FLUORIMETRIC QUANTITATION OF OPA-DERIVATIVES OF FUMONISINS B₁ AND B₂ IN CORN AND FUSARIUM MONILIFORME CULTURE EXTRACTS

QUALIFICAÇÃO FLUORIMÉTRICA DE DERIVADOS OPA DAS FUMONISINAS B₁ E B₂ DE MILHO E DE EXTRATOS DE CULTURA DE FUSARIUM MONILIFORME.

ABSTRACT

A system for HPLC analysis of fluorescent OPA-derivatives of fumonisins B₁ and B₂ was developed, using a reversed-phase C_{1 8} column and methanol-sodium acetate buffer as mobile phase. The system was applied successfully to the analysis of 195 com extracts and 42 culture extracts. FB₁ contents ranged from 0.88 to 49.32 µg/g in com samples, and from 20 to 2168 µg/g in culture samples. FB₂ contents ranged from 1.78 to 29.16 µg/g in com samples and from 10 to 380 µg/g in culture samples. Optimal retention times varied from 4.4 to 4.9 min. for FB₁ and 8.6 to 9.5 min. for FB₂ .

KEY WORDS:
Fumonisin; Fusarium moniliforme, com.

RESUMO

Um sistema para análise de fumonisinas B_I e B₂ por cromatografia líqüida de alta eficiência foi desenvolvido, usando-se uma coluna de fase reversa C_{1 8} e metanol-tampão acetato de sódio como fase móvel. O sistema foi aplicado com êxito à análise de 195 extratos de milho e 42 extratos de cultura de Fusarium monilifonne. Os teores de FB₁ variaram de 0,88 a 49,32 µg/g em amostras de milho e de 20 a 2168 µg/g em amostras de cultura. Os teores de FB₂ variaram de 1,78 a 29,16 µg/g em amostras de milho e de 10 a 380 µg/g em amostras de cultura. Tempos de retenção ótimos variaram de 4,4 a 4,9 min para FB₁ e de 8,6 a 9,5 min para FB₂ •

PALAVRAS-CHAVE:
Fumonisina; Fusarium moniliforme, milho.

INTRODUCTION

Fumonisins are toxic metabolites produced mainly by the widespread corn crop infectants Fusarium moniliforme and F. proliferatum (Bezuidenhout et al.,1988; Gelderblom et al., 1992). There has been an increasing interest in the development of analytical methods for fumonisins, dueto the evidence of their widespread natural occurrence in corn-based food and feed commodities, and to the association of fumonisins, especially fumonisin B₁ (FB₁ ) and fumonisin B₂ (FB₂ ) with animal diseases like equine leukoencephalomalacia (ELEM), porcine pulmonary edema (PPE) and nephrotoxicity and liver cancer in rats. Besides their toxicity also to poultry, fumonisins have been epidemiologically implicated in human esophageal cancer and have been declared as potentially carcinogenic to humans by the lntemational Agency for Research on Cancer (Ross et al., 1992; Shephard et al., 1996).

In Brazil, ELEM was first reported in São Paulo (Rego, 1950) associated with the ingestion of com, and its correlation to Fusarium moniliforme was established throughout the 1980s, when ELEM outbreaks were reported in several Brazilian states (Riet-Correa et al., 1982; Brito et al., 1982; Hirooka et al., 1990). From 1990 to 1993, 29 outbreaks of the disease were reported in Rio Grande do Sul, São Paulo and Minas Gerais (Meireles et al., 1994).

Literature reports analytical techniques using chromatography or enzyme-linked immunosorbent assay (ELISA) for quantitation of fumonisins in different matrixes, such as com, animal feeds, Fusarium spp. cultures and milk (Shephard et al., 1990; Stdenham et al., 1996; ROSS et al., 1991; Maragos & Richard, 1994; Trucksess et al., 1995). LC methods have proved more sensitive, efficient and accurate than TLC and ELISA, and less complex and time consuming than GC/ MS methods. As fumonisins do not absorb light in a selective UV-visible spectral region, nor do they fluoresce, derivatization procedures have been employed to make their detection possible. Derivatization with maleic anhydride has been reported to enable the quantitative determination of FB_I and F B₂ through LC with UV detection (Gelderblom et al., 1988). The sensitivity of the procedure, however, was limited to 10 µg/g. Reaction with ortho-phthaldialdehyde(OPA) (Shephard et al., 1990), fluorescamine (Ross et al.: 1991) or naphthalene-2,3-dicarboxaldehyde (NDA) (Maragos & Richard, 1994) has been investigated for the production of fluorescent derivatives of fumonisins. Pre-column OPA derivatization, followed by separation of fumonisins on reversed-phase LC columns and fluorescence detection has been applied successfully for the quantitation of FB_{1 ,} F B₂ and FB₃ in a variety of commodities. Most methods use methanol-water or acetonitrile-water, for the extraction of the toxins, and solid-phase extraction (either strong anion-exchange -SAX - or reversed-phase Cm cartridges) (Sydenham et al., 1996; RlCE et al., 1995) or immunochemical procedures (immunoaffinity colurnn) for cleanup of the extracts (Trucksess et al., 1995).

In this paper, we report an altemative system for LC isocratic separation of the derivatives of FB₁ and FB_2• For the extraction and purification of FB _I and FB₂ in com and F. moniliforme cultures an established procedure was applied (Ross et al., 1991). The toxins were analysed by fluorimetric detection of their OPA derivatives after separation on a reversed-phase LC column. Reaction conditions and chromatographic efficiency were optimized in order to achieve the shortest run times consistent with the resolution and sensitivity required. Assay accuracy and precision, and linearity of detector response to FB _I and FB₂ concentrations were determined.

MATERlALS AND METHODS

Safety note: As toxic metabolites, declared as potentially carcinogenic to humans by the lnternational Agency for Research on Cancer, fumonisins should be handled with caution. Laboratory spills should be washed with a 5% aqueous solution of commercial sodium hypochlorite followed by water, and waste disposed of according to applicable environmental rufes and regulations (Sydenham et al., 1996).

Fumonisin extracts from com and F. moniliforme cultures were obtained at the Departmento de Microbiologia do Instituto de Ciências Biomédicas, Universidade de São Paulo. These extracts were prepared and cleaned up according:tcg, ihe method of Ross (Ross et al., 1991 ), which uses acetonitrilewater solutions and solid-phase extraction through a C ₁₈ reversed-phase cartridge. FB₁ and FB₂ were eluted. with 2 mL of acetonitrile-water (7+3, v/v). The eluates were kept in a freezer, at temperatures below -20ºC. When necessary, they were diluted with methanol-water (7+3, v/v), prior to derivatization.

Fumonisin standard solutions: Standard fumonisins B₁ and B₂ were purchased from Sigma Chemical (St. Louis, MO, USA). Working standard solutions in different concentrations, used for construction of calibration curve, deterrnination of accuracy and precision, and other experiments, were prepared periodically by dilutions with methanol-water (7+3, v/v), from stock solutions of each fumonisin standard, sealed, and stored in a freezer, protected from light, at -20ºC. Using this procedure, we observed no toxin degradation in concentrated stock solutions (50 µg/mL, for FB_I and 25 µg/mL, for FB_{2 )} for a period longer than a year. For diluted standard solutions, at such low concentrations as 0.25 µg/mL, however, one should be aware about possibility of toxin degradation, as has been reported (Trucksess et al., 1995).

Derivatization reagent: Prepared with 40 mg ophthaldialdehyde(OPA), dissolved in 1 mL methanol, mixed with 5 mL filtered 0.1M disodium tetraborate solution and 50 µL 2-mercaptoethanol. Reagent is stable for ca. 1 week, stored in freezer, in capped amber vial (Sydenham et al., 1996).

LC mobile phase: Methanol-pH 3.5 sodium acetate buffer (77+23, v/v).

Sodium acetate buffer at pH 3.5 was prepared by mixing 460 mL of a 0.2M solution of acetic acid (12 mL CH₃ CO₂ H in 1,000 mL) and 40 mL of a 0.2M solution of sodium acetate (16.4g of CH, CO, Na or 27.2g of CH₃ CO₂ Na.3Hp in 1,000'mL), diluted with water to make 1,000 mL. When necessary, the pH of acetate buffer was adjusted with glacial acetic acid. Buffer solution was always filtered through nitrocellulose membrane filter, 0.45 µm (Sartorius, Goettingen, Germany).

Analytical-grade reagents were used in reactions and in mobile phase buffers.

Milli-Q water (Millipore, Belford, MA, USA) and LC grade methanol and acetonitrile (E.M. Science / Merck, Darmstadt, Germany) were used in ali preparations subjected to LC.

Liquid Chromatograph System: A Shimadzu (Kyoto, Japan) HPLC system, consisting of a controller (SCL-6B ), two pumps (LC-6AD), a Rheodyne injector with loop system calibrated to deliver 20 µ1, a column oven, a fluorescence detector (RF 551) anda data processar (C-R7A), was used in analysis and optimization experiments.

LC colum: A C_{1 8} reversed-phase column, 150x4.6mm ID, 5µm particle siz (Ultracarb 5 ODS 20 - Phenomenex, Torrance, USA) and in-line corresponding guard column was used in sample analysis.

UV/Visible Spectrophotomer: A Pye Unicam SPS-400 model (Cambridge, England), with wavelength adjusted at 335 nm, was used in reaction rate deterrnination.

Reaction time optimization: The reaction rate of fumonisins FB_I and B₂ with ortho-phthaldialdehyde was determined by measuring the UV-absorption at 335 nm of each derivative forrned during the time range of 0-20 min, after addition of 50 µL OPA-reéj.gent to separate 200 µL aliquots of standard solutions of FB_I and FB_{2 ,} at room temperature (22-23º C) and wavelength 335 nm.

LC system optimization: Solvent systems tested as mobile phases included methanol-phosphate buffer (pH 3.35); methanol-water-acetic acid; acetonitrile-water-acetic acid, and methanol-sodium acetate buffer (pH 3.3 and 3.5), in different proportions, with isocratic and gradient elution. The effect of variables such as room temperature (20-25 ºC), column oven temperature (25-30 ºC), column parameters (lengths of 25 and 15 cm, C ₁₈ or C₈ packing material) and flow rate (1-1.2 mL/min) on peak resolution and retention times here also deterrnined.

Derivatization reaction and LC analysis: A 200 µL aliquot of sample eluates or fumonisin standard working solutions was mixed in a small test tube with 50 µL OPA reagent and allowed to react for 1 minute, at room temperature. Then, 20 µL of derivatization mixture was injected into the LC system. OPA derivatives of fumonisins B_I and B₂ were separated and eluted isocratically, using methanol-pH 3.5 sodium acetate buffer (77+23, v/v), on a reversedphase C₁₈ column, under column oven temperature 27-30 ºC and flow rate 1.l - 1.2 mL/min. Room temperature was kept at 22-23 ºC.

Detection and Quantitation: Quantitation ofthe OPA derivatives of FB₁ and F B₂ was based on elicitation of their fluorescence, with excitation at 335 nm and detection at 440 nm. Fumonisin contents of samples were determined using a standard calibration curve constructed according to the leastsquares method, by means of an eletronic integrator. Five different solutions were used, containing both standards in concentrations ranging from 0.25-1.5 µg/ mL (FB₁ ) and 0.5-3.5 µg/mL (FBJ

RESULTS AND DISCUSSION

Chromatographic system: Efficiency and reproducibility of peak resolution, along with the shortest chromatographic run times, were obtained using sodium acetate buffer at pH 3.5, as the ionic component of mobile phase. Using other systems, such as methanol-phosphate buffer, wateracetonitrile-acetic acid, tested in different proportions, we failed to reproduce the sarne results.

The mobile phase methanol-pH 3.5 acetate buffer (77+23, v/v), with flow rate 1.0-1.2 mIJmin, in a C-18 column (l 50x4.6mm ID), resulted in a fast system, with optimal retention times of 4.4-4.9 min, for FB₁ , and 8.6-9.5 min, for FB_{2 •} Peak shape and retention time of FB₂ OPA-derivative were more strongly affected by buffer pH and buffer proportion in the mobile phase than the corresponding variables for FB₁ OPA-derivative. A proportion of acetate buffer higher than 25 %, in this system, resulted in a broader peak for FB_{2 •} With a lower buffer pH, such as 3.3, anda proportion of 27% sodium acetate buffer in mobile phase, retention times increased slightly, but better separation of fumonisin derivatives from matrix components was achieved. Typical chromatograms for com samples, obtained with these conditions are shown in Figure 1.

Column oven temperatures of 27-30 ºC allowed areduction of viscosity of mobile phase - which is generally high when buffer salt systems are used -- with better pump performance and lower pressure in the column. Within this temperature range, we observed no breakdown of fumonisin-derivatives in the column. It should be noted that we had a mixture mobile phase pH not lower than 4.0, which is a safer condition for column and pump-piston head preservation, and does not favor fumonisin artifact formation (Gelderblom et al., 1992), more likely to occur with low pH mobile phases.

Fumonisin analysis in coro and culture extracts: The system was used successfully in the analysis of 195 com extracts and 42 culture extracts, with FB₁ contents ranging from 0.88 to 49.32 µg/g, in corn samples, and from 20 to 2168 µg/g in culture samples. FB₂ contents ranged from 1.78 to 29.16 µg/g, in corn samples, and from 10 to 380 µg/g in culture samples (Corrêa et al., 1997).

Linearity, detection limit, accuracy and precision: To construct the standard calibration curve, 20 µL of each derivatization mixture were injected in the chromatograph, containing equivalent quantities of FB₁ and FB₂ in the range of 4 to 24 ng and 8 to 56 ng, respectively. Throughout this range, linearity was qbserved..

Fig. 1
Chromatograms of OPA derivatives of FB₁ and FB₂ obtained from a fumonisin standard solution (a) and a purified extract of corn sample containing 4.05 µg/g FB₁ and 2.30 µg/g FB₂ (b). Mobile phase: methanol - pH 3.3 sodium acetate buffer (73:27, v/v); flow rate, 1.0 mL/min.

The minimal concentrations detectable rn the conditions used were 1 O ng/µL for FB_I and FB_{2 ,} in com and in culture sample extracts.

Table 1 shows accuracy and precision p ta determined for the assay. ' ··

Table 1
Accuracy and precision data for the LC assay of fumonisins.

	Toxin	Spiked toxin (pg/mL)	Precision3 %	Accuracy6 %
Corn extracts	FB,	0.50	3.4	116
		0.25	5.2	88
	FB,	0.25	3.7	108
		0.125	5.6	94
(Data based on 4 determinations for each level of toxin added)
F moniliforme	FB,	0.50	3.8	117
culture extracts		0.25	4.9	97
	FB,	0.25	2.4	115
		0.125	2.1	84
(Data based on 3 determinations for each level of toxin added)

• ^a
  Coefficient of variation, expressed as relati ve standard deviation.
• ^b
  Percentage relation between the concentration of toxin measured and the concentration oftoxin added to blank com or culture extract.

Concentration and pH of-OPA reagent: Though the OPA reagent would be expected not to present any fluorescence, in most chromatograms we did observe a few very strong peaks in the range of ca. 2-4 min (Fig. 1), which can be attributed to residual contamination of its components. This was a constant and difficult problem to eliminate, but it could be minimized by reducing the proportion ofthe reagent in the derivatization mixture. An excess of OPA-reagent relative to toxin concentration was, however, always assured, and fumonisin concentrations in sample eluates to be analysed were maintai'i'led within the range of fumoni_sin concentrations in the Standard solutions used to .construct the calibratiori curve. When necessary, samplê eluates were diluteâ. An effitient derivatization rêaction was obtained using the OPA reagent in a volumetric proportion of 1 :4 relative.to sample extract or standard toxin solution, with reagent pH around 8.7.

When working with a calibration curve and automated quantitation, chromatographic conditions of flow, mobile phase polarity, room and column oven temperatures become criticai and must be kept under strict control. Analytical conditions used for sample eluates must, of course, always reproduce the conditions used for the standard solutions. The sarne happens with pH and concentration of reagent, reaction temperature and reaction time for derivatization.

Reaction time:Figure 2 shows the rate of formation of FB₁ and FB₂ OPA-derivatives in the conditions peculiar to our experiments. As reaction curves show, maximum absorptions of FB₁ and FB₂ derivatives are attained in ca. 10 and 7 minutes, respectively. The data obtained for FB_I formation agree closely with those reported by Rice et al. (1995), which preconize a reaction time of ten minutes for the derivatization of fumonisins. The orthophthaldialdehyde proportion relative to the amount of toxin in the reaction mixture, in their experiment, as well as in ours are lower than what is commonly used by most authors who preconize a reaction time of ca. 1 minute (Sydenham et al., 1996; Trucksess et al., 1995). These authors have not presented, however, reaction curves for FBr Our experiment showed that, when 1 minute was the fixed time between addition ofOPA-reagent and injection ofthe reaction product in the chromatograph, this variable deserved special attention, mainly concerning FB₂ analysis, as cán be deduced from Figure 2. Care must be taken to reproduce exactly the injection time for derivatized sample eluates and standard solutions. Whenever analysis speed is not a major requirement, and sensitivity of the method needs to be optimized, a Ionger reaction time would preferably be adopted. In our conditions, at 2 minutes of reaction, for example, we already had - 65% of maximum absorption for FB₁ derivative and -78% for FB₂ derivative. Both derivatives were sufficiently stable, and at 20 minutes of reaction we still had - 97% e - 92% of maximum response for FB _I and FB_2, respectively.

Fig. 2
Curves of reaction of fumonisins B₁ and B₂ with ÓPA.

CONCLUSIONS

The chromatographic system developed herein proved sufficiently sensitive and efficient for the analysis of fumonisins FB₁ and FB₂ in com samples and F. moniliforme cultures, with chromatographic run times faster than most systems reported in the literature using buffer stabilized mobile phase and isoeratie LC eonditions., FB₃ behaviour was not determined in this system, as its standard was not available. Analysis time may result longer if mobile, phase pR•and other-chromatographi<H., onditions have to be altered in order to provide an efficient separation of this toxin. However, in cases where a large amount of samples and the need of getting fast results makes speed of analysis crucial, a possible co-elution of FB₂ and FB₃ may become a minor disadvantage, and the system herein proposed will be a fast and satisfactory altemative for fumonisin analysis.

REFERENCES

• BEZUIDENHOUT, G.C.; GELDERBLOM, W.C.A.; GORST-ALLMAN, C.P.; HORAK, R.M.; MARASAS, W.F.O.; SPITELLER, G.; VLEGGAAR, R. Structure elucidation of the fumonisins, mycotoxins from Fusarium moniliforme. J. Chem. Soe. Chem. Commun., v.11, p. 743-745, 1988.
• BRITO, L.A.B.; NOGUEIRA, R.H.G.; PEREIRA, J.J.; CHQUILOFF, M.A.G.; BIONDINI, J. Leucoencefalomalácia em eqüino associada à ingestão de milho mofado. Arq. Ese. Vet., Univ. Fed. Minas Gerais, v.34, p. 49-53, 1982.
• CORRÊA, B.; ORSI, R.B.; MALLOZZI, M.A.B.; DIAS, S.M.C.; ISRAEL, W.M. Ocorrência natural de fumonisinas em híbridos de milho recém-colhidos e armazenados no Brasil. In: CONGRESO LATINOAMERICANO DE MICOTdXICOLOGIA, 2., 1997, Maracay, Venezuela. Resumenes. p. 83.
• GELDERBLOM, W.C.A.; JASKIEWICZ, K.; MARASAS, W.F.O.; THIEL, P.G.; HORAK, R.M.; VLEGGAAR, R.; KRIEK, N.P.J. Fumonisins - novel mycotoxins with cancer promoting activity produced by Fusarium moniliforme. Appl. Environ. Mierobiol., v. 54, p. 1806-1811, 1988.
• GELDERBLOM, W.C.A.; MARASAS, W.F.O.; VLEGGAAR, R.; THIEL, P.G.; CAWOOD, M.E. Fumonisins: Isolation, chemical characterization and biological effects. Myeopathologia, v. 117, p. 11-16, 1992.
• HIROOKA, E.Y.; VIOTTI, N.M.A.; MAROCHI, M.A.; ISHII, K.; UENO, Y. Leucoencefalomalácia em eqüinos no norte do Paraná. Rev. Mierobiol., v. 21, p. 223-227, 1990.
• MARAGOS, C.M. & RICHARD, J.L. Quantitation and stability of fttmonisins B_I and0 B₂ in milk. J. Assoe. Off. Anal. Chem., v. 77, p.1162-1167, 1994.
• MEIRELES, M.C.A.; CORRÊA, B.; FISCHMAN, O.; GAMBALE, W.; PAULA, C.R.; CHACONRECHE, N.O.; POZZR, C.R. Mycoflora of toxic feeds associated with equine leucoencephalomalacia (ELEM) outbreaks in Brazil. Myeopathologia, v. 127, p.183-188, 1994.
• REGO, C.L. Doença de sintomatologia nervosa causada por intoxicação pelo milho. Rev. Mil. Remonta Vet., v. 10, p.199-215, 1950.
• RICE, L.G.; ROSS, P.F.; DEJONG, J.; PLATTNER, R.D.; COATS, J.R. Evaluation of a liquid chromatographic method for the determination of fumonisins in com, poultry feed and Fusa rium culture material. J. Assoe. Off. Anal. Chem., v. 78, p.1002-1009, 1995.
• RIET-CORREA, F.; MEIRELES, M.C.A.; SOARES, J.M.; MACHADO, J.J.; ZAMBRANO, A.F. Leucoencefalomalácia em eqüinos associada à ingestão de milho mofado. Pesqui. Vet. Bras., v. 2, p.27-30, 1982.
• ROSS, P.F.; RICE, L.G.; PLATTNER, R.D.; OSWEILER, G.D.; WILSON, T.M.; OWENS, D.L.; NELSON, H.A.; RICHARD, J.L. Concentrations of fumonisin B_I in feeds associated with animal health problems. Myeopathologia, v. 114, p.129-135, 1991.
• ROSS, P.F.; RICE, L.G.; OSWEILER, G.D.; NELSON, P.E.; RICHARD, J.L.; WILSON, T.M. A review and update of animal toxicoses associated with fumonisin contaminated feeds and production of fumonisins by Fusariwn isolates. Myeopathologia, v. 117, p. 109-114, 1992.
• SHEPHARD, G.S.; SYDENHAM, E.W.; THIEL, P.G.; GELDERBLOM, W.C.A. Quantitative determination of fumonisins B, and B₂ by highperformance liquid chromatography with fluorescence detection. J. Liq. Chromatogr., v. 13,p.2077-2087, 1990.
• SHEPHARD, G.S.; THIEL, P.G.; STOCKENSTRÕM, S.; SYDENHAM, E.W. Worldwide survey of fumonisin contamination of com and com-based products. J. Assoe. Off. Anal. Chem., v. 79, p. 671-687, 1996.
• SYDENHAM. E.W.; SHEPHARD, G.S.; THIEL, P.G.; STOCKENSTRÕM, S.; SNIJMAN, P.W., VAN SCHALKWYK, D.J. Liquid chromatographic determination of fumonisins B₁ , B₂ and B₃ in com: AOAC IUPAC collaborative study. J. Assoe. Off. Anal. Chem., v. 79, p. 688-696, 1996.
• TRUCKSESS, M.W.; STACK, M.E.; ALLEN, S.; BARRION, N. Immunoaffinity column coupled with liquid chromatography for determination of fumonisin B, in canned and frozen sweet com. J. Assoe. Off. Anal. Chem., v. 78, p. 705-710, 1995.

Publication Dates

• Publication in this collection
  14 Feb 2025
• Date of issue
  Jul-Dec 1999

History

• Received
  15 Jan 1999