Chem. Res. Toxicol..1993,6, 97-101
97
Fusarin C: Isolation and Identification of Two Microsomal Metabolites Bing Zhut and Alan M. Jeffrey*ttJ Division of Environmental Sciences, Comprehensive Cancer Center, and Department of Pharmacology, Columbia University, New York,New York 10032 Received September 8,1992
Fusarin C (FC),a metabolite formed by Fusarium moniliforme,is, with microsomal activation, genotoxic. Two metabolites of FC, fusarin Z (FZ) and fusarin X (FX), have been isolated from an in vitro metabolic activation system using microsomal mixtures from phenobarbital-induced rat livers and their chemical structures identified. FZ and F X are 500 and 60 times more mutagenic than FC in the Ames test, respectively. Both result from hydroxylation a t the l-position of FC. FZ is a y-lactone involving isomerization of the 2,3-double bond and intramolecular transesterification between the 21-methyl ester and the newly formed l-hydroxyl group. FX, however, could not be converted to FZ by the microsomal system.
Introduction
20 21 COQCHI
Epidemiological evidence strongly suggests that environmental agents, many of which are likely to be found in the diet, are a major factor in cancer etiology (I). Fusarium moniliforme occurs in a wide variety of plant hosta and is one of the most prevalent fungi found on corn. A possible association with human esophageal cancer and consumption of corn contaminated with this fungus has been suggested, especially in the areas of Linxian county, China, and Southern Africa (2-4). Although fusarin C (FC)' contains an epoxide ring (5,6) (Figure I), it still requires enzymic activation to obtain maximal mutagenic activity (3, 7-9). A metabolite of FC, PM1, lacking the 21-methyl ester was much less mutagenic than FC (10).Inhibition of the esterase activity in microsomal mixtures which produce PM1 doubled the mutagenicity of FC (11).Neither study identified the active metabolite(8). The former reported water-soluble and relatively stable mutagenic metabolites, although attempts at their isolation were unsuccessful. We report here the isolation and identification of two active metabolites of FC, fusarin Z (FZ)and fusarin X (FX), and discuss their mechanism of formation.
Materials and Methods Caution: The extracts from F. moniliforme are mutagenic and carcinogenic in rodents. They should be handled accordingly. All experiments were performed under gold lights or in the dark. FC was isolated as previouslydescribed (12). Microsomes were prepared from phenobarbital-induced BDIX rats (7). Mutagenicity was assayed in the Ames test using Salmonella typhimurium strain TA 100,without adding microsomes (9,13) but with preincubation. Metabolic Activation of FC. FC was metabolicallyactivated as previouslydescribed (7)with slight modifications: diieopropyl fluorophosphate (25pL, 1 mM) was addedto 10mL of microsomal mixture to inhibit esterase activity (11). Efforts were made to
* To whom correspondence should be addressed.
Fax (212) 305-5328.
Tel(212) 305-6925;
t Division of Environmental Sciences and Comprehensive Cancer Center. t Department of Pharmacology. Abbreviations: FC, fusarin C; FX, fusarin X; FZ, fusarin Z.
'OH
0x1
yo\.
fusarin c
Figure 1. Structures of FC and ita metabolites, FZ and FX. optimize production of the active metabolite(s). The effects of concentrations of FC and the length of incubation time were examined. Samples of FC (3,4,and 6 pg) were incubated with 0.5-mL aliquots of microsomal mixture for 30 min or 4 pg of FC in 0.5 mL of microsomes for 5-60 min. After incubation, the microsomal mixtures were extracted twice with 1 mL of CHCL, the solvent was evaporated, and the residue was redissolved in 10 pL of sterile MezSO for mutagenicity assays. HPLC Analyses of Active Metabolites. FC (80 pg) was incubated in 10 mL of microsomes at 37 "C for 30 min before extraction twice with 10 mL of CHCb. The organic phases were combined and evaporated, and the residue was dissolved in 20 pL of methanol for HPLC analysis. Reverse-phase HPLC analyses were performed on an 850 Dupont Instrument (Du Pont Co, Wilmington, DE) equipped with an LKB (Pharmacia-LKB, Piscataway, NJ) 2140 rapid spectral detector. A 300- X 3.9." C-18Bondclone column (Phenomenex, Torrance, CA) was ueed with methanol in sodium phosphate buffer (10 mM, pH 7.4) as the mobile phase (45-75% over 40 min and a flow rate of 1mL/ min). Fractions were collected, the methanol evaporated, and the residual aqueous solution extracted three times with CHCL. The combined CHCb extracts were evaporated and redissolved in 40pL of MezSO of which 4pL/plate was used for mutagenicity assays. Isolation of FC Active Metabolites. FC (3 mg) waa incubated with 300 mL of microsomalmixture in a 6OO-mLflask for 30 min, with shaking at 150 rpm. The incubation mixture was then extracted twice with equal volumes of CHCL. The solvent was evaporated and the residue dissolvedin 2 mL of 30% methanol in water. The solution was fractionated on &p-Pak C-18 cartridges using 10 mL of 40% methanovwater and 10 mL of 50% methanol/water sequentially. The 50% methanovwater fraction was further separated by reverse HPLC as described above. Two compounds, eluting at 15.6 and 17.4 min, were collected separately, and most of the methanol was evaporated before extraction twice with equal volumes of CHCk. $3 1993 American Chemical Society
98 Chem. Res. Toxicol., Vol. 6, No. 1, 1993
Zhu and Jeffrey
Table I. Mutagenicity of Extracts of Microsomal Incubations in S.t y d h w r i r r m TA 100 (A) Effect of Microsome and FC Concentrations’ concentrations histidine revertanta/plateb control 103 f 6 123 f 4 control + MezSO 101 f 4 113 f 9
,
650
,
1.0
I
~
3 pg of FC/0.5 mL of microsomes 4 pg of FC/0.5 mL of microsomes 6 pg of FC/0.5 mL of microsomes 4 pg of FC/1.0 mL of microsomes
284f9 435 f 6 459 f 29 448 f 3
246 f 5 405 f 15 408 f 29 435 f 5
550
2
5
0.8
450
0
3
- 0.6
350
3c
-
i
~~
a
8
- 0.4
w
250
L
(B) Effect of Incubation Time‘ time (min)
control
control + MezSO 5 10 20 30 40 60
histidine 109 f 11 118 f 7 114 f 6 151 f 16 283 f 16 467 f 37 348 f 17 187 f 12
revertantslplateb 115 f 12 111 f 4 108 f 16 198 f 21 305 f 18 421 f 25 326 & 10 146 f 6
4 Metabolic activation of FC was performed at 37 O C for 30 min. Experiments wed duplicated plates (4 p L of MezSO solution/plate). All results were repeated once withindependent extra& and cultures.
b
4 pg of FC waa incubated with 0.5 mL of microsomes, at 37
10
5
0
25
20
15
Time lminl
Figure 2. HPLC analysis of metabolites of FC. A microsomal incubation extract was analyzed by reverse-phase HPLC with a gradient from 45 % to 75 % over 40 min. The HPLC eluate was fractionally collected, the methanol was evaporated, and the aqueous residues were extracted with CHCb. The CHCb solutions were evaporated and the residues dissolved in MeBO for mutagenicity assay (without microsomes). The compounds eluting at 15.6 and 17.4 min were FZ and FX, respectively.
OC.
The HPLC behavior of the microsomal metabolites was analyzed by coinjection with FX ( I d ) , a direct-acting mutagen isolated from cultures of F. moniliforme strain used to produced FC. Metabolism of [21-”C]FC. Samples containing 0.01 pCi of [21-14C]FC(15)and 400pg of FC in 40 mL of microsomal mixture were incubated for 30 min. The metabolites were extracted and analyzed by HPLC as described above. The eluate from 15to 18 min was collected,concentrated, and reanalyzed by HPLC using agradientfrom30%to50% methanolover60min,collectingl-min fractions to measure radioactivity. NMRand Mass Spectra. To identifythe structuresof active metabolites of FC, NMR and mass spectra were recorded with 400-MHz Bruker and Nermag 10-10 instruments, respectively.
Results We initially optimized conditions to produce the active metabolites. With CHCl3 extraction, all the mutagenic activity was found in the organic phase (data not shown). The best incubation conditions were 4 pg of FC per 0.5 mL of microsomal mixture, a t 37 “C, for 30 min (Table I). Increasing the concentration of FC or microsomes did not significantly increase the production of active metabolites. Decreasing the concentration of FC reduced yields. The length of incubation was critical: more or less than 30 min reduced yields significantly (Table I). To purify the active metabolites of FC, CHCl3 extracts were analyzed by reverse-phase HPLC and fractions collected for analysis of mutagenicity. Compounds eluting at 15.6 and 17.4 min comprised almost all the mutagenic activity (Figure 2) and >90% of mutagenicity of the original sample before HPLC. Later fractions contained no mutagenic activity (data not shown). To isolate FZ and FX and identify their chemical structures, a large-scale metabolic experiment was performed. The CHC13 extracts were first fractionated on Sep-Pak cartridges and then purified by HPLC. Both FZ and FX have absorption maxima at 358 nm. Assuming the same molar extinctions coefficients as FC (5),from 3 mg FC, we obtained 40 pg of FX and 200 pg of FZ. The chemical structure of FX was determined from ita NMR spectrum which, owing to the limited quantities available, wan not as clear as that obtained from the
1
0.8
1
n d
x 0.2
e
0.0 L 10
,
,
1
I
I
12
14
16
18
20
Time ,mini
Figure 3. HPLC identification of FX. A sample of FX (17.4 min) was prepared by microsomal oxidation of FC. This sample containedsome FZ (15.8min) when reanalyzed by HPLC (lower chromatogram, gradient from 45% to 75% methanol over 40 min). It was unresolved from standard FX prepared from F. moniliforme culture extracts (14) when mixed and coinjected using the same HPLC condition (upper chromatogram). compound isolated directly from the fungus, but the same characteristic signals were present. The long-range couplings and nuclear Overhauser effect studies used to determine the EIZ isomers in FC (5) could not be undertaken with the quantities of metabolite available. However, on the basis of those studies, such isomerization should have resulted in noticeable differences in the chemical shifts of the vinyl protons, which we did not see between the microsomal and fungal isolates of FX. Analysis by coinjection HPLC of two samples of FX, one prepared by microsomal oxidation of FC and the other by direct isolation from cultures of F. moniliforme (14), showed the two samples to be indistinguishable (Figure 3). The UV spectra of the two samples were also identical, at 358 nm. with, A The NMR and mass spectra of FZ are given in Tables I1 and 111, respectively. These suggested the 21-methyl group was missing, which was confirmed by isolation of the same metabolite using [21J4C1FC. Resulta (Figure 4) showed that FZ, eluting at 53.7 min, contained no radioactivity while the FX (58.5min) had the same specific activity as the FC used in the microsomal incubations. A comparison of the mutagenicities of FC, FX, and FC was difficult since they differed so widely in their reaponma.
Chem. Res. Toxicol., Vol. 6, No.1, 1993 99
Fusarin C Microsomal Metabolites Table 11. NMR Spectrum of FZ in CDtClS
FCo
FZ Droton
6 (mm)
1 2 4 6 8 9 10 14 18
4.89 7.33 6.20 6.34 6.79 6.71 7.49 4.04 -2.04
19
4.04 3.90
21 22 23 24
1.96 2.11 2.09
J (Hz)
6 (mm) 4.61b 6.96 6.07 6.30 6.79 6.67 7.49 4.06 2.06 2.11 4.05 3.90 3.74 1.73 2.09 1.98
~~
0 0
14 14/10 10 0 11/8/4 111614
J (Hz)
6.2b 7.2 15 15/11 11 2.1 151614 15/8/4 11/8/4 111614
Data from ref 5. FX diacetate is used for this comparison. Other lactones show similar chemical shifts: 4.77 (broaddoublet) and 7.12 ppm (triplet),J = 2 Hz (16);4.9 and 6.51 ppm, J = 0 Hz (17);4.85 and 7.51 ppm, J = 1.8 (18);4.86 and 7.26 (broad singlets) and 6.26 for the vinyl equivalent to C4 (19);4.76,4.80,4.80,and 7.18-7.35 for (20).The butyl,hexyl, and octyl2-(l-hydroxyalkyl)-2-buten-4-olides values for the protons on C-1of FC were 1.77 ppm and 7.2 Hz. 110
loo 90
3V
I
9
80 0.4
70
n 4
60
0.2
x
2
50 0.0
40
45
55
50
60
Time lminl
Figure 4. HPLC analysis of metabolites of [21-l4C1FC. Metabolites,prepared by incubation of [Pl-WIFCwith microsomes, were analyzed by reverse HPLC using a gradient from 30% to 50% methanol over 60 min. The eluate was collected as 1-min fractionsand radioactivitymeasured. The retentiontimes of FZ and FX were 53.8 and 57.6 min, respectively. The residual FC eluted much later under these conditions and was not collected.
From the data in Table IV, we calculated the relative doses needed to produce an equivalent increase above background. FX and FZ are approximately 60 and 500 times more mutagenic than FC.
Discussion The first indication that it might be possible to isolate the active metabolite(s) of FC was by Gelderblom et al. (10). They reported that these metabolites were water soluble since all the mutagenic activity was associated with the supernatant following a lOOOOOg centrifugation for 1 h. Initially, the active metabolites were fractionated using (2-18Sep-Pak cartridges. However, they were unsuccessful in isolating the active metabolites by HPLC using silica gel columns. Extending their observation that the metabolites were stable to reverse-phase chromatography, we have isolated two compounds designated FX and FZ. Their UV spectra suggested they retained the polyene structure of FC. FX was indistinguishable chromatographically, or from ita NMR spectrum, from a metabolite isolated directly from cultures of the F. moniliforme (14).
The second metabolite, FZ, has not yet been identified in fungal extracts. However, its NMR spectrum (Table 11) was similar to that of FX and FC with a few notable exceptions. The characteristic 21-methyl group signal w a ~ missing. The protons on C1 were shifted downfield and appeared similar to those in FX-diacetate (15) and those reported for similar lactones (16-20). We were unable to see the small, broadened coupling present in some, but not all, y-lactones. The (2-18protons were partly obscured by the 22- and 24-methyl groups and, particularly owing to the complex ABXz pattern they formed, exact chemical shifts and coupling constants could not be determined. The NMR spectrum is, however, completely consistent with the proposed structure for FZ (Figure 1). The mass spectrum of FZ (Table 111) showed a weak molecular ion and considerable fragmentation with cleavage of the polyene chain and loss of the lactone moiety. This too is consistent with the proposed structure. Confirmation of the loss of the methyl group in FZ was obtained using [21-l4C1FC. While the FX, as expected, retained the same specific activity as FC, the FZ contained no detectable radioactivity (Figure 4). Having established the structure of FZ, we were surprised at its HPLC behavior, eluting much earlier than FC but similar to FX: the retention times of FZ, FX, and FC were 15,17, and 30 min (ODs column, Du Pont Co., Wilmington, DE, 25 cm X 1cm, 4575% methanolin water, over 40 min). However, comparing these retention times with those of simple analogs (furanone, 8.5 min; crotyl alcohol, 3.2 min; and methyl crotonate, 35 min; the same ODS column eluted with water), the much shorter retention times of FX and FZ compared to FC are reasonable. Formation of FZ does not arise through FX. Incubation of the latter with microsomes did not yield any FZ. We propose, therefore, that, during the hydroxylation of FC, trans to cis isomerization of the 2-3 bond occurs which enables the 1-hydroxyl group to exchange with the 21methyl ester to produce the y-lactone. Numerous isomerizations have been observed during P450-catalyzed hydroxylations, but none, to our knowledge, involves cistrans isomerizations. The related methane monooxygenase from Methylococcus capsulatus oxidizes cis- and tranebub2-ene without isomerization (21). Enzymes for isomerizations of specific compounds are known, for example, maleylacetonecis-trans isomerase (22). We do not know whether the proposed isomerization of FC during the formation of FZ is an enzymic reaction. With its high mutagenic activity, FX is comparable stable: it could be extracted form corn cultures and purified through several chromatographic steps. In comparison to FX and FZ, FC shows little or no mutagenic activity without metabolic activation. With activation, on a molar basis, FC is almost as active as FZ despite the conversion of the former to the latter being
