Communication Preliminary study on fluorimetric detection of aflatoxins Q1, P1 and B1 using heptakis-di-O-methyl-b-cyclodextrin as post-column HPLC reagent B. I. Vázquez,*a C. A. Fente,a C. M. Franco,a A. Cepeda,a G. Mahuzierb and P. Prognonb a Laboratorio de Higiene e Inspección de los Alimentos, Dpto. de Química Analítica Nutrición y Bromatología, Facultad de Veterinaria, 2700 Lugo, Spain b Laboratoire de Chimie Analytique II—Bioanalyse, Faculté de Pharmacie, Université Paris-Sud, 5, rue J.B. Clément, 92290 Châtenay-Malabry, France Received 7th July 1998, Accepted 23rd November 1998 Post-column fluorimetric detection for the determination of aflatoxins Q1, P1 and B1 was carried out by using HPLC with a 2.0 mm id column. The post-column reagent consisted of a 1022 M aqueous solution of heptakis-di-O-methyl-b-cyclodextrin. The fluorescence enhancement achieved was 37- and 27-fold for AFQ1 and AFB1, respectively, whereas the AFP1 signal was increased just about 2-fold.With the proposed method, aflatoxins Q1, P1 and B1 can be simultaneously determined in human urine. The detection limits (S/N = 3) were as follows: 0.7 ng ml21 for AFQ1, 0.5 ng ml21 for AFP1 and 0.3 ng ml21 for AFB1. Introduction Aflatoxin B1 (AFB1) is a toxin produced by the widespread moulds Aspergillus flavus and Aspergillus parasiticus.1,2 Following consumption of food and feed contaminated with this mycotoxin, it is known that AFB1 is hydroxylated in the liver by several mammalian species, including humans, to yield aflatoxins Q1 (AFQ1) and P1 (AFP1), the two major metabolites produced (Fig. 1).3–6 Aflatoxins are powerful carcinogenic and mutagenic agents,7 and some of these metabolites have been explored as urinary biomarkers for hepatocellular carcinoma.8,9 All these reasons have focused the efforts of analytical chemists on studying highly sensitive methods to detect AFQ1, AFP1, AFB1 and AFM1. The use of pre-column TFA derivatization has been reported for their determination,10,11 as well as halogenation using bromine12,13 or iodine,14 in order to improve the fluorimetric detection.Nevertheless a major common drawback of these procedures is the instability of the derivatives and the fact that the use of bromine with AFP1 produces a damping of the native fluorescence of the compound.12 The use of cyclodextrins in a post-column reaction system has proved successful in improving the detection limit of the important fungal toxin B1, due to the high fluorescence enhancement obtained with b-CD derivatives.15,16 Our laboratory has determined, in previous work, some spectroscopic data (i.e., Stokes shifts, E(0-0) energy levels, relative quantum fluorescence yields, etc.) for the interaction between the hydroxylated aflatoxins Q1, M1 and P1, and different cyclodextrins. 17 AFM1 was deliberately excluded from the present study due to the demonstrated absence of fluorescence enhancement upon cyclodextrin addition.17 This absence of effect was observed with all cyclodextrins tested i.e., a-, b-, g-, hydroxypropyl-band heptakis-di-O-methyl-b-cyclodextrin. This feature was attributed to a lack of convenient fit between host and guest compound owing to the position of the furanic hydroxyl, which distinguishes AFM1 from AFB1.The aim of the present work was to develop a reversed phase liquid chromatographic system using post-column cyclodextrin derivatization with heptakis-di-O-methyl-b-cyclodextrin in order to enhance the signals of AFQ1, AFP1 and AFB1.A practical example of the feasibility of the proposed method was demonstrated by using human urine. Experimental Chemicals All reagents were of analytical grade. The purity of the organic solvents and ultra-pure water (MilliQ®-quality, Millipore, Molsheim, France) were checked via fluorescence prior to use. Aflatoxins Q1, P1 and B1 were obtained from Sigma (St. Louis, MO, USA), AFQ1 being supplied as a mixture of epimers. Heptakis-di-O-methyl-b-cyclodextrin (DIMEB) was purchased from Wacker (Munich, Germany).Solutions Individual aflatoxins, in crystalline form, were dissolved in acetonitrile as this solvent was recommended to avoid rapid aflatoxin degradation.10 Each solution, protected from light by aluminium foil, was kept at 4 °C for no longer than 1 month. Standards were prepared from these solutions, containing 9.8 3 104 ng ml21 for AFQ1, 4.5 3 103 ng ml21 for AFP1 and 9.4 3 105 ng ml21 for AFB1, and these were stored for one week at 4 °C.Successive dilutions were made with the mobile phase in order to achieve working concentrations and stored at 4 °C for no longer than 24 h. Fig 1 Chemical formulae of the three aflatoxins studied. Aflatoxins Q1 and P1 are hydroxylated derivatives resulting from aflatoxin B1 metabolism. Anal. Commun., 1999, 36, 5–7 5Stock aqueous solutions (1022 M) of the cyclodextrin cited above were prepared daily and maintained at room temperature before use.Urine samples Human urine samples were randomly collected and chromatographically tested for the absence of any of the aflatoxins assayed. Urine samples were prepared and extracted as follows: 50 ml of urine were spiked, by means of a syringe, with 100 ml (calibration limit of syringe) of appropriate amounts of mixtures of AFQ1, AFP1 and AFB1. After homogenization by vigorous shaking, chloroform (5 ml) was added, and then the tube was shaken mechanically for 3 min.After centrifugation for 5 min, the upper aqueous layer was discarded and the organic layer transferred into an amber screw-capped tube and evaporated to dryness under a stream of nitrogen. The residue was reconstituted in 0.5 ml of the mobile phase, and 20 ml injected. Chromatographic conditions The chromatographic system consisted of two LC6A metering pumps (Shimadzu, Kyoto, Japan), one for the mobile phase, a methanol–water mixture (40 + 60 v/v) at a flow rate of 0.2 ml min21, equipped with a Rheodyne® Model 7125, 20 ml loop injector (Cotati, CA, USA), and the second for introducing the post-column reagent, a 1022 M aqueous cyclodextrin solution pumped at a flow rate of 0.3 ml min21.The mixing of the mobile phase and the post-column reagent was performed with a Tee mixer (Supelco, Bellefonte, PA, USA). For chromatographic separation a C18 Nucleosil column (5 mm particle size, 120 Å pore size, 150 3 2.0 mm id) from Tecknochroma (Barcelona, Spain), was used.All measurements were made at room temperature (22 ± 2 °C). A Perkin-Elmer LC240 fluorescence detector (Norwalk, CT, USA) was programmed with the following excitation and emission wavelengths: lex = 365 nm, lem = 466 nm for AFQ1; lex = 365 nm, lem = 504 nm for AFP1; and lex = 360 nm, lem = 435 nm for AFB1, according to ref. 14. The chromatograms were recorded on a Shimadzu CR5A Chromatopac integrator (Kyoto, Japan).When cyclodextrins were added in the mobile phase, they were first dissolved in water and then methanol was added to yield the final desired concentration. Results and discussion Reversed phase HPLC methods (RP-HPLC) are the most widely used for aflatoxins. So, from our own experience with aflatoxins B and G, an isocratic mobile phase of methanol– water (40 + 60 v/v) was envisaged. Methanol was preferred over acetonitrile because of its lower association constant with cyclodextrins.18 A common C18 Nucleosil column (5 mm particle size, 120 Å pore size, 150 3 2.0 mm id) was employed.These toxins are simultaneously determined, in less than 30 min, and the elution order was Q1 < P1 < B1, in accordance with their polarities. Adding cyclodextrins in the mobile phase and post-column optimization In our previous work,17 it was spectroscopically observed that DIMEB, a b-cyclodextrin derivative, was the most suitable for the enhancement of the fluorescence emission of these three mycotoxins.Hence, DIMEB at a concentration of 1022 M, was directly added to the methanol–water mobile phase. Despite the increase in the signal, the analytical problems encountered with the direct introduction of cyclodextrin in the eluent are first, the decrease of the resolution, and second, the back pressure increase due to a higher viscosity of the eluent. Consequently, a post-column addition was preferentially used. For a mobile phase flow rate of 0.2 ml min21, and a fixed concentration of AFQ1, AFP1 and AFB1 (150 ng ml21 each), the post-column dilution effect was studied, first from 0.1 ml min21 to 0.5 ml min21 using deionized water as the reagent.A progressive decrease in the fluorescence signal from about 50 to 10% of the initial signal was observed for these aflatoxins. The decrease ratio (flow rate mobile phase : flow rate post-column solution) does not exactly correspond to that expected theoretically.This could be due to the addition of the water leading to quenching of the fluorescence signal, as already reported.18 Then, in the same experimental conditions, an aqueous 1022 M DIMEB solution (the optimal concentration, yielding the highest signal) was used as post-column reagent and the highest enhancement of the fluorescence signal was achieved at a postcolumn flow rate of 0.3 ml min21 (37-, 2- and 27-fold for AFQ1, AFP1 and AFB1, respectively, referred to the fluorescence signal obtained with post-column water).Although this postcolumn flow rate is larger than that of the mobile phase, the cyclodextrin concentration achieved seems to form inclusion complexes in an environment suitable for an optimum fluorescence signal. AFQ1 was the more highly promoted, even more than AFB1 (37- and 27-fold, respectively); in contrast, AFP1 remained nearly constant (about 2-fold), which may be due to its phenolic structure, which leads to some different spectroscopic behaviour.It should be noted that, although disappointing in comparison with the other aflatoxins tested, this finding allows us to determine AFP1, which is not possible by the destructive post-column halogenation derivatization.12 From a spectroscopic point of view, neither spectral shifts of the emission nor of the excitation have been noticed upon cyclodextrin addition. This indicates that the change in polarity of the eluent after post-column reagent addition has no effect on the environment of the aflatoxin included.Analytical figures of merit. Analysis of real examples The method was tested in real conditions by spiking human urine samples with standards containing a mixture of AFQ1, AFP1 and AFB1. Data from the extraction recovery and F-linearity test of the different aflatoxins, as well as repeatability and reproducibility of the method, are summarized in Table 1. The detection limits (S/N = 3) were as follows: 0.7 ng ml21 for AFQ1, 0.5 ng ml21 for AFP1 and 0.3 ng ml21 for AFB1, with 20 ml injections of the standard samples analysed by HPLC.Table 1 Analytical figures of merit Aflatoxin Extraction recovery (%)a RSD (n = 5) (%) F-linearity (r)a (p < 0.001) (n = 4) (r = ) Repeatabilityb RSD (n = 8) (%) Reproducibilityb RSD (n = 8) (%) LOD (S/N = 3) /ng ml21 Q1 96 9 0.998 9 11 0.7 P1 87 17 0.992 4 7 0.5 B1 94 6 0.996 5 5 0.3 a Extraction recovery and F-linearity test were demonstrated for: AFQ1 from 3 ng ml21 to 66 ng ml21; AFP1 from 2 ng ml21 to 60 ng ml21; and AFB1 from 2 ng ml21 to 31 ng ml21.b Repeatability and reproducibility were demonstrated for c = 15 ng ml21 for AFQ1, AFP1 and AFB1. 6 Anal. Commun., 1999, 36, 5–7These limits certainly do not reach the ng l21 levels for AFB1 and AFQ1 reported in other procedures in the literature (6.8 ng l21 and 18 ng l21, respectively).12,13 Nevertheless, taking into account the aflatoxins levels reported in human urine (0.5–16 ng ml21 for AFP1,3 0.1–4.8 ng ml21 for AFB1 19–21), the method should be suitable.It can be envisaged that with a higher volume of urine sample extracted and with a fluorescence detector of better performance, the limited accuracy of the proposed method will be significantly increased and thus will lie with the best limit of detection reported in the literature. Finally, Fig. 2 shows the typical chromatograms of a human urine blank (A) and a spiked urine (26 ng ml21 AFQ1, 3 ng ml21 AFP1 and 19 ng ml21 AFB1) (B). For both (A) and (B) deionized water was added post-column in order to get the times to match and to obtain the same quenching-water effect as in (C).(C) is the same spiked urine sample analysed by the proposed method with 1022 M DIMEB post-column reagent, showing the improvement of the fluorescence signal for these aflatoxins. References 1 P. M. Scott, J. Assoc. Off. Anal. Chem., 1987, 70, 276. 2 I. Dvorackova, in Aflatoxins and Human Health, CRC Press, Boca Raton, FL, USA, 1990, 458. 3 G. S. Qian, R. K. Ross, M .C. Yu, J. M. Yuan, Y. T. Gao, B. E. Henderson, G. N. Wogan and J. D. Groopman, Cancer. Epidemiol. Biomarkers. Prev., 1994, 3, 3. 4 B. D. Roebuck and G. N. Wogan, Cancer Res., 1977, 37, 1649. 5 G. H. Büchi, P. M. Müller, B. D. Roebuck and G. N. Wogan, Res. Commun. Chem. Pathol. Pharmacol., 1974, 8, 585. 6 G. E. Neal and P. J. Colley, Biochem. J., 1978, 174, 839. 7 W. F. Busby and G. N. Wogan, in Aflatoxins, ed.G. N. Searleed, American Chemical Society, Washington, DC, USA, 1985, p. 1. 8 R. K. Ross, J. M. Yuan, M. C. Yu, G. N. Wogan, G. S. Qian, J. T. Tu, J. D. Groopman, Y. T. Gao and B. E. Henderson, Lancet, 1992, 339, 943. 9 J. D. Groopman, J. Q. Zhu, P. R. Donahue, A. Pikul, L. S. Zhang, J. S. Chen and G. N. Wogan, Cancer Res., 1992, 52, 45. 10 D. L. Orti, J. Grainger, D. L. Ashley and R. H. Hill, Jr., J. Chromatogr., 1989, 462, 269. 11 H. Joshua, J. Chromatogr.A, 1993, 654, 247. 12 A. Kussak, B. Andersson and K. Andersson, J. Chromatogr. B, 1994, 656, 329. 13 A. Kussak, B. Andersson and K. Andersson, J. Chromatogr. B, 1995, 672, 253. 14 H. Jansen, R. Jansen, U. A. Th. Brinkman and R. W. Frei, Chromatographia, 1987, 24, 555. 15 A. Cepeda, C. M. Franco, C. A. Fente, B. I. Vázquez, J. L. Rodríguez, P. Prognon and G. Mahuzier, J. Chromatogr. A, 1996, 721, 69. 16 O. J. Francis, G. P. Kircheneuter, J. R. G. M. Ware, A. S. Carman and S. S. Kuan, J. Assoc. Off. Anal. Chem., 1988, 71, 725. 17 C. M. Franco, C. A. Fente, B. I. Vázquez, A. Cepeda, G. Mahuzier and P. Prognon, J. Chromatogr. A, 1998, 815, 21. 18 Y. Matsui and K. Mochida, Bull. Chem. Soc. Jpn., 1979, 52, 2808. 19 G. Niedwetzki, G. Lach and K. Geschwill, J. Chromatogr. A, 1994, 661, 175. 20 R. Guan, C. J. Oon, C. Wild and R. Motesano, Ann. Acad. Med. Singapore, 1986, 15(2), 201. 21 J. D. Groopman, A. J. Hall, H. Whittle, G. J. Hudson, G. N. Wogan, R. Motesano and C. P. Wild, Cancer Epidemiol., Biomarkers Prev., 1992, 1, 221. 8/09150A Fig 2 Chromatograms of human urine extracts: urine blank (A); spiked urine (26 ng ml21 AFQ1, 3 ng ml21 AFP1 and 19 ng ml21 AFB1) (B); the same spiked urine sample, detected with the proposed method, with DIMEB (1022 M) as post-column reagent (C). For detailed chromatographic conditions, see text. Anal. Commun., 1999, 36, 5–7 7