A Tool for Reliable Detection of Aflatoxin Biosynthetic Gene Clusters in Aflatoxigenic and Atoxigenic Aspergillus flavus Isolates

Molecular techniques and phenotypic characterisation have been used to differentiate aflatoxigenic and atoxigenic Aspergillus flavus strains. However, there is a lack of a consistent and reliable tool for discrimination between these strains of A. flavus. Here we report, an optimised real-time qPCR-based tool for reliable differentiation between aflatoxigenic and atoxigenic strains of A. flavus. Accordingly, expression profiles and deletion patterns of genes responsible for aflatoxin production in five representative aflatoxigenic and atoxigenic A. flavus strains (KSM012, KSM014, HB021, HB026 and HB027) were examined using the optimised real-time qPCR tool. We observed that under induced conditions, aflP, aflS, aflR and aflO transcripts were the most upregulated genes across the tested isolates while aflS and aflO were always expressed in both induced and uninduced isolates. However, aflR and aflP did not give clear distinctions between non-toxin and toxin producing isolates. The deletion patterns were prominent for aflD and aflR whereas alfO, aflS and aflP had no deletions among the isolates. Significant variation in transcript abundance for aflD, aflR and aflS were observed for aflatoxigenic isolate KSM014 under induced and uninduced states. False detection of aflD gene transcript in atoxigenic strain KSM012 was evident in both induced and uninduced conditions. With the exception of KSM012, aflP gene did not exhibit significant variation in expression in the isolates between induced and uninduced conditions. One-way ANOVA and Post-test analysis for linear trends revealed that aflatoxin biosynthetic cluster genes show significant (P < 0.05) differences between atoxigenic and aflatoxigenic isolates. Our optimized qPCR-based tool reliably discriminated between aflatoxigenic and atoxigenic A. flavus isolates and could complement existing detection methods.

1 Introduction 2 Certain fungi of the Aspergillus genus produce secondary metabolites termed aflatoxins, which are a 3 class of naturally-occurring mycotoxins (1). About 200 species of Aspergillus have been identified, of which 16 4 have been found to produce aflatoxins that are detrimental to both human and animals (1,2). A number of 5 Aspergillus species such as Aspergillus flavus, Aspergillus bombycis, Aspergillus nomius and Aspergillus niger 6 produce aflatoxins with high carcinogenic activity (3,4). Aflatoxin contamination has been detected in maize, 7 beans, cottonseed, peanuts and other grain crops (5,6). The contamination not only results in reduced crop value 8 but can cause health problems in both humans and animals that consume contaminated crops and feeds (7).

9
Although aflatoxin-producing Aspergillus species are found worldwide, they are of greater concern in 10 underdeveloped countries which lack appropriate infrastructure, management tools and resources required to 11 prevent, control or monitor their impact on the wider community (1). Environmental conditions: unseasonal rains 12 during harvesting, increased temperatures and moisture promotes fungal pathogen proliferation and mycotoxin 13 production (8,9). Additionally, increased risks of mycotoxin production and fungal growth is facilitated by 14 improper harvesting, poor storage facilities and sub-optimal temperatures during processing and marketing. These 15 environmental conditions and problems associated with food production and storage are common in most parts of 16 sub-Saharan Africa, where to date, the largest poisoning of mycotoxin epidemic has been reported (10,11).

17
The classification and identification of Aspergillus spp. based on phenotype is complemented by 18 molecular and chemotaxonomic characterisation (12,13). Phenotypic characterisation of A. flavus isolates may 19 identify an isolate as potentially aflatoxigenic, but this is neither definitive nor precise (14,15). Molecular 20 techniques, such as Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) have also been used to 21 differentiate aflatoxigenic from atoxigenic A. flavus strains, using the expression of regulatory and structural 22 aflatoxin pathway genes as markers for aflatoxin production (16,17). Despite the complexity of the aflatoxin 23 pathway involving at least 25 structural and two regulatory genes (18), some studies have found good agreement 24 between gene expression and aflatoxin production (17). Scherm et al. (17), reported the expression profiles of the 25 genes aflO, aflP and aflD were linked with the A. flavus strains capability for aflatoxins production.

26
In A. flavus, expression of nor-1 (aflD), the gene that encodes for an enzyme meant to catalyse the 27 conversion of the first stable biosynthesis of aflatoxins, (Fig. S1) norsolorinic acid, to averantin (19,20) is the 28 main structural gene in biosynthetic pathway of aflatoxins. Previous studies established the transcription of aflD 29 as a better marker for discrimination between atoxigenic and aflatoxigenic strains (21,22). In contrast, aflR 30 transcription, which expresses a regulator which is capable of activating majority of structural genes in the 31 aflatoxin biosynthetic pathway had poor discriminatory power (23).

32
The aflS gene codes for regulatory protein, laeA (24) whose function is not confirmed though, (25) 33 postulated that the interaction between aflR with laeA may support binding of DNA by the former. Du et al. and

34
Schmidt-heydt et al. (26,27), also observed that aflS may have the ability in modulation of aflR activity and that 35 environmental factors may influence the ratio of aflR to aflS. According to (28), in the case of some Aspergillus 36 spp., this modulation may have shifted the affinity of aflR towards the G-group-specific genes, ordA, cypA and

15
The designed primers were suitable for the current study and our findings (Tables 1; 2; 3  Scherm et al. (17), also found that a reduction in aflatoxin level was accompanied by a decrease in aflD transcripts.

23
Furthermore they found that the expression profiles of aflD (nor-1), aflO (omtB) and aflP (omtA) were consistently 24 correlated with the production of aflatoxins, whereas the expression of aflS (aflJ) and aflR were not.

25
Data analysis from Table 2 showed that aflP, aflS and aflO gene expression is not essential for  of aflD as a marker to distinguish aflatoxin producing and non-producing isolates from peanut, while aflR 31 expression could not be used to differentiate these phenotypes.

32
In the current study, an aflatoxin producing strain KSM014, consistently had higher transcript levels than 33 the other isolates across all the studied genes ( Figs.3; 4). For aflR and aflD, there was no significant increase in 34 transcript abundance in isolate KSM014 (uninduced state). The aflO and aflS genes were expressed at higher 1 significantly lower under induced conditions than uninduced conditions (Fig.4e). aflO expression decreased in 2 uninduced isolates of KSM012, HB021, HB026 and HB027 but increased for isolate KSM014 (Fig.4e).

3
Expression of the aflP gene did not vary significantly between induced and uninduced isolates KSM014, 4 HB021, HB026 and isolate HB027 according to 1-Way ANOVA and TMCT test (Fig.4d). aflP expression was 5 significantly higher for isolate KSM012 in both induced and un-induced states. Expression in the uninduced 6 isolate, HB026 was higher than in the induced isolate, but the difference was not significant based on a TMCT 7 test. This observation suggests that aflP is not a promising marker that can be used in discrimination of 8 aflatoxigenic and atoxigenic isolates (Fig.4d). aflD, aflR and aflS all differed significantly in expression in 9 KSM014 between induced and uninduced states (Figs.3; 4).

10
The aflD gene transcript was detected in both induced and uninduced isolate KSM012, which is 11 atoxigenic (Fig.4c; Table 2). False positive and negative transcription signals have been previously observed (34).

26
native almonds from Portuguese. Their conclusion was that expression of aflD was not a promising marker that 27 could be used in differentiation of non-aflatoxigenic and aflatoxigenic isolates. There was only one almond isolate 28 in their study that gave false positive transcript. We have also observed in the current study that aflD and aflR 29 transcripts were not consistent in giving a clear distinction between aflatoxigenic and atoxigenic strains, which is 30 in line with a related finding by (37). This could be due to either point mutations or large deletions in the aflatoxin 31 genome cluster involved in aflatoxin production ( Fig.3a; 4c).

32
Based on our results, aflO might be considered as a marker for distinguishing atoxigenic and 33 aflatoxigenic strains ( Fig.4e; Table 2). We used a more sensitive real-time RT-qPCR approach compared to other

4
Studies differentiating aflatoxigenic and atoxigenic strains have often relied on monitoring aflatoxin that lack of a tool/protocol for reliable discrimination between toxigenic and atoxigenic strains of A. flavus isolates 7 have not yet been successfully established. Furthermore, one should be aware that some genes are not exclusive 8 to the aflatoxin biosynthetic pathway, which could create false-positives from sterigmatocystin producing fungi 9 (40).
In conclusion, many enzymatic steps are involved in the aflatoxin biosynthesis pathway. Expression level 11 measurements of the genes that codes for the enzymes or the absence/presence of these genes could provide and their aflatoxigenic potential and makes the task challenging.
We found that certain genes in the aflatoxin biosynthetic pathway are expressed more highly than others       (Table 3) for one reference gene (β-tubulin) and five genes of interests (structural and regulatory) were 35 designed and assessed as previously described (30,41). The PCR and melt curve analysis were used to identify 36 both specific and non-specific amplification.
37 Table 31 List of primers and the corresponding target genes used in the study.  14 Expression stability analysis of aflatoxin biosynthetic genes

15
The qPCR reaction mixes and conditions were set up as described previously (30). To minimise      (Fig. 2.3a). In A. flavus KSM014, an aflatoxin producing strain, aflR transcript abundance increased under induced and decreased under uninduced conditions (Fig.3a). Thus, aflR could be a marker to differentiate toxin and non-toxin producers. When comparing the induced isolates, isolate HB026, had higher levels of expression of aflS and aflO. KSM012 and KSM014 had higher aflD and aflR transcript levels in induced than in uninduced cultures (Figs.3; 4; Table 2). AflR expression decreased significantly in uninduced isolate KSM012 (Fig.3a).  (44)(45)(46). Asterisks (red star) represents the specific genes studied. Table 1. Different gene expression profiles or deletion patterns exhibited by Aspergillus flavus strains.* Table 2. Clustered aflatoxin biosynthesis pathway genes showing enzymes involved, functions, statistical linear regression and efficiency. Table 3. List of primers and the corresponding target genes used in the study. Table S1 Integrity and quality of RNA assessed on Nano Drop spectrophotometer used for downstream analysis.