Abril_10_2012_ Evaluating Hypotheses for the Origin of Eukaryotes (1)

8/13/2019 Abril_10_2012_ Evaluating Hypotheses for the Origin of Eukaryotes (1)

1/11

Evaluating hypothesesfor the origin of eukaryotesAnthony M. Poole 1 * and David Penny 2

SummaryNumerous scenarios explain the origin of the eukaryotecell by fusion or endosymbiosis between an archaeonand a bacterium (and sometimes a third partner). Weevaluate these hypotheses using the following threecriteria. Can the data be explained by the null hypothesisthatnew featuresarise sequentiallyalong a stemlineage?Second, hypotheses involving an archaeon and a bacter-ium should undergo standard phylogenetic tests of genedistribution. Third, accounting for past events by pro-cesses observed in modern cells is preferable topostulating unknown processes that have never beenobserved. For example, there are many eukaryote exam-ples of bacteria as endosymbionts or endoparasites, butnoneknown in archaea.Strictlypost-hochypotheses thatignore this third criterion should be avoided. Applyingthese three criteria significantly narrows the number ofplausible hypotheses. Given current knowledge, ourconclusion is that the eukaryote lineage must havediverged from an ancestor of archaea well prior to theorigin of the mitochondrion. Significantly, the absence ofancestrally amitochondriate eukaryotes (archezoa) amongextant eukaryotes is neither evidencefor an archaealhostfor the ancestor of mitochondria, nor evidence against aeukaryotic host. BioEssays 29:7484, 2007.

2006 Wiley Periodicals, Inc.

IntroductionIt is now accepted that all known modern eukaryotes evolvedfrom a mitochondrion-bearing ancestor; (14) that is, there areno known living eukaryotes that never possessed a mitochon-drion (archezoa). Consequently, it is unclear whether othereukaryote-specific features such as the nucleus, endomem-brane system, mRNA splicing and linear chromosomespredate or postdate the origin of the mitochondrion. Genomic

comparisons reveal that eukaryote genes can be divided intoseveral classes: genes apparently specific to eukaryotes,genes that appear from gene trees to be most closely relatedto a specific bacterial group, genes with greatest sequencesimilarity to sequences in bacterial lineages, and genes mostsimilar to sequences in archaeal lineages. (512) Numeroushypotheses attempt to account for these patterns butdisagreeabout the nature of the host and the number of partnersinvolved in the origin of the eukaryote cell. Here we considerfour general models (Fig. 1). These are not exhaustive, but

instead aim to generalise the main features of a large numberof models (see Martin et al., (13) and Embley and Martin (3) fordetailed reviews of specific models).

(i) Fusion, where one partner is an archaeon, the otheris a bacterium. Fusion implies physical fusion of twocells, creating a single new cellular compartment bythe mixing of the cell contents of the two partners. Thisis separate to the origin of the mitochondrion (Fig. 1,panel i ).

(ii) Endosymbiosis, where the nucleus evolves directly froman engulfed archaeal or other type of cell. The nature ofthe engulfing cell likewise varies, but is never a eukaryoticcell (Fig. 1, panel ii ).

(iii) Endosymbiosis, where an archaeal cell takes up theancestor to the mitochondrion (Fig. 1, panel iii ).

(iv) Endosymbiosis, where a protoeukaryotic cell engulfs theancestor of the mitochondrion (Fig. 1, panel iv ).

There are several distinctions between the hypotheses.In contrast to model i , models iiiv specify a known process(endosymbiosis) wherein one lineage incorporates a secondcell type. Models i and ii require three partners: model i requires a fusionto explain theorigin of theeukaryotecell and,subsequently, endosymbiosis for the origin of mitochondria;model ii requires twoendosymbioses, one for the nucleus,onefor the mitochondrion. Models iii and iv , however, require only

two partners, a host cell which engulfs the ancestor of themitochondrion. Model iv , incontrastto models iiii , implies thatthe basic eukaryote cell had already evolved, and that theorigin of the mitochondrion may have been one of the finalsteps in the evolution of modern eukaryotes.

Suggestions that the nucleus itself is an endosymbiont(ii )(7,14) have been firmly rejected (15,16) and, likewise, thereis no strong genetic evidence for three-way fusions (i and

1Department of Molecular Biology and Functional Genomics, Stock-holm University, Sweden.2Allan Wilson Centre for Molecular Ecology & Evolution, Institute ofMolecular BioSciences, Massey University, Palmerston Nor th, NewZealand.Funding agency: A.M.P is supported by grants from the SwedishResearch Council and the Carl Tryggers Foundation.*Correspondence to: Anthony M. Poole, Department of MolecularBiology and Functional Genomics, Stockholm University, SE 106 91Stockholm, Sweden. E-mail: [email protected] 10.1002/bies.20516Published online in Wiley InterScience (www.interscience.wiley.com).

74 BioEssays 29.1 BioEssays 29:7484, 2006 Wiley Periodicals, Inc.

Problems and paradigms


2/11

ii).(11,13,17) The discovery of eukaryote-specific genes hasbeen used to argue that fusion must have involved threepartners; (6) theseauthorsprefer engulfment of an hypotheticalRNA-based chronocyte that evolved into the nucleus (variantof ii ) rather than standard models of gene gain and loss.A simpler explanation is of course that the eukaryote domainhad already split from the archaeal domain, as model iv suggests. (11,12) The confusion has arisen because two

aspects of an archezoa hypothesis were combined. Theseare (a) an early eukaryote cell engulfed the bacterial ancestorof mitochondria, and (b) some modern anaerobic eukaryotes,because they appeared to lack mitochondria, are examples ofsuch early eukaryotes.

Point b is now disproved through the death of theArchezoa hypothesis. (3,1820) Unfortunately, many have mis-takenly assumed that this also disproves the first part of the

Figure 1. Fourclassesof modelsfor theoriginof eukaryotes. Model i , topleftpanel: fusionbetweenan archaeonanda bacterium.Asusedhere,fusionimpliesphysical fusionoftwo cells leadingto a singlecellularcompartment anda singleintegratedgenome(whether or notthe genome is separated by a membrane). This process is different to the origin of the mitochondrion, meaning that modern eukaryotesemergefrom three cells, a fusion followedbyendosymbiosis.Thisparticularscenariowassuggestedby Zillig andcolleagues, (57) though theterm fusionis now used widelybut imprecisely. However, werestrictusage of theterm to cell fusion, as indicatedin model i . Model ii , topright panel: endosymbiosis with the nucleus evolving directly from an engulfed archaeal (or other cell). The nature of the engulfing cellvaries,but it is nevera eukaryoticcell.Thusmodelsvary onboththe natureof thehost andthe endosymbiontwhichevolves intothe nucleus.Note that modelsof this type require three cells in order to include themitochondrion, although some authors suggest thenucleus evolveddirectly from a virus. (6,7,14,84,85) Model iii , bottom left panel: endosymbiosis with an archaeal cell taking up the ancestor to themitochondrion. This model implies that the first step in the evolution of eukaryotes was the evolution of the mitochondrion; all eukaryote-specificfeatures thusevolved after thisevent. (32,45) Model iv , bottomright panel: endosymbiosis witha protoeukaryotic cellengulfingthe ancestorof themitochondrion. Thismodel implies thateukaryoteswere a separatelineage, distinctfrom archaea,and that thehostcellpossessed at least some eukaryote-specificfeatures.In some formulations,it canbe taken to imply that thefinal step in theevolutionof theeukaryote lineage was engulfment of the mitochondrial ancestor. (69)


BioEssays 29.1 75


3/11

hypothesis. Here we critically examine whether rejection ofpoint b necessarily leads us to an archaeal origin foreukaryotes.

A major problem is that there is no agreement on whatconstitutes a viable hypothesisunsurprisingly, there is noshortage of advocacy, but how should one evaluate thesealternatives?Weattemptto do thisby considering thefollowingpoints.

(1) The null hypothesis should be that the unique features ofan extant group evolved along its stem lineage (that is,prior to the last common ancestor of the groupFig. 2).

(2) Where possible, phylogenetic tests should be used toevaluate the various hypotheses (enabling rejection of thenull hypothesis in certain cases).

(3) Hypotheses that require hitherto undemonstrated bio-logical processes should be avoided when known

mechanisms areavailable.This is toprefer known causesto explain events in the past. (21)

We address each of these points in turn.

The null hypothesis:evolution along stem lineagesAs there are no known extant archezoa, the nature of thehostfor the mitochondrial endosymbiont must obviously bereconsidered. This is a classic example of stem lineages andcrown groups. Standard usage is that the last commonancestor of a group, plus all its descendents (living andextinct), is the crown group; the earlier groups form the stemlineage (Fig. 2a). For questions such as mammalian, bird,vertebrate or land plant evolution, we know from the fossilrecord that there has been successive acquisition of thecomplex characters seen in extant (crown group) species.

Figure 2. Stem and crown as applied to archaea and eukaryotes. A: If archaea and eukaryotes are each monophyletic and are sistergroupstheneach can bedefinedas consisting of a stem groupand a crowngroup. The crowngroupis thelastcommonancestor ofa group,plusall its descendents both living and extinct. The stemgroupare all earlier diverging, but now extinct, taxa. Adapted fromDonoghue. (86)

B: Given the absence of any fossil forms, features in either the archaeal or eukaryote crown groups cannot be confidently placed in thecommon ancestor. Thus, crown-specific features present in either archaea and eukaryotes could have been present in the commonancestor, or have evolved in one stem group. Current data cannot shed light on these alternatives with the exception of the mitochondrionwhich,bydefinition,was incorporatedinto theeukaryote lineagealongthe stem. Oncurrentknowledge,phagocytosisprecededtheorigin ofthe mitochondrion, though, again, it is not possible to say if phagocytosis evolved along the eukaryote stem or was a feature of the lastcommonancestorof eukaryotes andarchaea. C: Archaeaand eukaryotes evolve froma fullyformedarchaeal ancestor. Assumingarchaeaandeukaryotes aresister groups,the implicationsin termsof stemand crown features, areas follows.All featuresthat areuniversal amongcrown group archaeaevolved at someveryearly stage before thecommonancestor. On theother hand, all featuresuniversalamongcrowngroup eukaryotes evolved along the eukaryote stem, though the absence of any fossil data makes it impossible to divine the order ofemergence.Key featuresare indicatedschematicallyand include:phagocytosis,mitochondria, nucleusand endomembrane system,linearchromosomes, spliceosomal apparatus and introns. The postulated order of emergence differs depending on whether the archaeon isclaimed to be an endosymbiont that became the nucleus (model ii ), or whether it was the host that engulfed the ancestor of themitochondrion (model iii ).


76 BioEssays 29.1


4/11

However, if data from fossils are ignored onecould mistakenlyconclude that numerous traits have appeared in evolutionarybursts. (22) This is exactly the problem that we face regardingthe origin of eukaryotes: did eukaryote-specific features arisein successive stages, or in one burst? In one sense, theArchaea illustrate this point in that they are intermediate ininformationalgenes (suchas those involved in replication and

transcription) between Bacteria and Eukaryotes; (2325) with-out their discovery the apparent gulf would be even larger.The central question is thus: had eukaryotes become adistinct lineage by the time of the origin of mitochondria byendosymbiosis?

Under point 1 above, the null hypothesis is that featuresspecific to eukaryotes evolved in early eukaryotes after they

Figure 2. (Continued )


BioEssays 29.1 77


5/11


6/11

extant methanogens, these eukaryote genes grouping within the euryarchaeal part of the archaeal tree. Methanogen-likegenes in eukaryotes as judged by sequence similarity (17) aresuggestive, but insufficient to demonstrate an intra-archaealorigin for these eukaryote genes.

Importantly, sisterhood between archaea and eukaryotesdoes not support a methanogenic archaeal host; methano-genesis cannot be unequivocally placed in either the archaealcrown group ancestor, or the archaeaeukaryote commonancestor (Fig. 2a). To our knowledge, the only proposedphylogenetic result suggesting eukaryotes group withinarchaea is the eocyte tree, which instead groups eukaryoteswith crenarchaea, (5,33,34) not euryarchaeota. The eocyte treeis a source of ongoing debate (8,3540) and, rightly or wrongly,the prevailing consensus is that archaea and eukaryotes aresister groups. Controversy over tree topology and possibletree-building artefacts aside, neither of these topologiessupport a methanogenic host in the emergence of moderneukaryotes, in spite of suggestive associations such as

histone homologues in methanogens.(41)

Theseresults indicatea difficultywith includingan archaealpartner in a model for the origin of the eukaryote cell (modelsiiii ). Phylogenetic analyses strongly demonstrate bacterialorigins for mitochondria and plastids, (8,9,11,17,4244) so clearlyat thisphylogeneticdepth, there is sufficient signal to establishthe bacterial origin for the mitochondrion, and a later origin ofthe chloroplast. Why then would genes purported to be ofarchaeal origin not group within the diversity of modernarchaea? Numerous post hoc explanations can be invoked(as typified below) but, in reality, the data as they stand do notsupport an archaeal origin for either eukaryote genes (asfusion partner, model i ), or the nucleus (as endosymbiont,

model ii ) or as host (model iii , equivalent to fusion and theendosymbioticorigin for thenucleus in termsof a phylogenetictest for genes of archaeal origin, though only two cells arerequired). Furthermore, the mechanism of endocytosisresulting in the mitochondrion may be unspecified, as withfusion (model i ), or unprecedented, as in endosymbiosis(models ii and iii no extant archaeal species have thus farbeen shown to be either intracellular endosymbionts ofbacteria, or capable of engulfment).

One post-hoc explanation that could be invoked to rescuethesehypotheses is that the archaeal lineage that contributedgenetically to eukaryotes (as well as any other more basalarchaeal lineages) has since gone extinct. This would leave

both archaea andeukaryotesappearing monophyletic. As thedefining event in the origin of the eukaryote lineage is bydefinition the origin of the mitochondrion (under model iii ), theadditional requirement is that all eukaryote-specific featuresevolved after this event. (3,15,45) Thisrequires that, at minimum,one feature other than the mitochondrion was sufficientlyadvantageous to the individual in which it first appeared that aselective sweep eliminated even distantly related lineages.

(Excepting the unlikely possibility of rapid emergence of aselectively neutral complex feature becoming fixed by drift ina small otherwise undiversified population. (46) ) Alternatively,one could argue post-hoc, that the reason eukaryotes do notgroup within archaea is that the former are fast-evolving,meaning the monophyly of archaea is a consequence of long-branch attraction (seecommentby Martinin discussionon p85in Ref. (47)). Long-branch attraction has been invoked intesting the eocyte hypothesis, as the slowest-evolving sites inseveral protein data sets recover the eocyte topology. (33)

However, no evidence hasbeen presented to indicate that thislong-branch artefact accounts for the lack of a phylogeneticrelationship between methanogens and eukaryotes.

Until evidence to the contrary is produced, the nullhypothesis can only be that eukaryotes do not group withinarchaea, that is, archaea are strictly monophyletic. Thuscurrent theories invoking an archaeal partner (of the generaltype iiii ) are backed at best by no more than circumstantialevidence. The clear phylogenetic result seen with mitochon-

dria,(30)

chloroplasts(29)

and secondary endosymbioses(48)

has been strengthened as additional data have becomeavailable (as cited above). Indeed all endosymbioses orendoparasitisms, where such a phylogenetic test of ancestryhas been performed, provide a consistent picture. In contrast,phylogenetic data convincingly showing eukaryotes arisingwithin archaea are simply not available. The evolution offeatures unique to each of these domains must therefore beconsidered to have occurred in their respective stems from anancestor whose nature is not yet known (Fig. 2b).

Known mechanisms and endosymbioses

Science aims to explain events in the past by knownmechanisms. There are countless examples of endosym-bionts within eukaryote cells, and phagocytosis is a processwidespread among eukaryotes. Chloroplasts provide a clearancient exampleof endosymbiosisinvolvinga eukaryotichost,but nitrogen-fixing spheroid bodies in freshwater diatoms, (49)

the rhizobia of legumes, (50) bacterial endosymbionts ofinsects, (51) and red and green algal secondary and tertiaryendosymbionts (52,53) are among numerous other examples.To these examples we shall return, but first we examine theevidence that can be brought to bear on models proposingeukaryotes emerging from fusion or symbiosis between abacterium and an archaeon.

Model i predicts that genetic level fusion between archaeaandbacteriashould occur in nature. A possible example is thatof Thermotoga maritima (54) where, on sequence similarity, asmuch as 24% of the genome is suggested to be of archaealorigin. For a subset of such genes several indicators ofpotential xenology (homology via horizontal gene transfer)were reported, including local GC content, synteny, codonusage and the presence of flanking repeat sequences. (54)


BioEssays 29.1 79


7/11

While this example suggests that a significant number ofgenes can be transferred between the two domains, severalother points should be considered. Horizontal transfer couldpossibly produce a significant fusion signal at the genomiclevel, (55) but this is not equivalent to fusion of archaeal andbacterial compartments (model i ), and can be distinguishedfrom fusionif there is continualtransfer from a rangeof donors,as recently suggested. (55) In contrast to both horizontal genetransfer and endosymbiosis, no observations of fusion havethus farbeenreported. Second, the T. maritima genome resultargued for horizontal gene transfer for 24% of the genomebased on a blast analysis; the number of genes whereadditional data strongly support transfer are fewer andphylogenetic data, to our knowledge, exist only for one gene,reverse gyrase. (56) Third, a simple genomic fusion doesnot account for eukaryote cell structure, and all four modelsin Fig. 1 are in fact compatible with the identification ofbacterial and archaeal genes in eukaryotes, if the onlymeasure is sequence similarity. Thus, model i is the least

helpful in that it does not account for any eukaryotic cellarchitecture (not even the origin of the mitochondrion, whichoccurs subsequent to the fusion (57) ). It does however predictthe same phylogenetic relationship as models ii and iii forgenes originating from the two fusion partners (Fig. 3), asdiscussed above.

It is a prediction of models ii and iii above that archaeal/ bacterial endosymbioses would occur in nature. To ourknowledge only one example of endosymbiosis betweenprokaryotes is known; that of g-proteobacteria containedwithin b-proteobacteria, which are in turn contained withinthe cytoplasm of cells that make up the bacteriome, aspecialised organ in mealybugs that hosts endosymbiont

bacteria.(58)

The g-proteobacterial secondary endosymbiontsappear to have taken up residence on four independentoccasions and, subsequent to each infection event, havecoevolved with the b-proteobacterial primary symbionts. (59)

Access to the bacterial cytoplasm has been documentedpreviously, suggesting this is not a one-off observation. Forinstance, Daptobacter , a predatory bacterium, appears topenetrate the cytoplasm of its bacterial prey, (60,61) though toour knowledge no endosymbioses resulting from invasionhave been documented, and this genus is poorly studied.Several other examples suggest bacterial cells can take upresidence or at least parasitize other bacteria. The predatorybacterium, Bdellovibrio bacteriovorus ,(6062) invades the

periplasmof its bacterial hosts, where it reproduces, ultimatelybursting out of its host, which is lysed in the process. A similarmode of invasion is employed by a -proteobacterial symbiontsof the tick, Ixodes ricinus ; these bacteria invade and consumemitochondria of ovarian cells. (63,64)

Some authors have considered the mealybug example toindicate the plausibility of an archaealbacterial origin for theeukaryote cell. (47) However, this single example of a bacterial-

bacterial endosymbiosis is not equivalent to an archaealbacterial endosymbiosis; no examples of the latter have so farbeenreported, andunless evidenceappears thatdemonstrat-es this type of relationship, models ii and iii are not supported.

Furthermore, suggestions that the nucleus was an en-dosymbiont (model ii ) do not explain the unique membranestructure of the nuclear envelope (or for that matter any othereukaryote-specific structure) (15,16,65) since this has no coun-terpart in either archaea or bacteria. The internal membranesof bacterial planctomycetes, in particular members of thegenus Gemmata , are similar and possibly analogous to thenuclear envelope. However, current genomic data do notsupport an endosymbiotic or fusion scenario for planctomy-ceteevolution. Nordo eukaryote cells appear to have receivedgenes or organelles from this bacterial group. (66)

Endosymbiosis, where the engulfing cell is archaeal inorigin(model iii ), requires,on current knowledge, thatall extantarchaea have lost the capacity to internalise bacterialsymbionts. Modern examples of syntrophy between archaeal

methanogens and hydrogen-producing bacteria are invokedas indicative of the first step in the hydrogen (32) andsyntrophy (31,67) hypotheses for the origin of the eukaryotecell. While these hypotheses are persuasively argued from ametabolic viewpoint, neither explains the apparent absence ofcontemporary archaea harbouring endosymbionts, and theabsence of phagocytosis in prokaryotic lineages, archaea inparticular.

To recap, no archaea have been shown to carry bacterialendosymbionts (predicted by model iii and some variants ofmodel ii ), neitherhave anyarchaealendosymbionts of bacteriabeen observed (predicted by some variants of model ii ).Bacterial endosymbionts are extremely rare in bacteria, and

phagocytosis has so far not been demonstrated to be anattributeof archaea.Thus, whichever way onelooks at it,thereis currently no known precedent for endosymbiosis involvingan archaeon and a bacterium.

A eukaryotic ancestor capable of engulfment (i.e. endosy-mbiosis of the mitochondrial ancestor into a protoeukaryoticcell (model iv ), has been argued to present a common andknown mechanism for incorporation of a bacterial cell into aeukaryotic cell. (68,69) In Fig. 2a, this ancestor would correspondto the eukaryote crown common ancestor. Both engulfmentand endosymbiosis are widespread in eukaryotes andphagocytosis is common. However, simple engulfment of preyseems tomany tobe toosimplisticto account for themetabolic

interdependence between the eukaryote cell and its fledglingmitochondrion (perhaps explaining the focus of previousmodels (31,32,70) on the possible metabolic nature of theinteraction).

Having said that, as all subsequent endosymbiosesleading to establishment of organelles must have occurredvia phagocytosis, (71) it is hard to understand why the firstendosymbiosis need be the exception. Indeed, the more


80 BioEssays 29.1


8/11

recent (and incontrovertible) endosymbioses demolish anysuggestion that engulfment is too simplistic an explanation.

An importantaspect of theoriginofeukaryoticplastids (thatis likewise relevant for the origin of eukaryotes and mitochon-dria) is establishing that phagotrophic cells can enter intosymbiosis with their prey and evolve to become primaryproducers. A straightforward, thoughderived, example is givenbymixotrophiceukaryotesorganisms capableofbothphoto-trophy (on account of possessingphotosynthetic plastids) andphagotrophy (engulfment of microbes as food). A diverseassemblage of eukaryotes, including dinoflagellates, (72)

ciliates (73) and chlorarachniophytes, (74) are known to bemixotrophic. Consequently there is no controversy surround-ingthestatementthat phagotrophywascentral to theevolutionof obligately photosynthetic eukaryotes via primary, second-ary or tertiary endosymbioses. (52,71,75) Indeed, the spectrumof nutritional strategies, ranging from near-exclusive phago-trophy to near-exclusive phototrophy illustrates a feasible setof intermediates in the evolution of obligate phototrophs from

phagotrophic ancestors.(71,73)

It is beyond doubt that primaryproducerscanandhave evolved fromphagotrophic ancestors,and the process of endosymbionts evolving into organelles ineukaryotes has clearly occurred multiple times. (52)

Likewise, very transient symbiotic interactions can arisefrom preyengulfment; a salient example is that of kleptochlor-oplasts. Some dinoflagellates, ciliates and sea slugs areknown to engulf and digest photosynthetic algal cells, leavingonly the chloroplast, which is transiently retained in aphotosynthetically active state. After a short period, thekleptochloroplast is digested, further photosynthesis beingonly possible upon engulfment of additional photosyntheticeukaryote prey. (7678) The recently characterised flagellate

Hatena , which carries a plastid-bearing green algal symbiont,provides yet another example of this. Upon division, only onecell receives the engulfed symbiont; the cell without thesymbiont develops a feeding apparatus, enabling engulfmentof a new symbiont; this in turn leads to degeneration of thefeeding apparatus. (79)

These examples of mixotrophic eukaryotes preying onother eukaryotes demonstrate the feasibility of many inter-mediate stages in the evolution of photosynthetic eukaryotes.However, this establishes the mechanistic feasibility of thephagotrophic origin of the chloroplast or of secondary ortertiary endosymbionts. Given such a firm basis for theendosymbiotic photosynthetic model, we can move ahead to

consider a model for the origin of mitochondria.

A stepwise model for the origin of mitochondriaWe will now look at whether a stepwise model for the origin ofmitochondria by engulfment by the eukaryote crown commonancestor can be demonstrated by consideration of modernexamples.

A simple phagotroph-prey scenario for the emergence ofan obligate endosymbiosis would require the following steps:

(1) Phagotrophs engulfing prey cells via phagocytosis.(2) Emergence of individuals withinthe prey population which

are resistant to digestion, and which may escape from theinterior of the phagotrophic cell.

(3) Emergence of a facultative symbiotic relationshipbetweenphagotroph and prey; this couldbe mutualistic, commen-sal or a pathogenic interaction.

(4) Shift from a facultative to an obligate endosymbioticassociation.

(5) The obligate endosymbiont evolves into an organelle.

There is no shortage of examples of phagotropic eukar-yotes that engulf bacteria (step 1). Many types of amoebaeengulf and digest bacteria, and Enterobacteria are onecommon food source. (80) Engulfment can have several out-comes: digestion, resistance or escape. The latter two are

evolved traits amongst prey and often figure in strategiesemployed by pathogens, having developed resistance todestruction after engulfment (step2). Amoebae have receivedparticular attention, not least because they act as reservoirsfor pathogenic bacteria; numerous bacterial pathogenssurvive and proliferate in both amoebae and human macro-phages. (80)

Numerousexamples ofbacterialresistance todigestion arewell known from studies of the immune response in multi-cellularorganisms, (81) anda significantlisthas been collatedofbacteria resistant to engulfment by amoebae, (80) all serving toillustratethe feasibility of point 2. For instance,pathogens fromclinically important genera such as Burkholderia , Legionella ,

Listeria , Mycobacterium and Salmonella enter macrophagecells via the phagocytic pathway. Upon engulfment, thesesubsequently escape degradation by the phagosome througha variety of mechanisms, such as preventing fusion of thephagosome with the lysosome, or actively subverting theimmuneresponse. (81) Again, resistanceto degradation is seenin amoebae, leading to intracellular persistence, proliferationand, in some cases, to lysis of the engulfing cell. (80)

Resistance to engulfment might well provide a mechanismfrom which the establishment of mutualistic or commensalendosymbioses occurs, with the initial contact being aphagotrophprey interaction. As the above examples illus-trate, this can frequently turn into a parasitic endosymbiosis.

However, there are also cases where host cell lysis has beendocumented to be regulated by environmental cues, implyingthat a commensal relationship may emerge if favourableconditions prevail.

A spectrum of examples illustrating steps 3 and 4 areprovidedby a -proteobacterialendosymbionts of Acanthamoeba .In the case of Candidatus Odysella thessalonicensis , anincrease in incubation temperature from 22 8C to 3037 8C


BioEssays 29.1 81


9/11

results in a shift from stable intracellular occupation to lysis ofits amoebalhost, A. polyphaga .(82) Moreover, SSUrRNAtreesof Acanthamoeba hosts and their a -proteobacterial endosym-bionts (including Candidatus O. thessalonicensis ) are con-gruent, suggesting that there has been coevolution andcospeciation between hosts and endosymbionts. (83)

Step 5 is well established; salient examples are green-algal-derived secondary endosymbionts that gave rise toEuglenidsand Chlorarachniophytes, red-algalendosymbiontsof Cryptomonads, Dinoflagellates and others. Perhaps thebest-known example is given by Buchnera , vertically inheritedintracellular bacterial endosymbionts of aphids. While thisassociation is with a multicellular eukaryote, it illustrates a keyintermediate stage in the evolution of organelles from free-living bacteria. Buchnera genomes are far less reduced thaneither plastids or mitochondria, and it is not yet clear whetherany endosymbiont genes have been transferred to the hostnuclear genome. (75)

In conclusion, given the massive number and diversity of

endosymbiotic bacteria and organelles of bacterial-originresident in modern eukaryote groups, there is overwhelmingevidence for the establishment of endosymbioses betweenbacteria and eukaryotes. Again, there are currently noobservations of archaea either as hosts of bacterial endo-symbionts, no archaeal endosymbionts in bacteria, or forthat matter archaea residing with other archaea. On thecriterion of known processes, it is thus difficult to arguerationally for any other model than iv (Fig. 1).

ConclusionsThe first part of thearchezoa hypothesis, namely the idea thatthe eukaryote cell type had emerged before the incorporation

of mitochondria into eukaryotes, was initially taken as given.However, at that time it was assumed that some eukaryotesthat lacked obvious mitochondria were examples of eukar-yotes that had never had mitochondriathat is, they werearchezoa. With the realisation that extant archezoa aresecondarily amitochondriate (they are anaerobes and pos-sess vestigial mitochondrial forms) the two independentelements of the archezoan hypothesis were lumped together.It is nowrecognised, correctly, that extant archezoa have lostmitochondria, so the ancestor of all extant eukaryotes alsopossessed a mitochondrion. (3)

The first element of the archezoanhypothesis, theexisten-ce of an early protoeukaryote lineage, is what we have

addressed here. Does rejection of the second point, whetherthere are any living archezoa, logically lead us to an archaealorigin for eukaryotes? Most definitely not. As we point out,a definitive phylogenetic test is lackingmodern eukaryotesshould group within the diversity of modern archaea, in thesameway as is seenfor therelationshipbetweenmitochondriaand a-proteobacteria. Second, no contemporary evidenceexists for endosymbiontsin archaea,norarearchaeaknown to

be able to engulf cells. In contrast, there is abundant evidencefor both these phenomena in eukaryotes. The third point isthat, in addition to the mitochondrion, there are a number ofeukaryote-specific features for which no counterpart exists inarchaea.

Any theory that aims to explain the origin of mitochondriavia engulfment by an archaeal ancestor must not only seekevidence for the first two points, but also explain why theacquisition of mitochondria was the first step. If Archaea andEukaryotes are sister groups, as is widely held, similaritybetween these groups is expected on account of commonancestry, but such similarity does not allow us to deduce thenature of theancestor (i.e. eukaryoticor archaeal). To argue anarchaeal ancestry for eukaryotes, or a eukaryotic ancestry forarchaea, requires that one group falls within the phylogeneticdiversity of the other (Fig. 3); evidence for such a claim isbasically non-existent. Hypotheses involving fusion of threegroups seem superfluous; there is physical and geneticevidence for two cells, but no convincing genetic evidence

for the inclusion of a third.(3,55)

On the basis of known mechanisms of engulfment andknown examples of endosymbiosis, all data point to amechanism of cell engulfment being a prerequisite for theorigin of mitochondria. Again, on known mechanisms, engulf-ment can only have emerged during evolution along theeukaryote stem. The mealybug endosymbionts (a g-proteo-bacterium within a b-proteobacterium within a eukaryote cell)are not evidence for an archaeal-bacterial endosymbiosis.How a bacterial cell can gain entry into another bacterium inthis case is certainly an interesting and unsolved mystery, butthis is a known phenomenon, as illustrated by Bdellovibrio .However, thisdoesnot beara closeresemblance to eukaryotic

endosymbioses. To reiterate, no archaea are known to eitherhost endosymbionts or to be hosted as endosymbionts ofbacteria. Hence, it is reasonable to expect that there wouldbe phagocytotic predators before the existence of moderneukaryotes. The simplest hypothesis, based on currentknowledge, is that such cells were the stem lineage to moderneukaryotes.

To conclude, fusion and endosymbiotic hypotheses invol-ving an archaealand a bacterialpartner require that eukaryotegenes purported to be of archaeal origin have lost all signal oftheir origins, while this has not occurred for all eukaryoticgenes of mitochondrial origin. Phylogenetic data fit best withthe monophyly of the three domains, a common origin for

eukaryotesandarchaea, and acquisition of the endosymbioticprecursor of mitochondria early in eukaryote evolution byeukaryotic mechanisms of engulfment. 1 The alternativemodels require some series of untestable post hoc appealsto explain the data, such as all modern archaea losing their

1Note that this is not incompatible with interdomain transfer for specificgenessee Lester et al. (55) for a recent discussion.


82 BioEssays 29.1


10/11

abilities for endosymbiosis/engulfment, or genes evolvingwitha different mode and/or tempo on thearchaeal side of the treefrom the bacterial. Yes, archaea and eukaryotes appear toshare a common ancestor, and yes, modern eukaryotes are agenetic chimera with many genes being transferred to theeukaryote nucleus from the mitochondrion. However, explain-ing the numerous differences between modern eukaryotesand archaea requires evolution of domain-specific features onboth stems, archaeal and eukaryotic. Since this is a matterof the order of appearance of eukaryote-specific charactersalong the eukaryote stem lineage, and given the difficulties inaccounting for the ultimate engulfment of a bacterium by anarchaeon, the simplest explanation is that, by the time themitochondrion was incorporated into ancestral eukaryotes, atminimum, phagocytosis must have evolved. The suggestedspecific link between eukaryotes and methanogens owing tothe presence of histones the latter is further weakened by thediscovery of histones in crenarchaea. (87)

AcknowledgmentsWe thank Angela Douglas for many helpful comments on themanuscript.

References1. Embley TM, van der Giezen M, Horner DS, Dyal PL, Foster P. 2003.

Mitochondria and hydrogenosomes are two forms of the same funda-mental organelle. Philos Trans R Soc Lond B Biol Sci 358:191201;discussion 201192.

2. Tovar J, Leon-Avila G, Sanchez LB, Sutak R, Tachezy J, et al. 2003.Mitochondrial remnant organelles of Giardia function in iron-sulphurprotein maturation. Nature 426:172176.

3. Embley TM, Martin W. 2006. Eukaryotic evolution, changes andchallenges. Nature 440:623630.

4. van der Giezen M, Tovar J. 2005. Degenerate mitochondria. EMBO Rep6:525530.

5. Rivera MC, Jain R, Moore JE, Lake JA. 1998. Genomic evidence for twofunctionally distinct gene classes. Proc Natl Acad Sci USA 95:62396244.

6. Hartman H, Fedorov A. 2002. The origin of the eukaryotic cell: a genomicinvestigation. Proc Natl Acad Sci USA 99:14201425.

7. Horiike T, Hamada K, Kanaya S, Shinozawa T. 2001. Origin of eukaryoticcell nuclei by symbiosis of Archaea in Bacteria is revealed by homology-hit analysis. Nat Cell Biol 3:210214.

8. Rivera MC, Lake JA. 2004. The ring of life provides evidence for agenome fusion origin of eukaryotes. Nature 431:152155.

9. Gabaldon T, Huynen MA. 2003. Reconstruction of the proto-mitochon-drial metabolism. Science 301:609.

10. Canback B, Andersson SG, Kurland CG. 2002. The global phylogeny ofglycolytic enzymes. Proc Natl Acad Sci USA 99:60976102.

11. Karlberg O, Canback B, Kurland CG, Andersson SG. 2000. The dualorigin of the yeast mitochondrial proteome. Yeast 17:170187.

12. Andersson SG, Karlberg O, Canback B, Kurland CG. 2003. On the originof mitochondria: a genomics perspective. Philos Trans R Soc Lond B BiolSci 358:165177; discussion 177169.

13. Martin W, Hoffmeister M, Rotte C, Henze K. 2001. An overview ofendosymbiotic models for the origins of eukaryotes, their ATP-producingorganelles (mitochondria and hydrogenosomes), and their heterotrophiclifestyle. Biol Chem 382:15211539.

14. Lake JA, Rivera MC. 1994. Was the nucleus the first endosymbiont? ProcNatl Acad Sci USA 91:28802881.

15. Martin W. 1999. A briefly argued case that mitochondria and plastids aredescendants of endosymbionts, but that the nuclear compartment is not.Proc R Soc Lond B 266:13871395.

16. Poole A, Penny D. 2001. Does endosymbiosis explain the origin of thenucleus? Nat Cell Biol 3:E173174.

17. Esser C, Ahmadinejad N, Wiegand C, Rotte C, Sebastiani F, et al. 2004.A genome phylogeny for mitochondria among alpha-proteobacteria anda predominantly eubacterial ancestry of yeast nuclear genes. Mol BiolEvol 21:16431660.

18. Roger AJ. 1999. Reconstructing Early Events in Eukaryotic Evolution. AmNat 154:S146S163.

19. Keeling PJ, McFadden GI. 1998. Origins of microsporidia. TrendsMicrobiol 6:1923.

20. Embley TM, Hirt RP. 1998. Early branching eukaryotes? Curr Opin GenetDev 8:624629.

21. Lyell C. Principles of geology, being an attempt to explain the formerchanges of the earths surface, by reference to causes now in operation.London: John Murray; 18301833.

22. Donoghue PCJ, Purnell MA. 2005. Genome duplication, extinction andvertebrate evolution. Trends Ecol Evol 20:312319.

23. Cramer P. 2002. Multisubunit RNA polymerases. Curr Opin Struct Biol12:8997.

24. Keeling PJ, Doolittle WF. 1995. Archaea: narrowing the gap betweenprokaryotes and eukaryotes. Proc Natl Acad Sci USA 92:57615764.

25. Kelman LM, Kelman Z. 2004. Multiple origins of replication in archaea.Trends Microbiol 12:399401.

26. Forterre P, Philippe H. 1999. Where is the root of the universal tree of life?Bioessays 21:871879.

27. Bowler PJ. 1987. The Non-Darwinian Revolution. Baltimore: JohnHopkins University Press.

28. Dayrat B. 2003. The roots of phylogeny: how did Haeckel build his trees?Syst Biol 52:515527.

29. Douglas SE, Turner S. 1991. Molecular evidence for the origin of plastidsfrom a cyanobacterium-like ancestor. J Mol Evol 33:267273.

30. Yang D, Oyaizu Y, Oyaizu H, Olsen GJ, Woese CR. 1985. Mitochondrialorigins. Proc Natl Acad Sci USA 82:44434447.

31. Moreira D, Lopez-Garc a P. 1998. Symbiosis between methanogenicarchaea and delta-proteobacteria as the origin of eukaryotes: thesyntrophic hypothesis. J Mol Evol 47:517530.

32. Martin W, Muller M. 1998. The hydrogen hypothesis for the firsteukaryote. Nature 392:3741.

33. Tourasse NJ, Gouy M. 1999. Accounting for evolutionary rate variationamong sequence sites consistently changes universal phylogeniesdeduced from rRNA and protein-coding genes. Mol Phylogenet Evol13:159168.

34. Lake JA. 1998. Optimally recovering rate variation information fromgenomes and sequences: pattern filtering. Mol Biol Evol 15:12241231.

35. Brinkmann H, Philippe H. 1999. Archaea sister group of Bacteria?Indications from tree reconstruction artifacts in ancient phylogenies. MolBiol Evol 16:817825.

36. Brown JR, Douady CJ, Italia MJ, Marshall WE, Stanhope MJ. 2001.Universal trees based on large combined protein sequence data sets.Nat Genet 28:281285.

37. Snel B, Bork P, Huynen MA. 1999. Genome phylogeny based on genecontent. Nat Genet 21:108110.

38. Forterre P, Brochier C, Philippe H. 2002. Evolution of the Archaea. TheorPopul Biol 61:409422.

39. Baldauf SL, Palmer JD, Doolittle WF. 1996. The root of the universal treeand the origin of eukaryotes based on elongation factor phylogeny. ProcNatl Acad Sci USA 93:77497754.

40. Rivera MC, Lake JA. 1992. Evidence that eukaryotes and eocyteprokaryotes are immediate relatives. Science 257:7476.

41. Reeve JN, Sandman K, Daniels CJ. 1997. Archaeal histones, nucleo-somes, and transcription initiation. Cell 89:9991002.

42. Martin W, Stoebe B, Goremykin V, Hapsmann S, Hasegawa M, et al.1998. Gene transfer to the nucleus and the evolution of chloroplasts.Nature 393:162165.

43. Martin W, Rujan T, Richly E, Hansen A, Cornelsen S, et al. 2002.Evolutionary analysis of Arabidopsis, cyanobacterial, and chloroplastgenomes reveals plastid phylogeny and thousands of cyanobacterialgenes in the nucleus. Proc Natl Acad Sci USA 99:1224612251.

44. Lang BF, Gray MW, Burger G. 1999. Mitochondrial genome evolutionand the origin of eukaryotes. Annu Rev Genet 33:351397.


BioEssays 29.1 83


11/11

45. Martin W, Koonin EV. 2006. Introns and the origin of nucleus-cytosolcompartmentalization. Nature 440:41 45.

46. Stoltzfus A. 1999. On the possibility of constructive neutral evolution.J Mol Evol 49:169181.

47. Martin W, Russell MJ. 2003. On the origins of cells: a hypothesis for theevolutionary transitions from abiotic geochemistry to chemoautotrophicprokaryotes, and from prokaryotes to nucleated cells. Philos Trans R SocLond B Biol Sci 358:5983; discussion 8385.

48. Douglas SE, Murphy CA, Spencer DF, Gray MW. 1991. Cryptomonadalgae are evolutionary chimaeras of two phylogenetically distinctunicellular eukaryotes. Nature 350:148151.

49. Prechtl J, Kneip C, Lockhart P, Wenderoth K, Maier UG. 2004.Intracellular spheroid bodies of Rhopalodia gibba have nitrogen-fixing apparatus of cyanobacterial origin. Mol Biol Evol 21:14771481.

50. Parniske M. 2000. Intracellular accommodation of microbes by plants: acommon developmental program for symbiosis and disease? Curr OpinPlant Biol 3:320328.

51. Moran NA, Baumann P. 2000. Bacterial endosymbionts in animals. CurrOpin Microbiol 3:270275.

52. Archibald JM. 2005. Jumping genes and shrinking genomes - probingthe evolution of eukaryotic photosynthesis with genomics. IUBMB Life57:539547.

53. Gilson PR, McFadden GI. 2002. Jam packed genomesa preliminary,comparative analysis of nucleomorphs. Genetica 115:1328.

54. Nelson KE, Clayton RA, Gill SR, Gwinn ML, Dodson RJ, et al.1999. Evidence for lateral gene transfer between Archaea andbacteria from genome sequence of Thermotoga maritima. Nature 399:323329.

55. Lester L, Meade A, Pagel M. 2006. The slow road to the eukaryoticgenome. Bioessays 28:5764.

56. Forterre P, Bouthier De La Tour C, Philippe H, Duguet M. 2000. Reversegyrase from hyperthermophiles: probable transfer of a thermoadaptationtrait from archaea to bacteria. Trends Genet 16:152154.

57. Zillig W, Klenk HP, Palm P, Leffers H, Pu hler G, et al. 1989. Dideukaryotes originate by a fusion event? Endocyt Cell Res 6:125.

58. von Dohlen CD, Kohler S, Alsop ST, McManus WR. 2001. Mealybugbeta-proteobacterial endosymbionts contain gamma-proteobacterialsymbionts. Nature 412:433436.

59. Thao ML, Gullan PJ, Baumann P. 2002. Secondary (gamma-Proteobac-teria) endosymbionts infect the primary (beta-Proteobacteria) endosym-bionts of mealybugs multiple times and coevolve with their hosts. ApplEnviron Microbiol 68:31903197.

60. Martin MO. 2002. Predatory prokaryotes: an emerging researchopportunity. J Mol Microbiol Biotechnol 4:467477.

61. Guerrero R, Pedros-Alio C, Esteve I, Mas J, Chase D, et al. 1986.Predatory prokaryotes: predation and primary consumption evolved inbacteria. Proc Natl Acad Sci USA 83:21382142.

62. Rendulic S, Jagtap P, Rosinus A, Eppinger M, Baar C, et al. 2004. Apredator unmasked: life cycle of Bdellovibrio bacteriovorus from agenomic perspective. Science 303:689692.

63. Beninati T, Lo N, Sacchi L, Genchi C, Noda H, et al. 2004. A novel alpha-Proteobacterium resides in the mitochondria of ovarian cells of the tickIxodes ricinus. Appl Environ Microbiol 70:25962602.

64. Sacchi L, Bigliardi E, Corona S, Beninati T, Lo N, et al. 2004. A symbiontof the tick Ixodes ricinus invades and consumes mitochondria in a modesimilar to that of the parasitic bacterium Bdellovibrio bacteriovorus.Tissue Cell 36:4353.

65. Rotte C, Martin W. 2001. Does endosymbiosis explain the origin of thenucleus? Nat Cell Biol 3:E173174.

66. Fuerst JA. 2005. Intracellular compartmentation in planctomycetes. AnnuRev Microbiol 59:299328.

67. Lopez-Garcia P, Moreira D. 1999. Metabolic symbiosis at the origin ofeukaryotes. Trends Biochem Sci 24:8893.

68. Kurland CG, Collins LJ, Penny D. 2006. Genomics and the irreduciblenature of eukaryote cells. Science 312:10111014.

69. Cavalier-Smith T. 2002. The phagotrophic origin of eukaryotes andphylogenetic classification of Protozoa. Int J Syst Evol Microbiol 52:297354.

70. Andersson SG, Kurland CG. 1999. Origins of mitochondria andhydrogenosomes. Curr Opin Microbiol 2:535541.

71. Raven JA. 1997. Phagotrophy in phototrophs. Limnol Oceanogr 42:198205.

72. Stoecker DK. 1999. Mixotrophy among dinoflagellates. J EukaryotMicrobiol 46:397401.

73. Jones RI. 2000. Mixotrophy in planktonic protists: an overview. Fresh-water Biology 45:219226.

74. McFadden GI, Gilson PR, Hofmann CJ, Adcock GJ, Maier UG. 1994.Evidence that an amoeba acquired a chloroplast by retaining part of anengulfed eukaryotic alga. Proc Natl Acad Sci USA 91:36903694.

75. Douglas AE, Raven JA. 2003. Genomes at the interface betweenbacteria and organelles. Philos Trans R Soc Lond B Biol Sci 358:517;discussion 1718.

76. Skovgaard A. 1998. Role of chloroplast retention in a marine dino-flagellate. Aquat Microb Ecol 15:293301.

77. Gustafson DE Jr, Stoecker DK, Johnson MD, Van Heukelem WF, SneiderK. 2000. Cryptophyte algae are robbed of their organelles by the marineciliate Mesodinium rubrum. Nature 405:10491052.

78. Rumpho ME, Summer EJ, Green BJ, Fox TC, Manhart JR. 2001. Mollusc/ algal chloroplast symbiosis: how can isolated chloroplasts continue tofunction for months in the cytosol of a sea slug in the absence of an algalnucleus? Zoology (Jena) 104:303312.

79. Okamoto N, Inouye I. 2005. A secondary symbiosis in progress? Science310:287.

80. Greub G, Raoult D. 2004. Microorganisms resistant to free-livingamoebae. Clin Microbiol Rev 17:413433.

81. Rosenberger CM, Finlay BB. 2003. Phagocyte sabotage: disruption ofmacrophage signalling by bacterial pathogens. Nat Rev Mol Cell Biol4:385396.

82. Birtles RJ, Rowbotham TJ, Michel R, Pitcher DG, Lascola B, et al. 2000.Candidatus Odyssella thessalonicensis gen. nov., sp. nov., an obligateintracellular parasite of Acanthamoeba species. Int J Syst Evol Microbiol50 Pt 1:6372.

83. Beier CL, Horn M, Michel R, Schweikert M, Gortz HD, Wagner M. 2002.The genus Caedibacter comprises endosymbionts of Paramecium spp.related to the Rickettsiales (Alphaproteobacteria) and to Francisellatularensis (Gammaproteobacteria). Appl Environ Microbiol 68:60436050.

84. Takemura M. 2001. Poxviruses and the origin of the eukaryotic nucleus.J Mol Evol 52:419425.

85. Bell PJ. 2001. Viral eukaryogenesis: was the ancestor of the nucleus acomplex DNA virus? J Mol Evol 53:251256.

86. Donoghue PCJ. 2005. Saving the stem group - a contradiction in terms?Paleobiology 31:553 558.

87. C ubon ova L, Sandman K, Hallam SJ, DeLong EF, Reeve JN. 2005.Histones in Crenarchaea. J Bacteriol 187:54825485.


84 BioEssays 29.1

Abril_10_2012_ Evaluating Hypotheses for the Origin of Eukaryotes (1)

Documents

Transcript of Abril_10_2012_ Evaluating Hypotheses for the Origin of Eukaryotes (1)