Background

The genus Aspergillus, a member of the phylum Ascomycota, includes over 185 known species. Members of Aspergillus section Flavi are characterized by their ability to produce flavine-derived secondary metabolites which include the aflatoxins. Aflatoxins are mycotoxins associated with hepatotoxicity, mutagenicity and carcinogenicity in humans and animals [1, 2]. Species capable of producing aflatoxins include Aspergillus flavus, Aspergillus parasiticus, and several less common taxa such as Aspergillus nomius, Aspergillus tamarii, A. pseudotamarii, A. minisclerotigenes and A. bombycis [3, 4]. Other Aspergillus outside of the section Flavi are also known to produce aflatoxins and species with this ability have been shown to be more diverse than previously thought [5].

Grains of important food and export crops in Africa, such as maize and peanuts, are susceptible to contamination with different toxin-producing moulds. Of particular importance are aflatoxins that pose serious health risks to humans and animals. In Kenya, the toll on human health from periodic outbreaks that record case fatality rates of 39 % [6] remain common. In addition, food grains contaminated with aflatoxin-producing fungi also lose their quality based on nutritional value, taste, smell, appearance, and are no longer suitable for consumption or sale. The major challenge is that the contaminated cereals such as maize may appear just like normal grains without any outward physical signs of fungal infection [7].

The earliest reported case of aflatoxicosis in Kenya was in 1977, where several dogs and poultry died in Nairobi, Coast and Rift Valley Provinces after feeding on poorly stored grains, while in 1978, samples of human and dog food (flour) were found to be contaminated with [8]. In 1984, twelve people died in Machakos district of Kenya after consuming aflatoxin contaminated maize, while in 1995, about 172 samples of maize analyzed were found to have aflatoxin. In 2004, there was an outbreak of acute aflatoxicosis in Kenya, one of the most severe episodes of human aflatoxin poisoning in history. A total of 317 cases were reported by July 2004, with a case fatality rate of 39 % [9, 10]. This epidemic resulted from ingestion of contaminated maize [11]. Other outbreaks occurred in 1981, 2001, 2004, 2005, 2006, 2007 and 2008 resulting in sickness, death and destruction of contaminated maize [8, 9, 11–13]. Maize from the affected areas contained as much as 4400 ng g−1 aflatoxin B1, which is 440 times greater than the 10 ng g−1 tolerance level set by the Kenya Bureau of Standards. Most of the aflatoxin poisoning outbreaks occurred in remote villages and, therefore, the number of people affected could have been higher than reported [11]. The aflatoxin contamination in maize has been associated with drought combined with high temperatures as well as insect injury [14].

Aspergillus flavus and the closely related subspecies A. parasiticus have a world-wide distribution and normally occur as saprophytes in soil and on many kinds of decaying organic matter. They readily infest several important crops such as corn, cottonseeds, peanuts and tree nuts. They are widely distributed with greater quantities of the fungus occurring in warm climates [4, 15].

Gene flow within A. flavus is primarily limited by a vegetative compatibility system [16, 17] that delineates the species into numerous genetic groups called Vegetative Compatibility Groups (VCGs). A. flavus VCGs clonal lineages [18] that exist in complex communities are composed of many VCGs. VCGs vary in many characteristics; including aflatoxin-producing ability. Populations of A. flavus in an individual agricultural field contain isolates of many VCGs [4, 19]. VCGs with aflatoxin-producing potential are known to vary less among isolates within a VCG than among isolates from different VCGs [18]. Communities resident in different fields, areas, and regions may vary widely in average aflatoxin-producing ability [20–22]. Isolates and VCGs that do not produce aflatoxin, called atoxigenic, are common within A. flavus communities [20–23]. Surveys of A. flavus isolates from various geographic regions have revealed differences in the proportions of isolates that produce low, medium, and high amounts of aflatoxins [5, 24, 25]. For instance, in Argentina [26] and Iran [27] less than 30 % of the A. flavus isolates were capable of producing aflatoxin whereas in Nigeria the number was much higher, exceeding 50 % [28]. In the southern USA most A. flavus isolates produce aflatoxin [24, 25]. In the USA, there are currently several atoxigenic strains of A. flavus used to reduce crop contamination.

Biological control strategies directed at utilizing non-aflatoxin (atoxigenic) strains to limit aflatoxin contamination on crops have been pursued for almost two decades [29]. These strategies seek to give atoxigenic strains a competitive edge and exclude their aflatoxin-producing relatives, therefore decreasing the potential for contamination in crops and the environment. Successful strategies have been accomplished by using native non-aflatoxin-producing strains such as A. flavus [30].

In this study, we sought to identify the natural occurence of different strains/species of Aspergillus and to identify the ecological prevalence of both aflatoxin-producing and non-producing strains. Identification of the proportion of non-aflatoxin-producing strains in natural environments is an important step towards estimating their applicability in the biological control of the more medically and economically important toxigenic strains in Kenya.

Methods

Field survey locations

Field surveys were done in 6 agro-ecological zones (AEZ) in Kenya where maize and peanut crops are widely cultivated. The study covered 5 districts Makueni (1.8°S, 37.62°E), Makindu (−1°17′31.437″, 36°49′19.005″N/E), Kwale (3°3′4°45′, 38°31′39°31′), Kibwezi (−2°25′00″S, 37°58′00″E) and Msambweni (4°29′29.8″S, 39°28′33.7″E) (currently counties) selected due to their historical association with frequent and serious outbreaks of aflatoxin contamination of farm produce. To assess fungal populations, soil samples were collected from 57 fields arbitrarily chosen within the different districts and agroecological zones on the condition that fields were radially separated by at least 5–20 km from each other. A field was equivalent to a small-holder subsistence farm.

Sample collection and preparation

Soil samples were collected and processed as described by Dorner et al. [31]. From each field, three to five scoops were randomly taken using a sterile trowel from the top 4–6 cm of soil. Samples were transported in sterile containers to the Kenyatta University laboratory where they were sorted, pooled per field, oven-dried at 40 °C for 5 days, and stored in sealed plastic bags at room temperature (22–26 °C) until use. Aliquots of soil for pH measurements were suspended in distilled H2O (20 % soil, wt/vol).

Isolation and quantification of fungal species

For fungal isolation, soil samples were retrieved from storage, dried in a forced air oven (48–50 °C for 48 h), pulverized in plastic bags to remove clods and homogenized by hand-mixing.

Aspergillus section Flavi were isolated and quantified by the dilution plate technique on Modified Rose Bengal Agar (MRBA), [32]. In sterile 7 ml capacity polystyrene tubes, 1 g of soil was suspended in 3 ml of sterile water, mixed for 20 min on a rotary shaker, and plated on MRBA at appropriate dilutions. Dilution was to allow colony densities of less than 10 per plate. Plates were incubated in the dark for 3 days at 31 °C. Colonies of Aspergillus section Flavi were then identified by colony morphology. About 5–10 isolates per soil sample were transferred to 5/2 agar (5 % V-8 juice and 2 % agar, pH 5.2) and grown for 5 more days, unilluminated at 31 °C. Isolates were then classified on the basis of colony characteristics and conidial morphology at ×400 magnification. Aspergillus section Flavi colonies were identified by their characteristic growth pattern, retention of Rose Bengal within mycelia, and production of characteristic conidiophores after 3 days on MRBA. Isolates with abundant small sclerotia (average diameter <400 mm) were preliminarily classified as strain SBG [5] while isolates with smooth conidia and large sclerotia (average diameter over 400 mm) were classified as the L strains of A. flavus [33]. A. tamarii, A. nomious and A. parasiticus were initially identified by colony and spore morphology [3]. All preliminary identifications were confirmed by color reaction on AFPA (A. flavus and A. parasiticus agar) [34]. Quantities of Aspergillus section Flavi in soils were calculated as colony-forming units (CFU) per gram of soil. Several 3 mm plugs of sporulating culture were transferred to 4-dram vials containing 5 ml of sterile distilled H2O. These conidial suspensions were maintained at 4 °C for further analysis.

Aflatoxin-production by the fungal isolates

To determine the relative frequency of toxigenic and atoxigenic strains of Aspergillus section Flavi strain distribution across the country, isolates were randomly selected from each of the collection sites in order to determine their aflatoxin-production capability and also to establish the strains that produce the greatest quantities of aflatoxins and the frequency of occurrence of non-aflatoxin producers in Kenya. A total of 57 isolates (both L and S strains) were fermented in Adye and Matales medium with 22.4 mM urea as the sole nitrogen source and adjusted to pH 4.7 prior to autoclaving [5]. Vials (15 ml containing 5 ml A&M) were seeded with approximately 2 × 103 conidia suspended in 100 ml water. After incubation at 32 °C in the dark for 5 days, the pH of the medium was measured, 3 ml acetone was added, and the contents mixed by inversion. Vials were allowed to set for 1 h to allow lysis of fungal cells and extraction of aflatoxins from mycelia and conidia. Subsequently, the mycelia was collected on Whatman No. 4 filter paper, dried in a forced air oven (48 °C, 3 days), and weighed to quantify fungal biomass. The filtrate was diluted appropriately, spotted alongside standards of aflatoxin B1, B2, G1 and G2 (BORATEC limited, Kenya), and separated on thin layer chromatography (TLC) plates (silica gel 60, 20 mm) with the development solvent being a mix of diethyl ether, methanol and water in a 96:3:1 ratio. Aflatoxin was quantified in situ on the TLC plates with a scanning densitometer (model cs-930, Shimadzu Scientific Instruments, Inc., Tokyo). Extracts that initially did not show detectable aflatoxins were diluted with an equal volume of water and extracted with 3 ml methylene chloride to concentrate any latent aflatoxin. Aflatoxins were partitioned into the methylene chloride fraction, from which they can be retrieved by drying and the residues dissolved in 100 ml of methylene chloride. The solution was then subjected to TLC as already described.

Data analysis

Analyses were done with SPSS version 15.0 for Windows (release 15.0.0, Vista Hotfix Applied) and Microsoft Excel 2007. Pearson correlation coefficients were generated to assess relationships of ecological and biological variables with α = 0.05 and 0.01 for a 2-tailed t test. Analysis of variance was performed on all data with the general linear model (GLM), suitable for unbalanced data. The analyses for percentage values, CFU g−1, and aflatoxin concentrations were preformed with data transformed, using the arcsine of the square roots, the natural logarithms (ln), and the logs (count +1), respectively. Sampling sites and the agroecological zones were treated as class variables.

Results

Distribution of Aspergillus section Flavi in Kenya

Aspergillus section Flavi was detected in all the soil samples from 57 fields situated within the AEZs studied. A total of 220 section Flavi colonies were successfully transferred from MRBA to 5/2 agar and subsequently identified by macroscopic, microscopic and growth characteristics in AFPA medium. Out of all the Aspergillus section Flavi isolates, A. flavus was predominant (63.8 % of section Flavi), followed by A. parasiticus (28 %), A. tamarii (5.6 %) and about 2.3 % of the isolates assigned to A. nomius (a recently discovered strain among the Section Flavi) as shown in Table 1.

Table 1 Proportions of Aspergillus section Flavi of fields in five districts, soil pH, and colony-forming units per gram of soil in Kenya across the sampled agroecological zones
Full size table

The most common group (forming 85 % of the Aspergillus section Flavi isolated from the seven AEZs) were identified as L- strain with an average of 60.6 % in all the districts. The S- strains had a mean incidence of 39.4 %. Kwale district had the highest incidence of L strains (73 %) and the least of S-strain, while the reverse was true in Kibwezi district (L-strain being 48.5 % and S-strains being 51.5 %), (Table 1; Fig. 1).

Fig. 1
figure 1

Microscopic features of Aspergillus spp Conidial features (×400)

Full size image

The mean colony-forming units (CFU/g of soil) of the Aspergillus were extremely variable among the districts, ranging from 3.0 × 103 to 2.6 × 106. The CFU counts were significantly different between the districts. However, the CFU counts were not significantly different within the agroecological zones (p > 0.05), (Table 2).

Table 2 Pearson’s correlation coefficients of relationships among the quantity of Aspergillus section Flavi population in soil (CFU g−1), soil pH, AEZ and the isolates
Full size table

There was a significant positive correlation (0.594, p = 0.009) between pH and the proportion of S strains in the sampled colonies. On the other hand, the correlation of pH and the percentage of L strains was negative (−0.594, p = 0.005). The proportions of L and S strains in the CFU yield was negatively correlated (Table 2).

The mean soil pH among the districts sampled ranged from 6.12 in Kwale (weakly acidic) to 7.19 (weakly alkaline) in Kibwezi district (Table 1).

Incidences of toxigenic and atoxigenic strains of aspergillus section Flavi in Kenyan districts

A total of 57 isolates, each representing the collection sites were randomly examined for their aflatoxin-production potential. Out of all the isolates, 45.9 % were positive for different aflatoxins and were categorized as toxigenic. On the other hand, 54.1 % showed no detectable aflatoxins in their TLC profiles (Fig. 2), hence they were classified as atoxigenic strains. Msambweni district had the lowest atoxigenic/toxigenic strain ratio (41.7:58.3 %) while Makindu district had the highest atoxigenic/toxigenic strain ratio (75:25 %) (Table 3; Fig. 1).

Fig. 2
figure 2

An example of TLC profiles of Toxigenic and atoxigenic strains

Full size image
Table 3 Incidence of atoxigenic and toxigenic strains surveyed in all the 57 sites within the districts in Kenya
Full size table

Aflatoxin production by toxigenic strains and quantification

Aflatoxin-producing aspergilli were isolated from all the districts surveyed (Table 3). As expected, A. flavus and A. tamari produced only B aflatoxins while A. parasiticus produced both B and G subtypes. Aflatoxin-producing potential varied widely among isolates from the same districts, and agroecological zones (Table 3). Only two isolates of A. tamarii produced detectable aflatoxin (B). Aspergillus nomius isolates did not show any toxigenic potential.

The majority of aflatoxin producers were the L type members of section Flavi. However, the highest concentrations of aflatoxins were detected among the S type strains ranging from 2.25 × 105 to 4.65 × 106 ng g−1. Mean aflatoxin production varied significantly among the districts with the least being Kibwezi (4.0 × 103 ng g−1) and the highest being Kwale (4.65 × 106 ng g−1) (Tables 3, 4). The B aflatoxins were the most common metabolite (70.5 % as compared to the BG aflatoxins, 29.5 %) (Table 3).

Table 4 Aflatoxin production and quantification among isolates
Full size table

Discussion

Aspergillus section Flavi were detected in all 57 soil samples collected from various agroecological zones of maize and peanut growing areas in Kenya. This widespread distribution and prevalence conforms to previous findings of similar studies in Israel, Thailand, North America and West Africa [19, 28, 35]. Several members of the section Flavi including A. niger, A. fumigatus, A. clavatus, A. ochraceus, A. terreus and A. versicolor have been shown to be major mycotoxin-producing contaminants in various agricultural produce in Kenya. Our study supports a high incidence of Aspergillus section Flavi with A. flavus being the most predominant (63.2 %), A. flavus could indeed be linked to B1 aflatoxicosis contamination that has been shown to be prevalent as was determined by Probst et al. [9]. Members of A. parasiticus were less prevalent but nevertheless produced high concentrations of both B and G aflatoxins (Tables 3, 4). They, therefore, contributed significantly to the aflatoxicosis potential of resident fungal communities, a situation similar to that observed in West Africa by Donner et al. [28]. While all species produced B aflatoxin in different quantities, only A. parasiticus produced G aflatoxins among all groups studied in Kenya.

The negative correlation of pH and proportions of L strains could be an indication of the poorer adaptation of these strains to higher pH optima compared to the S strains. However, the adaptations of the two strains is not mutually exclusive as each of the two strains occur in different proportions within all the pH ranges observed. In addition, the range of soil pH variation across the examined regions was narrow, ranging from weak acidic to weak alkaline. The higher incidence of S strains in more alkaline soils has previously been observed in North American soils among isolates of A. parasiticus, A. tamarii and A. flavus [24]. However, in Nigeria [28], a negative correlation was observed. The effect of pH has been shown to play a lesser role in A. flavus population dynamics, with climate being cited as a more important determinant [35]. Therefore, other shared factors within the study areas such as temperature maxima of 40 °C, average humidity, semi aridity and transitional savannah ecosystems provide a range of biophysical conditions conducive for elevated levels of aflatoxin-producing fungi.

The aflatoxin-producing potential of Aspergillus communities is higher when S strain is present, as L strain isolates produce on average only 33 % of the toxin yield from S strain isolates [24]. In West Africa, S strain isolates produce greater quantities of aflatoxins than L strain isolates. This corroborates the finding of this study, in which the S strains produced the highest concentration of aflatoxins in Kenya. All the isolates of A. flavus (both L and S strains) did not produce the G aflatoxins but produced the B toxins. The percentage of A. flavus L-strain isolates that produced aflatoxins varied with geography and climate (Fig. 3).

Fig. 3
figure 3

Incidence of toxigenic and atoxigenic isolates of aspergillus section Flavi from soil samples of maize and pea nut growing areas in Kenya. S-strain L-strain

Full size image

Aflatoxin-producing fungi vary widely in many characteristics, including virulence for crops and aflatoxin-producing capacity [33]. Aspergillus flavus and A. parasiticus are most commonly implicated as causal agents of aflatoxin contamination. Aspergillus flavus has two morphotypes, the typical or L strain (sclerotia of >400 µm in diameter) and the S strain (sclerotia of <400 µm in diameter) [33, 36]. S-strain isolates produce more aflatoxins than L-strain isolates, on average. Many L-strain isolates produce no aflatoxins (“atoxigenic”) [24].

All members of A. flavus lack the ability to synthesize G aflatoxins due to a 0.8–1.5 kb deletion in the 28-gene aflatoxin biosynthesis cluster [18]. In contrast to cases in the United States, studies conducted in West Africa found out that an unnamed taxon (sometimes called strain SBG) is commonly implicated in contamination events [5]. Strain SBG is morphologically similar to the S strain of A. flavus, but DNA-based phylogenies reveal strain SBG to be a distinct species ancestral to both A. flavus and A. parasiticus [37, 38].

Aflatoxin levels unacceptable for human consumption may occur even in areas with relatively low frequencies of aflatoxin producers. In the current study, 59.3 % of aflatoxin-producing L-strain isolates produced more than 10 × 105 ng g−1 (14.3 ppb) aflatoxin B1. This combined with high incidences allows the L-strain to be the largest contributor to the average aflatoxin-producing ability of fungal communities in the five districts (Fig. 1; Table 4) and a potentially important causal agent of contamination in Kenya. Similar results were observed in Nigeria [28]. This is in contrast, with earlier studies conducted in Kenya where the S-strain was observed to be the primary cause of maize aflatoxin contamination [9]. The current study shows that S-strain is present in low frequency or absent in certain sites.

Incidences of atoxigenic strains of section Flavi varied widely among districts and agroecological zones (Table 1). The non-aflatoxin-producing strains are common in crop environments [21, 23, 24]. The atoxigenic strains of A. flavus and/or A. parasiticus have been employed as biopesticides directed at minimizing crop contamination with aflatoxins [4, 39]. Successful and effective biocontrol strategies necessitates that high ratios of atoxigenic to toxigenic strains be exhibited [40]. In the current study, high incidences of atoxigenic strains of section Flavi members were found in Kibwezi and Makueni districts both in lower midland zones (LM3, LM4 and LM5). These native populations provide potential biological control agents to manage the toxigenic strains.

Aflatoxin contamination of crops can be minimized by early harvest, prevention of insect damage and proper storage [41]. However, despite the careful management, unacceptable aflatoxin levels may occur from unpreventable insect damage to the developing crop or from exposure of the mature crop to moisture either prior to harvest or during storage in modules, handling, transportation or even use [33].

For many diseases, traditional chemical control methods are neither always economical nor are they effective, and fumigation as well as other chemical control methods may have unwanted health, safety and environmental risks in Kenya and many other developing countries. The antifungal abilities of some beneficial microbes have been known since the 1930s, and there have been extensive efforts to use them for plant disease control since then. However, they are only now being used commercially [33].

Aflatoxins cannot be readily removed from contaminated foods by detoxification. Therefore, there is interest in developing a biological control method that can increase crop safety by decreasing toxin content and this depends on the competitive displacement of toxigenic isolates using atoxigenic isolates of the same species. It has also been reported that aflatoxin production is inhibited by lactic acid bacteria, Bacillus subtilis and many moulds. This inhibition may result from many factors including competition for space and nutrients. In general, competition for nutrients required for aflatoxin production but not for growth and production of anti-aflatoxigenic metabolites by co-existing micro-organisms [42].

Currently, atoxigenic A. flavus L-strain isolates are used to competitively exclude aflatoxin producers during crop infection and thereby limit contamination in U.S. agriculture [4]. Such atoxigenic strains are highly effective against the S strain [4]. Adaptation and deployment of similar technologies in Africa could provide a promising strategy for prevention of future aflatoxicoses in East Africa while enhancing export possibilities for maize and peanuts [43].

Aspergillus section Flavi was resident in all sampled maize fields and quantities of section Flavi were higher on average in Kenya and compares to the ones in West Africa [35]. The incidences of the section Flavi in Kenyan soils demonstrates the fungal growth on crop-associated with organic matter. Corncobs, peanut pods and other crop debris harbor section Flavi for at least 3 years after harvest [44].

This study has reported high level of atoxigenic species of Aspergillus section Flavi. In vitro analysis showed that there is a growth competition between atoxigenic and toxigenic spp of Aspergillus section Flavi. The existing methods of biological control of aflatoxicosis by use of the atoxigenic strains can therefore be tested and adapted to Kenyan agriculture. The study has also established that the production of aflatoxins by these species depends on the prevailing physiological conditions such as temperature and moisture. Poor management practices such as poor storage, late harvesting, among others are predisposing conditions for the proliferation of the fungi, so there is need for education on the best practices so as to minimize on contamination. Furthermore, morphological analysis of Aspergillus section Flavi is limited with mechanisms of getting the true identity of the fungal species. There is need to adopt rapid and sensitive molecular techniques to support the morphological identifications. Finally, there is need to screen all the farm produce for possible aflatoxin contamination to reduce economic burdens related to the menace. We, therefore, hope this study will be taken into consideration by regulatory authorities in Kenya and beyond.

Conclusion

In conclusion, we report an ubiquitous presence of the members of the Aspergillus section Flavi within maize and peanut growing fields in Kenya. The diversity is characterized by a significant presence of atoxigenic A. flavus and A. parasiticus. Such atoxigenic variants have been previously applied in the biological suppression of natural populations of toxigenic strains and thereby minimizing aflatoxin contamination in peanuts [40], corn [45] and cotton [4]. The existing methods of biological control of aflatoxicosis by use of atoxigenic strains can therefore be tested and adapted to Kenyan agriculture. Control of natural populations of aflatoxin-producing fungi will lead to overall reduction of aflatoxicosis in farm produce as reported to be common in the regions studied [7, 9].