Abstract Detail


Lawrence, Travis Joseph [1], Ardell, David H [2].

tRNA Interaction Network Sheds Light on the Origin of Chloroplast.

More than one billion years ago the acquisition of a cyanobacterial endosymbiont led to the development of the eukaryotic chloroplast. Like the mitochondrion before it, the chloroplast profoundly changed the tree of life and, through the generation of oxygen during photosynthesis, our planet's entire ecosystem. Though a cyanobacterial plastid origin was proposed more than a century ago the precise point of plastid origin within the cyanobacterial tree (i.e. the identity of the cyanobacterial progenitor) is one of the last great questions regarding this landmark evolutionary event. Illuminating the evolutionary origin of the plastid progenitor is complicated by at least a billion years of vertical descent, including massive plastid genome reduction and a number of secondary and tertiary endosymbiosis events. Previous attempts to resolve the origin of plastids have arrived at conflicting results finding support for both a deep and a more recent origin of plastids. Here we present a novel approach for delving into the deepest evolutionary history of plastid origin. This method takes advantage of the unique constraints on the evolution of the tRNA-protein interaction network. Despite the very high structural similarity of all tRNAs each must interact productively with only one type of aminoacyl-tRNA synthetase (aaRS) to be charged with its cognate amino acid and must avoid interacting productively with others. The class (i.e. charging specificity) of a tRNA relies on a set of structural features called determinants that promote recognition by its cognate aaRS. These sets of structural features that determine a tRNA’s class, which we refer to as class-informative features (CIFs), are not static and vary widely across the tree of life. We have developed an analysis pipeline to bioinformatically identify tRNA CIFs from genomic sequence data and then use a novel neural network classifier to assign plastid genomes to known cyanobacterial clades. Using 113 cyanobacterial genomes we obtained 5,476 tRNAs, which were used to identify CIFs for cyanobacterial clades and train our neural network classifier. Using 14,319 plastid tRNAs from 423 plastid genomes we obtained a strong signal classifying plastids to a clade containing filamentous and/or N2-fixing cyanobacterial species that is within a more recent diversification of cyanobacteria. Our results suggest a relatively recent origin of the plastid lineage likely with close relatives among extant cyanobacteria and not a survivor of an extinct lineage.

1 - University of California, Merced, Quantitative and Systems Biology Graduate Group, 5200 North Lake Rd., Merced, CA, 95348, USA
2 - University of California, Merced, Molecular Cell Biology Unit, School of Natural Sciences, 5200 North Lake Rd., Merced, CA, 95343, United States

Chloroplast Progenitor
Machine Learning

Presentation Type: Oral Paper
Session: 29, Phylogenomics I
Location: Fort Worth Ballroom 4/Omni Hotel
Date: Tuesday, June 27th, 2017
Time: 1:45 PM
Number: 29002
Abstract ID:50
Candidate for Awards:None

Copyright © 2000-2017, Botanical Society of America. All rights reserved