 350 °C Drying gas flow rate 9 L/min Nebulizer pressure 35 psig Nozzle voltage 1,000 V Vcap 4,000 V positive

2 Sample preparation Results and Discussion Sample preparation is very straightforward, involving collec - tion of 10 mL samples in bottles containing preservatives, EPA Method 545 requirements adding internal standards, and filtering 1 mL of each sample EPA Method 545 calls for seven Laboratory Field Blanks into a vial using 0.2-µm PVDF filters and disposable syringes. (LFBs) and seven tap water samples, spiked near mid-range of Ascorbic acid (0.1 g/L as reducing agent for chlorine) and the calibration curve for each analyte. The resulting precision sodium bisulfate (1 g/L as microbial inhibitor) were used as expressed as percent relative standard deviations (RSDs) preservatives per EPA Method 545. must be ~ 20 % and accuracy expressed as recovery must be between 70 and 130 %. A Laboratory Reagent Blank (LRB) Analysis parameters must be analyzed directly following analysis of the highest The multiple reaction monitoring (MRM) transitions used calibration standard concentration for each analyte. Resulting for cylindrospermopsin, anatoxin-a, and the two internal peaks in the LRB must have areas < 1/3 of the lowest calibra - standards are shown in Table 2. tor concentration. Finally, the goal of this implementation of Method 545 would be to generate LCMRLs equal to or lower than those reported in the EPA method.

Table 2. Multiple Reaction Monitoring (MRM) Analysis Parameters

Precursor Product Fragmentor Collision Dwell Compound ion ion voltage energy (V) (msec) Polarity Type

Anatoxin-a 166.1 149.1* 95 12 150 Positive Target 131.1 95 12 150 Positive Target Cylindrospermopsin 416.1 194.1* 125 20 150 Positive Target 336.1 125 32 150 Positive Target Uracil-d4 † 115.1 98.0* 95 12 150 Positive ISTD Phenylalanine-d5 † 171.1 125.0* 95 8 150 Positive ISTD

† Internal standard *Transition used for quantitation

3 Method performance The extracted ion chromatogram (EIC) in Figure 1 illustrates complete resolution of the two cyanotoxins as well as the internal standards. Analysis of the LFB directly following the analysis of the highest calibration standard concentration (0.30 µg/L) revealed no carryover for either anatoxin-a or cylindrospermopsin (Figure 2).

×10 3

2.2 L-phenylalanine-d 5 7.47 2.0 1.8 1.6 1.4 s t

n 1.2 u o

C 1.0 Uracil-d 0.8 4 4.41 0.6 CYN ANA 6.41 0.4 6.12 0.2 0 00.51.01.52.02.53.03.54.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 Acquisition time (min) Figure 1. Extracted ion chromatogram (EIC) illustrating complete resolution of anatoxin-a (ANA) and cylindrospermopsin (CYN) as well as the two internal standards.

×10 2 ×10 2 Cal 1 at 0.02 µg/L Cal 1 at 0.02 µg/L 4 4 s s t 3 t 3 n n u u

o 2 o 2 C 6.11 C 6.41 1 178 1 125 0 0

×10 2 ×10 2 6.41 2,117 Cal 9 at 0.30 µg/L 6.12 Cal 9 at 0.30 µg/L 4 1,563 4 s s 3 3 t t n n u u o o 2 2 C C 1 1 0 0

×10 2 ×10 2 4 Reagent water blank 4 Reagent water blank following Cal 9 injection following Cal 9 injection s 3 s 3 t t n n u u o o 2 2

C 3.27 C 6.63 1 122 1 15 0 0 12345 6 7 12345 6 7 Acquisition time (min) Acquisition time (min) Figure 2. EICs of reagent water blanks for anatoxin–a and cylindrospermopsin run immediately following analysis of the highest calibrator concentration (Cal 9, 0.30 µg/L). No carryover was detected for either cyanotoxin.

4 The initial calibration curves were run from 0.030 to As required by Method 545, seven replicates of LFBs and 0.300 µg/L, but these proved to be too high for the subse - chlorinated tap water were spiked at 0.100 µg/L and preserv - quent calculation of the LCMRLs. Therefore, a second set of atives were added. Recovery (accuracy) and precision were calibration curves was run, from 0.005 to 0.050 µg/L. Both then determined for each set of samples (Table 3). Accuracy ranges yielded highly linear calibration, with all calibration and precision were well under the ± 30 % recovery and 20 % coefficients (R 2) ¡ 0.996 (Figure 3). RSD limits set by Method 545 for these values, in both reagent and tap water.

Anatoxin-a Cylindrospermopsin −1 −1 ×10 y = 1.342038*x − 0.001401 ×10 y = 0.587570*x − 0.001350 4.0 R2 = 0.9988 1.6 R2 = 0.9967

3.6 L s s

/ 1.4 e 3.2 e g s s µ n n 2.8 1.2 o o 0 p p 0

s 2.4 s 3 1.0 e e . r r

2.0 e e

o 0.8 v v t i i

t 1.6 t 0 a a

l l 0.6 3

e 1.2 e 0 . R R

0 0.8 0.4 0.4 0.2 0 0 0.01 0.03 0.05 0.07 0.09 0.11 0.13 0.15 0.17 0.19 0.21 0.23 0.25 0.27 0.29 0.31 0.01 0.03 0.05 0.07 0.09 0.11 0.13 0.15 0.17 0.19 0.21 0.23 0.25 0.27 0.29 0.31 Concentration (µg/L) Concentration (µg/L)

2 2 ×10 y = 10,485.406711*x − 2.469375 ×10 y = 9159.552518*x − 12.152276 5 R2 = 0.9984 5 R2 = 0.9960

L s s / e e g 4 s s 4 µ

n n 0 o o 5 p p s s 0 6 3

e e . r r

e e o v v t i 2 i 2

t t 5 a a l l 0 e e 0

R R

. 1 1 0 0 0 0 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 0 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 0.050 Concentration (µg/L) Concentration (µg/L) Figure 3. Linear calibration curves for two concentration ranges of the two cyanotoxins. The curve from 0.030 to 0.300 µg/L gave R 2 values of 0.9988 and 0.9967 for anatoxin-a (ANA) and cylindrospermopsin (CYN), respectively. The second calibration range (0.005–0.050 µg/L) gave R 2 values of 0.9984 and 0.9960 for ANA and CYN, respectively.

Table 3. Accuracy and Precision for Anatoxin-a (ANA) and Cylindrospermopsin (CYN) Analyses at 0.1 µg/L

Reagent water Tap water Replicate no. ANA* CYN* ANA* CYN* 1 0.115 0.108 0.089 0.104 2 0.100 0.097 0.083 0.106 3 0.096 0.097 0.086 0.107 4 0.092 0.106 0.090 0.110 5 0.087 0.100 0.097 0.105 6 0.096 0.096 0.086 0.115 7 0.087 0.102 0.099 0.094 Accuracy 96.1% 100.9% 90.0% 105.9% RSD† 10.0% 4.7% 6.6% 6.1%

* Recovery values, µg/L † Precision expressed as percent relative standard deviation (RSD)

5 LCMRL calculations calculations were done with a sample set ranging from Method 545 requires the calculation of the LCMRL, which is 0.030–0.250 µg/L, but this range was too high to determine accomplished by entering values in an EPA-supplied LCMRL an LCMRL for anatoxin-a. Calculator [3]. The LCMRL is the lowest spiking concentration An additional two sets of replicates were run for both cyan - at which recovery of between 50 and 150 % is expected 99 % otoxins, at 0.010 and 0.020 µg/L, using a calibration curve con - of the time by a single analyst. It requires a minimum of four structed from 0.005–0.050 µg/L. All nine fortification levels replicates at each of seven fortification levels, plus four were then input into the calculator for LCMRL determinations, Laboratory Reagent Blanks (LRBs). The LCMRL Calculator for both compounds. Figure 4 shows the LCMRL results, which constructs mean and variance models of measurement as a were equal to or lower (0.018 µg/L for ANA and 0.038 µg/L for function of spiking level, taking into account both precision CYN) than those cited in Method 545 (0.018 µg/L for ANA and and accuracy. The first LCMRL 0.063 µg/L for CYN).

Anatoxin-a

QC interval coverage plot LCMRL plot 1.00 Data e

g 0.35 LCMRL = 0.018 µg/L a r

e LCMRL = 0.018 µg/L Y = X ) v

0.98 L o Qual Lim: 50 −150 % / c 0.30 Regression

l g

a Coverage Prob: 0.99 µ 50 −150 % Recovery ( v

r n

e Lower/upper prediction limits t o 0.25 0.96 i n t i

a r C t Q n

f e 0.20 o c

n y 0.94 t o i l c i

d b 0.15 a e r b u o s r 0.92 a P 0.10 e M 0 0.05 0.1 0.15 0.2 0.25 0.05 Spike Concentration (µg/L) 0 0 0.05 0.1 0.15 0.2 0.25 Spike concentration (µg/L)

Cylindrospermopsin

QC interval coverage plot LCMRL plot 1.00 0.40

e Data

g LCMRL = 0.038 µg/L

a LCMRL = 0.038 µg/L r 0.95 Qual Lim: 50 −150 % e 0.35 ) v Y = X L

o Coverage Prob: 0.99 / c

Regression l g a 0.90 µ 0.30 ( v 50 −150 % Recovery

r n e t o Lower/upper prediction limits i n t i

0.85

a 0.25 r C t Q n

f e o c

0.80 0.20 n y t o i l c i

b d a 0.75 e 0.15 r b u o s r a P

0.70 e 0.10 M 0 0.05 0.1 0.15 0.2 0.25 0.05 Spike concentration (µg/L) 0 0 0.05 0.1 0.15 0.2 0.25 Spike concentration (µg/L) Figure 4. LCMRL results for anatoxin–a and cylindrospermopsin (0.018 and 0.038 µg/L respectively), which are equal to or lower than those reported in the EPA Method 545 (0.018 and 0.063 µg/L respectively).

6 Conclusions Acknowledgements

Using an Agilent 1290 Infinity LC coupled to the Agilent 6460 The authors would like to thank Joe Hedrick and Jason Link Triple Quadrupole LC/MS System with Agilent Jet Stream of Agilent Technologies for advice and the gift of the Agilent technology for the analysis of algal toxins in drinking water Polaris C18-Ether column. can enable laboratories to meet the stringent QC require - ments of EPA Method 545. Recoveries were between 90 and References 106 %, well within the ± 30 % required by EPA. Precision ranged from 4.7–10 % in reagent water and chlorinated tap 1. “Cyanobacterial Harmful Algal Blooms (CyanoHABs)”, water, also well below the EPA limit of 20 %. In addition, the United States Environmental Protection Agency, relative retention of Uracil-d4 on the Agilent Polaris C18-Ether http://www2.epa.gov/nutrient-policy-data/cyanobacter - column is greater than the retention on the column recom - ial-harmful-algal-blooms-cyanohabs, accessed April 23, mended in Method 545. The use of this column can help 2014. reduce potential ion suppression effects from preservatives and other non-retained compounds that may be present in 2. EPA Method 545 – Determination of Cylindrospermopsin drinking waters. LCMRLs that were equal to or below those and Anatoxin-a in Drinking Water by Liquid found in the EPA method were obtained with only 20-µL injec - Chromatography Electrospray Ionization Tandem Mass tions, instead of the 50-µL injections recommended in Spectrometry (LC/ESI-MS/MS); August 2013. Method 545. Injecting smaller sample amounts can help keep 3. S.D. Winslow, et al . “Statistical procedures for determina - the MS source cleaner during routine use, reducing tion and verification of minimum reporting levels for drink - maintenance requirements. ing water methods” Environ. Sci. Technol . 40 , 281-288, 2006.

For More Information

These data represent typical results. For more information on our products and services, visit our Web site at www.agilent.com/chem.

7 www.agilent.com/chem

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material.

Information, descriptions, and specifications in this publication are subject to change without notice.