Abstract 27 2.1 The Search for Primitive Eukaryotes 28 2.2 The Archezoa Hypothesis 29 2.3 Candidate Archezoa as Deep Branches 29 2.4 Challenges to the Archezoa Hypothesis 30

2.4.1 Genes of Mitochondrial Origin 30 2.4.2 Are the Deep Branches Truly Deep? 30

2.5 Excavata — A Home for Many Reformed Archezoa? 31 2.6 Morphological Evidence for Excavata 32 2.7 Molecular Phylogenies and Excavata 33 2.8 Relationships among Excavata 37

2.8.1Eopharyngia — Diplomonads and Retortamonads 37 2.8.2 Carpediemonas 38 2.8.3 Parabasalids 38 2.8.4 Preaxostyla — Oxymonads and Trimastix 38 2.8.5 Malawimonas 38 2.8.6 Heterolobosea, Euglenozoa and Jakobids 39 2.8.7 Relationships in Summary 39

2.9 Excavates as Early Eukaryotes? 39 2.9.1 Diplomonads and Parabasalids Early? 40 2.9.2 Jakobids Early? 41

2.10 Alternatives to Mitochondria in Excavates 42 2.11 Excavate Phylogeny and Mitochondrial Loss — Neoarchezoa 44 Acknowledgments 45 References 45

Ever since the distinction between prokaryotic and eukaryotic cells became clear (Stanier, 1970) there has been great interest in understanding the early history of the eukaryotic cell and the nature of the first eukaryotes. For much of the past three decades, attention has been focused on mitochondria as a key (and comparatively late) acquisition by the eukaryotic

lineage, and on mitochondrion-lacking groups as potentially primitive or deep-branching eukaryotes. This paradigm reached its zenith in the early 1990s with the small subunit ribosomal rRNA (SSUrRNA) supported version of the archezoa hypothesis. Over the past decade, better estimates of the relationships among eukaryotes and independent elucidation of the probable mitochondrial ancestry of many mitochondrion-lacking groups have forced a reevaluation of the history of mitochondrial mode in eukaryotes. The classic mitochondrion-lacking lineages originally considered to be archezoa probably fall within three major clades: opisthokonts, Amoebozoa and Excavata. Excavata includes diplomonads, retortamonads, oxymonads and parabasalids from the original archezoa, as well as two new amitochondriate taxa Carpediemonas and Trimastix, and several mitochondrion-bearing groups, namely jakobids, Malawimonas, Heterolobosea and Euglenozoa. Most excavates share a distinctive feeding groove, which identifies the group. Five groups — retortamonads, Carpediemonas, Trimastix, jakobids and Malawimonas — share several discrete cytoskeletal characters unique to excavates, strongly suggesting that they have descended from a similar common ancestor. Molecular phylogenies (SSUrRNA or multiple protein analyses, or both) indicate specific relationships between each of diplomonads, oxymonads, parabasalids, Heterolobosea and Euglenozoa, and at least one of the first five taxa. However, the same molecular phylogenies generally do not support the monophyly of Excavata as a whole. There are three possible scenarios: (1) Excavata is a clade and available molecular phylogenies are incorrect, (2) the excavate morphology has arisen convergently several times or (3) the excavates are actually the plesiomorphic stem group for most or all living eukaryotes. We favor the first of these scenarios, noting that many excavates have unusually divergent gene sequences that might be expected to foster artifacts in molecular phylogenetic analyses. Regardless, the complete resolution of the relationships among various Excavata should provide the phylogenetic basis for resolving many outstanding questions over the evolution of the amitochondriate state and perhaps the nature of early eukaryotic cells.