7435

NATURAL 3:294-298 (1995)

Detection of Produced in Fusarium moniliforme Cultures by HPLC With Eleetrospray MS and Evaporative Light Scattering Detectors

Ronald D. Plattner 8ioactive Constituents Research, National Center for Agricultural Utilization Research, Midwest Area, Agriculture Research SeNice, United States Department of Agriculture, Peoria, Illinois

ABSTRACT A variety of toxic secondary metabolites including fumonisins, can be produced by the fungus Fusarium moniliforme and closely related species in section Liseola in large amounts (g/kg in laboratory cultures). Underivatized fumonisins were detected by HPLC using either an evaporative light scattering detector or electrospray MS. Electrospray MS used together with NMR and GC/MS was used to identify a new

, fumonisin C4 , which corresponds in structure to fumonisin 84 with the C-1 terminal methyl group missing. Several novel strains of F. moniliforme mating population A were identified that produced little or no

fumonisin 81, but large amounts of either fumonisin 82, or 83 together with fumonisin 84 and C4 . These strains

which do not produce fumonisin 81 should prove useful in purification of fumonisin 82, 83, C4 , and 84 for toxicology studies. @ 1995 Wiley-Liss, Inc.'

Key Words: Fusarium moniliforme, Fumonisin, , Electrospray, MS

INTRODUCTION much lower levels. Alternatively, hydrolysis of the ester The fungus Fusarium moniliforme and several closely side-chains followed by gas chromatography/mass spectrom­ related species in section Liseola are frequently found to etry (GClMS) of the trimethyl silyl or trifluoroacetate deriv­ contaminate agricultural products. especially maize [Mara­ atives of the fumonisin backbones has been used [Plattner sas et al .. 1984]. Although these species of the genus Fusar­ et a!., 1990, 1992] to measure fumonisins. This method ium do not make . they do produce high levels gives results that agree well with those obtained by fluores­ of numerous secondary metabolites that are toxic to cence HPLC but does not distinguish between the A and B and animals. The most studied of the known toxins pro­ series offumonisins. A third method which detects the intact duced by F. monil(lorlne are the fumonisins (Fig. 1). This fumonisins without derivatization is fast atom bombardment! family ofmycotoxins is capable ofproducing toxins at gram! mass spectrometry (FAB/MS) [Korfmacher et a!., 1991]. kilogram levels when the fungus is grown on com in the Because FAB is a matrix sensitive technique, deuterium laboratory. Fumonisin BI' the predominant fumonisin, is labeled fumonisin BI has been used as an internal standard known to cause animal diseases including Equine Leucoen­ for quantitative measurements of fumonisin levels [Plattner cephalomalacia and Porcine Pulmonary Edema and are po­ and Branham, 1994]. The latter method gives excellent data tent liver toxins in most animal species [Nelson et a!., 1993]. for fumonisin BI; however, and B3 have the Additionally they are suspected human carcinogens [Gelder­ blom et a!., 1991]. Since fumonisins were fIrst reported in 1988 [Gelderblom et al., 1988]. there has been considerable Address reprint requests to Ronald D. Plattner, Bioactive Constituents effort to develop methods to detect and measure their con­ Research, National Center for Agricultural Utilization Research, Midwest centrations in agricultural commodities and to study their Area, Agricultural Research Service, United States Department of Agricul­ toxicology. This research has recengtly been reviewed [Nel­ ture, Peoria, IL 61604. son et a!.. 1993; Norred, 1993; Riley et a!., 1993]. The most Names are necessary to report factually on available data; however, the widely used analytical method to measure fumonisins is USDA neither guarantees nor warrants the standard of the produd, and the HPLC with fluorescence detection of a suitable derivative of use of the name by USDA implies no approval of the produd to the exclusion of others that may also be suitable. All programs and services of the free amine group [Shephard et a!., 1990]. This proce­ the U.s. Department ofAgriculture are offered on anondiscriminatory basis dure cannot detect the N-acetylated fumonisin analogs (fu­ without regard to race, color, national origin, religion, sex, age, marital monisin AI and A2) which are sometimes present, albeit at status, or handicap. © 1995 Wiley-Liss, Inc. 'This article is a US Govemment work and, as such, is in the public domain in the United States of America. TOXINS PRODUCED BY FUSARIUM MONILIFORME 295 ,6::H0 OH OH Rt o 0 NHRJ RJ CH'li~

0HO 0

R1 R2 R3 R. Fumonisin 8 1 OH OH H CH3

Fumonisin 82 OH H H CH3

Fumonisin 83 H OH H CH3

Fumonisin 8. H H H CH 3 Fumonisin A 1 OH OH OAe CH 3 Fumonisin~ OH H OAe CH3 Fumonisin C OH 1 OH H H2 Fumonisin C. H H H H2

Fig. 1. Structures of known A, B, and ( series fumonisins and the new fumonisin ((4) described in this article.

same molecular weight and in FAB/MS spectra are detected with column I and flow rate of 0.5 ml min of mobile phase, together. This article presents the analysis of underivatized unsplit, and a tube temperature of 75°C. The solvent con­ fumonisins by HPLC with detection with two types ofdetec­ sisted of a binary gradient of water (solvent A) and acetoni­ tors - the evaporative light-scattering detector (ELSD) and trile with 0.1 % trifluoroacetic acid (TFA) (solvent B). Col­ electrospray MS. umn 1 resolved all major fumonisins well with a gradient from 35% solvent B to 50% solvent Bin 15 min. The HPLC was equipped with a 25 I-Llloop injector. Injections of 1-10 MATERIALS AND METHODS I-LI of aqueous samples of fumonisins (nominal concentra­

Fumonisin Bl' B2 , B3 , B4 , and C4 were purified from tions in the range of 0.1-50 I-Lg/ml) were made and the cultures of isolates of F. moniliforme grown on cracked resulting detector response was measured with a Spectra­ maize using published procedures [Desjardins et aI., 1992; Physics Model 4270 integrator. Nelson et aI., 1993]. Strains used in this study were from the Alternatively the HPLC was coupled to a Finnigan-MAT culture collection at the Fusarium Research Center at the TSQ 700 mass spectrometer via the Finnigan electrospray Pennsylvania State University or from the collection of Dr. interface (Finnigan MAT, San Jose, CA). For electrospray John Leslie (Kansas State University). Extracts of culture experiments, the eluting solvent was a binary gradient with materials were made as reported previously [Nelson et aI., solvent A being methanol-water-acetic acid (50:50:1) and 1991 ]. solvent B being methanol-acetic acid (l00: 1). With the col­ HPLC experiments were performed on a Spectra-Physics umn 1, good separations were obtained with a gradient from 8700 liquid chromatograph. Either a C- I8 column (3 J.L 30% solvent A to 100% solvent B in 20 min; while with particle size 4.5 mm x 3 cm; column I-Perkin Elmer, Oak column 2, best separations were obtained with a gradient Brook, IL) with flow of 0.5 mllmin or a base deactivated from 80% solvent A to 50% A/50% B in 20 min with a final C-8 column (5-1-L particle size 4.5 mm x 25 cm; column hold of 10 min. The total HPLC eluent (either 0.5 or 0.2 2-YMC, Wilmington, NC) with flow of 0.2 mIlmin were mIlmin) was introduced into the detector. Mass spectra were used. The Varex MK III evaporative light-scattering detec­ obtained by scanning from rn/z 350/950 in 1-3 sec (full tor (Allech Associates, Deerfield, IL) was used as a detector scans) or by selected ion monitoring. 296 PLATTNER

Cl <.0 C o Q. <.0 f c (/) ..J W

5000000

o o 2 468 10 Fumonisin 81 injected

Fig. 2. Evaporative light scattering detector response. Plot of peak area vs. amount of fumonisin 8, injected (50 ng/-10 J.l.g).

722.4 a. 100 o. 2.12

.0

80

40

20

400 soo aoo 700 aoo

Fig. 3. Electrospray positive ion mass spectrum of fumonisin 8 "

RESULTS AND DISCUSSION 10 ng injected. Thus with injection of I fJ.,l of the standard The response of fumonisins in the ELSD were not linear extract of culture material (0.2 g/ml acetonitrile/water ex­ with concentration. The best fit between amount injected traction solvent) the effective detection limit of about 50 and detector response was obtained using a second order fit. ppm. While this is adequate for analysis ofculture materials Figure 2 shows the response curve for across where the fumonisin levels are frequently in the I,000's of the range from 50 ng to 10,000 ng injected into the HPLC. ppm levels, some sample clean-up will be needed to allow

Similar response curves were obtained for fumonisins B2 injection of larger aliquots ofextract to achieve the sensitiv­ and B3. Agreement was obtained between the amount of ity required for practical detection of fumonisins in naturally fumonisins Bl' B2 , and B3 in culture extracts using this contaminated extracts. procedure and the method of Shephard [1990] using the Electrospray MS is an ideal technique to detect and mea­ OPA derivative was generally within experimental error. sure fumonisins. The protonated molecule (m/z 722) of fu­ The detection limit for fumonisins with the ELSD was about monisin B I is the base peak in the mass spectrum (Fig. 3) TOXINS PRODUCED BY FUSARIUM MONILIFORME 297

a

b

c

I i I I 8:20 18:40 25;00 33:20

Fig. 4. Electrospray MS chromatograms of Fusarium moni/iforme culture extracts. Trace a is M-3125; trace b is K-816; trace c is K-819. One microliter (0.2 milligram equivalent of culture) injected on to a (-8 column and eluted with methanol/water/acetic acid [gradient from 60/40/1 to 75/2511 in 20 min hold at final] at a flow rate of

0.2 mUmin. Approximate retention times: Fumonisin 81, 11 min; B3, 17 min; 82, 23 min; (4,28 min; 84, 29 min. and little fragmentation is observed. Occasionally small than about 1 f.Lg of individual fumonisins result in column amounts of sodiated or potassiated molecules are seen at overload and the peaks begin to broaden, elute earlier, and higher levels, as are significant doubly charged protonated the response becomes non-linear. salt ions. By increasing the offset voltage between capillary The use of stable isotope internal standards with this tube and focusing quadrupole in the electrospray interface, method gave results that are similar to those reported for collisionally induced decomposition occurs and the resulting FAB/MS [Plattner and Branham, 1994] (data not shown) spectra (data not shown) shown fragments analogous to those when the samples were directly injected without chromatog­ reported in MS/MS spectra of FAB generated fumonisin raphy. When the column was used, however, a difficulty protonated molecules [Korfmacher et al., 1991; Plattner and was noticed. Carryover from previous injections is some­ Branham, 1994]. A full scan (m/z 600-850 in 0.5 sec) times observed after larger injections (> about 100 ng). This detection limit for fumonisins is lower than 100 pg when effect could invalidate measurements made on a subsequent directly coupled from the injector to the interface (i.e., no injection with small amounts of fumonisins because it can column) with a flow rate of 0.1 ml/min. A linear response alter the observed ratios of labeled and unlabeled fumoni­ for the protonated molecule of fumonisin B 1 (m/z 722) and sins. This problem has implications for both external and amount injected is observed for nine injections over the internal standard based method to measure fumonisins by range of 0.5 ng to 250 ng with a correlation coefficient of HPLC and needs to be carefully addressed in formulating r = 0.99832. When the HPLC column was inserted be­ any method of quantitation. tween the injector and the interface and chromatography is The analysis of culture material extracts from different done on either a column 1 or 2, injections of 0.5 ng can strains ofF. moniliforme generally shows similar patterns of easily be detected. A similar linear response of area vs. fumonisin homologs and isomers. However, rare isolates amount offumonisin injected is observed. Injections greater have been identified that produce little or no fumonisin B 1 298 PLATTNER and unusual ratios of the other fumonisins. Several strains of absence of fumonisin B t. These fumonisins are somewhat F. moniliforme isolated from maize in Nepal which produce more difficult to purify in the presence of much larger no detectable fumonisins have been identified [Desjardins amounts of fumonisin B] found in "normal" strains of F. et a!., 1992]. Two strains with unusual fumonisin ratio's moniliforme. discovered during screening of several hundred mating pop­ ulation A strains ofF. moniliforme for the ability to produce fumonisins are shown in Figure 4b,c along with a normal REFERENCES pattern of fumonisins shown in Figure 4a. One of these Branham BE, Plattner RD (1993): Isolation and characterization of a new unusual strains (KSU 819) (Fig. 4b) accumulates virtually fumonisin from liquid cultures of Fusarium moniliforme. J Nat Prod 56: 1630-1633. no fumonisin B] or B , but has high levels of fumonisin B 2 3 Desjardins AE. Plattner RD. Shackelford DD. Leslie JF. Nelson PE (1992): and fumonisin B4 , along with a new fumonisin which was Heritability of fumonisin B, production in Gibberella jujikuroi mating called fumonisin C4 . Fumonisin C4 is the analog of fumoni­ population A. Appl Environ Microbiol 58:2799-2805. sin B4 but missing the C-l methyl group (Fig. 1) based on its Oelderblom WCA. Jaskiewicz K. Marasas WFO. Thiel PO. Horak RM. NMR, electrospray MS, and hydrolysis GClMS spectra. Vleggaar R. Kriek NPJ (1988): Fumonisins-novel with The electrospray MS of fumonisin C has a strong signal at cancer-promoting activity produced by Fusarium moniliforme. Appl En­ 4 viron Microbiol 54: 1806-181 1. mJz 676 in the positive ion mode and mJz 674 in the negative Oelderblom WCA. Kriek NPJ. Marasas WFO. Thiel PO (1991): ion mode. This corresponds to the protonated molecule and and carcinogenicity of the Fusarium moniliforme metabolite fumonisins carboxylate anion of a fumonisin with the molecular weight B, in rats. Carcinogenesis 12:1247-1251. KorfmacherWA. Chiarelli MP. Lay JOJr, BlomJ. Holcomb M. McManus of 674 which is 14 daltons less than that of fumonisin B4 • The hydrolysis product was a C-19 aminotrioi. The EI-MS KT (1991): Characterization ofthe fumonisin B,: comparison of thermospray. fast-atom bombardment and electrospray mass spec­ fragmentation pattern of the TFA derivative of the hydro­ trometry. Rapid Commun Mass Spec 5:463--468. lyzed backbone of fumonisin C4 shows a shift in the signal Marasas WFO. Nelson PE. Toussoun TA (1984): "Toxigenic Fusarium from cleavage between C-2 and C-3 from mJz 140 to m/z Species: Identity and Mycotoxicology." University Park. PA: Pennsyl­ 126 indicating that the missing methyl group was at the vania State University Press, p 216. amino end of the fumonisin backbone. The strong methyl Nelson PE. Plattner RD. Shackelford DD. Desjardins AE (1991): Produc­ tions of fumonisins by Fusarium moniliforme strains from various sub­ doublet at about 1.2 ppm in the proton NMR spectrum strates and geographic areas. Appl Environ Microbiol 57:2410-2412. observed in the A and B series fumonisins is missing in the Nelson PE. Desjardins AE. Plattner RD (1993): Fumonisins. mycotoxins proton NMR spectra as it is in the spectrum of fumonisin C t produced by Fusarium species: biology. chemistry and significance. [Branham and Plattner, 1993]; and a double doublet from Annu Rev Phytopathol 31:233-252. two protons on the amino bearing carbon atoms is observed Norred WP (1993): Fumonisins-mycotoxins produced by Fusarium mo­ niliforme. J Toxicol Environ Health 38:309-328. at 3.0 ppm instead of the more complex single proton mul­ Plattner RD. Branham BE (1994): Labeled fumonisins: production and use tiplet that is observed when the single proton on C-2 is offumonisin B, containing stable isotopes. JAOAC Int 77:525-532. coupled to the three methyl protons on C-l and the single Plattner RD. Norred WP. Bacon CWo Voss KA, Peterson R. Powell RO proton on C-3. The carbon NMR spectrum shows 33 carbon (1990): A method of detection of fumonisins in corn samples associated atoms compared to 34 observed in the NMR spectrum of with field cases of equine leukoencephalomalacia. Mycologia 82:698­ 702. fumonisin B4 . The other strain (KSU 817) (Fig. 4c) made Plattner RD. Weisleder D, Shackelford DD. Peterson R, Powell RO (1992): little fumonisin B] or B3 but accumulated high levels of A new fumonisin from solid cultures of Fusarium moniliforme. Myco­ fumonisins B2 , B4 , and C4 . Full details of the survey for pathologia 117: 17-22. strains with unusual fumonisin production capabilities and Riley RT. Norred WP. Bacon CW (1993): Fungal toxins in foods: recent the genetics of crosses of these naturally occurring variants concerns. Annu Rev Nutr 13:167-189. Shephard OS. Sydenham EW. Thiel PO. Oelderblom WCA (1990): Quan­ and their implication in fumonisin biosynthesis will be re­ titative determination of fumonisins B, and B2 by high pressure liquid ported separately. These strains provide a convenient source chromatography with fluorescence detection. J Liquid Chromatogr 13: for production of large amounts offumonisin B2 or B3 in the 2077-2087.