Ancestral State Reconstructions Trace Mitochondria But Not Phagocytosis to the Last Eukaryotic Common Ancestor

Abstract

Two main theories have been put forward to explain the origin of mitochondria in eukaryotes: phagotrophic engulfment (undigested food) and microbial symbiosis (physiological interactions). The two theories generate mutually exclusive predictions about the order in which mitochondria and phagocytosis arose. To discriminate the alternatives, we have employed ancestral state reconstructions (ASR) for phagocytosis as a trait, phagotrophy as a feeding habit, the presence of mitochondria, the presence of plastids, and the multinucleated organization across major eukaryotic lineages. To mitigate the bias introduced by assuming a particular eukaryotic phylogeny, we reconstructed the appearance of these traits across 1789 different rooted gene trees, each having species from opisthokonts, mycetozoa, hacrobia, excavate, archeplastida, and Stramenopiles, Alveolates and Rhizaria. The trees reflect conflicting relationships and different positions of the root. We employed a novel phylogenomic test that summarizes ASR across trees which reconstructs a last eukaryotic common ancestor that possessed mitochondria, was multinucleated, lacked plastids, and was non-phagotrophic as well as non-phagocytic. This indicates that both phagocytosis and phagotrophy arose subsequent to the origin of mitochondria, consistent with findings from comparative physiology. Furthermore, our ASRs uncovered multiple origins of phagocytosis and of phagotrophy across eukaryotes, indicating that, like wings in animals, these traits are useful but neither ancestral nor homologous across groups. The data indicate that mitochondria preceded the origin of phagocytosis, such that phagocytosis cannot have been the mechanism by which mitochondria were acquired.

Keywords: ancestral state reconstruction; eukaryogenesis; last eukaryote common ancestor; origin of mitochondria; phagocytosis; phagotrophy.

Fig. 1.

Presence (filled circle) absence (empty circle) distribution of five traits in 150 eukaryotic species. Species with no circle for a given trait indicate missing annotation. The reference tree was inferred from the alignment of 18S RNA sequences, rooted on the Excavates branch, with the sole purpose of data display (see Materials and Methods). Tip labels are species codes (see supplemental table 1, Supplementary Material online, for complete species names and detailed trait annotations). The first character of the species codes indicates supergroup affiliation of the species: Excavates (E), Mycetozoa (M), Hacrobia (H), Archaeplastida (A), SAR (S) and Opisthokonta (O). The shades of gray show the clades of the six eukaryotic supergroups.

Fig. 2.

Distribution of marginal probabilities for alternative trait-states in LECA across single-copy gene trees (left; without paralogs) and multi-copy gene trees (right; with paralogs). Multinucleate, phagocytosis and phagotrophy were treated as binary traits, while plastids and mitochondria were treated as traits with three states each. For plastids the states were: absence, primary plastid or secondary plastid. For mitochondriam the states were as follows: mitosome, hydrogenosome, or canonical mitochondria (see Methods and supplemental table 1, Supplementary Material online for details). The number of trees used in the analyses are show in table 1. Trait-states with high probabilities in the trees have distributions (colored lines) that are right-shifted in the plots.

Fig. 3.

Distribution of supergroups descending from origin nodes across 1789 trees. For each internal node reconstructed as a trait origin, all the species (tips) descending from it were used to score an origin to the combination of supergroups (filled circles) to which the descending species belong. Origins at the root node (LECA) are shown in red.