Natural occurrence of Fusarium species and fumonisins in stored cassava chips

Cassava chips are transformed products obtained following the fermentation and drying of cassava fresh roots. Once produced, chips can be stored for more than 180 days in conditions conducive for the development of toxigenic moulds. In this study, the incidence of Fusarium spp. is assessed from a set of 72 home-stored samples of cassava chips collected from farmers, during a 2-month monitoring survey period. Results from this survey enabled the recovery of 298 isolates of Fusarium spp. The Fusarium taxa detected included: Fusarium chlamydosporum, Fusarium equiseti, Fusarium verticillioides, Fusarium oxysporum, Fusarium pallidoroseum and Fusarium solani. F. oxysporum was associated with the highest level of isolation frequency (25%), whereas F. solani isolated was scanty (5%). Their level of recovery was increasingly important as the moisture content of samples increased and/or the product was stored for longer periods. Specific associations were observed to exist between the most frequently isolated Fusarium spp. The level of fumonisins further assessed showed that only 5 samples hosted this mycotoxin at concentration levels ranging between 0.22 and 1.7 mg/kg. Among the parameters used to assess its incidence, only storage duration showed significant (P<0.05) relationships with fumonisins. The occurrence of fumonisins as natural contaminants of cassava-based products is reported here for the first time. This study shows that no known fumonisin producingFusarium species was related to toxin occurrence in the samples collected. It can be hypothesized that any species of Fusarium infesting the samples studied could have the potential capability of producing the toxin as a result of a possible horizontal transfer of functional fumonisin gene clusters from an ancestral gene. Consequently, the present study suggests both investigations on genes responsible for plausible fumonisin formation associated with the Fusarium spp. identified from the samples collected as well as the various factors allowing their expression.


INTRODUCTION
On a global scale, the contamination of agricultural products by Fusarium spp. is common and has been reported both in cold and warm ecologies (Allen and Kochman, 2001;Nganje et al., 2001;Poppenberger et al., 2003;Davis et al., 2006;Lemmens et al., 2005), suggesting that fungi in this genus can grow in a wide array of natural substrates.In general, Fusarium spp.have endophytic functions in certain plant materials such such as maize leaves and roots of lawn grass (Anonymous, 1998;Bacon et al., 2008;Latiffah and Chua, 2013), which are antagonistic to other Fusarium spp.(Van Wyk et al., 1988;Bancy et al., 2014), and have teleomorphs (if known) in Gibberella and other genera (Marasas et al., 1984;Leslie et al., 2004;Gagkaeva, 2008).However, one of the most significant aspects of this genus is related to its role as a plant pathogen for a large range of botanical species.To this effect, fungi in the genus Fusarium have been reported to cause damping-off and root rot of young conifer seedlings (Jane et al., 2006).In palm trees (Elaeis guineensis), banana (Musa spp.), cotton (Gossypium spp.) and several other crops of agricultural importance, they cause wilts, whereas pre-and post-emergence blights are often reported in cereals and other graminae (Marasas et al., 1984;Davis et al., 2006).In all respects, the threat connected with Fusarium spp.could be viewed in a multifaceted way: besides their implication in yield losses of agricultural crops, their role as deteriogens and significant mycotoxin producers in several food commodities has been evidenced in some studies (Kokkonen et al., 2010;Rubella and Kistler, 2005;Moss, 2002;Msikita et al., 1996).
According to the current literature, mycotoxins are natural secondary metabolite compounds of low molecular weight produced by some fungi in food and feedstuffs.They are toxic at low concentrations when consumed by humans and animals, causing diseases referred to as mycotoxicoses (Christensen and Meronuck, 1986;Bennett and Klich, 2003).For these reasons, mycotoxins are now recognized as an important health hazard in several countries worldwide (Anonymous, 1995a;FAO, 2001).
When Fusarium-toxins are considered, they have been associated with a wide array of health problems in animals and humans.During the past four decades, at least three important groups of mycotoxins have been connected with the genus Fusarium.These include: trichothecenes (T-2 and HT-2 toxins, diacetoxyscirpenol (DAS), deoxynivalenol (DON) and nivalenol (NIV)), zearalenone (ZEN) and fumonisins (Marasas, 1994;FAO, 2001;Bennett and Klich, 2003).
In all respects and based upon toxicological information gathered from several studies, fumonisins and especially fumonisins B1 are classified by the International Agency of Research on Cancer (IARC) as Group II carcinogen compounds, hence suggesting that these fungal toxicants are possibly carcinogenic to humans (FAO, 2001;IARC, 1993IARC, , 2002)).Accordingly, special care is to be devoted to food and feedstuffs contaminated with Fumonisinproducing Fusarium spp. as a result of their potential negative health impact on animals and/or humans consuming such products.Although, Alternaria alternata (f.sp.Lycopersici) has been implicated in the formation of fumonisin-like compounds in foodstuffs (Marasas et al., 2001), production of this mycotoxin as a result of the infestation of agricultural crops and food commodities by Fusarium spp.such as Fusarium graminearum, Fusarium nygamai, Fusarium proliferatum and Fusarium verticillioides is widely realized (Marasas, 1994;Moss, 2002).The earlier works by Gelderblom et al. (1988) which emphasised the role of Fusarium moniliforme (synonym: F. verticillioides (Sacc.)Nirenberg), as responsible for fumonisin formation in maize sampled in the Transkei province in South Africa seems, to our best knowledge, to be the first document describing and characterization of fumonisin occurrence in natural substrates.The major toxic effects of these mycotoxins in the lung, nervous system, liver and kidney are largely documented and mainly connected with their role in the disruption of sphingolipids metabolism (Wang et al., 1991;Moss, 2002;Marasas et al., 2004;Avishay-Abraham, 2005).Accordingly, fumonisins have been identified as the main etiological agent of leukoencephalomalacia, a fatal neurotoxic syndrome in horses and other equines (Gelderblom et al., 1988), whereas pulmonary edema has been reported in swine and apoptosis in rats (Harrison et al., 1990;Moss, 2002).In man, these fungal metabolites have been connected with cancer-promoting activities and especially in the oesophageal cancer occurrence in some regions of South Africa and China (Marasas et al., 1988;Peraica et al., 1999).Additionally, correlations between dietary exposure to fumonisins and neural tube defects have been established in some studies (Moss, 2002;Avishay-Abraham, 2005).
In 1995, following extensive studies conducted in some African countries, attention was drawn by food scientists of the international community on the deterioration of some stored food products such as maize (Anonymous, 1995a) and cassava (Anonymous, 1995b) as a result of their infestation by microbial entities among which are, toxigenic fungi.
As one of the main staple food cultivated in Africa, cassava bears high social and economic relevance.The crop ranks first in importance among major starchy foods consumed in sub-Saharan Africa (FAO, 2011).It is a cheap source of carbohydrates for an estimated 500-600 million people living in this part of the African continent (Nweke, 1998).Furthermore, its increasing role as a substitute for imported cereals has been acknowledged (FAO, 2011;Anaeto and Idighibe, 2011).Due to their short shelf life, cassava roots, the most widely edible part *Corresponding author.E-mail: germainessono@yahoo.fr.Tel: (237) 674275731.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License of this food crop cannot be stored for more than 4 days as a result of post-harvest physiological deterioration (Hahn, 1993).This intrinsic characteristic of the crop has given rise to the development of transformation techniques aimed at reducing food losses and increasing its market value (Hahn, 1989).Cassava chips represent such transformed products.They are obtained following the fermentation and drying of fresh roots, resulting either in small pieces 2 to 3 cm long, referred to as cassava pellets, or circular balls with a diameter more often greater than 10 cm.Previously, Anonymous (1992) described this food commodity as forming the bulk of cassava-based diets in Western and Central Africa.Surveys carried out in southern Cameroon ascertained the widespread importance of cassava chips in the food habits of local populations (Essono, 2008).In the frame of these surveys, observations were made that chips can be stored for more than 180 days and in conditions conducive for spoilage by a wide array of harmful microbes.In previous studies, a heavy prevalence of toxigenic fungi (among them Fusarium spp.) in stored cassava products has been reported (Anonymous, 1995b;Msikita, 1996;Essono et al., 2007).
In all these respects, because Fusarium spp.are important in plant pathogenesis, and responsible for a great deal of deterioration in stored food products, with subsequent formation of secondary toxic metabolites, this study was carried out with the aim of assessing the nature and relative importance of some species in this genus likely to occur in batches of cassava chips, a basic raw material commonly associated with the bulk of cassava-based foods consumed in Cameroon.However, emphasis is placed on the incidence of F. verticillioides (Sacc.)and fumonisins in the samples collected.

Study site and rationale for its choice
From December 2010 to March 2011, surveys were conducted in 45 villages of southern Cameroon (Central Africa) to collect information on processing practices and constraints from farmers cultivating cassava and transforming them into chips.These villages were chosen following a detailed macro-level characterization of the zone mentioned above and which led to the delimitation of a benchmark area.The benchmark concept is linked to the idea that there is a range of geographical and socioeconomic characteristics which can affect the habits of people in a given setting.In this respect, the most important criteria used to construct a framework for site selection in the benchmark area of southern Cameroon included: geographical distribution of main local crops, human population densities and infrastructural development with the emphasis on accessibility to major or nearest cities. Accordingly, it was assumed that climatic differences may strongly influence growth, production techniques, biological factors and processing methods associated with the cassava crop.Accessibility to major towns may also be an important factor influencing the degree of commercialization of cassava-based products, whereas population densities may change the pattern of resource uses such as labor allocation, fallow length and choice of crops.All these factors can in turn affect processing technologies as well as the level of consumption frequency of derived cassava products by the rural population.During the course of these surveys, different processing methods resulting in different chip types were observed, therefore suggesting probable differences in their microbial profiles.Within the areas surveyed, two villages were identified as producing the greatest quantities of cassava chips.Mengomo in the zone of Ebolowa was specialized in the production of cassava balls, whereas Nkometou III in the Yaoundé zone, mainly produced chips in the forms of pellets.In both zones, the annual rainfall is distributed in a bimodal pattern with the greatest accumulation in September-October and April-May, averaging 1876 mm at Ebolowa and 1654 mm at Yaoundé.

Sample collection and preservation
From June to August 2012, the two villages mentioned above were revisited during a second survey aimed at collecting home-stored samples of cassava chips from farmers and to monitor their fungal flora composition over time.Eligible farmers were those who had been residing in the village for the previous 24 months with farming activities.Within each village, six farmers referred to as working partners were randomly selected along a transect cutting across the village, and requested to produce cassava following their normal routine, with a minimum of instructions.Hence, a total of 12 working partners were identified.Chips produced by each working partner were referred to as "LOTS" that were to be harvested during two consecutive months.For each "LOT", the 2-month monitoring period was divided into six phases of sample collection: weekly collections were carried out for the first 4 weeks of storage then, once every 2 weeks from the fourth to the eighth week following processing and storage.Processing practices, storage facilities, chip type and the periodicity of collection were used as sampling criteria (Table 1a and b).Collection of samples was done from the top to the bottom of the sample package so as to obtain composite samples.A total of 72 composite samples each weighing 4 kg were obtained.Samples collected were sealed in polyethylene plastic bags, sub-divided into four 1 kg batches, and stored at 4°C in a cold chamber for up to 6 weeks.The first batch was kept aside as a backup; the second was directly used for water content determination.The third was used to estimate by sieving the insect population per unit of weight of cassava chips.The fourth for mycological analyses was preserved in sterile plastic bags at 4°C in a cold chamber.Water content, pH and fungal flora analyses were carried out on the day of collection.After determination of fungal flora composition, samples intended for mycoflora analyses were ground to a fine and homogeneous flour using a blender (Warring Commercial Blender), dried at 60°C for 48 h in an oven (Model: Gallenkamp Plus II), then stored in sealed sterile plastic bags at 4°C and 20% relative humidity, until use for further fumonisin analysis.

Moisture content and pH determination
At collection period, three 100 g samples were weighed out from each second batch of cassava chips using an analytical balance (Model: Sartorius CP 225D).These were dried at 60°C for 72 h to determine their moisture content (MC) on the basis of weight loss.pH was determined as follows: 5 g of chips from each third batch sample collected at a given period were ground to flour.The flour was mixed with 50 ml of sterile deionized water adjusted to pH 7.0 with 0.1 M KOH or 0.1 M H2SO4.The mixture was covered with Table 1a.Information related to cassava chips samples (pellets) collected (a) from farmers in Nkometou III (Yaoundé).

Isolation and identification of Fusarium species
Fusarium spp.were isolated from chips using the direct plating method on a selective pentachloronitrobenzene (PCNB) medium.Preliminary experiments (data not shown) carried out on a few samples in the frame of this study using surface sterilization with NaClO as suggested in some studies (Msikita et al., 1996), hardly allowed the recovery of mycoflora from sample of chips even when such samples were observed overgrown with fungi.We attributed this to the lack of a protective shell in cassava chips as it is the case for some food products such as maize, groundnuts, rice and other  similar food products.Such a protective barrier is useful in limiting or preventing the sterilizing substance used at an appropriate concentration, from diffusing into the inner parts of the analyzed product and killing the internal mycoflora.Heat sterilization using fire flame was further attempted, but did not show any significant differences in the recovery of fungal species as compared to the non-sterilization system finally adopted in this study.Although Sauer and Burroughs (1986) argued that the absence of sterilization is likely to bring about some overestimation in the actual internal microflora population when portions of studied samples are to be plated onto culture media for analyses, since it is reportedly known that mycotoxin formation in food products contaminated by toxigenic fungi requires a minimum of 48 h (Christensen and Meronuck, 1986), it was assumed that chips which were observed to be often dried for at least 2 days before storage, might have been contaminated with toxigenic fungi and their related toxin during this time frame.As a result, no prior disinfection of samples of chips collected was carried out.
For isolation procedures, cassava chips were arbitrarily selected from batch 4 of each sample package at a given collection period.These were split with a sterilized knife into small pieces (10 x 10 x 5 mm) and placed in the laminar flow hood.Inside the laminar flow cabinet, five pieces were removed with sterile forceps and plated in triplicates onto 9 cm diameter Petri dishes (at the center and across two perpendicular diameters of the Petri dishes) containing 20 ml of PCNB medium adjusted to pH 4.5 with 0.1 M sulfuric acid to suppress bacteria.The plates were incubated for 7 days at room temperature (25°C), with a 12-h photoperiod.
Following fungal isolations, single spore cultures of putative Fusarium spp.were purified onto full strength potato dextrose agar (PDA).Resulting isolates were examined macroscopically and microscopically according to Leslie and Summerell (2006).Isolates assigned to species were simultaneously shipped to International Mycological Institute (IMI) in the UK and also to the Program on Mycotoxin and Experimental Carcinogenesis (PROMEC) in South Africa for confirmation of identification (Table 2).Representative isolates of formerly identified Fusarium spp. or groups were transferred to slants containing full strength PDA.These were incubated as previously described for 6 days before storage at 4°C in a cold chamber.Further, Fusarium isolates were identified taking into account macroscopic and microscopic analogies with specimens obtained from IMI and PROMEC.The mean percentage occurrence of these fungi was recorded for the total number of chips plated to assess the pollution rate of each species.
The total number of Fusarium spp.isolates recovered from pieces of cassava chips belonging to samples obtained in both locations of collection were used to assess possible associations between all species during the 2-month monitoring period.

Fumonisin analysis
A direct competitive enzyme-linked immunosorbent assay (ELISA) based-method using fumonisin standards (Veratox® Neogen Corporation Langsin, MI) was used for fumonisin analyses.Since ELISA assays measure total fumonisin content and not individual subtypes, the relative proportions of different known fumonisins (FB1, FB2, FB3 and FB4) were not assessed.Extraction of fumonisins was carried out by blending 50 g of ground cassava chips resulting from batch five with 250 ml methanol : water (70:30 v/v) solvent for 2 min using a high speed blender (Warring Commercial Blendor).The solution was filtered through a Whatman No 1 filter paper.The resulting filtrate, known as the working sample, was collected in a sterilized 100 ml conical flask.Determination of fumonisins was carried out according to the specifications of the manufacturer in a microplate ELISA reader (Dynatech MR 250, Dynatech Inc, Guernesey channel, Islands), using fumonisins standards at 0, 1, 2, 4, and 6 mg/kg as controls.The detection limit of the method was < 0.2 mg/kg.

Statistical analysis
The proportion of chips within samples contaminated by Fusarium spp. was assessed after the cultures have been examined.The total number of samples contaminated (NSC) by each species and the total number of isolates (TNIL) of individual Fusarium spp.recovered from samples per location were recorded.The relative index frequency (RIF) of each Fusarium spp. was computed as the ratio of the number of isolates of individual Fusarium spp.obtained from each location over the total number of all Fusarium isolates obtained from chips within the same location and expressed as percentage.To increase normality, percentage data (X) were transformed to the Arc sin using the function Y = Arc sin(X/100) 1/2 before analysis (Gomez and Gomez, 1984).
To assess differences in the incidence of Fusarium spp.isolates among samples, data were subjected to analysis of variance (ANOVA) using the general linear model (GLM) procedure (Anonymous, 2011).Storage duration, chip type or location of sample collection, moisture content and pH were used as class variables.Pearson correlation coefficients (r) were further computed to derive possible associations between Fusarium spp.isolated from samples collected in both locations during the monitoring period, and also between the content in fumonisins and the parameters (moisture content, pH and storage duration) taken into consideration in the study.
All analyses were carried out with the Statistical Analysis System software (SAS Institute, Inc Cary, NC, USA, V.9.2 (32)).
Fusarium chlamydosporium and F. solani occurred only at Nkometou (Yaoundé), while F. equiseti, F. verticillioides, F. oxysporum, F. pallidoroseum and F. solani were isolated from some samples obtained in both zones.Numbers (Ncpc) and percentages (RIF) of isolates of each species from each location are given in Table 3.
F. oxysporum was the most commonly isolated species, and accounted for over 25% of the total fungal isolates belonging to this genus.It contaminated almost an equal number of samples in both zones although more isolates were obtained at Ebolowa.F. verticillioides, a fumonisin-producing fungus, and F. pallidoroseum were also common in both locations.Unlike F. oxysporum, they were not equally distributed in samples and locations.The level of recovery of F. verticillioides ranged from 14 isolates in 6 samples at Nkometou (Yaoundé), to 42 isolates in 16 samples at Mengomo (Ebolowa).Comparable figures were observed for F. pallidoroseum in both sites.F. equiseti was isolated from 20 samples overall and the maximum number of samples (18) hosting this fungus was observed in Yaoundé with only two samples in Ebolowa.For more details, Table 3 shows the distribution of the remaining Fusarium spp.
Throughout the monitoring period, the pH of samples varied from 6.08 to 3.93 at Nkometou (Yaoundé) and from 6.36 to 4.11 at Mengomo (Ebolowa) at the same time the moisture content ranged from 4.3 to 11.9% at Yaoundé, and from 4.5 to 46.3% at Ebolowa (Table 1a  and b).The analysis of variance performed on the species detected showed that their level of recovery was affected only by the pH, the moisture content and the duration of storage.The location of collection did not appear to have any significant relationship with the species isolated, suggesting that perhaps the specimens obtained were affected by the same environmental parameters (Table 4).
Correlation coefficients computed between pairs of these fungi based on total isolation figures obtained in both locations indicated significant associations.In both zones, and with the exception of F. solani, F. verticillioides was negatively related to the rest of Fusarium spp.isolated.But these relationships were not always significant.Similarly, F. oxysporum was positively associated with the majority of Fusarium spp.occurring in the samples examined (Table 5).

Incidence of fumonisins in the samples collected
Table 6 shows the pattern of evolution and distribution of F. verticillioides isolates within samples during the monitoring period across the two locations, whereas Table 7 shows the level of fumonisins associated with the samples analyzed.The results obtained suggest that five samples out of 72 were positive for fumonisins occurrence.The toxin was detected in chips starting from week 1 following storage, and the mean concentration in positive samples varied from 0.22 to 1.7 mg/kg.An important observation associated with these results is that all the samples containing fumonisins were not found to be contaminated with F verticillioides and even the sample most contaminated with F. verticillioides (33.3%)  (sample No 5 belonging to Lot No 11 for the sixth week of storage) in Table 6 was not found positive for occurrence of fumonisins at detectable levels.Using correlation analysis, the level of fumonisins detected in positive samples was assessed for their possible associations with moisture content, pH and storage duration.The results obtained suggest that no clear relationship was found between the fumonisins content and the majority of parameters used in this study to assess its formation.The relationship between the content of fumonisins and the moisture content was negative (r = -0.15)and non-significant.There were also positive but non-significant associations between the level of fumonisins and pH (r = 0.045).Only the relationships between the level of fumonisins and storage duration were significant (P < 0.001) and positive (r = 0.32) (Table 8).

DISCUSSION
This study is one of the first, to the author's knowledge, to report the occurrence of both Fusarium spp.and fumonisins in stored cassava products.In this respect, the six Fusarium spp.found contaminating the samples of cassava chips are well-known contaminants of several stored food products, and have been isolated from stored and pre-harvest food commodities elsewhere.In similar investigations, Bottalico et al. (1995) reported the occurrence of 10 Fusarium spp. in stored maize (Zea mays) and wheat (Triticum spp.) sampled in Italy, among which 4 species were associated with the present study chips samples.Recently, Ali et al. (2014) also reported the occurrence of seven Fusarium spp.from freshharvest and stored rice samples in Iran, including some of the species associated with the present study.Their recovery from stored chips is an indication that this food commodity is a suitable substrate for their growth and subsequent development.Consistent relationships were observed to exist between the pH, the moisture content, the duration of storage and the level of recovery of Fusarium spp. in the chips.This shows that the moisture contents connected with our samples fell within the range of those conducive to fungal development and that variations in pH, occurring in the chips during the monitoring period were sufficient enough to affect the level of recovery of the fungal specimens isolated.Changes in moisture contents and pH are important factors regulating the level of microbial entities in food commodities, and have been reported.While assessing the microbial composition of stored tubers of C. esculentus, Adebajo (1993) indicated significant relationships between the microflora incidence and both pH and moisture content and observed that the recovery of fungal propagules was important as the pH decreased and the moisture content increased.Drops in pH with time are a common feature in carbohydratebased products, and are related with organic acids produced by fungi during microbial metabolisms (Essono, 2008).The results obtained in the course of this study seem to support both hypotheses.
Although the level of recovery of Fusarium isolates varied little across the collection sites, the incidence of the species recorded was not equally important at all locations.These results are consistent with those obtained by Jeschke et al. (1990) following a study associated with Fusarium spp. in the Transkei region of South Africa.They reported a differential incidence in the occurrence of the Fusarium spp. that they isolated, suggesting for example that F. oxysporum was more abundant where F. equiseti had a low level of occurrence.They attributed this to a possible competition between species and, more importantly, to the mode of survival associated with each of the species they isolated.In this study, important relationships were identified between the Fusarium spp.isolated.Positive and significant associations were observed between F. oxysporum and F. equiseti and between F. oxysporum and F. pallidoroseum (Table 2).This coexistence could be due to the facts that these species have similar optimum temperatures (Burgess et al., 1988), share in common the potential capability to survive through chlamydospores (Leslie and Summarell, 2006), and due to the fact their colonies can grow together without apparent inhibition (Burgess et al., 1988).Conversely, there were negative but non-significant associations between F. verticillioides and the rest of the Fusarium species except F. solani.In this study, F verticillioides (formerly F. moniliforme) known to survive through thickened hyphae (Nyvall and Kommedahl 1968), was significantly associated with samples having lower moisture content levels.
From a mycotoxicological view point, the results obtained in this study show that mycotoxins can be detected from cassava-based products when present.The range of fumonisins detected varied from 0.22 to 1.7 mg/kg in positive samples.However, samples containing F. verticillioides were free of fumonisins, while only five samples from which F. verticillioides was not detected were positive for this mycotoxin.In this respect, 2 out of 6 "LOTs" from Mengomo were positive for fumonisins, while none was found in the 6 "LOTs" from Nkometou.It could be hypothesized that process C might be more conducive to fumonisin formation than processes A and B, and that these two latter might be more effective ways for storing cassava chips in a manner that minimises fumonisin production in storage (Tables 1a and 1b).Furthermore, albeit fumonisin was found only at day 7 in "LOT" 9 while they were found in increasing amounts from day 21 forward in "LOT" 8, it could be hypothesized that an heterogenous distribution of fungi and fumonisins or the randomness of the sampling procedure have affected the present results.
In the course of this study, F. verticillioides was not found in fumonisin positive "LOT" 9 at any time point.This suggests that the absence of toxigenic fungi in any food product does not guarantee that such a product is mycotoxin-free since the toxin may, as Christensen and Meronuck (1986) stated, persist after the fungi have disappeared.In addition, identification reports from South Africa indicated that the majority of isolates of F. verticillioides recovered from our samples were probably not fumonisins producers or were weak producers (Marasas, personal communication).More importantly, this might also imply that fungi other than F. verticillioides could have been responsible for fumonisins formation in the samples examined.
In recent years, formation of fumonisins by Fusarium species in food and feed substrates has been critically reviewed.The potential capability to synthesize fumonisins has in this respect, been reported to bear a discontinuous distribution across the genus Fusarium, as a result of the presence or absence of the fumonisin (FUM) clusters among various Fusarium species isolates (Proctor et al., 2004), and this genomic approach has been extensively used to explain the occurrence of fumonisins in several food substrates.
Among the six Fusarium spp.isolated from the samples collected, only F. verticillioides, and F. oxysporum are often connected with fumonisins formation in several food and feed substrates, whereas F. chlamydosporum, F. pallidororoseum, F. equisetti and F. solani are reported not to produce any fumonisin at all.However, according to recent advances in fumonisin research, the possibility that the above mentioned species could produce the toxin cannot be excluded.This hypothesis is based upon the assumption that there exists close relationships between secondary metabolites biosynthesis and gene clusters of several mycotoxin producing-fungi as a result of natural selection procedures probably related to a horizontal transfer originating from an ancestral cluster.In past studies, Desjardins and Proctor (2007) sequenced genes of a few Fusarium spp.and indicated that the fumonisins biosynthesis is usually carried out in several steps, and identified in this respect a total of 17 genes as responsible for the process.These genes are often organized as clusters, suggesting that they could probably be either co-transcribed, or transmitted together while still remaining functional.Although further research suggested that not all Fusarium spp.do not possess the fumonisin cluster, there is however increasing evidence indicating that a wide array of Fusarium spp.formerly considered as not connected with fumonisin production, now possess a functional cluster related to fumonisin biosynthesis.
Additionally, Aspergillus niger, was recently claimed to produce fumonisins in some food commodities (Frisvad et al., 2007;Mogensen et al., 2010).This species has also been identified as containing several clustered homologs of many of the F. verticillioides FUM genes upto-date identified (Nora and Wolfe, 2011).Prior to these findings, Desjardins and Proctor (2007) had already proposed the hypothesis of a horizontal transfer of the FUM cluster from a distantly related Euascomycete into A. niger.In the frame of the present study, over 60% of the samples collected from farmers in the locations investigated were found contaminated with A. niger (data not illustrated).Since none of the F. verticillioides containing samples was fumonisin positive, it can also be hypothesised that probably, the amounts of fumonisin detected in the samples analysed could possibly be due to A. niger.Accordingly, the lack of assessment of correlation relationships between A. niger-containing samples and fumonisin positive-samples could be considered as one of the major limitations of the present study.
The knowledge of fumonisins as toxic compounds associated with harmful effects in animal or human health is relatively recent.Gelderblom et al. (1988), related outbreaks of Equine Leucoencephalomalacia (ELEM), a fatal neurological disease in horses and equines with the consumption of maize contaminated by F. moniliforme (currently F. verticillioides).The International Agency for Research on Cancer (IARC) (Anonymous, 1993;2002), also indicated that fumonisins were associated with cancer-promoting activities in humans.However, the values in fumonisins levels connected to the samples dealt with in this research were not too high when contents often quantified in other food commodities such as maize, and other cereal-based foods are considered (Ritieni et al., 1997;Anonymous 2013).Although, the maximum tolerable levels of these mycotoxins in food commodities have not been internationally defined to the best of the authors' knowledge, the present results call for attention to the potential of stored cassava products as a suitable substrate for fumonisin formation.
In the present study, the samples collected throughout the research sites were analyzed only for fumonisins.It should however be recalled that other Fusarium toxins are also of concern and may be present in stored cassava chips.Both F. equiseti and F. oxysporum were isolated from our samples and share together the potential capability of producing the toxin moniliformin (Marika 2008).According to the same author, F. pallidoroseum has also often been connected with the production of the toxin beauvericin, so far known as associated with highly toxic effects in human cells, resulting in apoptosis (Ritieni et al., 1997).
In all respects, and based upon the present findings, further studies related to fumonisin occurrence in stored cassava chips are necessary and should be carried out, taking into consideration not only the incidence of fumonisin producing Fusarium species, but also the impact of A. niger and other related species that could allow fumonisin occurrence, using either species or genus specific clusters of functionally related genes.Such studies will also determine the extent to which contamination of stored cassava chips by several other toxigenic Fusarium species can contribute to mycotoxin formation using a large number of samples distributed over several geographical locations.
(a) Six samples were collected from each farmer.(b) see text for description.(c) C; steeping washed and peeled cassava tubers in hot water for 2 to 3 days before drying over the fireplace for 4 to 6 days.(d) moisture content.(e) days after storage.
within the same column (at each collection period) followed by different letters are significantly different (P < 0.05) according to Duncan multi-range test.

Table 2 .
Representative specimen of Fusarium isolates obtained in the frame of this study.Only IMI isolates for which code numbers were assigned are included in this table.Confirmation of these identification results was acknowledged from isolates forwarded to PROMEC.

Table 3 .
Occurrence a and distribution of Fusarium spp.during storage in two distinct locations of southern Cameroon.: data were back transformed after analysis of variance; b: Total isolation figures based on 36 samples and 580 pieces of plated cassava chips in each location.NSC: Number of samples contaminated; TNIL: Total number of isolates associated with each Fusarium species per location; RIF: relative index of the fungus: This parameter was calculated as the ratio of the number of cassava chips pieces contaminated by each Fusarium species obtained per location over the total number of Fusarium isolates, that is, 137 isolates in Yaoundé and 131 in Ebolowa. a

Table 4 .
Associations between pairs of Fusarium spp.collected in both locations (Yaoundé and Ebolowa) based on total isolation a figures computed using Pearson correlation coefficients (r).Total isolation figures based on 36 samples and 580 pieces of plated cassava chips in each location; b=not analyzed because one of the fungus included in the association was absent; *significant correlations at p<0.05. a=

Table 5 .
Analysis of variance of the relationship between the level of Fusarium spp.recovery and pH, moisture content (MC), storage duration (Stime), chip type or location of sampling (Tchips).

Table 6 .
Mean a F. verticillioides incidence* in cassava chips' samples during storage in Nkometou and Mengomo.

Table 7 .
Average total fumonisin (Fm) content (mg/kg) ± Standard deviation (SD) in cassava chips' samples processed and stored by the farmers of the two locations.

Table 8 .
Correlation analysis for the relationship between the content in fumonisin and the parameters associated with stored cassava chips in Mengomo village.