Analysis of structure, function, and activity of a benzene-degrading microbial community

Authors

  • Sven Jechalke,

    1. Department of Isotope Biogeochemistry, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Current affiliation:
    1. Julius Kühn–Institut – Federal Research Centre for Cultivated Plants (JKI), Institute for Epidemiology and Pathogen Diagnostics, Braunschweig, Germany
    Search for more papers by this author
    • SIP, stable isotope probing.

  • Alessandro G. Franchini,

    1. Department of Environmental Biotechnology, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Current affiliation:
    1. Environmental Microbiology, Institute of Biogeochemistry and Pollutant Dynamics, Swiss Federal Institute of Technology (ETH) Zurich, Zurich, Switzerland
    Search for more papers by this author
    • SIP, stable isotope probing.

  • Felipe Bastida,

    1. Department of Isotope Biogeochemistry, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Current affiliation:
    1. CEBAS-CSIC, Department of Soil and Water Conservation, Campus Universitario de Espinardo, Murcia, Spain
    Search for more papers by this author
  • Petra Bombach,

    1. Department of Isotope Biogeochemistry, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Search for more papers by this author
  • Mónica Rosell,

    1. Department of Isotope Biogeochemistry, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Current affiliation:
    1. Departament de Cristal°lografia, Mineralogia i Dipòsits Minerals, Facultat de Geologia, Universitat de Barcelona, Barcelona, Spain
    Search for more papers by this author
  • Jana Seifert,

    1. Department of Proteomics, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Search for more papers by this author
  • Martin von Bergen,

    1. Department of Proteomics, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    2. Department of Metabolomics, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    3. Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Aalborg, Denmark
    Search for more papers by this author
  • Carsten Vogt,

    Corresponding author
    • Department of Isotope Biogeochemistry, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Search for more papers by this author
  • Hans H. Richnow

    1. Department of Isotope Biogeochemistry, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany
    Search for more papers by this author

Correspondence: Carsten Vogt, Department of Isotope Biogeochemistry, UFZ – Helmholtz Centre for Environmental Research, Permoserstr. 15, 04318 Leipzig, Germany. Tel.: +49 341 235 1357; fax: +49 341 235 450822; e-mail: carsten.vogt@ufz.de

Abstract

We identified phylotypes performing distinct functions related to benzene degradation in complex microbial biofilms from an aerated treatment pond containing coconut textile. RNA- and protein-stable isotope probing (SIP) and compound-specific stable isotope analysis were applied to delineate bacteria and predominant pathways involved in the degradation of benzene. In laboratory microcosms, benzene was degraded at rates of ≥ 11 μM per day and per gram coconut textile under oxic conditions. Carbon isotope fractionation with isotopic enrichment factors (ε) of −0.6 to −1‰ and no significant hydrogen isotope fractionation indicated a dihydroxylation reaction for the initial ring attack. The incubation with [13C6]-benzene led to 13CO2 formation accompanied by 13C-labeling of RNA and proteins of the active biomass. Phylogenetic analysis of the 13C-labeled RNA revealed that phylotypes related to Zoogloea, Ferribacterium, Aquabacterium, and Hydrogenophaga within the Betaproteobacteria predominantly assimilated carbon from benzene. Although the phylogenetic classification of identified 13C-labeled proteins was biased by the incomplete metagenome information of public databases, it matched with RNA-SIP results at genus level. The detection of 13C-labeled proteins related to toluene dioxygenase and catechol 2,3-dioxygenase suggests benzene degradation by a dihydroxylation pathway with subsequent meta-cleavage of formed catechol.

Introduction

Aerated ponds offer a cost-efficient method for removal of organic matter in wastewater by microbial decomposition and sedimentation of settleable solids (Thörneby et al., 2006). The ponds harbor complex microbial communities that are often seen as a ‘black box’; knowledge on phylotypes performing distinct functions in pond systems is generally scarce. Recently, we could demonstrate that aerated ponds are a promising technique for the remediation of benzene-polluted water (Jechalke et al., 2010). As shown in this previous study, benzene concentrations in the first 14 months of operation were reduced on average from 20 mg L−1 at the inflow to 2 μg L−1 at the outflow of the system, and the benzene removal was assumed to be predominantly caused by aerobic biodegradation, because only 1% of the benzene escaped to the atmosphere (Jechalke et al., 2010). Coconut textile was chosen as a cheap and natural surface to increase biofilm formation in the pond. During pond operation, a substantial biofilm developed on the coconut textile, which contained high numbers of benzene-degrading microbes, as determined by the most probable number method (Jechalke et al., 2010). Denaturing gradient gel electrophoresis revealed that the biofilm was composed of a complex microbial community that was distinct to the water community (Jechalke et al., 2010). However, the used methods did not allow revealing the diversity of the benzene-degrading community within the biofilm community. Due to the continuously changing environmental conditions in the pond system (e.g. fluctuating temperature, light regime, and oxygen concentration), a heterogeneous benzene-degrading community might have developed, being robust against stress caused by regularly changing environmental conditions.

So far, microorganisms belonging to several different taxonomical lineages have been identified as aerobic benzene degraders, for example Acidovorax, Arthrobacter, Rhodococcus, Rhodoferax, Planococcus, Pseudomonas, Pseudoxanthomonas, Thermus, and Acinetobacter species [see Cao et al. (2009) for an overview]. To date, the number of microorganisms not cultivated – including those potentially degrading benzene – exceeds the number of pure cultures by several orders of magnitude (Keller & Zengler, 2004). Therefore, methods independent from the classical isolation approach are required to identify and characterize benzene degraders and to understand trophic interactions in complex environmental systems.

Over the last decade, stable isotope probing (SIP) approaches have provided an important link between ecosystem functioning and microbial physiology and have expanded the knowledge about the ecology of microorganisms in diverse environments (Manefield et al., 2002; Neufeld et al., 2007). With regard to biodegradation research, RNA-SIP techniques were successfully applied to identify phylotypes involved in, for example, the anaerobic degradation of benzene (Kasai et al., 2006; Sakai et al., 2009). Besides identification of phylotypes metabolizing pollutants, elucidation of biodegradation pathways in complex environments is especially relevant for a better understanding of biodegradation processes. Protein-SIP is a new technique providing both phylogenetic and functional information in microbial ecology studies by analysis of labeled and nonlabeled peptides (Jehmlich et al., 2010a; Seifert et al., 2012; Taubert et al., 2012). In addition, compound-specific stable isotope fractionation analysis (CSIA) has become a valuable tool for the assessment of biodegradation of groundwater contaminants (Meckenstock et al., 2004) and that also allows the identification of benzene degradation pathways (Fischer et al., 2008).

The goal of this study was to elucidate the dominant benzene degradation pathways and to identify benzene-degrading species from the aerated treatment pond by means of RNA-SIP, protein-SIP, and CSIA. Biofilms were taken from the pond and incubated under laboratory conditions with 13C-benzene. The flow of benzene carbon in this experiment along a food chain consisting of bacteria and eukaryotes, including chironomid larvae, was recently described elsewhere (Bastida et al., 2011).

Materials and methods

Site location and treatment pond setup

The aerated treatment pond system is set up next to a refinery plant in Leuna, Germany, in order to clean up the gasoline-contaminated groundwater which contains benzene, methyl tert-butyl ether, and ammonium as main contaminants, as described elsewhere (Jechalke et al., 2010). The system is an experimental plant and has been continuously operated since November 2007.

The oxygen content in the different compartments of the basins can be regulated by oxygen sensors coupled to aeration modules. Two different aeration periods were set up in the system to compare and characterize the degradation of the above target pollutants: (A1) a gradient of < 0.1 mg L−1 in the inflow area up to 1 mg L−1 minimum dissolved oxygen concentration in the outflow area, held from October 2007 to May 2009 (Jechalke et al., 2010); (A2) beginning from the 18th of May 2009, the aeration was increased, reaching maximum oxygen concentrations of 5.9 ± 0.5 mg L−1 in the basins in August 2009 (see details in Supporting Information, Fig. S1).

Chemicals

Chemicals, if not otherwise specified, were either obtained from Sigma-Aldrich (Munich, Germany) or Merck (Darmstadt, Germany) and were of analytical grade quality. [13C6]-benzene (≥ 99 atomic percent) was obtained from Campro Scientific (Berlin, Germany) and is referred to as 13C-benzene throughout the manuscript. Benzene was purchased from Merck (p.A. quality) with a natural carbon isotope composition of −25.2 ± 0.3‰ (corresponding to 1.08 atomic percent) and is referred to as 12C-benzene throughout the manuscript.

Compound-specific stable isotope analysis

Carbon and hydrogen enrichment factors were determined in laboratory microcosm experiments with coconut textile samples. Textile samples were taken in August 2008 and 2009 (corresponding to A1 and A2 periods, respectively) before and after the sampling for the SIP experiment (Fig. S1). All samples were taken from a depth of 60 cm. Details about the experimental design and the calculation of the isotope enrichment factors are given in the Supporting Information. In short, textile pieces were incubated at room temperature in mineral medium (Vogt et al., 2002) containing 50 mg L−1 benzene. Microcosms were sampled periodically in duplicates with gas-tight syringes to measure the benzene concentration and isotope composition of carbon and hydrogen. Sterile control experiments were set up without textile inoculum.

SIP experimental design

Biofilm samples for the SIP experiment were taken from the aerated treatment pond from a depth of 60 cm at the end of period A1 on the 12 May 2009. In this zone, benzene was continuously biodegraded under microaerobic conditions (Jechalke et al., 2010). Ten sterilized 1 L Duran® bottles were filled with 300 mL filtered (0.2 μm) benzene-containing inflow water and 9 g coconut textile with biofilm (wet weight, 64 ± 6% moisture) sampled in May 2009. Bottles were closed with Teflon-coated butyl-stoppers (VWR International, Darmstadt, Germany) and screw-caps. The benzene concentration in these microcosms was determined by headspace gas chromatography as described by Kleinsteuber et al. 2008). Total benzene concentrations were calculated using a Henry coefficient (Hi) of 0.153 for 15 °C as reported by Peng & Wan 1997). The residual benzene concentrations were between 0.12 and 0.15 mM per bottle. Subsequently, 12C- or 13C-benzene was additionally added from 20 mM aqueous stock solutions for reaching initial benzene concentrations between 0.23 and 0.34 mM.

Four microcosms were spiked with 13C-benzene and four with 12C-benzene. Two microcosms, autoclaved at 121 °C for 20 min and spiked with 13C-benzene, served as controls to monitor abiotic benzene losses. Due to the presence of background benzene, the 13C-benzene microcosms contained c. 50 atom% 13C-benzene at the beginning of the experiment (Fig. 1). To mimic field conditions as far as possible, the microcosms were incubated at daylight conditions at 15 °C, which is similar to the average annual temperature of the pond water. The bottles were shaken at low rotation speed (85 r.p.m.) on a horizontal shaker to enhance the oxygen diffusion into the liquid phase while preserving the integrity of the biofilms. After complete degradation of the quantity of benzene added, two 12C- and two 13C-benzene-spiked microcosms were sacrificed for further analysis (at time t1, after 2 days of incubation; Fig. 1). The remaining bottles were opened under a sterile laminar flow for 10 min to exchange the air in the headspace. After closing the bottles again, the spiking was repeated four times more with intermittent degradation of benzene to < 2 μM per bottle (Fig. 1). Hence, beginning with the second spiking, the 13C-benzene microcosms contained only 13C-labeled benzene (Fig. 1). Subsequently, additional two 12C- and two 13C-benzene microcosms were sacrificed (at time t2, after 8 days of incubation, Fig. 1). The benzene-pulsing strategy was used to reach high levels of 13C-labeling while reducing possible toxic effects of benzene to the indigenous microbial community. CO2 samples (2 mL) for carbon isotope analysis were taken by sterile syringes before and after each spiking with benzene (see below) as well as before and after headspace exchange. The headspace of the control bottles was not exchanged to prevent loss of benzene. CO2 samples of control bottles were taken at the beginning and at the end of the experiment. Each gas sample was injected into an evacuated 10 mL vial. A detailed description of the isotope analysis of the gas samples and calculation of mineralization is given in the Supporting Information. The pH was measured at the beginning and at the end of the incubations by taking 0.8-mL liquid samples (pH-Meter 765 Calimatic; Knick, Berlin, Germany).

Figure 1.

Cumulative-degraded benzene in microcosms incubated with 12C-benzene (solid diamonds), 13C-benzene (solid squares), and in autoclaved controls (solid triangles). Addition of 12C- or 13C-benzene (each after benzene degradation below detection limit) is indicated by arrows. 13C-benzene-amended microcosms contained initially 12C-benzene. For clarification, the amount of degraded 13C-benzene is additionally shown (open circles). Sampling time points (t0, t1, t2) are indicated by dashed lines. Each data point represents the average of two replicate bottles.

Nucleic acid extraction

For phylogenetic analysis of the entire biofilm community, DNA was extracted from 0.5 g (wet weight) of textile sample used as inoculum for the microcosms (t0) using the FastDNA® Kit (MP Biomedicals, Heidelberg, Germany). For analysis of the benzene-degrading community in the SIP experiment, RNA was extracted from 0.2 g textile (wet weight) sampled at the time points t0 (inoculum), t1, and t2 using the RNeasy Mini Kit (QIAGEN, Hilden, Germany) as described before (Nikolausz et al., 2007). Co-extracted DNA was digested using the DNA-free™ kit (Applied Biosystems Inc., Foster City, CA) according to the manufacturer's instructions.

Ultracentrifugation, amplification, fingerprinting, and phylogenetic analysis

Equilibrium density gradient centrifugation was performed using a cesium-trifluoroacetate gradient (CsTFA) as described elsewhere (Bombach et al., 2010). Gradients were loaded with 750 ng RNA which was quantified before using the Quant-iT™ RiboGreen® RNA Assay Kit (Invitrogen, Karlsruhe, Germany). Fifteen fractions were harvested from each gradient, analyzed for buoyant density (BD), and precipitated as described previously (Lueders et al., 2004; Whiteley et al., 2007). The RNA content of each fraction was quantified by a Quant-iT™ RiboGreen® RNA assay kit.

To characterize the phylogenetic structure of the total biofilm community, a clone library of 479 clones based on 16S rRNA genes was generated from the biofilm, taken at the beginning of the experiment (t0). Based on screening by amplified rDNA restriction analysis (ARDRA), 181 representative clones were selected for sequencing. Identical sequences were grouped together with 131 distinct sequences resulting in 72 OTUs at a genetic distance of 0.04 (Fig. S2). Assuming that identical ARDRA patterns represent similar bacterial sequence types, the estimated diversity sequence coverage according to Good (1953) was 81%, indicating that a majority of the bacterial sequence types were captured.

Procedures for amplification, cloning, and phylogenetic analysis of 16S rRNA genes of t0 and the cDNA obtained from the heavy rRNA fraction of the 13C-benzene microcosms (t2, fraction 5, BD of 1.83 g mL−1) as well as terminal-restriction fragment length polymorphism (T-RFLP) analysis of bacterial communities after digestion with MspI (BioLabs, New England) from each gradient fraction are described in the Supporting Information. Obtained nucleic acid sequences were deposited in the EMBL Nucleotide Sequence Database under the accession numbers FN824827FN824857 for sequences obtained from 16S rRNA genes (t0) and FN824958FN824975 for sequences obtained from cDNA of 16S rRNA genes (t2, fraction 5).

Protein extraction and LTQ-Orbitrap mass spectrometry

Textile subsamples from duplicate bottles at t1 and t2 were used for protein extraction using protocols modified from Benndorf et al. (2007, 2009). For each sample, 2 g (wet weight) of textile material was treated directly with 8 mL of phenol solution (10 g phenol in 1 mL water). The suspension was shaken for 5 min, ultrasonicated for 1 min (Bandelin Sonoplus HD70, cycle 50%), and then centrifuged for 15 min at 8243 g. The phenol phase was further processed as described by Benndorf et al. (2009). After protein precipitation with ammonium acetate, the obtained pellet was suspended in 70% ice-cold ethanol and incubated overnight at −20 °C. The obtained pellet was dissolved in 20% trichloroacetic acid solution. The mixture was incubated for 5 min at −20 °C and 15 min at 4 °C, followed by centrifugation at 15 710 g for 10 min. The resulting pellet was washed twice with 100% ice-cold acetone and incubated for 15 min at −20 °C. Subsequently, the suspension was centrifuged for 10 min at 15 710 g, and the pellet was dried at 37 °C for acetone evaporation. After quantification according to Bradford (1976), 50 μg of proteins was used for separation by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), as described elsewhere (Benndorf et al., 2007). Details regarding sample preparation and further LTQ-Orbitrap mass spectrometry are provided in the Supporting Information.

Mass spectrometric data analysis and calculation of relative isotope abundance

Raw data were applied to a database search using Thermo Proteome Discoverer software (v1.0 build 43) to carry out a tandem ion search algorithm from the Mascot house server (v2.2.1) by database comparison against all entries in the National Center for Biotechnology Information [NCBInr database 11–14 July 2009 (t1) 20–25th of August 2009 (t2)] with 10 p.p.m. tolerance for the precursor and 0.8 Da for MS2 fragments. Further, trypsin with a maximum of two missed cleavages was selected, and variable modifications, such as methionine oxidation and carbamidomethylation of cysteine, were allowed. Peptides were considered to be identified by Mascot when a probability < 0.05 (probability based ion scores threshold > 40) was achieved. Proteins were considered to be identified if at least one peptide was identified with an ion score higher than 45.

The calculation of relative isotope abundance was carried out using peptide information obtained from mass spectra of 12C-benzene samples and detecting the corresponding mass in a limited retention time window (± 5 min) in the spectra of the 13C samples. If possible, the identity of the mass (peptide) was checked by the fragment spectra. For calculation of the incorporation efficiency, we used the definition given by Snijders et al. 2006) which is based on the isotope pattern of the mass peaks. In addition, a Perl script provided in the supporting material of the study of Choudhary et al. 2006) was used. The calculation of the 13C incorporation was carried out as described by Jehmlich et al. (2010b).

Results

Phylogenetic analysis of the total biofilm community

A significant portion of sequences obtained at the beginning of the experiment (t0) belonged to Betaproteobacteria (34%), related to the orders of Burkholderiales (42%) and Rhodocyclales (44%; Fig. S3). Within these orders, most of the sequences identified at the genus level were affiliated with the genera Hydrogenophaga (9%) and Rhodoferax (12%). Another main group of the Betaproteobacteria sequences belonged to the genus Ferribacterium (9%), showing high sequence similarity with Dechloromonas aromatica RCB (98%) and Ferribacterium limneticum strain cda-1 (99%). A considerable amount of phylotypes affiliated to obligate anaerobic bacteria like Chlorobia (12%), Anaerolineae (6%), or Clostridia (5%) were also detected.

Carbon and hydrogen stable isotope fractionation during biodegradation of benzene

Carbon and hydrogen isotope fractionation during biodegradation of benzene by pond biofilms were determined in two microcosm experiments (Fig. S1). Benzene was degraded at a rate of 0.3–2.3 mM per day, corresponding to 0.06–0.46 mM per day and gram textile (wet weight). In sterile control experiments, the benzene concentration remained stable. Carbon enrichment factors for both experiments (± 95% confidence interval) ranged between −1.01 ± 0.09‰ and −0.64 ± 0.08‰, whereas no significant hydrogen isotope fractionation was observed after more than 90% of the benzene was degraded.

Benzene degradation and CO2 release in SIP experiments

When the textile biofilm was incubated with 13C-benzene, on average 0.08 and 0.3 mmol of benzene were degraded after 2 days of incubation (t1) and after 8 days of incubation (t2), respectively (Fig. 1). Benzene concentrations remained stable in the autoclaved control samples. Benzene concentrations, measured 1–3 days after spiking (see Fig. 1), were always below the detection limit, leading to average benzene degradation rates of ≥ 0.1 or ≥ 0.01 mM per day and gram textile (wet weight). Within the 8 days of incubation, 13CO2 was continuously released in the 13C-benzene-amended microcosms (Fig. S4), whereas δ13C values in 13C-benzene-spiked autoclaved controls and 12C-benzene microcosms remained constant or decreased slightly from −5‰ to −10‰ during benzene degradation, respectively. According to the amounts of produced 13CO2 and degraded benzene (Fig. S4), the average amount of mineralized benzene was 42 ± 8% (0.02 and 0.11 mmol for t1 and t2, respectively). The pH remained neutral (7.3–7.6) during the course of incubation.

Identification of benzene-degrading bacteria using RNA-SIP

RNA extracted from the 12C-microcosms accumulated at BD between 1.77 and 1.81 g mL−1, which is characteristic for nonlabeled rRNA for members of all three domains of Bacteria, Archaea, and Fungi (Lueders et al., 2004). At t1, RNA extracted from the 13C-benzene microcosms showed a slight shift to higher BD compared with the RNA from 12C-benzene microcosms (Fig. S5) indicating a small labeling of RNA, likely due to the low amount of 13C-benzene degraded at this stage of incubation. In contrast, a significant 13C-incorporation was evident at t2, where one peak appeared with a BD characteristic for unlabeled RNA, and a second peak with a BD ranging from 1.83 to 1.85 g mL−1 which is indicative of 13C-labeling of RNA. This average shift in BD corresponds to a 13C enrichment in RNA of about 50–75 atom%, according to the relationship between 13C enrichment and BD as reported by Manefield et al. (2002).

When comparing the T-RFLP profiles generated from the gradient fractions of the 13C-microcosms at t2, significant differences between the light and heavy fractions appeared. At least eight peaks (150, 174, 425–428, 441, 453, 477, 481, and 485 bp) were detected in the heavy fractions which were partially less abundant in the light fractions (Fig. 2). Notably, the same was true for the sampling point t1, although only a minor shift in the BD of the labeled compared with the unlabeled RNA was observed, indicating involvement in the initial attack of benzene. The relative abundances of at least four of these T-RFs (425–428, 453, 481, and 485 bp) clearly increased during the time course in the heavy fractions of the 13C-benzene microcosms and in the light fractions of the 12C-benzene microcosms, indicating an enrichment of these bacteria during incubation with benzene and thus representing key populations of aerobic benzene degraders in the biofilm community.

Figure 2.

Bacterial terminal-restriction fragment length polymorphism (T-RFLP) fingerprints of bacterial communities after digestion with MspI. Shown are the density-resolved RNA gradient fractions retrieved from 12C- or 13C-benzene microcosms after (a) 2 (t1) and (b) 8 days (t2) of incubation. Buoyant densities (g mL−1) of respective gradient fractions are given in parentheses. T-RFs appearing in the heavy fractions are indicated by arrows; abundant T-RFs are marked in bold. No PCR-products were obtained for fractions > 1.81 g mL−1 from the 12C-benzene microcosms.

Phylogenetic analysis of 13C-labeled 16S rRNA genes

A 16S rRNA clone library of 79 clones was created from the fraction with a BD of 1.83 g mL−1 at t2 (Fig. S5). Eighteen different sequence types were found by affiliation to closest relatives using the ncbi-blast algorithm (Table 1). Assuming that identical ARDRA patterns represent the same bacterial sequence types, clone coverage extended more than 80% of total abundance of sequence types according to calculations used previously (Good, 1953). Sequences were mainly related to the genera Zoogloea (52%), Ferribacterium (19%), Aquabacterium (10%), Hydrogenophaga (8%), Leptothrix (6%), Dechloromonas (3%), Rhodoferax (1%), and unclassified Comamonadaceae (1%) within the Betaproteobacteria (91%).

Table 1. Phylogenetic affiliation of clones obtained from the clone library at t2 after 8 days of incubation. The clone library was generated from one heavy rRNA fraction of the 13C-benzene microcosms (fraction 5, buoyant density of 1.83 g mL−1). Clones are sorted with descending frequencies
Clone accession no.Frequency of clones (%)MspI digested for T-RFLP (bp)Clone fragment size (bp)Closest sequenceGenusAccession no.Coverage (%)Max identity (%)
FN824960 244261311Zoogloea sp. EMB 62 Zoogloea DQ413151 9999
FN824962 184271316Dechloromonas aromatica RCB Dechloromonas CP000089 10098
Ferribacterium limneticum strain cda-1 Ferribacterium NR_026464 10098
FN824959 144851315Zoogloea ramigera strain 106 Zoogloea NR_026130 10099
FN824961 91361267Comamonadaceae bacterium MWH55 Aquabacterium AJ556799 10099
FN824958 84811337Hydrogenophaga sp. Rs71 Hydrogenophaga AM110076 9999
FN824974 5480431Eikelbloom type 1701 Leptothrix EU636006 10098
FN824975 54281321Thauera sp. R-28312 Thauera AM084110 10099
FN824965 34531262Zoogloea sp. EMB 62 Zoogloea DQ413151 10096
FN824968 31741305Dechloromonas hortensis strain MA-1 Dechloromonas AY277621 9997
Dechloromonas denitrificans ED1T  AJ318917 9997
FN824969 32021199Uncultured bacterium clone SHA-5 Unclassified “Bacteroidetes” AJ306736 10094
FN824963 11501273Pirellula sp. Pirellula X81945 9889
FN824964 14261040Zoogloea caeni strain EMB 43 Zoogloea DQ413148 10099
FN824966 14001218Sphingomonadaceae bacterium N Unclassified Sphingomonadaceae DQ497241 9995
FN824967 14841195Beta proteobacterium BAC49 Variovorax EU180529 9998
Curvibacter putative symbiont of Hydra magnipapillata Curvibacter FN543107 9998
FN824970 14771305 Rhodoferax fermentans Rhodoferax D16212 9996
FN824971 1396403Sinorhizobium sp. R-25067 Rhizobium AM084031 10099
FN824972 14011288Spirochaetes bacterium SA-10 Treponema AY695841 10091
FN824973 14411306 Methylobacter tundripaludum Methylobacter AJ414655 9997

To identify potential benzene degraders represented by the respective T-RFs of the gradient fractions, amplified 16S rRNA genes of sequenced clones were analyzed by T-RFLP. Based on this, the eight abundant T-RFs of the heavy fractions could be attributed mainly to phylotypes related to the genera Zoogloea, Dechloromonas, Ferribacterium, Aquabacterium, Hydrogenophaga, Thauera, or Leptothrix (see Table 1 for more details).

Protein identification and SIP of proteins

Protein yields obtained from textile biofilms did not change significantly from t1 to t2 (426 ± 67 and 311 ± 80 μg g−1, respectively), and also the protein patterns of SDS-PAGE gels (Fig. S6) were nearly identical. A total of 112 and 96 positive protein identifications were obtained at t1 and t2, respectively. The majority of proteins were related to Bacteria and Archaea, but some were also related to Eukaryotes, such as globin and tubulin from Plasmodium chabaudi chabaudi, myosin from Tribolium castaneum and photosynthetic proteins such as the photosystem II protein D2 from Peridinium foliaceum (Table S1). Phycocyanin-related proteins, typically abundant in Cyanobacteria, were also found.

Identified proteins were related by function and grouped according to the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (Table S1). Only small differences were found between t1 and t2. Proteins were mainly related to energy metabolism (t1: 29%; t2: 30%), photosynthesis (13%; 15%), biodegradation and metabolism of xenobiotics (10%; 11%), transcription or translation (10%; 11%), and membrane transport (10%; 11%). Among the proteins related to xenobiotic degradation in KEGG, we detected several dioxygenases, such as isopropylbenzene-, alkylbenzene-, chlorocatechol-, toluene-, glyoxilase- or catechol 2,3-dioxygenases, or methane/phenol/toluene hydroxylase (Table S1).

At t1, no incorporation of 13C was detected for any examined peptide. At t2, a strong incorporation was detected in peptides related to proteins involved in benzene biodegradation pathways [ring-hydroxylating dioxygenase (TodC2), catechol 2,3-dioxygenase] and house-keeping processes (e.g. DNA-directed RNA polymerase subunit alpha, isocitrate lyase, elongation factor Tu). These proteins had a 13C isotope abundance of more than 90 atom%, and the majority was affiliated to genera of the classes Betaproteobacteria (Dechloromonas, Azoarcus, Thauera) and Gammaproteobacteria (Pseudomonas, Shewanella, Halomonas). Peptides of proteins structurally affiliated with the soluble methane monooxygenase hydroxylase (sMMO) from Methylococcaceae were less enriched in 13C (22.5 atom%; Table 2).

Table 2. 13C-incorporation into proteins and peptides extracted from the textile at time point t2. Further parameters given in the table are as follows: peptide sequence, peptide incorporation (PepI), and relative isotope abundance (RIA). Best hit organism list has been carried out blasting every single peptide and taking those species with the highest and equal score; if agreement with clone libraries occurred, only genera found both in the protein analysis and the clone library of time point 0 (no. 1) or both time point 0 and RNA-SIP heavy fraction (no. 2) are shown. Organisms with identical families found in RNA-SIP heavy fractions are marked (no. 3), family name given in brackets were obtained from The Ribosomal Database Project website (http://rdp.cme.msu.edu/)
Protein nameBest hit organismsPeptide sequencePepI (%)RIA (%)
  1. SIP, stable isotope probing.

Putative cox2 cytochrome oxidase subunit 2 Dechloromonas aromatica 2 DVNLPDWSLSNTQLGAK94.095.5
NNIYDELSDIR95.0 
Isocitrate lyaseDechloromonas aromatica,2 Pseudomonas syringae,1 Pseudomonas fluorescens,1 Thauera sp.,2 Rhodoferax ferrireducens T1182SSVTALTGSTEEEQFH94.595.3
GYSAADVVR96.0 
DNA-directed RNA polymerase subunit alphaDechloromonas aromatica,2 Bordetella sp.,1 Thauera sp.,2 Azoarcus sp. BH72,1 Pseudomonas fluorescens,1 Pseudomonas stutzeri A15011IIDVQSVSPVQAK94.595.0
AENIYYIGDLIQR95.5 
Beta-ketothiolaseDechloromonas aromatica RCB2AAAAIDAGYFK96.095.5
KGDVVFDTDEHLKR95.0 
Glyoxalase/bleomycin resistance protein/dioxygenaseDechloromonas aromatica,2 Thauera sp.,2 Pseudomonas putida1FETFAQGYETYPDYPVK77.091.5
LDADLQAYGVK96.5 
IPAGELLETGER96.0 
QADSAGIDFFAFK96.5 
Soluble methane monooxygenase hydroxylase component alpha subunitMethylomonas PKSIII,1,3 Methylomonas sp.1,3YLNTDLNNAFWTQQK24.022.5
APVSVGAQEVHR21.0 
TodC2 (toluene dioxygenase)Thauera DNT-1,2 Pseudomonas putida,1 Pseudomonas fluorescens1MTDIYVGER96.096.0
LEVSSAFLCYR96.0 
Methane/phenol/toluene hydroxylaseThauera MZ1T2FILQQDPANVPIVQR95.095.0
Catechol 2,3 dioxygenaseThauera sp. MZ1T,2 Pseudomonas sp.1VLDLEEGINFYK96.095.7
NETFAGLGYLAQR95.0 
GETIYFFDPSGNR96.0 
OmpA/MotB domain proteinThauera MZ1T,2 Azoarcus sp. BH72,1 Dechloromonas aromatica RCB2LSADALFDFDK89.093.5
LEVILAVGHTDR98.0 
Acetoacetyl-CoA reductaseAzoarcus sp. BH72,1 Thauera MZ1T2GAFGQTNYAAAK95.094.0
SGVTVNTISPGYIGTK93.0 
Phosphate ABC transporter, periplasmic phosphate-binding proteinRhodoferax ferrireducens T118,2 Curvibacter2INYQSVGSGAGLK95.096.3
INYQSVGSGAGIK97.0 
VGEGTAVNWPVGAGGK97.0 
VSAIPDEAPTELQR96.0 
Malate dehydrogenaseAzoarcus sp. BH72,1 Thauera sp.,2 Azoarcus sp. BH71DQPVILQLLDLPQAQK96.094.2
VAVTGAAGQIGYSLLFR95.5 
GLSSAASAANAAIDHIR91.0 
DehydrogenaseBradyrhizobium japonicum USDA 110VVIVTGAGSGIGAAAAR95.095.0
NADH-quinone oxidoreductase Dechloromonas aromatica 2 VIYQPVTIEPR96.596.0
CatalaseHalomonas SK1, Vibrio rumoiensisNLTDAEAGELVAQDR93.593.8
LFNYADAQR94.0 
Probable amino acid ABC transporterLeptothrix cholodnii SP-61,3TAYGQGVADEFEK91.091.0
Alkyl hydroperoxide reductase/Thiol specific antioxidant/Mal allergenShewanella piezotolerans WP3, Cellvibrio japonicus Ueda107, Shewanella benthica KT99, Shewanella sediminis HAW-EB3, Aeromonas salmonicida subsp. salmonicida A449VAEIHDLGIGR93.592.3
IQYPMIGDPTGAITR91.0 
Acetaldehyde dehydrogenaseLeptothrix cholodnii SP-6,1 Pseudomonas putida,1 Pseudomonas pseudoalcaligenes,1 Dechloromonas aromatica2YAGNLDIMTAAAAR92.592.3
AIIILNPAEPPLIMR92.0 

Discussion

In the present study, we aimed to identify the benzene-degrading species and predominant degradation pathways in an aerated treatment pond for in situ remediation of benzene-contaminated groundwater. In laboratory microcosms amended with 13C-benzene, approximately half of the 13C-labeled carbon ended up in carbon dioxide, indicating that a significant part of the labeled benzene was mineralized by the microbial community. In a previous publication from this laboratory, we could demonstrate that Gram-negative bacteria played a key role in the benzene metabolization and that the benzene-derived carbon was further channeled to eukaryotes and Chironomus sp. larvae naturally feeding on the biofilm in the aerated treatment pond (Bastida et al., 2011). In the former study, the benzene degraders could not be taxonomically identified in high resolution by total lipid fatty acid-SIP analysis due to the methods' low potential for taxonomic assignment. The results of the present RNA-SIP experiment clearly reveal that benzene was metabolized by diverse members of Betaproteobacteria, a class that includes, like all Proteobacteria, exclusively Gram-negative bacteria, confirming the former results. In particular, Zoogloea phylotypes dominated within the library of the heavy fraction. Although Zoogloea strains are known for utilizing several aromatics using the meta-pathway (Unz & Farrah, 1972), a direct link of Zoogloea to benzene biodegradation has not yet been reported. Thus, our RNA-SIP results indicate that Zoogloea species, often found in activated sludge probably due to their ability to form dense cell aggregates (Shao et al., 2009), could be important benzene degraders in complex, biofilm-related microbial communities. Correspondingly, a possible involvement of Zoogloea phylotypes in benzene degradation was recently reported by Takahata et al. (2006), who detected a phylotype similar to Z. resiniphila in a benzene-contaminated aquifer by DGGE band sequencing. In addition, Weelink et al. (2007) observed a phylotype affiliated to Z. resiniphila PIV-3A2w enriched in a benzene-degrading, chlorate-reducing mixed culture. Remarkably, sequences affiliated to Zoogloea were only found in low abundance in the 16S rRNA gene library of the total biofilm community (see below). This indicates that metabolic processes other than benzene metabolism may play an important role in the biofilm growing on the geotextiles of the pond system which is likely due to the complex contaminant load providing other carbon substrates.

Additionally, sequences with a high similarity to D. aromatica RCB, F. limneticum cda-1 as well as Hydrogenophaga were found in the heavy fraction and congruently in high abundance in the clone library of the original geotextile material (t0; see discussion below). D. aromatica RCB is a facultative anaerobic bacterium and has been described for its ability to degrade benzene and other aromatic hydrocarbons under nitrate-reducing, chlorate-reducing, or oxic conditions (Chakraborty et al., 2005). Hydrogenophaga- and Rhodoferax-related species are frequent members of communities found in water treatment plants (Kampfer et al., 1993; Lemmer et al., 1997; Fahy et al., 2006) and have been already reported to be highly abundant in a benzene-degrading consortium (Fahy et al., 2006); later a BTX-degrading Hydrogenophaga strain was isolated (Fahy et al., 2008). Phylotypes affiliated to Pirellula sp. or Dechloromonas hortensis MA-1, and Dechloromonas denitrificans ED1T were only found in low abundance in the clone library, and accordingly, the respective T-RFs of 150 and 174 bp, respectively, were only marginally present in the overall T-RFLP profiles. Hence, these phylotypes probably played only a minor role in benzene degradation.

In agreement with the low labeling of RNA observed at t1, no labeled proteins were identified at this time point, possibly due to the small amount of benzene degraded and the dilution of the 13C-benzene by remaining 12C-benzene (Fig. 1). Among the identified proteins at t2 (Table S1), a total of 19 proteins were labeled (Table 2) from which 16 showed 13C-incorporation higher than 90 atom%. This is in good agreement with the observed high labeling of RNA at t2 and confirms the presence of phylotypes that grew by assimilation of benzene-derived carbon.

For nearly all identified 13C-enriched proteins, an identical match was found in the clone libraries at the genus or species-level, for example, proteins related to Daromatica RCB, Rhodoferax sp., and Thauera sp. (Table 2). However, several identified labeled proteins pointed to organisms that could not be detected in the clone libraries, for example Shewanella, Halomonas, and Bradyrhizobium. More important, none of the labeled proteins could be assigned to the genera Zoogloea and Hydrogenophaga, although phylotypes belonging to these taxa were obviously the dominant benzene assimilating bacteria according to RNA-SIP. This is likely caused by limited genetic information available for these genera (VerBerkmoes et al., 2009), as genomes of strains belonging to Zoogloea and Hydrogenophaga have not been sequenced so far. Only 112 and 94 protein entries, respectively, were represented in the NCBI database (May 2012) for Zoogloea and Hydrogenophaga; in contrast, protein entries for Thauera and Dechloromonas were almost two orders of magnitude higher. Thus, we cannot rule out the possibility that some of the proteins derive from Betaproteobacteria were effectively from Zoogloea. Indeed, a major drawback within environmental metaproteomic studies is that protein identifications are limited because the majority of microbial community members represent noncultured and thus nonsequenced phylotypes. However, the phylogenetic classification of proteins found in environmental samples containing a genetic inventory of unknown complexity, can be significantly improved by assembling databases for both the metaproteome and the metagenome (Morris et al., 2010), which will be a guideline in future studies.

Despite its taxonomy drawbacks, a clear benefit of protein-SIP is the functional information which might be gained from the metaproteomics data. In our study, the protein-SIP data suggest that benzene was mainly metabolized by organisms using a dihydroxylation reaction for the initial ring attack and the meta-cleavage pathway for further ring cleavage, as peptides related to toluene dioxygenase and catechol 2,3-dioxygenase were detected and enriched in 13C. In accordance with these results, parallel microcosm experiments using 12C-benzene showed that its degradation was associated with a moderate carbon isotope fractionation (εC: between −0.64 ± 0.08‰ and −1.01 ± 0.09‰), and no significant hydrogen isotope fractionation which is characteristic for dihydroxylation reactions (Fischer et al., 2008; Table S2).

Interestingly, one sequence related to Methylobacter tundripaludum was detected in the heavy RNA fraction, indicating that benzene-derived carbon was slightly assimilated also by organisms affiliated to type I methanotrophic bacteria. Notably, two peptides related to the sMMO of type I methanotrophs showed low 13C-incorporation (22.5 atom% on average, Table 2) demonstrating that the parent organism used diverse carbon sources and only feed to about 1/5 on benzene-derived carbon. In methanotrophs, the sMMO is expressed in the case of environmental copper concentrations below 0.8 μM (Hakemian & Rosenzweig, 2007) which is in agreement with the measured copper concentration of < 0.8 μM in the water from the inflow. Whereas it is known for a long time that sMMO oxidizes a wide range of organic substrates including benzene which is hydroxylated to phenol (Colby et al., 1977; Wilkins et al., 1994), an assimilation of benzene-derived phenol by sMMO-containing methanotrophs is not expected as these organisms appear to be obligate C-1 utilizers (Madigan & Martinko, 2009). Labeled methane was not detected as intermediate in the course of the experiment ruling out that methane derived from benzene was a source of carbon for methanotrophs or affiliated species. Hence, our results indicate that phylotypes affiliated to methanotrophic organisms can actually use other substrates than methane. Notably, a methanotrophic enrichment culture was described to be capable of degrading benzene when formate was added as source of reducing power and energy (Hanson & Hanson, 1996; Murrell et al., 2000; Hesselsoe et al., 2005).

In conclusion, our study shows that the combination of complementary techniques such as RNA-SIP, protein-SIP, and CSIA allows a coherent description of functionalities in a complex biofilm and provided added value to understand the ecology of contaminant degradation in biofilms. The benzene-degrading microbial community was composed of different phylotypes; such diversity might be beneficial for a continuous benzene removal under fluctuating environmental conditions.

Acknowledgements

This work was supported by the Helmholtz Centre for Environmental Research – UFZ in the scope of the SAFIRA II Research Programme (Revitalization of Contaminated Land and Groundwater at Megasites, subproject ‘Compartment Transfer – CoTra’). We thank Ute Lohse for technical assistance in the T-RFLP analyses and sequencing, Matthias Gehre and Ursula Günther for their support during the stable isotope analysis, Francesca Löper for technical assistance in system operation and sampling of textiles, and Stefanie Hinke, Ines Mäusezahl and Kerstin Ethner for laboratory help. F.B. and A.G.F. were supported by Host Fellowships for the Transfer of Knowledge (ToK) in the framework of the ISOTONIC project (MTKD-CT-2006-042758), and M.R. by a Beatriu de Pinós postdoctoral grant (2008 BP-A 00054) from the Autonomous Government of Catalonia (Agència de Gestió d'Ajuts Universitaris i de Recerca, AGAUR).

Author contribution

S.J. and A.G.F. contributed equally to this work.

Ancillary