Eukaryotes are clearly different from organisms in the bacterial domains.  They have internal membrane-bound structures (including the nucleus, the minimal requirement for a eukaryote), microtubular cytoskeletons, and characteristic ribosomes.  In addition, Roger (1999) summarizes synapomorphic characters that all Eukaryotes possess or once possessed and have lost through simplification (see Table 1).

Direct comparisons with bacterial domains are difficult.  Compared to the bacteria (prokaryotes), the eukaryotes are chimeroid living structures.  They have evolved through a variety of non-Darwinian means: endosymbiosis and fusion of cells (fusion of bacteria with nuclear host and the fusion of eukaryotes), to lateral gene transfer (Katz 1999).  The resulting changes from such a mode of construction would have led to large steps or apparent saltations separating the resulting taxa, the largest of which is the transition from bacteria (likely an ancestor to the present day Archaea) to the Eukarya (or Eukaryota).  Various theories have been proposed to account for the origin of eukaryotes.  Perhaps, the most generally accepted one is the Serial Endosymbiosis Theory (SET) of Taylor (1974).  SET indicates that eukayotes arose relatively quickly from bacterial ancestors as they came together in symbiotic communities, and their communal DNA coalesced as linear chromosomes within an internal membrane-bound nucleus.  Organelles like the mitochondrion (alpha proteobacterium) and chloroplast (cyanobacterium) retained their bacterial identity.  Other eukaryotic structures are more problematic.  Margulis (1990) modified SET to include a spirochaete symbiosis that gave rise to flagella (undulapodia), basal bodies, centrioles, and the cytoskeleton.  This modified SET is still hotly debated.  Even without the Margulis modification to SET, the origin of the eukaryotes is difficult to infer from current microbial eukaryotes.  

The Archaezoa Hypothesis of Cavalier-Smith (1983) suggested that the eukaryotes appeared prior to the endosymbiotic events that gave rise to mitochondrial eukaryotes.  According to Cavalier-Smith an archeal bacterium gave rise to a eukaryote through the elaboration of a cytoskeleton and the development of an internal membrane system that included the formation of a nucleus.  The flagella in his system evolved as external elaborations of the cytoskeleton.  Thus, in its original form, the Archaezoa Hypothesis implied that the Microsporidia, Metamonada, Parabasalia, and Archamoebae were extant members of the Archaezoa [see Figure 1 from Roger 1999].  The theory developed with what appeared to be a string of confirmations.  Most notable was the basal and deep branching positions of amitochondriate motile taxa like Giardia on molecular phylogenetic trees.  The development of the Archezoa Theory could be followed through a whole string of publications by Cavalier-Smith (1983, 1987a, 1987b, 1989, 1990, 1991a, 1991b, 1992, 1993, 1995, 1996-1997, 1998, 1999) and Cavalier-Smith and Chao (1995, 1996).  The group seemed to be confirmed such that the eukaryotes could be defined by mitochondrial lines that emerged from a premitochondrial archezoan ancestor [see Figure 2 from Gray et al. 1998].  This organization can be seen in Margulis and Schwartz (1999).  Roger (1999) gave a nice history of the Archezoa Hypothesis and ended with some uncomfortable details that led to the collapse of the theory.  For example, the apparent  primitive nature of the archezoa was an artifact of molecular trees and of the the parasitic/symbiotic simplification that occurred in the amitochondrial taxa.  Also, certain genetic sequences that must have come from the mitochondrion could be found in all of the archezoa.  Furthermore, taxa like the parabasalids have hydrogenosomes, anaerobic organelles that seem to have evolved from mitochondria.  If all extant eukaryotes have mitochondrial genes in their nuclear DNA, the most parsimonious solution would be that all living eukaryotes evolved from mitochondrial ancestors.  Thus, by 1999 the Archezoa Hypothesis was in trouble, and by 2002, it had died.  Keeling (2002) finally wrote the eulogy for the Archezoa Hypothesis.

Taylor (1999) insisted that molecular inference must be confirmed by phenotypic evidence.  He considered the proposal of Gray et al. (1998) in which eukaryotes were marked by 3 great lines of mitochondrial descent (as defined by mtDNA sequences, see Figure 2).  Taylor showed how the Gray et al. hypothesis was supported by ultrastructural evidence (see Figure 3).  This view also suggested that the mitochondrial symbiosis was one of the most fundamental steps in the appearance of the eukaryote lineages that are defined by mitochondria with discoid cristae, flattened cristae, and tubular cristae.

This view is both beautifully simple and simply too good to be true.  The mitochondrial monophyly theory is not consistent with the diversity of mitochondrial types in multicellular lineages.  For example, metazoans have both flattened and tubular cristate mitochondria depending on the tissues involved.  How can this be reconciled with the hypothesis of Gray et al. (1998)?  

The state of eukaryotic taxonomy at the higher taxonomic levels (particularly kingdoms) mirrored the changes in phylogeny.  The taxonomy went from the appearance of stability at four in the five kingdom system of Whittaker and Margulis (1978) to absolute chaos when molecular phylogenies began to be used in the decade of the 1990's.  Patterson (1999) demonstrated this problem in a dramatic way when he published more than 71 taxa (and taxonomic groups) without clear sister groups.   The chaos could be seen in the book by Tudge (2000) in which he defined many kingdoms of eukaryotes.  Some were based on single taxa like Giardia or Entamoeba.

Resolution to the taxonomic cacophony began to come quickly in the early years of the 21st century by use of molecular-structural data.  Baldauf (2003) reviewed the state of the literature relative to the Eukaryotic Tree and identified eight major lines or supergroups of eukaryotes.  Her analysis, grounded both in molecular and morphological taxonomy (see Figure 4), confirmed that the Archezoa Hypothesis in its early forms (Gray et al. 1998 and Keeling 1998) likely was dead.  In particular, she eliminated the concept of the "crown eukaryotes" as an artifact of single-gene comparisons.  The summary analysis of Keeling (2004) simplifies the supergroups to 5 (see Figure 5), a position that I have adopted.  The supergroup results helped to resolve the chaos as illustrated by Patterson (1999), and, at the same time, take into account the concerns of Taylor (1999) by identifying  the defining synapomorphies of those groups.  Stechmann and Cavalier-Smith (2004) have attempted to resurrect the Archaezoa Hypothesis in a new form by providing evidence that the eukaryotes and the archaea evolved from a Gram + bacterium like the Actinomycetes.  This will not be the last word in the argument about the Eukaryotic "Tree"; however, the approach of multiple ultrastructural characters coupled with and confirmed by molecular data is the only solution to untying the Gordian Knot of Eukaryotic lineages.

For further information, go to the Evolution of the Eukaryotic Supertree.

References

This page is maintained by Jack R. Holt. Last revised: 03/10/2008.