Phylogenomic fingerprinting of tempo and functions of horizontal gene transfer within ochrophytes

Abstract

Horizontal gene transfer (HGT) is an important source of novelty in eukaryotic genomes. This is particularly true for the ochrophytes, a diverse and important group of algae. Previous studies have shown that ochrophytes possess a mosaic of genes derived from bacteria and eukaryotic algae, acquired through chloroplast endosymbiosis and from HGTs, although understanding of the time points and mechanisms underpinning these transfers has been restricted by the depth of taxonomic sampling possible. We harness an expanded set of ochrophyte sequence libraries, alongside automated and manual phylogenetic annotation, in silico modeling, and experimental techniques, to assess the frequency and functions of HGT across this lineage. Through manual annotation of thousands of single-gene trees, we identify continuous bacterial HGT as the predominant source of recently arrived genes in the model diatom Phaeodactylum tricornutum Using a large-scale automated dataset, a multigene ochrophyte reference tree, and mathematical reconciliation of gene trees, we note a probable elevation of bacterial HGTs at foundational points in diatom evolution, following their divergence from other ochrophytes. Finally, we demonstrate that throughout ochrophyte evolutionary history, bacterial HGTs have been enriched in genes encoding secreted proteins. Our study provides insights into the sources and frequency of HGTs, and functional contributions that HGT has made to algal evolution.

Keywords: MMETSP; RNAseq; ornithine-urea cycle; phylogenomics; stramenopile.

Fig. 1.

Dynamics and identity of horizontally transferred genes in the P. tricornutum genome. (A) Evolutionary origins of 1,347 genes in the version 3 annotation of the P. tricornutum genome annotated by manual inspection of phylogenies as having been transferred into ochrophytes from nonstramenopile sources, and 1,771 gene transfers identified out of ochrophytes into nonstramenopile groups. Frequencies are shown for the most abundant contributors to each HGT category; full values are tabulated in Dataset S1. (B) Number of HGTs involving different groups of eukaryotic algae (shown on the left vertical axis) and bacteria (right axis) and ochrophytes, manually resolved to (left four) four time points in diatom evolution, and (right two) two earlier points in ochrophyte evolution.

Fig. 2.

Distribution of 2,796 bacterial HGTs automatically identified across ochrophyte genomes. This figure shows the consensus MrBayes (GTR, WAG) and RAxML (GTR, JTT, WAG) topology of a 213 taxa by 26,399-aa concatenated alignment, consisting of 63 genes with high occupancy in 161 ochrophyte genomes and transcriptomes, rooted on a set of 52 nonochrophyte outgroups (collapsed here to show only aplastidic stramenopile groups). Leaf names are shaded by evolutionary origin. Support values for each node are shown with shaded circles. The total number of bacterial HGTs attributed to each node by MRCA analysis is depicted by branch color and thickness; four branches with >100 HGTs are individually labeled. An analogous tree topology, showing only the frequencies of HGTs reconciled with >70% bootstrap support, is provided in SI Appendix, Fig. S18.

Fig. 3.

ALE inference of HGT origins in ochrophytes. (A) Distribution of the total number of duplications, losses, speciations, and intraochrophyte transfers averaged over the set of 100 ALE reconciliations for each of the 435 HGT protein clusters inferred as single bacterial HGT origins. (B) Distribution of the number of HGT origins per branch inferred by ALE for branches within diatoms (orange) or outside diatoms averaged over the 100 reconciliations. The difference (t test) between the distributions is reported on the top of the plot. (C) Ratio of the number of HGT origins inferred within diatoms in the original data (red) compared to the same ratio observed when the data are randomized (gray, null expectations obtained by shuffling the species names).

Fig. 4.

Continuous secretory enrichment of bacterial HGTs in ochrophytes. (A) Heatmap showing the percentage of all ochrophyte genes in the combined 162 species genome and transcriptome library, all genes in the 2,786 identified bacterial HGT clusters, and all HGT genes that originate at different time points in ochrophyte evolution that possess different targeting predictions. Cells are shaded by the significance of enrichment, inferred by (two-tailed) χ² tests against the total number of genes in species contained within the corresponding node. (B) GFP-fluorescence constructs of four exemplar genes in the P. tricornutum genome with endomembrane or secretory localizations and bacterial origins: Phatr3_J33371.p1 and Phatr3_J42871.p1, with inferred endoplasmic reticulum localizations (verified by ER-Tracker); and Phatr3_J50959.p1 and Phatr3_J85969.p1, proteins with partial plasma membrane or cell wall localizations. (Scale bars, 10 μm.)