On the origins of cells: a hypothesis for the evolutionary ... the origin of cells.pdf ·...

Published online 5 December 2002

On the origins of cells a hypothesis for theevolutionary transitions from abiotic geochemistry tochemoautotrophic prokaryotes and from prokaryotes

to nucleated cells

William Martin1 and Michael J Russell21Institut fur Botanik III Heinrich-Heine Universitaet Dusseldorf Universitatsstrasse 1 40225 Dusseldorf Germany

2Scottish Universities Environmental Research Centre Scottish Enterprise Technology Park Rankine Avenue East KilbrideGlasgow G75 0QF UK (mrussellsuercglaacuk)

All life is organized as cells Physical compartmentation from the environment and self-organization ofself-contained redox reactions are the most conserved attributes of living things hence inorganic matterwith such attributes would be lifersquos most likely forebear We propose that life evolved in structured ironmonosulphide precipitates in a seepage site hydrothermal mound at a redox pH and temperature gradientbetween sulphide-rich hydrothermal fluid and iron(II)-containing waters of the Hadean ocean floor Thenaturally arising three-dimensional compartmentation observed within fossilized seepage-site metal sul-phide precipitates indicates that these inorganic compartments were the precursors of cell walls and mem-branes found in free-living prokaryotes The known capability of FeS and NiS to catalyse the synthesisof the acetyl-methylsulphide from carbon monoxide and methylsulphide constituents of hydrothermalfluid indicates that pre-biotic syntheses occurred at the inner surfaces of these metal-sulphide-walledcompartments which furthermore restrained reacted products from diffusion into the ocean providingsufficient concentrations of reactants to forge the transition from geochemistry to biochemistry The chem-istry of what is known as the RNA-world could have taken place within these naturally forming catalytic-walled compartments to give rise to replicating systems Sufficient concentrations of precursors to supportreplication would have been synthesized in situ geochemically and biogeochemically with FeS (and NiS)centres playing the central catalytic role The universal ancestor we infer was not a free-living cell butrather was confined to the naturally chemiosmotic FeS compartments within which the synthesis of itsconstituents occurred The first free-living cells are suggested to have been eubacterial and archaebacterialchemoautotrophs that emerged more than 38 Gyr ago from their inorganic confines We propose thatthe emergence of these prokaryotic lineages from inorganic confines occurred independently facilitatedby the independent origins of membrane-lipid biosynthesis isoprenoid ether membranes in the archae-bacterial and fatty acid ester membranes in the eubacterial lineage The eukaryotes all of which areancestrally heterotrophs and possess eubacterial lipids are suggested to have arisen ca 2 Gyr ago throughsymbiosis involving an autotrophic archaebacterial host and a heterotrophic eubacterial symbiont thecommon ancestor of mitochondria and hydrogenosomes The attributes shared by all prokaryotes areviewed as inheritances from their confined universal ancestor The attributes that distinguish eubacteriaand archaebacteria yet are uniform within the groups are viewed as relics of their phase of differentiationafter divergence from the non-free-living universal ancestor and before the origin of the free-livingchemoautotrophic lifestyle The attributes shared by eukaryotes with eubacteria and archaebacteriarespectively are viewed as inheritances via symbiosis The attributes unique to eukaryotes are viewed asinventions specific to their lineage The origin of the eukaryotic endomembrane system and nuclear mem-brane are suggested to be the fortuitous result of the expression of genes for eubacterial membrane lipidsynthesis by an archaebacterial genetic apparatus in a compartment that was not fully prepared to accom-modate such compounds resulting in vesicles of eubacterial lipids that accumulated in the cytosol aroundtheir site of synthesis Under these premises the most ancient divide in the living world is that betweeneubacteria and archaebacteria yet the steepest evolutionary grade is that between prokaryotes and eukaryotes

Keywords hydrothermal vents acetyl-CoA pathway origin of life RNA-worldchemoautotrophic origins holochirality

Author for correspondence (wmartinuni-duesseldorfde)

One contribution of 21 to a Discussion Meeting Issue lsquoChloroplasts andmitochondria functional genomics and evolutionrsquo

Phil Trans R Soc Lond B (2003) 358 59ndash85 59 Oacute 2002 The Royal SocietyDOI 101098rstb20021183

1 INTRODUCTION

The Earth is 45 billion years (Gyr) old and the first oceanhad condensed by ca 44 Gyr (Wilde et al 2001) Thereare good reasons to believe that life arose here by ca

60 W Martin and M J Russell On the origins of cells

38 Gyr because carbon isotope data provide evidence forbiological CO2 fixation in sedimentary rocks of that age(Mojzsis et al 1996 Rosing 1999 Nisbet amp Sleep 2001Ueno et al 2002) By 35 Gyr stromatolites were presentpreserved microbial mats indicative of deposition by pho-tosynthetic prokaryotes (Walter 1983 Nisbet amp Sleep2001) By ca 15 Gyr so-called acritarchs became reason-ably abundant microfossils of unicellular organisms thatare almost certainly eukaryotes ( Javaux et al 2001) andprobably algae because of an easily preserved cell wall By12 Gyr spectacularly preserved multicellular organismsappear that were very probably red algae (Butterfield2000) There have been reports of more ancient remainsclaimed to be eukaryotes but they are equivocal Forexample Grypania is a 21 Gyr fossil (Han amp Runnegar1992) but could just as easily be a filamentous prokaryoteas a filamentous eukaryote because the cellular structureof the fossil is not preserved More recently steranes werefound in 27 Gyr sediments and were claimed to provideevidence for the existence of eukaryotes at that time(Brocks et al 1999) but several groups of prokaryotesincluding methanotrophic proteobacteria (see Schouten etal 2000) myxobacteria (Kohl et al 1983) and cyanobac-teria (Hai et al 1996) make the same kinds of compounds(for example cholesterol) claimed to be eukaryote-specificsuch that the sterane evidence for the age of eukaryotes isquestionable at best Ultra-light carbon material that bearsthe isotopic signature of biological consumption of bio-logically produced methane goes back to 27 Gyr (Hayes1994) providing evidence for the distinctness of eubac-teria (methanotrophs) and archaebacteria (methanogens)by that time at the very least and furthermore indicatingthat arguments for an origin of archaebacteria only850 million years (Myr) ago (Cavalier-Smith 2002) mustbe in error

Those geological benchmarks provide rough but robustorientation for the process of early evolution the origin ofprokaryotes by at least 35 Gyr and the origin of eukary-otes by at least 15 Gyr But biologists and geologists alikewould like to know a few more details What happenedduring those first 25 Gyr How did life get off the groundto begin with Why did it take eukaryotes so long toappear Why are the two groups of contemporary prokary-otes (eubacteria and archaebacteria) so genetically coher-ent yet so biochemically different Why are there nointermediate forms between prokaryotic and eukaryoticorganization Where did the nucleus come from Herewe discuss evolutionary processes that may underlie thedifferences between eubacteria and archaebacteria andthat may underlie the differences between prokaryotes andeukaryotes as well as the mosaic patterns of attributes(Zillig 1991) shared between the three

2 COMMONALITIES AND DIFFERENCES LEADQUICKLY TO DEEPER PROBLEMS

Understanding the evolutionary differences betweenarchaebacteria and eubacteria requires a concrete pictureof their last common ancestor This is because one canonly discuss the differentiation of their attributes on thebasis of the attributes possessed by their progenitor Butbecause there is no direct evidence for the nature of their

Phil Trans R Soc Lond B (2003)

last common ancestor its attributes can only be addressedthrough logical inference (Penny amp Poole 1999)

What did the ancestor of prokaryotes possess for sureIt certainly possessed the universal genetic code tRNA(Eigen amp Schuster 1978 Woese et al 1990 Eigen 1992)and ribosomes with a battery of ribosomal proteins(Woese 2002) most of which are neatly encoded in a con-served superoperon that is present and conserved in geneorder across many genomes (Wachtershauser 1998) Fur-thermore the universal ancestor surely possessed DNA(Reichard 1997 Poole et al 2000) and DNA polymerasesRNA polymerases and a battery of accessory translationfactors such as EF-Tu and EF-G that are present in allprokaryotes (Poole et al 1998 1999) probably along withother more or less ubiquitous prokarytotic proteins suchas F1ndashF0-type ATPases (Gogarten et al 1989) as well asprokaryotic forms of the signal recognition particle whichis an accessory to translation (Brinkmann amp Philippe1999) Furthermore it had to have had all of the pathwaysnecessary for efficient nucleotide and amino-acidbiosyntheses a core carbon metabolism to provide thebiosynthetic precursors and the cofactors required forthose biosynthetic pathways but this is the point wherethings get very complicated very quickly because althougharchaebacteria and eubacteria do many things similarlywhen it comes to the processing of genetic information(Woese 2002) they do many things fundamentally differ-ently when it comes to central metabolism

One simple example for such core metabolic differencesbetween the two groups of prokaryotes is the breakdownof glucose to pyruvate the glycolytic (EmbdenndashMeyerhof)pathway which is universal among eukaryotes but notamong prokaryotes If we look among prokaryotic gen-omes for the familiar glycolytic enzymes that occur ineukaryotes we see that they have good homologuesamong a wide spectrum of eubacterial genomes but manyof the corresponding homologues are missing in archaeb-acteria (figure 1) This observation might seem puzzlingto some (Canback et al 2002) but is easily attributable totwo simple factors First although archaebacteria possesspathways from glucose to pyruvate they have several alter-native pathways that involve different enzymatic stepshence different enzymes not present in glycolysis(Daniel amp Danson 1995 Schonheit amp Schafer 1995Kengen et al 1996 Selig et al 1997) These include (i)the modified EmbdenndashMeyerhof pathway which employsADP- instead of ATP-dependent steps and a glyceral-dehyde-3-phosphate oxidoreductase in place of the stepscatalysed by GAPDH and 3-phosphoglycerate kinase (ii)the EntnerndashDoudoroff pathway which does not requirethe enzymatic steps catalysed by GPI PFK FBA or TPIand (iii) the non-phosphorylated EntnerndashDoudoroff path-way which shares only two enzymatic steps in commonwith glycolysis (Selig et al 1997) those catalysed by enol-ase (Hannaert et al 2000) and pyruvate kinase (Schrammet al 2000) Second several archaebacteria have completeglycolytic pathways but for many steps they use totallydifferent enzymes that are unrelated to the familiar homol-ogues found in eubacteria (and eukaryotes) (Schonheit ampSchafer 1995 Selig et al 1997) These include (i) anADP-dependent (rather than ATP-dependent) hexokin-ase with no sequence similarity to the eubacterial enzyme(Ito et al 2001) (ii) an ADP-dependent PFK that lacks

On the origins of cells W Martin and M J Russell 61

amino acid identity inpairwise alignments

80_8970_7960_6950_5940_4930_3920_29

no hit or gt20

ArabidopsisTrypanosomaSpironucleus

HumanSaccharomyces

Arabidopsis (HK)Trypanosoma (HK)

Giardia (GK)Human (HK)

Saccharomyces (HK)

Arabidopsis (PPi)Trypanosoma (ATP)

Entamoeba (PPi)Human (ATP)

Saccharomyces (ATP)

Arabidopsis (I)Trypanosoma (I)

Giardia(II)Human (I)

Saccharomyces (II)

ArabidopsisTrypanosoma

GiardiaHuman

Saccharomyces


GiardiaHuman

Saccharomyces


HumanSaccharomyces

Arabidopsis (I)Trypanosoma (I)Spironucleus (I)

Human(D)Saccharomyces (D)


GiardiaHuman

Saccharomyces


GiardiaHuman

Saccharomyces

The

rmot

oga

Aqu

ifex

Bac

illus

Lact

ococ

cus

Myc

obac

teriu

mS

ynec

hocy

stis

Esc

heric

hia

Cau

loba

cter

Chl

amyd

iaB

orre

liaM

agne

toco

ccus

Agr

obac

teriu

mM

esor

hizo

bium

Sin

orhi

zobi

umP

seud

omon

as

Aer

opyr

mS

ulfo

lobu

sH

alob

acte

rium

Met

hnob

acte

rium

Met

hano

cocc

usP

yroc

occu

s ab

ysii

Pyr

ococ

cus

horik

oshi

iA

rcha

eogl

obus

The

rmop

lasm

a ac

idop

hilu

mT

herm

opla

sma

volc

anii

GKHK

GPI

PFK

FBA

TPI

GAP

PGK

PGM

Enolase

PYK

Figure 1 Global patterns of similarity between the enzymes of the glycolytic pathways in eukaryotes and the correspondingenzymes from several completely sequenced eubacterial and archaebacterial genomes The eukaryotic amino-acid sequenceswere compared with all peptides from the corresponding genome with gapped Blast (Altschul et al 1997) the percentageamino acid identity from the pairwise alignment to the best match was recorded and colour coded Only such matches werecounted as aligned over 100 amino acids or more Note that the eukaryotic sequences are much more similar to theireubacterial homologues (GK (EC 2712) HK hexokinase (EC 2711) GPI (EC 5319) PFK 6-phosphofructo-1-kinase(ATPndashPFK EC 27111 PPindashPFK EC 27190) FBA (EC 412113 with class I (Schiff-base intermediate) and class II(endiol intermediate) types) TPI (EC 5311) GAPDH (phosphorylating EC 12112) PGK 3-phospho-d-glycerate kinase(EC 2723) PGM 3-phospho-d-glycerate mutase (EC 5421 23-bisphospho-d-glycerate dependent (D) and independent(I) types) enolase 2-phospho-d-glycerate hydrolase (EC 42111) PYK pyruvate kinase (EC 27140))



sequence similarity to eukaryotic or eubacterial PFK butis related to archaebacterial ADP-dependent GK instead(Tuininga et al 1999) (iii) an aldolase that shares nosequence similarity with either of the typical eubacterialenzymes (Siebers et al 2001) and (iv) a TPI that sharesonly residual (ca 20) amino-acid sequence similarity tothe eubacterial enzyme (Kohlhoff et al 1996) So whatdid the universal ancestor possess in its pathway from glu-cose to pyruvate and where did the glucose come fromin the first place We will return to these questions shortly

A more striking and much more significant exampleinvolves lipid biosynthesis Lipid biosynthesis is importantbecause lipids are the constituents of membranes and allcellsmdashwith no exceptionmdashare surrounded by membranesBut archaebacteria and eubacteria possess completely dif-ferent lipids (Kandler 1998) Archaebacterial lipids consistof isoprenoids and various derivatives connected to sn-gly-cerol-1-phosphate by an ether bond Eubacterial lipidsconsist of fatty acids connected to sn-glycerol-3-phosphateby an ester bond (Kates 1979) The biochemical pathwaysleading to these two kinds of lipid could not differ moreTheir lowest common denominators are acetyl-CoA forthe non-polar component and dihydroxyacetone phos-phate which is non-chiral for the backbone Eubacteriasynthesize their fatty acids from acetyl-CoA by carboxylat-ing it to malonyl-CoA which is used to elongate the grow-ing chain by two carbon units at a time with reductionof the 3-ketoacyl intermediates Archaebacteria synthesizetheir isoprenoids via the condensation of IPP and its iso-mer dimethylallyl diposphate C5 units that are synthe-sized from acetyl-CoA via the MVA pathway (Langworthyet al 1982) Notably eubacteria also synthesize isopreno-ids but they do so through a completely unrelated path-way the DXP pathway intermediates of which areprecursors for thiamin diphosphate and pyridoxal diphos-phate biosynthesis (Lange et al 2000) Converselyarchaebacteria generally do not synthesize fatty acids atall Regarding the glycerol backbone sn-glycerol-1-phos-phate and sn-glycerol-3-phosphate are stereoisomersEubacteria synthesize sn-glycerol-3-phosphate from dihy-droxyacetone phosphate with sn-glycerol-3-phosphatedehydrogenase archaebacteria synthesize the correspond-ing stereoisomer with sn-glycerol-1-phosphate dehydro-genase which shares no sequence similarity with sn-glycerol-3-phosphate dehydrogenase (Koga et al 1998)Finally the ether and ester linkages are highly distinctchemical bonds Of course many archaebacteria possessthe derived tetraether lipids and various modifications ofthe polar head have long been known in both groups(Langworthy et al 1982 Zillig 1991) Add to this the find-ing that murein the typical peptidoglycan of the eubacter-ial cell wall is missing in archaebacteriamdashone of thecriteria through which the group was discovered(Kandler amp Konig 1978) they possess pseudomurein andother constituents instead (Kandler 1982)mdashand we arrivefrom the top-down perspective at an exceedingly severeexplanandum given that there is no similarity whatsoeverin the components with which archaebacteria and eubac-teria uniformly compartmentalize their cytosol from theenvironment what on Earth could have been the precur-sor of their membrane and cell wall in their last com-mon ancestor


This question leads directly to a second even more sev-ere problem from the bottom-up perspective given thatlife evolved from the geochemicals and given that some-thing like an RNA-world (Gilbert 1986 Poole et al 1998Bartel amp Unrau 1999 Gesteland et al 1999 but see alsoOrgel amp Crick 1993) ever existed in which catalytic self-replicating systems could get off the ground how on Earthwas it possible to attain sufficient and sustained concen-trations of the building blocks of life that are needed forany self-assembling prebiotic system to get to the stagethat it could perform a single round of replication

3 UNANSWERED BUT NOT UNANSWERABLEQUESTIONS

The second problem is quite well known in the litera-ture on the origin of life and is known as the concentrationproblem as addressed by De Duve (1991 1994) Mayn-ard-Smith amp Szathmary (1995) and others (Cairns-Smith1982 Shapiro 1986) Various solutions to this problemhave been proposed among which the currently mostpopular are probably surface catalysis that is the concen-tration of reactants on inorganic surfaces allowing themto react (Bernal 1951 Weber 1995 Ferris et al 1996Huber amp Wachtershauser 1998) and evaporites that isthe evaporation of terrestrial ponds containing organicprecursors synthesized by MillerndashUrey-like reactions(Nelson et al 2001) A fundamental drawback to the sur-face catalysis model is that once two molecules havereacted on a surface they diffuse away into the Hadeanocean never to react again A fundamental drawback tothe evaporite model is the sustainability of catalysis anypond so envisaged has to dry up and refill without a wash-out very many times before something like a living systemcan arise which is hard to imagine given the absence ofcontinents high and highly frequent tides and 10 km deepoceans at ca 4 Gyr (Gaffey 1997 Bounama et al 2001Glasby amp Kasahara 2001 Kamber et al 2001) Thenotion of Hadean oceans chock-full of Oparinrsquos prebioticsoup still enjoys some popularity (Bada amp Lazcano 2002)but the question remains of how a solution at equilibriumcan start doing chemistry Put another way once auto-claved a bowl of chicken soup left at any temperature willnever bring forth life Comets and meteorites containorganic compounds (Deamer 1985) as Oro (1961 1994)suggested as a possible source of some of lifersquos buildingblocks but it is not obvious how to get such compoundsto the necessary concentrations to generate the chemistryof life nor how to select those few active simple multi-purpose building blocks that make up the basis of life froman interplanetary chemical supply house

The severity of the problem concerning the differ-entness of archaebacterial versus eubacterial membranelipids and cell walls is with few exceptions (see Koga etal 1998) not as widely recognized as it perhaps shouldbe What is generally recognized as difficult is the step tocellular organization (Bernal 1960 De Duve 1991 Oro1994 Segret amp Lancet 2000) that is the origin of bonafide membrane-bounded cells Coacervates or spon-taneous self organization of abiotically formed lipidsthrough whatever prebiotic chemistry is imaginable(Deamer 1986 Oberholzer et al 1995) but what is notimaginable is putting inside that first little lipid droplet


exactly the right mixture of self-replicating molecules pos-sessing all that is needed (a source of energy for exampleonce isolated from the surrounding aqueous phase) tomake more of themselves including the lipids in whichthey would have been encased

Getting the universal ancestor into the membranous orother cloak that it has to have at some time under all mod-els for the origin of life and the origin of cells poses seem-ingly insurmountable problems In models for the origin ofcells the membranes that surround living systemsmdashunderwhatever chemical premise (Calvin 1969) and if they areaddressed at all (Deamer 1985 1986)mdashjust seem to comefrom thin air and the ultimate contents of cells so derivedhave to arise as a free-living cytosol that needs to getinside even in well-argued cases (Segret amp Lancet 2000)

Indeed if we look at all truly living things which arecomposed of cells (that is if we exclude viruses and thelike which are derived from cells upon which they dependfor replication) the principle of semi-permeable compart-mentation from the aqueous environment is even morestrictly conserved than the universal genetic code becausethere are rare deviations from the universal code (Hao etal 2002 Srinivasan et al 2002) but there are no excep-tions whatsoever to the principle of semi-permeable com-partmentation from the environment Furthermore anattribute of life that all cells possess (but viruses do not)is the capacity to perform redox chemistrymdashthe transferof electrons from an external donor to an availableacceptor In fact a very simple definition of a living systemmight be compartments separated from their surround-ings that spontaneously multiply with energy gleanedthrough self-contained thermodynamically favourableredox reactions If nothing about life is more conservedthan compartmentalized redox chemistry what was theprecursor compartment at the origin of life We have asuggestion

4 IRON MONOSULPHIDE 3D CELLS NOT JUST2D SURFACES

In 1981 360 Myr old hydrothermally formed iron sul-phide chimneys were reported from ore deposits near thetown of Silvermines Ireland (Larter et al 1981 Boyce etal 1983) These structures (figure 2a) now consist ofmostly of pyrite (FeS2) although they probably precipi-tated as the monosulphide (FeS) in the Carboniferous seaSome of the chimneys contain sphalerite (ZnS) Theywere formed through the hydrothermal exhalation ofmetal-bearing fluid originally derived from seawater thathad been convectively recycled through the crust (Bankset al 2002) Iron zinc and lead sulphides were precipi-tated as the hydrothermal fluid exhaled into a 60 degC brinepool containing bacteriogenic hydrogen sulphide (Boyceet al 1983 Russell 1983 Samson amp Russell 1987) to pro-duce the ore deposits found today These discoveries hadbeen inspired by the much hotter (ca 380 degC) mineral-laden black smoker fluids emanating from chimneys in theeastern Pacific photographed by Ballard amp Grassle (1979)(see also Spiess et al 1980) Although fundamentally simi-lar the chimneys developing at the hydrothermal vents atoceanic ridge crests were much larger than the centimetre-sized structures in Ireland Nevertheless the possibleevolutionary importance of both types of spring site as


100 mm

4020 mm 301012 mm

20 mm

(a) (b)

(c)

(d)

Figure 2 Iron monosulphide precipitates (a) A 360 Myr oldhydrothermally formed iron sulphide chimney fromSilvermines Ireland (Boyce et al 1983) (b) Electronmicrograph of a thin section of a 360 Myr old pyriteprecipitate found at the Tynagh ore deposit Ireland (Banks1985) (c) Enlargement of (b) (d ) Electron micrograph of astructure formed in the laboratory by injecting Na2S solution(500 mM representing hydrothermal fluid) into FeCl2solution (500 mM representing the iron-bearing Hadeanocean) (Russell amp Hall 1997) High concentrations wereused to produce an examinable structure Similar structuresare produced at submarine hydrothermal vents by porousclay (Geptner et al 2002)

chemically reactive lsquohatcheriesrsquo for the origin of life ratherquickly became a popular notion (Corliss et al 1981Russell et al 1988) although not without staunch critics(Miller amp Bada 1988)

Picking up on a long biochemical tradition of thoughtsconcerning the suspectedly primitive nature of redox reac-tions catalysed by iron sulphur centres and their possible


the universal ancestor(not free-living)

DNA era

RNP era

RNA era

prebioticchemistry

geochemistry

HS_

H2CO CH3SH

CN_

CH4

NH3H2 CO

hot reduced alkalinehydrothermal solutionca100 C pH ca10

crust

ocean

Ni2+

Fe2+gtgtFe3+

CO2

H+

P2O7

temperature- redox-and pH-gradient

cooler more oxidizedmore acid Hadean ocean

lt30 C pH ca55

amino acidssugars bases

reducedcarbon

compoundsCH3COSR

organic precursors

4_

FeS

NiS

deg

deg

Figure 3 A model for the origin of life at a redox pH and temperature gradient at a submarine hydrothermal vent SeeRussell amp Hall (1997) and Russell et al (2003) for details The terms RNA RNP and DNA era (instead of lsquoworldrsquo) are usedto emphasize that no nucleic acid evolution is possible without a supporting geochemistry later biogeochemistry and finallybiochemistry to provide a steady flow of adequate concentrations of polymerizeable precursors (for example nucleotides) andthus to underpin any sort of replication

role in early metabolism (Hall et al 1971 Marczak et al1983) but radically departing from notions of organicsoup electric discharge or photocatalysis for the origin ofthe first biomolecules Wachtershauser (1988a) suggestedthat redox reactions between iron sulphide mineralsresulting in FeS2 (pyrite) formation could have providedthe free energy and electrons needed for primordial carbonfixation More fully developed as a theory of surface


catalysis (Wachtershauser 1988b) this work laid down thecase for a chemoautotrophic origin of life but it is notwithout its problems The reduction of CO2 with an FeS-H2SFeS2 redox couple was shown to be a far less ener-getically favourable process than originally suggested byWachtershauser (Schoonen et al 1999) More importantlythe surface catalysis model still failed to solve the problemof how to achieve sufficient concentrations of reacted pro-


ducts to get life off the ground Criticizing the notion of2D life on surfaces and referring to the fate of reactedproducts on surfaces De Duve amp Miller (1991 p 10 015)noted lsquoThese once detached are considered irretriev-ably lostrsquo

At that point geochemists studying iron sulphide pre-cipitates as they arise and occur in nature took careful noteof these developments and suggested on the basis of goodevidence a different lsquohelliprole for iron sulphide that ofnucleating the membrane of the earliest cell wallsrsquo (Russellet al 1990 p 387) An important part of that good evi-dence is shown in figure 2 Iron monosulphide as it nat-urally precipitates from submarine hydrothermal exhalatesforms striking structures (Russell 1983) Figure 2b showsan electron micrograph of a thin section of a 360 Myr oldpyrite precipitate found at the Tynagh ore deposit Ireland(Banks 1985) and reveals a highly compartmentalizedinner fabric An enlargement of figure 2b is shown in fig-ure 2c which shows a spectrum of internal compartmentsthat are incompletely separated from one another bymembrane-like sheaths of iron sulphide precipitate

Iron sulphide structures of this type are familiar toscientists in the field of economic geology but it is veryworthwhile for biologists and biochemists to have a goodlook at such images because most biologists and chemistswould imagine an FeS precipitate as a mere clump of min-eral with no internal structure at the extreme a cube withsix faces for surface chemistry

Naturally formed geologically occurring 3D FeS struc-tures of this type are extremely relevant to the origins oflife As discussed in great detail (Russell amp Hall 1997)they provide a very plausible solution to the problem ofachieving concentrations of reactants synthesized throughFeS chemistry that are sufficient to generate a moleculepopulation of the complexity that prebiotic chemistryrequires if it is ever going to evolve life For example ironsulphides are known to have high affinities for organopho-sphates cyanide (Woods 1976 Leja 1982) amines(Wark amp Wark 1935) and formaldehyde (Rickard et al2001)

However the Tynagh and Silvermines chimneys (figure2andashc) were formed from ca 200 degC exhaling fluid contain-ing dissolved iron zinc and lead which precipitated as itmet cooler (ca 60 degC) sulphide-bearing brine during thelate Devonian At the origin of life some 4 Gyr ago theconditions were somewhat different as Fe21 concen-trations in the oceans were high and the exhaling fluidwould have contained hydrosulphide (HS2) (Russell ampHall 1997) Simulating Hadean conditions in the labora-tory by injecting Na2S solution (approximating hydrother-mal fluid) into FeCl2 solution (approximating the iron-bearing ocean) resulted in highly structured iron monosul-phide precipitates (Russell amp Hall 1997) an electronmicrograph of which is shown in figure 2d Of course theprecise conditions at seepage sites which existed 4 Gyrago and whether they would produce exactly such struc-tures is not known But deep ocean hydrothermal convec-tive currents and their seepage sites should have existedin the Hadean (Nisbet amp Sleep 2001) and strikingly struc-tured clay precipitates have been reported at active sub-marine seepage sites in Iceland (Geptner et al 2002) It ishardly unreasonable to assume that metal sulphides would


have formed at seepage site hydrothermal mounds in theHadean

5 ON THE ORIGIN OF LIFE AND CELLS

Founded in geology geochemistry and thermodyn-amics a model for the (chemoautotrophic) origin of lifewas proposed by Russell amp Hall (1997) the salientelements of which are as follows (i) life emerged as hotreduced alkaline sulphide-bearing submarine seepagewaters interfaced with the colder more oxidized moreacid Fe21 Agrave Fe31-bearing water at deep (ca 4 km) floorsof the Hadean ocean ca 4 Gyr ago (ii) the difference inacidity temperature and redox potential of these watersprovided a gradient of pH (ca four units) temperature(ca 60 degC) and redox potential (ca 500 mV) at the inter-face of these waters that was sustainable over geologicaltime-scales providing the continuity of conditions con-ducive to organic chemical reactions needed for the originof life (iii) the exhalate was part of a typical hydrothermalsystem that brought water from the ocean floor moder-ately deep into the crust (ca 5 km) where it was subjectedto very high pressures and temperatures and thus con-tained outgassing CO H2 N2 along with reduced nitro-gen (NH3 CN2) (Schulte amp Shock 1995) reducedcarbon (CH3COO2 H2CO short alkyl sulphides) CH4

(Kelley 1996 Appel et al 2001) and HS2 which arehardly unreasonable premises and which were individuallycited and justified on geochemical grounds (iv) colloidalFeS precipitate formed at the ocean floor interface wasinflated by the emergent exhalate to produce structurediron monosulphide precipitates possibly similar in mor-phology to those shown in figure 2 and (v) the sustainedredox gradient of about half a volt across the walls of theFeS precipitate which of course would have containedNiS and sulphides of other metals present in the Hadeanocean as well provided a persistent source of electrons forprebiotic chemistry in addition to catalytic surfaces ofgreat area and most importantly provided a continuouslyforming (lsquogrowingrsquo) network of 3D compartments withinwhich reactants could remain concentrated A slightlyadapted version of Russell amp Hallrsquos (1997) model ispresented in figure 3 Notably a proton-motive forcefounded in a natural pH gradient exists across the outerFeS membrane

Only few minor modifications with respect to the orig-inal model have been introduced here One of these con-cerns the possible temperature of the exhalate (ca 70ndash100 degC instead of ca 150 degC) and the temperature of theocean floor water (ca 0ndash30 degC versus ca 100 degC) (Russellet al 2003) Modern hydrothermal exhalates can be hotexceeding 350 degC or much cooler for example 70 degC(Kelley et al 2001 Geptner et al 2002) whereby the lat-ter have a pH of 98 or more Owing to the very poorstability of critical building blocks of life such as RNAtemperatures closer to 50 degC seem more conducive to theassembly and sustained stability of early organic moleculesand self-reproducing systems (Moulton et al 2000Bada amp Lazcano 2002) Thus it seems more reasonable toassume more moderate temperatures for the hydrothermalexhalate Regarding the temperature of the ocean there isconsiderable uncertainty Because the sun was much dim-mer 4 Gyr ago than it is now (Newman amp Rood 1977


Sagan amp Chyba 1997) it is possible that the Hadeanoceans were intermittently frozen in global glaciation inwhich case the ocean floor temperature would have beenvery cold or that they were fluid in a much warmer cli-mate due to high levels of greenhouse gases such as CO2CO and CH4 in the atmosphere (Nisbet amp Sleep 2001)The model would merely require the ocean temperatureto be below that of the exhalate A local gradient of tem-peratures exists at all submarine springs and seepages

Another change involves the removal of the reverse cit-ric acid (TCA or Krebs) cycle as an early CO2 fixationpathway Wachtershauser (1988b) made a strong case forthe reverse TCA cycle being perhaps the first biochemicalpathway based on its construeable similarity to FeSchemistry and thermodynamic attributes But like any cyc-lic pathway of CO2 fixation such as the Calvin cycle 3-hydroxypropionate pathway or pentose monophosphatecycle the reverse TCA cycle requires pre-existing stereo-chemically correct chemical intermediates of the cycle toget started A case can be made for the view that formolcondensation-like reactions could have fuelled early CO2-fixation (Russell et al 2003) However the acetyl-CoApathway of CO2 fixation better known to many as theWoodndashLjungdahl pathway is also very attractive as astarting point of metabolism because it is linear usingelectrons from H2 to reduce 2 mol of CO2 to produce anacetyl moiety that is transferred to the thiol group of coen-zyme A (Ljungdahl 1994) In the CO2-fixing directionthe pathway has the net reaction

2CO2 1 8[H] 1 CoASH CH3COSCoA 1 3H2O

with estimated thermodynamic values of DG9o= 2592 kJ mol21 if 2[H] = H2 or DG9o = 1132 kJ mol21 if2[H] = NADH (Fuchs 1994) The pathway is flexible andoccurs among various eubacteria and various archaebac-teria Among other things it can be (i) the basis ofchemoautotrophy in sulphate-reducing eubacteria andarchaebacteria in methanogens and in homoacetogenssuch as Clostridium thermoautotrophicum when growingchemoautotrophically (Ljungdahl 1994) (ii) the source ofchemiosmotic potential in homoacetogens when growingfermentatively (Ljungdahl 1994) (iii) a pathway of acetyl-CoA oxidation if CO2 and H2 levels are kept extremelylow through consumption by other microbes (iv) a path-way of acetate disproportionation to H2 and CO2 in ace-toclastic methanogenesis (v) a pathway for theassimilation of various C1 compounds and (vi) a sourceof electrons for carbon fixation through oxidation of COto CO2 (reviewed in Fuchs 1994 Ragsdale 1994 Ferry1995)

The key enzyme of the acetyl-CoA pathway is anextremely intriguing protein called ACSCODH(Ljungdahl 1994 Ragsdale 1994 Ferry 1995 Lindahl2002) In the CODH domain ACSCODH employs fiveFeS clusters including an unusual 4FeNi5S cluster thatbinds CO2 and reduces it to CO As part of the proposedreaction mechanism the ACS domain of ACSCODHcondenses Ni-bound methyl and Ni-bound CO yieldingan Ni-bound acetyl moiety that is transferred to CoASHyielding acetyl-CoA (Lindahl 2002) Emulating the ACSCODH enzymatic reaction Huber amp Wachtershauser(1997) showed inter alia that micromolar amounts ofmethyl thioacetate (CH3COSCH3) could be formed over-


night without pressure at 100 degC from 45 mmol CO and100 mmol CH3SH (Heinen amp Lauwers 1996) in thepresence of FeS and NiS making this type of reaction inFeS precipitates extremely plausible as a prebiotic sourceof basic carbon building blocks However the experimentproduced no pyrite CH3COSCH3 contains a thioesterbond and can be considered as homologous to the busi-ness end of acetyl-CoA In most organisms that use theacetyl-CoA pathway today the next step of CO2 fixationis catalysed by pyruvate synthase also called PFOR anFeS protein that combines the acetyl moiety of acetyl-CoAwith CO2 and two electrons from ferredoxin anotherFeS protein to yield pyruvate (Schoenheit amp Schafer1995 Yoon et al 1999) At very high temperature andpressure traces of pyruvate were synthesized from COand alkyl thiols by FeS catalysts (Cody et al 2000) Incontrast to other views (Wachtershauser 1988b Huber ampWachtershauser 1997) we see no need for the presenceof cyclic CO2 fixation pathways at this stage because alinear route of carbon fixation would suffice given con-tinuous geochemical input

There is a vast literature regarding plausible models forprebiotic syntheses of sugars (formol condensation(Quayle amp Ferenci 1978 Muller et al 1990)) bases(cyanide condensation (Oro amp Kimball 1962 Sanchez etal 1967)) amino acids (Hennet et al 1992 Marshall1994) and other molecules (Sutherland amp Whitfield1997) and a vast literature on the broad catalytic capabili-ties of RNA (Bartel amp Unrau 1999 Gesteland et al 1999)none of which we can deal with adequately here We sur-mise that it is conceivable and not implausible that usingsuch conceptual tools one could get to something like anRNA-world (Gilbert 1986) provided that its monomericbuilding blocks are steadily replenished and remain in suf-ficient concentrations conditions that structured FeS pre-cipitates could fulfil because of their sustained energyinput though the natural redox gradient and by their com-partmentalizing reactant-retaining nature (Russell ampHall 1997)

6 HOLOCHIRALITY

Another addition to the proposal of Russell amp Hall(1997) is that an unresolved problemmdashthe holochiralityproblemmdashneeds to be addressed (see also Russell et al(2003) for another perspective on this issue) Life todayconsists of l-amino acids and d-sugars but the moleculesat the very beginning of life almost certainly did not Ithas been argued that asymmetric surfaces could give riseto stereospecific prebiotic synthesis (Hazen et al 2001)but given the surfaces of iron sulphide we find it uncon-vincing that one or the other stereoisomer of either classshould tend to be synthesized at early stages Thus ourmodel up to this point could in principle lead to a sus-tainable and compartmentalized RNA-world but a badlyscrambled primitive one of racemic sugars coexisting witha racemate of amino acids and everything else A possibleway out of the racemic world could have involved the ori-gin of the peptidyl transferase reaction of the large riboso-mal subunit Among the myriad of molecules that existedin the RNA-world a population of them was involved inthe origin of translation whereby it has been sensibly sug-gested that the original function of the ribosome was RNA


replication via triplets not translation (Poole et al 1999)We can take it as a given that at some point a precursorof the peptidyl transferase reaction (Muth et al 2000) ofan RNA molecule did evolve and the chance stereochemi-stry of that molecule would determine the decision ofwhether it could catalyse the polymerization of d- or l-amino acids into peptides That filter would be sufficientto tip the holochirality scale because despite the presenceof a racemate only amino acids of the same a-carbon con-figuration would preferentially end up in peptides yieldinga population of distinctly handed peptides some of whichwould eventually need to feed back in a hypercycle-likemanner (Eigen amp Schuster 1978) into favoured synthesisof their stereochemically correct polymerizing template

Concerning d-sugars if we assume that the FeS-cata-lysed synthesis of acetyl-CoA and pyruvate was possibleas above then the origin of chirality should be sought atthe three-carbon stage Pyruvate is not chiral nor is itsphosphorylated derivative phosphoenolpyruvate Inorganisms that use the acetyl-CoA pathway the next stepof autotrophic carbon metabolism involves the stereospec-ific addition of a water molecule to the double bond inphosphoenolpyruvate yielding 2-phospho-d-glycerateToday that reaction is catalysed by the enzyme enolasewhich is by far the most highly conserved of all glycolyticenzymes (figure 1) and as far as we are aware is the onlyenzyme of core carbon metabolism that is truly ubiquitousamong all free-living organisms For the whole remainderof central carbon metabolism in all organisms the stereo-chemistry at C2 of 2-phospho-d-glycerate never changesThis approach to the holochirality problem implicates thepeptidyl transferase reaction of the ribosome as the pace-maker for amino-acid stereochemistry and a simpler pre-cursor of enolase as the pacemaker of sugarstereochemistry It is probably just curious coincidencethat enolase is the only enzyme of carbon metabolismother than adenylate kinase that balances AMP ADP andATP levels that is encoded in the highly conserved riboso-mal protein superoperon in archaebacteria (Hannaert etal 2000) which might also carry a faint trace of chiralorigins At any rate our suggestion for the origin of holo-chirality starts with the peptidyl transferase reaction andis thus largely congruent with Woesersquos (2002 p 8745)assessment lsquoThe evolution of modern cells then had tobegin with the onset of translationrsquo It was certainly adecisive step

One particularly interesting thought emerges from theforegoing namely that the primitive peptides synthesizedby the kind of a universal ancestor shown in figure 3(particularly the sulphhydryl-containing ones) would haveprovided a chiral proteinaceous surface associated withan FeS (or other metal sulphide) centre This would haveprovided dramatic synergy for biochemical diversity andshould have initiated an era of biochemical discovery withDarwinian backcoupling (via translation) to the encodingnucleic acids The very clear prediction of this notion isthat experiments of the type performed by Heinen ampLauwers (1996) or Huber amp Wachtershauser (1997)when supplemented with small synthetic holochiral pep-tides (random mixtures of varying complexity) shouldcatalyse the synthesis of a much more diverse spectra ofbiologically relevant molecules than control experimentslacking the synthetic peptides Experiments of this type


might provide insights as to how biochemistry evolvedand might even reveal various primitive analogues of con-temporary active centres

7 THE UNIVERSAL ANCESTOR THE LASTCOMMON ANCESTOR OF PROKARYOTES ONLY

By the time that a ribosome had evolved a genuineribonucleoprotein world the step to a DNA world and thelast common ancestor would not be far off But here wehave to depart completely from some long-held and widelyregarded views on the early evolution of cells (Woese2002) because the weight of evidence precludes the exist-ence of eukaryotes anywhere near the origin of cells Thereasons for this are at least threefold

First without question eukaryotes are ancestrally heter-otrophs the only autotrophy in the entire group havingbeen acquired through endosymbiosis during the origin ofplastids All eukaryotes that lack plastids and some thatpossess plastids are heterotrophs and without exceptionthe backbone of their ATP synthesis is the oxidativebreakdown of reduced carbon compounds through theglycolytic pathway (Martin amp Muller 1998) Embracingthe view of a chemoautotropic origin of life the firstorganisms necessarily had to be chemoautotrophs(Kandler 1998) and heterotrophy as a microbial lifestylecould only have arisen after autotrophs had evolved ametabolism sufficient to support the free-living lifestyleand had subsequently produced sufficient quantities ofreduced carbon compounds in the environment to provideheterotrophsmdashprokaryotic and eukaryoticmdashwith some-thing to eat

Second all of the eukaryotes whose origin needs to beexplained (contemporary ones) seem to have possessed amitochondrion in their evolutionary past (see Embley ampHirt 1998 Roger 1999 Williams et al 2002) The originof eukaryotes is thus hardly separable from the origin ofmitochondria (Gray et al 1999 Lang et al 1999) Thiswould mean that the origin of eukaryotes would necessar-ily post-date the diversification of the a-proteobacteriafrom which the mitochondrion is thought to descendmeaning that eukaryotes must substantially post-date pro-karyotes in origin

Third if we look beyond the diversity of the geneticsystem and consider the diversity of metabolic pathwaysthat prokaryotes use to synthesize the ATP that is thebackbone of their survival then we are faced with themore than 150 different redox reactions involving inor-ganic donors and acceptors used by the thermophilic pro-karyotes alone for their core energy metabolism (Amend ampShock 2001) versus the complete diversity of eukaryoticenergy metabolism which does not comprise more thanone single pathway of glucose oxidation to pyruvate andthree basic metabolic pathways of subsequent pyruvatemetabolism (Martin amp Muller 1998) The full diversity ofeukaryotic energy metabolism including the diversity ofenzymes catalysing the individual steps (figure 1) does notexceed that of a single generalist a-proteobacterium Withsuch a narrow biochemical diversity at the root of theirATP synthesis it seems completely at odds with commonsense to assume that eukaryotes are anywhere nearly asold as prokaryotes Prokaryotes are easily seen as hold-overs from the phase of evolution where novel biochemical


routes to glean energy from the environment were stillbeing invented and many such relics can be found amongprokaryotes today The argument by Kandler (1998) thatthe complete lack of chemoautotrophy among eukaryotesprescribes their origin subsequent to the origin of prokary-otes is compelling But the question of how eukaryotesand their defining features might have arisen is a differentmatter to be discussed in a later section

There are no eukaryotes that can live from the elementsalone without the help of prokaryotes (or prokaryoticendosymbionts) By contrast there are many eubacteriaand archaebacteria that can live from the elements alone(chemoautotrophs) and it is hence eminently reasonableto assume that organisms able to live from the elementsalone arose before those organisms that cannot The con-verse argument that eukaryotes arose before prokaryotes(as for example argued in Forterre amp Philippe (1999)) iscompletely unreasonable from the standpoint of microbialphysiology and is only tenable if one looks at genetic sys-tems and genetic systems only leaving microbial physi-ology aside But physiology is just as important for earlyevolution as genetic systems are because no genetic sys-tem operates without a physiological system to support itwith energy in the form of ATP and precursors for doingthe things that genetic systems do Arguing that eukary-otes predated prokaryotes which underwent some mys-terious reductive process is like arguing that land animalspredated land plants which nourished themselves fromanimal excrements The most fundamental differencebetween prokaryotic and eukaryotic physiology from thestandpoint of energy metabolism is illustrated in figure 4where only a very small sample of the various redoxcouples used by prokaryotes as summarized by Amend ampShock (2001) are contrasted to the entire diversity ofenergy metabolism found in eukaryotes (see also Schaferet al (1999) for an overview of energy metabolism inarchaebacteria) Clearly under our premise embracing achemoautotrophic origin of life eukaryotes simply cannothave arisen at even approximately the same time as proka-ryotes At any rate common sense and a bit of biochemi-cal reasoning dictates that eukaryotes must have arisenlong after prokaryotes did as most biologists have alwaysreasonably assumed and as the microfossil record indi-cates

Our model for the origin of life to this point has specifi-cally addressed the arguably most important problems thatearly self-organizing and self-replicating systems facednamely (i) where did the energy come from that is neededto continuously synthesize the building blocks of evolvinggenetic systems (natural redox and pH gradient) and (ii)how did the chemical constituents of life manage to staysufficiently concentrated to permit any kind of replicationto occur (structured FeS precipitates)

The universal ancestor as depicted in figure 3 was inour view an organismmdashthough not a free-living onemdashthatshould have possessed all of the attributes that are com-mon to all eubacterial and archaebacterial chemoauto-trophs the genetic code the ribosome DNA asupporting core and intermediate metabolism (nitrogenfixation modern amino-acid nucleic-acid and cofactorbiosyntheses) needed to supply the constituents of its rep-lication (doubling of mass) compartmentation from theenvironment redox chemistry and ATPases which could


have used the naturally existing proton gradient (alkalineinside acid outside) at the seepage site

This non-free-living universal ancestor would havefound itself at the dawn of the biochemical revolutionwhere genes and proteins were diversifying into a myriadof functions where RNA and metal sulphide catalystswere being replaced by proteins where new pathways andcofactors were being invented to augment and substitutetheir mineral and RNA precursors where FeS from themother lode was being incorporated into proteins as FeSclusters and where biochemistry started to diversify intothe forms that were both possible and useful From thestandpoint of protein structure this age of inventionwould have witnessed the origin of basic building blocksof biochemical function that (i) are conserved at the levelof a 3D structure among archaebacteria and eubacteriaand (ii) are recognizable as modules of function in variouselectron-transporting proteins as discussed by Baymann etal (2003) From the standpoint of amino-acid sequenceconservation this age of invention would have been aphase of molecular evolution where proteins were diver-sifying (eg Habenicht et al 1994) and getting better atwhat they do rather than not getting worse at what theydo as is thought to be the case for the evolution of todayrsquosproteins Put another way most modern proteins arethought to accept amino-acid substitutions because thenew amino acid does not significantly impair the pre-exist-ing function (neutral or nearly neutral molecularevolution) At the very beginning of protein evolutionpositively selected mutations that improve the function ofthe protein might have been the rule rather than theexception because without advantageous mutationsadaptive evolution cannot occur (Nei 1987) Accordinglythe mechanistic basis of sequence divergence among themost ancient proteins (intense positive selection) woulddiffer fundamentally from that in recently diverged pro-teins (near neutrality)

8 FROM A NON-FREE-LIVING UNIVERSALANCESTOR TO FREE-LIVING CELLS

At some point this biochemical diversification wouldhave included the invention of pathways that catalyse thesynthesis of lipids and cell wall constituents (figure 5)both of which would have begun to insulate the cytosolfrom its FeS encasement requiring a new category of pro-teins membrane-associated proteins involved in redoxchemistry These would have been rich in FeS (and othermetal sulphide) centres so as to replace the functions orig-inally associated with the original FeS redox chemistrygermane to the original FeS lsquocell wallrsquo As a consequencethe catalysis performed by (Fe Agrave Ni)S (Russell et al2003) would have been gradually separated from the min-eral wall and step-by-step incorporated into proteins animprint of which would be reflected in the FeS centres ofancient proteins under this view as schematically indi-cated in figure 6 Protein modules such as ferredoxins andhydrogenase are very common in todayrsquos membrane-asso-ciated electron transport chains (Baymann et al 2003)Only when a sufficient battery of redox carriers hadbecome assembled into the membrane to allow the com-partments to sustain ATP synthesis via chemiosmosis out-side the confines of their 3D (Russell amp Hall 1997 see also


Figure 4 A comparison of a narrow segment of the redox reactions involving inorganic donors and acceptors that a narrowsegment of prokaryotes use for sustained ATP synthesis (excerpted from Amend amp Shock (2001)) versus the entire breadth ofreactions that non-photosynthetic eukaryotes in toto use for sustained ATP synthesis Note that there are rare cases wheremitochondria can operate chemotrophically but only transiently not for sustained ATP production (Doeller et al 2001)

Koch amp Schmidt 1991) naturally chemiosmotic housingwould they have been able to make first attempts at lifein the wild With the invention of pathways for biosynth-esis of redox cofactors quinones hemes and other elec-tron-transporting centres these early membranes might


have begun to look reasonably similar to modern prokary-otic membranes at least in terms of basic architecture

Notably our model for the independent origins ofmembrane-bounded compartmentation in archaebacteriaand eubacteria predicts that the electron- and proton-


free-living cells(chemoautotrophs)

escape with conservationof redox chemistry

at cell surface

independent inventionof cell wall biochemistry

independent inventionof lipid biosynthesis

invention of mostbiochemistry


eubacteria

archaebacteria

Fe2+

H+

Fe3+

Ni2+

HCO3_

Figure 5 A model for the origin of membrane-bounded prokaryotic cells from iron monosulphide compartments within whichthe chemoautotrophic origin of life could have occurred The drawing implies (but does not show) very large populations ofreplicating systems and unspecified physical distances (within however a single submarine seepage site) between diversifyingreplicating systems en route to free-living eubacterial and archaebacterial cells Various kinds of genetic exchange that havebeen repeatedly postulated (Woese 2002) before the origin of a truly membrane-bounded cellular organization of prokaryotescould be easily accommodated by the model Note that iron monosulphide precipitates are initially colloidal (Russell amp Hall1997) and inflated to the general types of structure shown in figure 2 by exhaling hydrothermal fluid

transporting quinones of their independently evolvedmembranes should also differ substantially That predic-tion fares quite well The quinones of both groups of pro-karyotes possess isoprene side chains (Schutz et al 2000Berry 2002) However these isoprenesmdashlike the lipidsthemselvesmdashare synthesized by different completely


unrelated biochemical pathways in the two groups Eubac-teria synthesize IPP via the DXP pathway which startsfrom d-glyceraldehyde-3-phosphate and pyruvate via adecarboxylating transketolase reaction (DXP synthase)whereas archaebacteria synthesize IPP via the MVA path-way which starts with the condensation of three molecules


[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin

carbon monoxodedehydrogenase

fumarate reductase

pyruvate ferredoxin oxidoreductase

Figure 6 The position of FeS centres in some proteins thatcatalyse redox reactions Figures were sketched from [Fe]-hydrogenase (Peters 1999) PFOR (Chabriere et al 1999)[4Femiddot4S] ferredoxin (Macedo-Rieiro et al 2001) [2Femiddot2S]ferredoxin (Kostic et al 2002) CODH (Lindahl 2002) andfumarate reductase a membrane protein (Lancaster et al1999) Blue dots indicate active sites

of acetyl-CoA (Lange et al 2000) Not only do the iso-prene side chain pathways differ in the two groups theproton- and electron-transferring aromatic heads of thequinones also differ in archaebacteria and eubacteriaalthough some archaebacteria possess menaquinone andubiquinone like eubacteria others use sulphur-containingheterocyclics such as the benzothiophenes of Sulfolobusand Acidianus (Schutz et al 2000) whereby the methano-gens use methanophenazine which is not a quinone at all(Berry (2002) qv for quinone structural comparisons)Interestingly a key enzyme of the MVA pathway 3-hyd-roxy-3-methylglutaryl-CoA synthase is distantly related tob-ketoacyl-acyl carrier protein synthase III of eubacteria(Lange et al 2000) which catalyses a similar condensationreaction to produce the fatty acid precursor acetoacetyl-ACP from acetyl-ACP and malonyl-CoA as substratesindicating the presence of an ancestral acetyl-CoA con-densing protein in the non-free-living universal ancestorFurthermore DXP synthase of the DXP isoprenoid routeis distantly related to PFOR a ferredoxin-dependentenzyme catalysing the C2 C3 step subtending the acetyl-CoA pathway in many chemoautotrophs indicating thepresence of this primitive C2-transferring step in the non-free-living universal ancestor Because both DXP synthaseand PFOR require TPP in their C2-transferring catalysisthe prediction clearly follows that functional andor struc-tural analogues of the TPP coenzyme should be obtainablefrom mineral catalysts under Huber amp Wachtershauser


(1997) type conditions in the presence of suitablenitrogenous compounds

Under this view the evolutionary basis of the attributesshared by prokaryotes would be founded in their inherit-ance from the last common ancestor The evolutionarybasis for the fundamental differences that distinguish eub-acteria and archaebacteria but which more or less uni-formly unite the respective groups could be sought in theage of biochemical discovery during the persistence ofproto-eubacterial and protoarchaebacterial lineages livingwithin the confines of their energy-laden mineral incu-bator Such differences would include their use of manyfundamentally different cofactors such as pterin deriva-tives versus tetrahydrofolate coenzyme B methanofurancoenzyme F420 coenzyme M and cobamids in manyarchaebacteria (see DiMarco et al 1990 Thauer 1998White 2001) their use of different pathways for the samemetabolic intermediates such as IPP (Lange et al 2000)or their differences in purine biosynthesis (White 1997)their use (archaebacteria) or disuse (eubacteria) of smallnucleolar RNAs for the modification of ribosomes (Omeret al 2000) the differences between their DNA mainte-nance and repair machineries (Tye 2000) the differencesbetween their transcriptional regulatory apparatus(Thomm 1996 Bell et al 2001) and RNA polymerases(Langer et al 1995 Bell amp Jackson 1998) their quinones(Schutz et al 2000 Berry 2002) or and what is probablythe most important point for this paper the differencesbetween their membrane lipid biosynthetic pathways andcell-wall constituents (Kandler 1982) Many other suchdifferences could be listed The assembly of FeS-clustersinto proteins an ancient biochemical function (Tachezyet al 2001) would have been among those attributesessential to both lineages a function possibly provided bythe recently discovered SUF pathway of FeS clusterassembly (Takahashi amp Tokumoto 2002) Efficient meansof N2-reduction perhaps through primitive molybdenumndashiron vanadiumndashiron or iron-only nitrogenases (Chien etal 2000) would have also been prerequisite to the tran-sition to a free-living habit

Once the crucial step of membrane synthesis and a tur-gor-resistant cell wall had been taken independently thetwo kinds of prokaryotic stem lineage could have begunmaking their attempts at the transition to the free-livingstate (independently) and begun trying their luck at lifeasmdashchemoautotrophicmdashfree-living cells (figure 5) If onecell can escape so can the next and the next continu-ously until the well at the seepage site ran dry at whichpoint the origin of life would have been over Those free-living cells that possessed or invented the biochemicaltools needed to survive outside their initial confines wouldhave found themselves in a very standard kind of Darwin-ian struggle but one with a particular reward for ingenuityat tapping new kinds of inorganic redox resources in anuninhabited competitorless ocean-floor world Thiswould have been a time where much of the oxygen-inde-pendent metabolic diversity among the prokaryotes mighthave arisen (figure 5) The age of biochemical discoverywould have found a veritable gold mine at the oceansrsquossurface and the origin of chlorophyll-based photosyn-thesis in the eubacteria emerged apparently quite early(Dismukes et al 2001 Russell amp Hall 2001 Xiong ampBauer 2002) With that the prokaryotic world would have


been off and running changing the chemistry of theoceans and eventually supplying enough reduced carbonsediment to allow the origin of heterotrophic prokaryoticlifestyles At some early time oxygenic photosynthesiswould have been providing low local concentrations of anew and powerful electron acceptor (O2) with which earlymicrobes had to cope probably using exactly the samestrategies as modern organisms either through the use ofoxygen-detoxifying enzymes (Chapman et al 1999 Jennyet al 1999) through the avoidance of oxygen-containingniches or through the gradual adaptation to oxygen har-nessing its utilitous oxidative power as many oxygen-dependent microbes (figure 4) have done

9 EUKARYOTES CELLS WITH SOMETHING INSIDE

The foregoing model for the origin of cells (prokaryoticones) embraces the view of a chemoautotrophic origin oflife and metabolism in general and views organic chemis-try catalysed by FeS (and NiS) as ancient and ancestralin biochemistry It accounts for the attributes shared byeubacteria and archaebacteria as inheritances from theirnon-free-living common ancestor (figure 5) and accountsfor the differences of eubacteria and archaebacteriathrough divergence from that common ancestor In thissection we address the evolutionary basis for the attri-butes (i) that are shared by eubacteria and eukaryotes (ii)that are shared by archaebacteria and eukaryotes and (iii)that distinguish eukaryotes from prokaryotes

Thus this section boils down to the question of the ori-gin of eukaryotes one of biologyrsquos messiest problems Tointroduce the issue as briefly as possible one can try todistill the essence of what Otto Kandler and Carl Woesewho discovered the distinctness of archaebacteria fromother then-known cells and who proposed the three-domain classification of life (Woese et al 1990) haverecently written on the topic Considering the charactersshared by various lineages such as those considered byZillig (1991) Kandler (1998 p 20) wrote

hellipeach of the three statistically possible pairs among thethree domains appears to be a sister group dependingon the basic phenetic characters considered The quasi-random phenetic mosaicism among the three domainsthat this indicated was confirmed at the genomic levelwhen the orthologous genes of the archaeon Methanoc-occus jannaschii of the eukaryote Saccharomyces cerevisiaeand of several bacterial species were aligned (Clayton etal 1997) While the majority of the genes of Methanoc-occus are shared with either bacteria andor eukaryotesonly a minority of genes seem to be confined to thedomain Archaea (Bult et al 1996) Such a quasi-randomdistribution of genes among the three domains cannotbe explained satisfactorily by any order of dichotomousbranching of a common ancestral cell or by chimaerismoriginating from fusion or engulfment among prokary-otes at early evolutionary stages as was suggested by vari-ous authors (Zillig et al 1992 Sogin 1994 Golding ampGupta 1995 Doolittle 1996 Margulis 1996) The distri-bution pattern rather calls for a revival of the previousproposal of a tripartition at an early pre-cellular stage oflife (Fox et al 1980) resulting in three dichotomouslybranched separate lineages forming a phylogeneticlsquobushrsquo rather than a tree


Firmly embracing the notion of a chemoautotrophic originof life the nagging matter of lacking chemoautotrophy ineukaryotes prompted Kandler to note (p 25)

The domain Eukarya contains neither any genuinechemolithoautotrophic nor thermophilic members grow-ing optimally above 50 degC [hellip] Hence founder groupsof the domain Eukarya most likely consisted of highlyderived mesophilic heterotrophic pre-cells adapted to analmost lsquomodernrsquo at least late Archaean or early Protero-zoic environment of a predominantly mesophilic tem-perature regime favourable for the evolution ofintracellular membrane systems typical of eukaryotes

Woese (2002 p 8742) while considering the immenseproblems of cellular evolution surmized

Initial attempts to frame the issue have typically been inthe classical Darwinian mode and the focus to date hasalmost been exclusively on modeling the evolution of theeukaryotic cell The reason of course is clearmdashtheappeal of the endosymbiosis concept Because endosym-biosis has given rise to the chloroplast and mitochond-rion what else could it have done in the more remotepast Biologists have long toyed with an endosymbiotic(or cellular fusion) origin for the eukaryotic nucleus andeven for the entire eukaryotic cell These classical expla-nations have three characteristics they (i) invoke cellsthat are basically fully evolved (ii) evolve the essentialcomponents of the eukaryotic cell well after its archaealand bacterial counterparts (as has always been connotedby the term lsquoprokaryotersquo) and (iii) focus attention oneukaryotic cellular evolution which implies that theevolutions of the lsquoprokaryoticrsquo cell types are of a differ-ent charactermdashsimpler and it would seem less interest-ing [emphasis from the original]

Both authors dismiss endosymbiosis as possible explana-tory principles to account either for the characters sharedby eubacteria and eukaryotes or to account for thosecharacters shared by archaebacteria and eukaryotes Theyalso dismiss fusion models as do we because fusion mod-els are exactly that they entail the fusion of two cells eachsurrounded by its own membrane into a single cytoplasmsurrounded by one membrane like putting two smallersoap bubbles together to get one bigger one Eukaryotesdo occasionally fuse cytoplasms (for example atfertilization) but prokaryotes do not and there is no rea-son to suspect that they ever did in the past either Butboth authors do something that is very common in theliterature they lump endosymbiotic models and fusionmodels for the origin of eukaryotes into one basket Anexample of an explicit fusion model can be found in theright panel of fig 3 in Lopez-Garcia amp Moreira (1999)where a free-living consortium of prokaryotes just startsfusing away uniting up their cytoplasms

Nobody doubts any more that the principle of endosym-biosis as elaborated by Mereschkowsky (1905) con-demned by Wilson (1928) and revived by Margulis (Sagan1967) is real and that chloroplasts and mitochondria aredescended from free-living prokaryotes But the endosym-bionts did not fuse with their host rather they remainedintact as membrane-bounded cell compartments withintheir host a general principle in cell evolution termedmembrane heredity (Cavalier-Smith 2000) Yet rejectionsof the notion that endosymbiosis might have given rise not


only to the mitochondrion but also to the bipartite cellwhich became a eukaryote are sometimes reminiscent ofWilsonrsquos (1928) rejection of the notion that mitochondriawere endosymbionts to begin with

More recently Wallin (rsquo22) has maintained that chond-riosomes [mitochondria] may be regarded as symbioticbacteria whose association with the other cytoplasmiccomponents may have arisen in the earliest stages ofevolution (p 712) To many no doubt such specu-lations may appear too fantastic for present mention inpolite biological society nevertheless it is within therange of possibility that they may someday call for moreserious consideration (p 739)

The question at the root of the problem on the origin ofeukaryotesmdashand the reason why biologists have to this daybeen unable to connect up archaebacteria eubacteria andeukaryotes into a unified scheme of cell evolution that isgenerally accepted by allmdashconcerns the host the cell thatacquired the mitochondrion Less fundamentally prob-lematical but none the less a decade-long problem forendosymbiotic theory are hydrogenosomes pyruvate-oxidizing hydrogen- and ATP-producing organelles ofanaerobic eukaryotes that were discovered by Muller(Lindmark amp Muller 1973) Long suspected to be a thirdkind of endosymbiotic organelle distinct from mitochon-dria and chloroplast and perhaps descended from clostrid-ial ancestors (Whatley et al 1979 Muller 1988) recentadvances have shown that they are simply an anaerobicmanifestation of the same endosymbiont that gave rise tomitochondria (Embley et al 1997 Van der Giezen et al2002)

On the bottom line there are only two possibilities forthe nature of the host that acquired the organelle ancestralto mitochondria and hydrogenosomes it was either a pro-karyote or a eukaryote The long-held view that the hostwas a eukaryote drew support primarily from the existenceof eukaryotes without mitochondria (Cavalier-Smith1987) and from the finding that they branched deeply inthe eukaryotic branch of the lsquouniversal treersquo of ribosomalRNA (Vossbrink et al 1987 Sogin et al 1989) But all ofthe mitochondrion-lacking and suspectedly primitiveeukaryotes that built the basis for traditional views on theorigins of mitochondria for example microspridiansdiplomonads entamoebids and the like have turned outeither to have possessed mitochondria in their past or tostill possess the organelle in reduced form (Embley amp Hirt1998 Roger et al 1998 Roger 1999 Muller 2000 Phil-ippe et al 2000) They are thus not direct descendants ofan amitochondriate host Furthermore the seeminglydeep branching of such organisms in the lsquouniversal treersquois in many if not all cases apparently just an artefact of treereconstruction (Embley amp Hirt 1998 Philippe amp Laurent1998) A prime example of this problem are microsporidi-ans organisms that started many people believing thatmitochondrion-lacking eukaryotes are both ancient andprimitive on the basis of their rRNA phylogeny (Vossbricket al 1987) But the microsporidia are in fact highlyderived fungi (Keeling 2001) as revealed by genomesequencing (Katinka et al 2001) and the seemingly deepposition of their rRNA sequence is a phylogenetic artefact

However the lsquogreat similarity of the archaeal and eukar-yal information processing systemsrsquo (Woese 2002 p


8744) is definitely not an artefact it is an incontrovertibleobservation founded in the comparative analysis of tran-scription and translation machineries Yet in which lightthis observation should be interpreted is a different matterwere the host a eukaryote or were it an archaebacteriumwe would arrive at the same observation Models for theorigin of eukaryotes and the nature of the host haverecently been reviewed elsewhere (Martin et al 2001)

The starting point for the remainder of this section wasthe study of the phylogenies of various enzymes of carbo-hydrate metabolism in higher plants starting withGAPDH (Martin amp Cerff 1986 Martin et al 1993) in thecontext of their homologues in eubacteria and archaebac-teria and extended to the complete glycolytic pathway theCalvin cycle gluconeogenesis the oxidative pentose phos-phate pathway the glyoxylate cycle and the citric acidcycle as reviewed elsewhere (Martin amp Schnarrenberger1997 Martin amp Herrmann 1998 Henze et al 2001Schnarrenberger amp Martin 2002) At the same time simi-lar studies were being conducted on the glycolyticenzymes from amitochondriate protists as reviewed byMuller (1998) For every glycolytic enzyme (exceptenolase) (Hannaert et al 2000) the same observation wasscored the eukaryotic enzymes are more similar to theireubacterial homologues than they are to their archaebact-erial homologues as shown once more in figure 1 for theglycolytic pathway and in many cases the archaebacteriado not even possess the homologous enzyme

To account for the findings (i) that eukaryotes possessan archaebacterial genetic apparatus (ii) that eukaryotespossess eubacterial ATP-producing pathways in mito-chondria in hydrogenosomes and in the cytosol (iii) thatmitochondria and hydrogenosomes descend from a com-mon ancestral organelle and (iv) that amitochondriateeukaryotes possessed a mitochondrion in their evolution-ary past a hypothesis was formulated which explored thepossibility that the host might have been an archaebacter-ium an anaerobic and autotrophic one with a carbon andenergy metabolism that was dependent upon molecularhydrogen as an electron donor (Martin amp Muller 1998)The microbial physiology of the model was based onphenomena abundantly observable among modern proka-ryotes hydrogen transfer and anaerobic syntrophy(Fenchel amp Finlay 1995) The selective pressure associat-ing the host to its eubacterial symbiont was posited to havebeen the hostrsquos strict dependence upon H2 produced as awaste product of the fermentative metabolism of a facul-tatively anaerobic symbiont This model for the origin ofmitochondria and for the origin of the heterotrophic life-style in eukaryotes was fortuitously also a model for theorigin of eukaryotes It is shown in figure 7 which is larg-ely self-explanatory

Only few minor modifications with respect to the orig-inal figure but not to the original model have been intro-duced here First the original figures (Martin amp Muller1998) did not specifically indicate DNA or ribosomes inthe cells (corrected here) with the eubacterial componentsshaded blue and the archaebacterial components shadedred There is a limit to the amount of detail that can beincluded in such a figure hence the Gram-negative walland lipopolysaccharide layer of the symbiont are symbolized

Another change is that a model for the origin of theendomembrane system in eukaryotes including the


Figure 7 A summary of a model suggested for the origin of eukaryotes (Martin amp Muller 1998) including a possible route forthe principle underlying the origin of the nucleus under that model as discussed (Martin 1999a) where it was suggested thatthe origin of the endoplasmic reticulum endomembrane system preceded the origin of the nuclear compartment See text



nuclear membrane which is a single folded membranethat is contiguous with the endoplasmic reticulum issketched here in the top panel of the figure That modelfor the origin of the nuclear membrane (the vesicularmodel) was elaborated and contrasted with other modelselsewhere (Martin 1999a) its basic tenets are recapitu-lated in figure 7

Another change is that the eubacterial and archaebact-erial lipids were indicated somewhat more clearly althoughtheir relevance was stated in the original The lipids areimportant because according to textbook depictions ofcell evolution (the lsquouniversal treersquo) eukaryotes shouldhave archaebacterial lipids just like they have an archaeb-acterial genetic apparatus but they do not Eukaryoteshave eubacterial lipids Few evolutionary biologistsaddress this issue One suggestion to account for this isthat the common ancestor of all cells was a eubacteriumand that the single common ancestor of archaebacteriareinvented its entire lipid and cell wall chemistry sub-sequent to the divergence from the eukaryotic lineage onan rRNA-like tree (Van Valen amp Maiorana 1980 Cavalier-Smith 2002) Our explanation for the origin of eubacteriallipids in eukaryotes is that the genes for their biosynthesiswere transferred to the hostrsquos chromosomes and that path-way became functional there (acetyl-CoA NAD(P)H anddihydroxyacetone phosphate as the precursors is in amplesupply in all cells) and is somewhat more explicit but notprincipally different from either Zilligrsquos (1991) or Koga etalrsquos (1998) explanation Another change is that the pres-ence of eubacterial and archaebacterial genes in the cyto-solic chromosomes is more clearly indicated than in theoriginal where it was simply stated in words Note that theterm lsquogene transferrsquo is not exactly correct because genetranslocation from an organelle genome to the hostsrsquoschromosomes is always a duplicative process in the strictsense the transferred genes are copied not transferredbecause an active copy has to remain in the organelle untilits function can be substituted by the copy residing in thehostrsquos chromosomes (Allen 1993) During the phase ofevolution before the origin of a protein routing apparatusto direct cytosolic proteins into the mitochondrion(Henze amp Martin 2001) the process of gene copying fromthe organelle to the host which occasional organelle deathor lysis renders inevitable (Doolittle 1999) would havemeant a continuous downpour of eubacterial genes intothe hostrsquos chromosomes constantly providing new rawgenetic starting material for the evolution of eukaryotic-specific traits while providing the opportunity to retainoriginally encoded functions as well (The converse trans-fer process from host to organelle is extremely unfavour-able events of occasional host death or lysis do not leadto transfers they lead to no progeny) This view implicatesgene transfers from the organelle and the integration ofthose genomes within a single cellular confines (Herrmann1997) as major genetic mechanisms underlying the gener-ation of eukaryotic novelties

Another change is that histones which are present inmethanogens and relatives (Reeve et al 1997 Fahrner etal 2001) as well as in eukaryotes are sketched onto thehostrsquos chromosomes Another change is that inventionsspecific to the eukaryotic lineage are symbolized as greendots These include for example spliceosomal introns thatare probably descended from eubacterial group II introns


that were argued to have entered the eukaryotic lineagevia the mitochondrial symbiont (Cavalier-Smith 1991Roger amp Doolittle 1993) Other inventions include thecytoskeletal components tubulin and actin which areprobably descended from their prokaryotic homologuesFtsZ and MreB (Erickson 2001) the protein importapparatus of the two mitochondrial membranes (Truscottet al 2001 Pfanner amp Truscott 2002) the mitochondrialcarrier family of proteins (Van der Giezen et al 2002Voncken et al 2002) and what-all The cell inferredunder this model and sketched at the top of figure 7 hasan archaebacterial genetic apparatus eubacterial energymetabolism eubacterial lipids the seeds of an endomem-brane system from which the nucleus is inferrable andample genetic starting material (two highly divergent andpartially merged prokaryotic genomes) to evolve cyto-logical and genetic traits that are specific to the eukary-otic lineage

Thus to deliver the goods promised at the outset of thissection under the premise stated here (i) the evolutionarybasis for the attributes that are shared by eubacteria andeukaryotes lies in gene acquisition from the eubacterialcommon ancestor of mitochondria and hydrogenosomes(ii) the evolutionary basis for the attributes that are sharedby archaebacteria and eukaryotes lies in the archaebacter-ial nature of the host that acquired the mitochondrionand (iii) the evolutionary basis for the attributes that dis-tinguish eukaryotes from prokaryotes are founded innovelties that arose specifically in the eukaryotic lineagesubsequent to its origin which given the fossil evidenceprobably occurred between 15 and 2 Gyr ago

10 GETTING ACCUSTOMED TO THE POSSIBILITYOF A PROKARYOTIC AND AUTOTROPHIC HOST

The model presented here for mitochondrial (andeukaryote) origins differs in several fundamental aspectsfrom traditional models (Doolittle 1998) which usuallyentail a phagocytosing eukaryote as the host rather thanan H2-dependent autotrophic archaebacterium and assuch it has received welcome criticism (Cavalier-Smith2002) Regarding a descent of the host lineage from H2-dependent autotrophic archaebacteria it is indeed curi-ous that some proteins such as group II chaperonins(Leroux et al 1999) and histones (Sandmann amp Reeve1998) do link up eukaryotes with methanogens and theirrelatives but obviously genome-wide analyses are neededto derive further support for this view Another questionto some was whether anaerobic syntrophy would providesufficient selective pressure to select for the unsual kindsof prokaryotic morphologies required by the model toassociate an archaebacterium and a eubacterium but trulystriking anaerobic syntrophic associations involvingmethanogens and hydrogen dependence (albeit by sul-phate reducers) are becoming known in the kinds ofenvironments predicted (Boetius et al 2000 Michaelis etal 2002) Another criticism was that at the time the modelwas presented there was no precedence for any prokary-ote living inside another prokaryote raising questions asto whether it is possible that an endosymbiosis betweentwo prokaryotes would be possible (Doolittle 1998) Inthe meantime a clear example has become known whereone prokaryote lives stably as an endosymbiont within


another (von Dohlen et al 2001) Although the mech-anism of entry is still unknown that example demon-strates that it is possible to establish an endosymbiosiswith a non-phagocytotic prokaryotic host as predictedAnother prediction was that mitochondria should befound that possess and use both pyruvate dehydrogenaseand PFOR evidence that is forthcoming (M Hoffmeisterand C Rotte unpublished data) Another prediction wasthat evidence for a trace of glycolysis should be found inmitochondria which has been found (Liaud et al 2000)Another prediction was that autotrophic archaebacteriashould be able to acquire and express eubacterial genesand that such aquisitions may expand the biochemicalrepertoire of the recipient as the genome sequenceof Methanosarcina mazei a methanogen with ca 30eubacterial genes in its genome has been confirmed(Deppenmeier et al 2002) Furthermore the brunt ofgenes that M mazei has acquired from eubacteria are larg-ely associated with its nutritional lifestyle in line with thepredictions of the model (but also include proteins suchas GroEL and DnaK) The clearest most easily falsifiable(and most often overlooked) prediction of the hydrogenhypothesis is that it predicts all eukaryotes to possess amitochondrion (or hydrogenosome) or to have possessedone in their past Other current models for the origin ofeukaryotes do not generate that prediction

From the start it was clear that eukaryotes had archae-bacterial ribosomes linking these two lineages But it wasalso well-known early on (Zillig 1991) that equally numer-ous possibly even more (Rivera et al 1998) traits linkeukaryotes and eubacteria In some anaerobic protiststraces of the archaebacterial host lineage have beenretained in energy metabolism as well (see Hannaert etal 2000 Sanchez et al 2000 Suguri et al 2001) clearlyindicating mosaicism in eukaryotic biochemistry Zillig(1991) argued that eukaryotes acquired their eubacterialtraits including lipid biosynthesis and some of their glyco-lytic genes from a eubacterial symbiont but because hethought that Giardia was primitively amitochondriate heinvoked a eubacterial donor that was distinct from themitochondrion Today it is quite clear that Giardia in factdid possess a mitochondrion (Hashimoto et al 1998Roger et al 1998 Tachezy et al 2001) Had Zillig beenaware of that it is doubtful that he would have invokedthe additional endosymbiont The hydrogen hypothesis(figure 7) merely simplifies Zilligrsquos scheme of cell evol-ution by one endosymbiont accounts for the origin ofhydrogenosomes and specifies the possible ecologicalcontext of symbiosis by making predictions about the nat-ure of the cells involved But it uses the same explanatoryprinciple of gene acquisition from an endosymbiont thecommon ancestor of mitochondria and hydrogenosomesto account for the numerous biochemically eubacterialtraits in eukaryotes that are too often disregarded

11 HORIZONTAL GENE TRANSFERCOMPLICATING BUT NOT OBSCURING EARLY

EVOLUTION

The common ancestor of mitochondria and hydro-genosomes was a member of the a-proteobacteria (Yanget al 1985) and therefore it may have been photosyntheticor non-photosynthetic autotrophic or heterotrophic


anaerobic or aerobic or most probably all of the aboveas is the case for many contemporary representatives ofthe group such as Rhodobacter but it at least should havehad the physiological potential of a bacterium like Para-coccus ( John amp Whatley 1975) The a-proteobacteria area rather heterogeneous group (Woese 1987) some haveacquired hefty chunks of DNA from methanogens (seeChistoserdova et al 1998) and some have evolved veryelaborate mechanisms of lateral transfer for example thegene transfer agents of Rhodobacter which slice the gen-ome up into nicely transmissible 4 kilobase pieces undersuitable conditions (Lang amp Beatty 2001) These days ithas become popular to invoke an independent horizontalgene transfer from prokaryotes to eukaryotes from a newand different donor every time one sees a eukaryotic genethat is not archaebacterial-like but does not branch withhomologues from a-proteobacteria (Baughn amp Malamy2002) also by staunch opponents of lateral gene transfer(Kurland 2000) and particularly when it suits the argu-ment (Canback et al 2002) The main problem with suchad hoc use of that explanatory principle is that the eukary-otic genes involved usually share a single common eubac-terial origin (Rotte et al 2001 Horner et al 2002)meaning that eukaryotes acquired most of their eubacter-ial genes only once and in their common ancestor in linewith the model set out here Horiike et al (2001) analysedthe yeast genome and fulfilled the predictions set out byMartin amp Muller (1998) for the origins of functionalclasses of eukaryotic genes yet Horiike et al (2001) didnot notice that they had confirmed those predictions(Rotte amp Martin 2001)

There is another interesting variant of lateral gene trans-fer as an explanatory principle as it applies to eukaryoticgenes namely to take all of the eubacterial genes in eukar-yotes that do not branch with a-proteobacterial homol-ogues lump them together and attribute their origin toone single hypothetical donor usually an imaginary organ-ism (Hartman amp Federov 2002 Hedges et al 2001Horiike et al 2001) Given lateral transfer and the prob-lems of deep phylogenies it makes more sense to attributesuch genes to the mitochondrial endosymbiont Mito-chondria arose only once (Gray et al 1999 Lang et al1999) and lateral transfer is real among free-living prokar-yotes (see Doolittle 1999 Jain et al 1999 Eisen 2000Ochman et al 2000) Because lateral transfer occurstoday we should assume it to have also existed in the pastIf we incorporate that reasonable but mildly complicatingassumption to our views on early evolution it suddenlyseems silly to assume (i) that any contemporary a-proteo-bacterium should have exactly the same genes as it had2 Gyr ago and (ii) that all genes possessed by the singlea-proteobacterium that gave rise to mitochondria shouldbe found today only among a-proteobacterial genomes(Martin 1999b) The same relatively insevere problemapplies to cyanobacteria (Rujan amp Martin 2001) when itcomes to genes that plants acquired from plastids Hori-zontal gene transfer was lsquoa pain in the neckrsquo for molecularevolutionists 20 years ago (Dickerson 1980) and will stillbe in 20 more it has not dashed all hopes that genomesequences will help unravel evolutionary history it has justleft a bit of tea in the cup and swirling leaves are moredifficult to read After all the distinction between ancientgene transfers and outright phylogenetic artefacts in gene


trees that span 2 Gyr or more of evolution is anything butsimple (Forterre amp Philippe 1999 Poole et al 1999) indi-cating that both general (pathways) and specific (enzymes)biochemical characters in addition to broad-scale phylo-genetic comparisons which have revealed widespreadmosaicism in eukaryotes (Brown amp Doolittle 1997Doolittle 1997 Feng et al 1997) should weigh in evidenceof early cell evolution

12 CONCLUSION

In this paper we have undertaken the task of sketchingone possible route to get from the chemical elements toeukaryotes in ca 2 Gyr focusing on aspects of the originof life and the origin of cells that we consider to be parti-cularly important sustained sources of redox potential(energy) and physical compartments for every step alongthe way The catalytic capabilities of RNA (Cech 1986)and the importance of the ribosome possibly as a rep-licator first (Poole et al 1999) and certainly as a translat-ing coupler between genotype and phenotype (Woese2002) brought biology out of the chicken-and-egg prob-lem with respect to DNA and protein But where there isanything like replication going on something is doublingin mass so it should be kept in mind that a steady flowof precursors must be coming from somewhere that theirsynthesis requires sustained energy input sustained reduc-ing power in the form of electrons sustained concen-trations of precursors sufficient to allow any pair ofreactants to meet and physical compartments to retainthe products formed

We have argued strongly in favour of a chemoauto-trophic but mesophilic origin of life and upon that prem-ise we see no way in which eukaryotes none of which arechemoautotrophic as Kandler (1998) has pointed outcould have possibly arisen at the same time as prokaryoteslet alone before prokaryotes as some biologists currentlycontend We have argued in favour of a single origin oflife which gave rise to a non-free-living universal ancestorthat was confined to structured FeS precipitates at a sub-marine seepage site within which it arose We have arguedthat the archaebacterial and eubacterial lineages of prokar-yotes independently surmounted the transition to the free-living state based on the chemical dissimilarity of their lip-ids as Koga et al (1998) have suggested However theuniversal common ancestor we propose was not of a non-cellular organization because we consider it impossiblethat any form of life or self-replicating system regardlessof how primitive could glean energy from its environmentand retain chemical constituents for its proliferation with-out a pre-existing inorganic cellular scaffold We haveargued against lsquofusionrsquo or lsquochimaericrsquo models for the originof eukaryotes Instead we have argued for an endosymbi-otic origin of mitochondria in an archaebacterial hostUnder these premises the living worldrsquos most ancientevolutionary divide is that between eubacteria and archae-bacteria and its steepest evolutionary grade is that betweenprokaryotes and eukaryotes (Mayr 1998)

All life is organized as cellsmdashcompartments separatedfrom their surroundings that spontaneously multiply withenergy gleaned through self-contained thermodyn-amically favourable redox reactions From our viewpointphysical compartmentation from the environment and


self-organization of self-contained redox reactions are themost conserved attributes of living things hence inorganicmatter with such attributes would be lifersquos most likely fore-bear

We thank the Deutsche Forschungsgemeinschaft and BayerAG for support Martin Embley for making data availablebefore publication and Miklos Muller and Allan Hall for dis-cussions

REFERENCES

Allen J F 1993 Control of gene expression by redox potentialand the requirement for chloroplast and mitochondrial gen-omes J Theor Biol 165 609ndash631

Altschul S F Madden T L Schaffer A A Zhang J HZhang Z Miller W amp Lipman D J 1997 Gapped Blast

and PSI-Blast a new generation of protein database searchprograms Nucleic Acids Res 25 3389ndash3402

Amend J P amp Shock E L 2001 Energetics of overall meta-bolic reactions of thermophilic and hyperthermophilicArchaea and Bacteria FEMS Microbiol Rev 25 175ndash243

Appel P W U Rollinson H R amp Touret J L R 2001Remnants of an Early Archaean (375 Ga) sea-floorhydrothermal system in the Isua Greenstone Belt PrecambrRes 112 27ndash49

Bada J L amp Lazcano A 2002 Some like it hot but not thefirst biomolecules Science 296 1982ndash1983

Ballard R D amp Grassle J F 1979 Incredible world of deep-sea rifts Natl Geographic 156 680ndash705

Banks D 1985 A fossil hydrothermal worm assemblage fromthe Tynagh lead-zinc deposit in Ireland Nature 313 128ndash131

Banks D A Boyce A J amp Samson I M 2002 Constraintson the origins of fluids forming Irish ZnndashPbndashBa depositsevidence from the composition of fluid inclusions EconGeol 97 471ndash480

Bartel D P amp Unrau P J 1999 Constructing an RNA worldTrends Biochem Sci 24 M9ndashM12

Baughn A D amp Malamy M H 2002 A mitochondrial-likeaconitase in the bacterium Bacteroides fragilis implicationsfor the evolution of the mitochondrial Krebs cycle Proc NatlAcad Sci USA 99 4662ndash4667

Baymann F Lebrun E Brugna M Schoepp-Cothenet BGuidici-Oritconi M-T amp Nitschke W 2003 The redoxprotein construction kit pre-last universal common ancestorevolution of energy-conserving enzymes Phil Trans R SocLond B 358 267ndash274 (DOI 101098rstb20021184)

Bell S D amp Jackson S P 1998 Transcription and translationin Archaea a mosaic of eukaryal and bacterial featuresTrends Microbiol 6 222ndash228

Bell S D Magill C P amp Jackson S P 2001 Basal and regu-lated transcription in Archaea Biochem Soc Trans 29392ndash395

Bernal J D 1951 The physical basis of life London Routledgeand Kegan Paul

Bernal J D 1960 The problem of stages in biopoesis InAspects of the origin of life (ed M Florkin) pp 30ndash45 NewYork Pergamon Press

Berry S 2002 The chemical basis of membrane bioenergeticsJ Mol Evol 54 595ndash613

Boetius A Ravenschlag K Schubert C J Rickert DWiddel F Gieseke A Amann R Joslashrgensen B B WitteU amp Pfannkuche O 2000 A marine microbial consortiumapparently mediating anaerobic oxidation of methaneNature 407 623ndash626

Bounama C Franck S amp von Bloh W 2001 The fate ofthe Earthrsquos ocean Hydrol Earth Syst Sci 5 569ndash575


Boyce A J Coleman M L amp Russell M J 1983 Formationof fossil hydrothermal chimneys and mounds from Sil-vermines Ireland Nature 306 545ndash550

Brinkmann H amp Philippe H 1999 Archaea sister group ofBacteria Indications from tree reconstruction artifacts inancient phylogenies Mol Biol Evol 16 817ndash825

Brocks J J Logan G A Buick R amp Summons R E 1999Archean molecular fossils and the early rise of eukaryotesScience 285 1033ndash1036

Brown J R amp Doolittle W F 1997 Archaea and the prokary-ote-to-eukaryote transition Microbiol Mol Biol Rev 61456ndash502

Butterfield N J 2000 Bangiomorpha pubescens n gen n spimplications for the evolution of sex multicellularity andthe MesoproterozoicNeoproterozoic radiation of eukary-otes Palaeobiology 263 386ndash404

Cairns-Smith A G 1982 Genetic takeover and the mineral ori-gins of life Cambridge University Press

Calvin M 1969 Chemical evolution Oxford Clarendon PressCanback B Andersson S G amp Kurland C G 2002 The

global phylogeny of glycolytic enzymes Proc Natl Acad SciUSA 99 6097ndash6102

Cavalier-Smith T 1987 Eukaryotes with no mitochondriaNature 326 332ndash333

Cavalier-Smith T 1991 Intron phylogeny a new hypothesisTrends Genet 7 145ndash148

Cavalier-Smith T 2000 Membrane heredity and early chloro-plast evolution Trends Plant Sci 5 174ndash182

Cavalier-Smith T 2002 The neomuran origin of archaebac-teria the negibacterial root of the universal tree and bacterialmegaclassification Int J Syst Evol Microbiol 52 7ndash76

Cech T R 1986 The generality of self-splicing RNA relation-ship to nuclear mRNA splicing Cell 44 207ndash210

Chabriere E Charon M H Volbedam A Pieullem LHatchikian E C amp Fontecilla-Camps J C 1999 Crystalstructures of the key anaerobic enzyme pyruvateferredoxinoxidoreductase free and in complex with pyruvate NatStruct Biol 6 182ndash190

Chapman A Linstead D J amp Lloyd D 1999 Hydrogen per-oxide is a product of oxygen consumption by Trichomonasvaginalis J Biosci 24 339ndash344

Chien Y-T Auerbach V Brabban A D amp Zinder S H2000 Analysis of genes encoding an alternative nitrogenasein the archaeon Methanosarcina barkeri 277 J Bacteriol 1823247ndash3253

Chistoserdova L Vorholt J A Thauer R K amp LidstromM E 1998 C1 transfer enzymes and coenzymes linkingmethylotrophic bacteria and methanogenic Archaea Science281 99ndash102

Cody G D Boctor N Z Filley T R Hazen R M ScottJ H Sharma A amp Yoder Jr H S 2000 Primordial car-bonylated iron-sulfur compounds and the synthesis of pyruv-ate Science 289 1337ndash1340

Corliss J B Baross J A amp Hoffmann S E 1981 An hypoth-esis concerning the relationship between submarine hotsprings and the origin of life on Earth Oceanologica Acta4(Suppl) 59ndash69

Daniel R M amp Danson M J 1995 Did primitive micro-organisms use nonheme iron proteins in place of NADP JMol Evol 40 559ndash563

Deamer W D 1985 Boundary structures are formed byorganic components of the Murchison meteorite Nature317 792ndash794

Deamer W D 1986 Role of amphiphilic compounds in theevolution of membrane structure on the early Earth OriginsLife Evol Biosphere 17 3ndash25

De Duve C 1991 Blueprint for a cell the nature and origin oflife Burlington NC Neil Patterson Publishers


De Duve C 1994 Vital dust life as a cosmic imperative NewYork Basic Books HarperCollins

De Duve C amp Miller S L 1991 Two-dimensional life ProcNatl Acad Sci USA 88 10 014ndash10 017

Deppenmeier U (and 22 others) 2002 The genome ofMethanosarcina mazei evidence for lateral gene transferbetween bacteria and archaea J Mol Microbiol Biotechnol4 453ndash461

Dickerson R 1980 Evolution and gene transfer in purple pho-tosynthetic bacteria Nature 283 210ndash212

DiMarco A A Bobik T A amp Wolfe R S 1990 Unusualcoenzymes of methanogenesis A Rev Biochem 59 355ndash394

Dismukes G C Klimov V V Baranov S V KozlovY N DasGupta J amp Tyryshkin A 2001 The origin ofatmospheric oxygen on Earth the innovation of oxygenicphotosynthesis Proc Natl Acad Sci USA 99 2170ndash2175

Doeller J E Grieshaber M K amp Kraus D W 2001 Chem-olithoheterotrophy in a metazoan tissue thiosulfate pro-duction matches ATP demand in ciliated mussel gills JExp Biol 204 3755ndash3764

Doolittle W F 1997 Fun with genealogy Proc Natl AcadSci USA 94 12 751ndash12 753

Doolittle W F 1998 A paradigm gets shifty Nature 39215ndash16

Doolittle W F 1999 Phylogenetic classification and the uni-versal tree Science 284 2124ndash2128

Eigen M 1992 Steps towards life Oxford University PressEigen M amp Schuster P 1978 The hypercycle a principle

of natural sef-organization part C the realistic hypercycleNaturwissenschaften 65 347ndash369

Eisen J 2000 Horizontal gene transfer among microbial gen-omes new insights from complete genome analysis CurrOpin Genet Dev 10 606ndash611

Embley T M amp Hirt R P 1998 Early branching eukaryotesCurr Opin Genet Dev 8 655ndash661

Embley T M Horner D A amp Hirt R P 1997 Anaerobiceukaryote evolution hydrogenosomes as biochemicallymodified mitochondria Trends Ecol Evol 12 437ndash441

Erickson H P 2001 Cytoskeleton evolution in bacteriaNature 413 30

Fahrner R L Cascio D Lake J A amp Slesarev A 2001 Anancestral nuclear protein assembly crystal structure of theMethanopyrus kandleri histone Protein Sci 10 2002ndash2007

Fenchel T amp Finlay B J 1995 Ecology and evolution in anoxicworlds Oxford University Press

Feng D-F Cho G amp Doolittle R F 1997 Determiningdivergence times with a protein clock update and re-evalu-ation Proc Natl Acad Sci USA 94 13 028ndash13 033

Ferris J P Hill A R Liu R amp Orgel L E 1996 Synthesisof long prebiotic oligomers on mineral surfaces Nature 38159ndash61

Ferry J G 1995 CO dehydrogenase A Rev Microbiol 49305ndash333

Forterre P amp Philippe H 1999 Where is the root of the uni-versal tree of life BioEssays 21 871ndash879

Fuchs G 1994 Variations of the acetyl-CoA pathway indiversely related microorganisms that are not acetogens InAcetogenesis (ed H L Drake) pp 506ndash538 New YorkChapman amp Hall

Gaffey M J 1997 The early solar system Origins Life EvolBiosphere 27 185ndash203

Geptner A Kristmannsdottir H Kristjansson J amp Marte-insson V 2002 Biogenic saponite from an active submarinehot spring Iceland Clays and Clay Minerals 50 174ndash185

Gesteland R F Cech T R amp Atkins J F 1999 The RNAworld New York Cold Spring Harbor Press

Gilbert W 1986 The RNA world Nature 319 618


Glasby G P amp Kasahara J 2001 Influence of tidal effects onthe periodicity of earthquake activity in diverse geologicalsettings with particular emphasis on submarine systemsEarth Sci Rev 52 261ndash297

Gogarten J P (and 12 others) 1989 Evolution of the vacuolarH1-ATPase implications for the origin of eukaryotes ProcNatl Acad Sci USA 86 6661ndash6665

Gray M W Burger G amp Lang B F 1999 Mitochondrialevolution Science 283 1476ndash1481

Habenicht A Hellman U amp Cerff R 1994 Non-phos-phorylating GAPDH of higher plants is a member of thealdehyde dehydrogenase superfamily with no sequence hom-ology to phosphorylating GAPDH J Mol Biol 237 165ndash171

Hai T Schneider B Schmidt J amp Adam G 1996 Sterolsand triterpenoids from the cyanobacterium Anabaena hal-lensis Phytochemistry 41 1083ndash1084

Hall D O Cammack R amp Rao K K 1971 Role of ferre-doxins in the origin of life and biological evolution Nature233 136ndash138

Han T M amp Runnegar B 1992 Megascopic eukaryotic algaefrom the 21 billion-year-old Neeguanee iron formationMichigan Science 257 232ndash235

Hannaert V Brinkmann H Nowitzki U Lee J A AlbertM-A Sensen C Gaasterland T Muller M MichelsP amp Martin W 2000 Enolase from Trypanosoma bruceifrom the amitochondriate protist Mastigamoeba balamuthiand from the chloroplast and cytosol of Euglena gracilispieces in the evolutionary puzzle of the eukaryotic glycolyticpathway Mol Biol Evol 17 989ndash1000

Hao B Gong W Ferguson T K James C M KrzyckiJ A amp Chan M K 2002 A new UAG-encoded residue inthe structure of a methanogen methyltransferase Science296 1462ndash1466

Hartman H amp Fedorov A 2002 The origin of the eukaryoticcell a genomic investigation Proc Natl Acad Sci USA 991420ndash1425

Hashimoto T Sanchez L B Shirakura T Muller M ampHasegawa M 1998 Secondary absence of mitochondria inGiardia lamblia and Trichomonas vaginalis revealed by valyl-tRNA synthetase phylogeny Proc Natl Acad Sci USA 956860ndash6865

Hayes J M 1994 Global methanotrophy at the ArchaenndashPro-terozoic transition In Early life on Earth (ed S Bengston)pp 220ndash236 New York Columbia University Press

Hazen R M Filley T R amp Goodfriend G A 2001 Selec-tive adsorption of L- and D-amino acids on calcite impli-cations for biochemical homochirality Proc Natl Acad SciUSA 98 5487ndash5490

Hedges S B Chen H Kumar S Wang D Y ThompsonA S amp Watanabe H 2001 A genomic timescale for the ori-gin of eukaryotes BMC Evol Biol 1 4

Heinen W amp Lauwers A M 1996 Organic sulfur com-pounds resulting from the interaction of iron sulfide hydro-gen sulfide and carbon dioxide in an anaerobic aqueousenvironment Origins Life Evol Biosphere 26 131ndash150

Hennet R J-C Holm N G amp Engel M H 1992 Abioticsynthesis of amino acids under hydrothermal conditions andthe origin of life a perpetual phenomenon Naturwissensch-aften 79 361ndash365

Henze K amp Martin W 2001 How are mitochondrial genestransferred to the nucleus Trends Genet 17 383ndash387

Henze K Schnarrenberger C amp Martin W 2001 Endosym-biotic gene transfer a special case of horizontal gene transfergermane to endosymbiosis the origins of organelles and theorigins of eukaryotes In Horizontal gene transfer (ed MSyvanen amp C Kado) pp 343ndash352 London Academic


Herrmann R G 1997 Eukaryotism towards a new interpret-ation In Eukaryotism and symbiosis (ed H E A SchenkR G Herrmann K W Jeon amp W Schwemmler) pp 73ndash118 Heidelberg Germany Springer

Horiike T Hamada K Kanaya S amp Shinozawa T 2001Origin of eukaryotic cell nuclei by symbiosis of Archaea inBacteria is revealed by homology hit analysis Nature CellBiol 3 210ndash214

Horner D S Heil B Happe T amp Embley T M 2002 Ironhydrogenases ancient enzymes in modern eukaryotesTrends Biochem Sci 27 148ndash153

Huber C amp Wachtershauser G 1997 Activated acetic acidby carbon fixation on (FeNi)S under primordial conditionsScience 276 245ndash247

Huber C amp Wachtershauser G 1998 Peptides by activationof amino acids with CO on NiFe surfaces implications forthe origin of life Science 281 670ndash672

Ito S Fushinobu S Yoshioka I Koga S MatsuzawaH amp Wakagi T 2001 Structural basis for the ADP-speci-ficity of a novel glucokinase from a hyperthermophilicarchaeon Structure 9 205ndash214

Jain R Rivera M C amp Lake J A 1999 Horizontal transferamong genomes the complexity hypothesis Proc Natl AcadSci USA 96 3801ndash3806

Javaux E J Knoll A H amp Walter M R 2001 Morphologi-cal and ecological complexity in early eukaryotic ecosystemsNature 412 66ndash69

Jenney Jr F E Verhagen M F J M Cui X amp AdamsM W W 1999 Anaerobic microbes oxygen detoxificationwithout superoxide dismutase Science 286 306ndash309

John P amp Whatley F R 1975 Paracoccus denitrificans and theevolutionary origin of the mitochondrion Nature 254495ndash498

Kamber B Z Moorbath S amp Whitehouse M J 2001 Theoldest rocks on Earth time constraints and geologicalcontroversies In The age of the Earth from 4004 BC to AD2002 Special publications vol 190 (ed C L E Lewis ampS J Knell) pp 177ndash203 Geological Society of London

Kandler O 1982 Cell wall structures and their phylogeneticimplications In Archaebacteria (ed O Kandler) pp 149ndash160 Stuttgart Germany Gustav Fischer

Kandler O 1998 The early diversification of life and the originof the three domains a proposal In Thermophiles the keys tomolecular evolution and the origin of life (ed J Wiegel ampM W W Adams) pp 19ndash31 Washington DCTaylor amp Francis

Kandler O amp Konig H 1978 Chemical composition of thepeptidoglycan-free cell walls of methanogenic bacteria ArchMicrobiol 120 181ndash183

Kates M 1979 The phytanyl ether-linked polar lipids and iso-prenoid neutral lipids of extremely halophilic bacteria ProgChem Fats Other Lipids 15 301ndash342

Katinka M D (and 15 others) 2001 The genome of the intra-cellular parasite Encephalithozoon cuniculi Nature 414450ndash453

Keeling P J 2001 Parasites go the full monty Nature 414401ndash402

Kelley D S 1996 Methane-rich fluids in the oceanic crust JGeophys Res 101B 2943ndash2962

Kelley D S (and 10 others) 2001 An off-axis hydrothermalvent field near the Mid-Atlantic Ridge at 30 degrees NNature 412 145ndash149

Kengen S W M Stams A J M amp de Vos W M 1996Sugar metabolism of hyperthermophiles FEMS MicrobiolRev 18 119ndash137

Koch A L amp Schmidt T M 1991 The first cellular bioener-getic process primitive generation of a proton-motive forceJ Mol Evol 33 297ndash304


Koga Y Kyuragi T Nishihara M amp Sone N 1998 Didarchaeal and bacterial cells arise independently from non-cellular precursors A hypothesis stating that the advent ofmembrane phospholipid with enantiomeric glycerophosph-ate backbones caused the separation of the two lines ofdescent J Mol Evol 46 54ndash63

Kohl W Gloe A amp Reichenbach H 1983 Steroids from themyxobacterium Nannocystis exedens J Gen Microbiol 1291629ndash1635

Kohlhoff M Dahm A amp Hensel R 1996 Tetrameric triose-phosphate isomerase from hyperthermophilic ArchaeaFEBS Lett 383 245ndash250

Kostic M Pochapsky S S Obenauer J Mo H PaganiG M Pejchal R amp Pochapsky T C 2002 Comparison offunctional domains in vertebrate-type ferredoxins Biochem-istry 41 5978ndash5989

Kurland C G 2000 Something for everyone Horizontal genetransfer in evolution EMBO Reports 1 92ndash95

Lancaster C R D Kroger A Auer M amp Michel H 1999Structure of fumarate reductase from Wolinella succinogenesat 22 Ac resolution Nature 403 377ndash385

Lang A S amp Beatty J T 2001 The gene transfer agent ofRhodobacter capsulatus and lsquoconstitutive transductionrsquo in pro-karyotes Arch Microbiol 175 241ndash249

Lang B F Gray M W amp Burger G 1999 Mitochondrialgenome evolution and the origin of eukaryotes A RevGenet 33 351ndash397

Lange B M Rujan T Martin W amp Croteau R 2000 Isop-renoid biosynthesis the evolution of two ancient and distinctpathways across genomes Proc Natl Acad Sci USA 9713 172ndash13 177

Langer D Hain J Thuriaux P amp Zillig W 1995 Tran-scription in Archaea similarity to that in Eukarya Proc NatlAcad Sci USA 92 5768ndash5772

Langworthy T A Tornabene T G amp Holzer G 1982 Lip-ids of archaebacteria In Archaebacteria (ed O Kandler) pp228ndash244 Stuttgart Germany Gustav Fischer

Larter R C L Boyce A J amp Russell M J 1981 Hydrother-mal pyrite chimneys from the Ballynoe Baryte deposit Sil-vermines County Tipperary Ireland Mineralium Deposita16 309ndash318

Leja J 1982 Surface chemistry of froth flotation New YorkPlenum

Leroux M R Fandrich M Klunker D Siegers K LupasA N Brown J R Schiebel E Dobson C M amp HartlF U 1999 MtGimC a novel archaeal chaperone related tothe eukaryotic chaperonin cofactor GimCprefoldin EMBOJ 18 6730ndash6743

Liaud M-F Lichtle C Apt K Martin W amp Cerff R2000 Compartment-specific isoforms of TPI and GAPDHare imported into diatom mitochondria as a fusion proteinevidence in favor of a mitochondrial origin of the eukaryoticglycolytic pathway Mol Biol Evol 17 213ndash223

Lindahl P A 2002 The Ni-containing carbon monoxidedehydrogenase family light at the end of the tunnelBiochemistry 41 2097ndash2105

Lindmark D G amp Muller M 1973 Hydrogenosome a cyto-plasmic organelle of the anaerobic flagellate Tritrichomonasfoetus and its role in pyruvate metabolism J Biol Chem 2487724ndash7728

Ljungdahl L G 1994 The acetyl-CoA pathway and thechemiosmotic generation of ATP during acetogenesis InAcetogenesis (ed H L Drake) pp 63ndash87 New York Chap-man amp Hall

Lopez-Garcia P amp Moreira D 1999 Metabolic symbiosis atthe origin of eukaryotes Trends Biochem Sci 24 88ndash93

Macedo-Ribeiro S Martins B M Pereira P J Buse GHuber R amp Soulimane T 2001 New insights into the ther-mostability of bacterial ferredoxins high-resolution crystal


structure of the seven-iron ferredoxin from Thermus thermo-philus J Biol Inorg Chem 6 663ndash674

Marczak R Gorrell T E amp Muller M 1983 Hydrogenoso-mal ferredoxin of the anaerobic protozoon Tritrichomonasfoetus J Biol Chem 258 12 427ndash12 433

Marshall W L 1994 Hydrothermal synthesis of amino acidsGeochimica et Cosmochimica Acta 58 2099ndash2106

Martin W 1999a A briefly argued case that mitochondria andplastids are descendants of endosymbionts but that thenuclear compartment is not Proc R Soc Lond B 2661387ndash1395 (DOI 101098rspb19990792)

Martin W 1999b Mosaic bacterial chromosomes a challengeen route to a tree of genomes BioEssays 21 99ndash104

Martin W amp Cerff R 1986 Prokaryotic features of a nucleusencoded enzyme cDNA sequences for chloroplast and cyto-solyic glyceraldehyde-3-phosphate dehydrogenases frommustard Sinapis alba Eur J Biochem 159 323ndash331

Martin W amp Herrmann R G 1998 Gene transfer fromorganelles to the nucleus how much what happens andwhy Plant Physiol 118 9ndash17

Martin W amp Muller M 1998 The hydrogen hypothesis forthe first eukaryote Nature 392 37ndash41

Martin W amp Schnarrenberger C 1997 The evolution of theCalvin cycle from prokaryotic to eukaryotic chromosomesa case study of functional redundancy in ancient pathwaysthrough endosymbiosis Curr Genet 32 1ndash18

Martin W Brinkmann H Savonna C amp Cerff R 1993Evidence for a chimeric nature of nuclear genomes eubact-erial origin of eukaryotic glyceraldehyde-3-phosphatedehydrogenase genes Proc Natl Acad Sci USA 90 8692ndash8696

Martin W Hoffmeister M Rotte C amp Henze K 2001 Anoverview of endosymbiotic models for the origins of eukary-otes their ATP-producing organelles mitochondria andhydrogenosomes and their heterotrophic lifestyle BiolChem 382 1521ndash1539

Maynard-Smith J amp Szathmary E 1995 The major evolution-ary transitions Oxford University Press

Mayr E 1998 Two empires or three Proc Natl Acad SciUSA 95 9720ndash9724

Mereschkowsky C 1905 Uber Natur und Ursprung der Chro-matophoren im Pflanzenreiche Biol Centralbl 25 593ndash604(English translation in Eur J Phycol 34 287ndash295)

Michaelis W (and 16 others) 2002 Microbial reefs in theBlack Sea fueled by anaerobic oxidation of methane Science297 1013ndash1015

Miller S L amp Bada J L 1988 Submarine hot springs andthe origin of life Nature 334 609ndash611

Mojzsis S J Arrhenius G McKeegan K D HarrisonT M Nutman A P amp Friend C R 1996 Evidence for lifeon Earth before 3800 million years ago Nature 384 55ndash59

Moulton V Gardner P P Pointon R F Creamer L KJameson G B amp Penny D 2000 RNA folding arguesagainst a hot-start origin of life J Mol Evol 51 416ndash421

Muller D Pitsch Kittaka A Wagner E Wintner C E ampEschenmoser A 1990 Mineral induced formation of sugarphosphates Helvetica Chimica Acta 73 1410ndash1468

Muller M 1988 Energy metabolism of protozoa without mito-chondria A Rev Microbiol 42 465ndash488

Muller M 1998 Enzymes and compartmentation of coreenergy metabolism of anaerobic protists a special case ineukaryotic evolution In Evolutionary relationships amongProtozoa (ed G H Coombs K Vickermann M ASleigh amp A Warren) pp 109ndash131 Dordrecht The Nether-lands Kluwer

Muller M 2000 A mitochondrion in Entamoeba histolyticaParasitol Today 16 368ndash369


Muth G W Orteleva-Donnelly L amp Strobel S A 2000 Asingle adenosine with a neutral pKa in the ribosomal pepti-dyl transferase center Science 289 947ndash950

Nei M 1987 Molecular evolutionary genetics New YorkColumbia University Press

Nelson K E Robertson M P Levy M amp Miller S L2001 Concentration by evaporation and the prebiotic syn-thesis of cytosine Origins Life Evol Biosphere 31 221ndash229

Newman M J amp Rood R T 1977 Implications of solar evol-ution for the Earthrsquos early atmosphere Science 198 1035ndash1037

Nisbet E G amp Sleep N H 2001 The habitat and nature ofearly life Nature 409 1083ndash1091

Oberholzer T Albizio M amp Luisi P L 1995 Polymerasechain reaction in liposomes Chem Biol 2 677ndash682

Ochman H Lawrence J G amp Groisman E S 2000 Lateralgene transfer and the nature of bacterial innovation Nature405 299ndash304

Omer A D Lowe T M Russell A G Ebhardt H EddyS R amp Dennis P P 2000 Homologs of small nucleolarRNAs in Archaea Science 288 517ndash522

Orgel L E amp Crick F H C 1993 Anticipating an RNAworld some past speculations on the origin of life where arethey today Federation Proc Federation Am Soc Exp Biol 7238ndash239

Oro J 1961 Comets and the formation of biochemical com-pounds on the primitive Earth Nature 190 389ndash390

Oro J 1994 Early chemical stages in the origin of life In Earlylife on Earth (ed S Bengston) pp 48ndash59 New YorkColumbia University Press

Oro J amp Kimball A P 1962 Synthesis of purines under poss-ible primitive Earth conditions II Purine intermediates fromhydrogen cyanide Arch Biochem Biophys 96 293ndash313

Penny D amp Poole A 1999 The nature of the last universalcommon ancestor Curr Opin Genet Dev 9 672ndash677

Peters J W 1999 Structure and mechanism of iron-onlyhydrogenases Curr Opin Struct Biol 9 670ndash676

Pfanner N amp Truscott K N 2002 Powering mitochondrialprotein import Nat Struct Biol 9 234ndash236

Philippe H amp Laurent J 1998 How good are deep phylogen-etic trees Curr Opin Genet Dev 8 616ndash623

Philippe H Germot A amp Moreira D 2000 The new phy-logeny of eukaryotes Curr Opin Genet Dev 10 596ndash601

Poole A M Jeffares D amp Penny D 1999 Prokaryotes thenew kids on the block BioEssays 21 880ndash889

Poole A M Penny D amp Sjoberg B 2000 Methyl-RNA anevolutionary bridge between RNA and DNA Chem Biol 7R207ndashR216

Poole A M Jeffares D C amp Penny D 1998 The path fromthe RNA world J Mol Evol 46 1ndash17

Quayle R J amp Ferenci T 1978 Evolutionary aspects of auto-trophy Microbiol Rev 42 251ndash273

Ragsdale S W 1994 CO dehydrogenase and the central roleof this enzyme in the fixation of carbon dioxide by anaerobicbacteria In Acetogenesis (ed H L Drake) pp 88ndash126 NewYork Chapman amp Hall

Reeve J N Sandman K amp Daniels C J 1997 Archaeal his-tones nucleosomes and transcription initiation Cell 89999ndash1002

Reichard P 1997 The evolution of ribonucleotide reductionTrends Biochem Sci 22 81ndash85

Rickard D Butler I B amp Olroyd A 2001 A novel iron sul-phide switch and its implications for Earth and planetaryscience Earth Planetary Sci Lett 189 85ndash91

Rivera M C Jain R Moore J E amp Lake J A 1998 Gen-omic evidence for two functionally distinct gene classesProc Natl Acad Sci USA 95 6239ndash6244

Roger A J 1999 Reconstructing early events in eukaryoticevolution Am Nat 154 S146ndashS163


Roger A J amp Doolittle W F 1993 Why introns-in-piecesNature 364 289ndash290

Roger A J Svard S G Tovar J Clark C G SmithM W Gillin F D amp Sogin M L 1998 A mitochondrial-like chaperonin 60 gene in Giardia lamblia evidence thatdiplomonads once harbored an endosymbiont related to theprogenitor of mitochondria Proc Natl Acad Sci USA 95229ndash234

Rosing M T 1999 13C-depleted carbon microparticles in3700-Ma sea-floor sedimentary rocks from West Green-land Science 283 674ndash676

Rotte C amp Martin W 2001 Does endosymbiosis explain theorigin of the nucleus Nature Cell Biol 8 E173ndashE174

Rotte C Stejskal F Zhu G Keithly J S amp Martin W2001 PyruvateNADP1 oxidoreductase from the mitochond-rion of Euglena gracilis and from the apicomplexan Cryptospo-ridium parvum a fusion of pyruvateferredoxinoxidoreductase and NADPH-cytochrome P450 reductaseMol Biol Evol 18 710ndash720

Rujan T amp Martin W 2001 How many genes in Arabidopsiscome from cyanobacteria An estimate from 386 proteinphylogenies Trends Genet 17 113ndash120

Russell M J 1983 Major sediment-hosted zinc 1 leaddeposits formation from hydrothermal convection cells thatdeepen during crustal extension Short course in sediment-hosted stratiform lead-zinc deposits Mineralogical Associ-ation of Canada Short course handbook vol 8 pp 251ndash282

Russell M J amp Hall A J 1997 The emergence of life fromiron monosulphide bubbles at a submarine hydrothermalredox and pH front J Geol Soc Lond 154 377ndash402

Russell M J amp Hall A J 2001 The onset of life and thedawn of oxygenic photosynthesis respective roles of cubanecore structures [Fe4S4]21 and transient [Mn4O4]41[OCaO]2In Sixth Int Conf Carbon Dioxide Utilization September 9ndash14 2001 Breckenridge Colorado Abstracts p 49

Russell M J Hall A J Cairns-Smith A G amp BratermanP S 1988 Submarine hot springs and the origin of lifeNature 336 117

Russell M J Hall A J amp Gize A P 1990 Pyrite and theorigin of life Nature 344 387

Russell M J Hall A J amp Mellersh A R 2003 On the dissi-pation of thermal and chemical energies on the early Earththe onsets of hydrothermal convection chemiosmosis gen-etically regulated metabolism and oxygenic photosynthesisIn Natural and laboratory-simulated thermal geochemical pro-cesses (ed R Ikan) Dordrecht The Netherlands Kluwer(In the press)

Sagan L 1967 On the origin of mitosing cells J Theor Biol14 225ndash274

Sagan C amp Chyba C 1997 The early faint sun paradoxorganic shielding of ultraviolet-labile greenhouse gasesScience 276 1217ndash1221

Samson I M amp Russell M J 1987 Genesis of the Sil-vermines zincndashleadndashbarite deposit Ireland fluid inclusionand stable isotope evidence Econ Geol 82 371ndash394

Sanchez L B Galperin M Y amp Muller M 2000 Acetyl-CoA synthetase from the amitochondriate eukaryote Giardialamblia belongs to the newly recognized superfamily of acyl-CoA synthetases (Nucleoside diphosphate-forming) J BiolChem 275 5794ndash5803

Sanchez R A Ferris J P amp Orgel L E 1967 Studies inprebiotic synthesis II Synthesis of purine precursors andamino acids from aqueous hydrogen cyanide J Mol Biol30 223ndash253

Sandman K amp Reeve J N 1998 Origin of the eukaryoticnucleus Science 280 501ndash503

Schafer G Engelhard M amp Muller V 1999 Bioenergeticsof the Archaea Microbiol Mol Biol Rev 63 570ndash620


Schnarrenberger C amp Martin W 2002 Evolution of theenzymes of the citric acid cycle and the glyoxylate cycle ofhigher plants a case study of endosymbiotic gene transferEur J Biochem 269 868ndash883

Schonheit P amp Schafer T 1995 Metabolism of hyperthermo-philes World J Microbiol Biotechnol 11 26ndash57

Schoonen M A A Xu Y amp Bebie J 1999 Energetics andkinetics of the prebiotic synthesis of simple organic acids andamino acids with the FeS-H2SFeS2 redox couple as reduct-ant Origins Life Evol Biosphere 29 5ndash32

Schouten S Bowman J P Rijpstra W I amp SinningheDamste J S 2000 Sterols in a psychrophilic methanotrophMethylosphaera hansonii FEMS Microbiol Lett 186 193ndash195

Schramm A Siebers B Tjaden B Brinkmann H amp Hen-sel R 2000 Pyruvate kinase of the hyperthermophilic cren-archaeote Thermoproteus tenax physiological role andphylogenetic aspects J Bacteriol 182 2001ndash2009

Schulte M D amp Shock E L 1995 Thermodynamics ofStrecker synthesis in hydrothermal systems Origins Life EvolBiosphere 25 161ndash173

Schutz M (and 11 others) 2000 Early evolution of cyto-chrome bc complexes J Mol Biol 300 663-675

Segret D amp Lancet D 2000 Composing life EMBO Rep 1217ndash222

Selig M Xavier K B Santos H amp Schonheit P 1997Comparative analysis of the EmbdenndashMeyerhof and EntnerndashDoudoroff glycolytic pathways in hyperthermophilicArchaea and the bacterium Thermotoga Arch Microbiol 167217ndash232

Shapiro R 1986 Origins a skepticrsquos guide to the creation of lifeon Earth London Heinemann

Siebers B Brinkmann H Dorr C Tjaden B Lilie HVan der Oost J amp Verhees C H 2001 Archaeal fructose-16-bisphosphate aldolases constitute a new family ofarchaeal type class I aldolase J Biol Chem 276 28 710ndash28 718

Sogin M Gunderson J Elwood H Alonso R amp PeattieD 1989 Phylogenetic meaning of the kingdom concept anunusual ribosomal RNA from Giardia lamblia Science 24375ndash77

Spiess F N The RISE project group 1980 East Pacific Risehot springs and geophysical experiments Science 2071421ndash1433

Srinivasan G James C M amp Krzycki J A 2002 Pyrrolysineencoded by UAG in Archaea charging of a UAG-decodingspecialized tRNA Science 296 1459ndash1462

Suguri S Henze K Sanchez L B Moore D V amp MullerM 2001 Archaebacterial relationships of the phosphoenolp-yruvate carboxykinase gene reveal mosaicism of Giardiaintestinalis core metabolism J Eukaryot Microbiol 48493ndash497

Sutherland J D amp Whitfield J N 1997 Prebiotic chemistrya bioorganic perspective Tetrahedron 34 11 493ndash11 527

Tachezy J Sanchez L B amp Muller M 2001 Mitochondrialtype iron-sulfur cluster assembly in the amitochondriateeukaryotes Trichomonas vaginalis and Giardia intestinalis asindicated by the phylogeny of IscS Mol Biol Evol 181919ndash1928

Takahashi Y amp Tokumoto U 2002 A third bacterial systemfor the assembly of iron-sulfur clusters with homologs inarchaea and plastids J Biol Chem 277 28 380ndash28 383

Thauer R 1998 Biochemistry of methanogenesis a tribute toMarjory Stephensen Microbiology (Amsterdam) 144 2377ndash2406

Thomm M 1996 Archaeal transcription factors and their rolein transcription initiation FEMS Microbiol Rev 18 159ndash171


Truscott K N Pfanner N amp Voos W 2001 Transport ofproteins into mitochondria Rev Phys Biochem Pharmacol143 81ndash136

Tuininga J E Verhees C H Van der Oost J KengenS W M Stams A J M amp De Vos W M 1999 Molecularand biochemical characterization of the ADP-dependentphosphofructokinase from the hyperthermophilic archaeonPyrococcus furiosus J Biol Chem 274 21 023ndash21 028

Tye B K 2000 Insights into DNA replication from the thirddomain of life Proc Natl Acad Sci USA 97 2399ndash2401

Ueno Y Yurimoto H Yoshioka H Komiya T amp Maruy-ama S 2002 Ion microprobe analysis of graphite from ca38 Ga metasediments Isua crustal belt West Greenlandrelationship between metamorphism and carbon isotopiccomposition Geochimica Cosmochimica Acta 66 1257ndash1268

Van der Giezen M Slotboom D J Horner D S DyalP J Harding M Xue G P Embley T M amp KunjiE R S 2002 Conserved properties of hydrogenosomal andmitochondrial ADPATP carriers a common origin for bothorganelles EMBO J 21 572ndash579

Van Valen L M amp Maiorana V C 1980 The Archaebacteriaand eukaryotic origins Nature 287 248ndash250

von Dohlen C D Kohler S Alsop S T amp McManusW R 2001 Mealybug beta-proteobacterial endosymbiontscontain gamma-proteobacterial symbionts Nature 412433ndash436

Voncken F (and 11 others) 2002 Multiple origins of hydro-genosomes functional and phylogenetic evidence from theADPATP carrier of the anaerobic chytrid Neocallimastix spMol Microbiol 44 1441ndash1454

Vossbrinck C R Maddox J V Friedman S Debrunner-Vossbrinck B A amp Woese C R 1987 Ribosomal RNAsequence suggests microsporidia are extremely ancient euka-ryotes Nature 326 411ndash414

Wachtershauser G 1988a Pyrite formation the first energysource for life a hypothesis Syst Appl Microbiol 10 207ndash210

Wachtershauser G 1988b Before enzymes and templatestheory of surface metabolism Microbiol Rev 52 452ndash484

Wachtershauser G 1998 Towards a reconstruction of ances-tral genomes by gene cluster alignment Syst Appl Microbiol21 473ndash477

Walter M 1983 Archaean stromatolites evidence for theEarthrsquos earliest benthos In Earthrsquos earliest biosphere its originand evolution (ed J W Schopf) pp 187ndash213 PrincetonUniversity Press

Wark E E amp Wark I W 1935 The physical chemistry offlotation VI The adsorption of amines by sulfide mineralsJ Physical Chem 39 1021ndash1030

Weber A L 1995 Prebiotic polymerization of 23-dimer-capto-1-propanol on the surface of ironIII hydroxide oxideOrigins Life Evol Biosphere 25 53ndash60

Whatley J M John P amp Whatley F R 1979 From extra-cellular to intracellular the establishment of chloroplasts andmitochondria Proc R Soc Lond B 204 165ndash187

White R H 1997 Purine biosynthesis in the domain Archaeawithout folates or modified folates J Bacteriol 179 3374ndash3377

White R H 2001 Biosynthesis of the methanogenic cofactorsVitam Horm 61 299ndash337

Wilde S A Valley J W Peck W H amp Graham C M2001 Evidence from detrital zircons for the existence of con-tinental crust and oceans on the Earth 44 Gyr ago Nature409 175ndash178

Williams B A Hirt R P Lucocq J M amp Embley T M2002 A mitochondrial remnant in the microsporidian Trachi-pleistophora hominis Nature 418 865ndash869


Wilson E B 1928 The cell in development and heredity 3rdedn New York Macmillan (Reprinted in 1987 by GarlandPublishing New York)

Woese C R 1987 Bacterial evolution Microbiol Rev 51221ndash271

Woese C R 2002 On the evolution of cells Proc Natl AcadSci USA 99 8742ndash8747

Woese C R Kandler O amp Wheelis M L 1990 Towards anatural system of organisms proposal for the domainsArchaea Bacteria and Eukarya Proc Natl Acad Sci USA87 4576ndash4579

Woods R 1976 Electrochemistry of sulfide flotation In Flo-tation A M Gaudin memorial volume vol 1 (ed M CFurstenau) pp 298ndash333 New York American Institute ofMining Metallurgical and Petroleum Engineers Inc

Xiong J amp Bauer C E 2002 A cytochrome b origin of photo-synthetic reaction centers an evolutionary link between res-piration and photosynthesis J Mol Biol 322 1025ndash1037

Yang D Oyaizu Y Oyaizu H Olsen G J amp Woese C R1985 Mitochondrial origins Proc Natl Acad Sci USA 824443ndash4447

Yoon K-S Hille R Hemann C amp Tabita F R 1999 Rub-redoxin from the green sulfur bacterium Chlorobium tepidumfunctions as an electron acceptor for pyruvate ferredoxin oxi-doreductase J Biol Chem 274 29 772ndash29 778

Zillig W 1991 Comparative biochemistry of Archaea and Bac-teria Curr Opin Genet Dev 1 544ndash551

DiscussionD Horner (Dipartimento di Fisiologia e Biochimica Gen-

erali University of Milan Milan Italy) I want to under-stand why the core of glycolysis has to come from thissymbiont and also why apart from a parsimony argumentit is not possible for the endosymbiosis giving rise to themitochondrion to have happened more than once

W Martin Let me take the second comment first It isjustified and can be reworded as follows lsquoDr Martin ifyoursquore saying this is so likely that this could give rise toeukaryotes how come it doesnrsquot happen all the timersquoThat the eukaryotes arose only once that all eukaryotesthat have been studied to date have a mitochondrion andthat all mitochondria arose only once are observations Itis an observation it is an explanandum ie something thatneeds to be explained We have to generate a hypothesisthat can explain that observation We have heard fromTom [Cavalier-Smith] that secondary symbiosis is com-pared with primary symbiosis a very widespread phenom-enon Primary endosymbiosis happened twice amongcountless cells in 4 billion years It is excruciatingly rareI suspect that the rareness of a free-living bacteriumbecoming an organelle lies in the difficulties of sorting outthe genetic integration of these systems into somethingthat will survive because we know of many eukaryotesthat have bacterial symbionts but we do not know ofmany eukaryotes that have acquired independentorganelles It comes down to chloroplasts and mitochon-dria So I think the second comment simply boils downto an interesting observation

The first question is why did the core enzymes of gly-colysis have to come from the mitochondrial symbiontThat is the application of Occamrsquos razor to suggestionswhich are common in the literature that an inability toget a citric acid cycle enzyme to branch with the alpha-proteobacterial homologue means that another endosym-biotic event has to be invoked Every time they see a new


unusual gene whose phylogeny they cannot explain theyinvoke a new endosymbiont I have looked at a lot ofgenes and if I have to invoke a new symbiont every timeI see a strange branch then we would get a conglomer-ation that just does not make sense My suspicion is thatwhat is wrong is that our models for protein evolution inparticular for glycolytic enzymes are so wrong that theyare producing artefacts regularly and we just do notunderstand how these molecules are evolving So what weare doing is rather than entertaining a notion of a newendosymbiosis every time we see a new gene phylogenymdashI have been there done thatmdashwe say what is wrong arethe gene phylogenies and let us take a minimalist viewwhere we have one archaebacterial host and one eubacter-ial symbiont and see how much we can explain with that

R Blankenship (Department of Chemistry and Biochemis-try Arizona State University AZ USA) Do you concludefrom the fact that the archaeal glycolytic enzymes areapparently unrelated to the eubacterial ones that glycoly-sis was independently invented two separate times

W Martin Yes I embrace the view of a chemoauto-trophic origin of life that is we start with carbon monox-ide What do I think was the first pathway I think thefirst pathway was the acetyl-CoA pathway of CO2 fixationThis is a pathway you find among clostridia and it is wide-spread among archaebacteria Its key enzyme is carbonmonoxide dehydrogenase And what carbon monoxidedehydrogenase does is it reduces CO2 with molecularhydrogen to carbon monoxide takes a methyl group froma methyl donor adds that to an ironndashsulphur centretransfers the methyl group to the carbon monoxide andmakes an acetyl group all in one reaction Add a sulphidein the form of CoA and you get two molecules of CO2

in and one molecule of acetyl CoA out It is a linear reac-tion so that is a good start and that is exactly the reactionthat the HuberndashWachtershauser experiment modulates sostart with C2 The next reaction in metabolism in organ-isms that have the acetyl-CoA pathway is pyruvate syn-thase Pyruvate synthase is also an ironndashsulphur enzymethat just adds CO2 with electrons from ferredoxin to makepyruvate So we go up the glycolytic pathway OK so yesI think that if we embrace the notion of a chemoauto-trophic origin of life we have to embrace the notion thatthis common ancestor we infer living in its iron sulphidehousing that it embedded a myriad of ways to make carbonbackbones not just to break them down from Oparinrsquosorganic soup I would like to point out one very interestingfinding The most conserved enzyme of the glycolyticpathway is without doubt enolase Enolase catalyses thefirst stereochemically specific reaction in carbon meta-bolism It takes a water molecule and adds it to phos-phoenol-pyruvate which is not chiral and the product is2-D-phosphoglycerate and the stereochemistry at C2 ofthat carbon compound never changes for the rest ofmetabolism And strangely enolase the most conservedglycolytic enzyme and the only one that is universally dis-tributed is not encoded in glycolytic operons in archae-bacteria it sits in the same operon as the ribosomalproteins It sits in the same operon as the ribosomal pro-teins And if you ask me that is the key to the holochiralityproblem from sugars And the selection at the origin oftranslation for incorporation of only L-amino acids intoproteins from a racemate might be a possible solution


which we suggest in the paper for the holochirality prob-lem as it arises in the amino acids Short question longanswer Yes independent origins these enzymes and thewhole biochemistry are not related

T Cavalier-Smith (Department of Zoology University ofOxford Oxford UK ) I have a comment and a questionThe comment relates to the fact that you rightly criticizeCarl Woese for assuming that eukaryotes are as old aseubacteria and I have similarly criticized him forassuming that archaebacteria are that old also You havemade that same assumption and I think my criticisms ofhim apply to you as wellhellip

W Martin I will address that OK go aheadT Cavalier-Smith And you are aware of the idea that

I put forward in 1987 and since have written about morerecently which is that archaebacteria and eukaryotes aresisters and not that eukaryotes evolved from an archae-bacterial ancestor

W Martin May I replyT Cavalier-Smith YesW Martin What Tom is saying is that he has always

placed very great importance on the same explanandumthat we are addressing here namely that eubacteria haveone kind of lipid archaebacteria have a different kind oflipid and eukaryotes have eubacterial lipids To explainthat he has had something called the neomuran theorywhich in this element is not different from Lee vanValenrsquos and Morianarsquos 1980 paper seven years before inNature where they said exactly the same thing I am notcriticizing the novelty but I am just saying that thisexplanandum has been around for a long time Zillig hasalso addressed it Tom explains this by saying that eukary-otes and archaebacteria share a common ancestor and thatthat common ancestor was a eubacterium Thus hepostulated a eubacterial common ancestor of eukaryotesand archaebacteria that has many archaebacterial attri-butes but eubacterial lipids This ancestor is then arguedto have brought forth the common ancestor of all eukary-otes and the common ancestor of all archaebacteriawhereby the latter are argued to have completely rein-vented their lipid biochemical pathways prior to archae-bacterial diversification I do not believe that but wecannot necessarily prove that it is wrong What we canprove is wrong is an element of your most recent formu-lation of this model in which you propose that this hap-pened 800 million years ago Now some geochemists arehere and tell me if I am wrong the carbon isotope recordfor ultra-light carbon ie d1 3C values of ndash50permil or ndash60permilgoes back well beyond 27 billion years it goes right downinto the same carbon levels where Schidlowski found d1 3Cvalues that are typical of present-day CO2 fixation path-ways using inorganic C with present-day d13 C valuesThis ultra-light carbon is generally interpreted as there isno other process that makes it as the trace of methano-genesis And that means that archaebacteria as a groupmust be at least three billion years old as common sensewould dictate Another problem Tom with the view thatarchaebacteria are only 800 million years old is the thingsthey can do with a wide range of inorganic donors andacceptors You can say lsquoyesrsquo this is due to adaptation toextreme environments but just looking at the biochemicalcharacters which most people who are involved in thisearly cell evolution business do not do it is like taking the


group with the least diversity of characters and makingit the most ancient and saying that there has been rapidevolution of all of the characters that are clearly ancestralon biogeochemical grounds The kinds of chemistry weare talking about here are consuming pyrite or sphaleriteor even sustained ATP production using carbon monox-ide These are ancient pathways So that is how I wouldaddress those two criticisms

T Cavalier-Smith Well OK My question was that ina recent paper (Cavalier-Smith 2002) I mentioned a cou-ple of gene splittings which I think argue against the ideathat eukaryotes came from within archaebacteria and Iwonder how you would refute that argument As for thecarbon isotope data I would really like to hear from anygeochemists about the validity of these values

W Martin Dr Nisbet is raising his handT Cavalier-Smith Good but may I just make one

point about it The basic argument is not direct evidencefor archaebacteria it is based on an assumption about eco-logical interactions between archaebacteria and methylo-trophs It is a rather complex argument based on severalassumptions and I do not think that the geochemists havelooked critically at alternative hypotheses They haveassumed that archaebacteria are old therefore it wouldbe a reasonable assumption Strauss et al (1992) whenreviewing this very point do say that other explanationsare in principle possible

W Martin OK I would like to have Dr Nisbet replyto the evidence and at the same time recall that it is notonly methylotrophs that can produce this trace Theassociation between methanogens and sulphate-reducersactually causes methanogenesis to run backwards Soanaerobic methanotrophy works but it works throughreverse methanogenesis and it produces this trace withoutmethanotrophy in the strict sense

E Nisbet (Department of Geology Royal Holloway Uni-versity of London Egham Surrey UK ) I am a geologistand I do not know very much biology but certainly at 27billion years both the carbon and the sulphur isotopes lookvery comparable in many ways to what we get in the mod-ern Earth with a wide diversity of carbon pathwaysincluding things that fractionate carbon very substantiallyand also very substantial self-refractionation If you goback past 35 billion years the story is more complex andsome of the outcrops are beautiful At 37 billion years atIssua in Greenland there is clear evidence for light car-bon but the rocks are metamorphosed so you cannot saymuch more than that there was some sort of biologicalprocess It looks like acetyl CoA metabolism but that isa pure guess

T Cavalier-Smith That is not specifically archaebact-erial

E Nisbet No if you want to say it is specifically archae-bacterial you can say at least at 27 billion years Thereis also I should mention Brockrsquos finding of chemicals thatlook as if they are eukaryotic at 27 billion years We haveoil at 27 billion years

T Cavalier-Smith But sterols can be made by threegroups of bacteria

E Nisbet Yes I know they can be made by Archaeaas well

W Martin We agree on that


T Cavalier-SmithWhat about the gene splitting as evi-dence for eukaryotes being sisters not derived fromarchaebacteria

W Martin Well I do not know exactly which data youare referring to but we have seen operons put togetherindependently in many groups of organisms

T Cavalier-Smith This is splitting up genes into separ-ately coded proteins

W Martin Well Tommdashindependent In enolase wehave seen independent insertions at exactly the same siteI also consider independent splittings to be possible I washoping you would ask why do the eukaryotes in the riboso-mal RNA tree branch below the archaebacteria when ourmodel actually says they should branch within the archae-bacteria and the answer is exactly the same as for themicrosporidia The microsporidia belong at the top of theeukaryotic tree but because of odd evolutionary processesin their rRNA they branch artefactually at the bottom Inthe model for the origin of eukaryotes that I presentedarchaebacterial ribosomes are evolving in a chimaericcytosol Thus the same thing is expected to happen as inthe case of the microsporidian RNA molecule The ances-tral eukaryotic rRNA molecule goes through a phase ofdramatic change it acquires a very long branch relative toits archaebacterial homologues and branches artefactuallydeep on the archaebacterial side of the rRNA tree


Additional references

Cavalier-Smith T 2002 The phagotrophic origin of eukary-otes and phylogenetic classification of Protozoa Int J SystEvol Microbiol 52 297ndash354

Strauss H Des Marais D J Hayes J M amp Summons R E1992 The carbon-isotopic record In The Proterozoic biosphere(ed J W Schopf amp C Klein) pp 117ndash127 CambridgeUniversity Press

GLOSSARY

ACS acetyl-CoA synthaseCODH carbon monoxide dehydrogenaseDXP 1-deoxyxylulose-5-phosphateFBA fructose-16-bisphosphate aldolaseGAPDH glyceraldehyde-3-phosphate dehydrogenaseGK glucokinaseGPI glucose-6-phosphate isomeraseIPP isopentenyl diphosphateMVA mevanolatePFK phosphofructokinasePFOR pyruvate ferredoxin oxidoreductaseTCA tricarboxylic acidTPI triosephosphate isomeraseTPP thiamine pyrophosphate


38 Gyr because carbon isotope data provide evidence forbiological CO2 fixation in sedimentary rocks of that age(Mojzsis et al 1996 Rosing 1999 Nisbet amp Sleep 2001Ueno et al 2002) By 35 Gyr stromatolites were presentpreserved microbial mats indicative of deposition by pho-tosynthetic prokaryotes (Walter 1983 Nisbet amp Sleep2001) By ca 15 Gyr so-called acritarchs became reason-ably abundant microfossils of unicellular organisms thatare almost certainly eukaryotes ( Javaux et al 2001) andprobably algae because of an easily preserved cell wall By12 Gyr spectacularly preserved multicellular organismsappear that were very probably red algae (Butterfield2000) There have been reports of more ancient remainsclaimed to be eukaryotes but they are equivocal Forexample Grypania is a 21 Gyr fossil (Han amp Runnegar1992) but could just as easily be a filamentous prokaryoteas a filamentous eukaryote because the cellular structureof the fossil is not preserved More recently steranes werefound in 27 Gyr sediments and were claimed to provideevidence for the existence of eukaryotes at that time(Brocks et al 1999) but several groups of prokaryotesincluding methanotrophic proteobacteria (see Schouten etal 2000) myxobacteria (Kohl et al 1983) and cyanobac-teria (Hai et al 1996) make the same kinds of compounds(for example cholesterol) claimed to be eukaryote-specificsuch that the sterane evidence for the age of eukaryotes isquestionable at best Ultra-light carbon material that bearsthe isotopic signature of biological consumption of bio-logically produced methane goes back to 27 Gyr (Hayes1994) providing evidence for the distinctness of eubac-teria (methanotrophs) and archaebacteria (methanogens)by that time at the very least and furthermore indicatingthat arguments for an origin of archaebacteria only850 million years (Myr) ago (Cavalier-Smith 2002) mustbe in error

Those geological benchmarks provide rough but robustorientation for the process of early evolution the origin ofprokaryotes by at least 35 Gyr and the origin of eukary-otes by at least 15 Gyr But biologists and geologists alikewould like to know a few more details What happenedduring those first 25 Gyr How did life get off the groundto begin with Why did it take eukaryotes so long toappear Why are the two groups of contemporary prokary-otes (eubacteria and archaebacteria) so genetically coher-ent yet so biochemically different Why are there nointermediate forms between prokaryotic and eukaryoticorganization Where did the nucleus come from Herewe discuss evolutionary processes that may underlie thedifferences between eubacteria and archaebacteria andthat may underlie the differences between prokaryotes andeukaryotes as well as the mosaic patterns of attributes(Zillig 1991) shared between the three

2 COMMONALITIES AND DIFFERENCES LEADQUICKLY TO DEEPER PROBLEMS

Understanding the evolutionary differences betweenarchaebacteria and eubacteria requires a concrete pictureof their last common ancestor This is because one canonly discuss the differentiation of their attributes on thebasis of the attributes possessed by their progenitor Butbecause there is no direct evidence for the nature of their


last common ancestor its attributes can only be addressedthrough logical inference (Penny amp Poole 1999)

What did the ancestor of prokaryotes possess for sureIt certainly possessed the universal genetic code tRNA(Eigen amp Schuster 1978 Woese et al 1990 Eigen 1992)and ribosomes with a battery of ribosomal proteins(Woese 2002) most of which are neatly encoded in a con-served superoperon that is present and conserved in geneorder across many genomes (Wachtershauser 1998) Fur-thermore the universal ancestor surely possessed DNA(Reichard 1997 Poole et al 2000) and DNA polymerasesRNA polymerases and a battery of accessory translationfactors such as EF-Tu and EF-G that are present in allprokaryotes (Poole et al 1998 1999) probably along withother more or less ubiquitous prokarytotic proteins suchas F1ndashF0-type ATPases (Gogarten et al 1989) as well asprokaryotic forms of the signal recognition particle whichis an accessory to translation (Brinkmann amp Philippe1999) Furthermore it had to have had all of the pathwaysnecessary for efficient nucleotide and amino-acidbiosyntheses a core carbon metabolism to provide thebiosynthetic precursors and the cofactors required forthose biosynthetic pathways but this is the point wherethings get very complicated very quickly because althougharchaebacteria and eubacteria do many things similarlywhen it comes to the processing of genetic information(Woese 2002) they do many things fundamentally differ-ently when it comes to central metabolism

One simple example for such core metabolic differencesbetween the two groups of prokaryotes is the breakdownof glucose to pyruvate the glycolytic (EmbdenndashMeyerhof)pathway which is universal among eukaryotes but notamong prokaryotes If we look among prokaryotic gen-omes for the familiar glycolytic enzymes that occur ineukaryotes we see that they have good homologuesamong a wide spectrum of eubacterial genomes but manyof the corresponding homologues are missing in archaeb-acteria (figure 1) This observation might seem puzzlingto some (Canback et al 2002) but is easily attributable totwo simple factors First although archaebacteria possesspathways from glucose to pyruvate they have several alter-native pathways that involve different enzymatic stepshence different enzymes not present in glycolysis(Daniel amp Danson 1995 Schonheit amp Schafer 1995Kengen et al 1996 Selig et al 1997) These include (i)the modified EmbdenndashMeyerhof pathway which employsADP- instead of ATP-dependent steps and a glyceral-dehyde-3-phosphate oxidoreductase in place of the stepscatalysed by GAPDH and 3-phosphoglycerate kinase (ii)the EntnerndashDoudoroff pathway which does not requirethe enzymatic steps catalysed by GPI PFK FBA or TPIand (iii) the non-phosphorylated EntnerndashDoudoroff path-way which shares only two enzymatic steps in commonwith glycolysis (Selig et al 1997) those catalysed by enol-ase (Hannaert et al 2000) and pyruvate kinase (Schrammet al 2000) Second several archaebacteria have completeglycolytic pathways but for many steps they use totallydifferent enzymes that are unrelated to the familiar homol-ogues found in eubacteria (and eukaryotes) (Schonheit ampSchafer 1995 Selig et al 1997) These include (i) anADP-dependent (rather than ATP-dependent) hexokin-ase with no sequence similarity to the eubacterial enzyme(Ito et al 2001) (ii) an ADP-dependent PFK that lacks



80_8970_7960_6950_5940_4930_3920_29

no hit or gt20


HumanSaccharomyces



Saccharomyces (HK)



Saccharomyces (ATP)



Saccharomyces (II)


GiardiaHuman

Saccharomyces


GiardiaHuman

Saccharomyces


HumanSaccharomyces




GiardiaHuman

Saccharomyces


GiardiaHuman

Saccharomyces

The

rmot

oga

Aqu

ifex

Bac

illus

Lact

ococ

cus

Myc

obac

teriu

mS

ynec

hocy

stis

Esc

heric

hia

Cau

loba

cter

Chl

amyd

iaB

orre

liaM

agne

toco

ccus

Agr

obac

teriu

mM

esor

hizo

bium

Sin

orhi

zobi

umP

seud

omon

as

Aer

opyr

mS

ulfo

lobu

sH

alob

acte

rium

Met

hnob

acte

rium

Met

hano

cocc

usP

yroc

occu

s ab

ysii

Pyr

ococ

cus

horik

oshi

iA

rcha

eogl

obus

The

rmop

lasm

a ac

idop

hilu

mT

herm

opla

sma

volc

anii

GKHK

GPI

PFK

FBA

TPI

GAP

PGK

PGM

Enolase

PYK


















100 mm

4020 mm 301012 mm

20 mm

(a) (b)

(c)

(d)






DNA era

RNP era

RNA era

prebioticchemistry

geochemistry

HS_

H2CO CH3SH

CN_

CH4

NH3H2 CO


crust

ocean

Ni2+

Fe2+gtgtFe3+

CO2

H+

P2O7



lt30 C pH ca55


reducedcarbon

compoundsCH3COSR

organic precursors

4_

FeS

NiS

deg

deg


























6 HOLOCHIRALITY
































at cell surface





eubacteria

archaebacteria

Fe2+

H+

Fe3+

Ni2+

HCO3_






[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin


fumarate reductase

















































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY




80_8970_7960_6950_5940_4930_3920_29

no hit or gt20


HumanSaccharomyces



Saccharomyces (HK)



Saccharomyces (ATP)



Saccharomyces (II)


GiardiaHuman

Saccharomyces


GiardiaHuman

Saccharomyces


HumanSaccharomyces




GiardiaHuman

Saccharomyces


GiardiaHuman

Saccharomyces

The

rmot

oga

Aqu

ifex

Bac

illus

Lact

ococ

cus

Myc

obac

teriu

mS

ynec

hocy

stis

Esc

heric

hia

Cau

loba

cter

Chl

amyd

iaB

orre

liaM

agne

toco

ccus

Agr

obac

teriu

mM

esor

hizo

bium

Sin

orhi

zobi

umP

seud

omon

as

Aer

opyr

mS

ulfo

lobu

sH

alob

acte

rium

Met

hnob

acte

rium

Met

hano

cocc

usP

yroc

occu

s ab

ysii

Pyr

ococ

cus

horik

oshi

iA

rcha

eogl

obus

The

rmop

lasm

a ac

idop

hilu

mT

herm

opla

sma

volc

anii

GKHK

GPI

PFK

FBA

TPI

GAP

PGK

PGM

Enolase

PYK


















100 mm

4020 mm 301012 mm

20 mm

(a) (b)

(c)

(d)






DNA era

RNP era

RNA era

prebioticchemistry

geochemistry

HS_

H2CO CH3SH

CN_

CH4

NH3H2 CO


crust

ocean

Ni2+

Fe2+gtgtFe3+

CO2

H+

P2O7



lt30 C pH ca55


reducedcarbon

compoundsCH3COSR

organic precursors

4_

FeS

NiS

deg

deg


























6 HOLOCHIRALITY
































at cell surface





eubacteria

archaebacteria

Fe2+

H+

Fe3+

Ni2+

HCO3_






[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin


fumarate reductase

















































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY

















100 mm

4020 mm 301012 mm

20 mm

(a) (b)

(c)

(d)






DNA era

RNP era

RNA era

prebioticchemistry

geochemistry

HS_

H2CO CH3SH

CN_

CH4

NH3H2 CO


crust

ocean

Ni2+

Fe2+gtgtFe3+

CO2

H+

P2O7



lt30 C pH ca55


reducedcarbon

compoundsCH3COSR

organic precursors

4_

FeS

NiS

deg

deg


























6 HOLOCHIRALITY
































at cell surface





eubacteria

archaebacteria

Fe2+

H+

Fe3+

Ni2+

HCO3_






[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin


fumarate reductase

















































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY




DNA era

RNP era

RNA era

prebioticchemistry

geochemistry

HS_

H2CO CH3SH

CN_

CH4

NH3H2 CO


crust

ocean

Ni2+

Fe2+gtgtFe3+

CO2

H+

P2O7



lt30 C pH ca55


reducedcarbon

compoundsCH3COSR

organic precursors

4_

FeS

NiS

deg

deg


























6 HOLOCHIRALITY
































at cell surface





eubacteria

archaebacteria

Fe2+

H+

Fe3+

Ni2+

HCO3_






[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin


fumarate reductase

















































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY























6 HOLOCHIRALITY
































at cell surface





eubacteria

archaebacteria

Fe2+

H+

Fe3+

Ni2+

HCO3_






[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin


fumarate reductase

















































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY
































at cell surface





eubacteria

archaebacteria

Fe2+

H+

Fe3+

Ni2+

HCO3_






[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin


fumarate reductase

















































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY



[Fe] hydrogenase

[4Fe4S] ferredoxin

[2Fe2S] ferredoxin


fumarate reductase

















































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY











































EVOLUTION







12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY




12 CONCLUSION







REFERENCES






































































































































































































































































































GLOSSARY




















































































































































































































































































GLOSSARY


On the origins of cells: a hypothesis for the evolutionary ... the origin of cells.pdf ·...

Documents

Transcript of On the origins of cells: a hypothesis for the evolutionary ... the origin of cells.pdf ·...