Molecular evolution of glutathione S-transferases in the genus Drosophila

Abstract

As classical phase II detoxification enzymes, glutathione S-transferases (GSTs) have been implicated in insecticide resistance and may have evolved in response to toxins in the niche-defining feeding substrates of Drosophila species. We have annotated the GST genes of the 12 Drosophila species with recently sequenced genomes and analyzed their molecular evolution. Gene copy number variation is attributable mainly to unequal crossing-over events in the large delta and epsilon clusters. Within these gene clusters there are also GST genes with slowly diverging orthologs. This implies that they have their own unique functions or have spatial/temporal expression patterns that impose significant selective constraints. Searches for positively selected sites within the GSTs identified G171K in GSTD1, a protein that has previously been shown to be capable of metabolizing the insecticide DDT. We find that the same radical substitution (G171K) in the substrate-binding domain has occurred at least three times in the Drosophila radiation. Homology-modeling places site 171 distant from the active site but adjacent to an alternative DDT-binding site. We propose that the parallel evolution observed at this site is an adaptive response to an environmental toxin and that sequencing of historical alleles suggests that this toxin was not a synthetic insecticide.

Figure 1.—

The total number of GST genes and pseudogenes in the 12 Drosophila species. For each species, the number of GSTs genes (left) and pseudogenes (right) are shown with the different classes represented by different shading (see key). The tally of pseudogenes includes only those identified with our tblastn filtering criteria and does not account for gene losses inferred from the phylogenetic reconstruction in Figure 3.

Figure 2.—

The δ and ε GST clusters of the 12 Drosophila species are shown in A and B, respectively. Full-length GSTs are shown as arrowed boxes. The direction of the arrows represents the direction of transcription. The boxes are colored to indicate orthologs. Paralogous clades supported by bootstrap score ≥75 that lack orthologs are also colored (e.g., GstE15, GstE15a, GstE15b). The white boxes represent genes that do not occur in clades with a bootstrap score ≥75. The “ψ” and circled P signs below each box indicate GST pseudogenes and partial sequences, respectively. The numbers above the vertical bars represent the number of bases excluded to keep the rest of the figure to scale. The “| |” symbol represents sequence gaps in contigs.

Figure 3.—

GST gene gain and loss is represented on the species tree. Gene duplications that have 75% or greater bootstrap support are shown in boldface type above the branches of the species tree. Inferred gene losses are denoted with “ψ” or, if in a terminal branch of the species tree, are listed next to the species name. To account for all genes, duplications not supported by >75% bootstrap support are indicated in italics at the youngest possible branch where those duplications could have occurred. The number of δ and ε genes that are not found in clades with >75% supports are listed under “<75% bootstrap.” “Dψ” refers to the genes that have duplicated and become pseudogenes within single lineage. The prefix “GST” has been removed from GST δ and ε genes for clarity.

Figure 4.—

Outputs of model 3 of PAML are represented graphically, allowing comparison of selective constraint among GST genes. Model 3 uses a maximum-likelihood method to assign three ω-values to each gene set where ω₀, ω₁, and ω₂ have the proportions p₀, p₁, and p₂, respectively (we represent the proportions as percentages along the x-axis). While each gene set has only three ω-values, we have divided the ω-values observed across the whole 22 gene sets into four different ranges: highly constrained sites (0 ≤ ω ≤ 0.1), moderately constrained sites (0.1 ≤ ω ≤ 0.5), relaxed sites (0.5 ≤ ω ≤ 1.0), and positively selected sites (ω > 1). Genes with many sites evolving slowly have a high proportion of sites in the 0 ≤ ω ≤ 0.1 range and are mostly solid in the figure. These values were calculated for genes inferred to be present in the ancestral Drosophila (see Figure 3) and were calculated with all available orthologs. The CG10065 here refers only to the GST domain of this chimeric gene.

Figure 5.—

The state of the amino acid residue 171 of GSTD1 in the 12 Drosophila species. This site was estimated to be positively selected using empirical Bayes with probability >95% under model M8. The species tree is adapted from FlyBase.

Figure 6.—

Molecular docking of DDT and glutathione to the homology modeled D. melanogaster GSTD1 monomer showing the best-docked conformations when site of docking centered on (A) the GSH-binding site and (B) the positively selected residue K171.