Conservation and Loss of a Putative Iron Utilization Gene Cluster among Genotypes of Aspergillus flavus

Iron is an essential component for growth and development. Despite relative abundance in the environment, bioavailability of iron is limited due to oxidation by atmospheric oxygen into insoluble ferric iron. Filamentous fungi have developed diverse pathways to uptake and use iron. In the current study, a putative iron utilization gene cluster (IUC) in Aspergillus flavus was identified and characterized. Gene analyses indicate A. flavus may use reductive as well as siderophore-mediated iron uptake and utilization pathways. The ferroxidation and iron permeation process, in which iron transport depends on the coupling of these two activities, mediates the reductive pathway. The IUC identified in this work includes six genes and is located in a highly polymorphic region of the genome. Diversity among A. flavus genotypes is manifested in the structure of the IUC, which ranged from complete deletion to a region disabled by multiple indels. Molecular profiling of A. flavus populations suggests lineage-specific loss of IUC. The observed variation among A. flavus genotypes in iron utilization and the lineage-specific loss of the iron utilization genes in several A. flavus clonal lineages provide insight on evolution of iron acquisition and utilization within Aspergillus section Flavi. The potential divergence in capacity to acquire iron should be taken into account when selecting A. flavus active ingredients for biocontrol in niches where climate change may alter iron availability.


Introduction
Iron is an essential element for the majority of organisms, where it serves as cofactor for enzymatic reactions and catalyst for electron transport systems. Despite relative abundance in most environments, bioavailability of iron is limited due to oxidation by atmospheric oxygen into insoluble ferric (Fe 3+ ) oxyhydroxides, and iron-dependent organisms must reduce ferric iron to its ferrous (Fe 2+ ) state prior to absorbtion. At the same time, concentrations of this metal must be tightly regulated since it can catalyze the formation of highly reactive oxygen species and cause tissue damage [1][2][3]. In most organisms, cellular iron homeostasis is designed to tightly regulate the iron supply to prevent excess accumulation. Because of the twin needs of iron uptake and iron regulation, iron-dependent organisms have evolved tightly regulated acquisition and storage strategies [4,5], and iron often serves as a signal for gene expression involved in iron uptake and storage. Many prokaryotic and eukaryotic pathogens require iron for virulence [6,7] and employ multiple pathways for iron uptake, utilization, and storage [8].
Under iron starvation, many microorganisms, including fungi, utilize low-molecular mass (<1000 Da) compounds with high iron affinity termed 'siderophores' [9,10] for iron acquisition and storage. Biosynthesis of ferric-specific siderophores is regulated by iron availability [11]. High iron concentrations repress siderophore production. Most species of ensure genotypic diversity. Wildtype isolates from silica gel storage were cultivated on 5/2 agar (5% V8 juice, 2% agar, pH 5.2). Wildtype isolates were first transferred by single spore and then maintained in 4-mL vials by suspending plugs of 5/2 agar with abundant sporulation in sterile distilled water. Aspergillic acid production was assessed on Aspergillus flavus and parasiticus agar (AFPA) medium as previously described [54]. AFPA is used to identify fungi in Aspergillus section Flavi. [54]. When grown on AFPA agar (10g Bacto agar, 20g yeast extract, 10 g Bacto peptone and 0.5 g ferric ammonium citrate in 500 mL of water), A. flavus and A. parasiticus develop characteristic orange color on the reverse of plates from the reaction of ferric iron from ferric ammonium citrate (FAC) with aspergillic acid molecules [54]. Response of A. flavus isolates to different iron concentrations was tested on chemically defined Czapek's agar medium with 4, 2, 0.5, and 0 mM concentration of FAC. The medium was autoclaved (121 • C; 20 min), and filter-sterilized FAC solution was added prior to pouring. Reverse sides of plates were photographed after 5 days for color assessment. The color intensity was measured using ImageJ software [55]. At least three pictures were taken for each strain and FAC concentration. DO114-A Maize Atoxigenic USDA-ARS, Tucson [52] EC69-E Maize Atoxigenic USDA-ARS, Tucson [52] * Ability to produce aflatoxin.

Gene Identification and Characterization
IUCs were identified while analyzing the deleted regions from A. flavus through comparisons with A. oryzae RIB40 [56]. Deletions were predicted using DELLY [57], a variant detection program that predicts deletions by mapping paired-end reads from an interrogated isolate to a reference genome. Regions polymorphic in the five A. flavus isolates were further annotated using MAKER [58]. Deleted regions were characterized in reference to corresponding regions from the annotated genome of A. oryzae RIB40. Expression of genes in IUC was measured by mapping the transcripts from A. flavus NRRL3357 [59] to genomic regions from different A. flavus genotypes. Sequence reads from the Short Read Archive (SRA; https://www.ncbi.nlm.nih.gov/sra, accession number; PRJNA144055) were mapped to A. flavus genomic region using Bowtie [60] and TopHat [61]. Fragment per kilobase pair of exon model per million fragment mapped (FPKM) values were calculated using Cufflinks [62]. To identify the extra copies of genes in A. flavus genomes, iron permease (Ftr1) and ferrooxidoreductase (Fet3) genes from A. flavus NRRL3357 were downloaded from GenBank and used in the BLAST analysis. Functional analysis of the IUC was performed by using BLAST [63] and protein domains were identified with InterProScan [64,65]. Conserved domains were identified by comparing with NCBI's Conserved Domain Database (CDD) [66], and single nucleotide polymorphisms (SNPs) were called against the reference IUC of A. oryzae RIB40 using MAQ [67]. Multiple alignments were performed with CLUSTALW [68]. Sequences of IUC of the initial five genotypes are deposited in GenBank with accession numbers: KY586947-KY586951.

Neighbor-Net Network
In order to assess the extent of diversity in IUC among A. flavus genotypes, genetic diversity among 215 A. flavus genotypes (not including the 5 original isolates) from cotton and corn crops produced in Arizona and Texas was examined using previously identified simple sequence repeats (SSR) at 17 loci [69]. Amplicons containing the SSRs were scored with GeneMarker [70] from traces produced with an ABI Capillary Sequencer [71]. Genetic distance between genotypes was calculated across the 17 loci with the START2 program [72]. The distance matrix was analyzed drawn with the Neighbor-Net algorithm in SplitsTree v4.13.1 [73]. Edges were colored according to grouping of genotypes based on structures of IUC.

Phylogenetic Analysis
Phylogenetic analysis of the partial iron permease gene (637-bp) from the 5 original A. flavus isolates, A. nomius NRRL13137, A. oryzae RIB40, and A. parasiticus SU-1 was performed with the UPGMA algorithm using MEGA7 [74]. Multiple alignments were generated with CLUSTALW. Data sets were bootstrapped with 1000 replicates to generate branch confidence values, and bootstrap values <80% were omitted. Any gaps in the sequence were treated as missing data, and no out-group was applied to build the phylogeny.

PCR Profiling of the IUC
Whole genome sequencing of A. flavus isolates was done as previously described (Adhikari et al., 2016). Briefly, genomic DNA was isolated from conidia collected from a culture grown for 7 days (31 • C, dark) on 5/2 agar (5% V-8 vegetable juice, 2% salt, and 2% agar). The FastDNA SPIN Kit and the FastPrep Instrument were used following the manufacturer's instructions (MP Biomedicals LLC, Santa Ana, CA, USA). Small DNA fragments and other contaminants were removed by applying DNA from the FastDNA SPIN Kit was applied to a SPIN filter column following manufacturer's instructions. Quantification of genomic DNA was done by both spectrophotometer (modelND-1000, NanoDrop) and the Qubit dsDNA BR assay kit (Q32850) using the Qubit 1.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer's guidelines. PCR validation was done by designing primers for each IUC gene (Supplementary Table S1) and using the PCR conditions described previously [45]. Deletions that were predicted previously by DELLY were further validated by PCR, using the primers that either bridge the putative deletion or amplify flanking regions.

Identification, Characterization, and Validation of IUC
In order to identify the structural variants in the five initial A. flavus genotypes, genome sequences were compared with A. oryzae RIB40 reference [56]. Of the several regions with identified deletions, a 15 Kb region was highly variable among A. flavus isolates. The identified region contains 6 genes. Multiple alignments of the gene region show the location in a region homologous to the centromeric end of Superscaffold124 from A. oryzae RIB40. Genome-wide alignment indicates presence of only one such group of genes in the five initial A. flavus isolates, A. oryzae, and A. nomius while A. fumigatus, A. nidulans, and A. niger had no region with significant similarity. Functional analysis of the genes with BLAST and InterProScan indicated 6 genes ( Figure 1, Table 2). Gene 1 on the centromeric-end is present in all isolates and is predicted to be a hypothetical protein. Gene 2, a LysM domain containing protein, and Gene 3, a C6 transcription factor containing a GAL4 domain, are present in only two of the 5 A. flavus (BY18-A and DO114-A). Genes 1 through 3 have the closest homologs in A. oryzae RIB40. Gene 4 is a ferric reductase gene with a ferric reductase-like transmembrane component also found in A. oryzae RIB40. Gene 5 encodes a multicopper oxidase with ferroxidation activity, which is the closest homolog (e-value = 4 × 10 −169 ) of fet3 gene from A. oryzae RIB40. A comparison of the predicted amino acid sequences from A. flavus showed >95% similarity with fet3 Microorganisms 2021, 9, 137 5 of 18 from A. oryzae RIB40. The last gene (Gene 6) is an iron permease, which is the closest homolog to ftrA gene from A. oryzae RIB40 and has 95% similarity in predicted amino acid sequences to FtrA ( Table 2). Search of iron permease (ftr1) and ferrooxidoreductase (fet3) genes in A. flavus genomes showed presence of multiple ftr1 and fet3 genes in few A. flavus isolates outside the IUC. Few of the A. flavus isolates with complete IUC have ftr1 and fet3 homologs outside the IUC. A. flavus isolates with either partial or complete deletion of IUC, however, either lack the extra copies of genes or have partially deleted genes outside the IUC. Interestingly, none of the fet3 genes present outside the IUC were clustered with ftr1 genes as seen in IUC. Expression analysis of the genes within IUC showed that all six genes are expressed. FPKM values varied between genes, with the second set of 3 genes having higher expression values than the first set of 3 genes (Supplementary Table S2). To validate the genes, all six were PCR amplified from genomic DNA extracted from the 5 initial A. flavus isolates. BY18-A and DO114-A have all IUC genes intact, while CIA011 and EC69-E are missing the middle two genes (gene 3 and 4) and have lost portions of the last two genes through deletion.  Large deletion Small deletion Open arrows indicate primer-binding sites and direction of amplification for primers used to assess the distribution of genes within A. flavus communities and letters above the arrows indicate primer names.

Polymorphisms within IUC among A. flavus Isolates
Comparative analysis of the IUC from A. flavus isolates with reference A. oryzae RIB40 shows high levels of polymorphism ( Figure 1). CIA011 and EC69-E have the smallest IUC with >12.5 kb deleted and only portions of genes 1 and 6 maintained (IUC type A). Both isolates have the same deletion with identical flanking sequences. AF13, which has two deletions (9 kb and 2 kb), has completely lost three genes (LysM domain containing protein, C6 transcription factor, and ferric reductase) and partially lost the multicopper oxidase (IUC type B). The iron permease gene in AF13 remains intact. The remaining two isolates (BY18-A and DO114-A) have complete IUCs with all 6 genes intact (IUC type C). Mapping of A. flavus IUCs to the corresponding A. oryzae IUC shows high levels of polymorphism in the gene that codes for the hypothetical protein. Isolates in IUC type A have the greatest density of SNPs (65), followed by isolates in IUC type B (52) (Figure 1). Iron permease genes from BY18-A, CIA011, DO114-A, and EC69-E are almost identical to the genes in A. oryzae RIB40, with only 1-2 SNPs. The A. flavus AF13 iron permease gene is highly divergent, with 31 SNPs (Figure 2).

Production of Siderophores
Members of Aspergillus section Flavi produced a distinct bright orange color on the reverse of colonies grown on AFPA medium. The orange color results from reaction of ferric citrate with aspergillic acid, forming a colored complex [75]. All five of the initial A. flavus isolates used produced the characteristic orange color (Figure 3). Aspergillic acid, an N-hydroxylated pyrazine, is a naturally occurring hydroxamic siderophore [76]. A total of fifteen isolates (including the original five isolates) were tested, and similar results were obtained with all fifteen isolates grown on AFPA medium. Production of the characteristic orange color is used for taxonomic assessments [54] and indicates the ability to produce siderophores.

Production of Siderophores
Members of Aspergillus section Flavi produced a distinct bright orange color on the reverse of colonies grown on AFPA medium. The orange color results from reaction of ferric citrate with aspergillic acid, forming a colored complex [75]. All five of the initial A. flavus isolates used produced the characteristic orange color (Figure 3). Aspergillic acid, an N-hydroxylated pyrazine, is a naturally occurring hydroxamic siderophore [76]. A total of fifteen isolates (including the original five isolates) were tested, and similar results were obtained with all fifteen isolates grown on AFPA medium. Production of the characteristic orange color is used for taxonomic assessments [54] and indicates the ability to produce siderophores.

Response to Reduced Iron Media
The response of A. flavus isolates to reduced iron media was variable and dependent upon IUC composition. Isolates categorized as IUC type C (e.g., C6-E) did not produce siderophores, as indicated by lack of orange color (aspergillic acid) on the reverse of the plate (Figure 3). On the other hand, A. flavus isolates with IUC types A and B produced more siderophores than IUC type C in media with higher concentration of FAC. IUC type C isolates grown on 0 mM FAC produced no siderophore, while those from IUC types A and B produced large quantities of siderophore, as shown by a bright orange color on the reverse of the plate (Figure

Response to Reduced Iron Media
The response of A. flavus isolates to reduced iron media was variable and dependent upon IUC composition. Isolates categorized as IUC type C (e.g., C6-E) did not produce siderophores, as indicated by lack of orange color (aspergillic acid) on the reverse of the plate (Figure 3). On the other hand, A. flavus isolates with IUC types A and B produced more siderophores than IUC type C in media with higher concentration of FAC. IUC type C isolates grown on 0 mM FAC produced no siderophore, while those from IUC types A and B produced large quantities of siderophore, as shown by a bright orange color on the reverse of the plate (Figure 3). Increasing the concentration of FAC decreased siderophore production, and at 4 mM FAC concentration, siderophore production was greatly reduced. Measurement of red pixel count on the images of the reverse side of the plates showed higher number at both 0.5 mM and 2 mM concentrations as compared to 4 mM concentration for both isolates with partial or deleted IUC (Supplementary Table S3). The red pixel count is higher for isolate with partial or deleted IUC as compared to complete IUC.

Presence of Reductive Iron Assimilation (RIA) Pathway
Fungi use secreted iron chelators (siderophores) and reductive iron assimilation (RIA) mechanisms to acquire iron in a high affinity manner. RIA processes originally characterized in Saccharomyces cerevisiae [77] involved three different genes. Ferric iron is reduced by ferric reductase and then oxidized by the iron multicopper oxidase while being transported across the plasma membrane by the high-affinity iron permease. All three genes (ferric reductase, multicopper oxidase, and iron permease) are located on one scaffold of the A. flavus genome assembly and separated by intergenic spaces of 1300-bp and 618-bp respectively ( Figure 1). In isolates with IUC type A, both ferric reductase and fet3 genes are missing and a partial ftr1 gene is present (Table 3). ftr1 gene remnants in all isolates with IUC type A have identical sequence ends. IUC type B isolates lack ferric reductase but have a complete ftr1 gene with partial fet3. IUC type B isolates have remnants of various deletion events and are separated by the highest number of SNPs when compared with A. oryzae RIB40 or with each other (Figure 2).

Lineage-Specific Loss of IUC
Relationships among 215 genotypes of A. flavus were analyzed using Neighbor-Net network based on genetic distance estimated from SSR data from 17 loci distributed across 8 chromosomes of A. flavus [69]. Initial examination revealed that 79% of the 215 genotypes have IUC type C, 10% have IUC type B, and 11% have IUC type A (Table 3). In general, the Neighbor-Net network reflects IUC structure ( Figure 4). One well-resolved clade contains only genotypes with IUC type A with 81% of type A genotypes. Ten clades with 6 to 16 members contain IUC type C. Genotypes with IUC type B didn't group into large clades but were co-distributed with IUC type C genotypes in many smaller clades. For IUC types B and C, deletion locations and sizes often vary, as do sequences of flanking regions. When genotypes are divided into groups based on the sequence flanking the IUC deletion, IUC type B segregates into three distinct groups while IUC type A and C each belong to a single group (Table 3). clade contains only genotypes with IUC type A with 81% of type A genotypes. Ten clades with 6 to 16 members contain IUC type C. Genotypes with IUC type B didn't group into large clades but were co-distributed with IUC type C genotypes in many smaller clades. For IUC types B and C, deletion locations and sizes often vary, as do sequences of flanking regions. When genotypes are divided into groups based on the sequence flanking the IUC deletion, IUC type B segregates into three distinct groups while IUC type A and C each belong to a single group (Table 3). Network was generated by the split decomposition algorithm with allelic data from 17 SSR loci using SplitsTree4. Filled, half-filled, and empty circles represent genotypes with cluster type C, cluster type B, and cluster type A (Figure 1), respectively.

Evolutionary Relationship of Iron Permease Gene
Aspergillus flavus isolates with IUC type C have iron permease genes with 1345-bp while isolates with deletions have partial ftr1 with 637-bp. Phylogenetic analysis of these 637-bp from A. flavus isolates, A. nomius NRRL13137, A. oryzae RIB40, and A. parasiticus SU-1 resulted in a UPGMA tree with 5 well-supported clades ( Figure 5). A. parasiticus SU-1 and A. nomius formed a clade separate from A. flavus. A. flavus isolates with IUC type B (AF13) and IUC type A (CIA011 and EC69-E) were resolved into a clade distinct from A. oryzae and A. flavus isolates with IUC type C. Phylogenetic relationships based on the 637bp of iron permease gene recapitulate the grouping of the fungi based on the structure and presence/absence of genes in the IUC (Figure 2). A summary of the fungal isolates and the IUC type they belong to is provided in Supplementary Table S4. Network was generated by the split decomposition algorithm with allelic data from 17 SSR loci using SplitsTree4. Filled, half-filled, and empty circles represent genotypes with cluster type C, cluster type B, and cluster type A (Figure 1), respectively.

Evolutionary Relationship of Iron Permease Gene
Aspergillus flavus isolates with IUC type C have iron permease genes with 1345-bp while isolates with deletions have partial ftr1 with 637-bp. Phylogenetic analysis of these 637-bp from A. flavus isolates, A. nomius NRRL13137, A. oryzae RIB40, and A. parasiticus SU-1 resulted in a UPGMA tree with 5 well-supported clades ( Figure 5). A. parasiticus SU-1 and A. nomius formed a clade separate from A. flavus. A. flavus isolates with IUC type B (AF13) and IUC type A (CIA011 and EC69-E) were resolved into a clade distinct from A. oryzae and A. flavus isolates with IUC type C. Phylogenetic relationships based on the 637-bp of iron permease gene recapitulate the grouping of the fungi based on the structure and presence/absence of genes in the IUC (Figure 2). A summary of the fungal isolates and the IUC type they belong to is provided in Supplementary Table S4.

Presence of IUC in Other Fungi
We searched genome sequence databases both within and outside Aspergillus to find homologues of the IUC. BLAST searches revealed that the IUC of A. flavus has very high similarity to A. oryzae RIB40 IUC (evalue = 0.00) ( Table 2) followed by A. flavus NRRL3357. Recently sequenced A. flavus NRRL335 [78] and two other members of section Flavi, A. nomius, and A. parasiticus, both contain similar gene groups. When searched against larger datasets, except for A. nidulans, most Aspergillus genomes have at least one ftrA-fetC (Iron permease-ferroxidase) gene pair similar to the one found in A. flavus IUC [34]. The cooccurrence of the iron permease/multicopper oxidase genes has also been reported in many fungal genomes outside Aspergillus. For example, genes similar to fet3 and ftr1 have counterparts in many Ascomycete (Neofusicoccum parvum, evalue=4× 10 −116 ; Fusarium oxysporum, evalue = 5 × 10 −107 ) and Basidiomycete (Ustilago maydis, evalue = 1× 10 −125 ) fungi [8,79].

Discussion
Iron uptake and homeostasis functions are important for pathogenesis, and iron levels influence growth and development of A. flavus. There is considerable genetic diversity among isolates of A. flavus, and this variation is particularly frequent in genes encoding products used to produce secondary metabolites. In the current study, we applied whole genome comparative analysis to assess diversity among A. flavus isolates. In doing so, we identified group of genes potentially functional in iron uptake and utilization and forming a putative gene cluster (IUC) spanning 15 Kb. The putative IUC identified in this work contains 6 genes, including putative transcription factors. Among the various A. flavus isolates, multiple indels that disable genes in the IUC were detected in multiple genotypes, suggesting these mutations were fixed in the resulting clonal lineage. The current work indicates that A. flavus may have two different mechanisms for iron utilization, a siderophore-mediated iron utilization system and a reductive iron assimilation (RIA) pathway. Sequence similarity to genes responsible for high-affinity iron uptake in other fungi suggests a fully functional IUC in most A. flavus genotypes. Isolates missing genes in the IUC responded differently to iron starvation conditions by possibly employing a different pathway for iron uptake and utilization. These isolates produced siderophores when grown under low-iron conditions, while those with a complete IUC did not produce siderophores (Figure 3). These results, together with the A. flavus population profiling data, indicate lineage-specific loss of iron utilization genes and suggest the possibility that different genotypes of A. flavus have distinct iron uptake and utilization systems.
The IUC identified in the current study resides on a region homologous to Chromosome 5 of A. oryzae RIB40 and primarily contains genes with predicted functions in iron uptake and utilization (Figure 1). Functional prediction of IUC genes indicates possible involvement in the uptake and utilization of iron ( Table 2). The first two genes are hypothetical proteins with either a GAL4 domain or a fungal-specific LysM domain. LysM domain containing proteins act as effectors that play important roles in promoting virulence [80]. The third gene in the IUC contains a domain of unknown function with homology to fungal C6 transcription factors. Some members of the fungal-specific C6 transcription factor family are strongly associated with secondary metabolite gene clusters in A. flavus and A. parasiticus [81]. For example, aflR, one of the C6 transcription factors in A. flavus, is the key positive regulator of expression of most aflatoxin genes [82,83]. Based on the position of the gene in the IUC, we anticipate that this C6 transcription factor plays a role in the regulation of genes involved in iron uptake and utilization. The IUC ferric reductase gene has a ferric reductase like transmembrane domain. Ferric reductase facilitates the use of siderophore-bound iron. Ferric reductases in fungi play important roles in iron acquisition and virulence [84]. Another gene of particular interest is fet3, encoding the multicopper oxidase (MCO). Based on the domain structure of the predicted Fet3 protein, we anticipate this enzyme is involved in the ferroxidation process ( Figure 1). MCO plays an important role in high affinity iron uptake through Fe 2+ to Fe 3+ [85,86] conversion. The gene next to MCO in the IUC is a putative iron permease, involved in iron transport in iron-depleted environments. In many pathogenic fungi [8,87] products of similar iron permease genes are virulence factors. Clustering of genes encoding MCO and an iron permease is not unique to A. flavus. Such clustering is also found in many other fungi [27,88]. Although the link between ferroxidation and iron transport by the iron permeation depends on the coupling of these two enzymes (iron permease and multicopper oxidase), the process is not well understood [8]. The current results suggest reductive mobilization of iron is common in A. flavus.
The complete IUC was conserved in most (79%) A. flavus isolates analyzed in this study. Considering the wide distribution of complete IUC among A. flavus genotypes, RIA can be considered the index (basal) pathway for high affinity iron uptake [8]. RIA is an extracellular membrane-bound process whereby ferric iron is reduced and subsequently oxidized while being transported across the plasma membrane by the high-affinity iron permease Ftr1 [8,89]. As observed in A. flavus (Figure 1), Ftr1 and Fet3 are closely linked and co-dependent on the plasma membrane [25,90]. In S. cerevisiae vacuoles store iron, and homologs of Fet3/Ftr1 are reported to be involved in vacuolar iron storage in addition to their involvement in iron uptake [91][92][93]. Membrane-bound reductive iron assimilatory systems have been reported for a broad array of fungi [8,[27][28][29][30][31][32]94] but the existence of this system in A. flavus was previously unknown. Clustering of the genes in biosynthetic pathways is the hallmark of bacteria and filamentous fungi [95]. In that sense the potential clustering of the IUC genes of A. flavus is not surprising, but the magnitude of differences among closely related A. flavus genotypes reveals unexpected diversity within species of A. flavus in terms of iron uptake and utilization.
Discovery of new metabolic pathways and associated gene clusters through genome mining is greatly accelerated by the availability of sequenced genomes from multiple species of Aspergillus [96][97][98]. The genes identified within IUC and involved in iron ferroxidation and permeation have homologs in other fungi [79]. Homologs of fet3 and ftr1 genes were identified from A. flavus genomes but in our case none of those genes seem to be clustered outside the IUC. This suggests that fungal genera share common iron uptake utilization pathways, including genes involved in ferroxidation and permeation, siderophore synthesis, and transport. However, differences may exist in mechanisms through which iron utilization could be modulated by environmental factors. Phylogenetic analysis based on the partial iron permease gene ( Figure 5) reflects evolution of the IUC in Aspergillus section Flavi and strongly supports the relationship observed by multiple alignment of the IUC. The diversity of the IUC among A. flavus isolates suggest that A. flavus has iron utilization pathways similar to those observed in many other fungi and that these pathways could play important roles in adaptation to the diverse ecological niches this species is known to occupy.
To assess the extent of diversity in IUC among A. flavus genotypes, we characterized the patterns of genetic variation among genotypes from Arizona and Texas. A total of 215 genotypes representing A. flavus associated with either cotton or corn were profiled. The structure of the Neighbor-Net network largely reflects IUC structure (Figure 4), with clades including only genotypes lacking copies of all the IUC genes or only genotypes with all the IUC genes intact. Complete IUC deletions have common boarder sequences (Table 3), suggesting an ancient deletion event that has been retained through clonal evolution over tens of thousands of years. In contrast to the homogeneity in this clade, those containing partial IUCs have more diversity (Figure 4, Table 3). Sequencing of flanking regions from genotypes with partial IUC revealed a diversity of flanking sequences reflecting multiple deletion events (Table 3). In some cases, the same deletion occurred in multiple clades. This may have resulted from incomplete lineage sorting or gene flow. Repeated loss of IUC function and success of lineages where the IUC has been lost may suggest an adaptive advantage to this loss. Although specific ecological or environmental factors contributing to genetic differences among A. flavus genotypes are not known, soil types differing in iron content could have contributed to this differential adaptation. This pattern of multiple deletion events has been demonstrated for other biosynthetic gene clusters [52,[99][100][101] in Aspergillus section Flavi.
Under iron-limited conditions, most fungi synthesize and secrete siderophores, which are small organic compounds that bind ferric iron with high affinity and specificity [10]. Virtually all fungi express a nonreductive uptake system that is specific for siderophore iron chelates [102]. Our analysis revealed that A. flavus isolates with IUC type C are equipped with all genes required for an RIA system, while isolates with IUC type A and B rely on siderophore-mediated iron uptake system. To show the function of the IUC type identified in this study on iron utilization in A. flavus, isolates with different deletions in IUC were exposed to varying iron concentrations (Figure 3). When grown under iron-limited conditions, A. flavus isolates without IUC produced large quantities of siderophores indicating siderophore-mediated iron uptake and utilization. In contrast, isolates with complete IUC did not produce siderophores even under iron-limited conditions, suggesting their lack of reliance on siderophore-mediated iron uptake system. Nevertheless, these isolates have retained capacity to produce siderophores, as demonstrated by production on AFPA. It has been shown that deprivation of iron results in increased synthesis of siderophores in E. coli [103]; similar production of siderophores under iron-limited conditions in A. flavus suggests an analogous function. Although the genetic mechanisms that shape iron utilization pathways are not fully understood, one explanation for the observed diversity is that A. flavus, as an animal and plant pathogen, evolved in an environment that was populated by many other microorganisms. Diversity of uptake systems could have enabled A. flavus to effectively compete with other microorganisms for the limited amounts of available iron in the environment while facilitating facultative pathogenicity to plants and animals. Global climate change will influence not only the quantity of crops available but also compositions of edible crop components including reduction in iron content [104]. Competition for iron may be increasingly important to the ability of atoxigenic biocontrol strains of A. flavus to outcompete aflatoxin producers. The variability in iron acquisition processes described in the current work should be considered when selecting atoxigenic strains for aflatoxin management in altered climates.
Analysis of diversity and evolution within Aspergillus section Flavi suggests the possibility that A. flavus may have multiple pathways for iron acquisition and that the IUC has been modified in A. flavus lineages through multiple deletion events. Certain IUC deletions have been retained during lineage divergence. A component of the A. flavus RIA, the ferroxidation/permeation iron uptake system, shows similarities with other fungi where it is important for virulence [105][106][107]. One of the intriguing questions is why there is variation among A. flavus genotypes in iron acquisition genes despite evolutionary closeness. It is possible that the adaptation of A. flavus to different conditions may ultimately determine the manner in which iron is acquired from the environment. However, empirical data is required to test this hypothesis. Furthermore, A. flavus is the second leading cause of invasive aspergillosis and the genes identified in this study could be useful as targets for disease monitoring and management. Improved understanding of iron uptake and utiliza-tion might facilitate better understanding and management of aflatoxin contamination of food and feed.