Non-Sulfate-Reducing, Syntrophic Bacteria Affiliated with Desulfotomaculum Cluster I Are Widely Distributed in Methanogenic Environments

Abstract

The classical perception of members of the gram-positive Desulfotomaculum cluster I as sulfate-reducing bacteria was recently challenged by the isolation of new representatives lacking the ability for anaerobic sulfate respiration. For example, the two described syntrophic propionate-oxidizing species of the genus Pelotomaculum form the novel Desulfotomaculum subcluster Ih. In the present study, we applied a polyphasic approach by using cultivation-independent and culturing techniques in order to further characterize the occurrence, abundance, and physiological properties of subcluster Ih bacteria in low-sulfate, methanogenic environments. 16S rRNA (gene)-based cloning, quantitative fluorescence in situ hybridization, and real-time PCR analyses showed that the subcluster Ih population composed a considerable part of the Desulfotomaculum cluster I community in almost all samples examined. Additionally, five propionate-degrading syntrophic enrichments of subcluster Ih bacteria were successfully established, from one of which the new strain MGP was isolated in coculture with a hydrogenotrophic methanogen. None of the cultures analyzed, including previously described Pelotomaculum species and strain MGP, consumed sulfite, sulfate, or organosulfonates. In accordance with these phenotypic observations, a PCR-based screening for dsrAB (key genes of the sulfate respiration pathway encoding the alpha and beta subunits of the dissimilatory sulfite reductase) of all enrichments/(co)cultures was negative with one exception. Surprisingly, strain MGP contained dsrAB, which were transcribed in the presence and absence of sulfate. Based on these and previous findings, we hypothesize that members of Desulfotomaculum subcluster Ih have recently adopted a syntrophic lifestyle to thrive in low-sulfate, methanogenic environments and thus have lost their ancestral ability for dissimilatory sulfate/sulfite reduction.

Members of Desulfotomaculum cluster I have been generally known as gram-positive, spore-forming, sulfate-reducing bacteria (50, 56) and have frequently been found in various anoxic environments, such as sediments, rice paddy soil, human feces, and anaerobic sludges (for examples, see references 21, 42, 56, 57). The genus Desulfotomaculum includes over 20 validly described mesophilic and thermophilic species, all of which share the ability to oxidize various organic substances, like short-chain fatty acids, alcohols, and aromatic compounds with sulfate as a terminal electron acceptor (56). Due to these physiological traits, members of Desulfotomaculum cluster I have been considered important in sulfidogenic, anoxic environments, where they play crucial roles in the degradation of organic compounds, as well as in the biogeochemical cycling of sulfur (56).

On the basis of comparative 16S rRNA gene sequence analysis, Desulfotomaculum cluster I was considered to be comprised of seven well-separated subclusters, Ia to Ig (27, 50, 52, 57). Although it has been argued that each subcluster could be treated taxonomically as an individual genus, all subcluster members were until recently described as Desulfotomaculum species due to the absence of sufficient distinguishing phenotypic and molecular features (50). However, the two recently described species of the new genus Sporotomaculum lack the ability to respire anaerobically with sulfate, although they are members of Desulfotomaculum subcluster Ib (3, 44). In addition, Pelotomaculum thermopropionicum strain SI, representing the novel subcluster Ih within Desulfotomaculum cluster I (22, 23), is also lacking dissimilatory sulfate-reducing activity (22, 23). Importantly, strain SI grows syntrophically with hydrogenotrophic methanogens (22, 23), whereas only a few members of the genus Desulfotomaculum are able to grow in syntrophic association with methanogens in the absence of sulfate (26, 42). The collection of non-sulfate-reducing bacteria belonging to Desulfotomaculum subcluster Ih was recently extended by the successful isolation of further syntrophic strains in coculture with hydrogenotrophic methanogens: the propionate degrader Pelotomaculum schinkii (6, 7, 16) and the two phthalate-degrading strains JT and JI (42a, 43). Furthermore, Lueders et al. (37) reported that Pelotomaculum spp. (i.e., subcluster Ih microbes) were one of the active in situ syntrophic propionate-oxidizing populations in a rice field by rRNA-based stable-isotope probing.

The degradation of organic matter by Desulfotomaculum cluster I bacteria in methanogenic environments has been thought to be linked to dissimilatory sulfate reduction (21, 52). Based on the results of this study and the previous studies mentioned above, we propose that subcluster Ih bacteria have adapted to anoxic, low-sulfate conditions and thus constitute a significant fraction of the Desulfotomaculum cluster I population in methanogenic ecosystems. As a consequence of this evolutionary process, they have lost the capability for sulfate respiration and live in close proximity to hydrogen- and formate-consuming methanogens in order to maintain an energetically favorable, low-hydrogen partial pressure for the syntrophic oxidation of organic substrates.

MATERIALS AND METHODS

Microorganisms and cultivation media.

The following pure (co)cultures were used in this study. Pelotomaculum thermopropionicum strain SI was isolated in our previous study (22, 23). Strain MGP was isolated in this study as an anaerobic, mesophilic, syntrophic propionate-oxidizing bacterium. Methanosaeta thermophila strain P^T (DSM 6194), Methanothermobacter thermautotrophicus strain ΔH (DSM 1053), Methanospirillum hungatei strain JF1 (DSM 864), Methanosaeta concilii strain GP6 (DSM 3671), Methanobacterium bryantii strain M.o.H (DSM 863), Syntrophobacter fumaroxidans strain MPOB (DSM 10017), Desulfotomaculum thermobenzoicum strain TSB (DSM 6193), Desulfotomaculum thermosapovorans strain MLF (DSM 6562), and Desulfotomaculum thermocisternum strain ST90 (DSM 10259) were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany). Moorella thermoacetica (JCM 9319) was obtained from the Japan Collection of Microorganisms (Wako, Japan). Eschericha coli strain JM109 was purchased from TaKaRa Bio Inc. (Otsu, Shiga, Japan). Pelotomaculum schinkii strain HH was kindly provided by Frank de Bok (Wageningen University, Wageningen, The Netherlands). In addition, we also used the following microorganisms isolated in our laboratory: an anaerobic, syntrophic terephthalate-degrading bacterium, strain JT (43); an anaerobic, syntrophic isophtalate-degrading bacterium, strain JI (Qiu et al., unpublished); and a hydrogen- and formate-utilizing methanogen, Methanothermobacter thermautotrophicus strain type II. The culture media used for the enrichment and cultivation of the above-mentioned reference microorganisms were prepared as described previously (47). All cultivations were carried out at either 55°C or 37°C in 50-ml serum vials containing 20 ml of medium (pH at 25°C, 7.2) under an atmosphere of N₂-CO₂ (80/20 [vol/vol]) without shaking. The purity of strain MGP was routinely examined by microscopy and incubation of the cultures with medium containing 10 mM sucrose, 10 mM glucose, 10 mM lactate, and 10 mM pyruvate plus 0.1% yeast extract, thioglycolate medium, and AC medium (Difco) at 37 and 55°C.

Environmental samples.

Eight anaerobic-sludge samples and one sample from a rice paddy soil were collected and analyzed in this study (Table 1). The operational conditions for two laboratory-scale upflow anaerobic-sludge blanket reactors treating artificial wastewater (environmental TR2 and MR1) were described previously (46, 48). All the engineered and environmental samples displayed methanogenic activity.

TABLE 1.

Samples from methanogenic environments analyzed in this study

Sample no.	Sample	Reactor operation temp (°C)	Treating waste/wastewater	Sulfate concn in influent (mg/liter)	Influent CODcra (mg/liter)	Location
1	Thermophilic digested sludge TJ	55	Municipal solid waste	20	93,550	Niigata, Japan
2	Thermophilic digested sludge TO	53	Municipal sewage sludge	56	59,000	Osaka, Japan
3	Thermophilic granular sludge TR2	55	Artificial wastewater	157	4,000	Our laboratory
4	Thermophilic granular sludge TD	55	Wastewater from clear liquor manufacture	149	12,000	Our laboratory
5	Mesophilic granular sludge MR1	35	Artificial wastewater	81	2,000	Our laboratory
6	Mesophilic granular sludge MB	35	Wastewater from beer factory	55	3,000	Ibaraki, Japan
7	Mesophilic granular sludge MC	35	Wastewater from sugar factory	58	1,635	Nara, Japan
8	Mesophilic digested sludge MD	35	Municipal sewage sludge	62	57,600	Niigata, Japan
9	Rice paddy soil RP					Niigata, Japan

^aChemical oxygen demand as determined using dichromate.

Construction of 16S rRNA gene clone libraries.

Extraction of DNA from anaerobic environmental samples and enrichment cultures was done according to a previously described protocol (38). Before PCR amplification, the extracted DNA was purified with a GENE CLEAN Turbo kit (BIO101, CA). Amplification of 16S rRNA genes was carried out according to the method in a previous study (48). The Desulfotomaculum cluster I-specific primer pair DEM116f/DEM1164r (52) and the universal bacterial primer pair 341F/1490R (40, 55) were used for amplification of 16S rRNA genes from environmental samples and enrichments, respectively (52). The PCR products were purified with a MicroSpin Column (Amersham Biosciences) and subsequently cloned into E. coli using the TA cloning kit (Novagen). Ten clones were randomly picked from each clone library for sequencing. 16S rRNA gene sequences with a sequence similarity of 100% were grouped into a “phylotype.” The name of each phylotype is composed of the sample name and a number (for example, TJ-1 is clone number 1 recovered from the environmental sample TJ).

PCR amplification of dsrAB and apsA.

DNA extractions from pure (co)cultures (P. thermopropionicum, P. schinkii, S. fumaroxidans, and strains JT, JI, and MGP), as well as from enrichment cultures containing subcluster Ih bacteria, were performed according to the method of Hiraishi (20). The template quality of extracted DNA was tested by PCR with the 16S rRNA gene-specific primer pair 341F (40) and 907R (28). PCR amplifications of the dsrAB and apsA genes were performed as reported previously (9, 14, 25). Approximately 40 ng of template DNA was used in a 100-μl PCR mixture; the concentration of template DNA was measured by using the PicoGreen double-stranded DNA quantification kit (Molecular Probes) and a spectrofluorophotometer (RF-5300PC; Shimadzu). An approximately 1,900-bp fragment of dsrAB was amplified by using the primers DSR1Fdeg and DSR4Rdeg (25). The primer pairs APS-FW/APS-RV (9) and APS7-F (and its derivatives)/APS8-R (14) were used for amplification of approximately 390- and 900-bp fragments of apsA, respectively. Genomic DNA from S. fumaroxidans was used as the positive control in all dsrAB- and apsA-targeted PCR experiments (14). PCR products of dsrAB and apsA were purified with a MicroSpin Column (Amersham Biosciences) and either directly sequenced or cloned in E. coli with the TA cloning kit (Novagen) prior to being sequenced.

Sequencing and phylogenetic analysis.

The complete 16S rRNA gene sequence of strain MGP isolated in this study was determined as described previously (23). The sequences of the 16S rRNA, dsrAB, and apsA genes were determined by dye terminator cycle sequencing with a Quick Start kit (Beckman Coulter) and an automated sequence analyzer (CEQ2000XL; Beckman Coulter). If not otherwise indicated, phylogenetic analyses were performed with the ARB program (36). A 16S rRNAgene-based tree was constructed by applying the neighbor-joining method (45) for sequences having >1,000 nucleotides. Subsequently, shorter sequences were inserted into this tree without changing the tree topology by using the parsimony insertion tool of the ARB program. Bootstrap analysis (1,000 replicates) (12) was performed with the PAUP* 4.0 program package (53) to estimate the confidence of the 16S rRNA gene tree topology. New dsrAB and apsA sequences were aligned and phylogenetically analyzed as described previously (14, 35).

FISH.

The following 16S rRNA-targeted oligonucleotide probes were used in this study: TGP690 for P. thermopropionicum strain SI (23); EUB338, targeting most Bacteria (1); and DEM1164r for almost all members of Desulfotomaculum cluster I (52). A novel 16S rRNA-targeted oligonucleotide probe, Ih820 (5′-ACCTCCTACACCTAGCAC-3′; 820 to 837 in E. coli positions), targeting members of Desulfotomaculum subcluster Ih, was designed by using the ARB program. Probe Ih820 was applied with an unlabeled competitor probe (5′-ACCTCCTACACCTAGTAC-3′; 820 to 837 in E. coli positions) in order to avoid cross-hybridization with nontarget microorganisms. Further details of the cyanine 3.18 (Cy3)-labeled probes used in this study are available at probeBase (34). Fixation of cells in the environmental samples and enrichment cultures and whole-cell in situ hybridization were performed as described previously (45). Granular-sludge samples TR2, TD, MR1, MB, and MC were dispersed by a manually operated homogenizer and an ultrasonic homogenizer prior to fluorescence in situ hybridization (FISH). The stringency of hybridization was adjusted by adding formamide to the hybridization buffer (15% [vol/vol] for TGP690, 20% for EUB338, 10% for DEM1164r, and 20% for Ih820). Enumeration of probe-stained cells was performed with an epifluorescence microscope (Olympus BX50F) as described previously (22).

Real-time PCR quantification.

Quantitative PCR was performed with a LightCycler device (Roche, Mannheim, Germany) by using the TaKaRa Ex Taq R-PCR version (Takara Bio Inc., Otsu, Japan) with SYBR Green I (BioWhittaker Molecular Applications, Rockland, ME). A reaction mixture for PCR was prepared according to the manufacturer's instructions, including approximately 0.1 or 1 ng of template DNA and 0.0006% (vol/vol) SYBR Green I solution. The following 16S rRNA gene-targeted PCR primer sets were used: 519f and 907r (28, 51) for the domain Bacteria, Ar109f (5′-AMDGCTCAGTAACACGT-3′; a slightly modified version of a primer reported by Groβkopf et al. [15]) and Ar912r (15) for the domain Archaea, DEM116f and DEM1164r for members of Desulfotomaculum cluster I (51, 52), and DEM116f and Ih820 for Desulfotomaculum subcluster Ih bacteria. For the construction of template standards for each primer set, we used dilution series of 16S rRNA gene PCR products of P. thermopropionicum, E. coli, and M. bryantii (obtained with the bacterial primer pair 8F/1490R [55] or the archaeal primer pair Arch21F/1490R [8]). These PCR products were used in each real-time PCR analysis to calculate the copy number of 16S rRNA genes in the sample of interest. Template DNA was purified with a GENE CLEAN Turbo kit (BIO101, CA) and quantified spectrofluorometrically by using the PicoGreen double-stranded DNA Quantification kit (Molecular Probes). The optimal PCR conditions were determined by changing the Mg²⁺ concentration in the PCR buffer and the PCR annealing temperature for each primer set (Table 2). The PCR conditions were as follows: initial denaturation at 95°C for 2 min, followed by 40 cycles of denaturation at 95°C for 0 s, 10 s annealing (temperatures are shown in Table 2), and extension at 72°C for 45 s, 35 s, and 30 s for primer pairs DEM116f/DEM1164r and Ar109f/Ar912r and the other primers, respectively. Two tests were performed to confirm the specificity of the real-time PCR assays. First, melting-curve analysis was performed after each amplification step. Second, the sizes and the identities of PCR products were confirmed by gel electrophoresis and subsequent clone library analysis.

TABLE 2.

16S rRNA gene-targeted primers used for quantitative real-time PCR

Primer set	Target	Positive control	Fragment length (bp)	Mg2+ concn (mM)	Annealing temp (°C)
DEM116f-DEM1164r	Desulfotomaculum cluster I	P. thermopropionicum	1,054	3	60
DEM116f-Ih820	Desulfotomaculum subcluster Ih	P. thermopropionicum	706	3	65
519f-907r	Bacteria	E. coli	407	4	50
Ar109f-Ar912r	Archaea	M. bryantii	788	3	52

Expression analysis of dsrAB by RT-PCR.

For reverse transcription (RT)-PCR analysis, strain MGP was harvested under two different cultivation conditions, i.e., cells were incubated in the basal medium (20 mM propionate and 0.01% yeast extract) with and without sulfate (20 mM). Preparation of RNA from strain MGP was done according to the method in a previous report (41) with slight modification. The remaining DNA was digested with RNase-free DNase I (Promega). The absence of contaminating genomic DNA in the RNA extract was confirmed by PCR. RT-PCR was performed with a commercial RT-PCR kit (ReverTra Dash; TOYOBO) according to the manufacturer's instructions. Two primer sets were used for amplification of dsrAB from strain MGP. MGP1F (5′-ACGCACTGGAAACACG-3′) and MGP1R (5′-TATCAGCTATGTCGCAGT-3′) amplify an approximately 480-bp-long fragment covering the 3′ end of dsrA, the intergenic region, and the 5′ end of the dsrB gene. MGP9F (5′-CCAACCCGTACTTCTTCT-3′) and MGP3R (5′-GTTAGCGCATCCGGATAC-3′) amplify a 307-bp-long fragment of dsrB. Primers MGP1R and MGP3R were newly designed, while MGP1F and MGP9F are slightly modified versions of primers DSR1Fdeg and DSR9F, reported by Klein et al. (25) and Friedrich (14), respectively. PCR products obtained by the RT-PCR experiment were cloned with the TA cloning kit (Novagen) and sequenced.

Analytical methods.

Short-chain fatty acids, alcohols, methane, hydrogen, carbon dioxide, and other intermediate substances, such as succinate, fumarate, and lactate, were measured as described previously (22, 23). Terephthalate was determined by high-performance liquid chromatography using an Aspak DE413 column (Shodex; reverse phase) held at 50°C and a UV detector (Shimadzu) set at 230 nm. The mobile phase was a 40:60 (vol/vol) mixture of methanol and 1% aqueous acetic acid at a flow rate of 0.8 ml per min. Sulfate and sulfite were determined by ion chromatography (column, Shimadzu Shim-pack IC-A1; carrier, 25 mM phthalate hydrogen potassium; detector type, electrical conductivity detector).

Nucleotide sequence accession numbers.

Sequence data obtained in this study were deposited in the DDBJ/EMBL/GenBank databases under accession numbers AB154373 to AB154392.

RESULTS

Phylotype richness of Desulfotomaculum cluster I bacteria.

To extend our current knowledge of the diversity of bacteria of Desulfotomaculum cluster I in anaerobic environments, separate 16S rRNA gene clone libraries were established for nine methanogenic samples (Table 1) by using a cluster I-specific primer pair (52). Ten clones from each library were randomly selected, sequenced, and phylogenetically analyzed (Fig. 1). Most clones (76%) belonged to Desulfotomaculum subcluster Ih and were classified into 16 different phylotypes. The remaining 22 clones formed three phylotypes, which were affiliated with two other Desulfotomaculum I subclusters. In more detail, 10 clones each from anaerobic reactors treating brewery and sugar-manufacturing wastewater were grouped in phylotypes MB-1 and MC-1, respectively, both of which are most likely representatives of a novel Desulfotomaculum subcluster, Ii (Fig. 1). In addition, another phylotype, TJ-2, represented by two clones, was recovered from a thermophilic digested sludge and was affiliated with subcluster Ib. No clones belonged to lineages other than Desulfotomaculum cluster I.

FIG. 1.

Phylogenetic tree of Desulfotomaculum cluster I based on comparative analyses of 16S rRNA gene sequences and showing the phylogenetic positions of clones and strain MGP obtained in this study. Environmental clones obtained in this study are indicated in blue. Red indicates the isolated strain and clones from enrichment cultures established in this study. The initial tree was constructed with sequences greater than 1,000 nucleotides using the neighbor-joining method. Shorter sequences were subsequently inserted into the tree by using the parsimony insertion tool of ARB. The scale bar represents the estimated number of nucleotide changes per sequence position. The symbols at nodes show bootstrap values obtained after 1,000 resamplings.

Abundance of Desulfotomaculum subcluster Ih bacteria.

Based on all known 16S rRNA gene sequences of subcluster Ih bacteria and clone sequences obtained in this study, a new 16S rRNA-targeted oligonucleotide probe, Ih820, was designed for FISH. First, the specificity of probe Ih820 was experimentally evaluated by using the perfectly matching target organism P. thermopropionicum and the nontarget organism Moorella thermoacetica, which contains a single mismatch in the probe target site. Nonspecific binding of probe Ih820 to M. thermoacetica cells could be avoided by adding equimolar amounts of unlabeled competitor probe to the hybridization solution and by adjusting the formamide concentration to 20% (data not shown). Furthermore, probe DEM1164r (which was originally applied for PCR and membrane hybridization [52]) was optimized with suitable reference cultures (P. thermopropionicum, D. thermobenzoicum, D. thermosapovorans, and D. thermocisternum) for the specific detection of Desulfotomaculum cluster I bacteria by FISH. Subsequently, the relative abundances of bacteria belonging to Desulfotomaculum cluster I and subcluster Ih were determined for the nine environmental samples (Table 3). Subcluster Ih bacteria accounted for approximately 0.5% ± 0.2% and 4.1% ± 1.1% of the total DAPI (4′,6′-diamidino-2-phenylindole) counts in thermophilic-sludge samples TR2 and TD, respectively. The number of Ih820 probe-positive cells in these samples closely corresponded to the number of DEM1164r probe-positive cells (Table 3), confirming that subcluster Ih bacteria were the dominant members of the Desulfotomaculum clade in samples TR2 and TD. In contrast, no Ih820-positive cells were detected in the mesophilic-sludge samples MB and MC (Table 3). This observation is also consistent with the 16S rRNA gene library surveys, as all of the MB and MC clones analyzed were affiliated with the novel subcluster Ii.

TABLE 3.

Relative abundances of bacteria belonging to Desulfotomaculum cluster I and subcluster Ih as determined by FISH

No.	Environmental sample	Mean % of DAPI counts (SD)
		Desulfotomaculum cluster I (probe DEM1164r)	Desulfotomaculum subcluster Ih (probe Ih820)
1	TJ (thermophilic digested sludge)	−a	−
2	TO (thermophilic digested sludge)	−	−
3	TR2 (thermophilic granular sludge)	1.1 (0.8)	0.5 (0.2)
4	TD (thermophilic granular sludge)	4.3 (1.2)	4.1 (1.1)
5	MR1 (mesophilic granular sludge)	<0.1	<0.1
6	MB (mesophilic granular sludge)	<0.1	NDb
7	MC (mesophilic granular sludge)	<0.1	ND
8	MD (mesophilic digested sludge)	−	−
9	RP (rice paddy soil)	−	−

^a−, FISH-based cell counting was impossible due to high autofluorescence of the sample and/or low signal of probe-stained bacteria.

^bND, not detected (below detection limit).

Due to high autofluorescence or low probe-derived signals, microscopic quantification of the anaerobic digested-sludge samples TJ, TO, and MD and the rice paddy soil sample RP was not possible. Hence, additional quantitative real-time PCR assays were performed (Table 4). The relative abundance of each target group was calculated as the percentage of the total 16S rRNA gene copy number (defined as the sum of the gene copy numbers obtained by bacterial and archaeal primer sets) in the DNA extract. 16S rRNA genes of Desulfotomaculum cluster I bacteria were detected in all analyzed samples and comprised 0.03% to 4.8% of the total prokaryotic 16S rRNA gene pool. Subcluster Ih 16S rRNA genes could not be detected in samples MB and MC but accounted for 0.03% to 3.9% of the total 16S rRNA genes in the other samples.

TABLE 4.

Relative abundances of Archaea, Bacteria, and Desulfotomaculum cluster I and subcluster Ih as determined by real-time PCR

No.	Environmental sample	Abundance (%)
		Archaea	Bacteria	Desulfotomaculum cluster I	Desulfotomaculum subcluster Ih
1	TJ (thermophilic digested sludge)	0.5	99.5	1.7	0.3
2	TO (thermophilic digested sludge)	1.2	98.8	0.3	0.1
3	TR2 (thermophilic granular sludge)	26.5	73.5	1.2	1.1
4	TD (thermophilic granular sludge)	7.3	92.7	4.8	3.9
5	MR1 (mesophilic granular sludge)	31.4	68.6	0.1	0.1
6	MB (mesophilic granular sludge)	5.0	95.0	0.03	NDa
7	MC (mesophilic granular sludge)	7.3	92.7	0.1	ND
8	MD (mesophilic digested sludge)	1.0	99.0	0.2	0.3
9	RP (rice paddy soil)	2.4	97.6	0.1	0.03

^aND, not detected (below detection limit).

Cultivation of Desulfotomaculum subcluster Ih bacteria.

Taking advantage of known physiological properties of cultivated subcluster Ih bacteria, such as P. thermopropionicum (6, 7, 16, 22, 23), we also aimed to enrich/isolate these microorganisms from the environmental samples. Primary anaerobic enrichments were established at 37°C for the mesophilic-sludge samples and the paddy soil sample and at 55°C for the thermophilic-sludge samples. Because almost all members of subcluster Ih are syntrophic anaerobes, 20% (vol/vol) of a culture containing hydrogen- or formate-utilizing and aceticlastic methanogens was added to all primary enrichment cultures; thermophilic cultures were inoculated with M. thermautotrophicus strain ΔH, M. thermautorophicus strain type II, and M. thermophila, whereas mesophilic cultures were inoculated with cells of M. hungatei and M. concilii. In addition, because all known members of subcluster Ih can form spores, the inocula were pasteurized (80°C for 10 min) prior to the incubation. Enrichment cultures were set up with a basal medium supplemented with propionate (20 mM), ethanol (10 mM), and lactate (20 mM) because P. thermopropionicum can grow on these substrates in syntrophic coculture with hydrogenotrophic methanogens. FISH analysis of subcluster Ih bacteria in the 27 different enrichments showed that 10 enrichments were dominated by probe Ih820-positive cells. The other 17 cultures contained various morphotypes of cells that could not be identified as subcluster Ih bacteria. Five (2 propionate-degrading enrichments from samples TR2 and TD, 2 ethanol-degrading enrichments from samples TR2 and TD, and 1 lactate-degrading enrichment from sample TR2) of the 10 cultures enriched in subcluster Ih were found to contain P. thermopropionicum as a dominant population (as revealed by FISH with probe TGP690) and were thus not used for further experiments. The remaining five enrichments showed syntrophic propionate degradation (Table 5) and contained Ih820-positive cells, which did not hybridize with the P. thermopropionicum probe. Microscopic observation revealed that these cultures contained F₄₂₀-autofluorescent methanogen cells and spore-forming oval rods resembling Pelotomaculum species as dominant populations. Growth and methane formation of the cultures occurred after 40 to 90 days of incubation. Similar to cultures of other Pelotomaculum species (23, 43), it was difficult to maintain stable enrichments, i.e., growth and methane production sometimes stopped during the successive transfers of cultures. In order to minimize unstable growth, cultures were not permanently agitated (shaking of the vials for daily observation was done manually) for at least 40 days after starting the cultivation. The five cultures continuously produced methane with the concomitant degradation of propionate over six successive passages and contained probe Ih820-positive cells as the dominant bacterial population (Fig. 2).

TABLE 5.

Stoichiometry of propionate conversion by Desulfotomaculum subcluster Ih bacteria enrichment cultures^a

Origin of enrichment culture	Enrichment name	Cultivation temp (°C)	Propionate converted (mM)	Product formed (mM)		Electron recovery (%)
				Acetate	Methane
TJ	JP	55	5.8	0b	10.1	99
TO	OP	55	8.2	0b	14.1	99
MR1	MGP	37	15.5	12.4	13.5	96
MD	NP	37	7.3	0b	10.9	86
RP	FP	37	9.9	10.0	7.4	98

^aProducts were measured after 80 days of incubation.

^bAcetate was most likely converted to methane, because Methanosaeta-like cells were observed in the enrichment cultures under the microscope.

FIG.2.

In situ hybridization of novel subcluster Ih bacteria in syntrophic propionate-degrading enrichment cultures with Cy3-labeled Ih820 probe. Phase-contrast micrographs (A, C, E, G, and I) and fluorescence micrographs (B, D, F, H, and J) of the same fields are shown. (A and B) JP enrichment culture inoculated from environmental sample TJ; (C and D) OP enrichment culture from environmental sample TO; (E and F) MGP enrichment culture from environmental sample MR1; (G and H) NP enrichment culture from environmental sample MD; (I and J) FP enrichment culture from environmental sample RP. Bars, 10 μm.

We performed 16S rRNA gene-based clone analyses to characterize the dominant bacterial populations in these five enrichments in more detail. Clone libraries were constructed by using a universal bacterial primer set, and 10 clones were randomly selected and sequenced for each clone library. Most of the clones were affiliated with Desulfotomaculum subcluster Ih (Fig. 1). Four out of five phylotypes retrieved from the enrichments (MGP, NP, JP, and OP) perfectly matched or were very closely related to the phylotypes recovered from the respective original samples. The only exception was phylotype FP from the rice paddy enrichment, which was clearly different from phylotypes RP-1 and RP-2 from the original rice paddy soil sample (sequence similarities, 96.2 and 96.3%, respectively). None of the sequences obtained was similar to Syntrophobacter (2, 4, 17, 54) or Smithella (33) species, which are known mesophilic, syntrophic propionate-oxidizing Deltaproteobacteria.

Isolation of a novel syntrophic propionate-degrading strain.

We then attempted to isolate the subcluster Ih bacteria present in the enrichment cultures by employing a strategy successfully applied for the isolation of P. thermopropionicum (23). Since we found that after seven successive transfers some contaminating microorganisms, which probably did not participate in propionate degradation, were still present in the enrichment cultures, we again pasteurized the cultures at 85°C for 30 min or at 90°C for 20 min and serially diluted them into propionate medium containing hydrogenotrophic methanogens (M. thermautotrophicus strain ΔH and M. thermautotrophicus type II for thermophilic enrichments; M. hungatei for mesophilic enrichments). After a number of the serial dilution steps combined with pasteurization, a “pure” coculture with the hydrogenotrophic methanogen M. hungatei was obtained for enrichment MGP only at 37°C. The purity of this coculture was checked by microscopy and cultivation of the culture with a variety of media containing substrates such as pyruvate, lactate, sucrose, glucose, and yeast extract at 37°C or 55°C. Except for the propionate-containing enrichment medium, none of these media supported microbial growth.

Partial characterization of strain MGP.

Strain MGP was cultivated in pure coculture with M. hungatei from an anaerobic granular sludge treating an artificial wastewater (sample MR1) (Fig. 3). Cells of strain MGP were nonmotile and rod shaped, with a length of 2.7 to 5.5 μm and a width of 1 μm (Fig. 3). Spore formation was observed in late-log-phase cultures. The specific growth rate on 20 mM propionate medium in coculture with M. hungatei was approximately 0.2 day⁻¹ (calculated based on methane production). 16S rRNA gene sequence analysis revealed that the strain belonged to Desulfotomaculum subcluster Ih (Fig. 1). The most closely related validly described organism of strain MGP was P. schinkii (95% 16S rRNA sequence similarity with the P. schinkii rrnB 16S rRNA gene sequence) (6, 7).

FIG. 3.

Phase-contrast micrograph of strain MGP grown on propionate (20 mM) in coculture with M. hungatei. Bar, 10 μm.

Utilization of sulfate, sulfite, and organosulfonates by Desulfotomaculum subcluster Ih bacteria.

We evaluated the sulfate- and sulfite-reducing abilities of subcluster Ih bacteria that were highly enriched or isolated in this study, as well as those of other subcluster Ih bacteria, such as P. schinkii (6, 7) and strains JT and JI (43; Qiu et al., unpublished). All incubation experiments were performed in triplicate. Cells of enrichment cultures JP, OP, NP, and FP (after seven repeated passages) and the pure (co)cultures were transferred into medium supplemented with sulfate (20 mM) or sulfite (1 and 2 mM), propionate (20 mM), and 2-bromoethane sulfonate (5 mM) and incubated anaerobically at either 37°C or 55°C. However, no growth and no depletion of electron donors were found after prolonged incubations (over 6 months) and when dense cell suspensions were tested. In addition, sulfate and sulfite reduction also never occurred in sulfate (20 mM) or sulfite (1 and 2 mM) medium that was supplemented with propionate (20 mM) or terephthalate (1 mM) and external electron carriers, such as menadione (500 μg/liter), 1,4-naphthoquinone (500 μg/liter), vitamin K₁ (500 μg/liter), or hemin (50 μg/liter), and incubated for over 6 months (3, 19).

Moreover, we tested whether the growth of subcluster Ih bacteria was supported by organosulfonates, because some sulfate reducers and closely related non-sulfate reducers can utilize organosulfonates as electron acceptors for anaerobic respiration (29, 32). The pure (co)cultures P. thermopropionicum, P. schinkii, and strains JT, JI, and MGP and the enrichment cultures JP, OP, NP, and FP (after eight successive transfers) were transferred into taurine (10 mM), isethionate (10 mM), or cysteate (5 mM) medium supplemented with the respective electron donors (i.e., 10 mM propionate for P. thermopropionicum, P. schinkii, strain MGP, and the highly enriched cultures and 1 mM terephthalate for strains JT and JI). 2-Bromoethane sulfonate (5 mM) was added to the cultures to inhibit methanogens, except for the P. thermopropionicum cultures. However, growth and depletion of the electron donors were not observed after 3 months of incubation, although all experiments were conducted in duplicate culture vessels.

dsrAB and apsA sequence analyses.

To further investigate whether subcluster Ih microorganisms harbor genes required for dissimilatory reduction of sulfate or sulfite, such as the related Desulfotomaculum species, we screened for genes encoding the dissimilatory (bi)sulfite reductase (DSR) and adenosine-5′-phosphosulfate reductase. We used various PCR primers targeting conserved sites in the genes for the alpha and beta subunits of DSR (dsrAB) and the gene for the alpha subunit of adenosine-5′-phosphosulfate reductase (apsA) (9, 14, 25). A number of PCR experiments were performed under different reaction conditions (i.e., different Mg²⁺ concentrations in the PCR buffer, annealing temperatures from 45°C to 59°C, or reamplification of the initial PCR products) for the pure (co)cultures P. thermopropionicum, P. schinkii, and strains JT, JI, and MGP and the enrichment cultures JP, OP, NP, and FP (after eight successive transfers). However, only strain MGP and the thermophilic enrichment culture JP yielded a PCR product for dsrAB (1,797 bp) and apsA (377 bp), respectively. Sequencing and comparative analysis of both PCR products demonstrated that the sequences were homologous to the respective target genes of Desulfotomaculum species. In more detail, the dsrAB sequence from strain MGP formed a new lineage in the DsrAB tree and was related to sequences from other gram-positive bacteria (including the Desulfotomaculum subcluster Ia and If bacteria) (Fig. 4). The apsA sequence from the thermophilic enrichment culture JP formed a monophyletic lineage with the subcluster Ib Desulfotomaculum species Desulfotomaculum sapomandens and Desulfotomaculum geothermicum (deduced amino acid sequence similarity, 81.6%) in all trees (data not shown). However, it remains unclear whether this apsA sequence originated from a subcluster Ih bacterium, because (i) the amplified apsA fragment was too short for a reliable phylogenetic analysis and (ii) the enrichment culture JP still contained some contaminating microorganisms.

FIG. 4.

Phylogenetic consensus tree (based on FITCH distance matrix analysis) of DsrAB amino acid sequences deduced from dsrAB sequences of greater than 1,750 bases showing the affiliation of Pelotomaculum species strain MGP (indicated in boldface type). The bar indicates 10% estimated sequence divergence. Polytomic nodes connect branches for which a relative order could not be determined unambiguously by applying distance matrix, maximum-parsimony, and maximum-likelihood tree-building methods. Parsimony bootstrap values (100 resamplings) for branches are indicated by solid (>90%) or open (75 to 90%) circles. Branches without circles had bootstrap values of less than 75%. Gray boxes indicate different subclusters of Desulfotomaculum cluster I. LA-dsrAB Firmicutes is a low-G+C-content gram-positive bacterium with laterally acquired dsrAB genes.

dsrAB expression in strain MGP.

We performed RT-PCR experiments by using RNA extracts from strain MGP to check whether its dsrAB genes are transcribed. For this purpose, MGP cells were grown on propionate medium in the presence and absence of sulfate and subsequently subjected to RNA extraction. RT-PCR experiments using two newly designed primer pairs resulted in amplification of PCR products of the expected size, indicating that both cultivation conditions allowed strain MGP to transcribe its dsrAB genes (Fig. 5). The sequences of PCR products obtained in this RT-PCR experiment were completely identical to the dsrAB sequence of strain MGP that was obtained from DNA-based PCR amplification, thereby proving the specificity of the RT-PCR assay.

FIG. 5.

RT-PCR amplification of dsrAB mRNA and 16S rRNA from strain MGP. Lane M, marker; lanes 1, 4, and 7, DNA of strain MGP (positive controls); lanes 2, 5, and 8, RNA extracted from strain MGP grown on only propionate (20 mM); lanes 3, 6, and 9, RNA extracted from strain MGP grown on propionate (20 mM) plus sulfate (20 mM). PCR products were amplified by the following primer sets: MGP1F/MGP1R (lanes 1, 2, and 3); MGP9F/MGP3R (lanes 4, 5 and 6); and 341f/907r (lanes 7, 8, and 9).

DISCUSSION

Distribution and abundance of Desulfotomaculum subcluster Ih bacteria.

16S rRNA gene sequences from members of Desulfotomaculum cluster I were detected in all methanogenic samples analyzed in this study (Fig. 1). These findings are in accordance with previous 16S rRNA gene surveys of granular sludges (48, 49), anaerobic petrol-contaminated sites (11), phthalate isomer-degrading consortia (43), toluene-degrading consortia (13), and rice paddy soils (52, 57) and indicate that Desulfotomaculum cluster I bacteria are ubiquitously distributed in methanogenic environments. The relative natural abundance of this group varied from approximately 0.05 to 7% of the total bacterial or microbial populations, as determined previously by different molecular methods (10, 21, 51, 52). In comparison, our quantitative PCR data for Desulfotomaculum cluster I bacteria were also in the same range (0.03 to 4.8% of the total microbial 16S rRNA gene pool). Additionally, we showed that subcluster Ih-type cells represent the majority of the Desulfotomaculum cluster I cells in most of our samples. These observations are also consistent with previous reports (21, 51, 52). For example, quantitative dot blot and real-time PCR analyses of an Italian rice soil 16S rRNA gene clone cluster, which also belongs to subcluster Ih (Fig. 1) (52), demonstrated that its population size was almost equivalent to the size of the total Desulfotomaculum cluster I community (51, 52). At first glance, the proposed numerical dominance of subcluster Ih among environmental Desulfotomaculum cluster I members is inconsistent with the findings of Hristova et al. (21). They reported, based on quantitative dot blot hybridization with the 16S rRNA-targeted probe S-*-Dtm(bcd)-0230-a-A-18, that members of the subclusters Ib, Ic, and Id were the most abundant Desulfotomaculum cluster I bacteria in a variety of anoxic environments. However, reevaluation of the specificity of the probe showed that it also targets the majority of 16S rRNA gene sequences affiliated with subcluster Ih, suggesting that the results of Hristova et al. (21) and of our study are not necessarily contradictory. In summary, our findings show that Desulfotomaculum subcluster Ih constitutes an important fraction of the microbial community in a wide variety of methanogenic environments, highlighting its ecological impact in anoxic systems, usually with a low concentration of sulfate. However, the ubiquity of subcluster Ih bacteria will need to be further surveyed for other methanogenic environments, since our investigation was confined to eight anaerobic sludges and one rice paddy soil.

Ecophysiology of Desulfotomaculum subcluster Ih bacteria.

Their close phylogenetic relationship with classical Desulfotomaculum species suggests that members of Desulfotomaculum subcluster Ih are sulfate-reducing bacteria. However, all subcluster Ih-containing cultures (including P. thermopropionicum, P. schinkii, and strains JT and JI [42a, 43]) investigated in this study exhibited no growth with sulfate, sulfite, or organosulfonates, corroborating and extending previous findings on the physiological characteristics of this group (7, 22, 24). Consequently, environmentally retrieved 16S rRNA genes belonging to Desulfotomaculum cluster I are not always indicative of the presence of sulfate reducers and must be carefully interpreted.

Another interesting physiological feature of subcluster Ih bacteria is their ability to catabolize organic substances in syntrophic associations with hydrogenotrophic methanogens. With the exception of Cryptanaerobacter phenolicus (24), all previously isolated strains and cultures of subcluster Ih obtained in this study showed this unique phenotype, although the range of degradable substrates differed among strains/cultures. The physiology of the phenol degrader C. phenolicus (previously known as strain 7) has not been fully examined, and thus, it is currently unknown whether this strain can perform syntrophic substrate oxidation in cooperation with hydrogenotrophic microbes (24, 31). Given their recognized phenotypes and wide occurrence in low-sulfate, methanogenic environments, descendants of the Desulfotomaculum subcluster Ih branch most likely function as non-sulfate-reducing, syntrophic degraders of organic substrates in situ. This hypothesis received further independent support from a recent study, which showed, by using rRNA-based stable-isotope probing, that Pelotomaculum species were actively involved in syntrophic propionate oxidation in rice paddy soil (37).

Obligate syntrophic lifestyle of strain MGP?

Similar to previous cultivation attempts (22, 23, 43), the growth of subcluster Ih bacteria cultivated in this study was very slow and unstable. To enhance the stability of the growth of syntrophic, substrate-degrading cultures, a considerable amount (20% [vol/vol] inoculum) of pure cultures of hydrogen- or formate-utilizing methanogens and aceticlastic methanogens was added to all primary enrichment cultures. Unexpectedly, growth stability did not improve, and thus, only one propionate-degrading strain, MGP, was finally obtained in a binary mixed pure culture from an anaerobic granular sludge treating artificial wastewater. In contrast to many other propionate-oxidizing syntrophs, which are also able to respire sulfate (2, 4, 17, 42, 54) or dismutate crotonate in pure culture (33), this strain grew only in syntrophic association with hydrogenotrophic methanogens. Addition of various substrates did not support axenic growth of the strain. Similarly, the mesophile P. schinkii can also grow only on propionate in coculture with hydrogenotrophic methanogens and has thus been reported as the “first true obligately syntrophic propionate-oxidizing bacterium” (6, 7). The close phylogenetic relationship to P. schinkii and the currently known physiological capabilities suggest that strain MGP is also an obligate syntroph. Although we have recently isolated an additional strain, JI, which grows only on aromatic compounds in coculture with hydrogenotrophic methanogens and belongs to Desulfotomaculum subcluster Ih (Qiu et al. unpublished), obligate syntrophy is not a phenotype common to all members of this lineage. For example, C. phenolicus degrades phenol to benzoate (24), and P. thermopropionicum ferments pyruvate and fumarate (22, 23) in pure culture.

Presence and expression of dsrAB in strain MGP.

Although various PCR conditions were used, a dsrAB sequence was recovered only from the coculture containing strain MGP. The other Desulfotomaculum subcluster Ih bacteria either do not possess dsrAB or their dsrAB genes contained strong mismatches in the primer target sites, preventing amplification (58). Rapid accumulation of mutations in dsrAB would be particularly likely in these strains, as they apparently no longer use their dissimilatory (bi)sulfite reductase. Additional extensive experiments (for example, heterologous Southern hybridization analyses using dsrAB-targeted polynucleotide probes), which were beyond the scope of this study, will be required to clarify whether the genomes of these strains still carry dsrAB (pseudo)genes.

The dsrAB genes from strain MGP, representing the first entry of a Desulfotomaculum cluster Ih member in the dsrAB database, is affiliated with Desulfotomaculum species but formed a new lineage within the group (Fig. 4). This is in accordance with the 16S rRNA-based phylogeny of strain MGP. Previous comparative analyses of 16S rRNA- and DsrAB-based trees have shown that multiple lateral transfers of dsrAB have occurred among sulfate- or sulfite-reducing prokaryotes (25, 58). For example, members of Desulfotomaculum cluster I are phylogenetically separated in the DsrAB tree. Desulfotomaculum subclusters Ib to Ie form a monophyletic lineage with the deltaproteobacterial species Desulfobacterium anilini and have received their dsrAB genes laterally from a deltaproteobacterial ancestor (25, 58). In contrast, the dsrAB genes of strain MGP are affiliated with Desulfotomaculum subclusters Ia and If, suggesting that the strain acquired its dsrAB genes by vertical transmission from a gram-positive ancestor (Fig. 4). An interesting feature associated with the dsrAB genes of strain MGP is that these genes are transcribed in the presence and absence of sulfate (Fig. 5), although the MGP-containing coculture could not reduce sulfate, sulfite, or organosulfonates, which are typical (precursors of) substrates of dissimilatory (bi)sulfite reductases (29, 30, 32). The recovered dsrAB sequence fragment of strain MGP does not contain unusual errors (such as unexpected stop codons or frame shifts within the genes) indicative of pseudogenes. In addition, both subunits in the DsrAB amino acid sequence of strain MGP contain the conserved cysteine motif consensus sequences Cys-X₅-Cys and Cys-X₃-Cys that are essential for binding the [Fe₄S₄]-siroheme cofactor of sulfite reductases (5, 18). While these findings suggest that these dsrAB genes might still encode a functional protein, their potential roles in the physiology of strain MGP remains unresolved.

The apparent discrepancy between the presence of dsrAB and the inability to reduce sulfate/sulfite of strain MGP might reflect a relatively recent evolutionary transition. According to the 16S rRNAgene-based tree, subcluster Ih bacteria have only recently diverged from members of the other subclusters (Fig. 1), indicating a common endospore-forming, sulfate-reducing ancestor for all bacteria of the Desulfotomaculum cluster I branch. In order to cope with the usually limited availability of sulfate in methanogenic ecosystems, subcluster Ih bacteria could have specialized in living syntrophically with methanogenic archaea. Consequently, the capability for dissimilatory sulfate/sulfite reduction might have been lost in the course of this evolutionary process. In this respect, the presence of transcribed dsrAB genes might represent a genetic and/or gene-regulatory leftover in strain MGP. Inactivation and/or deletion of genes other than dsrAB, such as dsrMKJOP (hmeABCDE) and qmoABC, which might be essential for electron transfer during sulfate respiration (39), could be responsible for the apparent lack of this phenotype in subcluster Ih bacteria.

Similar evolutionary forces could have shaped Sporotomaculum species (3, 44), which are also not able to respire with sulfate or sulfite, although they belong to Desulfotomaculum subcluster Ib and possess dsrAB genes (Fig. 1 and 4). It is expected that comparative analysis of genome sequences of closely related sulfate-reducing and non-sulfate-reducing microorganisms will yield important insights into recent evolutionary adaptations of gram-positive sulfate reducers and the mechanisms underlying the apparent loss of their ancestral phenotype.