Anal Bioanal Chem (2011) 401:2617–2630 DOI 10.1007/s00216-011-5345-0
ORIGINAL PAPER
Development and validation of a rapid method for microcystins in fish and comparing LC-MS/MS results with ELISA Lucía Geis-Asteggiante & Steven J. Lehotay & Laurie L. Fortis & George Paoli & Chandi Wijey & Horacio Heinzen
Received: 6 July 2011 / Revised: 10 August 2011 / Accepted: 16 August 2011 / Published online: 1 September 2011 # Springer-Verlag (outside the USA) 2011
Abstract Microcystins (MCs) are the most common cyanotoxins found worldwide in freshwater, brackish, and marine environments. The rapid and accurate analysis of MCs and nodularin (Nod-R) in fish tissue is important for determining occurrence, following trends, and monitoring exposure for risk assessment and other purposes. The aim of this study was to develop a streamlined and reliable sample preparation method for eight MCs (MC-RR, MCYR, MC-LR, MC-WR, MC-LA, MC-LY, MC-LW, and MCLF) and Nod-R in fish, and conduct a validation of the new method using liquid chromatography–tandem mass spectrometry (LC-MS/MS) for analysis and compare the results with a commercial enzyme-linked immunosorbent assay Mention of brand or firm name does not constitute an endorsement by the US Department of Agriculture above others of a similar nature not mentioned. Electronic supplementary material The online version of this article (doi:10.1007/s00216-011-5345-0) contains supplementary material, which is available to authorized users. L. Geis-Asteggiante : S. J. Lehotay (*) : L. L. Fortis : G. Paoli : C. Wijey Agricultural Research Service, Eastern Regional Research Center, US Department of Agriculture, 600 East Mermaid Lane, Wyndmoor, PA 19038, USA e-mail:
[email protected] L. Geis-Asteggiante : H. Heinzen Cátedra de Farmacognosia y Productos Naturales, DQO, Facultad de Química, UdelaR, General Flores 2124, Montevideo 12800, Uruguay Present Address: L. L. Fortis USDA National Institute of Food and Agriculture, 1400 Independence Ave., SW, Stop 2201, Washington, DC 20250, USA
(ELISA) kit. Different sample preparation methods were compared, and a simple extraction protocol with acidified acetonitrile/water (3:1) followed by hexane partitioning cleanup was found to be most effective. Thorough validation of the final method was conducted, and 90– 115% recoveries were achieved for all analytes except for MC-RR, which gave 130% average recovery (isotopically labeled internal standards were unavailable to correct for possible biases). The use of electrospray ionization in the negative mode gave few interferences and minimal matrix effects in the LC-MS/MS analysis overall. Precision was typically 10–20% RSD among multiple days in experiments, detection limits were 3, and there cannot be any observed carry-over [50]. Using these criteria, the limits of quantification (LOQ) and identification were ≤10 ng/g for all analytes except MC-RR, for which the LOQ was between 10–25 ng/g. Linearity Matrix-matched and solvent-only calibration curves of 5, 10, 25, 50, and 100 ng/g equivalent concentrations were prepared. As shown in Table 4, the average linear regression (R2) values from several calibrations on different days were >0.990 for seven of the nine analytes under study. In the case of matrixmatched calibration curves, R2 values were consistently ≥0.995 for all analytes except MC-RR and MC-LW, where MC-RR was the first to elute (and least stable) and therefore was possibly affected by many polar matrix co-extractives, whereas MC-LW gave a broad peak that was difficult to integrate (see Fig. 2). Also, no carry-over was observed in the method (LC-MS/MS is often plagued by carry-over due to the high sensitivity and selectivity of the approach). Recoveries Recovery experiments were conducted at four levels with five replicates each on five different days. As shown in Table 5, eight of the nine analytes presented an overall average recovery between 90–115% showing an acceptable degree of trueness with the method. Also, the linearity of recoveries was acceptable with respect of concentration for all analytes. Only MC-RR, first analyte to elute, showed an
New method for analysis of microcystins in fish Table 4 Average linear regression for solvent-only (MeCN/H2O (62:38) in 0.4% formic acid) and matrix-matched calibration curves and matrix-effect (ME) for each analyte in catfish tissue using the final method (n=5)
Values in italics refer to R2 20%
Analyte
2625
Average calibration curve, no matrix
No matrix, average R2
Average calibration curve, matrix
Matrix-match average R2
ME±SD (%)
MC-RR
y=119x−343
0.986
y=113x−342
0.988
−5±17
Nod-R MC-YR MC-LR MC-WR MC-LA MC-LY MC-LW MC-LF
y=311x−231 y=155x−242 y=158x−197 y=168x−250 y=851x−1,217 y=588x−377 y=278x−367 y=555x−1,377
0.986 0.992 0.998 0.997 0.995 0.996 0.992 0.993
y=308x−240 y=149x−46 y=161x−155 y=173x−303 y=1,137x−947 y=622x−594 y=240x−41 y=518x+592
0.996 0.997 0.996 0.995 0.998 0.998 0.989 0.996
−2±12 −2±14 2±14 6±19 34±23 4±11 −1±26 21±51
overall average recovery >120%, which can be explained by its lower stability in the standards (see Fig. 1) and/or coelution of polar matrix components (therefore a stronger matrix effect). Ideally, an isotopically labeled internal standard would be used to improve the accuracy of the results, but there were no isotopically labeled MCs
commercially available. In this respect, the method and its validation results were considered to be very good taking into account that external calibration was employed. We decided to forego the filtration step because it did not seem to add value to the method, and as a precaution, we cleaned the ion spray shield after each sequence.
Fig. 2 Chromatograms of the quantifier ion of MCs standards at a concentration of 25 ng/g in catfish extract. Note: The time scale is the same for each chromatogram (from 0 to 6.5 min)
2626
L. Geis-Asteggiante et al.
Intra- and inter-day precision The evaluation of precision was carried out by spiking five replicates of catfish samples at four different levels (10, 25, 50, and 100 ng/g) on five different days. As shown in Table 5, the repeatabilities (within a sequence) were nearly always |20|% and for MC-LR, 14 of 20 samples showed ME>|20|%, mostly basa and catfish. For MC-LW, ten of 20 samples showed ME120% and RSD >20%
Day 3, 50 ng/g %Rec (RSD) 111 (8) 85 (6) 100 (6) 92 (8) 100 (8) 83 (6) 77 (7) 96 (8) 79 (6)
Days 1 and 2, 100 ng/g %Rec (RSD) 125 119 94 106 102 114
(5) (4) (5) (9) (4) (7)
130±16
102 125 118 117 79 105 81 109 94 124 90 117
(2) (4) (4) (9) (5) (5) (4) (6) (9) (8) (13) (8)
107±15
94±10 97±17
115±13 90±11 91±16 107±20 104±26
New method for analysis of microcystins in fish
2627
Fig. 3 Matrix effect variation for 20 different fish samples at 25 ng/g using the final method
unknown concentrations to the analyst of MCs ≥10 ng/g (or 0 for unknown blanks). The results are shown in Table 6, in which the reported results were 70–100% of the spiked concentration in 16 of the 20 samples. This is acceptable performance, especially considering that an internal standard was not available (and the results were not corrected for the validated recovery values). Only four samples (2, 3, 15, and 19) showed combined concen-
trations 120% of the added amounts. Most importantly, no false-negatives or false-positives were observed in LC-MS/MS. ELISA detection The ELISA procedure was also tested for the orthogonally selective confirmation of free MCs and Nod-R from fish
Table 6 Results for the LC-MS/MS and ELISA analysis of blind catfish samples by the final method Analytes samples 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Overall%rec (RSD)
MC-RR (ng/g)
Nod-R (ng/g)
MC-YR (ng/g)
MC-LR (ng/g)
35 (40) 14 (10)
15 54 10 36
72 (100)
86 (100)
(40) (75) (75) (40)
36 29 68 63 28
(50) (40) (75) (75) (40)
73 (100) 81 (25)
27 29 68 39 69
(25) (40) (75) (40) (75)
24 (50) 38 (40)
87 (23)
39 28 77 36 55
MC-LY (ng/g)
MC-LW (ng/g)
MC-LF (ng/g)
66 (100) 20 (25)
28 (50) 44 (60) 32 (40)
108 (17)
MC-LA (ng/g)
(10) (40) (10) (40)
27 (25) 47 77 59 52
MC-WR (ng/g)
(50) (40) (75) (40) (75)
44 (50) 52 (60) 31 (40)
86 (100) 94 (24)
36 89 36 54
(40) (75) (40) (75)
36 (50)
20 (25) 17 30 53 23 47
(25) (40) (75) (40) (75)
28 (50)
31 55 44 22
(40) (75) (75) (40)
28 (50)
29 (25) 28 51 71 65 25
(25) (40) (75) (75) (40)
49 79 34 64
(40) (75) (40) (75)
32 (50) 28 (40)
87 (20)
65 (100) 69 (100) 65 (8)
67 (18)
88 (29)
103 (17)
ELISA results – + + + + + + + + + + + + – + + + – + + N/A
No false-positives or false-negatives were found in LC-MS/MS, and spiking levels are in parentheses. A presumptive positive occurred in ELISA for sample 6 as shown in Fig. 6 N/A not applicable
2628
L. Geis-Asteggiante et al.
Fig. 4 Results of cross-reactivity test of MC–LA, MC–LY, MC–WR, and MC–LR
tissue. The ELISA is sensitive to low pH (formic acid) and/or the presence of MeOH (method 3) or MeCN (method 4) in the extracts. Thus, the organic solvents were evaporated under vacuum, and the final extracts were neutralized with buffer prior to conducting the ELISA. Cross-reactivity The Abraxis kit (Warminster, PA) included performance data in which selectivity was determined and crossreactivity measured. The Abraxis data showed complete cross-reactivity with Adda group of seven MCs including MC-YR, MC-LF, MC-LR, MC-RR, and MC-LW, and Nod-R to date. We evaluated the cross-reactivity of MC-LA, MC-LY, and MC-WR to complete the study for our compounds of interest by preparing calibration curves (0.15, 0.4, 1.0, 2.0, and 5.0 ng/mL) and comparing the plots obtained with MC-LR, which is the analyte by which equivalent MC concentrations are determined in ELISA. The results are presented in Fig. 4.
Fig. 5 a Calibration curves consisting of seven concentration levels (1.35, 3.6, 9, 18, 45, 90, 180, and 405 ng/g MC-LR equivalent) in reagent-only (water/Tris–HCl (50:50), pH=7) and fish extract (matrixmatched), B0 was determined for each matrix. b Study of 20 fish tissue extracts for setting the LOD (n=2)
Detectability As stated in “Materials and methods” calibration curves with standards were prepared in fish tissue (matrixmatched) and reagent only (water/Tris–HCl (50:50), pH=7). Figure 5a shows that the calibration curves had two sections depending on the concentration level; for lower concentration (1.35–9 ng/g), the ME is ≈−63%, and for higher concentration (9–405 ng/g), the ME is ≈63%, so the matrix-matched calibration curve is necessary if fish tissue extracts are to be used for quantification with the ELISA kit. For determination of the limits of detection (LODs), 20 fish tissues were extracted and analyzed by both LC-MS/MS and ELISA. In the latter, the values of the average response from the blanks (B0) and SD obtained were B0 =1.08±0.13.absorbance units. Thus, the percentage of response at the LOD relative to B0 is calculated as %B/B0 LOD =[B0 –(3×SD)]/B0. This LOD threshold value of %B/B0 was found to be 63%, which corresponded to
Fig. 6 Results of ELISA analysis of 20 blind-fortified catfish samples using the final method. The percentage of sample response was relative to the B0 of the matrix. See Table 6 for blind spiking scheme
New method for analysis of microcystins in fish
0.28 ng/mL in fish extract for ELISA or 1.7 ng/g equivalent in fish. This is the calculated LOD for ELISA in fish tissue samples in terms of MC-LR equivalent concentration, which is ≈5-fold less than our lowest spiking level of 10 ng/g (GV for 50 g/day fish consumption). Figure 5b shows the distribution of the response (%B/B0) of the 20 blanks; five of the 20 were presumptive positives (%B/B0 10 ng/g.
Conclusions The method development and validation results demonstrate that the final method for the quantitative and qualitative identification of eight MCs and Nod-R by LCMS/MS meet the data quality needs for monitoring of the free analytes in fish tissue. The method is rapid, easy to perform, and reliable. The detection limits were sufficiently lower than the calculated 10 ng/g exposure guidance value. The quantitative results could be improved through the use of isotopically labeled internal standards and evaluation of certified reference materials of cyanotoxins in fish, but these isotope standards or CRMs for the analytes were not commercially available. The confirmatory ELISA method
2629
was demonstrated to correctly detect all positive samples (>10 ng/g) in blind analyses. However, ELISA was not rapid enough for routine screening and could not distinguish between MC variants for quantitative purposes. Acknowledgments We thank Jennifer Cassidy and Kathleen Rajkowski for providing the different fish samples, Alan Lightfield for his support in LC-MS/MS usage, and Fernando Rubio of Abraxis for consultation about microcystins and ELISA. This research was funded in part by the USDA Food Safety Inspection Service ARS agreement number 60-1935-9-031 and the US–Israel Binational Agricultural Research and Development Fund number US-4273-09.
References 1. Sivonen K, Jones G (1999) Cyanobacterial toxins. In: Chorus I, Bartram J (eds) Toxic cyanobacteria in water: a guide to their public health consequences, monitoring and management. E & FN Spon, London 2. Carmichel WW (1994) The toxins of cyanobacteria. Sci Am 270:78–86 3. Meriluoto J, Codd GA (2005) Toxic: cyanobacterial monitoring and analysis. Åbo Akademi University Press, Finland, p 15 4. Zhang L, Ping X, Yang Z (2004) Talanta 62:193–200 5. Spoof L, Neffling M-R, Meriluoto J (2009) J Chromatogr B 877:3822–3830 6. Chorus I, Bartram J (1999) Toxic cyanobacteria in water—a guide to their public health consequences, monitoring and management. E & FN Spon. Published on behalf of World Health Organization, London 7. Codd GA, Lindsay J, Young FM, Morrison LF, Metcalf JS (2005) From mass mortalities to management measures. In: Huisman J, Matthijs HCP, Visser PM (eds) Harmful cyanobacteria. Springer, Aquatic Ecology Series, pp 1–23 8. Smith JL, Boyer GL (2009) Toxicon 53:238–245 9. Lance E, Neffling M-N, Gérard C, Meriluoto J, Bormans M (2010) Environ Pollut 158:674–680 10. WHO (1998) Guidelines for drinking water quality. World Health Organization, Geneva 11. Amorim A, Vasconcelos V (1999) Toxicon 37:1041–1052 12. Tencalla F, Dietrich D (1997) Toxicon 35:583–595 13. Thostrup L, Christoffersen K (1999) Arch Hydrobiol 145(4):447–467 14. Vasconcelos VM (1995) Aquat Toxicol 32:227–237 15. Deblois CP, Giani A, Bird DF (2011) Aquat Toxicol 103:63–70 16. Deblois CP, Aranda-Rodriguez R, Giani A, Bird DF (2008) Toxicon 51:435–448 17. Magalhães VF, Moraes SR, Azevedo S (2001) Toxicon 39:1077–1085 18. Yokoyama A, Park HD (2002) Environ Toxicol 17:424–433 19. Zurawell RW, Kotak BG, Prepas EE (1999) Freshw Biol 42:707–718 20. Chen J, Xie P (2005) Toxicon 45:615–625 21. MacKintosh C, Beattie KA, Klumpp S, Cohen P, Codd GA (1990) FEBS Lett 264(2):187–192 22. Zhang D, Ping X, Chen J (2010) Bull Environ Contam Toxicol 84:202–207 23. Soares RM, Magalhaes VF, Azevedo SMFO (2004) Aquat Toxicol 70:1–10 24. Karlsson KM, Spoof L, Meriluoto JA (2005) Environ Toxicol 20 (3):381–389 25. Kohoutek J, Adamovsky M, Oravec M, Simek Z, Palikova M, Kopp R, Blaha L (2010) Anal Bioanal Chem 398:1231–1237 26. Moreno IM (2005) J Chromatogr A 1080:199–203 27. Mohamed ZA, Carmichael WW, Hussein AA (2003) Environ Toxicol 18:137–141
2630 28. Xie L, Xie P, Ozawa K, Li L, Honma T, Yokoyama A, Park HD (2004) Environ Pollut 127:431–439 29. Xie L, Park H-D (2007) Aquaculture 271:530–536 30. Xie LQ, Xie P, Guo LG, Li L, Miyabara Y, Park HD (2005) Environ Toxicol 20:293–300 31. Adamovsky O, Kopp R, Hilscherova K, Babica P, Palikova M, Paskova V, Navratil S, Marsalek B, Blaha L (2007) Environ Toxicol Chem 26:2687–2693 32. Li XY, Chung IK, Kim JI, Lee JA, Sivonen K, Lahti K (2004) Toxicon 44:821–827 33. Rapala J, Erkomaa K, Kukkonen (2002) J Anal Chim Acta 466:213–231 34. Gago-Martinez A, Leao JM, Pineiro N, Carballal E, Vaquero E, Nogueiras M, Rodriguez-Vazquez JA (2003) J Environ Anal Chem 83:443–456 35. Gago-Martinez A, Pineiro N, Aguete EC, Vaquero E, Nogueiras M, Leao JM, Rodriguez-Vazquez JA (2003) J Chromatogr A 992:159–168 36. Zimba PV, Camus A, Allen EH, Burkholder JM (2006) Aquaculture 261:1048–1055 37. Neffling M-R, Spoof L, Quilliam M, Meriluoto J (2010) J Chromatogr B 878:2433–2441 38. Anastassiades M, Lehotay SJ, Stajnbaher D, Schenk FJ (2003) J AOAC Int 86:412–431 39. Schneider MJ, Mastovska K, Lehotay SJ, Lightfield AR, Kinsella B, Shultz CE (2009) Anal Chim Acta 637:290–297
L. Geis-Asteggiante et al. 40. Mol H, Plaza-Bolaños P, Zomer P, de Rijk TC, Stolker A, Mulder P (2008) Anal Chem 80:9450–9459 41. Bruker Daltonics (2001) Identification and quantification of microcystins in water samples or extracts from cyanobacterias using LC-ESI Ion Trap MS. Esquire series. Application Note #LCMS-17. 42. Gjølme N, Utkilen H (1996) Phycologia 35:80–82 43. Hyenstrand P, Metcalf JS, Beattie KA, Codd GA (2001) Wat Res 35:3508–3511 44. Kinsella B, Lehotay SJ, Mastovska K, Lightfield AR, Furey A, Danaher M (2009) Anal Chim Acta 637:196–207 45. Sano T, Takagi H, Nagano K, Nishikawa M, Kaya K (2011) Anal Bioanal Chem 399:2511–2516 46. Jacobs HL, Kahn HD, Stralka KA, Phan DB (1998) Risk Anal 18 (3):283–291 47. National Marine Fisheries Service, NOAA (2005) Fisheries of the United States–2005 http://www.st.nmfs.noaa.gov/st1/fus/fus05/ 08_perita2005.pdf Accessed on: 10 May 2011. 48. National Marine Fisheries Service, NOAA (2009). Fisheries of the United States–2009 http://www.st.nmfs.noaa.gov/st1/fus/fus09/ 08_perita2009.pdf Accessed on: 10 May 2011. 49. Dietrich D, Hoeger S (2005) Toxicol Appl Pharmacol 203:273–289 50. Lehotay SJ, Mastovska K, Amirav A, Fialkov AB, Alon T, de Martos PA, Kok A, Fernandez-Alba AR (2008) Trends Anal Chem 27:1070–1090 51. Chen J, Xie P (2005) Environ Toxicol 20:572–584
