ARTICLE pubs.acs.org/ac
Detailed LC-MS/MS Analysis of Ciguatoxins Revealing Distinct Regional and Species Characteristics in Fish and Causative Alga from the Pacific Kentaro Yogi,†,‡ Naomasa Oshiro,‡ Yasuo Inafuku,‡ Masahiro Hirama,§ and Takeshi Yasumoto*,|| †
Okinawa Science and Technology Promotion Center, 12-75 Suzaki, Uruma, Okinawa 904-2234, Japan Okinawa Prefectural Institute of Health and Environment, 2085 Aza-Ozato, Ozato, Nanjo, Okinawa 901-1202, Japan § Department of Chemistry, Graduate School of Science, Tohoku University, Sendai 980-8578, Japan Japan Food Research Laboratories, 6-11-10 Nagayama, Tama, Tokyo 206-0025, Japan
)
‡
bS Supporting Information ABSTRACT: Toxin profiles of representative ciguatera species caught at different locations of Japan were investigated in fish flesh by high-performance liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis. Identification and quantification of 16 toxins were facilitated by the use of 14 reference toxins prepared by either synthesis or isolation from natural sources and the previous LC-MS data thereof. Sodium adduct ions [M + Na]+ were used as parent and product ions. Distinct regional differences were unveiled: ciguatoxin-1B type toxins were found in snappers and groupers from Okinawa, ciguatoxin-3C type toxins were found in a spotted knifejaw, Oplegnathus punctatus, from Miyazaki located 730 km north of Okinawa, and both types of toxins were found in a red snapper, Lutjanus bohar, from Minamitorishima (Marcus) Island. Twelve toxins were identified in a dinoflagellate, Gambierdiscus toxicus, collected as the primary toxin source in French Polynesia. Occurrence of M-seco-toxins in fish and oxidized toxins in the dinoflagellate was confirmed for the first time. The present LC-MS/MS method is rapid, specific, and accurate. It not only outperforms the currently employed mouse bioassays but also enables the study of the toxin dynamics during the food chain transmission.
iguatera fish poisoning (CFP) refers to poisoning caused by ingestion of ciguatoxin-contaminated fish inhabiting tropical and subtropical regions. CFP is the largest source of acute food poisoning due to natural toxins, affecting 20 00060 000 patients annually. The causative toxins, collectively named ciguatoxins (CTXs), which are produced by an epiphytic dinoflagellate Gambierdiscus toxicus, enter the food chain via herbivorous fish and accumulate in the tissues of several hundred types of fish species.1,2 Thus, depending on the growth of the local G. toxicus population, the toxicity of individual fish may vary markedly in a regional, seasonal, and annual manner. This sporadic and unpredictable occurrence of toxicity in fish has posed a great difficulty in the monitoring of fish safety. Although CFP was a local phenomenon in the past, it has become a global problem due to expanding trade and tourism. There is a need for compromise between the increasing public concern for food safety on one hand and the urges to fully exploit coastal fish resources on the other hand. The safety level of CTXs in fish flesh for health protection is extremely low, as proposed in the United States at an action level of 0.01 parts per billion CTX1B equivalents.3 In vitro assays, especially the neuroblastoma (Neuro-2a) cell assay,4 have been vigorously pursued to replace the mouse bioassay (MBA) for monitoring fish toxicity in the U. S. and other countries. The compatibility of these methods with
C
r 2011 American Chemical Society
liquid chromatography tandem mass spectrometry (LC-MS/MS) methods is discussed in detail in a recent review article.5 However, MBA is still used widely. The method requires laborious steps for sample preparation and requires a long process time for obtaining results. More serious drawbacks of this method are its low sensitivity and specificity and the requirement of many mice for analysis. Development of proper methods for toxin detection and determination has been hampered by the extreme difficulty of isolating toxins as exemplified by the structural study of ciguatoxin, in which 0.35 mg of pure toxin was produced from 124 kg of the viscera of four tons of moray eels, which were collected and tested toxic over years from a wide area of French Polynesia.6,7 The chemical structures of a few major toxins were determined on the basis of NMR data, and those of minor toxins were evaluated by fast-atom bombardment tandem mass spectrometry using extremely small quantities.8,9 Thus, little sample remained for use as analytical standards or for measurements of biological activities. CTXs are composed of contiguous cyclic ether rings aligned in a ladder-like fashion. The skeletal structures of toxins isolated from Pacific samples are separable into two groups, the Received: March 30, 2011 Accepted: October 19, 2011 Published: October 19, 2011 8886
dx.doi.org/10.1021/ac200799j | Anal. Chem. 2011, 83, 8886–8891
Analytical Chemistry
ARTICLE
Chart 1. Structures of Ciguatoxins and Related Polyethers Used for LC-MS/MS Analysis
CTX1B type and the CTX3C type (Chart 1). The toxins undergo metabolic modification in fish and produce a large number of congeners, of which 23 structures have so far been determined. Hirama and co-workers recently succeeded in the chemical synthesis of three major toxins, CTX3C, 51-hydroxyCTX3C, and CTX1B,1012 enabling the quantification of minute samples from natural sources by means of comparison with these chemically well-defined references. These reference samples made it possible to assign complex and often minute signals with accuracy. The regional characteristics of the toxin profiles revealed in this study provide important information to other assay methods such as ELISA, another sensitive method, which was recently developed.13,14 A toxic polyether, gambierol,15 and potent antifungal polyethers, gambieric acids,16,17 are also included in this study because they are also products of some strains of G. toxicus and are detectable together with CTXs when present.
’ EXPERIMENTAL SECTION Reagents and Standards. All organic solvents, distilled water, formic acid, and ammonium formate used were of HPLC or analytical grade (Wako Pure Chemical Industries, Ltd., Tokyo). Out of the standard toxins shown in Chart 1, CTX1B (1), CTX3C (7), and 51-hydroxyCTX3C (9) were synthesized at Tohoku University.1012 CTX4A (4), CTX4B (5), 52-epi-54deoxyCTX1B (2), 54-deoxyCTX1B (3), 49-epiCTX3C (8), 2-hydroxyCTX3C (10), 2,3-dihydroxyCTX3C (11), M-secoCTX4A/B (6), M-seco-CTX3C (12), M-seco-CTX3C methyl acetal (13), gambierol (14), gambieric acid A (15), and gambieric acid B (16) were from natural sources and identified by spectral analysis.69,1517 The chromatographic and spectral
purities of the standard toxins are in the references cited. The acronyms for the toxins are the same as those in the literature.8,9 The M-seco-methyl acetals could be artifacts because methanol was used for extraction and storage. Specimens. Remnants of six fish implicated in CFP incidents were obtained from Okinawa Prefecture, where large carnivores are frequently toxic.18 Two additional specimens were recovered from meals at the time of the poisoning incidents in Miyazaki and Ibaraki Prefectures. The causative fish in Ibaraki was caught at Minamitorishima (Marcus) Island located 1800 km southeast of Tokyo (24°160 5900 N, 153°590 1100 E). The other fish were from the local fishing grounds. Fish species and epidemiological data are given in Table S1, Supporting Information. Only the flesh was used for analysis. The sites of catch, toxicity, and biological data of six specimens of L. bohar used to evaluate intraspecies variations are presented in Table S2, Supporting Information. The dinoflagellate, Gambierdiscus toxicus (RGI-1 strain), was collected at Rangiroa Atoll of French Polynesia and cultured at Tohoku University.19 Extraction. Fish samples were extracted according to the standard MBA method used in Japan for ciguatoxin detection.20 A flesh sample (240 g) was extracted twice by homogenizing in 700 mL of acetone. The combined filtrate was concentrated to an aqueous residue and extracted twice with 200 mL of diethyl ether. The combined ether extract was evaporated, and the residue was dissolved in 50 mL of MeOHwater (9:1). The aqueous methanol solution was defatted with hexane (100 mL) and evaporated. The evaporation of solvents was performed with a rotary evaporator under reduced pressure, and the volume of the solvents was adjusted on the basis of the sample weight to fit the proportions described above. The toxic crude extract was used for MBA and LC-MS/MS analysis. The cultured cells of G. toxicus were extracted with MeOH, and the residue was fractionated 8887
dx.doi.org/10.1021/ac200799j |Anal. Chem. 2011, 83, 8886–8891
Analytical Chemistry
ARTICLE
Figure 1. Chromatogram and retention time (Rt) data for 16 standard toxins. The solutions used for injection were 1 ng/mL, unless otherwise indicated with an asterisk (/).
Figure 2. Toxin profiles of fish flesh implicated in CFP in Okinawa. The sample solutions used for injection were equivalent to 5 g flesh/mL (a, b, d, e, f), except for 1 g/mL of V. louti (c).
on an ODS-Q3 (Wako, Osaka, Japan) column with aqueous methanol of varying ratios. Clean-up of Crude Extracts. The crude extracts were purified on solid phase extraction cartridges (GL Sciences INC, Tokyo) prior to LC-MS/MS analysis. An extract equivalent to 5 or 1 g of flesh was dissolved in 2 mL of EtOAcMeOH (9:1). The solution was passed through a Florisil cartridge (InertSep FLPR, 500 mg). The cartridge was eluted with 2 mL of the same solvent, and the combined eluate (4 mL) was dried under nitrogen stream at 40 °C. The residue was dissolved in 3 mL of MeCN and
applied to a PSA cartridge (InertSep PSA, 200 mg). The cartridge was washed with 3 mL of MeCN, and the target toxins were eluted with 3 mL of MeOH. The eluate was dried and dissolved in 1 mL or 200 μL of MeOH for LC-MS/MS analysis. The MeCN eluate was also required for analysis when less polar congeners were present. The ODS cartridges (Inert Sep C18, 200 mg) could also be used instead of PSA as follows: a toxin fraction from the Florisil cartridge was redissolved in 2 mL of MeOHwater (95:5), passed through an ODS cartridge equilibrated with the same solvent, and eluted with another 2 mL of 8888
dx.doi.org/10.1021/ac200799j |Anal. Chem. 2011, 83, 8886–8891
Analytical Chemistry
ARTICLE
Figure 5. Toxin profile in a red snapper from Minamitorishima. *2,3,51TrihydroxyCTX3C was deduced from [M + Na]+ (m/z 1095) and Rt.8
Figure 3. Correlation between LC-MS/MS and mouse bioassay results: fish actually implicated in poisoning incidents (a) and L. bohar tested by MBA in a monitoring study (b). The i.p. lethal dose in mice is 7 ng for CTX1B and 14 ng for 52-epi-54-deoxyCTX1B.69 The lethal doses for 54-deoxyCTX1B and M-seco-CTX4A/B are unknown and were not used for calculation.
Figure 4. Toxin profile in a spotted knifejaw from Miyazaki.
the same solvent. The combined eluate was dried and dissolved in MeOH for LC-MS/MS analysis. LC-MS/MS Analysis. The LC-MS/MS analysis was carried out on an Agilent 1200 series LC system coupled to an Agilent 6460 Triple Quadrupole LC/MS (Agilent Technologies, CA) equipped with an Agilent Jet Stream electrospray ionization source. LC separation was achieved using a Zorbax Eclipse Plus C18 column (2.1 50 mm, 1.8 μm, Agilent Technologies) at 40 °C using a linear gradient of mobile phases: 5 mM ammonium formate and 0.1% formic acid in water (A) and MeOH (B). The concentration of solution B was increased from 78% to 88% in 10 min and held for 4 min. The flow rate was 0.4 mL/min, and the injection volume was 5 μL.
The mass spectrometer was operated in a positive mode for the monitoring of sodium adduct ions ([M + Na]+). The collision energy was 40 eV for CTXs, 24 eV for gambierol, and 65 eV for gambieric acids. The [M + Na]+ ions were used as precursor ions and product ions, except for [M + 2Na H]+ of gambieric acids. The instrumental parameters were set as follows: drying gas, 10 L/ min of N2 at 300 °C; nebulizer gas, N2 at 50 psi; sheath gas, 11 L/ min of N2 at 400 °C; capillary voltage, 5000 V; nozzle voltage, 1000 V; fragmentor voltage, 350 V. Toxins were monitored as follows: 1 (m/z 1133.6), 2 and 3 (m/z 1117.6), 4 and 5 (m/z 1083.6), 6 (m/z 1101.6), 7 and 8 (m/z 1045.6), 9 (m/z 1061.6), 10 (m/z 1063.6), 11 (m/z 1079.6), 12 (m/z 1063.6), 13 (m/z 1077.6), 14 (m/z 779.5), 15 (m/z 1101.6), and 16 (m/z 1115.6).
’ RESULTS AND DISCUSSION LC-MS/MS Conditions for CTXs Analysis. The MS/MS conditions were optimized using 51-hydroxyCTX3C (9) because it was detectable at low levels. In H2O/MeOH mobile phases, the toxin produced a prominent [M + Na]+ ion, which upon collision at 40 eV produced the same ion of nearly unchanged intensity. In contrast, the use of a H2O/MeCN mobile phase gave rise to prominent peaks of the protonated molecule [M + H]+ and serially dehydrated ions thereof. The intensities of the parent [M + H]+ ion and product ions were not comparable to the intensity of the [M + Na]+ peak produced in the H2O/MeOH mobile phase. Selection of an ammonium adduct [M + NH4]+ did not improve the signal responses. Hence, the [M + Na]+ ion was selected for monitoring in the multiple reaction monitoring (MRM) analysis. Addition of ammonium formate to the mobile phase was effective at suppressing matrix interference. MRM Chromatogram of the Standard Toxins. The MRM chromatogram for 13 CTX standards, gambierol, and gambieric acid A and B is shown in Figure 1. The solutions for injection were adjusted to 1 ng/mL of the respective toxins. Because all of the toxins have molecular weights near 1000, the peak heights in Figure 1 reflect the molar response of the toxins. The concentrations for 3, 6, 8, 10, 12, 13, 15, and 16 were less accurate than those of the other CTXs, because they were determined using submicrogram orders of the toxins. The limits for detection (S/N > 3) and quantification (S/N > 10) of CTXs were 0.25 pg and 1.0 pg on-column, respectively. Recovery tests with spiked samples were not carried out because 8889
dx.doi.org/10.1021/ac200799j |Anal. Chem. 2011, 83, 8886–8891
Analytical Chemistry
ARTICLE
Figure 6. Toxin profile in the dinoflagellate, G. toxicus (RGI-1 strain). The inset displays magnification of trace amounts of toxins. *M-seco-CTX4A/B methyl acetal was deduced from [M + Na]+ (m/z 1115) and Rt.
of the limited availability of reference CTXs. However, the detection limits calculated from the data obtained with standard toxins indicated that this method could meet the required level of 0.01 ppb CTX1B equivalent in fish flesh. For example, the method could unambiguously detect CTX1B (1) and 52-epi-54-deoxyCTX1B (2) at 45 and 33 pg/g, respectively, in a flesh judged as nontoxic by MBA. As shown in Figure 1, the LC separation was difficult for several of the toxins but these could be identified in the MRM chromatograms. Identification of epimers, however, could not be achieved without the use of the standard toxins. Toxin Profiles in Fish from Okinawa. The toxin compositions in six species of fish, which were implicated in poisoning, characteristically show only the toxins of the CTX1B type (Figure 2af). The major toxin was 1, except in one specimen of P. laevis in which 1 was a minor toxin (Figure 2d). The possible individual and narrow regional fluctuations within the same species were tested with six specimens of L. bohar varying in collection sites, sizes, and toxicities. No significant fluctuations were observed (Figure S1, Supporting Information). Hence, the profile in Figure 2 can be regarded as representative of this species in this region. Interestingly, an M-seco toxin was detected, although in a trace amount, in fish for the first time. Under acidic conditions, the hemiacetal structure in M-seco toxins cyclizes to form spiro L/M rings.8,9 However, CTX4A and CTX4B were not detected, thereby ruling out the possible ring closure in fish. The absence of CTX3C type toxins indicates that G. toxicus growing in Okinawa produces only CTX1B type toxins. The mouse toxicity calculated from the LC-MS/MS analysis agreed well with the actual MBA result,18 indicating that the present LC-MS/MS method could serve as an excellent alternative to MBA (Figure 3). Moreover, the present method far exceeded MBA in sensitivity and specificity. Toxin Profiles in Fish from Miyazaki and Minamitorishima. The toxin profile of a spotted knifejaw, O. punctatus, from Miyazaki located 730 km northeast of Okinawa differed distinctly from that of fish from Okinawa in that it contained only CTX3C type toxins (Figure 4). This species is the main cause of CFP occurring outside of Okinawa. The present study is the first to reveal the characteristic profile as a regional distinction. The red snapper, L. bohar, from Minamitorishima contained both CTX1B and CTX3C type toxins (Figure 5). This complex profile clearly differed from that observed in other regions. Interestingly, all toxins were of oxidized forms, except for the trace amount of M-seco toxin. The mouse toxicity is calculable using data for the major toxins, 1, 7, 8, 9, and 11.
Toxin Profiles in G. toxicus. An RGI-1 strain of G. toxicus collected at Rangiroa Atoll showed a hybrid type profile (Figure 6). The primary toxins were 4, 5, 7, and 8. The less abundant M-seco analogs are likely biosynthetic precursors to the spiro L/M-ring toxins. The oxidized toxins, 2 and its 52-epimer, and 9, were detected for the first time in the dinoflagellate. Analysis of two additional extracts from the same strain reconfirmed the occurrence of these oxidized toxins (data not shown). Hence, the hypothesis that toxins are mainly oxidized in fish during food chain transmission should be amended. The complex toxin profile should be useful for distinguishing different strains.
’ CONCLUSIONS This study is the first to determine unambiguously as many as 16 ciguatoxins and related polyethers. With its high sensitivity, specificity, and rapidity, the present LC-MS/MS method could dramatically simplify monitoring of fish, currently implemented using MBA. Extracts from 15 g of flesh would sufficiently meet the required detection level of 0.01 ppb CTX1B equivalent for food. The most important element for the success of this study can be attributed to the use of reliable standard toxins produced through the concerted efforts of natural products and synthetic chemists. Previous attempts to determine CTXs by LC-MS/MS had to be carried out with only a few or no reference toxins.2124 The usefulness of the present method extends beyond the monitoring of fish. It enables the high-sensitivity and highaccuracy detection of the metabolic pathways of toxins in animals. Because the occurrence of CTX-producing G. toxicus strains is rare, the high sensitivity and specificity of the present method is a great advantage in identifying toxic strains. Only a small number of wild grown cells would be sufficient for the analysis, eliminating the need for laborious culturing currently undertaken to increase test material. An obstacle to the widespread use and further validation of the present method is the difficulty in obtaining standard toxins. Nevertheless, efforts are underway to remove this obstacle for two toxin groups considered in this study. Judging from the toxin profiles revealed, the method could be put into practical use with the addition of a few more standards. ’ ASSOCIATED CONTENT
bS
Supporting Information. Species, biological and epidemiological data of fish, and data of intraspecies variation of toxins.
8890
dx.doi.org/10.1021/ac200799j |Anal. Chem. 2011, 83, 8886–8891
Analytical Chemistry This material is available free of charge via the Internet at http:// pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
ARTICLE
(22) Stewart, I.; Eaglesham, G. K.; Poole, S.; Graham, G.; Paulo, C.; Wickramasinghe, W.; Sadler, R.; Shaw, G. R. Toxicon 2010, 56, 804–812. (23) Otero, P.; Perez, S.; Alfonso, A.; Vale, C.; Rodriguez, P.; Gouveia, N. N.; Gouveia, N.; Delgado, J.; Vale, P.; Hirama, M.; Ishihara, Y.; Molgo, J.; Botana, L. M. Anal. Chem. 2010, 82, 6032–6039. (24) Roeder, K.; Erler, K.; Kibler, S.; Tester, P.; Van The, H.; Nguyen-Ngoc, L.; Gerdts, G.; Luckas, B. Toxicon 2010, 56, 731–738.
’ ACKNOWLEDGMENT The authors thank Dr. Ritsuko Murakami (Ibaraki Prefectural Institute of Public Health) and Mr. Hirofumi Morioka (Miyazaki Prefectural Institute for Public Health and Environment) for donating fish specimens implicated in CFP, Dr. Hideaki Uchida (Agilent Technologies Japan, Ltd.) for technical advice on LCMS/MS analysis, and Ms. Satsuki Sakugawa, Mr. Hiroyuki Tamaki, and Mr. Koji Tamanaha (Okinawa Prefectural Institute of Health and Environment) for continuing support. This work was financially supported by the Regional Innovation Cluster Program “Okinawan Coastal Area”, the Ministry of Education, Culture, Sports, Science and Technology of Japan. ’ REFERENCES (1) Halstead, B. W. Poisonous and Venomous Marine Animals of the world, Vol. 2; U.S. Government Printing Office: Washington, D.C., 1967. (2) Yasumoto, T.; Murata, M. Chem. Rev. 1993, 93, 1897–1909. (3) Dickey, R. W.; Plakas, S. M. Toxicon 2010, 56, 123–136. (4) Manger, R. L.; Leja, L. S.; Lee, S. Y.; Hungerford, J. M.; Hokama, Y.; Dickey, R. W.; Granade, H. R.; Lewis, R.; Yasumoto, T.; Wekell, M. M. J. AOAC Int. 1995, 78, 521–527. (5) Caillaud, A.; de la Iglesia, P.; Darius, H. T.; Pauillac, S.; Aligizaki, K.; Fraga, S.; Chinain, M.; Diogene, J. Mar. Drugs 2010, 8, 1838–1907. (6) Murata, M.; Legrand, A. M.; Ishibashi, Y.; Yasumoto, T. J. Am. Chem. Soc. 1989, 111, 8929–8931. (7) Satake, M.; Morohashi, A.; Oguri, H.; Oishi, T.; Hirama, M.; Harada, N.; Yasumoto, T. J. Am. Chem. Soc. 1997, 119, 11325–11326. (8) Yasumoto, T.; Igarashi, T.; Legrand, A. M.; Cruchet, P.; Chinain, M.; Fujita, T.; Naoki, H. J. Am. Chem. Soc. 2000, 122, 4988–4989. (9) Yasumoto, T. Chem. Rec. 2001, 1, 228242, and references therein. (10) Hirama, M.; Oishi, T.; Uehara, H.; Inoue, M.; Maruyama, M.; Oguri, H.; Satake, M. Science 2001, 294, 1904–1907. (11) Inoue, M.; Miyazaki, K.; Uehara, H.; Maruyama, M.; Hirama, M. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 12013–12018. (12) Inoue, M.; Miyazaki, K.; Ishihara, Y.; Tatami, A.; Ohnuma, Y.; Kawada, Y.; Komano, K.; Yamashita, S.; Lee, N.; Hirama, M. J. Am. Chem. Soc. 2006, 128, 9352–9354. (13) Oguri, H.; Hirama, M.; Tsumuraya, T.; Fujii, I.; Maruyama, M.; Uehara, H.; Nagumo, Y. J. Am. Chem. Soc. 2003, 125, 7608–7612. (14) Tsumuraya, T.; Fujii, I.; Inoue, M.; Tatami, A.; Miyazaki, K.; Hirama, M. Toxicon 2006, 48, 287–294. (15) Satake, M.; Murata, M.; Yasumoto, T. J. Am. Chem. Soc. 1993, 115, 361–362. (16) Nagai, H.; Torigoe, K.; Satake, M.; Murata, M.; Yasumoto, T.; Hirota, H. J. Am. Chem. Soc. 1992, 114, 1102–1103. (17) Fuwa, H.; Goto, T.; Sasaki, M. Org. Lett. 2008, 10, 2211–2214. (18) Oshiro, N.; Yogi, K.; Asato, S.; Sasaki, T.; Tamanaha, K.; Hirama, M.; Yasumoto, T.; Inafuku, Y. Toxicon 2010, 56, 656–661. (19) Satake, M.; Murata, M.; Yasumoto, T. Tetrahedron Lett. 1993, 34, 1975–1978. (20) Satake, M. In Shokuhin Eisei Kensa Shishin Rikagaku-hen: Ciguatera; Ministry of Health, Labour and Welfare; Japan Food Hygiene Association: Tokyo, 2005; pp 691695 (in Japanese). (21) Chinain, M.; Darius, H. T.; Ung, A.; Cruchet, P.; Wang, Z.; Ponton, D.; Laurent, D.; Paulliac, S. Toxicon 2010, 56, 739–750. 8891
dx.doi.org/10.1021/ac200799j |Anal. Chem. 2011, 83, 8886–8891
