Diversity of Five Anaerobic Toluene-Degrading Microbial Communities Investigated Using Stable Isotope Probing

Abstract

Time-series DNA-stable isotope probing (SIP) was used to identify the microbes assimilating carbon from [¹³C]toluene under nitrate- or sulfate-amended conditions in a range of inoculum sources, including uncontaminated and contaminated soil and wastewater treatment samples. In all, five different phylotypes were found to be responsible for toluene degradation, and these included previously identified toluene degraders as well as novel toluene-degrading microorganisms. In microcosms constructed from granular sludge and amended with nitrate, the putative toluene degraders were classified in the genus Thauera, whereas in nitrate-amended microcosms constructed from a different source (agricultural soil), microorganisms in the family Comamonadaceae (genus unclassified) were the key putative degraders. In one set of sulfate-amended microcosms (agricultural soil), the putative toluene degraders were identified as belonging to the class Clostridia (genus Desulfosporosinus), while in other sulfate-amended microcosms, the putative degraders were in the class Deltaproteobacteria, within the family Syntrophobacteraceae (digester sludge) or Desulfobulbaceae (contaminated soil) (genus unclassified for both). Partial benzylsuccinate synthase gene (bssA, the functional gene for anaerobic toluene degradation) sequences were obtained for some samples, and quantitative PCR targeting this gene, along with SIP, was further used to confirm anaerobic toluene degradation by the identified species. The study illustrates the diversity of toluene degraders across different environments and highlights the utility of ribosomal and functional gene-based SIP for linking function with identity in microbial communities.

INTRODUCTION

Among petroleum-related environmental contaminants, benzene, toluene, ethylbenzene, and xylene (BTEX) are of particular concern because of their toxicity and easy migration in groundwater. Aerobic degradation of these chemicals is generally rapid (39, 56, 62); however, in the saturated subsurface, anaerobic conditions typically exist. Therefore, understanding anaerobic biodegradation is also relevant to site cleanup. In addition, many factors, such as electron acceptor availability, substrate competition between potential indigenous biodegraders, and cocontamination (e.g., ethanol), can affect anaerobic BTEX degradation. More knowledge on the diversity of degrading species in complex samples and the effect of electron acceptor availability on these species has the potential to enhance our understanding of the variability associated with anaerobic degradation.

With regard to BTEX contaminants, toluene degradation has been of great interest, with information available on pure and mixed cultures (based on the 16S rRNA gene) as well as the functional genes and enzymes involved. Toluene degradation has been observed over a range of electron-accepting conditions, with nitrate and sulfate being recognized as two important electron acceptors for this transformation. For example, relatively high levels of sulfate may occur naturally in some groundwater systems or may be added to enhance anaerobic degradation. A number of key toluene-degrading denitrifiers have been identified, belonging to the genera Azoarcus, Aromatoleum, Magnetospirillum, Dechloromonas, and Thauera (4, 14, 24, 27, 47, 48, 52, 64). Similarly, microorganisms have also been linked to toluene degradation under sulfate-reducing conditions, including, for example, microorganisms in the genera Desulfobacula, Desulfocapsa (41, 61), Desulfotomaculum (42), Desulfotignum (44), Desulfovibrio (3), and Desulfosporosinus (32, 36, 60).

Although numerous organisms have been linked to anaerobic toluene degradation in pure or mixed cultures, it is challenging to determine if these organisms are actually responsible for toluene degradation in complex samples. To address this, molecular methods have been developed to link function with identity in complex samples, enabling a greater understanding of microbial communities involved in contaminant removal or other biological processes. A key molecular method for this has been DNA-based stable isotope probing (SIP). This method has been used only recently to study anaerobic toluene degradation. To date, SIP has been applied in three different studies under sulfate-reducing conditions. In 2010, SIP was used to identify the active toluene degraders in aquifer sediment from a former gasworks site in Germany (11, 43, 60) and in a sulfate-reducing consortium developed from a BTEX-contaminated aquifer also in Germany (11, 43, 60). In 2011, SIP was also applied to contaminated sediment samples from Germany to identify active toluene-degrading species (46).

In the present study, we expanded this knowledge by applying DNA-based SIP to samples from a wider range of sources, including uncontaminated sites. SIP was applied under conditions of two electron acceptors (nitrate or sulfate amended) and at several time points during toluene degradation. SIP was applied to the ribosomal gene as well as the functional gene (bssA, encoding benzylsuccinate synthase) previously correlated to anaerobic toluene degradation. Benzylsuccinate synthase has been recognized as a key enzyme for anaerobic toluene biodegradation under nitrate-reducing (2, 30), sulfate-reducing (59–61), ferric iron-reducing (12, 26), and methanogenic enrichment cultures or environmental samples (58, 59).

Here, the overall aim was to determine the diversity of active anaerobic toluene degraders and bssA genes across different habitats, in both contaminated and uncontaminated samples, and compare these results to the current knowledge on pure cultures as well as the results of previous SIP studies involving only contaminated-site samples. We hypothesized that soil and sediments from BTEX-contaminated sites would more actively degrade toluene and would contain a wide variety of anaerobic degraders compared to material from uncontaminated sites. This work represents the first study to use SIP to examine the diversity of active anaerobic toluene degraders across diverse sample sources.

MATERIALS AND METHODS

Development of toluene-degrading microcosms.

A wide range of inoculum sources were investigated for toluene-degrading potential. These sources included agricultural soils (Michigan), subsurface soil from BTEX-contaminated sites (Michigan), sediments from a former gas-compressor site (25) (Oklahoma), digester sludge samples from wastewater treatment plants (Michigan), and anaerobic granular sludge (Washington). Triplicate samples of ∼10 g (wet weight) were incubated in sterile 160-ml serum bottles containing 50 ml anaerobic basal medium (63) and sealed with rubber stoppers and aluminum seals. Microcosms were prepared under strictly anaerobic conditions in an anaerobic chamber (Coy Laboratory Products Inc., Grass Lake, MI). Potassium nitrate and magnesium sulfate were added to the microcosms to serve as electron acceptors. From approximately 38 incubations, only two nitrate-amended and three sulfate-amended microcosms exhibited toluene degradation, and these were selected for the SIP experiments. The two active nitrate-amended microcosms were seeded from an agricultural soil (referred to here as AgN) and anaerobic granular sludge from a reactor in Washington (referred to here as GSN). The three sulfate-amended toluene-degrading microcosms were established from the same agricultural soil (referred to here as AgS), contaminated sediments were from a previous BTEX contaminated aquifer (referred to here as CSS) (25), and digester sludge from a wastewater treatment plant in St. Clair, Michigan (referred to here as DSS). For each microcosm, 3 g biomass (wet weight) from the enrichment cultures obtained at the screening stage were used to inoculate the AgN, AgS, and CSS microcosms. Also, 3 g biomass (wet weight) of freshly sampled sludges were used to construct the GSN and DSS microcosms. All microcosms were anoxically incubated in 60-ml serum bottles containing 25 ml of anaerobic basal medium, as described above. Potassium nitrate and magnesium sulfate were added to a final concentration of 1 g liter⁻¹ NO₃⁻ and SO₄²⁻. Each treatment involved triplicate abiotic controls, triplicate samples amended with unlabeled toluene (1 μl, 99%; Chem Service, West Chester, PA), and triplicate samples amended with labeled toluene (1 μl [ring-¹³C₆]toluene, 99%; Cambridge Isotope Laboratories, Inc., Andover, MA). These microcosms were incubated at room temperature (∼20°C) with reciprocal shaking.

Analytical techniques.

Toluene concentrations in headspace gas samples (200 μl) were typically determined weekly with a gas chromatograph (Perkin Elmer) equipped with a flame ionization detector and a capillary column (diameter, 0.53 mm; DB-624; J&W Scientific). The injector and detector temperatures were set at 200°C, and the column temperature was 120°C.

DNA extraction and ultracentrifugation.

Microcosms were sacrificed for DNA extraction at various time points during toluene depletion for AgN, GSN, AgS, and DSS to better understand the flow of carbon through these microbial communities. For CSS, an early stage (50% toluene removal) was investigated due to limited availability of active microcosms. A Powersoil DNA extraction kit (Mo Bio Laboratories, Inc., Carlsbad, CA) was used for total nucleic acid extraction from entire microcosms in each treatment according to the manufacturer's recommended procedure. Quantified (ND-1000 spectrometer; Nanodrop) DNA extracts (∼10 μg) were loaded into Quick-Seal polyallomer tubes (13 by 51 mm; 5.1 ml; Beckman Coulter) along with a Tris-EDTA (TE; pH 8.0)–CsCl solution. Prior to sealing (cordless quick-seal tube topper; Beckman), the buoyant density (BD) was determined with a model AR200 digital refractometer (Leica Microsystems Inc.) and adjusted by adding small volumes of CsCl solution or Tris-EDTA buffer with a final BD of 1.7300 mg liter⁻¹. The tubes were centrifuged at 178,000 × g (20°C) for 48 h in a Stepsaver 70 V6 vertical titanium rotor (8 samples, each with a 5.1-ml capacity) within a Sorvall WX 80 Ultra series centrifuge (Thermo Scientific). Following centrifugation, the tubes were placed onto a fraction recovery system (Beckman), and fractions (150 μl) were collected. The BD of each fraction was measured, and CsCl was removed by glycogen-assisted ethanol precipitation.

PCR and TRFLP.

The density-resolved fractions for each treatment were PCR amplified using 27F-FAM (5′-AGAGTTTGATCMTGGCTCAG, 5′-end labeled with carboxyfluorescein) and 1492R (5′-GGTTACCTTGTTACGACTT) (Operon Biotechnologies) as previously described (15). The presence of PCR products was confirmed by 1.5% agarose gel electrophoresis and subsequent staining of the gels with ethidium bromide. Only samples showing a band were subject to terminal restriction fragment length polymorphism (TRFLP). PCR products were purified with a QIAquick PCR purification kit (Qiagen Inc.), following the manufacturer's instructions, and approximately 150 ng was digested with HaeIII (New England BioLabs) with a 6-h incubation period. Additional digests (HhaI, MseI, Bsp1286I, BsrBI, etc.) for TRFLP analyses in a number of labeled heavy fractions were included to correlate the TRFLP fragment lengths to the in silico cutting sites of the cloned 16S rRNA gene sequences. DNA fragments were separated by capillary electrophoresis (ABI Prism 3100 genetic analyzer; Applied Biosystems) at the Research Technology Support Facility (RTSF) at Michigan State University. Data were analyzed with GeneScan software (Applied Biosystems), and the abundance of each fragment was determined [(individual fragment area/total area of all fragments) × 100].

Presence of bssA in microcosms and enumeration in SIP fractions.

The presence of the benzylsuccinate synthase alpha-subunit gene (bssA) was investigated using different primers pairs (Table 1) on DNA extracted from each of the five sample types. A gradient PCR was performed with annealing temperatures ranging from 45°C to 58°C. Only primers producing a strong and specific amplicon were selected for further investigation (see below). Four samples exhibited the presence of bssA genes (AgN, GSN, DSS, and CSS). For three of these (all except CSS, due to a lack of sufficient sample to extract an adequate amount of DNA from), quantitative PCR (qPCR) was used to enumerate bssA gene abundance in gradient fractions. The assay was conducted in a Chromo 4 real-time PCR cycler (Bio-Rad) using the primer sets 7772f-8546r for AgN and GSN and SRBf-SRBr for DSS. Both ¹²C- and ¹³C-labeled nucleic acid samples from density gradient fractions were subjected to quantification. In addition, the relative abundances of the dissimilatory sulfite reductase (dsrAB) gene and the denitrifying nitrite reductase gene (nirK) in total genomic DNAs extracted at different time points (from AgN, GSN, AgS, DSS, and CSS) were also determined with qPCR. The primer pair DSR1F (5′-ACSCACTGGAAGCACG)-DSR4R (5′-GTGTAGCAGTTACCGCA) (57) was used to target dsrAB, and the primer pair nirK876 (5′-ATYGGCGGVAYGGCGA)-nirK1040 (5′-GCCTCGATCAGRTTRTGGTT) (21) was used to quantify nirK. Each 20-μl PCR mixture contained 10 μl SYBR green real-time PCR solution (Applied Biosystems), 0.25 μM each primer, and 1 μl DNA template. The thermal protocol consisted of an initial denaturation (95°C, 15 min), 40 cycles of amplification (95°C, 15 s; 55°C, 20 s; 72°C 20 s), and a terminal extension step (72°C, 2 min). Melting curves were constructed from 55°C to 95°C and read every 0.6°C for 2 s. For each gradient fraction, 1 μl solution was diluted with 3 μl water as a template (to conserve the sample). Cloned plasmid DNA was utilized as a standard for quantification, and numbers of gene copies were determined as previously described (50) (plasmid size was 4,730 bp, including inserts of 774 bp with 7772f-8546r and 97 bp with SRBf-SRBr). The average qPCR efficiencies for bssA genes amplified from primer pairs 7772f-8546r and SRBf-SRBr were 94.7% and 97.6%, respectively. The average standardization slopes were −3.46 and −3.38 for 7772f-8546r and SRBf-SRBr, respectively. The average R² values were high (0.99) for both primer pairs, and the detection limits were ∼20 copies/μl and ∼5 copies/μl for 7772f-8546r and SRBr-SRBf, respectively.

Table 1.

Primers used in this study to investigate the presence of the bssA gene

Primer	Sequence (5′–3′)	Reference
6888f	AATTCATCGTCGGCTACCACG	61
7772f	GACATGACCGACGCSATYCT	61
8546r	TCGTCGTCRTTGCCCCAYTT	61
8828r	AGCAGRTTGSCCTTCTGGTT	61
SRBf	GTSCCCATGATGCGCAGC	7
SRBr	CGACATTGAACTGCACGTGRTCG	7
bssApd2f	CCTATGCGACGAGTAAGGTT	59
bssApd2r	TGATAGCAACCATGGAATTG	59
bssN2f	GGCTATCCGTCGATCAAGAA	59
bssN2r	GTTGCTGAGCGTGATTTCAA	59
bssAf	ACGACGGYGGCATTTCTC	6a
bssAr	GCATGATSGGYACCGACA	6a

Sequencing of partial bssA and 16S rRNA genes.

Clone libraries of the 16S rRNA genes were constructed for each treatment using DNA extracted following toluene depletion. The DNA was amplified with 27F-1492R as described above except that the forward primer was unlabeled and the final extension time was increased to 15 min. To reduce sequencing redundancy, restriction fragment length polymorphism (RFLP) analyses were performed, and specific operational taxonomic units (OTU) were selected for sequencing. In addition to 16S rRNA sequencing, amplicons generated with bssA primer pairs were also prepared for cloning and sequencing. The PCR products were purified with a QIAquick PCR purification kit (Qiagen Inc.) and cloned into the Escherichia coli TOP10 vector supplied with a TOPO TA cloning kit (Invitrogen Corporation). E. coli clones were grown on Luria-Bertani (LB) medium solidified with 15 g agar liter⁻¹ with 50 μg ampicillin liter⁻¹ for 16 h at 37°C. Colonies with inserts were verified by PCR with primers M13 F (5′-TGTAAAACGACGGCCAGT-3′) and M13 R (5′-AACAGCTATGACCATG-3′), plasmids were extracted from the positive clones with a QIAprep miniprep system (Qiagen, Inc.), and the insertions were sequenced at RTSF. The Ribosomal Database Project (RDP) (Center for Microbial Ecology, Michigan State University) analysis tool “classifier” was utilized to assign taxonomic identity. Phylogenetic trees for the partial bssA sequences along with the closest matches in GenBank were obtained by the neighbor-joining method using MEGA 5.0 software.

Nucleotide sequence accession numbers.

The 16S rRNA gene clone libraries and the partial bssA gene sequences obtained have been deposited in GenBank under accession numbers JN083436 to JN083463, JN806259 to JN806354 (partial 16S rRNA gene sequences), and JQ039466 to JQ039472 (partial bssA gene sequences).

RESULTS AND DISCUSSION

Previous SIP studies on BTEX biodegradation focused primarily on samples or biomass from formerly contaminated sites (22, 31, 60). To expand on this work and to discover novel toluene-degrading bacteria, the current study applied SIP to a wide range of inoculum sources, including agricultural soil, contaminated aquifer sediment, and anaerobic granular and digester sludge.

Occurrence of toluene degradation.

Toluene degradation achieved with inocula obtained from various sources, including agricultural soil, digester sludge, anaerobic granular sludge, and contaminated soils and sediments, was examined under nitrate- and sulfate-amended conditions. Toluene biodegradation was observed in 2 of 18 and in 3 of 20 experimental set-ups involving nitrate amendment and sulfate amendments, respectively (see Fig. S1 in the supplemental material). Active toluene microcosms were also tested for their benzene biodegradation; however, no biodegradation was found in any consortium. Although nitrate and sulfate removal was not investigated, the denitrifying nitrite reductase gene (nirK) and the dissimilatory sulfite reductase genes (dsrAB) (DSR1F/4R targeted both dsrA and dsrB) did increase in nitrate-amended and sulfate-amended samples, respectively, as toluene was degraded, as indicated by decreasing cycle threshold (C_T) values over time (see Fig. S2B and C in the supplemental material). These data suggest that toluene was likely being degraded under anaerobic conditions. In contrast, as would be expected, no increase in nirK was observed in the sulfate-amended samples (no change in threshold cycle [C_T] values) (see Fig. S2A), and no increase in dsrAB was noted in the nitrate-amended samples (see Fig. S2D). The organisms responsible for ¹³C assimilation were identified by the comparison of relative abundances of specific terminal restriction fragments (TRFs) between the control (amended with unlabeled toluene) and the sample (amended with labeled toluene) at selected time points for each fraction. The identities of each enriched TRF for each microcosm type were then determined using additional restriction enzyme digests and by comparison to predicted cut sites in each 16S rRNA gene clone library, as described below.

DNA-SIP of nitrate-amended microcosms.

Toluene biodegradation under nitrate-reducing conditions has been studied extensively (4, 10, 13, 49). To our knowledge, no detailed DNA-based SIP study has been conducted on nitrate-reducing, toluene-degrading mixed consortia. For the present study, a natural microbiota (agricultural soil) and an engineered microbial community (granular sludge) were investigated. These communities were targeted, in part, because it was likely that both had high levels of denitrifying microorganisms.

In the agricultural soil nitrate-amended microcosms (AgN), a 214-bp HaeIII TRF became the only dominant TRF (>80%) among labeled heavy fractions (banding between 1.7523 and 1.7448 g ml⁻¹) at all three sampling points (33%, 75%, and 100% toluene removal) (Fig. 1A). To identify the toluene-degrading bacteria based on the TRFLP results and to assign phylogenetic affiliation to distinct TRFs, the 16S rRNA clone library (see Table S1 in the supplemental material) derived from total DNA was inspected. Three different Burkholderiales-related microorganisms within the families Alcaligenaceae, Comamonadaceae, and Oxalobacteraceae all exhibited a predicted HaeIII TRF close to 214 bp (from the analysis of the clone sequences, each had a predicted HaeIII cut site of 219 bp) (see Table S1). Multiple digestions were then carried out to determine which of these three sequences was actually responsible for the 214-bp HaeIII TRF in the heavy fractions. From these digests (see Table S2 in the supplemental material), the phylotypes within the family Comamonadaceae were found to be directly correlated with carbon uptake from toluene and hence toluene degradation. Interestingly, the AgN clone library (see Table S1) contained a partial sequence related to the genus Azoarcus, which has been linked to toluene degradation under nitrate-amended conditions (2, 7, 24, 65). However, no enrichment of the appropriate Azoarcus TRF (77 bp) was seen in the labeled heavy fractions (Fig. 1B), indicating that these microbes were likely not the major degraders. These finding illustrates the importance of culture-independent approaches, compared to culture-dependent approaches, for understanding functions in mixed cultures.

Fig 1.

Percent relative abundance of fragments (digested by HaeIII) assigned to Comamonadaceae (A), Azoarcus (B) (nitrate-amended agricultural soil), Thauera (C) (nitrate-amended granular sludge), Desulfosporosinus (D and E) (sulfate-amended agricultural soil), Syntrophobacteraceae (F) (sulfate-amended digester sludge), and Desulfobulbaceae (G) (sulfate-amended contaminated site). Symbols: ▲, [¹³C]toluene, ∼33% degraded; ♦, [¹³C]toluene, ∼75% degraded (except for G, which is ∼50% degraded); ■, [¹³C]toluene, ∼100% toluene; ☐, [¹²C]toluene (∼100% toluene). TRFLP was not conducted with the heavier fractions of A, B, and C because gel electrophoresis indicated no PCR product in ¹²C-amended samples.

Although Comamonadaceae-related bacteria are often char-acterized as aerobic bacteria, some researchers have described members of the Comamonadaceae as denitrifying bacteria (16, 28). Comamonadaceae-related microorganisms have been correlated with cyclohexanol (40) and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) (28) biodegradation under nitrate-reducing conditions. As this is the first report of a Comamonadaceae-related microorganism degrading toluene under nitrate-amended conditions, additional work is needed (e.g., isolation and additional testing) to confirm that this organism can indeed degrade toluene without oxygen present. These results are potentially novel because the most commonly reported nitrate-reducing, toluene-degrading species (in the genera Azoarcus, Thauera, and Dechloromonas) fall within another Betaproteobacteria family, the Rhodocyclaceae.

The nitrate-amended microcosms constructed by using anaerobic granular sludge (GNS) exhibited a more diverse microbial community than those from AgN (see Table S1 in the supplemental material). TRFLP revealed significant ¹³C uptake for one TRF (Fig. 1C), a 72-bp fragment that dominated labeled heavy fractions (BD > 1.7448 g ml⁻¹). Additional enzyme digestions (see Table S2 in the supplemental material) in combination with the 16S rRNA clone library (see Table S1) were utilized to further distinguish this HaeIII TRF. The putative toluene degrader was classified in the genus Thauera (72 bp). Thauera-related phylotypes were strongly enriched in the heavy fractions at all three time points. Thauera spp. have been reported by others to be important toluene-degrading species under nitrate-reducing conditions (5, 10, 34, 53). The toluene metabolic pathway in Thauera is known to be initiated by the formation of benzylsuccinate from toluene and fumarate (10, 19, 20, 33). Sequencing of the partial bssA genes revealed that all 32 clones were closely related to the Thauera bssA gene (see below). The presence of Thauera-related bssA genes and the enrichment of these genes in ¹³C heavy fractions (see below) as quantified by qPCR strengthens the hypothesis that the Thauera phylotype was responsible for toluene degradation in this complex microbial community.

DNA-SIP of sulfate-amended microcosms.

In the sulfate-amended agricultural soil microcosms (AgS), the majority (30 of 37 clones) of the microbial community was classified within the Clostridia (see Table S3 in the supplemental material). Slight label assimilation was noted in two TRFs in the DNA extracted from microcosms which consumed ∼33% toluene; however, at the later two DNA extraction points (∼75% and ∼100% toluene removal), label assimilation was more pronounced (Fig. 1D and E). Two TRFLP fragments, 77 bp and 213 bp, were enriched in the heavy ¹³C-labeled fractions during the course of biodegradation, and no PCR products were found in the corresponding heavy ¹²C-labeled fractions. It is likely that both phylotypes were responsible for carbon uptake from toluene over the course of the incubations. The 16S rRNA clone library data (see Table S3) in combination with additional restriction digests (see Table S2 in the supplemental material) indicated that the two TRFLP fragments were both affiliated with the genus Desulfosporosinus (>98% sequence identity). The genus Desulfosporosinus has been linked to toluene degradation (32, 51). It was also identified as being able to assimilate [¹³C]toluene in a recent SIP project (60).

In the sulfate-amended digester sludge microcosms (DSS), a significant proportion (20 of 59 clones) of the microbial community were classified in the Deltaproteobacteria (see Table S3 in the supplemental material). The most abundant TRFLP fragment in ¹³C-labeled heavy fractions was a 204-bp TRF. This fragment was highly enriched in heavy fractions at the three extraction times, and the effect was most obvious for the last two time points (∼75% and ∼100% toluene depletion) (Fig. 1F). The maximum relative abundances of the 204-bp fragment were 27% (at a BD of 1.7360 g ml⁻¹), 43% (at a BD of 1.7480 g ml⁻¹), and 50% (at a BD of 1.7502 g ml⁻¹) for the three extraction points, indicating an increase in label uptake with time. In contrast, the relative abundance of this fragment was less than 5% in each unlabeled control gradient fraction (Fig. 1F). The BDs of these heavy peaks are very close to that of fully ¹³C-labeled Methylobacterium extorquens (1.757 g ml⁻¹) (38), indicating a high degree of label assimilation. The bacterial clone library generated (see Table S3) indicated that the 204-bp TRF belonged to the family Syntrophobacteraceae. As described above, additional digestion of the heavy fractions was performed to confirm this correlation between the 204-bp TRF and the associated 16S rRNA sequence (see Table S2 in the supplemental material).

Members of Syntrophobacteraceae were previously identified as sulfate-reducing bacteria (17, 37) and can degrade long-chain fatty acids (54, 55) and propionate (6) under sulfate-reducing conditions. The closest relative of the sequences represented by the 204-bp TRF with a validly published name is Syntrophobacter wolinii (18), sharing 91% sequence similarity. S. wolinii has been reported as a sulfate-reducing bacterium, and its closest relatives are Desulfomonile tiedjei and Desulfoarculus baarsii. The 204-bp TRF also shared a more distant similarity (88% 16S rRNA similarity) with the sulfate-reducing strain PRTOL1, a toluene sulfate reducer isolated from fuel-contaminated subsurface soil (8, 9). Syntrophobacteraceae were identified in a toluene-degrading sulfate-reducing bacterial consortium, but in that study, the Desulfobulbaceae were identified as key organisms of toluene degradation within the consortium (43). In another study, polar lipid-derived fatty acid SIP in combination with whole-cell hybridization linked toluene degradation to the Desulfobacter-like populations, while Syntrophobacteraceae were present in the microbiota but not responsible for toluene biodegradation (45). The Syntrophobacteraceae were also observed in a benzene-degrading in situ microcosm but were not linked to benzene biodegradation (23). To date, to our knowledge, there is no direct evidence to correlate members of the family of Syntrophobacteraceae with primary anaerobic toluene biodegradation. The presence of bssA genes similar to sulfate-reducing strain PRTOL1-related bssA genes (see below) in DNA extracted from the digester sludge treatment is consistent with the hypothesis that the novel Syntrophobacteraceae clade played an active role in toluene biodegradation in these sulfate-amended samples.

The third sulfate-amended sample involved microcosms (referred to as CSS) inoculated with sediment from a former gas compressor site (25). Previous research indicated that sulfate was the major terminal electron acceptor at this site. The clone library for this treatment indicated the dominance (91 of 106 clones) of microorganisms within the Deltaproteobacteria (orders Desulfobacterales and Desulfuromonadales) (see Table S3 in the supplemental material). Unfortunately, because of sample limitations, SIP was performed only on a sample from an early toluene-degrading stage (∼50% toluene removal). A 202-bp TRF was enriched in the heavy ¹³C-labeled fractions (relative abundance as ∼70% with a BD of 1.7469 g ml⁻¹) (Fig. 1G). The clone library for this microbial community (see Table S3) illustrated a dominance of Desulfobulbaceae-affiliated organisms (80 of 106 clones), which correlated with the 203-bp TRF. Multiple enzyme digestions (see Table S2 in the supplemental material) confirmed that the Desulfobulbaceae-affiliated 16S rRNA gene sequence was responsible for the enriched HaeIII 203-bp TRF. These data indicate that the Desulfobulbaceae-affiliated microorganisms were responsible for toluene degradation. Desulfobulbaceae have been classified as BTEX-degrading sulfate reducers (1, 29, 31, 35, 43). They were also reported in the DNA-SIP project described above (60) but were not identified as the primary toluene degraders. Interestingly, Desulfosporosinus-related microorganisms (identified as primary toluene degraders in AgS) were present in the CSS clone library (see Table S3) but were not enriched in the ¹³C-labeled heavy fractions.

Partial sequencing of the bssA gene.

A number of primers successfully amplified partial bssA genes from four of the five treatments (see Table S4 in the supplemental material). Amplicons from AgN, GSN, DSS, and CSS were selected for sequencing. Primer set 7772f-8546r produced PCR products of the expected size (∼774 bp) within the two nitrate-amended consortia. A total of 32 clones from agricultural soil and granular sludge nitrate-amended enrichment cultures were digested, and representative clones (as indicated by restriction digests) were selected for sequencing. All 64 clones showed the same OTU, indicating sequence similarity to Thauera-related bssA genes. The primer pair SRBf-SRBr displayed good coverage of the sulfate-amended enrichment cultures (see Table S4). Representative OTU from the DSS and CSS were sequenced, and three amplicons of each sample showed ∼92% (90 of 97 bp) similarity to the partial bssA gene from the sulfate-reducing bacterium PRTOL1 (EU780921.1), which was congruent with the template sequences of the primer set. In addition, for CSS, longer partial bssA genes were obtained with the primer pair 7772f-8828r. Of these, 19 were closely related to bssA of strain TRM1 (99% sequence similarity) (61). Three clones could not be classified with any known bssA sequences. Unfortunately, unspecific PCR products or no PCR products were produced from the sulfate-amended agricultural soil (AgS) from the primers tested (Table 1). Representative sequences produced from primers 7772f-8546r and 7772f-8828r were compared to their closest matches in GenBank and aligned in a phylogenetic tree (515 bp) (Fig. 2).

Fig 2.

Phylogenetic tree of bssA partial sequences (515 bp) from nitrate-amended agricultural soil (primer set 7772f-8546r), nitrate-amended granular sludge (primer set 7772f-8546r), and sulfate-amended contaminated soil microcosms (primer set 7772f-8828r), along with the closest matches in GenBank, constructed with MEGA 5.0 software using the neighbor-joining method.

Quantification of bssA genes in SIP fractions.

To further confirm labeling of anaerobic toluene degraders in the nitrate- and sulfate-amended samples, the bssA genes were quantified from ¹²C and ¹³C gradient fractions at the last time point using the primer sets 7772f-8546r and SRBf-SRBr. Quantitative PCR analysis of gradient fractions detected separation of bssA genes in labeled and unlabeled samples (Fig. 3). In GSN, quantitative label assimilation was very evident in the ¹³C-labeled fractions, where bulk bssA gene moved to a heavier fraction, with a BD of 1.7513 g ml^−1, compared to that of the ¹²C-labeled fractions (1.7208 g ml⁻¹) (Fig. 3A). In AgN, highly ¹³C-labeled bssA genes were present at a BD of 1.7415 g ml⁻¹, while such heavy bssA genes were not found in the ¹²C-labeled fractions, but the peak occurred at a lighter fraction, with a BD of 1.7099 g ml⁻¹. Also, a tail of a bssA gene formed in lighter ¹³C-labeled fractions, with a BD of 1.7046 g ml⁻¹ (Fig. 3B). In DSS, a separation of the ¹²C and ¹³C peaks was also seen (Fig. 3C). Quantitative PCR was not performed on the AgS fractions (no bssA primers were suitable for these microcosms) or the CSS fractions (limited sample was available).

Fig 3.

Difference between abundance of bssA gene copies in ultracentrifugation fractions from labeled ([¹³C]toluene) and unlabeled toluene-amended microcosms from the granular sludge nitrate-amended microcosms (A), the agricultural soil nitrate-amended microcosms (B), and the digester sludge sulfate-amended microcosms (C), as determined by qPCR. Symbols: ■, [¹³C]toluene (∼100% toluene degraded); ☐, [¹²C]toluene (∼100% toluene degraded).

In summary, five distinct phylotypes were identified as the active toluene-degrading bacteria under either nitrate- or sulfate-amended conditions for five different microbial communities. Interestingly, of all samples tested, including BTEX-contaminated soil, agricultural soil and wastewater treatment samples were the most fruitful sources of toluene degraders. These findings could imply that these systems should be investigated for novel degraders of other key contaminants. For three of the treatments investigated here, the phylotypes were similar to those of previously identified toluene degraders (Thauera-, Desulfosporosinus-, and Desulfobulbaceae-related phylotypes), whereas the other two treatments produced novel toluene degraders (Comamonadaceae- and Syntrophobacteraceae-related phylotypes). The discovery of two novel toluene degraders indicates the importance of culture-independent approaches for identifying the active microorganisms in complex samples. In addition, this study provided information on the diversity of bssA sequences and the utility of a number of primer pairs for detecting the bssA gene. Further, the work highlights the value of combining ribosomal- and functional-gene-based SIP to link function with identity for complex microbial samples and adds to our understanding of the microbial ecology of toluene-degrading communities from various environments.