The Anopheles gambiae glutathione transferase supergene family: annotation, phylogeny and expression profiles

Abstract

Background: Twenty-eight genes putatively encoding cytosolic glutathione transferases have been identified in the Anopheles gambiae genome. We manually annotated these genes and then confirmed the annotation by sequencing of A. gambiae cDNAs. Phylogenetic analysis with the 37 putative GST genes from Drosophila and representative GSTs from other taxa was undertaken to develop a nomenclature for insect GSTs. The epsilon class of insect GSTs has previously been implicated in conferring insecticide resistance in several insect species. We compared the expression level of all members of this GST class in two strains of A. gambiae to determine whether epsilon GST expression is correlated with insecticide resistance status.

Results: Two A. gambiae GSTs are alternatively spliced resulting in a maximum number of 32 transcripts encoding cytosolic GSTs. We detected cDNAs for 31 of these in adult mosquitoes. There are at least six different classes of GSTs in insects but 20 of the A. gambiae GSTs belong to the two insect specific classes, delta and epsilon. Members of these two GST classes are clustered on chromosome arms 2L and 3R respectively. Two members of the GST supergene family are intronless. Amongst the remainder, there are 13 unique introns positions but within the epsilon and delta class, there is considerable conservation of intron positions. Five of the eight epsilon GSTs are overexpressed in a DDT resistant strain of A. gambiae.

Conclusions: The GST supergene family in A. gambiae is extensive and regulation of transcription of these genes is complex. Expression profiling of the epsilon class supports earlier predictions that this class is important in conferring insecticide resistance.

Figure 1

Alignment of A. gambiae GSTs. ClustalW alignment of all putative subunits of A. gambiae cytosolic GSTs. Gaps introduced to maximise sequence similarity are shown by horizontal dashes. The regions of the alignment that were removed prior to phylogenetic analysis are shown in light grey. A vertical line demarks the putative end of the N-terminal domain. Residues to the left of this boundary were removed from the alignment in construction of the phylogenetic tree shown in Figure 2 (see text for further details).

Figure 2

Neighbour-joining tree and introns positions of A. gambiae GSTs. The putative amino acid sequences of all 32 GST subunits in A. gambiae were aligned with GST-1 from C. elegans (Accession number CAA78471) using ClustalW (see text for details). The tree was constructed by the neighbour-joining method from a similarity matrix of pairwise comparisons made using the Jukes-Cantor algorithm. Nodes with distance bootstrap values (500 replicates) of > 70% are marked by *. The positions and phase of introns are shown on the right of the dendogram. Phase 0 introns are shown by a solid black line, phase 1 by dotted lines and phase 2 by grey solid lines.

Figure 3

Neighbour-joining tree illustrating the relationship between representative insect and mammalian GST classes. GSTs from the non-insect specific classes from A. gambiae, D. melanogaster were aligned with representatives from mammalian classes using ClustalW. The tree was constructed by the neighbour-joining method from a similarity matrix of pairwise comparisons made using the Jukes-Cantor algorithm. Sequences shown in blue are from A. gambiae. Dm = D. melanogaster, Hs = Homo sapiens, Rn = Rattus norvegicus. Nodes with distance bootstrap values (500 replicates) of > 70% are marked by*.

Figure 4

Histogram showing the frequency of intron sizes in the A. gambiae GST supergene family.

Figure 5

Schematic diagram indicating the organisation of GST genes in the A. gambiae genome. The numbers represent polytene chromosome divisions. Those genes marked with * are alternatively spliced to produce multiple transcripts.

Figure 6

A: Neighbour-joining tree illustrating the relationship between A. gambiae and D. melanogaster epsilon class GSTs. ClustalW was used to align the putative amino acid sequences of the 8 epsilon class GST subunits in A. gambiae with the 14 putative epsilon class GSTs from D. melanogaster and with GSTS1 from D. melanogaster (as an outgroup). The tree was constructed by the neighbour-joining method from a similarity matrix of pairwise comparisons made using the Jukes-Cantor algorithm. Nodes with distance bootstrap values (500 replicates) of >70% are marked by *. B, Epsilon class GST arrangement in A. gambiae and D. melanogaster. Arrows mark directions of transcription. The solid bars denote the genes and the vertical lines within these mark the approximate introns position.

Figure 7

Neighbour-joining tree illustrating the relationship between A. gambiae and D. melanogaster delta class GSTs. ClustalW was used to align the putative amino acid sequences of the 18 delta GST subunits in A. gambiae with the 11 putative delta class GSTs from D. melanogaster and with GSTS1 from D. melanogaster (as an outgroup). The tree was constructed by the neighbour-joining method from a similarity matrix of pairwise comparisons made using the Jukes-Cantor algorithm. Nodes with distance bootstrap values (500 replicates) of >70% are marked by *. B, Delta class GST arrangement in A. gambiae and D. melanogaster. Arrows mark directions of transcription. The solid rectangles denote the exons of the genes. Agcp2459 and agcp2052 are two putative genes, identified by the automatic analysis of the A. gambiae genome, that interrupt the delta GST cluster on chromosome 2R, division 18B.

Figure 8

Schematic diagram showing the organisation of the sigma GST genes in A. gambiae and D. melanogaster. Each rectangle above the scale bar represents a different exon. The empty rectangles indicate 5'UTR regions whilst the rectangles in solid colours are coding sequence. The cDNAs produced by splicing of these genes are shown beneath the scale bar.

Figure 9

Expression levels of the epsilon class GSTs in two strains of A. gambiae. A Transcript copy number determined by qPCR using cDNA as a template. Results represent the average of three independent cDNA samples. B gene copy number, determined by using three independent aliquots of mass homogenates of gDNA as a template. In both cases, copy numbers were normalised against the copy number of the S7 genes and error bars represent standard deviations. Statistically significant differences (*p < 0.05, **p < 0.01) between ZAN/U and Kisumu are indicated. Kisumu is an insecticide susceptible strain, ZAN/U is resistant to DDT.