Phylogenetic analyses of cyanobacterial genomes: quantification of horizontal gene transfer events

Abstract

Using 1128 protein-coding gene families from 11 completely sequenced cyanobacterial genomes, we attempt to quantify horizontal gene transfer events within cyanobacteria, as well as between cyanobacteria and other phyla. A novel method of detecting and enumerating potential horizontal gene transfer events within a group of organisms based on analyses of "embedded quartets" allows us to identify phylogenetic signal consistent with a plurality of gene families, as well as to delineate cases of conflict to the plurality signal, which include horizontally transferred genes. To infer horizontal gene transfer events between cyanobacteria and other phyla, we added homologs from 168 available genomes. We screened phylogenetic trees reconstructed for each of these extended gene families for highly supported monophyly of cyanobacteria (or lack of it). Cyanobacterial genomes reveal a complex evolutionary history, which cannot be represented by a single strictly bifurcating tree for all genes or even most genes, although a single completely resolved phylogeny was recovered from the quartets' plurality signals. We find more conflicts within cyanobacteria than between cyanobacteria and other phyla. We also find that genes from all functional categories are subject to transfer. However, in interphylum as compared to intraphylum transfers, the proportion of metabolic (operational) gene transfers increases, while the proportion of informational gene transfers decreases.

Figure 1.

Illustration of an embedded quartet in a gene tree. In each 11-taxon unrooted gene tree (shown in gray thin lines), we look at the relationship of any four taxa at a time (shown in thick black lines is an example of an embedded tree for taxa 1, 4, 9, and 10), which we call an embedded quartet. In one gene tree, an embedded quartet can support only one of three possible phylogenetic relationships among the four taxa. However, in other gene trees, an alternative phylogenetic relationship for this quartet can be observed. In each 11-taxon gene tree, all possible (₄ ¹¹) four-taxon trees embedded within the gene tree are examined.

Figure 2.

Quartet decomposition analysis of cyanobacteria. Panel A illustrates a component of quartet decomposition analysis. Each embedded quartet is represented by a vertical bar and a black dot. The black dot indicates how many data sets contain this embedded quartet. The vertical bar shows the number of data sets having the topology of the quartet that is supported by a plurality of gene families (value above zero) and the number of data sets having one of the other two quartet topologies (value below zero). The bar is color-coded with respect to bootstrap support. Only quartets that resolve the quartet relationships in at least 30% of analyzed data sets are visualized. The value of 30% was chosen based on simulations (see text for details). Panel B shows the quartet spectrum of 1128 sets of orthologous genes from cyanobacteria. Columns are sorted according to the number of supporting data sets with at least 80% bootstrap support. Quartets with a very short internal branch or very long external branches were excluded from the analyses to minimize artifacts of phylogenetic reconstruction. Quartets above the x-axis are combined into a plurality signal (see Fig. 3 and text for more details). Quartets below the x-axis are embedded into 685 unique sets of orthologous genes. The appearance of discontinuities in the spectrogram corresponds to the uneven phylogenetic distances among 11 analyzed genomes. For example, quartets containing two very closely related taxa (such as the ones containing both Nostoc and Anabaena) will almost unanimously agree on one of three possible quartet topologies.

Figure 3.

Visualization of the evolutionary history of cyanobacteria as inferred from quartet decomposition analyses. The unrooted tree topology was calculated from the embedded quartets supported by the plurality of sets of orthologous genes, and it is shown in black. Conflicts to the plurality topology that involve two particular taxa grouping together with at least 80% bootstrap support are shown as gray lines connecting two taxa in conflict. The number of conflicting data sets that have the two connected taxa grouping together are shown on the corresponding gray line. Note that the number of conflicting sets shown represents only a small subset of observed conflicts to the plurality signal, since not all topological incongruencies are of this kind. The genes involved in conflicts, as well as corresponding topologies, are provided as Supplemental material. (A) Anabaena sp. PCC7120; (Tr) Trichodesmium erythraeum IMS101; (S) Synechocystis sp. PCC6803; (1P) Prochlorococcus marinus CCMP1375; (2P) Prochlorococcus marinus MED4; (3P) Prochlorococcus marinus MIT9313; (mS) marine Synechococcus WH8102; (Th) Thermosynechococcus elongatus BP-1; (G) Gloeobacter violaceus PCC7421; (N) Nostoc punctiforme ATCC29133; (C) Crocosphaera watsonii WH8501.

Figure 4.

Example of intraphylum transfer: a hemolysin-like protein. This example of horizontal gene transfer was extracted from the list of data sets exhibiting conflicts with the plurality signal. This gene family has detectable homologs in other phyla, but phylogenetic analysis of an extended data set shows that cyanobacteria form a coherent phylogenetic group. Interestingly, in cyanobacteria this protein seems to have acquired a different function (Nagai et al. 2001). This suggests that this protein is probably not being exchanged with the organisms outside of cyanobacteria, which makes it a good example of intraphylum transfer. In this phylogeny Thermosynechococcus and Crocosphaera are sister taxa, which contradicts the relationships observed in the plurality tree. The arrow indicates placement of the Thermosynechococcus sequences in the consensus tree. The tree topology and branch lengths were calculated in PhyML (Guindon and Gascuel 2003) under the JTT+Gamma model. The bootstrap support values are from TREE-PUZZLE + NEIGHBOR phylogenetic analyses as described in Methods. Only bootstrap support values above 80% are shown. Additional examples of putative intraphylum transfers are available in the Supplemental material.

Figure 5.

Example of horizontal gene transfer to cyanobacteria: threonyl tRNA synthetase. This is a phylogenetic tree reconstructed from a data set in which the Anabaena sp. genome did not have a detectable homolog in its annotation. In this tree, sequences of three Prochlorococcus spp. group within Gamma-proteobacteria with high bootstrap support. The presence of an ancestrally transferred gene constitutes a shared derived character for the descendents (Andersson et al. 2005; Huang et al. 2005), suggesting that the three Prochlorococcus spp. form a monophyletic group. Note that despite coherency of the remaining cyanobacteria as a group, the relationships within cyanobacteria do not follow the plurality topology. tRNA synthetases are known to have complex evolutionary histories involving multiple HGT events (e.g., Wolf et al. 1999). Cyano-bacteria are shown in bold. To obtain bootstrap support values, a consensus tree was generated from 100 bootstrap sampled trees (see Methods for more details). Only bootstrap support values above 80% are shown. Additional examples of interphylum transfers are available in the Supplemental material.

Figure 6.

Distribution of cyanobacterial sets of orthologous genes across functional categories. The functional categories are according to the COG database, March 2003 release (Tatusov et al. 2003). Panel A shows the distribution of all 1128 analyzed genes, while panel B shows the distribution of 685 genes that conflict with the plurality signal. Conflicts with the plurality signal are observed in sets of orthologs across all functional categories, including genes involved in translation and transcription, and no particular functional category shows bias toward conflicts.

Figure 7.

Distribution of phylogenetically useful extended genes across functional categories. Panel A shows the distribution of 700 phylogenetically useful extended data sets. Panel B shows the distribution of 160 sets where cyanobacteria do not form a monophyletic group (putative interphylum transfers). Panel C shows the distribution of 294 sets where cyanobacteria form a monophyletic group, but they conflict with the plurality signal (putative intraphylum transfers). Numbers inside the pie graphs refer to the number of sets of orthologous genes in each corresponding functional category. Across short phylogenetic distances, all types of genes appear to be equally affected by transfer, while across large phylogenetic distances, genes encoding metabolic functions are more frequently transferred, and genes in transcription and translation are transferred less frequently.