Peanut ( Arachis hypogaea L.) is the second most cultivated crop in Zambia. In 2013, groundnut was cultivated on 207,249 Ha, mostly by subsistence farmers, and the average yield was 515 Kg/Ha (FAOSTAT, 2013). Within Zambia, peanut production is highest in the Eastern Province where the crop is grown on 80,000 Ha in different agroecological zones, from the cooler and wetter plateau, i.e., Chipata District, to the hotter and drier Luangwa Valley, i.e., Mambwe District. Like in most countries in southern Africa, the cropping season starts at the onset of the rains in December, and peanut is grown in rotation with other staples and cash crops such as maize, cotton, sunflower, tobacco, etc. Farmers, however, prioritize sowing corn, cotton, tobacco, and sunflower over peanut, which are planted up to 3 weeks after the rains commence.

Aflatoxins are chemical compounds produced by the opportunistic pathogens Aspergillus flavus Link ex Fries, (Teleomorph Petromyces flavus) A. parasiticus Speare (Teleomorph P. parasiticus), and A. nomius (Teleomorph P. nomius) (Abbas et al., 2004; Amaike and Keller, 2011; Horn et al., 2009a; Horn et al., 2009b; Horn et al., 2011; Paterson et al., 1997), however, only A. flavus and A. parasiticus have an impact on agriculture (Horn, 2003; Klich, 2007). Oil seed crops such as corn, peanut, cotton, and sunflower are susceptible to aflatoxin contamination. Aflatoxins are a common contaminant of foods, especially the staple diets of many in developing countries (Williams et al., 2004) and the US Food and Drug Administration considers aflatoxins to be unavoidable contaminant of foods (Williams et al., 2004). Dietary exposure to aflatoxin can lead to aflatoxicosis, defined by Williams et al. (2004), as poisoning from ingesting aflatoxins. Aflatoxicosis, can either be acute, which is severe intoxication that leads to direct liver damage followed by sickness or death (Azziz-Baumgartner et al., 2005), or chronic, which is subsymptomatic exposure that may lead to nutritional and immunological consequences (Williams et al., 2004). However, all doses i.e., chronic and acute, have an accumulative effect on the risk for cancer (Williams et al., 2004). In developing countries, exposure to aflatoxins may begin early, when infants are exposed tog weaning and adult foods (Gong et al., 2002).

Aflatoxin contamination is a serious problem in Africa (Atehnkeng et al., 2008; Kaaya and Warren, 2005; Matumba et al., 2015; Monyo et al., 2012; Mutegi et al., 2009; Setamou et al., 1997; Wagacha and Muthomi, 2008) and also worldwide (Amaike and Keller, 2011; Munkvold, 2003). Despite the magnitude of the aflatoxin problem, only a few studies have attempted large scale comparisons of soil populations of aflatoxigenic fungi over large geographic areas which represent differences in climatic, soil type, and cropping systems (Horn, 2003). Reports on characterization of aflatoxigenic fungi have been published mostly from the southeastern US (Abbas et al., 2004; Boyd and Cotty, 2001; Horn, 2003; Horn, 2005; Horn et al., 1995; Sweany et al., 2011; Jamie-Garcia and Cotty, 2006). In Africa, there have been reports from Kenya (Okoth et al., 2012) where the authors tested isolates from two different agroecologies – the hotter and drier Makueni, where the acute outbreak of aflatoxicosis occurred (Lewis et al., 2005), and the cooler and wetter Nandi, a corn growing region in the Rift Valley. Okoth et al. (2012) reported that there were no differences in the occurrence of A. flavus between the two tested regions, however, S-strain A. flavus was more prevalent in Makueni compared to Nandi, where L-stains were more prevalent. In Nigeria, Atehnkeng et al. (2008) collected maize samples from three agroecological zones and determined the identity of isolated species and also the incidence of toxigenic versus atoxigenic strains. However, most reports from Africa document the prevalence of aflatoxin contamination in different commodities and researchers often do not differentiate or characterize Aspergillus species (Monyo et al., 2012; Matumba et al., 2015).

It is important to study and document factors that influence aflatoxin contamination, such as understanding the activities and population structure of aflatoxigenic fungi in the soil, because this is a pre-requisite for developing effective measures to control aflatoxin contamination (Horn, 2003). Other important factors also should be considered when developing preharvest aflatoxin mitigation measures, i.e., the crops grown and their resistances to aflatoxin contamination, soil characteristics, agronomic and cultural practices, weather, and the occurrence of drought. Taken together, these would be used to develop a holistic approach to aflatoxin mitigation.

To our knowledge, there are no published reports from Zambia on the population density of A. flavus , and characterization of aflatoxigenic fungi. Therefore, the objective of this study was to identify aflatoxigenic fungi from soils cropped to peanut, estimate the population densities of A. flavus , and to characterize isolates for aflatoxin production.

In this study, we documented the population densities of A. flavus across two agroecologies in eastern Zambia. Population densities of A. flavus varied amoung districts. The mean population density of A. flavus was 2.6, 1.8, 2.0, and 2.4 log CFU/g of dry soil in Chipata, Mambwe, Nyimba, and Petauke Districts, respectively. Mambwe and Nyimba Districts, which fall in the comparatively warmer Luangwa Valley, had lower populations of A. flavus than Chipata and Petauke Districts, located in the cooler and wetter plateau.

Several studies have documented population densities at different scales—from single fields to larger geographical zones, at different sampling frequencies, and also from fields with different cropping histories. Horn (2005) sampled two cultivated fields in Georgia at five different times and showed that for populations of A. flavus L strain, field A had lower mean densities at 2.7 log CFU/g compared to field B at 3.1 log CFU/g. A similar observation was also reported on population densities of A. flavus S strains from these fields (Horn, 2005). From an earlier study, also conducted in Georgia, Horn (1995) found that populations of A. flavus and A. parasiticus did not differ significantly at the beginning of the season and remained stable during the season in fields cropped to corn and peanuts. However, populations of A. flavus relative to A. parasiticus increased significantly in one drought stressed corn field after harvest and debris dispersal. In Virginia, which has relatively cooler weather, populations of A. flavus did not exceed 2.0 log CFU/g soil in peanut cultivated soil (Griffin and Garren, 1974).

Variations in population densities of A. flavus , both between and within districts were therefore expected. However, we expected higher populations in the hotter and drier Luangwa Valley compared to the cooler and wetter plateau. Several factors, such as flooding, could explain for the documented differences in population densities of A. flavus . Parts of the Luangwa Valley are prone to floods. Farmers interviewed during sample collection noted that some fields cropped to groundnuts had been exposed to recurrent flooding events, with water standing in these fields for a few days. For example, in the sampled Chikowa region of Mambwe District, the Kasengwa, Lwuo, and Kabilubilu rivers flooded during the 2011/2012 growing season, whereas in the Masumba region of Mambwe District, the Msandile river flooded affecting fields cropped to groundnuts and other crops. In such areas, it is common practice to plant along the flood plains of the rivers, because farmers try to take advantage of the fertile alluvial soils deposited whenever these rivers break their banks.

Population densities of A. flavus , an aerobic fungus found in the upper profiles of the soil (Horn, 2003), could therefore have been reduced because of these flooding events. Colony forming units estimated through soil dilution plating mostly arise from conidia and hyphae (Horn, 2003). Conidia are asexual spores that are hydrophobic and are mostly found in the upper 6 cm of the soil (Horn, 2003). However, in agricultural soils, Aspergillus spp. can be found up to 30 cm deep, possibly due to ploughing activities that result in mixing of soil at such depths (Horn, 2003). Flooding events could reduce the densities of both conidia and hyphae in the upper soil profiles by either depositing soils from the flooding river, or by washing down these propagules further down the soil profile.

Many papers on characterizing Aspergillus in both agricultural and non-agricultural fields have been published (Boyd and Cotty, 2001; Karthikeyan et al., 2009; Okoth et al., 2012; Solozarno et al., 2014). Boyd and Cotty (2001) reported that A. flavus and A. tamarii were the only species isolated from plant debris and pod samples collected from the dessert in Arizona. In contrast, Okoth et al. (2014), reported that A. flavus was the most abundant species from section Flavi isolated from corn in two districts in different agroecologies in Kenya. A. tamarii and A. parasiticus were also isolated. Interestingly, Okoth et al. (2014) reported that A. parasiticus and A. tamarii were isolated more frequently in the cooler district of Nandi compared to Makueni which is also receives significantly lower rainfall. Of the 91 Aspergillus spp. isolates characterized in our study, the most abundant species was A. flavus (47%), followed by A. parasiticus (21%). Other species identified were A. tamarii (7%), A. oryzae (9%), A. terreus (7%), A. nomius (7%) and A. niger (3%) (Table 1).

A. flavus populations are genetically diverse (Chang et al., 2005; Karthikeyan et al., 2009) and is evident in the different capabilities of isolates producing toxin (Karthikeyan et al., 2009; Okoth et al., 2014). There are many methods, both cultural and analytical that can be used to determine toxigenicity. Abbas et al. (2004) reported that cultural methods were equally as effective in determining toxigenicity and were cheaper than testing for toxin using HPLC or ELISA.

In our study 100% of the S-strain A. flavus isolates produced AFB₁ at 4 and 20 ppb, whereas 86% and 56% of the L-strain isolates produced AFB₁ at 4 and 20 ppb, respectively. Non-toxigenic isolates have sequence breakpoints in the aflatoxin gene cluster, and Chang et al. (2005) showed that such deletions are not a rare occurrence in natural populations. Recently, Solozarno et al. (2014) reported that 60% of A. flavus isolates from a corn field in Mississippi were toxigenic however they did not differentiate isolates based on S- and L- morphotypes. Interestingly, S morphotypes have been identified in Asia, United States, Africa, South America, but not in Europe (Boyd and Cotty, 2001; Okoth et al., 2014; Perrone et al., 2014). Karthikeyan et al. (2009) also tested isolates of A. flavus recovered from corn for toxin production and reported that 63% (32 out of 52 isolates) produced AFB₁ and the rest did not.

Our results also showed that 79% and 64% of tested A. parasiticus isolates produced aflatoxins at 4 and 20 ppb, respectively. A. nomius isolates were also tested for toxin production and all isolates produced aflatoxins at 4 and 20 ppb (Table 1). However, by adopting for this study, a cut-off lower detection limit of 4 ppb, we could be under-reporting on the percentage of toxigenic isolates present. Interestingly, A. flavus isolates may lose the ability to produce aflatoxin in-vitro following successive culture transfers (Sweany et al., 2011) and this too, could have affected our results. Therefore, to better determine toxigenicity, a combination of assays may be used. Abbas et al. (2004), reported that out of 517 isolates tested, combining cultural assays reduced false positives for aflatoxigencity to 0% and false negatives to 7%.

In this study, we also sought to identify and characterize for the first time in eastern Zambia, fungi from Aspergillus spp. isolated from soils cropped to groundnut. 91 Aspergillus spp. isolates, randomly selected from soil dilution plates (Fig. 1), were identified and characterized for aflatoxin production. We identified A. flavus (small (S) and large (L) sclerotial strains), A. parasiticus, A. niger, A. nomius, A. oryzae , A. tamarii, and A. terreus . A. flavus colonies were identified as greenish-yellow to yellow-white colonies on CzA (Fig. 2A), but were mostly green on MEA (Fig. 3A), and white-green colonies on PDA (Fig. 4A). Some A. flavus colonies were sclerotial with sparse hyphal growth on nutrient agar. A. flavus S- and L-strain isolates had sclerotia ≤400 μm or >400 μm in diameter, respectively. Under microscopic examination, conidia of A. flavus colonies were mostly smooth-walled and were borne on phialades, which were attached to metulae on globose vesicles (Fig. 5A). A. flavus is a common fungus in the soil and has been isolated in arid, temperate, and humid agroecologies worldwide (Abbas et al., 2004; Boyd and Cotty, 2001; Griffin and Garren, 1974; Horn, 2005; Horn 2003; Okoth 2014; Sweany et al., 2011)

In contrast, A. parasiticus colonies were pale green on both CzA (Fig. 2B) and MEA (Fig. 3B), and dark green on PDA (Fig. 4B). Under microscopic examination, conidial walls of A. parasiticus were echinulate and the stalks of conidiophores were rough-walled. Conidia were borne on phialades which lack metulae (Fig. 5B). A. parasiticus is also a common soil inhabitant which is mostly isolated in cooler temperate areas but is distributed worldwide (Domsch et al. 2007).

A. tamarii were comparatively easier to identify. Colonies of A. tamarii were chocolate-brown to cinnamon-brown on CzA (Fig. 2C and D), whereas on MEA, colonies were greenish with tints of brown (Fig. 3C), and green with more pronounced brown on PDA (Fig. 4C). Under microscopic observation, conidia of A. tamarii were echinulate and characteristically attached in chains, radiating from phialades lacking metulae (Fig. 5C). A. tamarii does not produce aflatoxin but produces another neurotoxin—cyclopiazonic acid (CPA) (Boyd and Cotty, 2001). The impact of CPA contamination in food is however an area not well researched into, especially in Africa. In the US, A. tamarii is mostly isolated from soil but occasionally isolated from plants (Horn, 2007). However, Horn (2007) discusses that the distribution of A. tamarii is not well defined in the US, but in contrast, Okoth et al. (2011) reported that A. tamarii was mostly found in the cooler areas of Nandi in Kenya, compared to the hotter and drier regions of Makueni.

For A. oryzae , colonies on CzA varied from yellow-white (Fig. 2E) to green-brown (Fig. 2F), whereas on MEA colonies were white (Fig. 3D) or green (Fig. 3E). Compared to A. flavus , conidia of A. oryzae were echinulate, and were borne on conidiophores with characteristically long stalks (Fig. 6A). In addition, unlike for A. flavus , A. oryzae colonies did not form sclerotia. A. oryzae is important in food processing in Asia where it is used to ferment products and in the making of soy sauce, sake, etc (Domsch et al., 2007). The fungus has also been isolated worldwide in soils from cultivated fields and also in agricultural products.

For A. nomius, colonies were green to yellow with white sectors on CzA (Fig. 2G), light green on MEA (Fig. 3F) and white fluffy colonies on PDA (Fig. 4E). A. nomius colonies formed sclerotia (Figs. 2G and 3F), mostly larger than 400 μm in diameter. Under the microscope, conidia were slightly echinulate with globose vesicles and had roughened conidiophores (Fig. 6B). There are few reports on the isolation and identification of A. nomius, a comparatively newer aflatoxin producing species. In the US, A. nomius is most reliably found in agricultural soils of the Mississippi Delta (Horn, 2007).

For A. terreus , colonies were cinnamon to light brown on CzA (Fig. 2H) and light green on both MEA (Fig. 3H) and on PDA (Fig. 4F). Under microscopic observation, vesicles of A. terreus were semi-spherical and had conidia arising from phialades attached to metulae (Fig. 6C). Conidia heads of A. terreus were strongly columnar arising from the top sides of the vesicles (Fig. 6C). Lastly, colonies of A. niger were characteristically black, both on CzA (Fig 2I) and on PDA. A. niger (Fig. 7) is an important spoilage fungus that can also attack young peanut seedling in the field causing crown rot (Melouk and Damicone, 1997), and produces ochratoxins (Perrone et al., 2007). Its importance in causing crown rot and reducing plant stand has not been documented in Zambia. A niger is ubiquitous in soil (Melouk and Damicone, 1997).

Despite the importance of aflatoxin contamination and increased focus on mitigation efforts, information on the identification of aflatoxigenic fungi in Africa is still limited. To our knowledge, this is the first peer reviewed report from Zambia documenting the population densities of A. flavus across different agroecologies. In addition, it is also the first report on the identification of different Aspergillus spp., such as A. nomius, A. terreus , A. oryzae , and A. tamarii from Zambia. However, our findings on the presence of these different Aspergillus spp. is not surprising since these fungi have been isolated in many different ecologies worldwide (Abbas et al., 2004; Boyd and Cotty, 2001; Domsch et al., 2007; Horn, 2003; Horn, 2005; Horn et al., 1995; Jamie-Garcia and Cotty, 2006; Monyo et al., 2012; Okoth et al., 2012). The significance of our research is the documentation of the presence of different toxigenic species, such as the S-strain A. flavus and also A. nomius, but also L- strain A. flavus and A. parasiticus.