| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Max-Planck-Institute for Marine Microbiology, Celsiusstrasse 1, D-28359 Bremen, Germany
Correspondence
Jan Kuever
kuever{at}mpa-bremen.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
Present address: Bremen Institute for Materials Testing, Paul-Feller-Strasse 1, D-28199 Bremen, Germany.
The GenBank/EMBL/DDBJ accession numbers for the nucleotide sequence data reported in this study are EF442876–EF442976 (apr) and EF442977–EF442994 (16S rRNA).
Supplementary data are available with the online version of this paper.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
). Despite its suggested antiquity, the capability for dissimilatory sulfate reduction is patchily distributed, and occurs solely within members of six bacterial and two distinct archaeal lineages (Itoh et al., 1999; Mori et al., 2003; Moussard et al., 2004; Rabus et al., 1999). In all recognized sulfate-reducing prokaryotes (SRP), the dissimilatory process is mediated by three key enzymes. After the activation of the chemically inert sulfate to adenosine-5'-phosphosulfate (APS) by ATP sulfurylase (Sat), the second enzyme, APS reductase (Apr), converts APS to AMP and sulfite, which is finally reduced to sulfide by the activity of the sulfite reductase (Dsr) (Rabus et al., 1999). Homologous proteins are also present in the anoxygenic photolithotrophic and chemolithotrophic sulfur-oxidizing bacteria (SOB) (Dahl & Trüper, 1994; Hipp et al., 1997; Schedel & Trüper, 1979, 1980; Sperling et al., 1998; Trüper & Fischer, 1982). Indeed, earlier studies have confirmed that the dissimilatory APS reductases are highly conserved among SRP and SOB and form heterodimers with one alpha subunit (75–80 kDa, one FAD) and one beta subunit (18–23 kDa, two [4Fe–4S] centres) which are encoded by the aprBA gene loci (Fritz et al., 2000; Hipp et al., 1997; Lampreia et al., 1994; Molitor et al., 1998; Speich et al., 1994). In contrast, the proteins that mediate the electron transport between the cytoplasmic AprBA and DsrAB and the quinol/quinone pool in the membrane are still poorly characterized. However, experimental evidence is accumulating that the membrane-bound redox complexes HmeABCDE [heterodisulfide reductase (Hdr)-like menaquinol-oxidizing enzyme] and homologous DsrMKJOP, as well as QmoABC (quinone-interacting membrane-bound oxidoreductase), are involved (Dahl et al., 2005; Mander et al., 2002; Pires et al., 2003, 2006; Sander et al., 2006). The Qmo redox complex from Desulfovibrio desulfuricans ATCC 27774 has been described as consisting of two soluble, HdrA-homologous, FAD-containing proteins, QmoA and QmoB, and membrane integral, HdrC/E-homologous protein, QmoC, containing two haem b groups and a putative quinone-binding site. Based on experimental results with Desulfovibrio desulfuricans ATCC 27774, the QmoABC complex has been postulated to be the missing link between the membrane menaquinol/menaquinone pool and the cytoplasmic reduction of APS by acting as a menaquinol/APS reductase oxidoreductase (Pires et al., 2003).
Although lateral gene transfer (LGT) between distantly related phylogenetic lineages and domains is well-documented, especially for genes that encode proteins of metabolic pathways (Boucher et al., 2003), the general impact of LGT as a major driving force in the genome evolution of prokaryotes is still debated (Daubin & Ochman, 2004; Gogarten & Townsend, 2005; Jain et al., 2003; Kurland et al., 2003; Lerat et al., 2005). The evolution of the sulfate-respiration process in SRP has primarily been investigated by comparative phylogenetic studies of the dissimilatory sulfite reductase DsrAB, and this has confirmed the occurrence of multiple events of LGT of dsrAB among members of this physiological group (Klein et al., 2001; Mussmann et al., 2005; Zverlov et al., 2005). A widespread dispersal via ‘metabolic islands’ has recently been discussed as responsible for the polyphyletic distribution of this metabolic trait (Klein et al., 2001; Mussmann et al., 2005). In contrast to the comprehensive analysis of DsrAB phylogeny, knowledge concerning the evolution of the dissimilatory APS reductase is currently restricted to a single phylogenetic study based on the limited sequence information of a minor part of the aprA gene-coding region (
0·9 kb) from 60 taxonomically different SRP species (Friedrich, 2002). The AprA tree topology of that study differed partially from the 16S rRNA gene-based tree, and led the author to suggest frequent events of inter- and intradomain LGT of the apr genes involving members of the Gram-positive sulfate-reducing bacteria (SRB), Syntrophobacteraceae, Syntrophaceae and Archaeoglobus, as well as the thermophilic Thermodesulfovibrio islandicus and representatives of the genus Thermodesulfobacterium (Friedrich, 2002).
The aim of this study was to increase the available genetic information for the apr gene locus by developing new PCR primer pairs for amplification and sequencing of nearly the entire coding region of the dissimilatory APS reductase (aprBA) genes from reference strains of all currently known SRP lineages. The phylogeny of both subunits AprB and AprA of the dissimilatory APS reductase from 103 different SRP (including some selected sequences of reverse-operating APS reductases of SOB) was compared with their 16S rRNA gene-based phylogeny to reveal novel events of LGT among the examined sulfate-reducing species. In addition, the results of the Apr phylogenetic analyses are discussed within the context of (1) the DsrAB phylogeny, (2) the collected genomic data concerning the presence and genomic arrangement of genes coding for putative functionally associated proteins (Qmo complex, AprM) at the apr gene locus of SRP and SOB, and (3) the geochemical data, in order to elucidate the origin and evolution of dissimilatory sulfate reduction/sulfite oxidation in prokaryotes.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
) were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ) either as lyophilized or as actively growing cultures. If necessary, strains were cultivated as recommended by the DSMZ. Actively growing cultures of the following reference strains were provided by O. Kniemeyer (SRB strain EbS7), A. Galushko (strain NaphS2), G. Harms (strain mXyS1), H. T. Dinh (‘Desulfovibrio ferrophilus’ strain IS5, ‘Desulfobacterium corrodens’ strain IS4 and SRB strain QLNR1), K. Nauhaus (Desulfofrigus sp. strain HRS-LA3X and Desulfovibrio sp. strain HRS-LA4), M. Könnecke (SRB strain JHA1) and J. Detmers (Desulfovibio sp. strain JD-160, Desulfotomaculum sp. strains JD-175 and JD-176) at the Max-Planck-Institute for Marine Microbiology, Bremen, and by H. P. Goorissen (Desulfotomaculum solfataricum) and T. A. Hansen (Desulfobacterium niacini strain PM4) at the University of Groningen. Desulfovibrio sp. strains LB1/LA1(H2)/LA2/49MC/sponge tissue 85CD as well as Desulfobulbus sp. strain LB2 were isolated from sediment and seawater samples of the Caribbean Sea (RV Sonne cruise SO-154, Caribflux project).
|
). Genomic DNAs were extracted from the investigated strains using the DNAeasy kit (Qiagen) according to the manufacturer's instructions. The DNA concentration was measured using a spectrophotometer. DNA quality was checked by PCR amplification with the 16S rRNA gene-targeting primer sets GM3F/GM4R and Arch21F/Arch958R, as described elsewhere (DeLong, 1992; Muyzer et al., 1995).
PCR primers.
Two sets of degenerate primers that anneal to conserved aprBA gene regions of SRP (Table 2) were newly designed based on comparison of Desulfovibrio vulgaris, Desulfovibrio desulfuricans, Archaeoglobus fulgidus (see Table 1
for GenBank accession numbers), Allochromatium vinosum (U84759) and Chlorobaculum tepidum (NC_002932) full-length apr sequences. Since the aprA alignment revealed a limited number of suitable primer target sites and a generally low degree of conserved nucleotide positions in the 3' terminal gene region, two reverse PCR primers of differing degeneracy (AprA-10-RV, AprA-11-RV) were developed that were complementary to the target site of either the Desulfovibrio species or the Archaeoglobus/Gram-positive SRB. The PCR primer pair AprB-1-FW/AprA-5-RV yielded a 1·2–1·35 kb aprBA amplicon from the investigated SRP, while an 1·4 kb gene fragment of the 3' terminal aprA gene region was amplified using the forward primer AprA-1-FW in combination with the reverse primers AprA-10-RV or AprA-11-RV. The aprBA and aprA PCR products overlap in sequence by 400 bp (corresponding to aprA nucleotide positions 1236–1631 from Desulfovibrio vulgaris; Table 2).
|
Cloning of PCR products.
The 1·3 kb aprBA gene fragment of Thermacetogenium phaeum was ligated into pCR 2·1-TOPO vectors (TOPO TA cloning systems, Invitrogen) and transformed into chemically competent Escherichia coli TOP10 cells following the manufacturer's instructions. Clone plasmids with inserts of approximately 1·3 kb were selected and screened by PCR amplification with subsequent RFLP analysis of the amplicons. RFLP patterns were visualized on 2 % (w/v) agarose gel runs in 1x TBE buffer and stained with ethidium bromide. Plasmids of representative clones from unique RFLP patterns were recovered with the QIAprep Spin kit (Qiagen).
Nucleotide sequencing.
The PCR products (aprBA, aprA and 16S rDNA amplicons) were directly sequenced in both directions using the respective PCR amplification primers and the ABI BigDye Terminator Cycle Sequencing kit (Applied Biosystems). Sequencing reactions were run on an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems).
Sequence data analysis and phylogenetic tree inference.
The DNA sequence data of the aprBA and aprA amplicons were assembled and manually corrected using the BioEdit (version 7.0.5) sequence alignment editor (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). BLAST searches for homologous sequences were performed at the NCBI web site (http://www.ncbi.nlm.nih.gov/BLAST/). Searches on the preliminary data of ongoing sequencing projects of SRP and SOB genomes were performed at The Institute for Genomic Research (TIGR) web site (http://www.tigr.org) and at the US Department of Energy (DOE) Joint Genome Institute web site (http://img.jgi.doe.gov/cgi-bin/pub/main.cgi). The deduced partial amino acid sequences of this study and the publicly available full-length AprBA sequences (summarized in Table 3) were automatically aligned using the web server Tcoffee@igs (http://igs-server.cnrs-mrs.fr/Tcoffee/). The frameshifted aprB/A sequences of the Pyrobaculum aerophilum genome were manually corrected before inclusion into the datasets. The nucleotide sequences were aligned according to the corrected amino acid alignment. The AprB and AprA datasets were analysed separately with the phylogeny inference methods included in the ARB software package (http://www.arb-home.de). Alignment regions of insertions and deletions (indels) were not considered in the phylogenetic analysis. Unrooted phylogenetic trees were calculated based on 103 (AprB) and 555/413 (AprA) amino acid positions using the ARB implemented program package (distance matrix, Fitch analysis; maximum-parsimony, ProtPars; maximum-likelihood, ProML and PUZZLE) and the PhyML program (maximum-likelihood method; http://atgc.lirmm.fr/phyml). Maximum-likelihood trees were constructed using the WAG or JTT amino acid substitution model matrices. The robustness of inferred trees was tested by bootstrap analysis with 100 resamplings. Tree reconstruction with PUZZLE analysis was performed by 20 000 Quartet Puzzling steps employing the WAG amino acid replacement model with either a unique rate of evolution or a mixed four-category discrete gamma heterogeneity model (approximation of parameters done using a neighbour-joining tree). Predictions of potential promoters, termination sites and operons in genome data were performed using the web versions FGENESB, BPROM and BTERM of the Softberry program package available at (http://www.softberry.com/berry.phtml). Protein secondary structure analysis and transmembrane helix prediction were done using the tools available at http://us.expasy.org/tools/#secondary.
|
for GenBank accession nos) were loaded into the ARB database and aligned using the ARB_ALIGN tool. The maximum-parsimony method with a 50 % positional conservation filter and the Kimura two-parameter model of the ARB program package was used for phylogenetic tree calculations based on 1400 nucleotide positions. Partial 16S rDNA gene sequences were added to the calculated initial trees by applying the parsimony insertion tool of ARB without allowing changes in the overall tree topology. The robustness of the inferred tree was tested by bootstrap analysis with 100 resamplings using the PhyML program. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
). No aprBA amplicon was obtained from any of the six representatives of Desulfotomaculum subcluster Ib (for Desulfotomaculum subcluster assignment, see Fig. 1), Thermodesulforhabdus norvegica or Caldivirga maquilingensis, while direct nucleotide sequencing of the Desulfosporosinus orientis, Desulfobacterium niacini and ‘Desulfobacterium oleovorans’ strain HxD3 amplicon was unsuccessful with the reverse primer AprA-5-RV. The second primer set, AprA-1-FW combined with AprA-10-RV or AprA-11-RV, amplified the 3'-terminal aprA gene fragment from 93 out of 97 tested SRP species (Table 1). No aprA amplification product was obtained from the investigated Desulfosporosinus strains, Desulfotomaculum acetoxidans or Caldivirga maquilingensis. Direct nucleotide sequencing of the Desulfotomaculum halophilum, Desulforhabdus amnigena, Desulfotalea psychrophila, Desulfobacterium catecholicum, SRB strain Kette, Desulfocella sp. (DSM 2056) and Thermacetogenium phaeum aprA amplicon resulted in only partially evaluable sequences due to sequence ambiguities. The failure to amplify and sequence apr gene fragments from certain SRB species indicated that they do not contain target sites fully complementary to the applied primers. In fact, compositional analysis of the aprBA gene locus of Desulfotalea psychrophila (Rabus et al., 2004) revealed a deviating codon usage that resulted in four mismatches at the primer binding site of Apr-10-RV. Nevertheless, the newly developed PCR primer sets enabled the enlargement of the present 0·9 kb aprA database (Friedrich, 2002) to a comprehensive aprBA database which comprises 2·2–2·4 kb (equivalent to 91–93 %) of the coding region of the dissimilatory APS reductase from 75 SRP species over a wide taxonomical range.
|
) revealed a strictly conserved gene locus arrangement of aprB preceding the aprA gene and the absence of terminator-like signals downstream of the aprB stop codon, demonstrating that both genes form a transcriptional unit. The aprBA genes of SRP (and Chlorobiaceae) are generally separated by an intercistronic sequence region (16–148 nt) of a varying, family-specific length interval. As expected for highly expressed genes, the aprA genes from most investigated SRP possess Shine–Dalgarno (SD) sequences with entire complementarity (GGAGG) to the anti-SD core sequence and an optimal space distance between 6 and 12 nt (Kozak, 1999). AUG is used to start translation in all aprA sequences, with the exception of the members of Desulfotomaculum subcluster Ic, which utilize the less potent initiator GUG. The alternative start codon UUG (accession no. NC_006138, nucleotide position 1 246 432 in the genome sequence) annotated for the aprA gene of the Desulfotalea psychrophila genome (Rabus et al., 2004) has to be corrected in our opinion to AUG (nucleotide position 1 246 438), which fulfils the required criteria (e.g. space distance of 9 nt instead of the annotated 3 nt to the SD sequence). In contrast, no intergenetic sequence exists in the Allochromatium vinosum-related sequence group as a consequence of the general overlapping of the aprB stop codon with the start codon of aprA in the sequence UGAUG or UAAUG.
Phylogeny of dissimilatory APS reductases from SRP
The AprB and AprA trees presented in this study were based on 110 AprB (Fig. 2) and 93 AprA sequences (Fig. 3), with 103 and 555 compared amino acid positions, respectively. AprA subtrees were calculated (413 compared positions) to include in the phylogenetic analysis partial sequences obtained from e.g. Desulfotomaculum subcluster Ib (see Supplementary Fig. S1). The Allochromatium vinosum-related sequence group (including ‘Candidatus Pelagibacter ubique’, uncultured bacterium EBAC2C11 and Thiobacillus denitrificans; see Table 3) and Pyrobaculum aerophilum were used to root the trees. The comparison of the AprB- and AprA-based tree topologies of the investigated SRP revealed a consistent separation into four distinct phylogenetic sequence clusters consisting of (1) Archaeoglobus species, (2) the Gram-positive SRB associated with members of the deltaproteobacterial Syntrophobacterales, Desulfarcales and the Desulfobacterium anilini-related SRB group, (3) the thermophilic SRB in affiliation with the Chlorobaculum tepidum-related sequences group (comprising Chlorobium phaeobacteroides, Chlorobium chlorochromatium and Chlorobium clathratiforme; see Table 3), and (4) the representatives of Desulfovibrionales and Desulfobacterales (Figs 2 and 3). The intracluster-branching order of taxa was consistent between the subunit trees, irrespective of the tree inference method or dataset used. However, topological discrepancies were obtained with respect to the relative ordering of the major SRP groups in the AprB and AprA trees, e.g. the varying position of the Thermodesulfobacteriaceae and Archaeoglobus. This might be explained by the limited amount of phylogenetic information contained in the smaller beta subunit, which is most likely insufficient to resolve branching orders at deeper phylogenetic levels. Nevertheless, the inferred phylogenies indicated the co-evolution of the beta and alpha subunits of the dissimilatory APS reductase in the investigated SRP.
|
|
, Supplementary Fig. S1) are topologically congruent with the tree of the earlier phylogenetic analysis which was based on 252 compared positions from AprA sequences of 60 SRP species and support the major LGT events postulated earlier (Friedrich, 2002). However, the AprA datasets used in the two phylogenetic studies overlap in only 30 investigated reference strains. The newly investigated genera and species, e.g. representatives of Desulfosporosinus and Desulfotomaculum subclusters Ic, Id and If (see Table 1), complete the previous Apr database for phylogenetic analyses and provide a more comprehensive framework for molecular ecology studies (Figs 2 and 3, Supplementary Fig. S1). The enlarged species coverage of this study refined and emphasized by novel results the general 16S rRNA gene congruence of the Apr-based inter- and intrafamily SRP relationships, e.g. the formation of the Desulfovibrionaceae, Desulfonatronumaceae and Desulfomicrobiaceae subclusters within the Desulfovibrionales. Interestingly, the Apr trees showed the Desulfohalobiaceae to be subdivided into two groups consisting of (1) Desulfohalobium retbaense related to the Desulfomicrobiaceae, and (2) Desulfothermus naphthae/Desulfonatronovibrio hydrogenovorans/Desulfonauticus sp. forming a basal-positioned group. The proposed 16S rRNA gene-based classification of the Desulfotomaculum and Desulfosporosinus members into the subclusters Ia–If and II (Kuever et al., 1999) (Fig. 1) was supported by the Apr trees (Figs 2 and 3, Supplementary Fig. S1). In contrast with the earlier study (Friedrich, 2002), Desulfarculus baarsii, Desulfomonile tiedjei and the Desulfobacterium anilini-related SRB group branched separately within the Gram-positive AprB/A lineage. Furthermore, the family Syntrophaceae was not monophyletic, since Desulfomonile tiedjei formed a separate lineage within the Desulfotomaculum radiation, while the newly provided AprB/A sequences of Desulfobacca acetoxidans were by 16S rRNA gene-based phylogeny discordantly associated with the Gram-negative, thermophilic Thermodesulfovibrio yellowstonii. Another newly discovered deviation from the 16S rRNA gene-based phylogeny was the Apr-based, close relationship of the Gram-positive Thermacetogenium phaeum (Hattori et al., 2000) with Thermodesulfovibrio yellowstonii. The cloning results of the Thermacetogenium phaeum aprBA amplicon demonstrated the absence of further gene copies in this reference strain. Direct nucleotide sequencing of its PCR-amplified 16S rDNA resulted in unambiguous nucleotide sequences. Indeed, the uncontaminated pure-culture status of the type strain of Thermacetogenium phaeum was also confirmed by the DSMZ (S. Spring, personal communication).
Additional evidence for lateral transfer of aprBA genes among SRP
The presence of indels at identical positions within the Apr sequences was used as additional evidence for the occurrence of the inferred LGT events. In addition, the new, enlarged aprBA sequence dataset was checked for recent LGTs by identification of (1) atypical aprBA sequence characteristics, such as significant G+C-content deviations (Lawrence & Ochman, 1997), and variations in codon usage with respect to the recipient genome and the dsrAB gene, and (2) sequence similarity of the aprBA intercistronic region between distantly related species. In this way, the separate phylogenetic position of the Allochromatium vinosum-related sequences group was confirmed by the presence of four unique indels in each subunit which were absent from AprB and AprA sequences of all SRP and the affiliated Chlorobiaceae cluster (see Supplementary Fig. S2 for amino acid sequence alignment). The proposed LGT of aprBA genes from Gram-positive SRB donor strains to the Syntrophobacteraceae, Desulfomonile tiedjei (including the unclassified putative SRB), the Desulfobacterium anilini-related SRB group, Desulfarculus baarsii as well as Archaeoglobus species were confirmed by the presence of six shared indels in the respective Apr sequences. The separate, basal branching point of the archaeal genus in AprB/A-based trees was supported by two additional unique indels. The close relationship of Desulfobacca acetoxidans to the Thermodesulfovibrio–Chlorobaculum cluster was confirmed by three unique indels located in the AprA sequence and the shared short C-terminal sequence of the AprB protein. No atypical sequence characteristics were detected in members of the above recipient lineages that would indicate the recent occurrence of the proposed LGT events. However, the Apr-based close relationship of Thermacetogenium phaeum with Thermodesulfovibrio might be the first example of a recent LGT of apr genes among distantly related SRB reference strains. A recent lateral transfer of the aprBA gene to Thermacetogenium phaeum was supported by the presence of identical indel positions and lengths in the AprBA sequences of Thermacetogenium phaeum and Thermodesulfovibrio (see Supplementary Fig. S2), similar aprBA gene G+C content, codon usage, and the nearly identical length and nucleotide sequence of the intercistronic region.
Genomic arrangement of genes coding for the dissimilatory APS reductase and functionally associated proteins
The sat/aprBA/qmoABC gene organization of the available genomes reflected the phylogenetic divergence of the investigated SRP species into the four major AprBA lineages and their affiliation to the green sulfur bacteria (see Table 3 for gene locus numbers in genomes, and Fig. 4 for a graphical representation). The genomic arrangement in the thermophilic SRB, e.g. Thermodesulfovibrio yellowstonii and Thermodesulfobacterium commune, and the Chlorobiaceae was identical; the genes are most probably regulated in two separate operons, sat-aprBA and qmoABC. Notably, the sat gene is followed in all thermophilic SRP genomes by an additional ORF that encodes a protein of unknown function (Archaeoglobus fulgidus, AF1668; Pyrobaculum aerophilum, PAE2610). The aprBA and qmoABC gene arrangement in the genomes of the deltaproteobacterial representatives was similar to that of the thermophilic SRB and Chlorobiaceae genomes; however, the sat gene was separately located and regulated. Interestingly, the close relationship of the AprBA of Syntrophobacter fumaroxidans to the Gram-positive SRB lineage (e.g. Desulfotomaculum reducens MI-1) was reflected in their identical sat-aprBA and qmoAB operon structure; a sequence homologous to qmoC adjacent to this gene locus could not be identified in the preliminary genome data of these species. However, Syntrophobacter fumaroxidans harbours a second and functionally complete qmoABC operon that is not associated with the sat-aprBA-qmoAB gene cluster. A separately transcribed qmoC homologue is present near the qmoAB and sat-aprBA gene loci in the metagenomic sequences of the LGT-affected unclassified putative SRB strains (fosws39f7, fosws7f8) (Mussmann et al., 2005). Interestingly, these strains are also closely related in the QmoA- and QmoB-based phylogenetic trees (not shown). This might indicate a concerted lateral transfer of the entire sat-aprBA-qmoAB gene cluster from the Gram-positive donor lineage (representative strain Desulfotomaculum reducens) to the deltaproteobacterial Syntrophobacter fumaroxidans and both uncultivated SRB strains (fosws39f7, fosws7f8), with subsequent intragenomic rearrangements (Suyama & Bork, 2001) in the uncultivated strains. Nevertheless, the presence of a (second) LGT-affected and separately located qmoABC gene locus in the Syntrophobacter fumaroxidans and Archaeoglobus fulgidus genomes demonstrates that independent lateral transfer of the apr and qmo genes has occurred.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
; Molitor et al., 1998; Zverlov et al., 2005). The topology of the AprB/A-based trees of this study is congruent with the topology of an AprA-based tree of an earlier work (Friedrich, 2002). However, the two studies are based on different Apr datasets with respect to the SRP species coverage (102 versus 60 species) and the amount of phylogenetic information used for tree inference (2·2–2·4 kb versus 0·9 kb of the apr gene locus). The inferred phylogenetic positions of the Apr sequences from 72 newly investigated SRP species of this work were generally consistent with the major LGT events of apr genes proposed by Friedrich (2002) to have affected the sulfate-reducing members of the Syntrophobacterales, Thermodesulfobacterium and Thermodesulfovibrio species, as well as the archaeal genus Archaeoglobus. Nevertheless, the newly provided Apr sequences (Figs 2 and 3, Supplementary Fig. S1) refined and emphasized the 16S rRNA gene-congruent inter- and intrafamily clustering of the SRP taxa by the higher number of examined species. The results of this phylogenetic study prove the principal vertical transmission of the aprBA genes in the major SRP lineages. In contrast to the findings of the earlier analysis (Friedrich, 2002), the LGT-affected members of the Syntrophobacteraceae, the Desulfobacterium anilini-related SRB group, Desulfomonile tiedjei and Desulfarculus baarsii form four separately branching Apr lineages in affiliation with the monophyletic Gram-positive SRB group. Their separate phylogenetic placement points to four independent LGT events involving apr genes, most probably with different Gram-positive SRB species serving as donor strains. The different phylogenetic relationships of the xenologous aprBA genes from Desulfomonile and Desulfobacca indicate the incorrect systematic classification of both genera within the ‘Syntrophaceae’, along with members that are not capable of sulfate or sulfite reduction (Smithella and Syntrophus) (Kuever et al., 2005). Therefore, the AprB/A phylogenies support (1) the proposed taxonomical placement of Desulfarculus baarsii as well as Desulfobacterium anilini and its relatives into individual orders distinct from the Syntrophobacterales (Kuever et al., 2005), and (2) the need for reclassification and validation of Desulfomonile tiedjei and Desulfobacca acetoxidans at the family level within the Deltaproteobacteria.
Two novel cases of LGT are proposed with respect to the xenologous Apr sequences of Desulfobacca acetoxidans and Thermacetogenium phaeum (also confirmed by indel pattern; see Supplementary Fig. S2). The apr gene composition analysis supported an ancient occurrence of the LGT event in Desulfobacca acetoxidans, whereas Thermacetogenium phaeum is the first reference strain reported to have been affected by a recent lateral transfer of aprBA genes. The type strain of Thermacetogenium phaeum was isolated from a thermophilic anaerobic methanogenic reactor (UASB), and to date is the only recognized sulfate-respiring strain within the Gram-positive Thermoanaerobacteriales (Hattori et al., 2000). Its phylogenetic position in DsrAB-based trees has not yet been investigated and is therefore unknown. In support of a recent lateral aprBA transfer, the co-occurrence and predominance of Thermacetogenium phaeum and Thermodesulfovibrio species (recently classified as Thermodesulfovibrio aggregans and Thermodesulfovibrio ethanolicus; Y. Sekiguchi & H. Imachi, unpublished data) have been demonstrated by 16S rRNA analysis within the same granular sludge pellets of a thermophilic UASB reactor (Sekiguchi et al., 1998). This close cell-to-cell contact would have increased the probability of interspecies LGT. The successful acquisition and functional implementation of this novel metabolic trait in the ancestor of Thermacetogenium phaeum might have been accelerated by certain genetic and physiological prerequisites: (1) the genomic arrangement (sat-aprBA-qmoABC) in the Thermodesulfovibrio donor strain (see genomic structure of Thermodesulfovibrio yellowstonii; Fig. 4b), which would have enabled a concomitant lateral transfer of all relevant genes in one single event; (2) the pre-existing capability for sulfite respiration in the ancestor of Thermacetogenium phaeum [see Supplementary Table S1 for the presence of dsr genes in the genomes of the related Moorella thermoacetica and Carboxydothermus hydrogenoformans; the capability of dissimilatory sulfite reduction has also been proven for Carboxydothermus hydrogenoformans (Henstra & Stams, 2004)]; and (3) the utilization of the same major membrane quinone component in the donor and recipient, menaquinone-7. The latter would have allowed an immediate linkage of the newly acquired sulfate-respiration system to the electron-transport processes in the ancestor of Thermacetogenium phaeum. Although the stable, specific conditions within the UASB reactor might have favoured the artificial generation of the SRB strain Thermacetogenium phaeum, it is a representative example of the permanently ongoing evolutionary process of gene flux among the genomes of free-living prokaryotes in nature (Daubin & Ochman, 2004; Jain et al., 2003; Lerat et al., 2005; Zhaxybayeva et al., 2004).
Comparison of AprBA and DsrAB phylogeny from SRP
The topological comparison of AprB/A- and DsrAB-based trees revealed members of the same SRB lineages to be involved in lateral transfer of the dsr and apr genes. However, in the DsrAB-based tree, the deltaproteobacterial SRB group was monophyletic and the Desulfobacterium anilini-related SRB group was suggested to be the donor lineage for the xenologous dsrAB genes of the Desulfotomaculum subclusters Ib, Ic, Id and Ie, as well as Moorella thermoacetica (member of the Thermoanaerobacteriales). Desulfosporosinus and the Desulfotomaculum subclusters Ia, If and Ih have been postulated to represent the DsrAB-‘authentic’ Gram-positive SRB clades that were not affected by LGT (Imachi et al., 2006; Klein et al., 2001; Zverlov et al., 2005). The xenologous gene displacements of the orthologous dsrAB genes in subclusters Ib–Ie are supported by the AprB/A phylogeny of this study, because Desulfotomaculum subclusters Ia/If and Desulfosporosinus consistently branch close to the root of the DsrAB- and AprB/A-based trees. According to the AprB/A phylogeny, the Gram-positive SRB species are monophyletic and present a relative branching order of taxa that is congruent with 16S rRNA and unaffected by LGT. Conversely, the acquisition of the apr gene of the deltaproteobacterial SRB lineages (Syntrophobacteraceae, the Desulfobacterium anilini-related SRB group, Desulfomonile tiedjei and Desulfarculus baarsii) postulated by four independent events of xenologous gene displacement is supported by the moderate relationships among their orthologous DsrAB sequences (Zverlov et al., 2005). Indeed, this class of LGT has been reported to have influenced the evolutionary path of 10–15 % of the orthologous genes in the bacterial domain (Novichkov et al., 2004). The consistent xenologous branching positions from Thermodesulfobacteriaceae as well as Archaeoglobus in the AprA- and DsrAB-based trees (Klein et al., 2001) point to a concerted acquisition of the novel capabilities to respire sulfite and sulfate by ancient, dual LGT events from unknown donor lineages; however, an early lateral transfer from Gram-positive SRB is suggested for the latter genus, since Archaeoglobus species seemed to be affiliated most closely with this lineage. In contrast, the ancestors of Thermodesulfovibrio might initially have been sulfite reducers, as indicated by the 16S rRNA-congruent branching of their DsrAB sequences (Klein et al., 2001), and received their ability to respire sulfate later on their evolutionary path. A donor lineage for the LGT of their apr genes is not apparent. Interestingly, a congruent taxonomical classification of both uncultivated putative SRB strains (fosws39f7, fosws7f8) (Mussmann et al., 2005) was not possible on the basis of the Dsr and Apr sequences. A distinct origin is proposed for their xenologous apr and dsr genes, which are arranged as a metabolic island.
Correlation between Apr phylogeny and presence of proteins functionally associated with AprBA (Qmo and AprM) in SRP and SOB
The genome analysis indicated the development of two unrelated protein (complexes), AprM and QmoABC, mediating the electron transfer between the cytoplasmic sulfite oxidation/APS reduction process and the membrane quinone pool in correlation with the Apr phylogeny-based divergence of the SRP and SOB. The membrane-bound redox complex QmoABC that has been investigated and characterized from Desulfovibrio desulfuricans ATCC 27774 is proposed to act as a menaquinol/APS reductase oxidoreductase (Pires et al., 2003). In support, further experimental studies have demonstrated the coordinated down-regulation of the apr and qmo gene sets in a Desulfovibrio vulgaris strain Hildenborough mutant in the presence of nitrite (Haveman et al., 2004), and the dependence of the sulfate-reduction capability of Desulfotomaculum aeronauticum on amendment with the quinone precursor menadione (Hippe et al., 1997). Indeed, (1) the presence of the qmoABC genes and their close proximity to the aprBA genes in all genomes of SRP (see Table 3, Fig. 4) (whereas they are absent in sulfite- or thiosulfate-reducing species which lack the aprBA genes; Supplementary Table S1), and (2) the congruent protein phylogenies (trees not shown), support the proposed functional association of the Qmo complex and the APS reductase in SRP. The presence of SRP-related apr and qmo genes in Chlorobiaceae genomes is an indication of an electron transfer system similar to that of SRP in the green sulfur bacteria. In contrast, in genomes of SOB harbouring Allochromatium vinosum-related Apr sequences, the aprBA genes are always co-regulated with a preceding aprM (encoding a membrane-integral protein), whereas ORFs homologous to qmoABC are absent (see Table 3, Fig. 4b). The strictly conserved arrangement and the AprBA-congruent AprM tree topology (not shown) point to an essential role for AprM in Allochromatium vinosum and other Apr-related SOB. An in vivo function as a structural component (membrane anchor) for the cytoplasmic APS reductase or even a direct functional involvement in electron transfer, by analogy with the menaquinol : fumarate oxidoreductase of E. coli (Cecchini et al., 2002; Lancaster, 2003), are possible.
Unlike the APS reductase, the transmembrane redox complex that is functionally associated with the dissimilatory sulfite reductase (DsrAB) seems to be identical among the SRP and SOB. The homologous HmeABCDE and DsrMKJOP protein complexes have been proposed to operate in the electron transfer pathway between cytoplasmic DsrAB and the membrane-integral quinol/quinone pool (Dahl et al., 2005; Mander et al., 2002; Pires et al., 2006; Sander et al., 2006). The genome analyses in this study revealed DsrMKJOP homologues to be present in all sulfate- and sulfite-reducing as well as several sulfur-oxidizing bacteria (Supplementary Table S1).
Evolutionary scenario for the dissimilatory APS reductase as a key enzyme of the sulfate-reduction pathway in the light of Apr and Dsr phylogeny and geochemical data of Archean Earth
The global appearance of mass-independent isotope fractionation (MIF) in sedimentary sulfides and sulfates older than 2·45 billion years is consistent with low oxygen levels in the atmosphere (pO2 <10–5 present atmospheric level; PAL) and a global sulfur cycle dominated by atmospheric reactions during the Archean era (4·0–2·4 billion years ago) (Farquhar et al., 2000). Atmospheric sulfate generated by photolysis of SO2 would have represented the only significant abiogenic sulfate load of the ocean (Canfield, 2005; Farquhar et al., 2000; Strauss, 2003). Consistently, the absence of significant mass-dependent isotope fractionation between Archean sulfate and sulfide is an indication of a global sulfate concentration below 200 µM (Canfield et al., 2000; Canfield, 2001; Habicht et al., 2002) and the absence of an active global sulfur cycle in the Archean hydrosphere (Strauss, 2003). Sulfate in concentrations sufficient for energy conservation by sulfate respiration could have only been supplied by the metabolic activity of sulfur-oxidizing anoxygenic phototrophic bacteria (Canfield & Raiswell, 1999; Shen et al., 2001). Indeed, photosynthetic microbial mats dominated by anoxygenic phototrophic bacteria existed at 3·4 billion years ago (Tice & Lowe, 2004). The ancestors of the modern SRP could have originated in dependence on the favourable conditions provided by the sulfur-oxidizing anoxygenic phototrophic bacteria, and the sulfate-reduction process might have become established and geochemically expressed as early as 3·4 billion years ago (Shen et al., 2001) at local, restricted sites.
In agreement with the proposed geochemical setting of the Archean Earth, the Dsr and Apr phylogenies (Boucher et al., 2003; this study) consistently point to their origin and evolution as oxidative-operating enzymes. Since the DsrAB has been suggested to be of ancient origin (Dhillon et al., 2005), the reverse sulfate-reduction pathway might have evolved successively (sulfite reductase preceding the APS reductase and ATP sulfurylase) in an ancestral anoxygenic phototrophic bacterium. The development of DsrAB and its functionally associated DsrMKJOP complex (Dahl et al., 2005; Mander et al., 2002; Sander et al., 2006) would have allowed the utilization of hydrothermal-derived sulfide as a reductant for anoxygenic photosynthesis. The subsequent phylogenetic divergence into three DsrAB lineages (Boucher et al., 2003) might have been the result of two early, independent LGTs of the progenotic dsrAB (in concert with the dsrMKJOP genes) from the ancestral sulfide-oxidizing bacterial donor lineage to the ancestors of SRP and sulfur-respiring archaeal Pyrobaculum. In the same way as their modern equivalents, the ancient microbial mats might have presented hot spots for the metabolic diversification of the microbial community by LGT (Molin & Tolker-Nielsen, 2003; Sorensen et al., 2005). The ancestors of the SRB lineages might initially have been sulfite respirers by adaptive reversal of the oxidative-operating Dsr protein sets. The SRP-related DsrMKJOP of the Chlorobiaceae (Sander et al., 2006) might have arisen from an early xenologous replacement of the dsrMKJOP genes of an ancestral green sulfur bacterium with those from a sulfite reducer.
Since the end product of Dsr-mediated sulfide/sulfur oxidation is sulfite, the reverse APS reductase might primarily have been developed by the ancestral sulfide oxidizers for detoxification instead of energy conservation. Indeed, a sulfite-oxidation pathway via the intermediate APS would have allowed the simultaneous generation of energy (ATP) by substrate phosphorylation, which would have contributed to relieving energy limitation in a primitive anoxygenic phototrophic bacterium. Two conflicting scenarios would explain the DsrAB-incongruent AprBA tree topology and the appearance of two different membrane protein(s) systems that interact with APS reductase. First, after phylogenetic separation of the progenotic detoxifying APS reductase into the lineages of the anoxygenic phototrophic green and purple sulfur bacteria, the protein sets QmoABC and AprM emerged independently in these groups and allowed the utilization of sulfite as a reductant in anoxygenic photosynthesis. The concurrent presence of apr and qmo genes in Chlorobiaceae and SRP genomes would have been the result of a subsequent, concerted LGT from an ancestral green sulfur bacterial donor to the sulfite-respiring ancestors of the SRP lineages. The second, alternative scenario implies an early divergence of the progenotic detoxifying AprBA into the phylogenetic lineages of the SOB and the SRP analogous to the postulated evolutionary path of DsrAB (Boucher et al., 2003; Molitor et al., 1998). The Qmo complex would then have originated within an ancestral sulfite-reducing bacterial lineage and not in an ancestor of the green sulfur bacteria, whereas AprM developed in ancestral purple sulfur bacteria. Because of the restricted distribution of the apr and qmo genes within the Chlorobiaceae (B. Meyer and J. Kuever, unpublished results), their ancestors either never possessed or lost their ancestral apr genes due to functional replacement by convergently evolved proteins. Indeed, a thermophilic sulfate-reducing lineage (e.g. Thermodesulfovibrio) could have served as donor for the later acquisition or reacquisition of the entire gene locus by LGT. Irrespective of possible evolutionary scenarios, the basal branching AprBA lineage of Pyrobaculum aerophilum might represent the APS reductase type most closely related to the progenotic detoxifying form, because no gene coding for any of the proposed functionally associated proteins is present in the genome. The patchy and polyphyletic distribution of the sulfate-reduction pathway among prokaryotes and the late diversification of the major SRP lineages, despite the postulated early origin of the respiration process (Shen et al., 2001), might be the result of the persistently low sulfate content of ocean waters until 2·4 billion years ago (Canfield et al., 2000; Canfield, 2005; Farquhar et al., 2000; Habicht et al., 2002; Strauss, 2003), which restricted the abundance and ecological significance of this physiological group in the Archean era. The radiation of the SRP might have started with the beginning of the oxygenation of the atmosphere (Canfield et al., 2000; Canfield, 2005; Farquhar et al., 2000), which resulted in an increasing oceanic sulfate concentration during the Proterozoic era (2·5–0·54 billion years ago) (Canfield, 2005; Kah et al., 2004). A widespread lateral distribution of the sulfate-reduction pathway via mobilizable metabolic islands has been suggested (Friedrich, 2002; Mussmann et al., 2005). However, (1) the absence of characteristic mobility elements indicative of classical genomic islands in the metagenome sequences (Mussmann et al., 2005), (2) the generally scattered arrangement of the dsr and apr genes in the genomes of validated SRP, and (3) the differing phylogenies of DsrAB (Boucher et al., 2003; Klein et al., 2001; Zverlov et al., 2005) compared with those of AprBA (this study) and Sat (Sperling et al., 1998) caused by non-parallel LGT events, seem to contradict this hypothesis for the evolution of the dissimilatory sulfate-reduction pathway.
|
NOTE ADDED IN PROOF |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
Edited by: G. Muyzer
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
Canfield, D. E. (2001). Isotope fractionation by natural populations of sulfate-reducing bacteria. Geochim Cosmochim Acta 65, 1117–1124.[CrossRef]
Canfield, D. E. (2005). The early history of atmospheric oxygen: homage to Robert A. Garrels. Annu Rev Earth Planet Sci 33, 1–36.[CrossRef]
Canfield, D. E. & Raiswell, R. (1999). The evolution of the sulfur cycle. Am J Sci 299, 697–723.
Canfield, D. E., Habicht, K. S. & Thamdrup, B. (2000). The Archean sulfur cycle and the early history of atmospheric oxygen. Science 288, 658–661.[CrossRef][Medline]
Cecchini, G., Schröder, I., Gunsalus, R. P. & Maklashina, E. (2002). Succinate dehydrogenase and fumarate reductase from Escherichia coli. Biochim Biophys Acta 1553, 140–157.[Medline]
Dahl, C. & Trüper, H. G. (1994). Enzymes of dissimilatory sulfide oxidation in phototrophic sulfur bacteria. Methods Enzymol 243, 400–421.
Dahl, C., Engels, S., Pott-Sperling, A. S., Schulte, A., Sander, J., Lubbe, Y., Deuster, O. & Brune, D. C. (2005). Novel genes of the dsr gene cluster and evidence for close interaction of Dsr proteins during sulfur oxidation in the phototrophic sulfur bacterium Allochromatium vinosum. J Bacteriol 187, 1392–1404.
Daubin, V. & Ochman, H. (2004). Bacterial genomes as new gene homes: the genealogy of ORFans in E. coli. Genome Res 14, 1036–1042.
DeLong, E. F. (1992). Archaea in coastal marine environments. Proc Natl Acad Sci U S A 89, 5685–5689.
Dhillon, A., Goswami, S., Riley, M., Teske, A. & Sogin, M. (2005). Domain evolution and functional diversification of sulfite reductases. Astrobiology 5, 18–29.[CrossRef][Medline]
Farquhar, J., Bao, H. M. & Thiemens, M. (2000). Atmospheric influence of Earth's earliest sulfur cycle. Science 289, 756–758.[CrossRef][Medline]
Friedrich, M. W. (2002). Phylogenetic analysis reveals multiple lateral transfers of adenosine-5'-phosphosulfate reductase genes among sulfate-reducing microorganisms. J Bacteriol 184, 278–289.
Fritz, G., Buchert, T., Huber, H., Stetter, K. O. & Kroneck, P. M. H. (2000). Adenylylsulfate reductases from archaea and bacteria are 1 : 1 alpha beta-heterodimeric iron-sulfur flavoenzymes – high similarity of molecular properties emphasizes their central role in sulfur metabolism. FEBS Lett 473, 63–66.[CrossRef][Medline]
Gogarten, J. P. & Townsend, J. P. (2005). Horizontal gene transfer, genome innovation and evolution. Nat Rev Microbiol 3, 679–687.[CrossRef]
