Gene duplication, modularity and adaptation in the evolution of the aflatoxin gene cluster

Abstract

Background: The biosynthesis of aflatoxin (AF) involves over 20 enzymatic reactions in a complex polyketide pathway that converts acetate and malonate to the intermediates sterigmatocystin (ST) and O-methylsterigmatocystin (OMST), the respective penultimate and ultimate precursors of AF. Although these precursors are chemically and structurally very similar, their accumulation differs at the species level for Aspergilli. Notable examples are A. nidulans that synthesizes only ST, A. flavus that makes predominantly AF, and A. parasiticus that generally produces either AF or OMST. Whether these differences are important in the evolutionary/ecological processes of species adaptation and diversification is unknown. Equally unknown are the specific genomic mechanisms responsible for ordering and clustering of genes in the AF pathway of Aspergillus.

Results: To elucidate the mechanisms that have driven formation of these clusters, we performed systematic searches of aflatoxin cluster homologs across five Aspergillus genomes. We found a high level of gene duplication and identified seven modules consisting of highly correlated gene pairs (aflA/aflB, aflR/aflS, aflX/aflY, aflF/aflE, aflT/aflQ, aflC/aflW, and aflG/aflL). With the exception of A. nomius, contrasts of mean Ka/Ks values across all cluster genes showed significant differences in selective pressure between section Flavi and non-section Flavi species. A. nomius mean Ka/Ks values were more similar to partial clusters in A. fumigatus and A. terreus. Overall, mean Ka/Ks values were significantly higher for section Flavi than for non-section Flavi species.

Conclusion: Our results implicate several genomic mechanisms in the evolution of ST, OMST and AF cluster genes. Gene modules may arise from duplications of a single gene, whereby the function of the pre-duplication gene is retained in the copy (aflF/aflE) or the copies may partition the ancestral function (aflA/aflB). In some gene modules, the duplicated copy may simply augment/supplement a specific pathway function (aflR/aflS and aflX/aflY) or the duplicated copy may evolve a completely new function (aflT/aflQ and aflC/aflW). Gene modules that are contiguous in one species and noncontiguous in others point to possible rearrangements of cluster genes in the evolution of these species. Significantly higher mean Ka/Ks values in section Flavi compared to non-section Flavi species indicate increased positive selection acting in the evolution of genes in OMST and AF gene clusters.

Figure 1

Precursor and terminal metabolites in AF biosynthesis. Sterigmatocystin (ST), O-methylsterigmatocystin (OMST) and aflatoxins (AF) are synthesized as end products by numerous ascomycetes. There are four major aflatoxins: B₁, B₂, G₁ and G₂. Aflatoxins B₂ and G₂ are missing the double bond (indicated in red), which is present in B₁ and G₁. A. parasiticus produces B₁, B₂, G₁ and G₂; nonaflatoxigenic A. parasiticus strains commonly accumulate OMST. The gene aflU is required for the formation of G aflatoxins [10]; aflQ is required for the formation of B aflatoxins [17]; and aflP is required for the conversion of ST to OMST [17]. A. flavus, A. parasiticus, A. nomius [68], A. pseudotamarii [69] and A. bombycis [68] belong to Aspergillus section Flavi. Emericella is a teleomorphic genus for the sexual stage of Aspergillus. Monocillium is an anamorphic name associated with a Niesslia teleomorph, also in the Phylum Ascomycota. The Ascomycota comprise highly divergent fungal lineages that span 450 million years of evolutionary history [70].

Figure 2

Genome-wide tallies of aflatoxin gene duplicates, correlations among gene duplicates and inferred gene modules. A. The histogram plot shows the total number of putative aflatoxin gene cluster duplicates on y-axis across five Aspergillus genomes. The gene order in the histogram follows the order of genes in the A. flavus cluster (see cluster schematic below histogram). B. Hierarchical cluster dendrogram showing the correlations among gene duplicates in Figure 2A. Correlations are based on a dissimilarity measure of (1-r²) in which correlation values are assigned "distance" values ranging from 0.0 (completely correlated, r² = 1) to 1.0 (completely uncorrelated, r² = 0). The y-axis represents the height or distance between the gene groups divided at that point. The dendrogram shows seven putative gene modules listed from left to right as: aflX/aflY, aflJ/aflR/aflS, aflC/aflW, aflA/aflB, aflF/aflE, aflT/aflQ and aflG/aflL that are highly correlated (0.80 <r² < 1) across the five Aspergillus genomes. We consider aflR/aflS/aflJ as correlated since only aflH separates aflR/aflS from aflJ. These correlated pairs are the inferred gene modules, color coded in Figure 3.

Figure 3

Gene module reorganization in complete clusters and modularity in partial clusters. The cluster schematic shows the chromosomal location, gene order and direction of transcription of genes in ST, AF and partial gene clusters. To simplify comparisons among AF and ST clusters we adopt the new AF gene nomenclature throughout [17]. The seven inferred gene modules are color coded. The arrows in the cluster at the top indicate the location of noncontiguous recombination blocks in the A. parasiticus gene cluster [40]. The intergenic regions indicated by the black arrows share a common evolutionary history and can be concatenated into a single contiguous block such that aflB and aflL are adjacent in a hypothetical ancestor. Similarly the intergenic regions shown with grey arrows can be reunited such that aflE and aflW are adjacent. Overall this reorganization mirrors the order of these genes in the A. nidulans ST cluster and highlights the importance of gene module reorganization in the evolution of AF and ST clusters. A partial cluster duplication has been reported for some strains of A. parasiticus [71]. Syntenic partial clusters of five genes (aflC, aflS, aflR, aflX and aflY) were identified in A. fumigatus and A. terreus.

Figure 4

Ka/Ks analysis for AF cluster orthologs. Plot of mean Ka/Ks values on y-axis for cluster orthologs in A. parasiticus, A. flavus, A. oryzae, A. nomius and A. nidulans, as well as for putative partial clusters in A. parasiticus, A. fumigatus and A. terreus. Mean Ka/Ks values for each gene are based on all pairwise comparisons with each species designated separately as the reference sequence.