Aphids acquired symbiotic genes via lateral gene transfer

Abstract

Background: Aphids possess bacteriocytes, which are cells specifically differentiated to harbour the obligate mutualist Buchnera aphidicola (gamma-Proteobacteria). Buchnera has lost many of the genes that appear to be essential for bacterial life. From the bacteriocyte of the pea aphid Acyrthosiphon pisum, we previously identified two clusters of expressed sequence tags that display similarity only to bacterial genes. Southern blot analysis demonstrated that they are encoded in the aphid genome. In this study, in order to assess the possibility of lateral gene transfer, we determined the full-length sequences of these transcripts, and performed detailed structural and phylogenetic analyses. We further examined their expression levels in the bacteriocyte using real-time quantitative RT-PCR.

Results: Sequence similarity searches demonstrated that these fully sequenced transcripts are significantly similar to the bacterial genes ldcA (product, LD-carboxypeptidase) and rlpA (product, rare lipoprotein A), respectively. Buchnera lacks these genes, whereas many other bacteria, including Escherichia coli, a close relative of Buchnera, possess both ldcA and rlpA. Molecular phylogenetic analysis clearly demonstrated that the aphid ldcA was derived from a rickettsial bacterium closely related to the extant Wolbachia spp. (alpha-Proteobacteria, Rickettsiales), which are intracellular symbionts of various lineages of arthropods. The evolutionary origin of rlpA was not fully resolved, but it was clearly demonstrated that its double-psi beta-barrel domain is of bacterial origin. Real-time quantitative RT-PCR demonstrated that ldcA and rlpA are expressed 11.6 and 154-fold higher in the bacteriocyte than in the whole body, respectively. LdcA is an enzyme required for recycling murein (peptidoglycan), which is a component of the bacterial cell wall. As Buchnera possesses a cell wall composed of murein but lacks ldcA, a high level of expression of the aphid ldcA in the bacteriocyte may be essential to maintain Buchnera. Although the function of RlpA is not well known, conspicuous up-regulation of the aphid rlpA in the bacteriocyte implies that this gene is also essential for Buchnera.

Conclusion: In this study, we obtained several lines of evidence indicating that aphids acquired genes from bacteria via lateral gene transfer and that these genes are used to maintain the obligately mutualistic bacterium, Buchnera.

Figure 1

Structure of the aphid LdcA. (A) ClustalX alignment of amino acid sequences of LdcAs. Residues conserved in all lineages, four lineages, and three lineages are shaded black, dark gray, and light gray, respectively. Arrowheads indicate the residues required for LdcA activity. The long form of the aphid LdcA is used for the alignment. (B) Domain structure of the aphid LdcA protein and structures of the corresponding mRNA and genomic DNA.

Figure 2

Structure of the aphid RlpA. (A) ClustalX alignment of amino acid sequences of RlpAs. Residues conserved in all lineages, four lineages, and three lineages are shaded black, dark gray, and light gray, respectively. Residues contributing to the domain structures are boxed. A residue that is different between type I and type II of A. pisum RlpA is denoted with an asterisk. (B) Domain structure of the aphid RlpA protein and structures of the corresponding mRNA and genomic DNA. For reference, the domain structure of E. coli RlpA is also shown. (C) Alignment of ICK motifs of the aphid RlpA with those of three antimicrobial peptides of the harlequin beetle, Acrocinus longimanus. Asterisks indicate the residues conserved in all the sequences. The grey background indicates conserved cysteines. The percentage of identity and E-value of bl2seq performed between each sequence and the ICK motif-1 (the one on the N- terminal side) of the pea aphid RlpA are shown on the right. Dashes (-) indicate alignment gaps. Dots (.) represent residues identical to those of the pea aphid RlpA.

Figure 3

Phylogenetic position of the aphid LdcA. A total of 246 aligned amino acid sites were subjected to the analysis. A neighbour-joining tree is shown, while the ML tree and BP tree exhibited substantially the same topologies. On each node, bootstrap support values over 50% are shown (NJ above, ML below). Thickened nodes indicate the Bayesian posterior probabilities are > 0.95. Taxonomic positions (eubacterial taxonomy unless otherwise stated) are shown in brackets. α, β, γ, and δ indicate proteobacterial classes. The A. pisum-Rickettsiales cluster is shown in red. The sequence from the fungus G. zeae is shown in green.

Figure 4

Phylogenetic position of aphid RlpA. A total of 76 aligned amino acid sites were subjected to the analysis. A neighbour-joining tree is shown, while the ML tree and BP tree exhibited substantially the same topologies. On each node, bootstrap support values over 50% are shown (NJ above, ML below). Thickened nodes indicate the Bayesian posterior probabilities are > 0.95. Taxonomic positions (eubacterial taxonomy unless otherwise stated) are shown in brackets. α, β, γ, and ε indicate proteobacterial classes.

Figure 5

Expression levels of ldcA and rlpA in the bacteriocyte. Ivory columns, expression levels in the whole body; blue columns, expression levels in the bacteriocyte; bars, standard errors (n = 6). The expression levels are shown in terms of mRNA copies of target genes per copy of mRNA for RpL7. Asterisks indicate statistically significant differences (Mann-Whitney U-test; **, p < 0.01).