The natural history of molecular functions inferred from an extensive phylogenomic analysis of gene ontology data

Abstract

The origin and natural history of molecular functions hold the key to the emergence of cellular organization and modern biochemistry. Here we use a genomic census of Gene Ontology (GO) terms to reconstruct phylogenies at the three highest (1, 2 and 3) and the lowest (terminal) levels of the hierarchy of molecular functions, which reflect the broadest and the most specific GO definitions, respectively. These phylogenies define evolutionary timelines of functional innovation. We analyzed 249 free-living organisms comprising the three superkingdoms of life, Archaea, Bacteria, and Eukarya. Phylogenies indicate catalytic, binding and transport functions were the oldest, suggesting a ‘metabolism-first’ origin scenario for biochemistry. Metabolism made use of increasingly complicated organic chemistry. Primordial features of ancient molecular functions and functional recruitments were further distilled by studying the oldest child terms of the oldest level 1 GO definitions. Network analyses showed the existence of an hourglass pattern of enzyme recruitment in the molecular functions of the directed acyclic graph of molecular functions. Older high-level molecular functions were thoroughly recruited at younger lower levels, while very young high-level functions were used throughout the timeline. This pattern repeated in every one of the three mappings, which gave a criss-cross pattern. The timelines and their mappings were remarkable. They revealed the progressive evolutionary development of functional toolkits, starting with the early rise of metabolic activities, followed chronologically by the rise of macromolecular biosynthesis, the establishment of controlled interactions with the environment and self, adaptation to oxygen, and enzyme coordinated regulation, and ending with the rise of structural and cellular complexity. This historical account holds important clues for dissection of the emergence of biomcomplexity and life.

Archaea‐First and the Co‐Evolutionary Diversification of Domains of Life

“…These phases dissected patterns of FF sharing among Archaea, Bacteria, and Eukarya unfolding along the evolutionary timeline of FF domain innovation (Figure C). Similar patterns and phases were obtained from an analysis of CATH topologies and homologous superfamily domain structures and an analysis of molecular functions at different levels of GO database classification …”

“…Molecular features include the structure of highly conserved RNA molecules such as tRNA, 5S rRNA, and RNase P RNA, proteins structural domains in proteomes . GO definitions of molecular functions, and many additional lines of molecular evidence . Cellular features include the ether‐linked lipid membranes of Archaea, which markedly differ from the ester‐linked lipid membranes that unify Bacteria and Eukarya and makes a membrane fusion scenario or the independent gain of a same membrane composition unlikely because membranes are an ancient essential requirement for all cells.…”

Giant viruses: spore‐like missing links between Rickettsia and mitochondria?

“…16,17 Such parsimony-based analyses reconstruct very ancient phylogeny from shared structures 18 and/or functions. [19][20][21][22][23] Shared synteny to evaluate evolutionary relatedness This method using the presence/absence of shared structures is a viable alternative to the otherwise doi: 10.1111/nyas.14022 powerful alignment−homology methodology most phyletic reconstructions use nowadays. 24,25 It applies to molecular properties the methods that classical naturalists use for phyletic reconstructions based on (frequently morphological) character states (for an example for psammophile lizards, see Ref.…”

“…The most comprehensive analysis giving support to the hypothesis that viruses have a monophyletic cellular origin comes from deep phylogenetic reconstructions based on comparisons among protein fold families, which are not confounded by artifacts due to phylogenetic reconstruction methods . Such parsimony‐based analyses reconstruct very ancient phylogeny from shared structures and/or functions …”

The tree of life describes a tripartite cellular world

“…[34] Similarly, alignment-free trees of molecular functions and trees of functionomes (entire set of molecular functions of an organism) have been derived from a genomic census of GO terms defined by the GO Consortium at different levels of ontological classification. [47][48][49] All of these analyses have consistently recovered 3D trees no matter the choice of characters (e.g., superfamilies, [45] families, [46] GO functions [47][48][49] ), the size of datasets, [50,51] or the classification scheme used for annotating protein domains. [52,53] Such robustness over a wide range of alignmentfree reconstructions is remarkable.…”