Widespread Lateral Gene Transfer from Intracellular Bacteria to Multicellular Eukaryotes

See allHide authors and affiliations

Science  21 Sep 2007:
Vol. 317, Issue 5845, pp. 1753-1756
DOI: 10.1126/science.1142490


Although common among bacteria, lateral gene transfer—the movement of genes between distantly related organisms—is thought to occur only rarely between bacteria and multicellular eukaryotes. However, the presence of endosymbionts, such as Wolbachia pipientis, within some eukaryotic germlines may facilitate bacterial gene transfers to eukaryotic host genomes. We therefore examined host genomes for evidence of gene transfer events from Wolbachia bacteria to their hosts. We found and confirmed transfers into the genomes of four insect and four nematode species that range from nearly the entire Wolbachia genome (>1 megabase) to short (<500 base pairs) insertions. Potential Wolbachia-to-host transfers were also detected computationally in three additional sequenced insect genomes. We also show that some of these inserted Wolbachia genes are transcribed within eukaryotic cells lacking endosymbionts. Therefore, heritable lateral gene transfer occurs into eukaryotic hosts from their prokaryote symbionts, potentially providing a mechanism for acquisition of new genes and functions.

The transfer of DNA between diverse organisms, lateral gene transfer (LGT), facilitates the acquisition of novel gene functions. Among Eubacteria, LGT is involved in the evolution of antibiotic resistance, pathogenicity, and metabolic pathways (1). Rare LGT events have also been identified between higher eukaryotes with segregated germ cells (2), demonstrating that even these organisms can acquire novel DNA. Although most described LGT events occur within a single domain of life, LGT has been described both between Eubacteria and Archaea (3) and between prokaryotes and phagotrophic unicellular eukaryotes (4, 5). However, few interdomain transfers involving higher multicellular eukaryotes have been found.

Wolbachia pipientis is a maternally inherited endosymbiont that infects a wide range of arthropods, including at least 20% of insect species, as well as filarial nematodes (6). It is present in developing gametes (6) and so provides circumstances conducive for heritable transfer of bacterial genes to the eukaryotic hosts. Wolbachia-host transfer has been described in the bean beetle Callosobruchus chinensis (7) and in the filarial nematode Onchocerca spp. (8).

We have found Wolbachia inserts in the genomes of additional diverse invertebrate taxa, including fruit flies, wasps, and nematodes. A comparison of the published genome of the Wolbachia endosymbiont of Drosophila melanogaster (9) and assemblies of Wolbachia clone mates (10) from fruit fly whole-genome shotgun sequencing data revealed a large Wolbachia insert in the genome of the widespread tropical fruit fly D. ananassae. Numerous contiguous, overlapping, clone sequences (contigs) were found that harbored junctions between Drosophila retrotransposons and Wolbachia genes. The large number of these junctions and the deep sequencing coverage across the junctions indicated that these inserts were probably not due to chimeric libraries or assemblies. To validate these observations, we amplified five Drosophila-Wolbachia junctions with polymerase chain reaction (PCR) and verified the end sequences for three of them. Fluorescence in situ hybridization (FISH) of banded polytene chromosomes with fluorescein-labeled probes of two Wolbachia genes (11) revealed the presence of Wolbachia genes on the 2L chromosome of D. ananassae (Fig. 1).

Fig. 1.

Fluorescence microscopy evidence supporting Wolbachia/host LGT. DNA in the polytene chromosomes of D. ananassae were stained with propidium iodide (red), whereas a probe for the Wolbachia gene WD_0484 bound to a unique location (green, arrow) on chromosome 2L.

We found that nearly the entire Wolbachia genome was transferred to the fly nuclear genome, as evidenced by the presence of PCR-amplified products from 44 of 45 physically distant Wolbachia genes in cured strains of D. ananassae Hawaii verified by microscopy to be lacking the endosymbiont after treatment with antibiotics (fig. S1) (11). In contrast, only spurious, incorrectly sized, and weak amplification was detected from a cured control line lacking these inserts (Townsville). The 45 genes assayed (table S1) are spaced throughout the Wolbachia genome. Thus, the high proportion of amplified genes suggests gene transfer of nearly the entire Wolbachia genome to the insect genome.

A 14-kb region containing four Wolbachia genes with two retrotransposon insertions was sequenced (11) from a single bacteria artificial chromosome (BAC), constituting an independent source of DNA as compared with the largely plasmid-derived whole-genome sequence of D. ananassae. The two retroelements each contained 5–base pair (bp) target site duplications (9/10 bp identical), long terminal repeats, and gag-pol genes (Fig. 2A) indicating that the Wolbachia insert is accumulating retroelements. Insertion of this region appears to be recent, as shown by the nearly identical target site duplications and the >90% nucleotide identity between corresponding endosymbiont genes and sequenced homologs in the D. ananassae chromosome.

Fig. 2.

Schematics of Wolbachia inserts in host chromosomes. (A) Contigs containing Wolbachia sequences generated from the D. ananassae Hawaii shotgun sequencing project are segregated into sequences coming from the endosymbiont (wAna) or from the D. ananassae chromosome (Dana) on the basis of presence or absence of eukaryotic genes in the contigs. These are compared to those from the reference D. melanogaster Wolbachia genome (wMel) and a D. ananassae BAC. NAD, nicotinamide adenine dinucleotide. (B) Fragments of the Wolbachia gene WD_0024 gene have inserted into different positions in the N. giraulti (NG) and N. vitripennis (NV) genomes with unique insertions in each lineage, including N. longicornis (NL). (C) A region in the D. immitis genome (Dg2) that is transcribed has introns similar to sequences from the Wolbachia infecting B. malayi (wBm). All matches in (A) and (B) have >90% nucleotide identity; those in (C) have >75% nucleotide identity. TPR, tetratrico peptide repeat; CDS, coding sequence.

Crosses between Wolbachia-free Hawaii males (with the insert) and Wolbachia-free Mexico females (without the insert) revealed that the insert is paternally inherited by offspring of both sexes, confirming that Wolbachia genes are inserted into an autosome. Because Wolbachia infections are maternally inherited, this also confirms that PCR amplification in the antibiotic-treated line is not due to an undetectable infection. Furthermore, the Hawaii and Mexico crosses revealed Mendelian, autosomal inheritance of Wolbachia inserts [paternal N = 57, proportion of offspring with Wolbachia genes (k) = 0.49; maternal N = 40, k = 0.58]. Six physically distant, inserted Wolbachia genes perfectly cosegregated in F2 maternal inheritance crosses (11), suggesting they also are closely linked.

PCR amplification and sequencing (11) of 45 Wolbachia loci in 14 D. ananassae lines from widely dispersed geographic locations revealed large Wolbachia inserts in lines from Hawaii, Malaysia, Indonesia, and India (table S2). Sequence comparisons of the amplicons from these four lines revealed that all open reading frames (ORFs) remained intact with >99.9% identity between inserts. This is compared to an average of 97.7% identity for the inserts compared with wMel, the Wolbachia endosymbiont of D. melanogaster. These results indicate the widespread prevalence of D. ananassae strains with similar inserts of the Wolbachia genome, probably because of a single insertion from a common ancestor.

In addition, reverse transcription PCR (RT-PCR) followed by sequencing (11) demonstrated that ∼2% of Wolbachia genes (28 of 1206 genes assayed; table S3) are transcribed in cured adult males and females of D. ananassae Hawaii. The complete 5′ sequence of one of the transcripts, WD_0336, was obtained with 5′–rapid amplification of cDNA ends (RACE) on uninfected flies (11), suggesting that this transcript has a 5′ mRNA cap, a form of eukaryotic posttranscriptional modification. Analysis of the transcript quantities of inserted Wolbachia genes with quantitative RT-PCR (qRT-PCR) (11) revealed that they are 104 times to 107 times less abundant than the fly's highly transcribed actin gene (act5C; table S3). There is no cutoff that defines a biologically relevant amount of transcription, and assessment of transcription in whole insects can obscure important tissue-specific transcription. Therefore, it is unclear whether these transcripts are biologically meaningful, and further work is needed to determine their importance.

Screening of public shotgun sequencing data sets has identified several additional cases of LGT in different invertebrate species. In Wolbachia-cured strains of the wasp Nasonia,six small Wolbachia inserts (<500 bp) were verified by PCR and sequencing (11) that have >96% nucleotide identity to native Wolbachia sequences, in some cases with short insertion site duplications. These include four in Nasonia vitripennis, one in N. giraulti, and one in N. longicornis (table S4 and Fig. 2B). Amplification and sequencing of 14 to 18 geographically diverse strains of each species indicated that the inserts are species-specific. For example, three Wolbachia inserts in N. vitripennis are not found in the closely related species N. giraulti or N. longicornis, which diversified ∼1 million years ago (12). These data suggest that the Wolbachia gene inserts are of relatively recent origin, similar to the inserts in D. ananassae.

Nematode genomes also contain inserted Wolbachia sequences. Because Wolbachia infection is required for fertility and development of the worm Brugia malayi, the genomes of both organisms were sequenced simultaneously, complicating assemblies and leading to the removal of Wolbachia reads during genome assembly [>98% identity over 90% of the read length on the basis of the independent BAC-based genome sequence of wBm, the Wolbachia endosymbiont of B. malayi (13)]. Despite this, the genome of B. malayi contains 249 contigs with Wolbachia sequences (e value < 10–40), nine of which were confirmed by long-range PCR and end-sequencing (11). These include eight large scaffolds containing >1-kb Wolbachia fragments within 8 kb of a B. malayi gene (table S5). Comparisons of wBm homologs to these regions suggested that all of these Wolbachia genes within the B. malayi genome are degenerate. In addition, a single region <1 kb was examined that contains a degenerate fragment of the Wolbachia aspartate aminotransferase gene (Wbm0002). Its location was confirmed by PCR and sequencing in B. malayi as well as in B. timori and B. pahangi (11).

Of the remaining 21 arthropod and nematode genomes in the trace repositories (11), we found six containing Wolbachia sequences. Potential Wolbachia-host LGT was detected in three: D. sechellia, D. simulans,and Culex pipiens (Table 1), as revealed by the presence of reads containing homology to both endosymbiont and host genomes (11).

Table 1.

Summary of Wolbachia sequences and evidence for LGT in public databases. Junctions were validated by PCR amplification and sequencing (11), with the number of successful reactions compared to the number attempted. Species marked with a plus sign are described in the literature as being infected with Wolbachia. All whole-genome shotgun sequencing reads were downloaded for 26 arthropod and nematode genomes (11). Organisms identified as lacking Wolbachia sequences either had no match or matches only to the prokaryotic ribosomal RNA. Because the Nasonia genomes are from antibiotic-cured insects, they were identified as having a putative LGT event merely on identification of Wolbachia sequences in a read. All other organisms were considered to have putative LGT events if the trace repository contained ≥1 read with (i) >80% nucleotide identity over 10% of the read to a characterized eukaryotic gene, (ii) >80% identity over 10% of the read to a Wolbachia gene, and (iii) manual review of the BLAST results for 1 to 20 reads to ensure significance (11). NA, not applicable.

View this table:

The sequencing of wBm also facilitated the discovery of a Wolbachia insertion in Dirofilaria immitis (dog heartworm). The D. immitis Dg2 chromosomal region encoding the D34 immunodominant antigen (14, 15) contains Wolbachia DNA within its introns and in the 5′ untranslated region (5′-UTR) (Fig. 2C). These Wolbachia genomic fragments have maintained synteny with the wBm genome (13), suggesting they may have inserted as a single unit and regions were replaced by exons of Dg2. A second gene (DgK) has been identified in other D. immitis lines that has 91% nucleotide identity in the exon sequences but contains differing number, position, size, and sequence of introns (16) and has no homology to known Wolbachia sequences.

Whole eukaryote genome sequencing projects routinely exclude bacterial sequences on the assumption that these represent contamination. For example, the publicly available assembly of D. ananassae does not include any of the Wolbachia sequences described here. Therefore, the argument that the lack of bacterial genes in these assembled genomes indicates that bacterial LGT does not occur is circular and invalid. Recent bacterial LGT to eukaryotic genomes will continue to be difficult to detect if bacterial sequences are routinely excluded from assemblies without experimental verification. And these LGT events will remain understudied despite their potential to provide novel gene functions and affect arthropod and nematode genome evolution. Because W. pipientis is among the most abundant intracellular bacteria (17, 18) and its hosts are among the most abundant animal phyla, the view that prokaryote-to-eukaryote transfers are uncommon and unimportant needs to be reevaluated.

Supporting Online Material

Materials and Methods

Fig. S1

Tables S1 to S5


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article