Lateral gene transfers and the origins of the eukaryote proteome: a view from microbial parasites

Abstract

Our knowledge of the extent and functional impact of lateral gene transfer (LGT) from prokaryotes to eukaryotes, outside of endosymbiosis, is still rather limited. Here we review the recent literature, focusing mainly on microbial parasites, indicating that LGT from diverse prokaryotes has played a significant role in the evolution of a number of lineages, and by extension throughout eukaryotic evolution. As might be expected, taxonomic biases for donor prokaryotes indicate that shared habitat is a major factor driving transfers. The LGTs identified predominantly affect enzymes from metabolic pathways, but over a third of LGT are genes for putative proteins of unknown function. Finally, we discuss the difficulties in analysing LGT among eukaryotes and suggest that high-throughput methodologies integrating different approaches are needed to achieve a more global understanding of the importance of LGT in eukaryotic evolution.

Figure 1

Candidate LGT among parasitic microbial eukaryotes. (a) Taxonomy of donor lineages for candidate LGTs. Comparison of the prokaryotic lineages inferred to be donating genes to the extracellular mucosal parasites Entamoeba histolytica, Trichomonas vaginalis, and Giardia lamblia compared with the inferred donor lineages for the insect-transmitted blood parasites Trypanosoma brucei, T. cruzi, Plasmodium falciparum, P. vivax and P. yoelii yoelii (top panel). Comparison of the prokaryotic lineages inferred to be donating genes to the parasite E. histolytica and its free-living amoebozoan relative D. discoideum (bottom panel). ‘Other bacteria’ comprise the Actinobacteria, Aquificae, Fusobacteria, Plantomycetes, Spirochaetes, or Tenericutes. Fisher's exact test was performed to test the null hypothesis that the taxonomy of the donors is distributed equally between the compared taxa. The P-values for the tests are indicated; they both reject the null hypothesis. The numbers of LGTs considered for each set of taxa are indicated between brackets. (b) Assessment of gains and losses of lateral gene transfer (LGTs) during parasite speciation. Maximum parsimony was used to map candidate LGTs on the species trees for taxa among (a) Trypanosomatidae and (b) Apicomplexa. Gains and losses are indicated as green and orange bars respectively. Characters were analysed using Dollo parsimony, so each character is allowed to have only a single gain, but may have multiple losses. It is inferred that 45 LGTs occurred (over 75 genes affected by LGT) before the divergence of the three parasitic Trypanosomatidae lineages. Interestingly, we detected 26 of the same LGTs in the genome of the free-living kinetoplastid Bodo saltans [45] using Blast similarity scores, suggesting these transfers may predate the transition to parasitism. Figures in panel (a) and (b) are derived from [12^•]. (c) The mapping of annotated genes (red and blue genes indicate the differential orientation of the inferred open reading frames) on the scaffold DS113827 from the genome sequence data of T. vaginalis strain G3 [46]. A 32 kbp fragment (orange bar) was shown to be highly similar to the Firmicutes Peptoniphilus harei and encode 27 annotated genes. A matching gene cluster was found in all four additional investigated strains of the parasite [24]. Entries labelled with RG in their locus tags correspond to highly repetitive gene families, which are known to litter the genome of T. vaginalis [46]. The figure in panel c was generated using TrichDB [47].

Figure 2

Evidence for transcription of LGT in Trichomonas vaginalis. The level of transcription of different gene sets was contrasted through histograms with 11 bins reflecting no read (bin 1, zero mean 3′-normalised reads across 11 conditions in [28]) to the highest level of transcription (bin 11, >100,000 to 300,000 mean reads). The inset in the top panel shows all bins with different levels of transcription expressed as the mean of 3-‘normalised reads across 11 tested growth conditions. The top panel illustrates the variation in transcription level between all annotated protein coding genes from the T. vaginalis G3 genome sequence data [46]. The middle panel contrasts protein-coding genes annotated as ‘hypotheticals’ (blue bars, with no BlastP hits in databases) versus ‘hypothetical conserved’ (green bars, with BlastP hits in databases). The bottom panel indicates the level of transcription of all LGT cases identified in [12^•]. The LGT genes are notably skewed towards the right hand side of the histogram (higher level of transcription) compared to ‘hypotheticals’. This suggests that the majority of LGTs are likely to be functionally integrated into the biology of the parasite whereas the great majority of ‘hypotheticals’ are not and might represent pseudogenes or miss-annotations of spurious genes.