Distribution of aflatoxins in shelling and milling fractions of naturally contaminated rice.

The objective of this study was to determine the distribution of an economically important class of mycotoxins, the aflatoxins, in rice milling fractions. Rice plants grown under field production conditions are frequently infected with types of pathogenic fungi that produce toxic metabolites (mycotoxins). Paddy (seeds) rice from healthy plants in the field was collected and stored on a farm under humid, poorly ventilated conditions. Samples were milled into four fractions (hulls, brown rice, bran and white rice) and analysed for aflatoxins (B(1), B(2), G(1) and G(2)) using a validated method. Rice fractions from healthy plants, which contained low levels of aflatoxins (less than 1 µg kg(-1)), were used to determine the efficiency of the extraction method. Seeds stored under poor conditions were found to be contaminated with aflatoxins B(1) and B(2) as were the fractions. The sums of AFB(1) and AFB(2) in stored paddy rice, hulls, brown rice, bran and white rice were 141, 39, 158, 367 and 56 µg kg(-1), respectively. The ratio of aflatoxin B(1) and B(2) was about 10 : 1. AFG(1) and AFG(2) were less than 1 µg kg(-1). Thus, brown rice contained 92.9% of the aflatoxins in paddy rice, whereas white rice contained only 27.9%.

Comparison of cultural and analytical methods for determination of aflatoxin production by Mississippi Delta Aspergillus isolates.

This study compared cultural and analytical methods for detecting aflatoxin production by Aspergillus species. Aspergillus isolates were obtained from various Mississippi Delta crops (corn, peanut, rice, cotton) and soils. Most of the isolates (99%) were A. flavus and the remainder comprised A. parasiticus and A. nomius. The following three cultural methods were evaluated on potato dextrose agar: fluorescence (FL) on beta-cyclodextrin-containing media (CD), yellow pigment (YP) formation in mycelium and medium, and color change after ammonium hydroxide vapor exposure (AV). Aflatoxins in culture extracts were confirmed by thin-layer chromatography (TLC) and quantified by enzyme-linked immunosorbent assay (ELISA). Of the 517 isolates, 314 produced greater than 20 ng/g of total aflatoxin based on ELISA, and 180 produced greater than 10 000 ng/g of aflatoxin in the medium. Almost all the toxigenic isolates (97%) were confirmed by TLC as producers. Of the toxigenic isolates, as determined by ELISA, 93%, 73%, and 70% gave positive FL, YP, and AV responses, respectively. Of the 203 isolates producing less than 20 ng/g of aflatoxin, 20%, 6%, and 0% of respective FL, YP, and AV methods gave false-positive responses. The 9% false-positive results from TLC fall within this range. This study showed good agreement among all tested cultural methods. However, these cultural techniques did not detect aflatoxin in all cultures that were found to produce aflatoxins by ELISA, LC/MS, and TLC. The best results were obtained when the AV color change and CD fluorescence methods were used together, yielding an overall success rate comparable to TLC but without the need for chemical extraction and the time and expense of TLC.

Comparison of phytotoxicity and mammalian cytotoxicity of nontrichothecene mycotoxins.

The phytotoxicity and mammalian cytotoxicity of four nontrichothecene mycotoxins (apicidin, sambutoxin, wortmannin, HC-toxin) were compared. Phytotoxicity was evaluated in terms of electrolyte leakage, growth inhibition, and reduction in chlorophyll content. Based on the parameters evaluated, the relative order of phytotoxicity to duckweed (Lemna pausicostata L.) was wortmannin > HC-toxin > apicidin >> sambutoxin. A 48-hr exposure to 10 microM wortmannin, HC-toxin or apicidin caused electrolyte leakage from duckweed. The IC50 values for growth inhibition and chlorophyll reduction for wortmannin, HC-toxin, and apicidin were 0.2 and 2.6 microM, 15.4 and 12.6 microM, and 27.7 and 45.3 microM, respectively. Based on the parameters measured, a 72-hr exposure to 100 microM sambutoxin was not toxic to duckweed. Kudzu (Pueraria lobata L.) leaf disc assays revealed a similar trend in relative toxicities, but higher mycotoxin concentrations were required to elicit phytotoxic effects compared to duckweed. All four mycotoxins were cytotoxic to four mammalian cell cultures. However, in contrast to plants, wortmannin was the least toxic (IC50 = 10 to 20 microM) and sambutoxin exhibited a high level of toxicity (IC50 = 0.5 to 1 microM).

Phytotoxicity and ultrastructural effects of gymnopusin from the orchid Maxillaria densa on duckweed (Lemna pausicostata) frond and root tissues.

Two phenanthrene derivatives, characterized as erianthridin (9,10-dihydro-2,7-dihydroxy-3,4-dimethoxyphenanthrene) and gymnopusin (2,7-dihydroxy-3,4,9-trimethoxyphenanthrene), were isolated from an extract of the orchid Maxillaria densa, using phytotoxicity with amaranth (Amaranthus hypochondriacus) to guide fractionation. Gymnopusin and erianthridin inhibited radicle elongation of A. hypochondriacus seedlings with IC(50) values of 330 and 58.2 microM, respectively. The phytoxicity of the two phenanthrene derivatives was also assessed on duckweed (Lemna pausicostata), and compared with mammalian toxicity estimated in vitro with four mammalian cell lines. On duckweed, both phenanthrene derivatives caused electrolyte leakage, chlorophyll loss and photobleaching. Ultrastructural examination of duckweed frond and root tissues treated with gymnopusin (100 microM) revealed membrane damage to the tonoplast after 12 h of exposure. Effects on membrane integrity followed a time course similar to that of electrolyte leakage. Both phenanthrene derivatives exhibited moderate cytotoxicity to all mammalian cells tested, which precludes their use as a bioherbicide.

Phytotoxicity and mammalian cytotoxicity of macrocyclic trichothecene mycotoxins from Myrothecium verrucaria.

Macrocyclic trichothecene toxins produced by Myrothecium verrucaria (a phytopathogen of interest in biological weed control) and the non-trichothecene toxin atranone B from Stachybotiys atra were tested for phytotoxicity in duckweed (Lemna pausicostata L.) plantlet cultures and kudzu (Pueraria lobata L.) leaf disc assays, and for mammalian cytotoxicity in four cultured cell lines. Roridin E and H, epi-isororidin E, and verrucarin A and J were phytotoxic (half-maximal effect in the concentration range 0.1-9.7 microM on duckweed and 1.5->80 microM on kudzu) and cytotoxic to mammalian cell lines (half-maximal inhibition of proliferation in the concentration range 1-35 nM). Trichoverrins A and B and atranone B were moderately phytotoxic (half-maximal effect in the concentration range 1 9-69 microM on duckweed and 13->80 microM on kudzu) and weakly cytotoxic with mammalian cell lines (half-maximal inhibition of proliferation in the concentration range 0.3->2 microM).

Macrocyclic trichothecenes are undetectable in kudzu (Pueraria montana) plants treated with a high-producing isolate of Myrothecium verrucaria.

Myrothecium verrucaria was found to be an effective pathogen against kudzu grown in the greenhouse and the field. M. verrucaria produced large amounts of macrocyclic trichothecenes when cultured on solid rice medium, including epiroridin E (16.8 mg/g crude extract), epiisororidin E (1 mg/g), roridin E (8.7 mg/g), roridin H (31.3 mg/g), trichoverrin A (0.6 mg/g), trichoverrin B (0.1 mg/g), verrucarin A (37.4 mg/g), and verrucarin J (2.2 mg/g). Most of these toxins were also isolated from M. verrucaria spores and mycelia grown on potato dextrose agar medium, including epiroridin E (32.3 mg/g), epiisororidin E (28.6 mg/g), roridin E (0 mg/g), roridin H (60 mg/g), trichoverrin A (1.3 mg/g), trichoverrin B (1.8 mg/g), verrucarin A (13.8 mg/g), and verrucarin J (131 mg/g). When M. verrucaria was cultured on liquid media, the numbers but not the amounts of toxins decreased. Only epiroridin E (28.3 mg/g), epiisororidin E (29.6 mg/g), verrucarin B (195 mg/g) and verrucarin J (52.6 mg/g) were measured when the fungus was cultured on cornsteep medium. On soyflour-cornmeal broth M. verrucaria produced several toxins, including epiroridin E (58.1 mg/g), epiisororidin E (5.8 mg/g), verrucarin B (29.9 mg/g) and verrucarin J (32 mg/g). In contrast, no macrocyclic trichothecenes were detected by HPLC analysis of plant tissues of kudzu, sicklepod, and soybean treated with aqueous suspensions of M. verrucaria spores formulated with a surfactant. Chloroform-methanol extracts of kudzu leaves and stems treated with M. verrucaria spores were less cytotoxic to four cultured mammalian cell lines than the corresponding extracts from control plants. Purified macrocyclic trichothecenes (verrucarin A and T-2 toxin) were very cytotoxic to the same cell lines (< or = 2 ng/ml). These results show that neither intact macrocyclic trichothecenes nor toxic metabolites could be detected in plant tissues after treatment with M. verrucaria spores. These results argue for both safety and efficacy for the use of M. verrucaria in biological control of kudzu and other noxious weeds, and support proceeding to animal feeding trials for further evaluation of safety.

Structure-activity relationships for human estrogenic activity in zearalenone mycotoxins.

Zearalenones are estrogenic Fusarium mycotoxins consisting of a resorcinol moiety fused to a 14-member macrocyclic lactone. Using an improved MCF7 human breast cell proliferation assay, we have compared the estrogenicity of 17 chromatographically-homogeneous zearalenones. Both similarities and substantial differences from published results in intact animal systems were observed. Substantial human estrogenicity was retained even in analogs lacking hydroxylation on the aromatic and macrocyclic rings.

Antimalarial, antiviral, and antitoxoplasmosis norsesterterpene peroxide acids from the Red Sea sponge Diacarnus erythraeanus.

A new norsesterterpene acid, named muqubilone (1), along with the known sigmosceptrellin-B and muqubilin were isolated from the Red Sea sponge Diacarnus erythraeanus. The structure determination of 1 was based primarily on 1D and 2D NMR analyses. Sigmosceptrellin-B exhibits significant in vitro antimalarial activity against Plasmodium falciparum (D6 and W2 clones) with IC(50) values of 1200 and 3400 ng/mL, respectively. Muqubilin and 1 show in vitro antiviral activity against herpes simplex type 1 (HSV-1) with ED(50) values of 7.5 and 30 microg/mL, respectively. Muqubilin and sigmosceptrellin-B display potent in vitro activity against Toxoplasma gondii at a concentration of 0.1 microM without significant toxicity.

Fumonisin B1 from the fungus Fusarium moniliforme causes contact toxicity in plants: evidence from studies with biosynthetically labeled toxin.

Fumonisin B1 (FB1) is the most abundant of a series of sphingosine analog mycotoxins produced by the fungus Fusarium moniliforme, a ubiquitous contaminant of stored corn (maize) worldwide. FB1 exhibits a variety of biological activities including phytotoxicity, which is of particular interest for its potential role as a virulence factor to facilitate invasion of plant tissues by the fungus. Droplets of FB1 solution applied to the leaf surface of jimsonweed, black nightshade, and susceptible tomatoes caused necrosis, growth inhibition, and death. With Arabidopsis thaliana grown on agar plates, an IC50 (concentration causing half maximal phytotoxicity) of less than 1 ppm was observed. [3H]FB1 was prepared by biosynthetic incorporation of commercially-available radiolabeled presumptive precursors into the toxin in rice medium solid cultures of F. moniliforme JW#1. The labeled toxin produced by incorporation of [9,10(-3)H]palmitate induced phytotoxic symptoms identical to unlabeled material, indicating it had full biological activity. The area of necrosis on treated leaves was similar in light and dark treated plants. Using liquid scintillation counting to quantify radioactivity in excised plant parts, over 95% of the [3H]FB1 radioactivity applied to leaves of light or dark-treated plants was recovered from the treated leaf. When [3H]FB1 was applied to a wound site on target plants, severe damage occurred at the site of FB1 application and in tissue above the site. These results indicate that FB1 applied to intact surfaces of target plants exhibits primarily contact activity. Translocation of FB1 is limited, occurring only when FB1 is applied to a wound site, and it results in damage to tissue above the point of application, indicating that FB1 is xylem mobile.

Fumonisin B1 from the fungus Fusarium moniliforme causes contact toxicity in plants: evidence from studies with biosynthetically labeled toxin.

Fumonisin B1 (FB1) is the most abundant of a series of sphingosine analog mycotoxins produced by the fungus Fusarium moniliforme, a ubiquitous contaminant of stored corn (maize) worldwide. FB1 exhibits a variety of biological activities including phytotoxicity, which is of particular interest for its potential role as a virulence factor to facilitate invasion of plant tissues by the fungus. Droplets of FB1 solution applied to the leaf surface of jimsonweed, black nightshade, and susceptible tomatoes caused necrosis, growth inhibition, and death. With Arabidopsis thaliana grown on agar plates, an IC50 (concentration causing half maximal phytotoxicity) of less than 1 ppm was observed. [3H]FB1 was prepared by biosynthetic incorporation of commercially-available radiolabeled presumptive precursors into the toxin in rice medium solid cultures of F. moniliforme JW#1. The labeled toxin produced by incorporation of [9,10-3H]palmitate induced phytotoxic symptoms identical to unlabeled material, indicating it had full biological activity. The area of necrosis on treated leaves was similar in light and dark treated plants. Using liquid scintillation counting to quantify radioactivity in excised plant parts, over 95% of the [3H]FB1 radioactivity applied to leaves of light or dark-treated plants was recovered from the treated leaf. When [3H]FB1 was applied to a wound site on target plants, severe damage occurred at the site of FB1 application and in tissue above the site. These results indicate that FB1 applied to intact surfaces of target plants exhibits primarily contact activity. Translocation of FB1 is limited, occurring only when FB1 is applied to a wound site, and it results in damage to tissue above the point of application, indicating that FB1 is xylem mobile.

Inhibition of oncogene product enzyme activity as an approach to cancer chemoprevention. Tyrosine-specific protein kinase inhibition by purpurogallin from Quercus sp. nutgall.

Inhibitors of oncogene product enzyme activity were sought as a prescreen for potential cancer chemopreventive agents. Purpurogallin, a polyphenol from Quercus sp. nutgall, was found to inhibit the tyrosine-specific protein kinase of the human erb-b oncogene product (epidermal growth factor receptor) for both autophosphorylation (IC50 = 27.5 microM) and phosphorylation of an exogenous substrate (IC50 = 45.3 microM). An examination of enzyme kinetics indicated that purpurogallin is a competitive inhibitor of both ATP (Ki = 54.9 microM for autophosphorylation, Ki = 33.9 microM for phosphorylation of exogenous substrate) and the tyrosine-containing acceptor substrate poly(glutamate, alanine, tyrosine) 6:3:1 (Ki = 83.7 microM).

Mycotoxin production by Fusarium proliferatum isolates from rice with Fusarium sheath rot disease.

Twenty samples of unpolished (rough) rice collected in Arkansas and Texas during the 1995 harvesting season from fields exhibiting Fusarium sheath rot disease or panicle blight were previously shown to include 8 samples positive for fumonisin B1 (FB1) in the range 2.2-5.2 ppm, and moniliformin (MON), but no beauvericin (BEA), deoxynivalenol, its derivatives or zearalenone were detected. Fifteen cultures of F. proliferatum were established from the 20 rough rice samples. Single spore isolates of each culture were grown on rice and tested for the production of fumonisins (FB1, FB2, FB3, etc.), MON and BEA. All 15 isolates produced FB1, FB2, MON and BEA in culture on rice. No deoxynivalenol, its derivatives or zearalenone were detected. Seven cultures produced FB1 at > 50 ppm (range 80-230 ppm), with the rest producing FB1 in the range 14-43 ppm. FB2 was produced in the range 5-47 ppm, and those cultures which produced the most FB1 also produced the most FB2. Of the 15 cultures producing MON, 11 produced it at > 100 ppm in the range 188-6018 ppm, with the rest producing in the range 7-64 ppm. BEA was produced in the range 109-1350 ppm. Other derivatives of fumonisins, including FA1, FA2 and partially hydrolyzed FB1, as well as several unknown metabolites including a compound with MW 414, were identified in culture extracts by continuous flow fast atom bombardment with ion spray mass spectrometry (CF/FAB/MS). Further study is needed to identify the factors that control production of FB1, MON and BEA by F. proliferatum in culture and in field samples.

Ultrastructural effects of AAL-toxin TA from the fungus Alternaria alternata on black nightshade (Solanum nigrum L.) leaf discs and correlation with biochemical measures of toxicity.

Ultrastructural effects of AAL-toxin TA from Alternaria alternata on black nightshade (Solanum, nigrum L.) leaf discs and correlation with biochemical measures of toxicity. In black nightshade (Solanum nigrum L.) leaf discs floating in solutions of AAL-toxin TA (0.01-200 microM) under continuous light at 25 degrees C, electrolyte leakage, chlorophyll loss, autolysis, and photobleaching were observed within 24 h. Electrolyte leakage, measured by the conductivity increase in the culture medium, began after 12 h with 200 microM AAL-toxin T(A), but was observed after 24 h with 0.01 to 50 microM AAL-toxin T(A), when it ranged from 25%) to 63% of total releasable electrolytes, respectively. After 48 h incubation, leakage ranged from 39% to 79% of total for 0.01 to 200 microM AAL-toxin T(A), respectively, while chlorophyll loss ranged from 5% to 32% of total, respectively. Ultrastructural examination of black night-shade leaf discs floating in 10 microM AAL-toxin TA under continuous light at 25 degrees C revealed cytological damage beginning at 30 h, consistent with the time electrolyte leakage and chlorophyll reduction were observed. After 30 h incubation chloroplast starch grains were enlarged in control leaf discs, but not in AAL-toxin T(A)-treated discs, and the thylakoids of treated tissue contained structural abnormalities. After 36-48 h incubation with 10 microM AAL-toxin T(A), all tissues were destroyed with only cell walls, starch grains, and thylakoid fragments remaining. Toxicity was light-dependent, because leaf discs incubated with AAL-toxin T(A) in darkness for up to 72 h showed little phytotoxic damage. Within 6 h of exposure to > or =0.5 microM toxin, phytosphingosine and sphinganine in black nightshade leaf discs increased markedly, and continued to increase up to 24 h exposure. Thus, phy siological and ultrastructural changes occurred in parallel with disruption of sphingolipid synthesis, consistent with the hypothesis that AAL-toxin T(A) causes phytotoxicity by interrupting sphingolipid biosynthesis, thereby damaging cellular membranes.

Phytotoxicity and cytotoxicity of the fumonisin C and P series of mycotoxins from Fusarium spp. fungi.

Fumonisin C (FC) and P (FP) are two recently identified series of sphingosine-analog mycotoxins, for which biological activities have not previously been reported. FC1, FC2 and OH-FC1 (1 microM) exhibited strong phytotoxicity comparable to the standard FB1 in duckweed (Lemna pausicotata L.) cultures, whereas FC3 and FC4 were moderately phytotoxic. Conversely, FP1 exhibited weak phytotoxicity only at higher concentrations (> or =10 microM). These mycotoxins exhibited a similar pattern of cytotoxicity with FB1-sensitive cultured mammalian cell lines, H4TG and MDCK.

Cytotoxicity of fumonisin B1: implication of lipid peroxidation and inhibition of protein and DNA syntheses.

The effects of fumonisin B1 (FB1) from Fusarium moniliforme on lipid peroxidation and protein and DNA syntheses were studied in monkey kidney cells (Vero cells). FB1 was found to be a potent inducer of malondialdehyde (MDA), one of the secondary products formed during lipid peroxidation. At 0.14 microM (0.1 microg/ml), FB1 induced 0.496 +/- 0.1 nmoles of MDA/ mg protein, compared to the control level 0.134 +/- 0.01 nmoles of MDA/mg protein (P < 0.005). No inhibition of protein or DNA synthesis was observed at this concentration of FB1. Inhibition of protein and DNA syntheses was observed at FB1 concentrations > 14 microM (10 microg/ml) with an IC50 of 33 microM for both protein synthesis and DNA synthesis. These results indicate that lipid peroxidation is a very sensitive cellular response to the mycotoxin fumonisin B1 observed at concentrations lower than that required to inhibit cellular synthesis of macromolecules, protein and DNA. This oxidative damage induced by FB1 concentrations encountered in naturally contaminated foodstuffs and feed might lead to mutagenicity and genotoxicity.

Natural Occurrence of Fumonisins in Rice with Fusarium Sheath Rot Disease.

Twenty samples of rough rice (Oryza sativa) (unpolished kernels) collected during the 1995 harvest season from Arkansas (seven samples) and Texas (13 samples) were obtained from rice fields known to include plants with symptoms of Fusarium sheath rot putatively caused by Fusarium proliferatum. Samples were analyzed for fumonisin B1 (FB1) at three laboratories using three different extracting solvents by high-performance liquid chromatography (HPLC) or enzyme-linked immunosorbent assay (ELISA) methods. Forty percent of the samples were positive for FB1 at levels ≤4.3 µg/g by HPLC. The same samples contained FB1 at ≤3.6 µg/g when measured by an ELISA method. Most samples that were positive for FB1 were positive for fumonisin B2 (FB2) and fumonisin B3 (FB3) by HPLC at levels ≤1.2 µg/g. Very good agreement was obtained among the two laboratories using HPLC methods and the third using ELISA. Shelling of the unpolished rice results in hull and brown rice fractions. In a sample that contained 4.3 µg/g in whole kernels, the fumonisin level was very high in hulls (≤16.8 µg/g) and low in brown rice (≤0.9 µg/g). Milling of brown rice results in bran and white rice fractions. Fumonisins were found in bran at a level of ≤3.7 µg/g but were below the level of detection by HPLC in white rice. The presence of fumonisins (FB1, FB2, and FB3) was confirmed by fast atom bombardment/mass spectrometry. This is the first report of fumonisins in naturally contaminated rice in the United States.

Immunohistochemical localization of insulin-degrading enzyme along the rat intestine, in the human colon adenocarcinoma cell line (Caco-2), and in human ileum.

Insulin-degrading enzyme (IDE) has been implicated in the intracellular degradation of insulin in insulin target cells. Knowledge of the existence of this enzyme in the intestine will be beneficial to the achievement of clinical oral efficacy of insulin. A comparative study was conducted with rat intestine, human colon adenocarcinoma (Caco-2) cells, and human ileum. Confocal microscopy analysis using the anti-IDE antibody showed that IDE was localized in the mucosal cells of rat and human intestines, as well as in Caco-2 cells. Immunostaining of this enzyme was homogeneous throughout the cell excluding nucleus, indicating a typical cytosolic distribution in rat and human enterocytes and in Caco-2 cells.

Activation of mitogen-activated protein kinase by the carcinogenic mycotoxin fumonisin B1.

Recent studies indicate that the mycotoxin and putative carcinogen fumonisin B1 stimulates mitogenesis in Swiss 3T3 fibroblasts. Because mitogen-activated protein kinase (MAPK) is a key enzyme in the signal transduction pathways activated by many mitogens, we hypothesized that fumonisin B1 might modulate MAPK activity. Consistent with other reports, we observed that fumonisin B1 is synergistic with insulin for mitogenesis in Swiss 3T3 cells. We also observed that: (1) fumonisin B1 stimulates rapid, transient activation of MAPK; and (2) the dose-response curve for fumonisin B1-stimulated activation of MAPK is similar to the dose-response curve for the mitogenic effects of fumonisin B1. In contrast to fumonisin B1, insulin stimulates minimal activation of MAPK in Swiss 3T3 cells. The observation that fumonisin B1 and insulin differ with respect to the modulation of MAPK activity suggests a possible mechanism for their synergistic stimulation of mitogenesis in Swiss 3T3 cells.

Biological activities of synthetic analogues of Alternaria alternata toxin (AAL-toxin) and fumonisin in plant and mammalian cell cultures.

In a search for an analogue of AAL-toxin with high phytotoxicity and low mammalian toxicity, aminopentols [(AP1), hexacetyl AP1 and N-acetyl AP1], and nine analogues (1-9), were tested for toxicity to duckweed (Lemna pausicostata), susceptible tomato (asc/asc) leaf discs, black nightshade leaf discs and mammalian cell lines, including dog kidney (MDCK), rat liver hepatoma (H4TG) and mouse fibroblasts (NIH3T3). These were compared with AAL-toxin and fumonisin B1 (FB1). Analogue 9 at 10 microM increased cellular leakage and chlorophyll loss from both tomato and black nightshade leaf discs. The diester 9 was the most active in the duckweed bioassay, but it was much less toxic to MDCK and H4TG cells with an IC50 of 200 microM compared to 10 microM for FB1. Analogue 9 and FB1 showed similar low toxicities (IC50 = 150 microM) to NIH3T3 cells. Among the substances tested, only analogue 9 had significant phytotoxicity and low mammalian toxicity, indicating some potential for development of safe and effective natural herbicides.