MedicineLakex

 

Mbev-20-09-17 1564.1574

Single Eubacterial Origin of Eukaryotic Sulfide:Quinone Oxidoreductase,a Mitochondrial Enzyme Conserved from the Early Evolution of EukaryotesDuring Anoxic and Sulfidic Times Ursula Theissen,* Meike Hoffmeister,* Manfred Grieshaber,  and William Martin**Institute of Botany III and  Institut fu¨r Zoophysiologie, University of Du¨sseldorf, Du¨sseldorf, Germany Mitochondria occur as aerobic, facultatively anaerobic, and, in the case of hydrogenosomes, strictly anaerobic forms.
This physiological diversity of mitochondrial oxygen requirement is paralleled by that of free-living a-proteobacteria, thegroup of eubacteria from which mitochondria arose, many of which are facultative anaerobes. Although ATP synthesis inmitochondria usually involves the oxidation of reduced carbon compounds, many a-proteobacteria and somemitochondria are known to use sulfide (H2S) as an electron donor for the respiratory chain and its associated ATPsynthesis. In many eubacteria, the oxidation of sulfide involves the enzyme sulfide:quinone oxidoreductase (SQR).
Nuclear-encoded homologs of SQR are found in several eukaryotic genomes. Here we show that eukaryotic SQR genescharacterized to date can be traced to a single acquisition from a eubacterial donor in the common ancestor of animalsand fungi. Yet, SQR is not a well-conserved protein, and our analyses suggest that the SQR gene has furthermoreundergone some lateral transfer among prokaryotes during evolution, leaving the precise eubacterial lineage from whicheukaryotes obtained their SQR difficult to discern with phylogenetic methods. Newer geochemical data and microfossilevidence indicate that major phases of early eukaryotic diversification occurred during a period of the Earth's historyfrom 1 to 2 billion years before present in which the subsurface ocean waters contained almost no oxygen but containedhigh concentrations of sulfide, suggesting that the ability to deal with sulfide was essential for prokaryotes and eukaryotesduring that time. Notwithstanding poor resolution in deep SQR phylogeny and lack of a specifically a-protebacterialbranch for the eukaryotic enzyme on the basis of current lineage sampling, a single eubacterial origin of eukaryotic SQRand the evident need of ancient eukaryotes to deal with sulfide, a process today germane to mitochondrial quinonereduction, are compatible with the view that eukaryotic SQR was an acquisition from the mitochondrial endosymbiont.
ATP synthesis in most mitochondria involves the oxygen respiration (National Research Council 1979; generation of a proton gradient with the help of the Grieshaber and Vo¨lkel 1998). At sulfide concentrations electron transport chain through respiratory complexes I to below approximately 20 lM, electrons from sulfide can be IV. Yet, numerous exceptions to that rule occur among donated to O2 as the terminal acceptor. Sulfide concen- anaerobic mitochondria, which can bypass complex IV or trations in the range of 10 to 50 lM inhibit the electron use alternative terminal acceptors other than O2, such as transfer from cytochrome c to complex IV (Bagarinao and nitrate or fumarate (Tielens et al. 2002). Further exceptions Vetter 1990; Grieshaber and Vo¨lkel 1998). At higher are hydrogenosomes, anaerobic mitochondria that generate sulfide concentrations, mitochondrial complex IV is ATP through substrate level phosphorylation without the blocked and the electrons are donated to alternative, yet help of a proton gradient (Martin and Mu¨ller 1998). In all unknown, acceptors, possibly involving a similar alterna- hydrogenosomes and in most mitochondria, ATP synthesis tive oxidase as is found in plants (Vo¨lkel and Grieshaber involves the use of reduced carbon compounds as the 1996a; Parrino, Kraus, and Doeller 2000). Since sulfide electron donor in one or a series of redox reactions. Yet concentrations can reach approximately 20 mM in en- some mitochondria can synthesize ATP chemolithotroph- vironments inhabited by marine invertebrates (Fenchel and ically, using proton gradients that are generated with the Riedel 1970; Vo¨lkel, Hauschild, and Grieshaber 1995), help of electrons taken not from carbon compounds, but and since sulfide is a potent toxin, both the energy- rather from sulfide (H2S, HS
, and S2
), which is abundant producing and detoxifying functions of sulfide oxidation in many anaerobic environments, such as marine sedi- are essential to mitochondrial function in these organisms.
ments (Vo¨lkel, Hauschild, and Grieshaber 1995). In In eubacteria, sulfide oxidation is commonly catalyzed mitochondria of marine invertebrates from such environ- by the flavoprotein, sulfide:quinone oxidoreductase (SQR) ments, electrons from sulfide can be transferred to (also called sulfide quinone reductase). The biochemistry of quinones as their entry point into the mitochondrial eubacterial SQR has been characterized in some detail respiratory chain for ATP synthesis (Doeller et al. 1999;Doeller, Grieshaber, and Kraus 2001).
(Reinartz et al. 1998; Griesbeck, Hauska, and Schu¨tz The terminal acceptor for electrons stemming from 2000). Eubacterial SQR catalyzes the reaction H2S þ sulfide in mitochondria depends upon the sulfide concen- Ubiquinone $ [S60] þ UbiquinoneH2 (Griesbeck et al.
tration itself. This is because sulfide is a strong inhibitor of 2002). The enzyme has been purified and cloned fromRhodobacter capsulatus (Schu¨tz et al. 1997) and Oscilla-toria limnetica (Arieli et al. 1994) and biochemically char- Key words: sulfide-quinone reductase, endosymbiosis, hydrogeno- acterized in Chlorobium limicola (Shahak et al. 1992), somes, electron transport chain, anaerobiosis.
Rhodobacter capsulatus (Shahak et al. 1994), Paracoccus denitrificans (Schu¨tz et al. 1998), Allochromatium vinosum Mol. Biol. Evol. 20(9):1564–1574. 2003 (Reinartz et al. 1998), and Aquifex aeolicus (Nu¨bel et al. 2000) (reviewed in Griesbeck, Hauska, and Schu¨tz Molecular Biology and Evolution, Vol. 20, No. 9,
Society for Molecular Biology and Evolution 2003; all rights reserved.
2000). Bacterial SQR is a single polypeptide with an Single Eubacterial Origin of Eukaryotic Sulfide apparent molecular mass of 48 to 55 kDa, is possibly active reviewed in Martin et al. 2001). In general, there are four as a dimer, is membrane associated, belongs to the simple possibilities for the origin of mitochondrial sulfide glutathione reductase family of flavoproteins, and is oxidation in eukaryotes. (1) The host that acquired the inhibited by quinone analogs at micromolar or nanomolar mitochondrion could have possessed an SQR enzyme that concentrations (Arieli et al. 1994; Schu¨tz et al. 1997; was retargeted to the mitochondrion to become functional Griesbeck et al. 2002). Rhodobacter SQR was shown to there; in this case eukaryotic SQR should be related to reside in the periplasm (Schu¨tz et al. 1997). In Chlorobium archaebacterial SQR because the DNA replication (Tye and Rhodobacter, electrons from sulfide enter into the 2000), translation (Lecompte et al. 2002), transcription electron transport chain of anaerobic photosynthesis (Reeve 2003), and chromatin-packaging systems (Reeve through SQR (Griesbeck, Hauska, and Schu¨tz 2000; 2003) of eukaryotes are specifically related to their Griesbeck et al. 2002). In the nonphotosynthetic a- archaebacterial homologs. (2) The mitochondrial symbiont proteobacterium Paracoccus denitrificans (Schu¨tz et al.
could have possessed the SQR enzyme, whereby the gene 1998) and in Aquifex aeolicus (Nu¨bel et al. 2000), SQR must have been transferred to the host's chromosomes, introduces electrons from sulfide into the respiratory chain since SQR is not encoded in any mitochondrial DNA; in (reviewed in Griesbeck, Hauska, and Schu¨tz 2000).
this case eukaryotic SQR should reveal a single eubacterial Mitochondrial SQR activity has been indirectly origin. (3) Neither host nor symbiont may have possessed measured in many organisms through the mitochodrion- SQR, and the SQR gene could have been acquired through dependent formation of thiosulfate from sulfide, for horizontal gene transfer in organisms that inhabit sulfidic example, in the annelids Heteromastus filiformis and environments; in this case eukaryotic SQR should reveal Arenicola marina as well as in the molluscs Solemya reidi multiple origins from diverse prokaryotic donors. (4) and Geukensia demissa (Oeschger and Visman 1994; Eukaryotic SQR is an invention specific to the eukaryotic reviewed in Grieshaber and Vo¨lkel 1998), but has not been lineage; in this case eukaryotic SQR should be unrelated to purified to homogeneity from any multicellular eukaryote.
prokaryotic SQR. The phylogenetic distribution of SQR- A functional mitochondrial SQR was, however, recently related enzymes among various genomes has been cloned and characterized from the ascomycete Schizo- previously studied (Vande Weghe and Ow 1999; Gries- saccharomyces pombe (Vande Weghe and Ow 1999). S.
beck, Hauska, and Schu¨tz 2000), but the phylogeny of pombe SQR showed marked sequence similarity to the SQR itself has not.
SQR purified and extensively characterized at the bio- SQR was almost certainly an essential and possibly chemical level from the a-proteobacterium Rhodobacter ubiquitous enzyme during the phase of eukaryotic evo- capsulatus (Schu¨tz et al. 1997) and was furthermore lution from 2 to 1 billion years ago, since newer geo- shown to be imported into and functional in S. pombe chemical evidence indicates that the Earth's ocean waters mitochondria (Vande Weghe and Ow 1999). Mitochon- were anoxic and very sulfidic during that time (Canfield drial SQR activity has been most extensively studied in 1998; Shen, Buick, and Canfield 2001; Anbar and Knoll marine invertebrates that inhabit sulfide-rich intertidal 2002), findings that underscore the evolutionary impor- sediments, most notably the annelid lugworm Arenicola tance of anaerobic biochemistry in both ancient and marina and the ribbed mussel Geukensia demissa (Vo¨lkel modern eukaryotes (Tielens et al. 2002; Embley et al.
and Grieshaber 1996a; Parrino, Kraus, and Doeller 2000).
2003). Here, we report the occurrence of SQR and SQR- Both in those organisms (Doeller et al. 1999; Doeller, related enzymes among genomes of eubacteria, archaebac- Grieshaber, and Kraus 2001; Vo¨lkel and Grieshaber teria, and eukaryotes and examine the phylogeny of SQR, 1996b) and in recent biochemical studies of SQR from with particular attention to the evolutionary origin of the chicken mitochondria (Yong and Searcy 2001) it was eukaryotic nuclear genes for mitochondrial sulfide:quinone shown that mitochondrial sulfide consumption was coupled to ATP synthesis. Since chickens do not inhabitsulfide-rich environments, the role of SQR in their Materials and Methods mitochondria is probably not ATP production, but mayinvolve detoxification. The primary oxidation product of Database searching was performed using the se- sulfide produced by mitochondrial SQR is still not known quences for the functionally characterized SQR from with certainty. The two-electron reaction would yield S. pombe (Vande Weghe and Ow 1999) and from Rb.
either elemental sulfur (S60) or sulfanes (HSSnH) as the capsulatus (Schu¨tz et al. 1997) as queries against GenBank primary oxidation product. The four-electron reaction and against unfinished microbial genomes listed by the would yield thiosulfate (S ), which is the most Institute for Genome Research (http://tigrblast.tigr.org/ commonly detected oxidation product (O'Brien and Vetter ufmg/). Sequence handling, data formatting, and alignment 1990; Vo¨lkel and Grieshaber 1992; Johns et al. 1997).
were performed with programs of the GCG Package Recent results by Yong and Searcy (2001) suggest that version 10.3 (Genetics Computer Group, Madison, Wis.) sulfanes might be produced during mitochondrial sulfide and with ClustalW (Thompson, Higgins, and Gibson oxidation, but sulfanes have still not been directly 1994). Alignments were reinspected and manually adjust- ed. For phylogenetic analyses, programs of the PHYLIP The ability of mitochondria to perform sulfide package (Felsenstein 1998), the MOLPHY package oxidation for ATP synthesis raises the question as to the (Adachi and Hasegawa 1996), and Puzzle (Strimmer and evolutionary origin of eukaryotic SQR genes, particularly von Haeseler 1996) were used. Uncorrected proportions from the standpoint of endosymbiotic theory (recently of differences between sequences (p-distances) were Theissen et al.
FIG. 1.—Conserved regions from the alignment of SQR sequences (accession numbers given) corresponding to Cys 159, the FAD binding domain III, and Cys 353 using the numbering of Griesbeck et al. (2002). The residue number of the cysteines and the conserved glycine in FAD binding domainIII (boldface type) in each sequence is given. Group designations (I, II, and III) refer to families of sequence similarity. Species names designated asSQR indicate that the sequence is known to encode SQR acitvity; those marked with an asterisk possess the N-terminal translational fusion indicated infigure 2. Gaps are indicated as dashes. Nostoc sp. PCC 7120 is synonymous with Anabaena PCC 7120.
calculated with ClustalW (Thompson, Higgins, and a guide, and using the terminology of Griesbeck et al.
Gibson 1994). Logdet distances were calculated with the (2002), we manually identified the conserved regions LDDist program available at the Web site http://artedi.ebc.
corresponding to SQR fingerprint 2, containing Cys 159, and SQR fingerprint 5, containing Cys 353, as well as and Moulton 2002) networks for representing the data FAD binding domain III in all sequences shown in figure 1.
were constructed with the software available at the Web These anchor points of sequence conservation were used for further manual refinement of the alignment. Residues ProtML and Puzzle, the JTT-F matrix was used. Quartet- Cys 159 and Cys 353, originally identified by Bronstein et puzzling (QP) was employed using a gamma distribution al. (2000), refer to their positions in the SQR sequence and eight categories of rate heterogeneity.
from Rb. capsulatus (Griesbeck et al. 2002). The spacingof these conserved motifs is generally uniform throughoutthe alignment. Two further sequence motifs identified by Griesbeck et al. (2002) in the analysis of seven prokaryotic Conservation and Fusions Among SQR-Related Proteins SQR homologs surrounding the residues Cys 127 and His Using Blast, SQR homologs were identified in ge- 196 are not conserved across all 37 SQR homologs shown nomes of eubacteria, archaebacteria, and eukaryotes and in figure 1 and hence are not shown. Patterns of sequence retrieved from the databases. Because of poor sequence similarity visible in the alignment around Cys 159, Cys conservation, the alignment of these sequences is challeng- 353, and the FAD binding domain suggested the presence ing. For example, the functionally characterized SQR of three distinct groups of sequence diversity which we proteins from S. pombe and Rb. capsulatus share only 24% term groups I, II, and III. Although not thoroughly amino acid identity in the pairwise Needleman-Wunsch exhaustive with respect to all unfinished genomes, this alignment. As a consequence, the automatic alignment database search provides a representative overview of programs ClustalW and Pileup align neither the FAD sequence diversity for the SQR family, including all binding domains nor strictly conserved cysteine residues functionally characterized members. Across groups I, II, identified by Bronstein et al. (2000) and Griesbeck et al.
and III, amino acid identity in pairwise Needleman- (2002) in their analyses of SQR and related sequences Wunsch comparisons of unaligned sequences was approx- from prokaryotes. Using those conserved domains as imately 25% on average. Of the sequences indicated in Single Eubacterial Origin of Eukaryotic Sulfide FIG. 2.—Conserved regions in an ORF N-terminally fused present to several SQR homologs (indicated with an asterisk next to the accessionnumber; see also fig. 1) and present as an independent ORF in severalgenomes (no asterisk). Sequences marked with two asterisks possess attheir C-terminus not SQR, but rather an unrelated ORF that is annotatedin several genomes on the basis of weak sequence similarity as a b- FIG. 3.—Schematic overview of fusions involving SQR. The lactamase (see text). The strictly conserved cysteine is marked with ‘‘*,'' positions of conserved cysteine residues is indicated by dotted lines. (a) and the partially conserved cysteine is marked with ‘‘j.'' Their positions SQR as it occurs in most prokaryotes. (b) SQR as it occurs in S. pombe. T are indicated as in figure 1.
indicates the mitochondrial transit peptide identified (Vande Weghe andOw 1999). (c) SQR as it occurs in sequences marked with an asterisk infigures 1 and 2. (d) Unfused ORF present in many prokaryotic genomes figure 1, only the homologs from Rb. capsulatus (Schu¨tz et as indicated in figure 2. (e) Fusion of the ORF with an SQR-unrelatedsequence as it occurs in sequences marked with two asterisks in figure 2.
al. 1997), from S. pombe (Vande Weghe and Ow 1999),and from the cyanobacteria Aphanonthece and Oscillatoria function (fig. 3) that shares no similarity with SQR and (Bronstein et al. 2000) have been shown to represent active that is annotated in several entries as a member of the SQR enzymes.
metallo-b-lactamase family by virtue of its similarity to the The partial sequence of an SQR homolog from PFAM (Bateman et al. 2002) lactamase B family Dictyostelium discoideum that contained the three strictly (PF00753; http://pfam.wustl.edu/), which includes several conserved motifs was assembled from unannotated sequence data (http://db.dictybase.org/) but lacked 125N-terminal residues relative to the S. pombe sequence (fig.
1). One additional SQR homolog each was identified in theCaenorhabditis elegans (GenBank accession number Phylogenetic analyses started with the complete NP_500688) and S. pombe (GenBank accession number alignment containing 36 OTUs (operational taxonomical T43278) genomes, but these sequences appeared to be C- units; sequences) and 715 positions, only seven of which terminally truncated and lacked the Cys 353 region. The were constant. Recalling that these sequences are highly Caenorhabditis homolog shown (GenBank accession divergent (,25% identity in many comparisons), and number NP_502729) possesses a unique approximately recalling that neither Pileup using the PAM250 matrix nor 50–amino acid insertion at about position 230 that might ClustalW using the Blosum matrices recovered the represent a translated intron, but we were unable to conserved cysteines (fig. 1) for all sequences, we started identify possible intron donor and acceptor sites in the analyses with a NeighborNet network of protein logdet nucleotide sequence that would justify its removal.
distances to obtain graphical representation of the data (fig.
Searches among the genome sequence data for the 4a). Neighbor-joining (NJ) (Saitou and Nei 1987) trees anaerobic protists Giardia intestinalis and Entamoeba using the uncorrected proportion of amino acid differences histolytica revealed no identifiable SQR homologs. No (p-distance or Hamming distance) as a distance measure SQR homologs were detected among photosynthetic (NJP) gave a very similar picture. Although the p-distance eukaryotes, although SQR detects glutathione reductase is a rough measure, it has been shown to perform well in as a distant relative in several eukaryotic genomes in computer simulations when sequence divergence is high BLAST searches (data not shown).
because it has a lower variance than correction procedures The SQR homologs from Nitrosomonas europaea, (Nei 1996; Nei and Kumar 2000). At the 95% bootstrap Burkholderia fungorum, and Ralstonia metallidurans proportion (BP) level, the NJ tree of p-distances recovered possess a well-conserved approximately 180–amino acid all branches that protein logdet did plus only one open reading frame (ORF) N-terminal to the SQR domain additional branch—that joining Archaeoglobus and Mag- that is present as an independent ORF of unknown netospirillum, which was found at BP ¼ 93% using protein function in several prokaryotic genomes (fig. 2) but does not always co-occur with SQR. This N-terminal ORF The groups I, II, and III identifiable in the alignment contains one strictly conserved cysteine and an additional were recovered in the NJP topology, although the BP for cysteine conserved in some sequences (fig. 2). Homologs group III was only 73% using NJP (48% using logdet).
of the N-terminal ORF from Agrobacterium, Sinorhi- Local rearrangement using ProtML starting from the NJ zobium, Mesorhizobium, and Xylella are fused with yet tree of ML distances and using the JTT-F matrix strongly another approximate 240–amino acid ORF of unknown separated groups I, II, and III at BP greater than 95% each Theissen et al.
FIG. 4.—Phylogenetic analyses of SQR sequences. (a) NeighborNet (NNet) network of protein logdet distances among SQR sequences. The NNet network depicts splits in the data as series of parallel lines, conflicting or non–tree-like signals for a given taxon or group of taxa can thus berepresented. For example, in addition to the split linking Rhodospirillum and Pasteurella, there is also one linking Rhodospirillum and Sulfolobustokadaii (a conflicting or non–tree-like signal). The splits joining members of group I and group II, respectively, are indicated. Rhodobacter isabbreviated as ‘‘Rb.'' (b) ProtML tree using the JTT-F matrix for sequences that pass the significance test for amino acid compositional equilibrium atP ¼ 0.95. (c) ProtML tree using the JTT-F matrix for sequences from group II that pass the significance test for amino acid compositional equilibrium atP ¼ 0.95. Bootstrap support for the monophyly of eukaryotic SQR in with various methods and for various subsets of the data is summarized in the boxbelow the tree in (c) (see text). Branches supported at a BP 95% with the method indicated are labeled with a dot; for the NNet graph, dots indicate Single Eubacterial Origin of Eukaryotic Sulfide and furthermore grouped the eukaryotic sequences to- disulfide oxidoreductase flavoprotein family (fig. 5). For gether (not shown) but with a low BP (59%). However, the example, in Blast comparisons to GenBank, FCC members amino acid composition of 14 sequences in the 36 OTU detect each other at E-values roughly less than 10
30 and data deviated from the expectation at P ¼ 0.95 as estimated usually share more than 25% sequence identity in pairwise with Puzzle: Thermosynechococcus, Magnetococcus, Rb.
comparisons, but FCC members detect SQR members at sphaeroides, P. syringae, P. aeruginosa, R. metallidurans, E-values roughly greater than 10
10, with which they R. solanacearum, Staphylococcus, Rhodospirillum, Ar- usually share less than 20% sequence identity in pairwise chaeoglobus, Desulfovibrio, Ferroplasma, S. tokodaii, and comparisons, and vice versa. Other members of the DiSR Thermoplasma. Removing these from the alignment (22 family were detected at much lower similarity levels. This OTU data) and rechecking revealed that Burkholderia indicates that SQR and FCC are distinct but specifically deviated, upon removal of which all remaining 21 related subfamilies within the larger family of DiSR sequences passed the significance test for amino acid flavoproteins. Notably, many of the prokaryotic genomes composition. ProtML analysis of that data after excluding surveyed here encode both SQR and FCC homologs: gapped sites (21 OTUs, 343 positions) produced the Chlorobium (one SQR and two FCCs), Magnetospirillum, topology in figure 4b, which had a very low BP (, 50%) R. solanacaerum, R. metallidurans, Magnetococcus, for the monophyly of the eukaryotic sequences, support for Aquifex, Rb. sphaeroides, Paracoccus, Solfolobus, and which increased to 86% when gapped sites were included (not shown).
However, the 21 OTU data (fig. 4b) was still highly Sequence Conservation and Motifs in SQR divergent, with between group (I, II, and III) amino acididentity of sequences from the multiple alignment in the Sequences such as SQR that share only about 20% range of only 20%. To examine the possible monophyly of identity in many pairwise comparisons pose a challenge to eukaryotic SQR more closely, we investigated the phylogenetic analysis. In such cases, phylogeny is strongly phylogeny of the 18 sequences belonging to group II in aided by information from three-dimensional structures figure 4a. This data set (18 OTUs, 681 sites including (Schu¨tz et al. 2000; Baymann et al. 2003). Crystal gaps) was inspected for amino acid compositional bias, structures are available for several members of the DiSR whereby P. aeruginosa, R. solanacearum, and Staphylo- flavoprotein family, including of FCC from Chromatium coccus failed the frequency distribution test at P ¼ 0.95.
vinosum (Chen et al. 1994), but there are currently no three- Removal of these sequences from the data and rechecking dimensional structures available for SQR. In studies of revealed that Rhodospirillum failed, leaving 14 OTUs, all ancient protein phylogeny or poorly conserved sequences, of which passed the amino acid frequency distribution test.
as with the present SQR data, the discrepancy between The resulting alignment (14 OTUs, 681 sites) contained sequence conservation and structure conservation can better sequence conservation with 43 invariant sites and all become severe, as underscored by Rost (1997), who found aligned sequences being at least 30% identical in all that the majority of pairwise comparisons among protein comparisons, although this is still generally poor sequence sharing a common structure reveal only about 8% to 9% conservation. Nonetheless, analysis with ML, NJ, and QP, sequence identity. At low sequence identity, many also after exclusion of gapped positions (14 OTUs, 386 assumptions of phylogenetic methods are almost certainly sites), provided good support for the monophyly of violated. For example, recent computer simulations showed eukaryotic SQR (branch E in figure 4c) with all methods that the fraction of sites determined to be neutral by the except MP (fig. 4c).
criterion of protein folding thermodynamics fluctuates ina manner that depends upon the randomly chosen neutralmutations accumulated by the sequence as it mutates through sequence space (Bastolla et al. 2002). Such find- Sulfide:quinone oxidoreductase belongs to the larger ings suggest that at levels of sequence similarity where family of disulfide oxidoreductase (DiSR) flavoproteins structural constraints can cause different fractions of sites in that includes glutathion reductases, the lipoamid dehy- a protein to become neutral, current phylogenetic inference drogenase (E3) subunit of pyruvate dehydrogenase, thio- methods, which are founded in neutral theory, will reflect redoxin reductase, and—importantly—flavocytochrome c patterns of shared sequence similarity but will unlikely (Fcc or FCC) (Schu¨tz et al. 1997; Griesbeck et al. 2002).
recover the true tree. Despite these problems, if one wishes FCC is used by numerous prokaryotes for sulfide oxidation to study the evolution of SQR, one has to work with the as an alternative to SQR (Schu¨tz et al. 1997; Griesbeck degree sequence conservation that SQR has to offer.
et al. 2002). However, sequence comparisons revealed that Based on spectroscopic and mutational analyses, all SQR sequences examined here are more similar to each Griesbeck et al. (2002) proposed a mechanism for the SQR other than they are to FCC or other members of the that involved the participation of three cysteine residues, splits with a BP 95% in NJ trees of logdet distances. Eubacterial taxon designations are indicated for major recognized groups: a, b, c, d, a-proteobacteria, etc.; cy indicates cyanobacteria, and Gþ indicates gram-positives. Sequence groups I, II, and III are indicated at the periphery of the treein (a) and (b). Species names designated as SQR indicate that the sequence is known to encode SQR activity. Archaebacterial sequences are boxed.
Eukaryotic sequences are indicated in bold. The scale bar indicates 0.1 substitution per site with the respective method.
Theissen et al.
Cys 127, Cys 159, and Cys 353; however, they also notedthat a catalytic mechanism involving only two cysteineresidues instead of three would be compatible with theavailable data. We found that Cys 159 and Cys 353 arestrictly conserved in SQR homologs investigated here,but we found no evidence for a conservation of Cys 127outside of group I. Cysteine residues were lacking inseveral SQR sequences within 50 amino acids N-terminalof Cys 127 to Cys 159 (alignment available upon request)and also in the S. pombe SQR, the only sequence outsideof group I that has been shown to be directly involved insulfide oxidation.
Noting that Cys 127 is missing in S. pombe SQR, Griesbeck et al. (2002) pointed out that the Km values ofroughly 2 mM each for sulfide and quinone measured forS. pombe SQR are 1000-fold higher than for the FIG. 5.—Schematic phylogeny depicting sequence similarity shared between SQR and FCC relative to other members of the DiSR family as eubacterial enzymes. This low substrate affinity could, in estimated by Blast results and sequence identity in pairwise comparisons.
principle, cast doubt on the functional identity of S. pombeSQR (HMT2, the product of the hmt2 gene) as a functionalSQR enzyme. However, Vande Weghe and Ow (1999)showed (1) that isolated mitochondria from hmt2þ S.
SQR homologs detected in sequenced archaebacterial pombe cells could reduce exogenous quinones with genomes in addition to eubacterial homologs from the sulfide, whereas hmt2
could not, (2) that hmt2þ S. pombe sulfate reducer Desulfovibrio, from the anaerobic, photo- cells could oxidize endogenously produced sulfide, synthetic, green sulfur bacterium Chlorobium, and from whereas hmt2
could not, (3) that HMT2 produced in E.
the a-proteobacterium Magnetospirillum. SQR sequences coli is a flavoprotein, (4) that his-tagged HMT2 purified in group II (fig. 4a and b) comprise eubacterial and from E. coli reduces quinones in a sulfide-dependent manner in vitro, albeit with poor kinetic constants, and (5) Lateral gene transfer (LGT) exists among prokaryotes that HMT2 resides in mitochondria. Thus, despite the high and has become a major issue in gene and genome evolution (Gogarten, Doolittle, and Lawrence 2002). The m values measured for S. pombe HMT2 produced in E.
coli (which lacks SQR), the brunt of evidence indicates present analysis suggests that also SQR genes may have that S. pombe HMT2 is an active mitochondrial SQR, but transferred among prokaryotes during evolution, judging there remains the possibility that it requires an additional from the interleaving of eubacterial and archae distribution subunit or factor not required by group I SQR for full of a-proteobacterial homologs (fig. 4a). However, when activity. This possibility and the lack of Cys 127 in several sequences possessing significant amino acid bias are SQR sequences drew our attention to the ORF of unknown removed from the data (fig. 4b), the degree of interleaving function translationally fused to the N-terminus of group I also decreases, suggesting that a phylogenetic argument SQR from Nitrosomonas europaea, Burkholderia fungo- for unrestricted LGT of SQR genes among prokaryotes rum, and Ralstonia metallidurans. These possess a strictly cannot be made for these data. However, the presence of conserved cysteine residue at position 40 of the Nitro- four robustly clustering cyanobacterial SQR genes in somonas sequence (fig. 2). Database searching revealed group I and the presence of a single cyanobacterial SQR in that this ORF is present in numerous eubacterial genomes group II (Synechocystis) suggests that the Oscillatoria- and that Cys 40 is strictly conserved in all homologs, type SQR (group I) might represent the endogenous perhaps suggesting that it might be able to assume the cyanobacterial gene, whereas Synechocystis may have function of Cys 127 in SQR from group II. However we picked up its SQR from a proteobacterial donor (fig. 4c).
could not identify this ORF in all genomes whose group II The interleaving of b-proteobacterial and c-proteobacterial SQR lacks Cys 127, for example, Pasteurella multocida, homologs in figure 4c indicates further probable workings leaving the question of whether two or three cysteine of LGT for SQR among prokaryotes.
residues are involved in the SQR catalytic mechanism(Griesbeck et al. 2002) open from this standpoint.
A Single Origin of Eukaryotic SQR, a Eubacterial Relictfrom the Anoxic and Sulfidic Past SQR Sequence Diversity and Lateral Gene Transfer The present analyses provide evidence for a single Among Prokaryotes eubacterial origin of eukaryotic SQR, indicating that SQR homologs encompass three groups of sequence eukaryotes sampled here acquired the gene for mitochon- diversity that are nonuniformly distributed across eubac- drial SQR once in evolution from a eubacterial donor (fig.
teria, archaebacteria, and eukaryotes (figs. 1 and 4a and b).
4c). The nature of that eubacterial donor is highly relevant Group I contains the functionally characterized eubacterial to the issue of mitochondrial evolution and eukaryote SQR enzymes from cyanobacteria and Rhodobacter origins. There are two simple possiblities: the donor of capsulatus and furthermore contains only eubacterial the SQR gene either was the ancestor of mitochondria or homologs. Group III contains functionally uncharacterized Single Eubacterial Origin of Eukaryotic Sulfide On the one hand, arguing meekly against the view that (Tielens et al. 2002). The inheritance by mitochondria of the SQR donor was the mitochondrial endosymbiont is preexisting and functioning aerobic and anaerobic compo- a single finding, namely that eukaryotic SQR does not nents in the same electron transport chain from a faculta- specifically branch with a-proteobacterial SQR. Indeed, it tively anaerobic ancestor of mitochondria that was perhaps has recently been argued that any eukaryotic nuclear gene similar in overall physiology to facultatively anaerobic a- that is to be inferred to be of mitochondrial origin must proteobacteria, such as Rhodospirillum, Paracoccus de- be shown to branch specifically with a-proteobacterial nitrificans (John and Whatley 1975), or Rhodobacter, homologs (Kurland and Andersson 2000; Canback, which possess SQR (Schu¨tz et al. 1998; Griesbeck, Andersson, and Kurland 2002). However, that view is Hauska, and Schu¨tz 2000), seems much more likely than probably too simplistic for several reasons. First, SQR is not the piece-by-piece addition during eukaryotic evolution of a highly conserved protein, such that the early evolution of anaerobic components (SQR, rhodoquinone, etc.) to an this eukaryotic gene as viewed from the perspective of (hypothetical) ancestrally aerobic mitochondrial electron phylogenetics may have simply been obscured by mutation.
transport chain as envisaged by those who argue for an Sequence conservation in SQR permits one to trace the origin of mitochondria from strictly aerobic Rickettsia-like origins of the eukaryotic gene to eubacteria, but tracing it to parasites (Kurland and Andersson 2000).
any particular eubacterial lineage on the basis of 30% Second, the nuclear gene for SQR was apparently sequence identity is probably asking too much of acquired once in eukaryotic evolution, not several times as phylogenetic inference methods. Such loss of phylogenetic would be predicted under models that envisage lateral gene signal among poorly conserved genes has been well acquisition from food bacteria as the major source of documented in genome-wide phylogenies involving genes eubacterial, but apparently non–a-proteobacterial, genes in acquired from chloroplasts (Martin et al. 2002). Second, the eukaryotes (Doolittle 1998). Such single acquisition with overall pattern of sequence similarity in figure 4a suggests a phylogenetically unresolved eubacterial origin as seen that contemporary a-proteobacteria may themselves have for SQR is also observed for several other proteins acquired their SQR genes from different sources, as involved in anaerobic ATP synthesis in eukaryotes, for evidenced by the presence of a-proteobacterial SQR example, [Fe]-hydrogenase (Horner, Foster, and Embley homologs in group I (Rhodobacter), group II (Rhodospi- 2000; Horner et al. 2002), pyruvate:ferredoxin oxidore- rillum, very close to a c-proteobacterial homolog), and ductase (Horner, Hirt, and Embley 1999; Rotte et al. 2001; group III (Magnetospirillum, branching among archaebac- Embley et al. 2003), many glycolytic enzymes (Hannaert terial homologs). It is reasonable to assume that a- et al. 2000), and NADH oxidase (Nixon et al. 2002), not proteobacteria were undergoing LGT, also for SQR genes, to mention many other eukaryotic proteins that are not at the time of mitochondrial origins and subsequently.
involved directly in ATP synthesis, such as proteasome Incorporating LGT into evolutionary thinking thus makes it homologs HslV and HslU (Couvreur et al. 2002). Notably, difficult to pinpoint exactly which genes the ancestor of [Fe]-hydrogenase and pyruvate:ferredoxin oxidoreductase mitochondria possessed and/or contributed to eukaryotes possess several FeS clusters (Chabriere et al. 1999; on the basis of today's sequence comparisons. In other Peters 1999), as do several proteins of the mitochon- words, allowing for the existence of LGT during pro- drial respiratory chain (Burger et al. 1996; Friedrich and karyotic evolution (Gogarten, Doolittle, and Lawrence Schiede 2000). Recent findings indicate that many proteins 2002), no single contemporary a-proteobacterium can be required for the assembly of FeS clusters are localized in expected to contain exactly the same set of orthologous mitochondria (Lill and Kispal 2000) and related organelles genes as the ancestral mitochondrial endosymbiont did such as hydrogenosomes (Tachezy, Sanchez, and Mu¨ller (Rotte et al. 2001). Third, the current sampling of eu- 2001) and mitosomes (Katinka et al. 2001; Williams et al.
bacterial lineages is currently quite sparse; in time, eubac- 2002). The emerging monophyly of FeS cluster assembly terial homologs that are more closely related to eukaryotic in eukaryotes suggests that it was acquired en bloc from SQR might be found. Thus, the lack of an a-proteobacterial the ancestor of mitochondria (Huynen et al. 2001), as branch for this poorly conserved and laterally transferred we suggest here for an SQR-containing mitochondrial gene (SQR) does not constitute clear evidence against its respiratory chain.
Third, newer evidence suggests that during the On the other hand, several findings argue in favor period of Earth's history from 2 billion years ago to 1 of the view that the eukaryotic SQR gene was acquired billion years ago (2 to 1 Ga) SQR must have been very from the ancestor of mitochondria. First, eukaryotic SQR important, if not essential, for most, if not all, eukayrotes, functions in the mitochondrial membrane the same way at least the ones that inhabited the oceans. This is that a-proteobacterial SQR functions in the eubacterial because the sulfur isotope record indicates that biological membrane, donating electrons from sulfide to quinones.
sulfate reduction, which produces sulfide, was highly Hence, a eubacterium with a diversified (facultatively) active and globally widespread during that time (Canfield anaerobic electron transport chain would be the most likely 1998; Shen, Buick, and Canfield 2001; Anbar and Knoll SQR gene donor, for example one that could use fumarate 2002). The consequence is that Earth's oceans subsurface as an electron acceptor. Eubacteria such as the a- water would have been both anoxic (without oxygen) and proteobacterium Rhodospirillum rubrum, which possesses sulfidic (laden with sulfide) during that time. Anbar and SQR, commonly use rhodoquinone (Okayama et al. 1968) Knoll (2002) discussed this anoxic, sulfidic marine alternatively to ubiquinone in their anaerobic electron environment in the context of low resulting copper and transport chain, just like anaerobic mitochondria do today molybdenum concentrations, which they argued to have Theissen et al.
possibly impaired eukaryotic diversity, because these are important trace elements for eukaryotes. However, from We thank the Deutsche Forschungsgemeinschaft for financial support, Carmen Rotte for critical reading, and immediate problem posed by such environments for Gu¨nter Hauska for many helpful comments on the early eukaryotes would have been (1) ATP production without oxygen and (2) dealing with high concentrationsof sulfide. Put another way, only osmotrophic eukaryotessuch as fungi would have been limited by trace elementavailability—phagocytosing eukaryotes would have been able to obtain their trace elements from ingested prey, but Adachi, J., and M. Hasegawa. 1996. Computer science all subsurface eukaryotes during the period from 1 to 2 monographs, No. 28. MOLPHY version 2.3: programs for Ga would have been confronted with high sulfide molecular phylogenetics based on maximum likelihood.
concentrations. SQR is the mechanism that contemporary Institute of Statistical Mathematics, Tokyo.
eukaryotes use to deal with high sulfide concentrations Anbar, A. D., and A. H. Knoll. 2002. Proterozoic ocean today, both in terms of detoxification and in terms of chemistry and evolution: a bioinorganic bridge. Science utilizing sulfide for mitochondrial ATP synthesis (Grie- shaber and Vo¨lkel 1998). It is therefore reasonable to Arieli, B., Y. Shahak, D. Taglicht, G. Hauska, and E. Padan.
assume that ancient eukaryotes dealt with sulfide the 1994. Purification and characerization of sulfide-quinone same way as contemporary eukaryotes do, namely with reductase, a novel enzyme driving anoxygenic photosynthesisin Oscillatoria limnetica. J. Biol. Chem. 269:5705–5711.
mitochondrial SQR. Hence, the sulfidic and anoxic phase Bagarinao, T., and R. D. Vetter. 1990. Oxidative detoxification of of Earth's history revealed by the sulfur isotope record sulfide by mitochondria of the California killifish Fundulus does not lead to the prediction of limited eukaryotic parvipinnis and the speckled sanddab Citharichthys stig- diversity during the period from 1 to 2 Ga as suggested maeus. J. Comp. Physiol. 160B:519–527.
by Anbar and Knoll (2002), rather it leads to the Bastolla, U., M. Porto, H. E. Roman, and M. Vendruscolo. 2002.
prediction that eukaryotes diversified during anaerobic Lack of self-averaging in neutral evolution of proteins. Phys.
times and therefore that they should have preserved Rev. Lett. 89: art. no. 208101.
abundant traces of that anaerobic past—which they have, Bateman, A., E. Birney, L. Cerruti, R. Durbin, L. Etwiller, S. R.
particularly in their mitochondria and hydrogenosomes Eddy, S. Griffiths-Jones, K. L. Howe, M. Marshall, and E. L.
(Martin and Mu¨ller 1998; Tielens et al. 2002; Embley et L. Sonnhammer. 2002. The Pfam protein families database.
al. 2003) and also in the form of mitochondrial SQR.
Nucleic Acids Res. 30:276–280.
Unicellular eukaryotes are at least 1.5 Ga old (Javaux, Baymann, F., E. Lebrun, M. Brugna, B. Schoepp, M.-T. Guidici- Oritconi, and W. Nitschke. 2003. The redox protein Knoll, and Walter 2001) and multicellular red algae are construction kit: pre-last universal common ancestor evolution at least 1.2 Ga old (Butterfield 2000), meaning that dif- of energy conserving enzymes. Philos. Trans. R. Soc. Lond. B ferentiation of eukaryotic lineages below the plant lineage Biol. Sci. 358:267–274.
occurred in an anoxic and sulfidic world. Thus, eukaryotes Bronstein, M., M. Schu¨tz, G. Hauska, E. Padan, and Y. Shahak.
that today inhabit anoxic and sulfidic marine environments 2000. Cyanobacterial sulfide-quinone reductase: cloning and did not necessarily have to become especially adapted to heterologous expression. J. Bacteriol. 182:3336–3344.
such conditions, nor did they need to acquire SQR genes by Bryant, D., and V. Moulton. 2002. NeighborNet: an agglomer- lateral transfer to do so. Rather, it seems that they ‘‘grew ative method for the construction of planar phylogenetic up'' in an anoxic and sulfidic world and that mitochondrial networks. Proceedings of the 2nd Workshop in Algorithms SQR is simply a relic retained from that phase of eukaryotic history, whereby it still fulfills those same essential Burger, G., B. F. Lang, M. Reith, and M. W. Gray. 1996. Genes functions in modern eukaryotes from sulfidic habitats.
encoding the same three subunits of respiratory complex II are Newer data indicate the fungal-animal divergence to be present in the mitochondrial DNA of two phylogenetically among the deepest branches in the eukaryotic tree distant eukaryotes. Proc. Natl. Acad. Sci. USA 93:2328–2332.
(Stechmann and Cavalier-Smith 2002), such that the animal Butterfield, N. J. 2000. Bangiomorpha pubescens n. gen., n. sp.: and fungal lineages sampled here cover much of the depth implications for the evolution of sex, multicellularity, and the but not the breadth of eukaryotic diversity. In eukaryotes Mesoproterozoic/Neoproterozoic radiation of eukaryotes.
from aerobic and/or nonsulfidic habitats, such as S. pombe, the SQR gene and activity have nonetheless been retained Canback, B., S. G. Andersson, and C. G. Kurland. 2002. The (Vande Weghe and Ow 1999), perhaps for detoxification global phylogeny of glycolytic enzymes. Proc. Natl. Acad.
functions, and the SQR gene has apparently been lost in Sci. USA 99:6097–6102.
many lineages, among them Arabidopsis and Saccharo- Canfield, D. E. 1998. A new model for Proterozoic ocean myces. No SQR homologs have yet been sequenced from chemistry. Nature 396:450–453.
Chabriere, E., M. H. Charon, A. Volbeda, L. Pieulle, E. C.
those eukaryotes in which mitochondrial SQR has been Hatchikian, and J. C. Fontecilla-Camps. 1999. Crystal most extensively characterized at the biochemical level: structures of the key anaerobic enzyme pyruvate:ferredoxin marine invertebrates (Grieshaber and Vo¨lkel 1998; Doeller, oxidoreductase, free and in complex with pyruvate. Nature Grieshaber, and Kraus 2001) and chicken (Yong and Struct. Biol. 6:182–190.
Searcy 2001). However, work on the marine invertebrates Chen, Z. W., M. Koh, G. Van Driessche, J. J. Van Beeumen, R.
is ongoing. Clearly, our prediction is that SQR from these G. Bartsch, T. E. Meyer, M. A. Cusanovich, and F. S.
eukaryotes will share the same origin as S. pombe SQR.
Mathews. 1994. The structure of flavocytochrome c sulfide Single Eubacterial Origin of Eukaryotic Sulfide dehydrogenase from a purple phototrophic bacterium. Science John, P., and F. R. Whatley. 1975. Paracoccus denitrificans and the evolutionary origin of the mitochondrion. Nature Couvreur, B., R. Wattiez, A. Bollen, P. Falmagne, D. Le Ray, and J. C. Dujardin. 2002. Eubacterial hslV and hslU subunits Johns, A. R., A. C. Taylor, R. J. A. Atkinson, and M. K.
homologs in primordial eukaryotes. Mol. Biol. Evol.
Grieshaber. 1997. Sulphide metabolism in thalassinidean crustacea. J. Mar. Biol. Ass. UK 77:127–144.
Doeller, J. E., B. K. Gaschen, V. Parrino, and D. W. Kraus. 1999.
Huynen M. A., B. Snel, P. Bork and T. J. Gibson. 2001. The Chemolithoheterotrophy in a metazoan tissue: sulfide supports phylogenetic distribution of frataxin indicates a role in iron- cellular work in ciliated mussel gills. J. Exp. Biol. 202:1953– sulfur cluster protein assembly. Hum. Mol. Genet. 10:2463– Doeller, J. E., M. K. Grieshaber, and D. W. Kraus. 2001.
Katinka, M.D., S. Duprat, E. Cornillot et al. (17 co-authors).
Chemolithoheterotrophy in a metazoan tissue: thiosulfate 2001. The genome of the intracellular parasite, Encephalitho- production matches ATP demand in ciliated mussel gills.
zoon cuniculi. Nature 414:450–453.
J. Exp. Biol. 204:3755–3764.
Kurland, C. G., and S. G. Andersson. 2000. Origin and evolution Doolittle, W. F. 1998. You are what you eat: a gene transfer of the mitochondrial proteome. Microbiol. Mol. Biol. Rev.
ratchet could account for bacterial genes in eukaryotic nuclear genomes. Trend. Genet. 14:307–311.
Lecompte, O., R. Ripp, J. C. Thierry, D. Moras, and O. Poch.
Embley, T. M., M. van der Giezen, D. S. Horner, P. L. Dyal, and 2002. Comparative analysis of ribosomal proteins in complete P. Foster. 2003. Hydrogenosomes and mitochondria: pheno- genomes: an example of reductive evolution at the domain typic variants of the same fundamental organelle. Philos.
scale. Nucleic Acids Res. 30:5382–5390.
Trans. R. Soc. Lond. B Biol. Sci. 358:191–203.
Lill, R., and G. Kispal. 2000. Maturation of cellular Fe-S Felsenstein, J. 1998. PHYLIP (phylogeny inference package).
proteins: an essential function of mitochondria. Trend.
Distributed by the author, Department of Genetics, University Biochem. Sci. 25:352–356.
of Washington, Seattle.
Martin, W., M. Hoffmeister, C. Rotte, and K. Henze. 2001. An Fenchel, T. M., and R. J. Riedl. 1970. The sulfide system: a new overview of endosymbiotic models for the origins of biotic community underneath the oxidized layer of marine eukaryotes, their ATP-producing organelles (mitochondria sand bottoms. Mar. Biol. 7:255–268.
and hydrogenosomes), and their heterotrophic lifestyle. Biol.
Friedrich T, and D. Schiede. 2000. The respiratory complex I of bacteria, archaea, and eukarya and its module in common with Martin, W., and M. Mu¨ller. 1998. The hydrogen hypothesis for membrane-bound multisubunit hydrogenases. FEBS Lett.
the first eukaryote. Nature 392:37–41.
Martin, W., T. Rujan, E. Richly, A. Hansen, S. Cornelsen, T.
Gogarten, P. J., W. F. Doolittle, and J. G. Lawrence. 2002.
Lins, D. Leister, B. Stoebe, M. Hasegawa, and D. Penny.
Prokaryotic evolution in light of lateral gene transfer. Mol.
2002. Evolutionary analysis of Arabidopsis, cyanobacterial, Biol. Evol. 19:2226–2238.
and chloroplast genomes reveals plastid phylogeny and Griesbeck, C., G. Hauska, and M. Schu¨tz. 2000. Biological thousands of cyanobacterial genes in the nucleus. Proc. Natl.
sulfide oxidation: sulfide-quinone reductase (SQR), the Acad. Sci. USA 99:12246–12251.
primary reaction. Pp. 179–203 in S. G. Pandalai, ed. Recent National Research Council. 1979. Hydrogen sulfide. University research developments in microbiology, Vol 4. Research Park Press, Baltimore.
Signpost, Trivadrum, India.
Nei, M. 1996. Phylogenetic analysis in molecular evolutionary Griesbeck, C., M. Schu¨tz, T. Scho¨dl, S. Bathe, L. Nausch, N.
genetics. Ann. Rev. Genet. 30:371–403.
Mederer, M. Vielreicher, and G. Hauska. 2002. Mechanism of Nei, M., and S. Kumar. 2000. Molecular evolution and sulfide-quinone reductase investigated using site-directedmutagenesis and sulfur analysis. Biochemistry 41:11552– phylogenetics. Oxford University Press, New York.
Nixon, J. E. J., A. Wang, J. Field, H. G. Morrison, A. G.
Grieshaber, M. K., and S. Vo¨lkel. 1998. Animal adaptations for McArthur, M. L. Sogin, B. J. Loftus, and J. Samuelson. 2002.
tolerance and exploitation of poisonous sulfide. Ann. Rev.
Evidence for lateral transfer of genes encoding ferredoxins, nitroreductases, NADH oxidase, and alcohol dehydrogenase 3 Hannaert, V., H. Brinkmann, U. Nowitzki, J. A. Lee, M. A.
from anaerobic prokaryotes to Giardia lamblia and Entamo- Albert, C. W. Sensen, T. Gaasterland, M. Muller, P. Michels, eba histolytica. Euk. Cell 1:181–190.
and W. Martin. 2000. Enolase from Trypanosoma brucei, Nu¨bel, T., C. Klughammer, R. Huber, G. Hauska, and M. Schu¨tz.
from the amitochondriate protist Mastigamoeba balamuthi, 2000. Sulfide:quinone oxidoreductase in membranes of the and from the chloroplast and cytosol of Euglena gracilis: hyperthermophilic bacterium Aquifex aeolicus (VF5) Arch.
pieces in the evolutionary puzzle of the eukaryotic glycolytic pathway. Mol. Biol. Evol. 17:989–1000.
O'Brien, J., and R. D. Vetter. 1990. Production of thiosulfate Horner, D. S., P. D. Foster, and T. M. Embley. 2000. Iron during sulphide oxidation by mitochondria of the symbiont- hydrogenase and the evolution of anaerobic eukaryotes. Mol.
containing bivalve Solemya reidi. J. Exp. Biol. 149:133–148.
Biol. Evol. 17:1695–1709.
Oeschger, R., and B. Vismann. 1994. Sulphide tolerance in Horner, D. S., B. Heil, T. Happe, and T. M. Embley. 2002. Iron Heteromastus filiformis (Polychaeta): mitochondrial adapta- hydrogenases—ancient enzymes in modern eukaryotes.
tions. Ophelia 40:147–158.
Trends Biochem. Sci. 27:148–153.
Okayama, S., N. Yamamoto, K. Nishikawa, and T. Horio. 1968.
Horner, D. S., R. P. Hirt, and T. M. Embley. 1999. A single Roles of ubiquinone-10 and rhodoquinone in photosynthetic eubacterial origin of eukaryotic pyruvate: ferredoxin oxidore- formation of adenosine triphosphate by chromatophores from ductase genes: implications for the evolution of anaerobic Rhodospirillum rubrum. J. Biol. Chem. 243:2995–2999.
eukaryotes. Mol. Biol. Evol. 16:1280–1291.
Parrino, V., D. W. Kraus, and J. E. Doeller. 2000. ATP Javaux, E. J., A. H. Knoll, and M. R. Walter. 2001.
Production from the oxidation of sulfide in gill mitochondria Morphological and ecological complexity in early eukaryotic of the ribbed mussel Geukensia demissa. J. Exp. Biol.
ecosystems. Nature 412:66–69.
Theissen et al.
Peters, J. W. 1999 Structure and mechanism of iron-only Tachezy, J., L. B. Sanchez, and M. Mu¨ller. 2001. Mitochondrial hydrogenases. Curr. Opin. Struct. Biol. 9:670–676.
type iron-sulfur cluster assembly in the amitochondriate Reeve, J. 2003. Archaeal chromatin and transcription. Mol.
eukaryotes Trichomonas vaginalis and Giardia intestinalis, as indicated by the phylogeny of IscS. Mol. Biol. Evol.
Reinartz, M., T. Tscha¨pe, T. Bru¨ser, H. G. Tru¨per, and C. Dahl.
1998. Sulfide oxidation in the phototrophic bacterium Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994.
Chromatium vinosum. Arch. Microbiol. 170:59–68.
ClustalW: improving the sensitivity of progressive multiple Rost, B. 1997. Protein structures sustain evolutionary drift. Fold.
sequence alignment through sequence weighting, position- Des. 2:S19–S24.
specific gap penalties and weight matrix choice. Nucleic Rotte, C., F. Stejskal, G. Zhu, J. S. Keithly, and W. Martin. 2001.
Acids Res. 22:4673–4680.
Pyruvate:NADPþ oxidoreductase from the mitochondrion of Tielens, A. G. M., C. Rotte, J. van Hellemond, and W. Martin.
Euglena gracilis and from the apicomplexan Cryptosporidium 2002. Mitochondria as we don't know them. Trends Biochem.
parvum: a fusion of pyruvate:ferredoxin oxidoreductase andNADPH-cytochrome P450 reductase. Mol. Biol. Evol.
Sci. 27:564–572.
Tye, B. K. 2000. Insights into DNA replication from the third Saitou, N., and M. Nei. 1987. The neighbor-joining method: domain of life. Proc. Natl. Acad. Sci. USA 97:2399–2401.
a new method for reconstructing phylogenetic trees. Mol.
Vande Weghe, J. G., and D. W. Ow. 1999. A fission yeast gene Biol. Evol. 4:406–425.
for mitochondrial sulfide oxidation. J. Biol. Chem. 274: Schu¨tz, M., M. Brugna, E. Lebrun et al. (12 co-authors). 2000.
Early evolution of cytochrome bc complexes. J. Mol. Biol.
Vo¨lkel, S., and M. K. Grieshaber. 1992. Mechanisms of sulphide tolerance in the peanut worm, Sipunculus nudus (Sipunculi- Schu¨tz, M., C. Klughammer, C. Griesbeck, A. Quentmeier, C. G.
dae) and in the lugworm, Arenicola marina (Polychaeta).
Friedrich, and G. Hauska. 1998. Sulfide-quinone reductase J. Comp. Physiol. B Biochem. Syst. Environ. Physiol. 162: activity in membranes of the chemotrophic bacterium Para- coccus denitrificans GB17. Arch. Microbiol. 170:353–360.
———. 1996a. Mitochondrial sulfide oxidation in Arenicola Schu¨tz, M., Y. Shahak, E. Padan, and G. Hauska. 1997. Sulfide- marina: evidence for alternative electron pathways. Eur. J.
quinone reductase from Rhodobacter capsulatus. J. Biol.
———. 1996b. Sulphide oxidation and oxidative phosphoryla- Shahak, Y., B. Arieli, E. Padan, and G. Hauska. 1992. Sulfide tion in the mitochondria of the lugworm Arenicola marina. J.
quinone reductase (SQR) activity in Chlorobium. FEBS Lett.
Exp. Biol. 200:83–92.
Vo¨lkel, S., K. Hauschild, and M. K. Grieshaber. 1995. Sulfide Shahak, Y., C. Klughammer, U. Schreiber, E. Padan, I.
stress and tolerance in the lugworm Arenicola marina during Herrmann, and G. Hauska. 1994. Sulfide-quinone and low tide. Mar. Ecol. Prog. Ser. 122:205–215.
sulfide-cytochrome reduction in Rhodobacter capsulatus.
Williams, B.A., R. P. Hirt, J. M. Lucocq, and T. M. Embley.
Photosynthesis Res. 39:175–181.
2002 A mitochondrial remnant in the microsporidian Shen, Y., R. Buick, and D. E. Canfield. 2001. Isotopic evidence for microbial sulphate reduction in the early Archaean era.
Trachipleistophora hominis. Nature 418:865–869.
Nature 410:77–81.
Yong, R., and D. G. Searcy. 2001. Sulfide oxidation coupled to Stechmann, A. and T. Cavalier-Smith. 2002. Rooting the ATP synthesis in chicken liver mitochondria. Comp. Bio- eukaryote tree by using a derived gene fusion. Science 297: chem. Physiol. B 129:129–137.
Strimmer, K., and A. von Haeseler. 1996. Quartet puzzling: Geoffrey McFadden, Associate Editor a quartet maximum-likelihood method for reconstructing treetopologies. Mol. Biol. Evol. 13:964–969.
Accepted May 8, 2003

Source: http://sulfide-life.info/mtobler/images/stories/readings/theissen%202003%20mol%20biol%20evol.pdf