Factors that affect the occurrence of fumonisin.

The two important Fusarium ear rots of corn, Gibberella ear rot (Fusarium graminearum, formally F. moniliforme and allied species) and Fusarium ear rot (F. verticillioides and allied species) grow under different environmental conditions. F. graminearum grows well only between 26 and 28 degrees C and requires rain both at silking and during disease progression. F. verticillioides grows well at higher temperatures, and ear rot and fumonisin accumulation are associated with drought and insect stress and growing hybrids outside their areas of adaptation. In southern Transkei, where esophageal cancer has been associated with the consumption of F. verticillioides and fumonisin-contaminated corn, environmental conditions favor this fungus in most years. In the nearby areas where the soils, crops, food consumption, and populations are the same and where esophageal cancer is low, temperatures are cooler and F. graminearum is favored. Although F. verticillioides is associated with a disease of corn, it may be that this fungus is a mutualistic endophyte of the plant. Perhaps because of this, breeding for resistance to Fusarium ear rot has produced inconclusive results to date. The best available strategies for reducing the risk of fumonisin contents of maize are to ensure that hybrids are adapted to the environment and to limit drought stress and insect herbivory. It may also be necessary to make use of alternative strategies such as producing hybrids that contain enzymes to degrade fumonisin as it is produced.


Factors Affecting Fumonisin Production
Fumonisin (FB 1 ) is a mycotoxin produced by Fusarium verticillioides and some related species. Mycotoxins, produced by filamentous fungi, are chemicals that affect human and animal health. By convention, this excludes mushroom poisons. Most mycotoxins, including FB 1 , are secondary metabolites of the fungi concerned, that is, compounds produced after one or more nutrients become limiting (1)(2)(3). Secondary metabolites are produced from one or more primary metabolites. Primary metabolites are compounds produced after nutrients exterior to the cell are absorbed or transported inside, catabolized, then metabolized to compounds such as acetate, amino acids, or glutamic acid. Secondary metabolites are not produced from reacting preexisting compounds in or on the substrate. The occurrence of secondary metabolites from fungi is governed entirely by the existence of conditions that favor the growth of the fungus concerned, which will be the principal focus of this review.
Fungal mycelia are threadlike structures with a 3-to 4-µm diameter that make up the fungal colony for most taxa. With exceptions, only the terminal few cells of a mycelium are biologically active. It is unambiguously known that mycotoxins are produced by the penultimate cells of each mycelium as it grows. This means that the production of a mycotoxin occurs on the micro level of each individual threadlike structure. The accumulation (release) of mycotoxin occurs as the result of millions of cells on the crop and is a function of total fungal biomass rather than of time (4). That is, mycotoxins are produced within hours of germination, and their detection is a function of their production by millions/billions of cells. In stirred jar fermentations, this same process takes place, except that billions of producing cells are grown more or less in a synchronous fashion. The occurrence of mycotoxins in crops is governed entirely by the existence of conditions that favor the growth of the fungi concerned. Under environmental conditions, different fungal species are favored as diseases of crop plants or as saprophytes on stored crops (5). When conditions favor the growth of toxigenic species, it is an invariable and unfortunate rule that one or more of the compounds for which the fungus has the genetic potential are produced.
In vitro, FB 1 is produced under conditions known to that favor the production of polyketides and sesquiterpenes (3). The toxin was optimally produced in media that has moderate water activity and is nitrogen limited. Fumonisin production doubled roughly every 48 hr as long as mycelial dry weight increased. The ratio of FB 1 to FB 2 significantly increased from approximately 4.2 to 5.4 during the fermentation. This reflects the biosynthesis of FB 1 and FB 2 backbones from independent polyketide backbones. Fumonisin is produced under relatively high oxygen tensions but has an unusual apparent requirement for low pH (approximately 2) for optimal production.
The F. graminearum polyketide metabolite zearalenone also has an apparent requirement for high oxygen tension (6). Zearalenone is typically accumulated in corn only in the fall, sometimes after the crop has died, when oxygen tensions are higher than in living plants and the environment is cool. (The solubility of oxygen in water is increased in colder versus warmer water.) In contrast, the F. graminearum metabolite deoxynivalenol is produced in vitro under conditions of low oxygen tension (7). Deoxynivalenol (also a phytotoxin) is seen in corn kernels concurrent with development of the infection when the plant is living and little zearalenone is being produced (8). Oxygen is tightly scavenged in photosynthesizing plants, hence oxygen tensions are, as noted, low in living plant tissue. This physiologic evidence suggests that fumonisin is optimally produced in nonliving tissue.
The unusual requirement for a low pH for optimal biosynthesis of fumonisins may also provide some insight into the plantfungus relationship. The pH of well-rotted corn is low because of organic acids produced by metabolism of the starch. Fumonisins bind to various cations at neutral to alkaine pH values that affect their biologic activity (9).
Stable isotope labeling of fumonisin has shown that the backbones are of polyketide origin. The methyl groups at C12 and C16 are derived from methionine; the carboxylic acid side chains are derived from glutamic acid; and the amino group is derived from serine [(10); Figure 1)]. Polyketidases have been cloned from F. verticillioides as the first step in finding the genes for this metabolite (11). This is among the more complicated biosynthetic schemes in the known array of Fusarium metabolites.
The biology of fumonisin production as gleaned from fermentation studies suggests that fumonisin is produced in senescent corn tissue and, further, that fumonisin per se would not be particularly stable or biologically active under the conditions of growing corn tissue.
The two important Fusarium ear rots of corn, Gibberella ear rot (Fusarium graminearum, formally F. moniliforme and allied species) and Fusarium ear rot (F. verticillioides and allied species) grow under different environmental conditions. F. graminearum grows well only between 26 and 28°C and requires rain both at silking and during disease progression. F. verticillioides grows well at higher temperatures, and ear rot and fumonisin accumulation are associated with drought and insect stress and growing hybrids outside their areas of adaptation. In southern Transkei, where esophageal cancer has been associated with the consumption of F. verticillioides and fumonisin-contaminated corn, environmental conditions favor this fungus in most years. In the nearby areas where the soils, crops, food consumption, and populations are the same and where esophageal cancer is low, temperatures are cooler and F. graminearum is favored. Although F. verticillioides is associated with a disease of corn, it may be that this fungus is a mutualistic endophyte of the plant. Perhaps because of this, breeding for resistance to Fusarium ear rot has produced inconclusive results to date. The best available strategies for reducing the risk of fumonisin contents of maize are to ensure that hybrids are adapted to the environment and to limit drought stress and insect herbivory. It may also be necessary to make use of alternative strategies such as producing hybrids that contain enzymes to degrade fumonisin as it is produced.

The Relationship of F. verticillioides and Allied Species to Zea mais
F. verticillioides and F. proliferatum are the most common fungi associated with maize. For many years, F. verticillioides has been known to occur systemically in leaves, stems, roots, and kernels (12). These fungi can be recovered from virtually all corn kernels worldwide including those that are healthy (13)(14)(15)(16)(17). Some data suggest that the relationship between the fungus and maize is mutualistic, with the fungus producing metabolites such as fusaric acid and gibberellins that are beneficial to the plant (17). F. verticillioides has been reported to suppress the growth of other ear fungi (18). Studies that report differences in recoveries from corn kernels are based mostly on plating of surface-disinfected kernels. This is an insensitive method and detects actively growing mycelia in proportion to its biomass. In other words, the few F. verticillioides cells in the tip cap of the kernel are often not detected. Detection is roughly proportional to the number of living cells in kernels. Even where kernels have been inoculated, it appears that the inoculated strain may not prevail (19). The wound alone is causing damage to the ear, allowing the endemic strain to proliferate.
This might place F. verticillioides as an endophyte of corn. Strictly defined, fungal endophytes are symptomless "infections" of (typically) leaves of plants that confer some ecologic benefit to the plant (20). It is common for endophytes to produce disease symptoms in the affected plants as they become senescent. Mutualistic endophytes occur in seaweeds, grasses, flowing plants, conifers, and deciduous trees. In these cases, the primary known benefit is to limit herbivory.
The best understood endophytes are grass endophytes. The association between Balansia endophytes of various grass species greatly attenuates herbivory by insect pests, and commercial products based on this premise are sold in the United States and Canada. The principal mechanism for this activity is the accumulation of endophyteproduced insecticidal alkaloids. A number of additional benefits accrue to the host grass from such associations, including increased rates of vegetative reproduction, improved drought tolerance, and increased resistance to fungal diseases (21). Conifer needle endophytes produce anti-insectan compounds, and when such strains are present in a stand, the fitness of the stand is increased (22,23).
The biology of the F. verticillioides-corn association is not known, i.e., which of the several possible mutualisms that might apply is not known (24). It is known that in conifers, grasses, and flowering plants the endophyte-plant association is under plant genetic control. If this is true in corn, it may open up avenues for the production of genotypes with lower potential to accumulate fumonisin. If corn does not grow well in the absence of F. verticillioides, strategies to degrade fumonisin as it is produced in vivo may be useful (25).
Even from regions where fumonisin accumulations in corn are less likely, strains of F. verticillioides isolated from corn have the potential to produce this toxin. These regions include Africa, Asia, Europe, Canada, the United States, Mexico, and South America (26)(27)(28)(29)(30).
Fumonisins are potent phytotoxins that cause electrolyte loss and interfere with the formation of complex phytosphingolipids (31). In crosses of high-and low-fumonisinproducing strains of G. fujikuroi, only progeny that produced high concentrations of fumonisin in vitro caused significant stem rot (32).

Fusarium Ear Rots
Gibberella ear rot, or pink ear rot, is prevalent in northern temperate climates, especially in wet years, and is caused by G. zeae and F. culmorum. Fusarium ear rot/Fusarium kernel rot is associated with warm, dry years and insect damage and is caused by F. subglutinans (G. subglutinans), F. verticillioides (= G. fujikuroi), and F. proliferatum (33). F. culmorum is more important as a disease-causing species in eastern Europe (as opposed to F. graminearum, which is responsible for disease in the northern U.S., Canada, and Europe) (34,35). In warmer parts of the United States and lowland tropics, F. verticillioides is one of the most important ear diseases (36).
Factors affecting the epidemiology of Gibberella ear rot are reviewed in Miller (37) and Sutton (38). Epidemics require the congruence of three factors: airborne or insectborne spores, inoculum at the correct time, and appropriate moisture and temperature. F. graminearum disease incidence is affected by moisture at silk emergence, and prevalence is increased with warm, wet weather later in the season. If it is cool and wet, epidemics do not take place, regardless of initial infections (37,39). Continuous rain in northern temperate areas is usually associated with cold temperatures not conducive to the growth of F. graminearum (38). Monitoring of the growth of F. graminearum in experimentally infected ears showed the growth rate of the fungus was indeed sensitive to temperature. A period where the average "growing degree days > 5" was approximately 10 virtually halted growth of the fungus. An average value of approximately 15 "growing degree days > 5" resulted in rapid growth of F. graminearum (8).
Temperature is the signal-controlling factor for cereal diseases caused by Fusarium species (37,40), not least of which are the two fusaria considered here: F. graminearum and F. verticillioides. Measurement of the growth rate of F. graminearum and F. verticillioides in corn tissues was made using the 14 C ergosterol method of Gessner and Newell (41). This method allows the growth rate of fungi to be determined in natural materials such as plant tissues by determining the incorporation of 14 C-acetate into the unique fungal sterol, ergosterol. As suggested by field measurements noted above (8), F. graminearum has a very narrow temperature window for growth ( Figure 2). The optimal temperature is between 26 and 28°C. Growth rate at 24°i s one-quarter that at 26-28°C; at 28°C, it is about one-half. In contrast, F. verticillioides grows well above 26°C (42). Temperature is a categorical variable with respect to the distribution of the two ear diseases discussed here.

Epidemiology of Fusarium Ear Rot
As discussed by others in this volume, a very high incidence of esophageal cancer has been reported among the black population of Transkei (43). Corn meal porridge is the staple diet. Adults also consume beer deliberately made from moldy corn containing up to 118 mg/kg fumonisins (44). Such beers could contain fumonisin concentrations of 30 mg/L, based on experiments conducted on beer made from worts containing fumonisincontaminated corn adjuncts (45). The incidence of esophageal cancer was higher in both sexes in the south than in the northern parts of the region.
The occurrence of F. graminearum in maize kernels was found to be greater in lowrisk areas for esophageal cancer (46,47). Sydenham et al. (48) found that deoxynivalenol and zearalenone were approximately 3 times higher in home-grown maize in areas with a low incidence of esophageal cancer. In contrast, the occurrence of F. verticillioides in maize kernels significantly correlated to esophageal cancer rates. The prevalence of F. verticillioides was greater in households where home-grown maize was collected in the high incidence area compared to a similar collection in a low-incidence area (47). Studies conducted after the characterization of the fumonisins in 1988 also found significantly higher levels of F. verticillioides and fumonisins (20×) in areas with high esophageal cancer rates compared to those with low esophageal cancer rates in Transkei (44,48,49).
The high-and low-incidence sites are about 200 km apart. Based on the performance of hybrids in experimental plots, maize grows well in both areas. Although some soil fertility factors differed between the areas with high and low incidences of esophageal cancer, there is no evidence that any nutrient was limiting, at least for hybrid maize production (50). Farmers grow open-pollinated maize of varying genotypes passed on from farm to farm and season to season. Kernel types include large flour-maize kernels as well as dent and flint-type white, yellow, and blue kernels.
The important difference between the areas with low and high incidences of cancer is that the northern area with a low incidence of esophageal cancer is at an altitude of approximately 500 m higher and is hence cooler. The environmental conditions that prevail in the areas with a high incidence of esophageal cancer in Transkei clearly favor colonization of corn by F. verticillioides, especially when there is drought. After temperature, there are four factors that can be considered in relation to the occurrence of F. verticillioides ear rot of corn: drought stress, insect damage, other fungal diseases, and corn genotype.

Drought
Studies of the occurrence of fumonisin from natural occurrence and experimental infections clearly demonstrate the importance of drought rather than temperature stress on the occurrence of fumonisin. The occurrence of fumonisin in Ontario, Canada (a cool maize-growing region), was limited to drought-stressed fields. In the 1993 corn crop, the three counties with the highest average FB 1 concentrations had 1.4 µg/g, and the three counties with the lowest average FB 1 had 0.4 µg/g. Temperature was similar at 104 and 107% of the 30-year average, respectively. Rainfall in the high group was only 49% of normal and in the low group, 95% of normal values (51).
In experimental inoculations of 14 corn genotypes performed in Poland, the year of highest FB 1 accumulation was 117% of 30-year average values; the year of lowest FB 1 accumulation was 102% of normal. Rainfall in the highest FB 1 year was 6% of normal and in the lowest FB 1 year was 65% of typical values (52).
A study of fumonisin occurrence in hybrids grown in the United States indicated that hybrids grown outside their range of adaptation had higher fumonisin concentrations. Fumonisin concentrations were inversely proportional to June rainfall (53), again suggesting the important role of drought stress. Data from samples collected in Africa, Italy, and Croatia also indicate fumonisin accumulation in lines grown outside their area of adaptation, which includes tolerance to moisture stress (54,55). Hybrids with an increased propensity for kernel splitting had more Fusarium kernel rot (56). Kernel splitting is generally worse under drought conditions.

Insect Damage
As a generalization, drought stress results in greater insect herbivory on corn, hence it is not possible to totally separate these variables. Further, Fusarium ear rot severities were related to wound size in studies of experimental inoculation methods (57). It appears that kernel damage alone promotes the disease and fumonisin accumulation (19,58). However, there is a strong relationship between insect damage and Fusarium ear rot. A field survey in Austria demonstrated that the incidence of the European corn borer increased F. verticillioides disease and fumonisin concentrations. In contrast, corn borer incidence was not correlated to incidence of F. graminearum (59). Disease incidence was also shown to correlate to populations of thrips (Frankliniella occidentalis) (60). Hybrids with a thin kernel pericarp were more susceptible to insect wounds, which allowed easier access to the fungus (61). Corn genotypes containing the anti-insectan Bt protein had lowered recoveries of F. verticillioides and fumonisin (62,63).

Other Fungal Diseases
Corn infected by other ear-damaging pathogens such as F. graminearum may be predisposed to F. verticillioides infection and fumonisin accumulation. Ear wounds inoculated by F. graminearum, F. verticillioides, and F. subglutinans produced visible symptoms on a 1-9 scale of 7.3, 4.4, and 4.7, respectively. Despite that F. graminearum and F. subglutinans do not produce fumonisin, these ears contained 42 and 3 µg/g FB 1 , respectively (35). As noted above, even when F. verticillioides is inoculated, it does not follow that the introduced strain causes ear damage and fumonisin production, because the fungus is always present.

Breeding
Partly because the problem of fumonisin is new, categorical statements about the impact of plant breeding on reducing Fusarium ear rot cannot be made. Three crucial matters have not been assessed by researchers: a) F. verticillioides is essentially ubiquitous in corn kernel, b) fumonisin can be present in concentrations in symptomless kernels, and c) relationships between fumonisin concentration are not strong in ears with low or moderate damage (24). Results of a field (natural infection) study showed that inbred differences for asymptomatic and symptomatic kernel infections are expressed in their hybrids (64). Within areas of adaptation, there are apparent differences of symptom response (55,64,65). Studies that determine Fusarium ear rot after inoculation have shown some reductions in disease symptoms. After many cycles of selection, there were slight to some improvements in symptom expression in some tropical late yellow flint genotypes (36).
It is more obvious that factors that control insects (63), resistance to other ear diseases, and area of adaptation (including drought and temperature tolerance) are important in reducing the risk of fumonisin accumulations in corn. Hybrids grown outside their area of Environmental Health Perspectives • VOLUME 109 | SUPPLEMENT 2 | May 2001

Summary
The weight of evidence is that F. verticillioides is endemic in corn kernels. The biology of fumonisin production suggests that fumonisin is produced in material quantities only in senescent corn tissue and that fumonisin would not be biologically active in growing corn tissue. This is consonant with other examples of mutualistic endophytes.
Incidence of Fusarium kernel rot is higher in warmer climates under dry conditions. In such environments, insect damage is well recognized as a collateral factor. Regardless of moisture stress, insects appear to promote F. verticillioides occurrence. Breeding for resistance to Fusarium ear rot has produced inconclusive results to date. It may also be necessary to make use of alternative strategies, such as producing hybrids that contain enzymes to degrade fumonisin as it is produced. The best available strategies for reducing the risk of fumonisin contents of maize are to ensure that hybrids are adapted to the environment and to limit drought stress and insect herbivory.