MBE Advance Access originally published online on November 6, 2008
Molecular Biology and Evolution 2009 26(2):367-374; doi:10.1093/molbev/msn253
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
An Ancient Horizontal Gene Transfer between Mosquito and the Endosymbiotic Bacterium Wolbachia pipientis
School of Integrative Biology, University of Queensland, Brisbane, Queensland, Australia
E-mail: scott.oneill{at}uq.edu.au.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The extent and biological relevance of horizontal gene transfer (HGT) in eukaryotic evolution remain highly controversial. Recent studies have demonstrated frequent and large-scale HGT from endosymbiotic bacteria to their hosts, but the great majority of these transferred genes rapidly become nonfunctional in the recipient genome. Here, we investigate an ancient HGT between a host metazoan and an endosymbiotic bacterium, Wolbachia pipientis. The transferred gene has so far been found only in mosquitoes and Wolbachia. In mosquitoes, it is a member of a gene family encoding candidate receptors required for malaria sporozoite invasion of the mosquito salivary gland. The gene copy in Wolbachia has substantially diverged in sequence from the mosquito homolog, is evolving under purifying selection, and is expressed, suggesting that this gene is also functional in the bacterial genome. Several lines of evidence indicate that the gene may have been transferred from eukaryotic host to bacterial endosymbiont. Regardless of the direction of transfer, however, these results demonstrate that interdomain HGT may give rise to functional, persistent, and possibly evolutionarily significant new genes.
Key Words: endosymbiont • symbiosis • Wolbachia pipientis • Aedes aegypti • Anopheles gambiae • SGS
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Vertically transmitted endosymbiotic bacteria are an ideal model system for investigating horizontal gene transfer (HGT) between prokaryotes and eukaryotes. The transovarial route of transmission of these bacteria means that they are intimately associated with host germ tissue over long periods of evolutionary time, which is expected to enhance the frequency of HGT events. Biological interdependence of host and symbiont may also increase the likelihood that transferred genes are functionally relevant to the recipient and thus maintained in its genome. Wolbachia pipientis is an
-proteobacterial obligate endosymbiont of arthropods and nematodes that is estimated to infect at least 20% of insect species (Werren and Windsor 2000
) and possibly as many as 66% (Hilgenboecker et al. 2008). Recent studies have demonstrated that genetic sequences, ranging in size from single genes to entire bacterial genomes, have been transferred from Wolbachia to many of their insect hosts (Kondo et al. 2002; Fenn et al. 2006; Dunning Hotopp et al. 2007). Given Wolbachias estimated prevalence, interdomain gene transfer may occur much more frequently than has previously been thought.
Most of the genes transferred between Wolbachia and their hosts are nonfunctional in the recipient genome, as these gene copies often contain stop codons, frameshifts, or retroelement insertions and/or are transcriptionally inactive (Dunning Hotopp et al. 2007; Nikoh et al. 2008). The transferred genes therefore do not typically fulfill the basic criteria for evolutionary significance: longevity and integration into the biology of the recipient taxon (Blaxter 2007). In contrast, the family of salivary gland surface (SGS) proteins previously characterized in Aedes aegypti and Anopheles gambiae meet both of these criteria.
SGS genes have no known nonmosquito eukaryotic homologs but share moderate sequence similarity with the hypothetical gene WD0513 from wMel, the Wolbachia endosymbiont of Drosophila melanogaster (Wu et al. 2004). This has led to the suggestion that these genes arose as a result of an ancient HGT, probably from Wolbachia to mosquito (Arca et al. 2005; Korochkina et al. 2006). The proteins encoded by SGS genes play a role in insect–malaria interactions (Korochkina et al. 2006). After a female mosquito takes a malaria-infected blood meal, malarial sporozoites must invade the salivary glands before the parasite can be transmitted to a new host. This invasion is thought to require specific interactions between sporozoite ligands and mosquito salivary gland receptors (Brennan et al. 2000). Aedes aegypti aaSGS1 (encoded by AAEL009993), for example, localizes to the basal lamina of the salivary glands of adult female mosquitoes, and inoculation of mosquitoes with anti-aaSGS1 antibodies reduces sporozoite invasion by approximately 65% (Korochkina et al. 2006), indicating that sporozoites interact with aaSGS1 during salivary gland invasion. SGS homologs are also expressed in adult female salivary glands in Aedes albopictus (Arca et al. 2007), An. gambiae (Arca et al. 2005; Korochkina et al. 2006), and Anopheles funestus (Calvo et al. 2007), where they may perform similar functions.
Here, we use the complete genome sequences of the mosquitoes Ae. aegypti, An. gambiae, and Culex quinquefasciatus, and of Wolbachia strains infecting D. melanogaster (wMel) and Cx. quinquefasciatus (wPip), to confirm that the SGS genes have been horizontally transferred between these taxa. To determine the direction of HGT, we then consider data on the taxonomic distribution, phylogenetic relationships, gene length, GC content, synonymous codon usage, and intronic content of SGS genes. Taken together, we believe that the balance of evidence supports a direction of transfer from eukaryotic host to Wolbachia endosymbiont.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Sequence Data
To identify all SGS gene homologs in mosquitoes, we used the amino acid sequences of aaSGS1 (GenBank accession number AAV28546 [GenBank] ) and WD0513 (GenBank NP_966293 [GenBank] ) as BLASTP queries against the predicted peptide databases of Ae. aegypti (gene build AaegL1.1), An. gambiae (gene build AgamP3.4), and Cx. quinquefasciatus (gene build CpipJ1.1), downloaded from Vectorbase (http://www.vectorbase.org). To identify any unannotated homologs, we also used these two sequences as TBLASTN queries against the three assembled mosquito genome sequences, downloaded from Vectorbase. We extracted all hits from each search with E-values better than 1e–05. To confirm that hits were more closely related to SGS genes than to members of other gene families, we used their translations as BLASTP queries against the nonredundant (NR) database and checked that their best hit was either another mosquito SGS gene or WD0513. All mosquito homologs are listed in supplementary table S1, Supplementary Material online.
During this analysis, we noticed that many of the mosquito SGS genes returned a match to WD0512 in addition to WD0513. These genes are adjacent and overlapping (by 86 bp) in the wMel genome. We used WD0512 as a TBLASTN query against the mosquito genomes and found that the first 300–500 bp of this gene was homologous to the 3' end (or sometimes the 3' untranslated region [UTR]) of each SGS gene. The 3'-UTR of AAEL004181 was homologous to the full length of WD0512.
To identify homologs of WD0513 in other Wolbachia strains, we used the amino acid sequence encoded by this gene as a BLASTP query against the wMel and wBm proteomes and as a TBLASTN query against the genome assemblies of the Wolbachia strains wWil, wAna, and wSim (all http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi), wOv (www.sanger.ac.uk/Projects/Wolbachia/), and wPip (http://www.sanger.ac.uk/Projects/W_pipientis/). Only a single homolog was identified, in wPip. This gene, recently annotated as WP1346 (Klasson et al. 2008), is homologous to almost the full length of WD0513 and WD0512.
Phylogenetic Analysis
To identify possible outgroup sequences for our phylogenetic analysis, we used the amino acid sequences of WD0513 and each of the mosquito SGS proteins as BLASTP or TBLASTN queries against the NR, genomic survey sequences (GSS), whole genome shotgun reads, high throughput genomic sequences, and environmental sequence databases on NCBI (http://www.ncbi.nlm.nih.gov). We chose as outgroups four proteins that were in the top 10 nonmosquito hits for all queries against the NR database (no closer outgroups were identified in any of the other databases): Flavobacterium johnsoniae YD-repeat-containing protein, GenBank accession number YP_001196062.1; Bacteroides thetaiotaomicron conserved hypothetical protein, NP_811839.1
[GenBank]
; Bacteroides ovatus hypothetical protein, EDO13775
[GenBank]
.1; and Cellulophaga sp. cell wall associated RhsD protein, ZP_01049542.1.
We excluded both of the unannotated SGS homologs we identified (supplementary table S1, Supplementary Material online) from the phylogenetic analysis, as only fragmentary gene sequence was available for each. We also excluded three of the annotated genes, AAEL003694, CPIJ007816, and CPIJ004852 (although inclusion of these genes does not change our conclusions with regard to the direction of transfer; supplementary fig. S1, Supplementary Material online). AAEL003694 is a very short, truncated gene (coding for 846 residues, in comparison with, e.g., aaSGS1 with 3,060 residues). CPIJ007816 contains two tandem segments, each homologous to almost the full length of WD0513. We believe that CPIJ007816 probably represents two separate genes, but the true gene boundaries cannot be clearly determined. CPIJ004852 is a highly divergent gene that is very difficult to align with confidence and whose phylogenetic position varies with the method of alignment and outgroup choice.
The remaining sequences were translated into amino acids and aligned using T-coffee (Notredame et al. 2000). Regions of good sequence conservation were observed along the full length of the ingroup genes. Only the most highly conserved domain could be confidently aligned between ingroup and outgroup genes, and so this region alone was used for the phylogenetic analysis. Ingroup topology remains the same if outgroups are not included and a full-length alignment of ingroup sequences is used (data not shown). The alignment was edited by hand in Se-Al (Rambaut 1996) and then trimmed using Gblocks (Castresana 2000) with a range of stringency settings. All settings gave very similar results after phylogenetic analysis; we report results of settings –b3 = 15 and –b5 = a, which allow the retention of sites at which some sequences have a gap and removes any regions of the alignment containing more than 15 contiguous poorly conserved sites. This produced an alignment of 454 aa.
We constructed Bayesian phylogenetic trees using MrBayes (Huelsenbeck and Ronquist 2001), performing five runs each of four chains, sampling every 100th iteration over 10,000,000 generations after a burn-in of 25,000 samples and allowing mixed models of amino acid substitution. Maximum likelihood trees were constructed with PHYML (Guindon and Gascuel 2003), using the Jones, Taylor, and Thorton model of substitution and a gamma distribution of substitution rates with four rate classes, with the parameter estimated from the data. Shimodaira–Hasegawa tests of alternate topologies (Shimodaira and Hasegawa 1999; Goldman et al. 2000) were performed using Tree-Puzzle (Schmidt et al. 2002).
Gene Characteristics
We calculated the lengths of SGS genes based on annotated gene boundaries. For each mosquito species, we calculated the GC content of the annotated coding regions of each SGS gene and compared these values with the mean GC content of the other protein-coding genes in the genomes. To examine synonymous codon usage in SGS genes compared with other protein-coding genes in mosquito and Wolbachia genomes, we used R (http://www.r-project.org/) to implement an internal correspondence analysis of codon usage that corrects for differences in amino acid usage between proteins (Lobry and Chessel 2003). We examined the length, location, and conservation of introns annotated in mosquito SGS genes (http://www.vectorbase.org/).
Tests for Selection
To test whether WD0513 is evolving neutrally or under selection, we aligned the nucleotide sequence of this gene with the homologous portion of WP1346 and AAEL004181. We then performed two likelihood ratio tests. First, we compared a model in which all three branches had the same dN/dS ratio with one in which each branch had its own ratio, to determine whether it was appropriate to use branch-specific tests for selection. As the free-ratio model was a significantly better fit to the data (P > 0.0001), we then performed a second set of tests, comparing a model in which the dN/dS of the wMel or wPip branch was fixed to 1 with a model in which the dN/dS values of all branches were freely estimated. Model likelihoods were estimated using PAML (Yang 1997).
Homologs of Other Genes in the wMel Genomic Island
To test whether homologs of any of the other genes in the wMel genomic island surrounding WD0513 were present in wPip or in any of the mosquito genomes, we used the amino acid sequences encoded by each of the genes in the island as BLASTP queries against the mosquito gene builds and TBLASTN queries against the genome assemblies of the mosquitoes and wPip. As WD0514 is a gene containing ankyrin repeat domains, which retrieve spurious BLAST hits based on domain sequence similarity rather than overall gene homology, we also used the amino acid sequence of this gene without the ankyrin domains as a query. For the WD0514 homolog identified in Ae. aegypti and the homologs of multiple island genes identified in wPip, we confirmed that they were likely orthologs of the query genes by using them as BLAST queries back against the wMel protein set and confirming that they were reciprocal best BLAST hits.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
We used BLASTP and TBLASTN analyses to identify all SGS homologs present in publicly available sequence databases. There are at least seven SGS genes in the genomes of both Ae. aegypti and Cx. quinquefasciatus and at least four in An. gambiae (supplementary table S1, Supplementary Material online). We also identified a single homologous genomic region in both wMel and wPip (WD0512–WD0513 and WP1346, respectively). No other closely related eukaryotic or prokaryotic homologs were identified in any sequence database; the closest outgroups were relatively distant bacterial proteins (
23% amino acid identity over 20% of the length of the protein).
Phylogenetic Analysis
We then reconstructed the phylogeny of these genes using Bayesian and maximum likelihood methods (fig. 1; both methods inferred the same topology). If an SGS gene had been horizontally transferred from Wolbachia to mosquito, as previously suggested, we would expect the Wolbachia genes to cluster at the base of the tree. The earliest branching genes in the ingroup, however, are all from mosquitoes. In contrast, the two Wolbachia genes are nested high in the tree within a clade of mosquito genes. Some of the basal ingroup nodes of the tree have only moderate clade confidence and bootstrap values, but this topology is a significantly better fit to the data than one in which the two Wolbachia sequences branch at the base of the ingroup (P = 0.02, Shimodaira–Hasegawa test).
|
Because AAEL004181, the mosquito gene most closely related to the Wolbachia genes, is not nested directly within previously characterized SGS genes (Korochkina et al. 2006
) in this tree, we performed a test of genetic distances to be confident that it was a member of this same gene family. We aligned the amino acid sequences of AAEL004181, aaSGS1 (encoded by AAEL009993), agSGS2 (AGAP009916), agSGS4 (AGAP009917), and agSGS5 (AGAP009918) using T-coffee and then trimmed this alignment with Gblocks as before. We then used Tree-Puzzle to produce a matrix of gamma-corrected maximum likelihood distances between these sequences. Distances between AAEL004181 and the SGS genes ranged from 1.53 to 1.70 substitutions per residue, whereas distances between SGS genes ranged from 0.51 (aaSGS4–aaSGS5) to 1.75 (aaSGS1–agSGS2). Distances between AAEL004181 and the SGS genes are smaller than those between some of the previously annotated SGS genes themselves, and we are therefore confident that AAEL004181 is a member of this gene family. The two Wolbachia genes are not monophyletic in the tree: WP1346 and AAEL004181 cluster together, and WD0513 branches basally to them, possibly indicating that multiple horizontal transfers may have occurred. However, this topology is not a significantly better fit to the data than one in which AAEL004181 branches outside monophyletic Wolbachia sequences (P = 0.08, Shimodaira–Hasegawa test), and it is probably most conservative to consider that these three sequences form an unresolved trichotomy.
Gene Characteristics
To further investigate the direction of horizontal transfer, we tested whether WD0513 and the mosquito SGS genes have characteristics typical of eukaryotic genes (gene length, GC content, and codon usage consistent with the whole mosquito genomes and introns present) or prokaryotic genes (gene length, GC content, and codon usage more similar to that of the Wolbachia genome, no introns). Results are summarized in table 1.
|
The mean length of the mosquito SGS genes is 12,222 bp, which is on the order of the average length of all genes across the three mosquito genomes (8,367 bp). In contrast, WD0513 is 8,532 bp, over 10 times longer than the mean gene length in the wMel genome (852 bp). WP1346 is 11,646 bp in length, again over 10 times longer than the mean gene length of 944 bp in wPip.
Although the GC content of SGS genes is typically lower than the average GC for other genes in the mosquito genomes, they are rarely extreme outliers. Two SGS genes in Ae. aegypti have 42% GC content, placing them in the fourth percentile of genes for GC% in the Ae. aegypti genome, but many other SGS genes have GC content entirely consistent with the rest of the genome, for example, all of the Cx. quinquefasciatus SGS genes (41–50% GC, 21st–53rd percentiles). The GC content of WD0513, 35%, is the same as the mean GC across the entire wMel genome, and an almost identical pattern is observed for WP1346 and wPip. Similarly, our correspondence analysis also shows that the synonymous codon usage of the SGS genes is typical for their respective mosquito genomes and that the synonymous codon usage of WD0513 is consistent with that of the wMel genome (fig. 2).
|
Almost all SGS genes have one or more annotated introns (http://www.vectorbase.org). However, it seems likely that many or all of these introns may be artifacts of annotation. If the annotated introns are genuine, we would expect to see one of two possible patterns of intron conservation between recently diverged SGS genes. For orthologs or paralogs created without an mRNA intermediate, the position and length of introns should be largely consistent between gene copies. For paralogs created by retrotransposition of mature or partially processed mRNA, the duplicate gene should either be intronless or should show loss of 5' introns but conservation of 3' introns (as splicing proceeds 5' to 3'). We examined the position and length of Vectorbase-annotated introns in three sets of closely related SGS gene paralogs, one from each mosquito species, to determine whether we observe either conservation or partial or complete loss of introns in these genes. No such patterns were seen (Supplementary Material online). As these introns are also currently unsupported by expression data, we believe that most or all of the introns annotated in SGS genes are artifactual.
Selection
We investigated whether WD0513 and WP1346 are evolving neutrally or under selection by aligning these genes with AAEL004181 and examining the ratio of nonsynonymous to synonymous changes (dN/dS) along each branch of the triplet. WD0513 has a dN/dS of 0.17, and WP1346 has a dN/dS of 0.14. Both these values are significantly lower than a neutral dN/dS of 1 (P > 0.0001, likelihood ratio test, Bonferroni correction for multiple tests), indicating that these genes are probably evolving under purifying selection. Recombination (as is known to occur between strains of Wolbachia, Baldo et al. 2005) can affect analyses of selection (Anisimova et al. 2003). However, recombination tends to bias these analyses toward detection of positive rather than purifying selection, and so our estimate of dN/dS is likely to be conservative.
The genomic region in wPip homologous to the 5' end of WD0513 has been disrupted by the insertion of a transposable element of the IS110 family, and the start codon of WP1346 corresponds to codon 154 in WD0513. It is possible that this slightly truncated protein remains functional in wPip, and the low dN/dS value for this gene reflects current purifying selection. Alternatively, the IS110 element may be a recent insertion that has only recently disrupted this gene; such recent activity would be consistent with the fact that the three copies of this element in the wPip genome are identical to one another at the nucleotide level. Nonetheless, the absence of stop codons in WD0513 and WP1346, together with the low estimated values of dN/dS, offers further evidence for the action of long-term purifying selection on these genes.
Multiple Genes Transferred from Mosquito to Wolbachia
WD0513 in wMel is part of a larger genomic island of 13 genes (WD0506–WD0518) that is present in wMel and two closely related A-group strains of Wolbachia but absent from other A- and B-group strains examined (Iturbe-Ormaetxe et al. 2005). Given this unusual taxonomic distribution, and the presence of multiple transposable element genes in the island, we tested whether other island genes might also have originated by HGT from mosquito. Most genes in the genomic island had either no mosquito homologs or had many closer prokaryotic than eukaryotic homologs, indicating that no interdomain HGT of these genes has taken place.
We did, however, identify likely ancestors of WD0512 and WD0514 in the mosquito genomes. The 3' end or the 3'-UTR of every mosquito SGS gene we identified is homologous to the first 300–500 bp of WD0512, and the 3'-UTR of AAEL004181 is homologous to the full length of WD0512. WD0514 has a single mosquito homolog, AAEL004188, which is located directly upstream of AAEL004181 in the Ae. aegypti genome. The presence of full-length homologs of WD0512 and WD0514 on either side of AAEL004181 offers further support for the close phylogenetic relationship between this SGS gene and WD0513 (fig. 1).
Horizontal Transfer between Wolbachia Strains
Orthologs of numerous genes from the genomic island were found clustered in the wPip genome. It seems likely that the entire island was at one time present in the wPip genome but that it has subsequently been subject to partial or complete duplication followed by gene loss (data not shown). This island is absent from all A- and B-group Wolbachia tested other than the closely related strains wMel, wMelCS, and wMelPop (A-group) (Iturbe-Ormaetxe et al. 2005) and wPip (B-group). This discontinuous taxonomic distribution suggests either that the island was present in the ancestor of A-group and B-group Wolbachia and has subsequently been lost in the majority of sampled Wolbachia strains or that it has been horizontally transferred between strains more recently. Several lines of evidence support interstrain HGT as a more likely explanation. First, the wMel and wPip orthologs of island genes have anomalously high sequence similarity (up to 96% nucleotide identity for each ortholog pair's longest BLASTN high-scoring segment pair) compared with the rest of the two genomes. Furthermore, numerous independent losses would be required to produce the genomic island's current taxonomic distribution if it had been present in the ancestor of A- and B-group Wolbachia (Iturbe-Ormaetxe et al. 2005). Finally, mechanisms for genetic transfer between A- and B-group Wolbachia clearly exist, given the previously reported examples of intergroup HGT (Masui et al. 2000; Bordenstein and Wernegreen 2004).
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Our analyses support the claim of Korochkina et al. (2006) that the sequence similarity observed between the SGS genes and WD0513 is due to an HGT event. Genes may have similar sequences due either to common origin or to convergence, and there are two reasons to believe that the data examined here are not an example of convergence. First, convergent similarity is likely to be restricted to relatively short functional regions of genes rather than complete gene sequences (Eisen 1998
). In contrast to this, we observe sequence similarity along the full lengths of WD0513 and AAEL004181. Secondly, our putatively transferred sequence fragment contains multiple genes: WD0512 and WD0514, the genes flanking WD0513 in the wMel genome, show strong similarity to the 3'-UTR of AAEL004181 and to the upstream Ae. aegypti gene AAEL004188, respectively. Convergence across the lengths of multiple neighboring genes and noncoding regions seems extremely unlikely.
The direction in which this horizontal transfer occurred is less clear, but we believe that the balance of evidence (table 1) supports a transfer from a eukaryotic host (probably mosquito) to Wolbachia. SGS gene homologs have been identified in every mosquito species for which we have whole genomes or salivary gland expression libraries. In contrast, homologs have been found in only a minority of the Wolbachia strains sampled (Iturbe-Ormaetxe et al. 2005). This sparse distribution in Wolbachia suggests that it was the recipient rather than the donor of the transferred gene (Andersson 2005). This direction of transfer is also strongly implied by the phylogeny of these genes (fig. 1), as the basal position of mosquito sequences, and the nesting of the Wolbachia homologs within a mosquito clade, is supported both by multiple tree reconstruction methods and by a maximum likelihood test of alternative topologies.
Finally, a direction of transfer from eukaryote to prokaryote is also supported by the lengths of the SGS genes and their Wolbachia homologs. The mosquito SGS genes are relatively long even for the eukaryotic genomes in which they are found; even the shortest of them would be exceptionally long for a gene derived from a bacterium. WD0513, at 8,532 bp, is the longest open reading frame in the wMel genome and an extreme outlier in terms of length. Only one other wMel gene, WD0024, which is a fusion of two RNA polymerase subunit genes, approaches this length; the third longest gene, by comparison, is only 4,725 bp. The mean gene length in wMel is only 852 bp, less than one-tenth of the length of WD0513. Similarly, WP1346 is 11,649 bp long, over 10 times as long as the mean gene length of 944 bp in wPip. In this regard, WD0513 and WP1346 appear to be more like typical eukaryotic than prokaryotic genes.
If WD0513 has been transferred from a eukaryotic host, we might expect it to retain other characteristics of its native genome, such as GC content and patterns of synonymous codon usage. WD0513, however, does not differ from the rest of the wMel genome for these traits. The SGS genes do not show any consistent deviation from the GC content or codon usage of the mosquito genomes either, and these traits are therefore uninformative for inferring the direction of transfer. This lack of signal is presumably due to amelioration of the original base composition and codon usage of the transferred sequences to reflect the mutational biases and/or selective pressures of the recipient genome. As this transfer event was an ancient one (WD0513 is only 50% identical at the amino acid level to AAEL004181), there has been substantial evolutionary time for this amelioration to take place.
Only one set of data potentially supports a transfer in the opposite direction, from Wolbachia to host. It is rare for mosquito genes to lack introns—90% of Ae. aegypti genes have introns, and the proportion rises to almost 94% in An. gambiae (Nene et al. 2007). If our assessment is correct, and the introns currently annotated in SGS genes are artifactual, then horizontal transfer of these genes from a bacterial source is a possible explanation for their lack of introns. On the other hand, very few intronless eukaryotic genes have arisen from HGT: most are thought to have been generated by retrotransposition followed by genomic reintegration or other mechanisms requiring only vertical descent (Sakharkar et al. 2006). This indication of directionality is therefore not compelling.
Recent findings of substantial HGT between prokaryotic endosymbionts and their eukaryotic hosts (Kondo et al. 2002; Dunning Hotopp et al. 2007; Nikoh et al. 2008) indicate that interdomain gene transfers may be unexpectedly frequent. Such transfers have been assumed to be unidirectional, from prokaryote to eukaryote, but multiple lines of evidence presented here support the conclusion that WD0513 is the result of eukaryote-to-prokaryote HGT. Gene transfers in this direction are far less common than from prokaryote to eukaryote (Andersson 2005), but given the frequency with which hosts and endosymbionts exchange genetic material over evolutionary time scales (Dunning Hotopp et al. 2007), it is entirely possible that one or more host-to-Wolbachia transfers may have occurred. Host genomes may therefore be occasional, but important, sources of novel genetic material for their bacterial endosymbionts.
Regardless of the direction in which this HGT occurred, it remains an event of strong evolutionary interest. Unlike many of the interdomain transfers previously described (Dunning Hotopp et al. 2007; Nikoh et al. 2008), the genetic distance between WD0513 and the SGS genes indicates that the transferred copy has been retained over a substantial period of evolutionary time, and both donor and recipient genes appear to be functional. Some members of the SGS gene family have been previously characterized (Korochkina et al. 2006), and although the function of WD0513 is currently unknown, it has diverged substantially from its mosquito homologs without pseudogenization, is evolving under purifying selection, and is expressed (Iturbe-Ormaetxe et al. 2005), indicating that it probably plays some role in the biology of the Wolbachia strains in which it is found. This HGT thus fulfills the criteria of evolutionary significance (Blaxter 2007), demonstrating that interdomain HGTs are potentially powerful sources of evolutionary innovation.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Supplementary table S1 and supplementary figure S1 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
We thank Marie Sémon, Lyn Cook, Mike Crisp, Mark Ragan, Steve Sinkins, members of the O'Neill and McGraw labs, and attendees of the fifth International Wolbachia Conference for valuable discussion and comments on the analyses. This work was supported by grants from the Australian Research Council (ARC; DP0557987) and National Health and Medical Research Council (NHMRC; 496601), and a UQ postdoctoral fellowship to M.W.
|
Footnotes |
|---|
Laura Katz, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Andersson JO. Lateral gene transfer in eukaryotes. CMLS—Cell Mol Life Sci (2005) 62:1182–1197.[CrossRef]
Anisimova M, Nielsen R, Yang Z. Effect of recombination on the accuracy of the likelihood method for detecting positive selection at amino acid sites. Genetics (2003) 164:1229–1236.
Arca B, Lombardo F, Francischetti IMB, Pham VM, Mestres-Simon M, Andersen JF, Ribeiro JMC. An insight into the sialome of the adult female mosquito Aedes albopictus. Insect Biochem Mol Biol (2007) 37:107–127.[CrossRef][Web of Science][Medline]
Arca B, Lombardo F, Valenzuela JG, Francischetti IMB, Marinotti O, Coluzzi M, Ribeiro JAC. An updated catalogue of salivary gland transcripts in the adult female mosquito, Anopheles gambiae. J Exp Biol (2005) 208:3971–3986.
Baldo L, Bordenstein S, Wernegreen JJ, Werren JH. Widespread recombination throughout Wolbachia genomes. Mol Biol Evol (2005) 23:437–449.[CrossRef][Web of Science][Medline]
Blaxter M. Symbiont genes in host genomes: fragments with a future? Cell Host Microbe (2007) 2:211–213.[CrossRef][Web of Science][Medline]
Bordenstein SR, Wernegreen JJ. Bacteriophage flux in endosymbionts (Wolbachia): infection frequency, lateral transfer, and recombination rates. Mol Biol Evol (2004) 21:1981–1991.
Brennan JDG, Kent M, Dhar R, Fujioka H, Kumar N. Anopheles gambiae salivary gland proteins as putative targets for blocking transmission of malaria parasites. Proc Natl Acad Sci USA (2000) 97:13859–13864.
Calvo E, Dao A, Pham VM, Ribeiro JMC. An insight into the sialome of Anopheles funestus reveals an emerging pattern in anopheline salivary protein families. Insect Biochem Mol Biol (2007) 37:164–175.[CrossRef][Web of Science][Medline]
Castresana J. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol (2000) 17:540–552.
Dunning Hotopp JC, Clark ME, Oliveira DCSG, et al(20 co-authors). Widespread lateral gene transfer from intracellular bacteria to multicellular eukaryotes. Science (2007) 317:1753–1756.
Eisen JA. Phylogenomics: improving functional predictions for uncharacterized genes by evolutionary analysis. Genome Res (1998) 8:163–167.
Fenn K, Conlon C, Jones M, Quail MA, Holroyd NE, Parkhill J, Blaxter M. Phylogenetic relationships of the Wolbachia of nematodes and arthropods. PLoS Pathog (2006) 2:887–899.[Web of Science]
Goldman N, Anderson JP, Rodrigo AG. Likelihood-based tests of topologies in phylogenetics. Syst Biol (2000) 49:652–670.
Guindon S, Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol (2003) 52:696–704.
Hilgenboecker K, Hammerstein P, Schlattmann P, Telschow A, Werren JH. How many species are infected with Wolbachia?—a statistical analysis of current data. FEMS Microbiol Lett (2008) 281:215–220.[CrossRef][Web of Science][Medline]
Huelsenbeck JP, Ronquist F. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics (2001) 17:754–755.
Iturbe-Ormaetxe I, Burke GR, Rieger M, O'Neil SL. Distribution, expression and motif variability of ankyrin domain genes in Wolbachia pipientis. J Bacteriol (2005) 187:5136–5145.
Klasson L, Walker T, Sebaihia M, et al(12 co-authors). Genome evolution of Wolbachia strain wPip from the Culex pipiens group. Mol Biol Evol (2008) 25:1877–1887.
Kondo N, Nikoh N, Ijichi N, Shimada M, Fukatsu T. Genome fragment of Wolbachia endosymbiont transferred to X chromosome of host insect. Proc Natl Acad Sci USA (2002) 99:14280–14285.
Korochkina S, Barreau C, Pradel G, Jeffery E, Li J, Natarajan R, Shabanowitz J, Hunt D, Frevert U, Vernick KD. A mosquito-specific protein family includes candidate receptors for malaria sporozoite invasion of salivary glands. Cell Microbiol (2006) 8:163–175.[CrossRef][Web of Science][Medline]
Lobry JR, Chessel D. Internal correspondence analysis of codon and amino-acid usage in thermophilic bacteria. J Appl Genet (2003) 44:235–261.[Medline]
Masui S, Kamoda S, Sasaki T, Ishikawa H. Distribution and evolution of bacteriophage WO in Wolbachia, the endosymbiont causing sexual alterations in arthropods. J Mol Evol (2000) 51:491–497.[Web of Science][Medline]
Nene V, Wortman JR, Lawson D, et al(95 co-authors). Genome sequence of Aedes aegypti, a major arbovirus vector. Science (2007) 316:1718–1723.
Nikoh N, Tanaka K, Shibata F, Kondo N, Hizume M, Shimada M, Fukatsu T. Wolbachia genome integrated in an insect chromosome: evolution and fate of laterally transferred endosymbiont genes. Genome Res (2008) 18:272–280.
Notredame C, Higgins DG, Heringa J. T-coffee: a novel method for fast and accurate multiple sequence alignment. J Mol Biol (2000) 302:205–217.[CrossRef][Web of Science][Medline]
Rambaut A. Se-Al: sequence alignment editor v2.0. (1996) Available from: http://evolve.zoo.ox.ac.uk/.
Sakharkar KR, Sakharkar MK, Culiat CT, Chow VTK, Pervaiz S. Functional and evolutionary analyses on expressed intronless genes in the mouse genome. FEBS Lett (2006) 580:1472–1478.[CrossRef][Web of Science][Medline]
Schmidt HA, Strimmer K, Vingron M, von Haeseler A. TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics (2002) 18:502–504.
Shimodaira H, Hasegawa M. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol Biol Evol (1999) 16:1114–1116.[Web of Science]
Werren JH, Windsor DM. Wolbachia infection frequencies in insects: evidence of a global equilibrium? Proc R Soc Lond Ser B Biol Sci (2000) 267:1277–1285.
Wu M, Sun LV, Vamathevan J, et al(30 co-authors). Phylogenomics of the reproductive parasite Wolbachia pipientis wMel: a streamlined genome overrun by mobile genetic elements. PLoS Biol (2004) 2:327–341.[CrossRef][Web of Science]
Yang ZH. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci (1997) 13:555–556.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Aikawa, H. Anbutsu, N. Nikoh, T. Kikuchi, F. Shibata, and T. Fukatsu Longicorn beetle that vectors pinewood nematode carries many Wolbachia genes on an autosome Proc R Soc B, November 7, 2009; 276(1674): 3791 - 3798. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Ishmael, J. C. D. Hotopp, P. Ioannidis, S. Biber, J. Sakamoto, S. Siozios, V. Nene, J. Werren, K. Bourtzis, S. R. Bordenstein, et al. Extensive genomic diversity of closely related Wolbachia strains Microbiology, July 1, 2009; 155(7): 2211 - 2222. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||