Infection of maize by Fusarium species and contamination with fumonisin in africa

Programme on Agricultural and Food Technology, National Institute of Agricultural Research of Benin, P. O. Box 128, Porto-Novo, Benin. International Institute of Tropical Agriculture (IITA), P. O. Box: 08-0932 Tri Postal, Cotonou, Benin. Programme on Mycotoxins and Experimental Carcinogenesis (PROMEC), Medical Research Council, P. O. Box 19070, Tygerberg 7505, South Africa. Forestry and Agricultural Biotechnology Institute (FABI), Faculty of Biological and Agricultural Sciences, University of Pretoria, Pretoria 0002, South Africa.


INTRODUCTION
Maize (Zea mays L.) is a cereal crop grown throughout the word. Maize plays an important role in the diet of millions of African people due to its high yields per hectare, its ease of cultivation and adaptability to different agro-ecological zones, versatile food uses and storage characteristics (Asiedu, 1989). The total production of *Corresponding author. Tel: +229 21 41 60. E-mail: lta@intnet.bj. Africa in 2001 was estimated to be about 42 millions tons (FAO, 2002).
In the field as well as in the store, many pests and parasites attack maize and during the storage period. Insects are most often considered as the principal cause of grain losses (Gwinner et al., 1996). However, fungi are also important and rank second as the cause of deterioration and loss of maize (Ominski et al., 1994). Kossou and Aho (1993) reported that fungi could cause about 50 -80 % of damage on farmers' maize during the storage period if conditions are favourable for their development. The major genera commonly encountered on maize in tropical regions are Fusarium, Aspergillus and Penicillium (Samson, 1991;Orsi et al., 2000).

FUSARIUM SPECIES AND THEIR IMPORTANCE IN MAIZE
Fusarium species are ubiquitous in soils. They are commonly considered as field fungi invading more than 50% of maize grains before harvest (Robledo-Robledo, 1991). Several phytopathogenic species of Fusarium are found to be associated with maize including F. verticillioides (Sacc.) Nirenberg, F. proliferatum (Matsushina) Nirenberg, F. graminearum Schwabe and F. anthophilum (A. Braun) Wollenweber (Lawrence et al., 1981;Scott, 1993;Munkvold and Desjardins, 1997). Among them, F. verticillioides is likely to be the most common species isolated worldwide from diseased maize (Munkvold and Desjardins, 1997). Doko et al. (1996) reported F. verticillioides as the most frequently isolated fungus from maize and maize-based commodities in France, Spain and Italy. Likewise, Orsi et al. (2000) found in Brazil that F. verticillioides was the predominant Fusarium species on maize. In general in Africa, very little information is available on F. verticillioides occurrence on maize. Reports of surveys conducted in some African countries however showed it as the most prevalent fungus on maize Allah Fadl, 1998;Baba-Moussa, 1998;Kedera et al., 1999).
F. verticillioides is an endophyte of maize establishing long-term associations with the plant (Baba- Moussa, 1998;Pitt and Hocking, 1999). Symptomless infection can exist throughout the plant in leaves, stems, roots, grains, and the presence of the fungus is in many cases ignored because it does not cause visible damage to the plant (Munkvold and Desjardins, 1997). This suggests that some strains of F. verticillioides produce disease in maize and others do not (Bacon and Williamson, 1992).
F. verticillioides infects maize at all stages of plant development, either via infected seeds, the silk channel or wounds, causing grain rot during both the pre-and postharvest periods (Munkvold and Desjardins, 1997). Figure 1 shows Fusarium spp. damage on maize cob. A diagrammatic illustration of the disease cycle of F. verticillioides in maize is proposed on Figure 2 showing the following possible infection pathways: • Infection from seed to cob and further to grain through systemic movement in stalk, • Infection from root to grain through stalk and cob, • Infection from airborne or water-splashed conidia to silk and further to grain, • Infection through wounds caused by insects that can also act as vectors of inoculum (Munkvold and Desjardins, 1997).

FUMONISINS AND THEIR TOXICOLOGICAL EFFECTS
Maize contamination by fungi not only renders grains unfit for human consumption by discoloration and reduction of nutritional value, but can also lead to mycotoxin production. Mycotoxins are poisonous secondary metabolites produced by some fungi in staple foods and foodstuffs. Many of them are considered to be important worldwide, but the five most often reported and well documented are deoxynivalenol/nivalenol, zearalenone, ochratoxin, aflatoxins and fumonisins (Pittet, 1998;Pitt, 2000). There is ample evidence that mycotoxin problems affect the agricultural economies of many countries in the world, mainly the African countries. The FAO estimated that each year, between 25% and 50% of the world's food crops are contaminated by mycotoxins (Mannon and Johnson, 1985). The direct impact of mycotoxins on the staple product quality constitutes an important danger for human health and among them fumonisins produced by some toxigenic Fusarium species on maize and maize-based foods and feeds increase the risk.
Fumonisins are recently discovered mycotoxins. In 1988, their chemical structure and biological activity were elucidated in South Africa (Gelderblom et al., 1988;Marasas, 2001). Since the discovery of these toxins, numerous research works have been undertaken to investigate further about them. The interesting results found so far have been thoroughly reviewed (Norred, 1993;Riley et al., 1993;Cardwell and Miller, 1996;Gelderblom et al., 1996;Shephard et al., 1996;Marasas, 1996;IPCS, 2000;Bolger et al., 2001;Marasas, 2001;WHO, 2002). These reviews mainly highlighted: 572 Afr. J. Biotechnol. • Events leading to the discovery of the fumonisins, • The toxicological effects of these toxins, • Their worldwide occurrence in maize and maizebased foods and feeds, • Their association with animal and human diseases, • Their impact on animal and human health.
Fumonisins have been found as very common contaminants of maize-based foods and feeds in the United States of America, China, Europe, South America and Africa (Sydenham et al., 1991;Thiel et al., 1992, Visconti andDoko, 1994;Shephard et al., 1996). To date, a total of 28 fumonisin analogs have been identified and characterised (Rheeder et al., 2002). But the most abundantly found in naturally contaminated foods and feeds are FB 1 , FB 2 , FB 3 (Shephard et al., 1996;Rheeder et al., 2002).
Fumonisins are produced by several Fusarium species (Marasas, 2001)  Amongst these, F. verticillioides and F. proliferatum are by far the most prolific fumonisin producers (Shephard et al., 1996). They produce the highest amounts of toxins: up to 17900 µg/g of FB 1 have been recorded in cultures for the former, and 31000 µg/g FB 1 for the latter (Rheeder et al., 2002). Maize is the product in which fumonisins are most abundant (Shephard et al., 1996). Fumonisins have been also detected but at lower levels in sorghum (Shetty and Bhat, 1997;Leslie and Marasas, 2001), rice (Abbas et al., 1998) and spices (Pittet, 1998). Fumonisins can contaminate maize foods and feeds as a result of the Fusarium invasion before and after harvest (Doko et al., 1995).
Fumonisins have emerged as a highly visible animal and human health safety concern since they have been associated with many animal diseases such as leukoencephalomalacia (LEM) in horses (Marasas, 1996), pulmonary oedema syndrome (PES) in pigs (Harrison et al., 1990;Colvin and Harrison, 1992) and hepatocarcinogenesis in rats . With respect to humans, studies on the prevalence of oesophageal cancer in regions of South Africa, China, Italy and Iran, revealed an association between this disease and the consumption of maize contaminated by Fusarium spp (Franceschi et al., 1990;Rheeder et al., 1992;Chu and Li, 1994;Marasas, 1996;Ueno et al., 1997;Shephard et al., 2000;Wang et al., 2000). The International Agency for Research on Cancer (IARC) evaluated in 1992 the toxins derived from F. verticillioides as possibly carcinogenic to humans (IARC, 1993). More recently, based on the research results obtained so far, FB 1 has been evaluated as possibly carcinogenic to humans (class 2B) (IARC, 2002).
Although the effects of fumonisins on humans are not yet well understood, legislation is being put in place to regulate commercial exchanges of fumonisincontaminated maize and maize-based foods. The US Food and Drug Administration (FDA) recommended that the fumonisin levels should not be higher than 4 µg/g in human foods (FDA, 2000a;FDA, 2000b). In Switzerland, tolerance levels for fumonisins of 1 µg/g in dry maize products intended for human consumption were proposed . The Joint FAO/WHO Expert Committee on Food Additives (JECFA) allocated a group provisional maximum tolerable daily intake (PMTDI) of 2 µg/g for FB 1 , FB 2 and FB 3 , alone or in combination (WHO, 2002). With respect to animals, the total fumonisin maximal tolerable levels recommended by FDA in maize-based feeds are 5 µg/g for horses, 5 µg/g for rabbit, 60 µg/g for ruminants (cattle, sheep, goat) and Fandohan et al. 573 100 µg/g for poultry (chicken, turkey, duckling) (FDA, 2000c). The mechanism of action of fumonisins in animal diseases is quite complex, but it appears that the toxins mainly cause disruption of lipid metabolism in cells, an event that can lead to cellular deregulation or toxic cell injury and finally to cell death (Wang et al., 1991;Riley, 1998;Marasas et al., 2001).
Fumonisins are found to be phytotoxic. FB 1 can indeed damage a wide variety of plants including maize (Scott, 1993;Lamprecht et al., 1994). Doehlert et al. (1994) showed that the presence of high levels of fumonisins in maize seeds might have deleterious effects on seedling emergence. Elongation of maize radicles was inhibited by about 75% after 48 h of imbibition in 100 µg/g of fumonisins and amylase activities in seeds significantly decreased as well.
Fumonisins are also found to be relatively heat stable (Alberts et al., 1990;Howard et al., 1998) and light stable (IARC, 1993). They are also stable in stored products when these are kept in airtight at very low temperatures (Gelderblom et al., PROMEC Unit, Medical Research Council, Tygerberg, South Africa, 2002, unpublished data), or γ-irradiated . However, instability of fumonisins in contaminated products over time has been shown (Scott et al., 1999;Kim et al., 2002). Fumonisins are also water soluble (IPCS, 2000).

FACTORS INFLUENCING INFECTION OF MAIZE WITH FUSARIUM SPECIES AND FUMONISIN DEVELOPMENT
Infection of maize with Fusarium species and its contamination by fumonisins are generally influenced by many factors including environmental conditions (climate, temperature, humidity), insect infestation and pre-and postharvest handling. These factors do not influence infection independently but most often there are complex interactions.

A. Environmental factors
Worldwide surveys showed high levels of fumonisins associated with warmer and drier climates (Shephard et al., 1996) and when weather conditions are favourable for Fusarium infection . At the same location, fumonisin contamination is not necessarily the same from one year to another. Hennigen et al. (2000) found in Argentina a marked difference in terms of fumonisin contamination for the same maize varieties during two consecutive growing seasons, due to the fact that environmental conditions may differ from one growing season to another.
Studying the effect of climatic conditions on fumonisin occurrence in freshly harvested maize in different regions of the State of Parana in Brazil, Ono et al. (1999) detected higher fumonisins levels in maize samples from the Northern and Central-Western regions compared to that from the South. The authors suggested that it could be due to the differences in rainfall levels during the month preceding harvest (92.8 mm in South, 202 mm in North).
Physiological stress during the period just preceding maize harvest, due to drastic oscillations in rainfall and relative humidity, is likely to create favourable conditions for fumonisin production (Visconti, 1996). Shelby et al. (1994) suggested that dry weather at or just prior to pollination of maize might be an important factor for fumonisin production in maize. All this leads to the conclusion that some climatic events such as changes in rainfall patterns or stress during the last stages of maize plant development in the field are likely to have a great influence on fumonisin production in maize before harvest.
Furthermore, temperature and moisture conditions during the growing season as well as during storage are often pointed out to affect maize infection by Fusarium spp. and fumonisin synthesis. In this connection, water activity (a w ), the water available for fungal growth, plays a key role. Velluti et al. (2000) working in vitro on fungal competition on maize found that the growth rate of F. verticillioides was higher at a temperature of 25°C, whereas at 15°C, growth was much lower. These researchers also found that at a constant temperature, the growth rate of F. verticillioides increased with water activity. Scott (1993) suggested that the best temperature for production of fumonisin B 1 in maize is 20°C. Marin et al. (1999) rather found that the toxin was optimally produced at 30°C and 0.98 a w . However, Alberts et al. (1990) showed that the mean FB 1 yield obtained at 25°C (9.5 g/kg) was significantly higher than that at 20°C (8.7 g/kg) and 30°C (0.6 g/kg). Munkvold and Desjardins (1997) reported that F. verticillioides generally grows in grain when moisture content is more than 18 -20%.

B. Agricultural practices
It has been reported that late planting of maize with harvesting in wet conditions favours disease caused by F. verticillioides (Bilgrami and Choudhary, 1998), and the prevalence of this fungus is considerably increased with wet weather later in the season (Al-Heeti, 1987). Moreover, repeated planting of maize and other cereal crops in the same or in nearby fields favours fungal infection by increasing the fungal inoculum and insect population that attack maize plants (Bilgrami and Choudhary, 1998). Lipps and Deep (1991) found that the rotation maize/nonhost crop of Fusarium was better than maize/maize, as the former was less favourable to Fusarium disease outbreak than the latter. Weed control also affects fungal infection in maize fields because it helps to eliminate nonhost weeds on which Fusarium can also be found (Bilgrami and Choudhary, 1998).

C. Maize characteristics
The type of maize cultivar and grain characteristics such as colour, endosperm type, chemical composition and stage of development may also influence fungal infection and subsequent fumonisin production. Late-maturing maize cultivars in which grain moisture content decreases slowly below 30% are most susceptible to Fusarium disease (Manninger, 1979). It is thought that maize cultivars with upright cobs, tight husks (Emerson and Hunter, 1980), thin grain pericarp (Riley and Norred, 1999), and an increased propensity for grain splitting (Odvody et al., 1990) are likely to be more susceptible to Fusarium infection. Tight-husked varieties favour Fusarium problems because of slow drying (Dowd, 1998).
Fumonisins are found more concentrated in the pericarp and germ of the grain than in the endosperm, so that removal of those outer parts by mechanical processes such as dehulling can significantly reduce the toxin in maize (Charmley and Prelusky, 1995;Sydenham et al., 1995;FDA, 2000b). However, influence of maize grain colour on fumonisin contamination does not seem to be clear. Shephard et al. (1996) reported that in some years, fumonisin levels were significantly lower in yellow than in white maize, but the reverse situation was observed in other years. Hennigen et al. (2000) compared contamination of maize varieties of flint endosperm to that of dent type and did not find significant differences. Shelby et al. (1994) tested fifteen maize hybrids and found no significant correlation between starch, lipid, fibre, and protein contents and fumonisin production in maize.
Grain age may also influence fumonisin production in maize. Warfield and Gilchrist (1999) found higher levels of fumonisins in maize grains at the dent stage and significantly lower levels in grains at the immature stage, suggesting that production of the toxin may begin early in cob development and increase as the grains reach physiological maturity. Likewise, Chulze et al. (1996) reported that contamination of maize by fumonisins was greater after physiological maturity.

D. Postharvest operations
Postharvest handling and processing (sorting, washing, dehulling, milling, fermentation, cooking) favourably or unfavourably affect fungal infection and fumonisin production in maize. Mechanical damage during and after harvest may offer entry to the fungal spores either in maize cobs or grains. Dharmaputra et al. (1994) found that motorised shellers can cause mechanical damage on grains providing entry points to fungal spores. Substantial amounts of fumonisins (up to 74%) can be removed by simply washing maize grains, immersing them in water and by removing the upper floating fraction, as contaminated grains generally have a low density (Shetty and Bhat, 1999). These authors also found that removal of the toxin is more significant (about 86%) if salt is added to the water during that process. Likewise, sorting and removal of small, broken and visibly contaminated grains during processing can significantly reduce toxin levels (Charmley and Prelusky, 1995). Steeping maize grains in water has also been found effective in reducing fumonisin content (Canela et al., 1996). In contrast, fermentation of maize does not seem to reduce fumonisin levels (Shephard et al., 1996;Desjardins et al., 2000).
As for milling, Bennett et al. (1996) found that by wetmilling fumonisin-contaminated maize, the toxin distribution in the different fractions is as follows: very little or no fumonisin in the starch fraction, but detectable fumonisins in fibre, germ and steep water fractions. This indicates that maize-based foods derived from the starch fraction are likely to contain less fumonisins than that derived from the other fractions. After dry-milling contaminated maize, fumonisins levels were found lower in grits and higher in germ, bran and fines (Bolger et al., 2001). It has also been shown that fumonisin levels decrease as the level of refinement of maize meal during milling increases (Shephard et al., 1996).
Regarding cooking, it has been observed that fumonisins are fairly heat-stable and that ordinary cooking does not substantially reduce the toxin (Alberts et al., 1990;Scott, 1993). Significant removal of fumonisins is more likely to occur only when temperature during cooking is more than 150°C (Bolger et al., 2001).
Although some processing methods potentially can be selected as favourable ways to reduce fumonisin levels in maize-based products, it is important to keep in mind that their success would depend on many factors including the moisture content of the product, the degree of contamination and distribution of the toxin in the product, and the presence of additives (Charmley and Prelusky, 1995;Bolger et al., 2001).

A. Storage insects
Insects also play an important role in infection of maize by Fusarium spp. They can act as wounding agents or as vectors spreading the fungus from origin of inoculum to plants (Dowd, 1998). Wounding by insects may provide an opportunity for the fungus to circumvent the natural protection of the integument and establish infection sites in the vulnerable interior (Bilgrami and Choudhary, 1998). Borers and insects of the family Nitidulidae are most often cited as favouring maize infection by Fusarium spp. Fandohan et al. 575 They include among others the lepidopteran stem and cob borers (Ostrinia nubilalis, Sesamia calamistis, Eldana saccharina, Mussidia nigrivenella and Busseola fusca), thrips and sap beetles (family Nitidulidae) (Flett and Van Rensburg, 1992;Munkvold and Desjardins, 1997;Cardwell et al., 2000;Ako et al., 2003). Sobek and Munkvold (1995) found in the USA that damage caused by the European maize borer Ostrinia nubilalis increased infection by F. verticillioides by three-to ninefold over those with simple mechanical damage. Moreover, larvae of O. nubilalis can also act as vectors of F. verticillioides by carrying inoculum from plant surfaces into maize cobs . In South Africa, Flett and Van Rensburg (1992) showed that Busseola fusca infestation significantly increased the incidence of F. verticillioidesinfected maize cobs, irrespective of whether the cobs are artificially inoculated with the fungus or not. In a recent study in Benin, it has been observed that cob/stem infection by F. verticillioides positively correlated with infestation of Eldana saccharina, Cryptophlebia leucotreta, Mussidia nigrivenella and Sesamia calamistis (Schulthess et al., 2002).
Regarding the beetles, it has been shown that not only nitidulid beetles are strongly implicated in F. verticillioides infection, but also cucurlionid and silvanid species positively correlated with fungal infection (Cardwell et al., 2000). All these findings pose the problem of cause and effect relationships between fungal infection and insect infestation on maize plants. It is likely that the presence of F. verticillioides promotes insect attacks (Schulthess et al., 2002) and insect infestation favours fungal infection (Dowd, 1998). F. verticillioides may be introduced into the stem and cob via insects (Munkvold and Carlton, 1997). Likewise, incidence of infection by F. verticillioides in maize stems is a source for cob infection by the fungus, not only through movement of the fungus, but also through increased activity of stem borers (Baba-Moussa, 1998). On the other hand, F. verticillioides produced volatiles that are quite attractive to nitidulid beetles (Bartelt and Wicklow, 1999). It has been shown that fecundity, laying of eggs and survival of larvae of Eldana saccharina were significantly higher on inoculated maize plants (Ako et al., 2003). The authors also found that development time of Carpophilus dimidiatus was lower and its fecundity higher on infected grain than on noninfected grain. Schulthess et al. (2002) suggested that keeping the plant free of the fungus could be an effective way to reduce insect damage to both stem and grain. On the other hand, any action also to avoid insect infestation is useful for reducing infection of maize by F. verticillioides (Riley and Norred, 1999).

B. Fungal interactions
Interactions among fungi in maize also constitute an important factor influencing fungal infection and subsequent mycotoxin production. Harvested maize grains in the tropical zones contain mycelium and spores of several fungal species including mainly Fusarium, Aspergillus and Penicillium that can come into contact, grow and compete for food if environmental conditions are favourable. As far as Fusarium species are concerned, many research reports highlighted their interaction with other fungi. Velluti et al. (2000) showed that populations of F. verticillioides and F. proliferatum, the most important fumonisin producers, are markedly reduced by the presence of F. graminearum, and that fumonisin B 1 (FB 1 ) production by them can be significantly inhibited as well in the presence of F. graminearum. On the other hand, Marin et al. (1998) found that F. verticillioides and F. proliferatum are generally very competitive and dominant against Aspergillus flavus and Penicillium spp., especially at a w more than 0.96. This inhibition can lead to significantly reduced aflatoxin contamination in infected grains (Zummo and Scott, 1992).

ATTEMPTS TO CONTROL F. VERTICILLIOIDES AND TO DETOXIFY OR REDUCE FUMONISIN LEVELS IN MAIZE
There is strong evidence that due to its endophytic habit, control of F. verticillioides in the field is very difficult. Novel control strategies are being investigated and some reported technologies include: • The use of an endophytic bacterium (e.g. Bacillus mojavensis) as a biological control agent on maize seed (Bacon and Hinton, 2000). • The use of an iodine-based product called Plantpro 45 as a biocompatible control of the fungus. The active ingredient of that product has been used as a disinfectant in human and animal health care products (Yates et al., 2000).
• The use of non-producing strains of F. verticillioides aiming to minimise fumonisin levels in maize .
Additional investigations are however needed to render some of those technologies more applicable. Decontamination of fumonisins in maize and maizebased products by means of chemical reactions is the object of many research studies. Fumonisins are quite stable molecules and their destruction is likely to be also quite difficult. Ammoniation, initially used for detoxify products from aflatoxins has been investigated for fumonisin reduction but does not always give satisfactory results. Scott (1993) reported that treatment of F. verticillioides culture material with 2% of ammonium hydroxide at 50°C decreased fumonisin concentration by 89%, but only 32% of toxin reduction were later measured after four days air-drying. In contrast, nixtamalization, the alkaline cooking of maize for tortilla production in Central America, significantly reduces fumonisin concentration in maize (Dombrink-Kurtzman et al., 2000). However, Voss et al. (1996) found that nixtamalized F. verticillioides culture remained toxic. This indicates that reduction in detectable fumonisins does not necessarily result in reduced toxicity.
It is therefore clear that detoxification of mycotoxins in foods is not so easy. Sinha (1998) suggested that it must be economical, simple, easy to be applied by unskilled person, not too time-consuming, capable of removing all traces of the active toxin without hazardous chemical residues in the decontaminated food, and does not impair the nutritional quality of the food.
Considering the above-mentioned review of existing findings on fumonisin contamination, several points arise and need to be emphasised.

Contamination of food commodities by fumonisins has
become a serious food safety problem throughout the world. People are more and more aware that the fumonisins, in addition to aflatoxins, constitute a real threat to human and animal health. However, in contrast to aflatoxins, fumonisins are less documented. Indeed, they are recently discovered mycotoxins and more research studies are urgently needed in order to understand more about them. 2. Some information is available on factors contributing to fumonisin production and on those able to reduce fumonisin levels in foods. However, research results on some factors remain uncertain, or are not applicable to a developing country situation. The need for more information about environmental and agroecological influences, fumonisin toxicology in respect to human and animal health, prevention methods against fungal infection and fumonisin contamination, methods to use for reducing the toxin in foods and other aspects of fumonisins, is great enough to challenge scientists to undertake many research studies. 3. To date in Africa, apart from South Africa, very little information is available on the natural occurrence of both Fusarium and fumonisins, although this part of the world is most often suspected of having potentially higher levels of fumonisins due to its position in tropical and subtropical zones. Work undertaken so far in a few African countries basically consisted of sporadic surveys of farmers' stores and retail markets, mostly basing data measurements on a relatively small number of samples (Shephard et al., 1996). It is a matter of great concern that in Africa, millions of people are consuming contaminated maize and maize-based foods daily without being aware of the danger. Efforts are, however, to be saluted in investigating fumonisins contamination in maize and maize-based foods in some African countries other than South Africa such as Benin, Cameroon, Ghana, Kenya, Zambia and Zimbabwe (Shephard et al., 1996;Doko et al., 1995;Hell et al., 1995;Kedera et al., 1999;Kpodo et al., 2000;Gamanya and Sibanda, 2001;Ngoko et al., 2001). Consequently, there is great need for more investigations on the continent, mainly in the maize production and consumption zones.