Retrieval of first genome data for rice cluster I methanogens by a combination of cultivation and molecular techniques

Authors


*Corresponding author. Tel.: +49 6421 178 720; fax: +49 6421 178 809, E-mail address: liesack@staff.uni-marburg.de

Abstract

We report first insights into a representative genome of rice cluster I (RC-I), a major group of as-yet uncultured methanogens. The starting point of our study was the methanogenic consortium MRE50 that had been stably maintained for 3 years by consecutive transfers to fresh medium and anaerobic incubation at 50 °C. Process-oriented measurements provided evidence for hydrogenotrophic CO2-reducing methanogenesis. Assessment of the diversity of consortium MRE50 suggested members of the families Thermoanaerobacteriaceae and Clostridiaceae to constitute the major bacterial component, while the archaeal population was represented entirely by RC-I. The RC-I population amounted to more than 50% of total cells, as concluded from fluorescence in situ hybridization using specific probes for either Bacteria or Archaea. The high enrichment status of RC-I prompted construction of a large insert fosmid library from consortium MRE50. Comparative sequence analysis of internal transcribed spacer (ITS) regions revealed that three different RC-I rrn operon variants were present in the fosmid library. Three, approximately 40-kb genomic fragments, each representative for one of the three different rrn operon variants, were recovered and sequenced. Computational analysis of the sequence data resulted in two major findings: (i) consortium MRE50 most likely harbours only a single RC-I genotype, which is characterized by multiple rrn operon copies; (ii) seven genes were identified to possess a strong phylogenetic signal (eIF2a, dnaG, priA, pcrA, gatD, gatE, and a gene encoding a putative RNA-binding protein). Trees exemplarily computed for the deduced amino acid sequences of eIF2a, dnaG, and priA corroborated a specific phylogenetic association of RC-I with the Methanosarcinales.

1Introduction

The phylogenetic diversity of methanogens is reflected by the current classification that divides them into five orders, including Methanosarcinales, Methanomicrobiales, Methanobacteriales, Methanococcales, and Methanopyrales[1]. Differentiation can be made, depending on the compounds used for methanogenic conversions, between the CO2-reducing (H2, formate, and alcohols except methanol), methylotrophic (methylated compounds), and acetoclastic (acetate) pathways. All methanogens capable of utilizing acetate as substrate for growth belong to the Methanosarcinales. Members of the genus Methanosarcina can be regarded as the most versatile methanogens possessing both the methylotrophic and acetoclastic pathways. In addition, some Methanosarcina spp. are also able to reduce CO2. In contrast, Methanosaeta spp. are obligately acetoclastic methanogens [2]. With the exception of some organisms belonging to the Methanosarcinales, all methanogens are able to use molecular hydrogen as a source of electrons for the reduction of their substrates to CH4[3]. The principal similarities, but also the differences between the methanogenic pathways have been characterized in detail by biochemical studies and comparative genomics. Complete genomes have been determined for representative methanogen types, such as Methanosarcina mazei Goe1, M. acetivorans C2A, Methanocaldococcus jannaschii DSM 2661, and Methanothermobacter thermautotrophicusΔH.

Despite the large number of cultured representatives, the cultivation-independent retrieval of euryarchaeal 16S rRNA genes has shown that members of novel methanogen groups still elude isolation [4–6]. One of these as-yet uncultured groups has been designated rice cluster I (RC-I) and has been shown to be one of the predominant methanogen populations in rice paddy soils of various geographic locations [7]. The environmental significance of RC-I is also underlined by its more recent detection in a wide variety of anoxic habitats, such as peatlands [8–10], water columns and sediments of freshwater lakes [11,12], and contaminated aquifers [13,14].

RC-I forms a clade within the phylogenetic radiation characterized by members of the orders Methanosarcinales and Methanomicrobiales. Comparative analyses of both 16S rRNA and mcrA gene sequences suggest a specific phylogenetic association of RC-I with the Methanosarcinales[5,15]. The mcrA gene encodes the active site of methyl-coenzyme M reductase, the key enzyme of methanogenesis. Nonetheless, pure cultures of RC-I would have to be classified as representatives of a new family or even new order due to the phylogenetic distinction between RC-I and members of both the Methanosarcinales and Methanomicrobiales[5].

Several attempts have been made to obtain RC-I representatives in pure culture [10,15]. Although conventional procedures such as dilution series and roll-tube enrichment techniques combined with antibiotic treatments have been thoroughly applied, these attempts failed. However, one of these approaches resulted in the enrichment of a methanogenic consortium (MRE50) that contained a phylogenetically diverse population of RC-I among other microbial components [15]. Consortium MRE50 was originally established in the year 2000, using Italian rice field soil as the inoculum. Since then it has been stably maintained by consecutive transfers to fresh medium.

Here, we report on a combined approach of molecular ecology techniques and laboratory cultivation to examine whether the enrichment status of RC-I in culture MRE50 has changed during subculture. Both its high enrichment status and low intracluster diversity prompted recovery of large genome fragments of RC-I representatives from biomass of culture MRE50. Computational analysis of three fully sequenced genomic fragments provided further support for the proposal of a specific phylogenetic association of RC-I with the Methanosarcinales.

2Materials and methods

2.1Methanogenic consortium MRE50

The methanogenic consortium was maintained in a total volume of 500 ml of anoxic mineral medium supplemented with 500 mg of yeast extract under a headspace of H2/CO2 (80:20 v/v, 150 kPa). The mineral medium was prepared as described earlier [16] with slight modifications. These included the additions of riboflavin (0.05 mg l−1) and selenite/tungstate solution [17], and usage of half the concentration of trace element solution. Na2S (1.5 mM) was used as the reducing agent. The pH was adjusted to approximately 7.0 at 25 °C. Culture MRE50 was transferred fortnightly to fresh medium in 1-liter serum bottles (10% inoculum) and was incubated at 50 °C. Optical density was measured with a spectrophotometer (Biophotometer, Eppendorf, Hamburg, Germany) at 600 nm. Methane formation and consumption of hydrogen and carbon dioxide were followed by gas chromatography [18]. Organic acids and alcohols in liquid samples were analyzed using high-pressure liquid chromatography as described previously [19].

2.2DNA extraction

Cells from culture MRE50 were harvested at late stationary growth phase (14 days after inoculation). Total DNA was extracted and purified using two different protocols. For PCR and cloning, cells were lysed by mechanical agitation in a bead beater (Mikro-Dismembrator S; B. Braun Biotech International, Melsungen, Germany). A 2-ml screw-cap reaction vessel filled with glass beads (0.17 mm in diameter) and a suspension of cells and phosphate buffer (pH 8.0) containing 1% sodium dodecyl sulfate (SDS) were shaken twice for 60 s at 2500 rpm. If DNA was extracted for real-time PCR, SDS was omitted from the lysis buffer. The crude DNA extract was purified using the Wizard® DNA Clean-Up-Kit (Promega, Mannheim, Germany) according to the manufacturer’ s instructions and stored at 4 °C.

For construction of a fosmid clone library, high-molecular-weight (HMW) DNA was extracted using a newly developed cryogenic grinding protocol. Harvested cells were resuspended in cell suspension buffer (10 mM Tris–HCl, pH 7.2, 20 mM NaCl, 50 mM EDTA) to a final density of 6 × 109 cells ml−1, transferred to a 3-ml Teflon container (B. Braun Biotech International), and frozen in liquid nitrogen. After addition of a 8.0-mm-diameter stainless steel bowl (B. Braun Biotech International), the frozen cells were lysed in the Mikro-Dismembrator S cell disruptor by grinding for 60 s at 2500 rpm. Further purification of the crude HMW DNA extract, including proteinase K treatment, protein denaturation, separation of low-molecular-weight impurities by passage through an anion-exchange resin column (Qiagen genomic-tip 500/G, Qiagen, Hilden, Germany) and isopropanol precipitation, was carried out according to the protocol given in the Qiagen Genomic DNA Handbook for the isolation of genomic DNA from bacteria. Finally, the HMW DNA was resuspended in 1 ml of Tris–EDTA buffer (pH 8.0) and stored at 4 °C for further analyses.

2.3Real-time PCR

The PCR assays used in real-time PCR are specified in Table 1. The PCR assay C was developed using the “probe design” tool of the ARB program package [24]. 16S rRNA gene sequences of RC-I and other Euryarchaeota currently available in public databases were used for reference. The optimal annealing temperature was determined for positive control DNA (see below) in relation to those negative control DNAs, which exhibited a low number of mismatches in the target regions. This included DNA extracts from Methanothermobacter marburgensis (DSM 2133), Methanococcus maripaludis (DSM 277), and Methanoplanus endosymbiosus (DSM 3599) (three mismatches for RC-If].

Table 1.  PCR assays used in this study
AssayTarget groupMarker geneFragment length (bp)Primer setSequence (5′ to 3′)aAnnealing temperature (°C)Reference
  1. aAccording to IUB code.

ABacteria16S rRNA400519fxCAG CMG CCG CGG TAA NWC50(x20,21]
    907bCCG TCA ATT CMT TTR AGT  
BArchaea16S rRNA800A109fACK GCT CAG TAA CAC GT52[4]
    A934bGTG CTC CCC CGC CAA TTC CT  
CRC-I16S rRNA250RC-IfAGG CAA CTG CGA TAG GGG63This study
    RC-IbCCC AGT CCC AAG CAA TGT  
DBacteria16S rRNA15009fGAG TTT GMT CCT GGC TCA G48[21,22]
    1492bACG GYT ACC TTG TTA CGA CTT  
EMethanogensmcrA500MCRfTAY GAY CAR ATH TGG YT50[23]
    MCRbACR TTC ATN GCR TAR TT  

Real-time PCR mixtures contained approximately 50 ng of DNA, 2.5 μl of 10 × PCR buffer (Invitrogen, Karlsruhe, Germany), 1.5 mM MgCl2, 50 μM of each deoxynucleoside triphosphate (dNTPs, Amersham Pharmacia Biotech, Freiburg, Germany), 0.33 μM of each primer (MWG Biotech, Ebersberg, Germany), 0.25 μl of 500-fold diluted SybrGreen™ I (Biozym, Hess. Oldendorf, Germany), 40 μg of bovine serum albumin (Roche Biochemicals, Mannheim, Germany), and 0.3 U of iTaq Polymerase (Invitrogen). Amplification was performed in a total volume of 25 μl in Thermo-fast 96 PCR plates (PeqLab, Erlangen, Germany) sealed with iCycler IQ optical quality tapes (Bio-Rad, Munich, Germany), using a DNA thermal cycler (model iCycler IQ; Bio-Rad). The thermal profile was as follows: initial denaturation for 5 min at 94 °C and 40 cycles consisting of denaturation at 94 °C for 60 s, primer annealing for 50 s at a temperature as specified for each target gene (Table 1), and elongation at 72 °C for 60 s. Each measurement was carried out in triplicate. Data analysis was carried out in relation to standard calibration curves using the iCycler software (Version 2.3.1370; Bio-Rad) [20,25]. Calibration curves for the different PCR assays were generated using dilution series of quantitative standards as follows: genomic DNA of Escherichia coli (Roche Diagnostics) for PCR assay A and a single RC-I 16S rRNA gene amplicon for PCR assays B and C. Precise quantification of DNA standards was performed using the PicoGreen double-stranded DNA quantification kit (Molecular Probes, Leiden, The Netherlands) according to manufacturer’ s instructions. The amounts of PCR amplicons measured in real-time PCR were converted into target gene numbers per milliliter of culture MRE50.

2.4Single gene clone libraries

PCR assays B, D, and E were used to generate PCR products for cloning (Table 1). Except for mcrA, the reaction mixture contained 1–5 ng of DNA, 5 μl of 10 × PCR buffer (Promega), 1.5 mM of MgCl2, 50 μM of each dNTP (Amersham Pharmacia Biotech), 0.33 μM of each primer (MWG Biotech), and 1 U of Taq DNA polymerase (Promega). For PCR amplification of mcrA, PCR buffer (Promega), MgCl2, and dNTPs were replaced with 1 × MasterAmp PCR PreMix B (Epicentre, Madison, WI, USA). Amplification was performed in a total volume of 50 μl in 0.2-ml reaction tubes, using a DNA thermal cycler (model GenAmp 9700; Applied Biosystems, Weiterstadt, Germany). The thermal PCR profile was as follows: initial denaturation for 5 min at 94 °C and 30 cycles consisting of denaturation at 94 °C for 1 min, primer annealing for 1 min at a temperature as specified for each target gene (Table 1), and elongation at 72 °C for 2 min. The final elongation step was extended to 7 min. PCR amplicons were purified and cloned using commercial kits according to the manufacturer’ s instructions (QIAquick PCR purification kit (Qiagen) and TOPO-TA Cloning® Kit (Invitrogen)]. Randomly selected clones were checked for correct insert size via standard vector-targeted PCR and agarose gel electrophoresis. Sequencing was performed on an ABI Prism 377 DNA sequencer using dye terminator chemistry as specified by the manufacturer (Applied Biosystems).

2.5Terminal restriction fragment length polymorphism (T-RFLP) analysis

Bacterial and archaeal 16S rRNA gene amplicons (assays B and D, Table 1) were generated from total DNA of culture MRE50 as specified above except that primer 9f (assay D) or A934b (assay B) was used labelled with 6-carboxyfluorescein. Restriction of the fluorescently labelled amplicons with MspI (Bacteria) or TaqI (Archaea), size-separation of the terminal restriction fragments (T-RFs), and data analysis have been described elsewhere [26,27].

2.6Whole-cell hybridization

Generally, fixation of cells and hybridization with labelled probes followed a procedure described earlier [28]. Briefly, cells were harvested during the exponential growth phase, resuspended in 0.5 ml of phosphate-buffered saline, and fixed with a freshly prepared paraformaldehyde solution (4%). Whole-cell hybridization was carried out on Teflon-coated slides with eight wells. Aliquots (1–2 μl per well) of the fixed cell suspensions were dried and passed through a dilution series of 50%, 80%, and 100% ethanol (v/v]. For detection of bacterial cells, the slide was hybridized with the 5(6)-carboxyfluorescein-N-hydroxysuccinimide ester (FLUOS) labelled probe EUB338 [29]. Archaeal cells were stained with a mixture of the indocarbocyanine dye (Cy3) labelled probes ARC915/ARC344 [30]. The concentration of formamide within the hybridization buffer was 30%.

2.7Construction of a fosmid library

HMW DNA was size-fractionated by pulse field gel electrophoresis (PFGE) in a CHEF-DR® III apparatus (Bio-Rad) using the following parameters: 1–12 s switch time, 6 V/cm, 15 h. To enable the recovery of DNA fragments in the desired size range, the central region of the 1% pulse-field agarose gel was replaced by 1% low-melting point agarose (Bio-Rad). Ethidium bromide-stained gel slices containing electrophoresed DNA and the Low Range PFG-marker (New England Biolabs, Beverly, MA, USA) allowed recovery of HWM DNA of appropriate size (approximately 40–50 kb) from the unstained agarose gel. The HMW DNA was purified by agarose digestion with β-agarase I as described by the manufacturer (New England Biolabs).

The library was constructed using the CopyControl™ Fosmid Library Production Kit according to the manufacturer’ s instructions (Epicentre). Briefly, blunt-ended, 5′-phosphorylated DNA fragments were generated and ligated into the fosmid vector pCC1FOS™ (Epicentre). The molar ratio of insert to vector DNA was approximately 1:10. Further steps included the packaging of ligation products into phage particles and transformation of competent cells by infection. Single colonies plated on LB plates (12.5 μg ml−1 chloramphenicol (cam]) were randomly picked and transferred into 96-well microtitre plates containing 150 μl of LB/cam medium per well. Plates were incubated at 37 °C overnight. Grown cells were replicated and incubated overnight in 96-well plates containing freeze medium (2.5% w/v granulated LB broth, 13 mM KH2PO4, 36 mM K2HPO4, 1.7 mM sodium citrate, 6.8 mM (NH4)2SO4, 4.4% v/v glycerol, 0.4 mM MgSO4, 12.5 μg ml−1 cam) and stored at −80 °C.

2.8PCR-based screening for archaeal genome fragments

Aliquots (10 μl) of transformed E. coli cells grown in one row of a 96-well microtiter plate were combined to generate row pools. Pooled cells were lysed by boiling for 15 min. Aliquots (20 μl) of the row pool lysates were combined to generate plate pools. The lysates of row and plate pools were stored at −80 °C.

Microtitre plates that contained at least one clone carrying an archaeal 16S rRNA gene were identified by PCR-based screening (assay B, Table 1), using 1 μl of the supernatant of each plate pool as the template. Row pools of test-positive plates were screened in the same way. Finally, clones of single rows that had been tested positively were analyzed separately. Test-positive fosmid clones were purified from a 2-ml overnight culture by boiling for 5 min. The cell debris was subsequently pelleted by centrifugation with a microcentrifuge at 8000 ×g for 5 min. PCR amplicons of archaeal 16S rRNA genes were sequenced as specified above.

The internal transcribed spacer (ITS) regions of 16S–23S rRNA genes were analyzed by cycle sequencing of test-positive fosmid clones as described previously [31]. Fosmid DNA was extracted using the QIAprep Spin Miniprep kit (Qiagen) according to the manufacturer’ s instructions. Sequencing was carried out with a newly designed oligonucleotide primer (5′-GCCGTAGGGGAATCTGCGGC-3′) that targeted a stretch at the 5′-end of the 23S rRNA gene. After completion of the reaction, cycle sequencing products were purified with AutoSeq™ G-50 columns (Amersham Pharmacia Biotech Inc., Piscataway, USA).

2.9DNA sequencing of fosmid clones

The fosmid clones 4B7mr3, 1B6mr2, and 3D10mr4 were completely sequenced using a shotgun approach. Fosmid DNA was sonicated, end-repaired, size-selected, ligated into the SmaI-digested pUC19 (Fermentas, St. Leon-Rot, Germany) and transferred into Escherichia coli (strain DH10B; Gibco, Karlsruhe, Germany) via electroporation. Shotgun clone inserts were amplified by PCR [32] and sequenced from both sides using dye terminator chemistry and 3730 ABI capillary sequencer systems (Applied Biosystems). Regions of weak quality were improved by resequencing and primer walking to achieve finished sequence quality of at least phred 30, which corresponds to 99.9% accuracy of the base call (i.e., a 1 in 1000 probability that the base is called wrong).

2.10Computational analyses

2.10.1Phylogenetic placement

Phylogenetic trees of nearly complete bacterial 16S rRNA gene sequences (1400 bp) as well as partial archaeal 16S rRNA gene sequences (900 bp) were constructed using fastDNAml [33] and the ARB program package [24]. We established databases for selected proteins (length of the respective protein sequence used for tree construction is given in parentheses) as follows: McrA (147 aa), eIF2a (264 aa), DnaG (428 aa), and PriA (263 aa). Protein sequences were retrieved with standard software tools and the BLAST server of the National Center for Biotechnology Information using the BLAST algorithm [34] with non-redundant protein databases (compiled from SWISSPROT, TrEMBL and PIR). Proteins previously annotated as conserved hypotheticals were included into the databases based on alignment homologies. Each of the four sequence databases was integrated into the ARB program package. Protein trees were reconstructed by performing maximum likelihood analyses (protml of the Molphy package [35] implemented in ARB). Identity and similarity values were calculated using the appropriate tool of the ARB package.

2.10.2Gene annotation

Annotation was carried out with ORPHEUS for gene prediction [36], HTGA for a first automated annotation [32] and Artemis for the final manual annotation (Sanger Institute, Hinxton, UK). Homology searches of translated open reading frames (ORFs) were carried out in HTGA using the BLAST algorithm [34] against a non-redundant protein database (compiled from SWISSPROT, TrEMBL and PIR) and other integrated databases [32]. Only ORFs larger than 200 bp were considered for annotation. Gene densities were calculated by subtracting the sum of lengths of the noncoding sequence regions from the total size of the fosmid inserts. Based on comparative analysis of the putative gene products, ORFs without any significant similarity to public database entries were designated to encode hypothetical proteins. ORFs exhibiting similarities to genes of unknown function were accounted as conserved hypotheticals.

2.11Nucleotide accession numbers

Bacterial 16S rRNA gene sequences (clones MRE50b01 to MRE50b30) retrieved from culture MRE50 have been deposited in the GenBank, EMBL, and DDBJ databases under Accession Nos. AY684072 to AY684101. The accession number of the single mcrA gene sequence type (clone MRE50mcrA) is AY683452. The accession numbers of the three completely sequenced fosmid clones 4B7mr3, 1B6mr2, and 3D10mr4 are CR626856, CR626858, and CR626857, respectively.

3Results

3.1Characterization of culture MRE50

3.1.1Gas dynamics

Based on the turbidity of culture MRE50, microbial growth could be confirmed visually within 2 days of transfer to fresh medium. Methane production started approximately 24 h after inoculation and increased steadily with time. Concomitant consumption of H2 and CO2 (Fig. 1(a)) could be recorded only after day 2, while a slight accumulation of H2 and CO2 was observed during the first 48 h of incubation. The initial methane production rate was 0.6 μmol CH4 ml−1 day−1, as calculated between days 1 and 4 after transfer to fresh medium. Methane was the predominant product. Acetate, propionate, butyrate, caproate, formate, and ethanol were detected only at low concentrations (0.1–1.2 μmol ml−1). The concentration of formate (0.2 μmol ml−1) decreased between days 2 and 5 to undetectable amounts, whereas the concentrations of acetate and butyrate increased slightly until day 5 (0.3 and 1.0 μmol ml−1, respectively). The stoichiometry of the conversion of H2 and CO2 to CH4 was calculated at several time points during the incubation (Fig. 1). Approximately, 3–4 mol H2 and 1 mol CO2 were consumed per mol of CH4 formed. Other experiments resulted in similar stoichiometries (data not shown).

Figure 1.

Growth dynamics of culture MRE50. (a) Headspace gas analysis measured for methane (▪), hydrogen (▴), and carbon dioxide (Δ) concentrations during incubation for 35 days. (b) Optical density measurements at 600 nm (•, dashed line) and target numbers of bacterial (▾), archaeal (▪), and RC-I-specific (□) 16S rRNA genes. Target numbers were measured in triplicate at five different time points (0, 2, 5, 9, and 24 days).

Consortium MRE50 grew only with H2/CO2 as energy substrate. No growth was obtained using acetate under a gas atmosphere of N2/CO2. However, the addition of small amounts of acetate and also of cysteine–HCl stimulated overall growth on H2/CO2 (data not shown).

3.1.2Real-time PCR

This method was used to assess changes in the target numbers of bacterial and archaeal 16S rRNA genes over time (Fig. 1(b)). The number of bacterial 16S rRNA genes increased significantly from (6.6 ± 0.2) × 105 to (9.2 ± 0.2) × 107 targets ml−1 within the first 48 h of inoculation. A lag phase prior to bacterial growth was not observed. After 48 h, the number of bacterial 16S rRNA gene targets decreased rapidly, suggesting that the bacterial cells lysed and decomposed. The number of archaeal 16S rRNA genes increased from (8.8 ± 0.1) × 106 to (3.9 ± 0.3) × 107 targets ml−1 between day 2 and 5. The maximum number of archaeal 16S rRNA genes was observed on day 9 ((4.3 ± 0.1) × 107 targets ml−1 of culture]. The target number of 16S rRNA genes determined either for archaea or specifically for members of RC-I was nearly identical. The target numbers of mcrA genes were approximately three times lower (data not shown) than those determined for archaeal 16S rRNA genes for each sampling point.

3.1.3Bacterial diversity

Phylogenetic analysis assigned 30 randomly sampled 16S rRNA gene sequences to the families Thermoanaerobacteriaceae and Clostridiaceae, two phylogenetically defined taxa of the low mol% G + C gram-positive phylum [1] (Fig. 2). The majority of these sequence types belonged to a novel sublineage of uncultured bacteria within the Thermoanaerobacteriaceae. So far, this sublineage has been defined by 16S rRNA gene sequences retrieved from various anoxic habitats, including rice paddy soil, hot springs, and a dechlorinating enrichment culture. The intracluster diversity of this sublineage is high, exhibiting 16S rRNA gene sequence dissimilarity values of up to 19.1%.

Figure 2.

Maximum likelihood dendrogram showing the bacterial 16S rRNA gene sequences recovered from consortium MRE50 in relation to clone sequences (BSV81, OPB54, SHA-16, ST12, BSV-72, BIri47, IA-10, 16SX-2, and BA089) and cultured representatives of the low mol% G + C gram-positive phylum. The habitat sources of the environmental sequences are indicated. Accession numbers (GenBank, EMBL, and DDBJ entries) of reference sequences are given in parentheses. The root was determined by using archaeal 16S rRNA gene sequences as an outgroup. The scale bar represents 10% sequence divergence.

Other clone sequences retrieved from culture MRE50 exhibited moderate relationships to cultured bacteria (closest relatives and 16S rRNA gene sequence identity values are given in parentheses), including clones MRE50b24 (Moorella glycerini, 89.9%) and MRE50b25 (Gelria glutamica, 92.6%) in the Thermoanaerobacteriaceae, clones MRE50b28 and MRE50b29 in the clostridial cluster III (Clostridium papyrosolvens, approximately 92.0%) [37], and MRE50b26 (Clostridium sp. FCB90-3, 94.2%) in the Clostridiaceae. The T-RFs predicted in silico for the bacterial clone sequences matched exactly with the major T-RFs observed in a T-RFLP fingerprint pattern that had been generated for the bacterial component of culture MRE50 (data not shown).

3.1.4Archaeal diversity

Only two slightly different RC-I sequence types were identified in a set of 32 archaeal 16S rRNA genes that had been randomly sampled for analysis. Type I (G-type) contained a single guanidine at RC-I 16S rRNA gene nucleotide position 174, whereas type II (T-type) exhibited a thymidine triplet at the corresponding position. Both sequence types grouped within a cluster that was defined by the environmental clone sequences AS08-16, ARR24, pST-6, and 348/B-2 (Fig. 3(a)). The intracluster identity values were 97–98%. The T-RFLP fingerprint pattern generated for the archaeal component of culture MRE50 exhibited only a single T-RF (data not shown). This 392-bp T-RF was previously shown to be indicative of RC-I sequence types [26].

Figure 3.

Comparison of maximum likelihood dendrograms constructed for (a) 16S rRNA gene and (b) McrA sequences. The MRE50 sequence types were recovered from both single gene libraries and the fosmid library. They are shown in relation to environmental RC-I sequences (I) and cultured methanogens of the orders Methanosarcinales (II), Methanomicrobiales (III), Methanopyrales (IV), Methanobacteriales (V), and Methanococcales (VI). The habitat sources of the environmental sequences are indicated. Accession numbers (GenBank, EMBL, and DDBJ entries) of reference sequences are given in parentheses. The root of the 16S rRNA gene tree was determined by using the 16S rRNA sequence of E. coli as an outgroup, while the McrA tree is unrooted. The scale bars represent 10% sequence divergence.

Nineteen mcrA clones randomly selected for analysis were found to exhibit complete sequence identity on a stretch of 463 bp (designated clone MRE50mcrA). Phylogenetic analysis of its deduced amino acid sequence assigned MRE50mcrA to RC-I. MRE50mcrA is most closely related to rice paddy soil clone RS-MCR22 (97.2% identity, Fig. 3(b)).

3.1.5Whole-cell hybridization

Phase-contrast microscopy revealed a diverse microbial population consisting of mainly rod-shaped cell types (Fig. 4). Staining of the bacterial component identified various morphotypes, including rods of 3–10 μm in length, filaments, and rods with terminally located spores (the latter morphotype is not illustrated). All cells stained by the archaeal probes were rods approximately 1 μm wide and 5 μm long. More than 50% of total cells could be assigned to the Archaea.

Figure 4.

Whole-cell hybridization of culture MRE50. (a) Epifluorescence micrograph showing an overlay of whole-cell hybridizations with specific probes for either Bacteria (green) or Archaea (red). (b) Phase-contrast image. The scale bar (10 μm) applies to both images.

3.2Retrieval and analysis of genome fragments from RC-I

3.2.1Construction of a large insert clone library

Preparation of HMW DNA applying the cryogenic grinding method resulted in DNA fragments 30–70 kb in size. A fosmid library was constructed that consisted of 41,088 clones with insert sizes ranging from 36 to 45 kb (data not shown). This corresponded to a total of 1.64 Gbp or approximately 410 genome equivalents, assuming 4 Mb of relative genome size.

3.2.2Screening for RC-I genome fragments

Clones carrying an archaeal 16S rRNA gene were detected by PCR-based screening with high frequency (3.2% of 8800 clones tested). Thirty-two test-positive genome fragments were selected for further analysis. All of them matched exactly one of the two slightly different RC-I 16S rRNA gene sequence types that had been retrieved from culture MRE50 by construction of single gene clone libraries (Fig. 3(a)).

Comparative analysis of the ITS regions of the 16S–23S rRNA genes revealed only two distinct ITS types for the 32 test-positive fosmid clones. These two types differed mainly in the absence or presence of a tRNA-Ala gene. The combined interpretation of the 16S rRNA gene and ITS sequence data led to the conclusion that the 32 clones corresponded to at least three different sequence variants of the ribosomal RNA (rrn) operon. A representative clone of each putative variant was chosen for detailed analysis (Fig. 5).

Figure 5.

Comparison of rrn operon variants located on the RC-I genomic fragments 4B7mr3, 1B6mr2, and 3D10mr4. (a) Position of the rrn operon sequences (shaded) on the fosmid clone inserts. Designation and size of the fragments are indicated. (b) Magnification of a highly conserved 5414-bp sequence region, encoding the three ribosomal RNAs as well as an additional tRNA-Ala (“t”). The latter was not detected on clone insert 3D10mr4. The lengths (bp) of defined subregions are given. The number of mismatches within these subregions between insert 3D10mr4 and those of inserts 4B7mr3 and 1B6mr2 are marked with asterisks.

3.2.3Annotation of fosmid clones 4B7mr3, 1B6mr2, and 3D10mr4

The complete sequences of the three fosmid clones were determined, resulting in 122.5 kb of RC-I sequence information. A total of 108 open reading frames (ORFs) were identified on the three RC-I genome fragments. Significant similarities to genes encoding proteins with known functions were found for 49 ORFs. Nine ORFs showed significant similarities to genes encoding conserved hypotheticals. Of these 58 proteins, 46 exhibited similar sequences to predicted proteins of archaea, 11 to those of bacteria, and only one showed similarity to a predicted eukaryotic protein (Tables 2–4). The organismal source of those archaeal proteins that exhibited the lowest e values (see footnote to Tables 2–4) to the 46 counterparts was as follows: Methanosarcina spp. (30 proteins), other methanogens (six proteins), and non-methanogenic archaea (10 proteins). Fifty putative proteins showed no similarity to public database entries.

Table 2.  Predicted RNA and protein-encoding genes of fosmid insert 4B7mr3
ORFProtein size (AA)Predicted function (gene) (EC)COGa/IPRbPathwayBest BLAST hite valuec
  1. Nineteen hypothetical proteins without any similarity to BLAST entries were excluded.

  2. aCluster of orthologous groups (COG) of proteins, i.e., the proteins that comprise each COG are assumed to have evolved from an ancestral protein, and are therefore homologs.

  3. bInterPro entry (InterPro is a database of protein families, domains and functional sites in which identifiable features found in known proteins can be applied to unknown protein sequences.)

  4. cExpect value: probability that the associated match is due to randomness; the lower the e value, the more significant the match.

  5. dORFs predicted to represent informational genes that possess a strong phylogenetic signal (Table 5).

3430Oligosaccharide repeat unit transporter2244/−Transport of organic compoundsMethanosarcina mazei Goe17e − 45
7613Conserved hypothetical proteinMethanosarcina mazei Goe1e − 19
9336N-terminal acetyltransferase complex subunit Ard1 (RimI)0456/000182Fatty acid metabolismMethanothermobacter thermautotrophicusΔH2e − 12
17795Alkaline serine protease (Peptidase S8)−/000209Protein modification and degradationPyrococcus furiosus3e − 59
20384Conserved hypothetical proteinPyrococcus furiosuse − 82
22497Argininosuccinate lyase (arginosuccinase) (ASAL) (ArgH) (EC 4.3.2.1)0165/−Amino acid metabolism (alanine, aspartate and glutamate)Methanosarcina mazei Goe1e − 142
23419Glutamyl-tRNA (Gln) amidotransferase subunit D (GatD) (EC 6.3.5.-)d0252/006034Translation factorMethanosarcina mazei Goe1e − 136
24701Glutamyl-tRNA (Gln) amidotransferase subunit E (GatE) (EC 6.3.5.-)d−/004413Translation factorMethanosarcina acetivorans C2A0
26269Formamidopyrimidine-DNA glycosylase (Nei) (EC 3.2.2.23)0266/000191DNA metabolism, modification and replicationStreptococcus agalactiaee − 34
27241Isochorismatase (EC 3.3.2.1)−/000868Ubiquinone biosynthesisBacillus cereus (strain atcc 14579)2e − 57
rRNA5S ribosomal RNArRNA  
rRNA23S ribosomal RNArRNA  
tRNAtRNA-AlatRNA  
rRNA16S ribosomal RNArRNA  
30138Conserved hypothetical proteinArchaeoglobus fulgidus2 e − 17
31120Predicted acetyltransferase (RimI) (EC 2.3.1.-)0456/−Fatty acid metabolismMethanocaldococcus jannaschii2 e − 11
3291Conserved hypothetical proteinMethanosarcina acetivorans C2Ae − 11
33481DnaG family protein (DnaG) (EC 2.7.7.-)d0358/−RNA metabolismMethanosarcina mazei Goe1e − 119
34116Conserved hypothetical proteinMethanosarcina mazei Goe12 e − 20
35104Conserved hypothetical proteinMethanosarcina mazei Goe13 e − 22
Table 3.  Predicted RNA and protein-encoding genes of fosmid insert 1B6mr2
ORFProtein size (AA)Predicted function (gene) (EC)COGa/ IPRbPathwayBest BLAST hite valuec
  1. Fourteen hypothetical proteins without any similarity to BLAST entries were excluded.

  2. aCluster of orthologous groups (COG) of proteins, i.e., the proteins that comprise each COG are assumed to have evolved from an ancestral protein, and are therefore homologs.

  3. bInterPro entry (InterPro is a database of protein families, domains and functional sites in which identifiable features found in known proteins can be applied to unknown protein sequences.)

  4. cExpect value: probability that the associated match is due to randomness; the lower the e value, the more significant the match.

3780Anaerobic ribonucleoside-triphosphate reductase (NrdD) (EC 1.17.4.2)1328/−Purine metabolismMethanothermobacter thermautotrophicusΔH0.0
4262Pyruvate-formate lyase-activating enzyme (PflA)1180/−Pyruvate and acetyl-CoA metabolism, purine metabolismMethanosarcina mazei Goe17e − 42
7316Probable dephospho-kinase (CoaE)UnclassifiedArchaeoglobus fulgiduse − 72
8259Predicted sugar phosphate isomerase/epimerase (IolE)1082/−Gluconeogenesis (glycolysis)Methanosarcina acetivorans C2A4e − 60
9397Permease of the major facilitator superfamily similar to fosmidomycin resistance protein (YfnC)0477/005829Drug and analog sensitivityBacillus subtilise − 34
10406Predicted fosmidomycin resistance protein−/007114Drug and analog sensitivityDeinococcus radiodurans4e − 34
12164Conserved hypothetical proteinHalobacterium sp. (strain nrc-1)2e − 22
13312Metal-dependent hydrolase of β-lactamase superfamily III (ElaC)1234/−UnclassifiedMethanosarcina mazei Goe1e − 97
17344Aspartate-semialdehyde dehydrogenase (Asd) (EC 1.2.1.11)−/000319Amino acid metabolism (lysine)Methanosarcina mazei Goe1e − 134
18277Conserved hypothetical proteinMethanothermobacter thermautotrophicusΔH8e − 71
19384Predicted permeases of the major facilitator superfamily0477/007114UnclassifiedBacillus haloduranse − 05
20330Predicted lipoate synthase0320/−Cofactor metabolismMethanosarcina mazei Goe12 e − 88
23391Thiamine biosynthesis protein (ATP pyrophosphatase) (ThiI)0301/003720Cofactor metabolismMethanosarcina mazei Goe1e − 110
25257Dihydrodipicolinate reductase (DapB) (EC 1.3.1.26)0289/000846Amino acid metabolism (lysine)Methanosarcina mazei Goe12e − 70
26319Dihydrodipicolinate synthase (DapA) (EC 4.2.1.52)0329/002220Amino acid metabolism (lysine)Methanosarcina mazei Goe13e − 79
276530S ribosomal protein S17E1383 / 001210Ribosomal proteinPyrococcus abyssie − 08
28478Partial threonyl-tRNA synthetase (HrrS) (EC 6.1.1.3)−/002320Aminoacyl tRNA synthetasePyrococcus furiosus4 e − 93
29286Glycerate dehydrogenase (PgdH) (EC 1.1.1.95)−/006140Glyoxylate and dicarboxylate metabolismMethanosarcina acetivorans C2A4e − 67
rRNA5S ribosomal RNArRNA  
rRNA23S ribosomal RNArRNA  
tRNAtRNA-AlatRNA  
rRNA16S ribosomal RNArRNA  
31234Predicted nucleotidyl transferase (EC 2.7.7.-)2413/−UnclassifiedArchaeoglobus fulgidus2e − 49
32104Uncharacterized protein involved in tolerance to divalent cations (CutA1)1324/004323UnclassifiedMethanothermobacter thermautotrophicusΔHe − 20
33565Membrane metalloprotease (metallopeptidase M50)0750/008915Protein modification and degradationMethanosarcina mazei Goe17e − 89
35282Predicted flavodoxin (domain: flavodoxin/nitric oxide synthase)−/001226UnclassifiedBacteroides thetaiotaomicron3e − 08
36294Hypothetical phosphoribosylaminoribosylaminoimidazole succinocarboxamide synthetase (SAICAR)0152/001636UnclassifiedXenopus laevis4e − 28
37178Hypothetical protein (probable glutelin, multicopper oxidase type 1)−/000480Unclassified
39589Pyruvate carboxylase subunit B (PycB) (EC 6.4.1.1)−/000891Pyruvate and acetyl-CoA metabolismMethanosarcina mazei Goe10.0
40722Pyruvate carboxylase subunit A (PycA) (EC 6.4.1.1)−/004549Pyruvate and acetyl-CoA metabolismMethanothermobacter thermautotrophicusΔHe − 161
41185Signal sequence peptidase0681/−Protein and peptide secretionMethanosarcina acetivorans C2A7e − 37
Table 4.  Predicted RNA and protein-encoding genes of fosmid insert 3D10mr4
ORFProtein size (AA)Predicted function (gene) (EC)COGa/IPRbPathwayBest BLAST hite valuec
  1. Fifteen hypothetical proteins without any similarity to BLAST entries were excluded.

  2. aCluster of orthologous groups (COG) of proteins, i.e., the proteins that comprise each COG are assumed to have evolved from an ancestral protein, and are therefore homologs.

  3. bInterPro entry (InterPro is a database of protein families, domains and functional sites in which identifiable features found in known proteins can be applied to unknown protein sequences.)

  4. cExpect value: probability that the associated match is due to randomness; the lower the e value, the more significant the match.

  5. dORFs predicted to represent informational genes that possess a strong phylogenetic signal (Table 5).

RRNA5S ribosomal RNArRNA  
rRNA23S ribosomal RNArRNA  
rRNA16S ribosomal RNArRNA  
1583Pyruvate kinase (Pyk) (EC 2.7.1.40)−/001697Pyruvate and acetyl-CoA metabolismThermoanaerobacter tengcongensise − 151
2312dTDP-glucose 4,6-dehydratase (CpsN) (EC 4.2.1.46)−/001509Nucleotide sugar metabolismMethanosarcina acetivorans C2Ae − 114
3259Serine/threonine protein kinase1718/−Amino acid metabolism (alanine, aspartate and glutamate)Methanosarcina acetivorans C2A2e − 67
4181RNA-binding proteind1094/004087Transcription and RNA processingMethanosarcina mazei Goe13e − 46
62543-isopropylmalate dehydratase (EC 4.2.1.33)Amino acid metabolism (valine, leucine and isoleucine)Methanosarcina acetivorans C2A7e − 68
7267Translation initiation factor 2, subunit alpha (eIF2a)d1093/−Translation factorMethanosarcina acetivorans C2A2e − 82
89650S ribosomal protein L44E1631/−Ribosomal proteinMethanosarcina mazei Goe17e − 25
10355Eukaryotic-like DNA primase, small subunit (PriA) (EC 2.7.7.-)d−/002755DNA metabolism, modification and replicationMethanosarcina mazei Goe18e − 76
14814Valyl-tRNA synthetase (ValS) (EC 6.1.1.9)−/002303Aminoacyl tRNA synthetaseThermoplasma volcanium0
15264Hypothetical transcriptional regulator1378/−Transcription and RNA processingPyrococcus furiosus4e − 18
16506Predicted permease of the major facilitator superfamily0477 / 007114UnclassifiedClostridium perfringens5e − 50
18582Glutamyl-tRNA synthetase (GlnS) (EC 6.1.1.17)0008/004526Aminoacyl tRNA synthetaseMethanosarcina acetivorans C2Ae − 175
211075ATP-dependent DNA helicase (PcrA) (EC 3.6.1.-)d−/000212DNA metabolism, modification and replicationMethanosarcina acetivorans C2Ae − 100
24186Hypothetical protein (similar to aldehyde dehydrogenase)−/002086Gluconeogenesis (glycolysis) Pyruvate and acetyl-CoA metabolism
26399Conserved hypothetical proteinClostridium tetanie − 45
27316DNA adenine methylase (Dam) (EC 2.1.1.72)0338/002294DNA metabolism, modification and replicationGloeobacter violaceus pcc 74215e − 89
30102Hypothetical protein (similar to zinc carboxypeptidase A (Metalloprotease M14])−/000834Protein modification and degradation

Our further studies focused on two aspects: (i) characterization of the rrn operon variants and (ii) the phylogeny of informational genes predicted to encode proteins that are involved in essential functions of the cell machinery [38].

3.2.4rrn operon variants

Comparative sequence analysis of the fosmid clones 4B7mr3, 1B6mr2, and 3D10mr4 showed that (i) the gene arrangement within the rrn operons followed the order 16S–23S–5S and (ii) the 23S and 5S rRNA gene sequences are identical among the three rrn operon variants (Fig. 5). While the rrn operon variants represented by the clones 4B7mr3 and 1B6mr2 were almost identical (G-type vs. T-type of the 16S rRNA gene sequence), the third variant (clone 3D10mr4) differed by the absence of a tRNA-Ala gene and, in addition, by sequence differences in regions that flank the rRNA genes.

3.2.5Informational genes

Genes that are predicted to encode components of the cellular information processing machinery, e.g. DNA replication/modification and transcription/translation, are termed informational genes [38]. Seven ORFs of RC-I were predicted to represent such informational genes. All of them showed significant similarities to predicted proteins of Methanosarcina spp. (Table 5). The deduced amino acid sequences of three informational genes, including eIF2a, dnaG and priA, were chosen for phylogenetic reconstructions. In each of the protein trees, RC-I exhibited the same phylogenetic affiliations: (i) It showed a specific association to the Methanosarcinales and (ii) belonged to a monophyletic supergroup characterized by the Methanosarcinales, Archaeoglobales, Halobacteriales, and Thermoplasmatales[39,40] (Fig. 6).

Table 5.  Affinities of informational genes predicted for RC-I to Methanosarcina spp.
Predicted function (gene]Fosmid (ORF]Best BLAST hit (UniProt/TrEMBL entry]Identity (%)Similarity (%)
Translation initiation factor 2 subunit alpha (eIF2a]3D10mr4 [7]M. acetivorans C2A (Q8TSZ9]57.971.8
DnaG family protein (dnaG]4B7mr3 [33]M. mazei Goe1 (Q8PXC6]51.867.6
Eukaryotic-like DNA primase small subunit (priA]3D10mr4 [10]M. mazei Goe1 (Q8PVZ4]41.360.3
ATP-dependent DNA helicase (pcrA]3D10mr4 [21]M. acetivorans C2A (Q8TND2]37.359.5
Glutamyl-tRNA amidotransferase subunit D (gatD]4B7mr3 [23]M. mazei Goe1 (Q8PUM7]57.072.8
Glutamyl-tRNA amidotransferase subunit E (gatE]4B7mr3 [24]M. acetivorans C2A (Q8TM08]57.673.9
RNA-binding protein3D10mr4 [4]M. mazei Goe1 (Q8PVF3]52.274.5
Figure 6.

Unrooted maximum likelihood dendrograms showing the phylogenies of (a) translation initiation factor 2 subunit alpha (eIF2a), (b) DnaG family protein (DnaG, EC 2.7.7.-), and (c) eukaryotic-like DNA primase small subunit (PriA, EC 2.7.7.-). The positions of RC-I sequence types are shown in relation to the positions of a representative selection of Euryarchaeota. Accession numbers (GenBank, EMBL, and DDBJ entries) of the reference sequences are given in parentheses. The scale bars represent a sequence divergence of 10%.

4Discussion

4.1Growth characteristics of culture MRE50

During growth of MRE50, methane was produced in stoichiometries that were highly indicative of hydrogenotrophic methanogenesis, i.e. 4H2+ CO2 CH4+ 2H2O. The growth yield of RC-I methanogens was 3 × 107 16S rRNA targets per 6.5 μmol CH4 produced (Fig. 1). Assumption of three rrn operon copies per cell (see discussion below) and a specific dry weight (dw) of 0.5 × 10−12 g cell−1 leads to a growth yield of approximately 0.8 g dw mol−1 CH4. This value compares well with the growth yields of pure cultures of methanogens under similar growth conditions, being in the order of about 0.6 g dw mol−1 CH4[41]. The finding that RC-I methanogens in culture MRE50 are hydrogenotrophic methanogens agrees well with two previous studies on RC-I, which showed that methane was formed only from H2/CO2. In the first, Fey et al. [42] observed that, during the anaerobic incubation of paddy soil slurries, RC-I populations exhibit an optimum in methane production at temperatures around 50 °C. This finding allowed us to enrich members of RC-I selectively from an indigenous methanogenic community that is otherwise predominated by mesophilic populations of the Methanosarcinales and Methanobacteriales[43]. In the second study, RC-I representatives were enriched from peat bog under acidic conditions at 28 °C (consortium K, [10]). Comparative analysis of 16S rRNA gene sequences showed that consortium K is a mixed population that consisted, besides a bacterial component, of RC-I as the major methanogenic group (98% of total archaeal cells, clone K-5a2, Fig. 3(a)) and members of the Methanomicrobiales. The sequence of clone K-5a2 exhibited 93% identity to the RC-I sequence types retrieved from culture MRE50.

After transfer of culture MRE50 to fresh medium, an immediate growth response of bacteria was observed, as indicated by the rapid increase of the bacterial 16S rRNA gene target numbers (Fig. 1(b)) and by the concomitant accumulation of H2 and CO2 during the first two days of incubation (Fig. 1(a)). The rapid growth response suggests that the bacterial populations contain multiple rrn operon copies [44]. As a consequence, the number of bacterial cells might be 5–10 times lower than the number of 16S rRNA gene targets measured by real-time PCR. In correspondence to the temperature used for enrichment, members of the family Thermoanaerobacteriaceae (Fig. 2) were identified to be prevalent in culture MRE50. Cultured representatives of this taxon are strict anaerobes and thermophiles [45,46]. These bacteria probably grew on the expense of yeast extract that had been included into the medium. Alternatively, some of them may have grown chemolithoautotrophically on H2/CO2, e.g., homoacetogenic Moorella spp. [37]. The fact that only small amounts of acetate accumulated in culture MRE50 might be attributed to its consumption by the methanogens (i.e., assimilation as carbon source) and/or the bacterial component (e.g., syntrophic oxidation). This assumption corresponds well to the observation that the addition of small amounts of acetate to culture MRE50 stimulated overall cell growth. During the later stages of incubation of culture MRE50, the bacteria apparently lysed and may have served as substrate for consortium MRE50. It is presently unclear which physiological role the bacterial community plays in co-culture with the RC-I methanogens. However, all attempts to isolate RC-I methanogens in pure culture by serial dilutions in both liquid culture and roll-tubes failed. These attempts also included the use of lysozyme and various antibiotics to eliminate the bacterial component. The antibiotics, including ampicillin, streptomycin, vancomycin, and chloramphenicol, were used separately or in combination.

The bacterial composition of culture MRE50 was clearly different from that of consortium K, which might reflect the different incubation temperatures used for enrichment [10]. The bacterial component of consortium K made up 5–10% of total cells and was composed of uncharacterized members of the Alpha- and Deltaproteobacteria, Acidobacteria, and green non-sulfur bacteria. Several attempts were made to isolate RC-I representatives from consortium K using Hungate techniques in combination with serial dilutions on basic +N and −N media, kanamycin treatment to suppress the bacterial component and plating into tubes or vials with agar or agarose. However, as we observed for culture MRE50, RC-I methanogens failed to grow separately from bacterial satellites.

4.2Enrichment status of RC-I

Comparative sequence analyses of both the 16S rRNA and mcrA genes provided evidence that the archaeal component of culture MRE50 is composed only of RC-I. This view is also supported by the detection of only a single 392-bp T-RF that was highly indicative of RC-I 16S rRNA genes. Assuming that RC-I is the only archaeal component of consortium MRE50, the RC-I population amounts to more than 50% of total cells, as concluded from the FISH data. The RC-I population exhibited identity values above 97% with 16S rRNA gene clone sequences that had been retrieved from rice paddy soils in Italy (clones AS08-19 and ARR24) and Japan (clone pST-6), and from swampy marshland in Florida, USA (clone 384/B-2) (Fig. 3(a)). This finding suggests that the RC-I population studied here is environmentally significant.

A previous diversity assessment of consortium MRE50 had revealed an archaeal population phylogenetically more diverse [15] than the present study, including members of the Methanobacteriaceae. The RC-I 16S rRNA gene sequence types retrieved from culture MRE50 in the previous study were only moderately related to those recovered in this study. The highest identity values observed between the two 16S rRNA gene sequence types obtained in this study and a representative of the previous study (clone MRE11) were found to be 94–95% (Fig. 3(a)). At the mcrA level, the highest identity value was found to be 83% (clone MRE50mcrA vs. clone MRE-ME5, Fig. 3(b)). These findings clearly indicate that, while the methanogenic activity of culture MRE50 was stably maintained, a major shift in the composition of the archaeal component has occurred over the last years.

4.3RC-I intracluster diversity

The high enrichment status of a phylogenetically rather homogeneous RC-I population prompted us to investigate the RC-I genospecies diversity in more detail. The ITS region has been suggested to be an appropriate marker for studying microdiversity and sequence heterogeneity at the genospecies level [47]. PCR-based retrieval of this marker from the fosmid library followed by comparative sequence analysis showed that the RC-I population was characterized by three different rrn operon variants.

The detection of three different rrn operon variants among 32 test-positive fosmid clones raised the question of whether they are indicative of the occurrence of two or three closely related but distinct genotypes or whether they represent multiple copies within a single RC-I genotype. Based on the current knowledge of the presence of multiple rrn operons in methanogens, we favour the assumption of a single RC-I genotype for the following reasons:

  • (i)Multiple rrn operons are found in the majority of methanogens studied so far, including Methanocaldococcus jannaschii DSM 2661 and Methanothermobacter thermautotrophicusΔH (two copies), Methanosarcina mazei Goe1 and Methanosarcina acetivorans C2A (three copies), and Methanococcus vannielii EY33 (four copies) [48]. However, a known exception is Methanopyrus kandleri. Its genome harbours only single 16S, 23S, and 5S rRNA genes. These genes are unlinked, while in most Euryarchaeota the genes encoding ribosomal RNAs are linked in the order 16S–23S–5S [49].
  • (ii)Sequence and structural heterogeneities are observed among different rrn operons of a single genome. For example, one of the three rrn operon copies in the Methanosarcina strains Goe1 and C2A contains a tRNA-Ala gene in the ITS region, while the other two copies do not [50].
  • (iii)The gene content in the regions flanking the three different rrn operons in the genome of M. mazei Goe1 is highly diverse. This holds true for the regions flanking the three different rrn operons in the genome of M. acetivorans C2A. However, both gene content and gene organization of the regions that flank the homologous rrn operon copies in the genomes of strains Goe1 and C2A are highly conserved (ERGO™ bioinformatics suite; http://ergo.integratedgenomics.com/ERGO). No conservation could be observed among the three RC-I fosmid inserts for the total gene content apart from the rrn operons.
  • (iv)All methanogen genomes studied so far are characterized by the presence of only a single mcrA gene. Based on this finding, the target numbers of RC-I 16S rRNA genes in relation to those determined for mcrA suggest that the RC-I genome studied here contains three rrn operon copies.

4.4Phylogenetic considerations

Prior to the era of comparative genomics, the evolution of microorganisms has been extensively analyzed only by single gene phylogenies. However, phylogenies constructed for multiple sets of genes and deduced protein sequences, or even complete genomes give reason to conclude that single gene phylogenies often do not reflect the organismal evolution but reflect only the phylogeny of the respective gene itself. Horizontal gene transfer was considered to be the major driving force for substantial gene exchanges even between domains. However, various findings suggested the presence of an evolutionary stable core of mainly informational genes that were thought to be rarely affected by horizontal gene transfer [51]. Consequently, these genes were considered to be appropriate candidates for reconstruction of the organismal evolution [52]. More recently, the low frequency of informational genes to be involved in horizontal gene transfer was an issue of controversy [53,54]. Daubin et al. [54] proposed that genes should be classified into two major groups, designated acquired and ortholog classes. Acquired genes are frequently affected by horizontal gene transfer and mostly encode proteins of unknown function. Orthologous genes are single copy genes that mostly encode proteins of known function. The ortholog class was shown to constitute only 10–15% of those genes previously characterized as informational genes. Orthologous genes were further differentiated, based on the presence or absence of a sufficient phylogenetic signal.

Among the 41 ORFs of RC-I to which putative functions could be assigned, seven ORFs were predicted to encode components of the cellular information processing machinery, and thus were classified as informational genes (Table 5). The deduced amino acid sequences of the seven genes all showed highest similarities to putative proteins of M. mazei Goe1 and M. acetivorans C2A. The seven genes also belong to the ortholog class of genes, as evidenced by comparative genomics. The deduced protein sequences of three genes (eIF2a, dnaG, and priA) were used to reconstruct the phylogeny of RC-I (Fig. 6). The tree topologies provide further support that RC-I has a specific phylogenetic association with members of the Methanosarcinales, as already suggested by the trees reconstructed for the 16S rRNA gene and McrA (Fig. 3).

5Perspectives

The results obtained for the enrichment status and intracluster diversity of RC-I showed that genome fragments of only a single or a very few closely related RC-I genotypes are deposited in the fosmid library. Recently, reconstructions of nearly complete genomes of as-yet uncultured microorganisms from Sargasso Sea [55] and biofilms [56] were reported. The assembling of these environmental genomes was achieved by shotgun sequencing of small insert clone libraries. Based on these reports, we anticipate that the reconstruction of a major portion or even a complete RC-I genome should be possible. However, the approach would have to focus on a directed screening for overlapping RC-I genome fragments within the large insert fosmid library of consortium MRE50. This approach seems to be feasible taking into consideration that all hydrogenotrophic methanogens studied so far are characterized by genome sizes of only 1.6–1.8 Mb [57]. In contrast, Methanosarcina spp. exhibit genome sizes greater than 4 Mb. The differences in genome size are also reflected by coding region densities. These were determined for hydrogenotrophic methanogens to constitute 88–90% of total genomic information, while the corresponding values for the genomes of Methanosarcina spp. are clearly lower (74–75%). The gene density determined for the RC-I genome fragments 1B6mr2, 3D10mr4, and 4B7mr3 (86.3%) is consistent with those of hydrogenotrophic methanogens.

Acknowledgements

We thank Sonja Fleissner and Sabrina Patzak for excellent technical assistance. This work was supported by the Competence Network Göttingen “Genome research on bacteria” (GenoMik) financed by the German Federal Ministry of Education and Research (contract 031U213A).

Ancillary