Dominance of Geobacteraceae in BTX-degrading enrichments from an iron-reducing aquifer


  • Sabrina Botton,

    1. Earth Surface Processes and Materials Department, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Marijn Van Harmelen,

    1. Earth Surface Processes and Materials Department, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Martin Braster,

    1. Molecular Cell Physiology, Faculty of Earth and Life Sciences, Vrije Universiteit, De Boelelaan, Amsterdam, The Netherlands
    Search for more papers by this author
  • John R. Parsons,

    1. Earth Surface Processes and Materials Department, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Amsterdam, The Netherlands
    Search for more papers by this author
  • Wilfred F.M. Röling

    1. Molecular Cell Physiology, Faculty of Earth and Life Sciences, Vrije Universiteit, De Boelelaan, Amsterdam, The Netherlands
    Search for more papers by this author

  • Editor: Alfons Stams

Correspondence: Sabrina Botton, Earth Surface Processes and Materials Department, IBED – Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands. Tel.: +31 0 20 525 6566; fax: +31 0 20 525 7431; e-mail:


Microbial community structure was linked to degradation potential in benzene-, toluene- or xylene- (BTX) degrading, iron-reducing enrichments derived from an iron-reducing aquifer polluted with landfill leachate. Enrichments were characterized using 16S rRNA gene-based analysis, targeting of the benzylsuccinate synthase-encoding bssA gene and phospholipid fatty acid (PLFA) profiling in combination with tracking of labelled substrate. 16S rRNA gene analysis indicated the dominance of Geobacteraceae, and one phylotype in particular, in all enrichments inoculated with polluted aquifer material. Upon cultivation, progressively higher degradation rates with a concomitant decrease in species richness occurred in all primary incubations and successive enrichments. Yet, the same Geobacteraceae phylotype remained common and dominant, indicating its involvement in BTX degradation. However, the bssA gene sequences in BTX degrading enrichments differed considerably from those of Geobacter isolates, suggesting that the first steps of toluene, but also benzene and xylene oxidation, are carried out by another member of the enrichments. Therefore, BTX would be synthrophically degraded by a bacterial consortium in which Geobacteraceae utilized intermediate metabolites. PLFA analysis in combination with 13C-toluene indicated that the enriched Geobacteraceae were assimilating carbon originally present in toluene. Combined with previous studies, this research suggests that Geobacteraceae play a key role in the natural attenuation of each BTX compound in situ.


Benzene, toluene, ethylbenzene and xylenes (BTEX) are a group of monoaromatic petroleum-derived pollutants often found in contaminated aquifers as a consequence of illegal discharges or losses during spills. The response of indigenous microbial communities towards chemical pollutants entering the subsurface represents an important parameter for assessing the potential effects of contamination on the surrounding environment.

Natural attenuation of BTEX under anaerobic, iron-reducing conditions has been reported previously for the contaminated aquifer located downgradient of the Banisveld landfill, Boxtel, the Netherlands (van Breukelen et al., 2003; van Breukelen & Griffioen, 2004). Cultivation-independent analysis of microbial communities present in this aquifer indicated a strong dominance of members of the Geobacteraceae (Röling et al., 2001; Lin et al., 2005). Particularly, the presence of one dominantly occurring phylotype related to the occurrence of natural attenuation of organic pollutants (Röling et al., 2001; Lin et al., 2005), suggesting the involvement of this microorganism in biodegradation.

Members of the family Geobacteraceae are also the only iron-reducers capable of BTEX degradation described so far (Lovley et al., 1993, 2004; Coates et al., 2001). Their occurrence was mostly observed at petroleum contaminated sites, where a direct link could be established between the presence of these bacteria and high concentrations of petroleum-derived pollutants, such as the BTEX contaminants (Anderson et al., 1998; Rooney-Varga et al., 1999). In contrast, landfill leachate contains mainly dissolved organic matter (DOC), but also inorganic macrocomponents, heavy metals and xenobiotic organic compounds, which in general constitute a minor component of the DOC (Christensen et al., 2001). BTEX compounds (<221 μg L−1) represent only 0.1% of DOC in leachate from the Banisveld landfill (van Breukelen et al., 2003). Therefore, a direct relationship between phylogeny data, indicating the dominance of Geobacteraceae in the polluted aquifer (Röling et al., 2001; Lin et al., 2005), and in situ BTEX degradation has still not been ascertained. The strong presence of Geobacteraceae may well solely be due to growth on the nonhydrocarbon fraction of DOC.

Previously, the microbial potential towards BTEX degradation had been assessed at the Banisveld site by means of microcosms established with aquifer material. Enrichments capable of complete degradation of benzene, toluene or xylene (BTX) were obtained under iron-reducing conditions (Botton & Parsons, 2006, 2007). In the present study, the aim was to improve one's understanding of the natural attenuation processes occurring within the Banisveld iron-reducing aquifer, and under iron-reducing conditions in general, by linking function – here, the ability to degrade the individual BTX compounds – with microbial community structure in these previously established enrichments. The BTX-degrading communities were characterized in terms of phospholipid fatty acids (PLFA) (Rooney-Varga et al., 1999; Green & Scow, 2000) and 16S rRNA genes-based denaturing gradient gel electrophoresis (DGGE) patterns. PLFA analysis represents an effective tool to characterize microbial communities in terms of physiology and identity, for example by tracking the carbon source of interest into the growing biomass of specific microorganisms (Green & Scow, 2000; Boschker & Middelburg, 2002). 16S rRNA gene data provide information about the individual members present in the BTX degrading-consortia, not just the microorganisms directly degrading the BTX, but also bacteria performing other functions (e.g. removal of intermediates).

In addition, a functional gene was addressed in relation to the specific metabolic features of the enriched cultures, namely the ability to degrade BTX: the bssA gene, which codes for benzylsuccinate synthase required for the first step of the anaerobic degradation of toluene and xylenes (Biegert et al., 1996; Beller & Spormann, 1999; Achong et al., 2001; Kane et al., 2002). The metabolic pathway that bacteria utilize to gain energy from toluene and xylene under anaerobic conditions appears to be well conserved: benzylsuccinate is formed by the addition of fumarate to toluene by benzylsuccinate synthase and then undergoes further transformation to benzoyl-CoA and finally to CO2, as described for many species (Biegert et al., 1996; Beller & Spormann, 1997; Beller & Edwards, 2000), including iron-reducing bacteria (Kane et al., 2002; Botton & Parsons, 2007). Similarly, xylenes are transformed into their corresponding methylbenzylsuccinates (Beller et al., 1995; Krieger et al., 1999; Beller & Edwards, 2000; Reusser et al., 2002; Botton & Parsons, 2007). In contrast, benzene degradation by the iron-reducing bacteria enriched from Banisveld aquifer has been proven to occur via phenol as an intermediate (Botton & Parsons, 2007).

Materials and methods

Microcosms and enrichment cultures

Enrichment cultures were inoculated with polluted groundwater or sediment collected from the contaminated aquifer located downstream of the Banisveld landfill, Boxtel, the Netherlands (Fig. 1).

Figure 1.

 Schematic overview of the investigated polluted aquifer downgradient from the Banisveld landfill. The arrows indicate the positions from which polluted groundwater (GW) and sediment (Sed) were collected and used as inoculum in primary enrichment cultures. Light grey in the picture indicates the position of the plume of pollution (van Breukelen et al., 2003).

The aquifer is characterized by iron-reducing conditions and an exhaustive description of the sampling site was published previously in van Breukelen et al. (2003) and van Breukelen & Griffioen (2004). Sampling procedure and methodology for establishing primary enrichment cultures (microcosms) can be found elsewhere (Botton & Parsons, 2006); secondary enrichments were obtained by transferring 5–10 mL of liquid of the parent microcosms into 40 mL of fresh anaerobic medium as described in Botton & Parsons (2007).

Briefly, the various microcosms under investigation were inoculated with (for sampling locations, see Fig. 1): (1) polluted groundwater (GW, 75 mL) or sediment (Sed, 15–50 g), and amended with one of the BTX compounds (20–100 μM as the final concentration in the liquid phase; see Botton & Parsons, (2006) for details) as the sole carbon source; (2) polluted sediment as inoculum but without any BTX added; and (3) unpolluted sediment and either spiked with one of the BTX compounds or left unamended. Each enrichment was provided with 10 mM of amorphous iron-oxyhydroxides as the sole electron acceptor, prepared by neutralizing FeCl3 with NaOH until pH=7 (Lovley & Lonergan, 1990), and electron donor and acceptor were added again upon depletion. Microcosms were established with groundwater or sediment as inoculum, anaerobic sterile growth medium with carbonate buffer (30 mM, final pH=6.7), trace elements and vitamin solutions. Enrichments were spiked at least twice with one of the BTX compounds and the anaerobic incubations were kept at room temperature (20±2 °C).

Degradation of benzene, toluene and each of the xylene isomers was observed in enrichment cultures incubated with polluted sediment whereas no contaminant removal was detected when unpolluted sediment was used as the inoculum. Trends of contaminant removal coupled to Fe(III) reduction in these first-generation cultures were reported in Fig. 2 of (Botton & Parsons, 2006), whereas degradation trends of subsequent second- and third-generation enrichments are depicted in Fig. 1 and summarized in Table 1 of Botton & Parsons (2007). Samples for molecular analysis (this study) and chemical analysis (Botton & Parsons, 2007) in order to determine the degradation rates, amount of contaminant removed and electron balances reported here were obtained at the same time. The corresponding days of incubation are indicated in Fig. 2. Some chemical data, e.g. the rate of degradation at the time of sampling, are summarized in Table 2 for the convenience of the reader, and a schematic overview of the enrichments used in this study is depicted in Fig. 5.

Figure 2.

 Cluster analysis of DGGE fingerprints of bacterial communities in benzene (a)-, toluene (b)- and xylene (c)-degrading enrichments obtained with groundwater (GW) or sediment (Sed) as inoculum. Numbers 1, 2 and 3 designate primary, secondary and tertiary enrichments, respectively, capital letters refer to different sets of enrichments and small letters indicate duplicates obtained from the same parent culture. Days of incubation refer to the moment in which samples for molecular analyses were withdrawn from the enrichments.

Table 1.   DNA sequence homology between bssA gene sequences, encoding the α subunit of benzylsuccinate synthase, of anaerobic toluene-degrading strains, at the positions targeted by degenerate primers used in this study to amplify bssA genes
(Genbank Accession no.)
Forward primer position
(bp 1772-1790)
Reverse primer position
(bp 2269-2285)
  1. Y is C or T, R is A or G, nucleotides indicated in bold and underlined refer to mismatches with primers. The primer design was performed based on alignment of bssA sequences available in GenBank on 18 November 2003, the Azoarcus EbN1 and Magnetospirillum sequences were released by GenBank after this data.

Azoarcus sp. T (AY032676)----C--C-------------------G--------
Azoarcus sp. EbN1 (CR55380)----C--C----------- --------A--T-----
Thauera aromatica K172 (AJ001848)----C--C-------------------G--------
Thauera sp.DNT-1 (AB066263)----C--C-------------------G--------
Thauera aromatica T1(AF113168)----C--C-------------------G--------
Magnetospirillum sp. TS-6 (AB167725)----C--C--G----------T-----G--------
G. metallireducens (AF441130)----T--T-------------------A--------
Table 2.   Variation in degradation rates and richness index in primary and secondary enrichments originally inoculated with polluted groundwater or sediment
Primary enrichments (En. 1)Secondary enrichments (En. 2)
(μM) (number
of spikes)
rates (μM day−1)
(μM) (number
of spikes)
rates (1)
(μM day−1)
  • Values represent averages of three cultures grown under the same conditions and with one of the BTX compounds as sole carbon source.

  • Concentration degraded and average degradation rates were measured at the moment in which samples for molecular analyses were withdrawn from active enrichments. (no. of spikes) indicates the number of times cultures were amended with one of the BTX compounds.

  • *

    Values calculated as averages of second- and third-generation enrichments.

Groundwater (GW)
 Benzene68.5 (2)0.29.5 ± 0.7   
 Toluene141.0 (3)0.614.066.8 (4)1.711.0 ± 2.8
Sediment (Sed)
 Benzene29.2 (2)0.113.3 ± 2.915.0 (2)0.26.2 ± 2.5
 Toluene190.0 (3)0.412.3 ± 1.8344.0 (3)*3.0*9.2 ± 3.0*
 ortho-xylene101.0 (2)0.16.67 ± 2.950.7 (1)0.77.0
 meta-Xylene154.0 (3)0.46.0470.0 (3)3.64.3 ± 0.3
 para-Xylene44.0 (2)0.214.0 ± 4.435.3 (1)0.23.0 ± 1.4
Figure 5.

 Schematic overview of the enrichments used in this study. Polluted groundwater (GW) and polluted sediment of clean sediment were used as inoculum to establish primary enrichments (first-generation microcosms). The second generation of enrichment cultures was obtained by subsequent transfer of the liquid fraction (5–10 mL) of actively BTX degrading first-generation enrichments. With the same procedure, third-generation enrichments were also established. Each enrichment culture was monitored at least in duplicate and compared with sterile controls.

Molecular analysis

Molecular analysis of BTX-degrading cultures was carried out using, as a template for PCR amplification, 1 μL of: (1) a well-mixed bacterial suspension from sediment-free enrichments or (2) DNA extracted from 1 mL of microcosm incubations with Bio 101 FastDNA® SPIN Kit for soil according to the manufacturer's instructions. Bacterial and Geobacteraceae 16S rRNA gene fragments were amplified with a nested approach, in order to obtain fingerprints of the overall bacteria community and the Geobacteraceae community, respectively. The comparison between direct and nested amplification methods using general bacterial primers showed that their corresponding DGGE profiles presented the same bands. With the direct amplification, however, the product yield was sometimes lower than with the nested method; hence the latter was applied for all samples. In the first round of amplification, in order to obtain bacteria-specific 16S rRNA gene fragments, bacterial primers F8 (Felske et al., 1997) and R518 (Muyzer et al., 2003), corresponding, respectively, to positions 8–36 and 534–517 in the Escherichia coli 16S rRNA gene, were added (0.4 μM each) to a 25 μL reaction vial along with 0.4 mM dNTPs, 10 μg bovine serum albumin (Biolabs, UK) and 2.5 U Taq polymerase (MRC Holland, the Netherlands). Amplification was performed according to the following programme: 94 °C for 4 min, followed by 35 cycles of 94 °C for 0.5 min, 54 °C for 1 min and 72 °C for 1 min, with a final elongation phase at 72 °C for 5 min.

A Geobacteraceae-specific PCR (Lin et al., 2005) amplified a 0.8-kb 16S rRNA gene fragment with F8 (Felske et al., 1998) and R825 (corresponding to 825–841 in E. coli 16S rRNA gene) (Snoeyenbos-West et al., 2000) primers, using the following programme: 94 °C for 4 min, and then touch-town primer annealing from 65 to 56 °C (decreasing 1 °C per two cycles), followed by 15 cycles at 55 °C annealing temperature and a final elongation at 72 °C for 5 min. Diluted products (1/100) of the bacteria or Geobacteraceae-specific first-round amplification were used as templates and amplified with primers F357-GC (corresponding to positions 341–358 in the E. coli 16S rRNA gene) and R518 (Muyzer et al., 2003) in the second round of PCR.

DGGE was carried out with the Bio-Rad DCode system. PCR products (20 μL) were loaded onto 1 mm thick 8% (w/v) polyacrylamide (37.5 : 1 acrylamide : bisacrylamide) gels containing a 30–55% linear denaturing gradient, where 100% denaturant is defined as 7 M urea and 40% (v/v) formamide. Gels were run in 1 × TAE buffer (40 mM Tris-acetate, 1 mM Na-EDTA, pH 8.0) at 200 V for 4 h and further stained in 1 × TAE buffer containing 1 μg mL−1 ethidium bromide and recorded with a CCD camera system (The imager, Appligen, Illkirch, France).

Gel images were converted, normalized and analysed by the gelcompar II software package (Applied Maths, Kortrijk, Belgium). A marker consisting of 12 clones was included in DGGE profiles in order to facilitate the conversion and normalization of gel images, after which DGGE profiles were compared using a band assignment-independent method (Pearson product–moment correlation coefficient). Richness (number of species) in 16S rRNA genes was calculated after band-based analysis of gel images.

Sequence data were obtained from DGGE bands by excising the appropriate bands from the DGGE gels and reamplifying them with primers F357-GC and R518 (Muyzer et al., 2003). 16S rRNA gene fragments from the DGGE analysis were sequenced using F357 and R518 (Muyzer et al., 2003).

In order to design specific primers for the bssA gene, all bssA gene sequences available in GenBank (October 2003) were aligned (Table 1). Based on well-conserved parts in the alignment, two degenerate primers, bssA_f (TCGA(C/T)GA(C/T)GGCTGCATGGA) and bssA_r (TTCTGGTT(T/C)TTCTGCAC), were formulated. These primers were used to amplify a 0.57-kb fragment, using the same PCR conditions as described above. Amplification was performed according to the following programme: 94 °C for 4 min, followed by 35 cycles of 94 °C for 0.5 min, 44 °C for 1 min and 72 °C for 1 min, with a final elongation phase at 72 °C for 5 min. Most-probable number PCR (MPN-PCR), to determine the detection limit, was performed in triplicate on aliquots of serial 10-fold dilutions of a DNA extract with a known amount of bssA gene copies. bssA gene copies were calculated after quantifying the amount of extracted DNA using a molecular size marker (Precision molecular mass standard; Biorad, Hercules, CA), and assuming a genome size of about 4 Mb and a single bssA gene copy per genome (Rabus et al., 2005). Phylogenetic analysis of the obtained sequences was carried out by comparing sequencing data with GenBank database and using the blast algorithm to attain the most closely related sequences (Altschul et al., 1990). The bssA gene sequences were aligned using clustalw, distance analysis with the Jukes–Cantor correction and bootstrap resampling (100 times) were carried out with the treecon package (van de Peer & de Wachter, 1994) and the distance matrix was used to construct a tree by neighbour joining (Saitou & Nei, 1986).

PLFA analysis

PLFA analysis was carried out on second-generation toluene-degrading cultures enriched from polluted groundwater and on a pure culture of Geobacter metallireducens (DSM7210). Geobacter metallireducens was cultivated under the same conditions as the enrichments: growth medium, electron acceptor and donor (toluene) and incubation temperature were identical for both cultures to avoid influence of incubation conditions on the final PLFA pattern (Petersen & Klug, 1994). Samples (c. 8 mL) were taken at regular intervals during an incubation time of 100 days. After 80 μM of unlabelled toluene had been degraded (day 20, T0) the first sample was taken and cultures were further amended with 80 μM of 13C1-toluene, labelled on the methyl group. Upon depletion, labelled toluene was readded to the cultures: 90 and 200 μM were added at days 39 and 48, respectively. During the incubation, three more samples were withdrawn from the cultures for PLFA analysis (days 29, 57 and 90, with the latter corresponding to Tend).

The total lipids were extracted with chloroform and divided into fractions of different polarity on a silicic acid column (Guckert et al., 1985). The methanol fraction, containing the polar lipids (PLFA), was then derivatized and the correspondent fatty acid methyl esters (FAMEs) were analysed by GC-MS. Dimethyl disulphide derivatization was used to determine the double-bond position of fatty acid methyl esters (Nichols et al., 1986) and the quantification of the FAMEs was based on the C19 : 0 internal standard.

The GC-MS analyses were performed with a Thermo Quest Trace GC equipped with a 30 m DB-5 column (0.32 mm internal diameter; 0.25 μm film thickness) coupled to a Finnigan Trace quadrupole mass spectrometer (MS). Two microlitres of samples were injected on-column (ThermoQuest AS 2000 Autosampler) and the following temperature programme was applied: 80 °C for 1 min, 4 up to 220 °C and 15 °C min−1 until 300 °C.

Nucleotide sequence accession numbers

Nucleotide sequences of the bssA gene have been deposited in the GenBank database under accession numbers EF190465 and EF190466.


Molecular analysis of BTX-degrading cultures

Samples withdrawn from enrichment cultures over time, either within the same microcosm but also in consecutive generations further obtained, were profiled by DGGE of amplified 16S rRNA gene fragments. In order to detect variations in overall bacteria community composition in relation to the exposure and adaptation of the cultures towards a specific pollutant, samples were grouped on the basis of the monoaromatic hydrocarbon being degraded; see Table 2. A schematic overview of the enrichments used in this study is depicted in Fig. 5.

Molecular fingerprinting of benzene-degrading cultures

Cluster analysis of fingerprints of the bacterial communities in benzene-degrading cultures (Fig. 2a) shows that enrichments derived from groundwater (GW in Fig. 2) grouped separately from enrichments derived from sediment (Sed in Fig. 2) – similarity lower than 20%– indicating the presence of diverse bacterial communities in the two systems. In addition, a distinctive band (band 1 in Fig. 2a) was present in all incubations and dominated the DGGE profile for each benzene-degrading sediment enrichment. DGGE profiling after amplification with Geobacteraceae-specific primers revealed a band with similar mobility to band 1 in Fig. 2 (data not shown). Excision of this band and sequencing indicated that the band was derived from a member of the Geobacteraceae, its sequence being identical to five sequences in GenBank that are closest related to Geobacter spp. (accession numbers: AY013634, AY752749, AY752751, AY752753, AY752762), including the dominant phylotype detected previously using culture-independent approaches in polluted groundwater samples obtained from the aquifer neighbouring the Banisveld landfill (Röling et al., 2001; Lin et al., 2005). Almost all DGGE profiles revealed a second intense band (band 2 in Fig. 2a); however, a clear sequence for it could not be obtained. Close to band 1, almost overlapping it, another band could be observed in the profile corresponding to the first-generation culture derived from sediment (Sed 1 in Fig. 2a) after 345 days of incubation, band 3 in Fig. 2. As for band 1, Geobacteraceae-specific profiling revealed a band with equal mobility to band 3 in DGGE and sequencing indicated 99% similarity to the same uncultured Geobacter sp. mentioned above.

Primary enrichments were characterized by a more diverse community than the secondary enrichments and the decrease in number species in secondary enrichments, as indicated by the richness index reported in Table 2, corresponded to an increased benzene degradation rate.

Molecular fingerprinting of toluene-degrading cultures

DGGE profiles of toluene-degrading enrichments (Fig. 2b) were also strongly dominated by the same intense Geobacteraceae bands – bands 1 and 3 – as observed for benzene-oxidizing enrichments (Fig. 2a) inoculated with polluted sediment. Also, the DGGE patterns observed for groundwater (GW) and sediment (Sed) samples were again indicative of dissimilar communities (similarity<20%) being enriched, although in both cases variation over time in community structure was observed (Fig. 2b, Table 2). The richness index decreased over a time span of 1 year in first-generation enrichments (Sed 1) and the same trend could be observed in successive cultures (Sed 2, 3). For example, in one of the sediment cultures (Sed 1-B), after c. 200 days of incubation, richness was equal to 13.7±1.5 and 6 months later decreased to 11.0±0.0. In secondary enrichments (Sed 2-B), a value of 12.0 – measured at the beginning of the incubation – declined to 6.3±1.2 when cultures were further transferred to fresh medium (Sed 3-B). As in the case of benzene-degrading enrichments, the decrease in the number of species was concomitant with a steady increase in the average toluene degradation rate.

Second-generation enrichments were still dominated by Geobacteraceae (band 1, Fig. 2b) but also other bands seemed to be characteristic for toluene- and xylene-degrading systems, i.e. bands 4 and 6 in Fig 2. The sequencing of band 4 revealed 97% similarity to Desulphomonile sp. (accession number: AY167452) whereas band 6 had 97% similarity to an uncultured δ proteobacterium (accession number: AF529133).

Third-generation enrichments in which band 5 was detected were amplified with Geobacteraceae-specific primers and DGGE analysis indicated a band with a mobility similar to band 5 of Fig. 2b (data not shown). Indeed, sequencing of band 5 in Fig. 2b indicated 96% similarity to Geobacter argillaceus (accession number: DQ145534) and an uncultured Geobacter sp. (accession number: AY752759) – with the latter sequence being previously detected in the aquifer downstream of the landfill (Lin et al., 2005).

Molecular fingerprinting of xylene-degrading cultures

Cluster analysis indicated that the DGGE profiles of xylene-degrading enrichments generally grouped isomer specifically: enrichments in which meta-xylene was removed formed one cluster (top of Fig. 2c), whereas ortho-xylene-degrading incubations formed a distinct cluster (bottom of Fig. 2c). However, the profiles obtained from para-xylene-degrading enrichments did not gather into a single group but were instead more dispersed. Richness decreased in each secondary enrichment (Table 2), with the only exception being for the ortho-xylene-degrading enrichment for which, however, the ability to degrade this pollutant was maintained only in one of the second-generation transfers (ortho-xyl 2-C in Fig. 2c).

DGGE profiles of enrichments revealed the same intense Geobacteraceae bands (bands 1 and 3 in Fig. 2c) as detected in the DGGE profiles derived from benzene- and toluene-degrading cultures (Fig. 2a and b). Band 2, the Desulfomonile-like (band 4) and Delta-proteobacterium-like (band 6) sequences previously observed in toluene-degrading consortia, were again observed in DGGE profiles of enrichments in which xylene was removed.

Comparison with molecular fingerprints of unamended microcosms and microcosms derived from unpolluted sediment

Microcosms established with unpolluted (or clean) sediment were either amended with one of the BTX compounds or left unamended and incubated under the same conditions as the primary enrichments based on polluted sediment. Molecular analysis, after c. 200 days of incubation during which no degradation could be observed (Botton & Parsons, 2007), showed that DGGE profiles obtained from amended and unamended microcosms initiated with clean sediment presented identical DGGE profiles. These profiles derived from microcosms established with unpolluted sediment differed from actively BTX-degrading enrichments obtained from polluted sediment (data not shown). Only one common band (band 4) corresponding to Desulfomonile could be detected.

In contrast, DGGE profiles obtained from enrichments inoculated with polluted sediment that were not spiked with BTX compounds revealed the same intense Geobacteraceae bands (bands 1 and 3 in Fig. 2) – already reported in Lin et al. (2005) after culture-independent analysis of polluted sediment – as well as band 4.

Diagnostics of the gene encoding benzylsuccinate synthase

The set of primers designed to amplify bssA gene fragments (Table 1) was tested on and gave positive amplification with the denitrifying bacteria Thauera aromatica (DSM 6984) and Azoarcus sp. T (DSM 9506) and with the iron-reducing bacteria G. metallireducens (DSM 7210) and Geobacter grbiciae (DSM 13690). No amplification was observed with DNA from non-toluene-degrading E. coli and Paracoccus denitrificans. MPN-PCR revealed that a single copy per PCR reaction could be detected.

Positive amplification of bssA gene fragments was obtained for all toluene- and xylene-degrading enrichments but also for first-generation enrichments from polluted sediment not amended with BTX. In contrast, the bssA gene could not be detected in samples of secondary enrichments exposed to benzene and enrichments initiated with unpolluted sediment despite the high sensitivity of the MPN-PCR method. Sequences for PCR products obtained from toluene-degrading enrichments (GW 3-A-a and -b in Fig. 2b) were compared with bssA gene sequences of several toluene-degrading isolates deposited in GenBank (Fig. 3).

Figure 3.

 Phylogenetic relationships between the nucleotide sequences of bssA genes encoding the α subunit of benzylsuccinate synthase obtained from anaerobic toluene-degrading enrichments, derived from the aquifer polluted by the Banisveld landfill, and pure cultures of toluene-degrading bacteria. Only bootstrap values above 50% are shown.

The DNA sequences of bssA gene fragment of iron-reducing toluene-degrading cultures enriched from the polluted aquifer were clearly separated (60% similarity) from G. metallireducens (accession number: AF441130), a bacterium known to degrade toluene under iron-reducing conditions (Kane et al., 2002). A closer relationship was observed with toluene-oxidizing denitrifiers, a 75–80% similarity to the cluster consisting of Thauera sp. DNT, Thauera aromatica T1 and Azoarcus sp. T (Fig. 3), and between 70% to 75% similarity with a second group of denitrifying bacteria, consisting of T. aromatica and Azoarcus sp. EbN1 (Fig. 3). However, phylotypes relating to Thauera or Azoarcus were not retrieved in the enrichments by 16S rRNA gene-based analysis. Also, xylene-degrading consortia were found to harbour a bssA gene highly similar (>95%) to those observed in the investigated toluene-degrading microcosms (data not shown).

PLFA-based community analysis and tracking of 13C-labelled toluene

PLFA-based community analysis was used in this study as an additional screening tool, independent of the 16S rRNA gene amplification-based profiling, which can produce biased results (von Wintzingerode et al., 1997). PLFA profiling was conducted on second-generation cultures of toluene-degrading microcosms. After 100 days of incubation, during which c. 400 μM toluene were degraded with iron as a terminal electron acceptor, PLFA concentrations were compared with the PLFA profile of a similarly incubated pure culture of G. metallireducens, Fig. 4a, as the 16S rRNA gene-based analysis suggested a dominance of microbial communities by members of the Geobacteraceae.

Figure 4.

 PLFA profiles of Geobacter metallireducens and enrichment cultures growing on toluene under iron-reducing conditions (a). Increase in time of the most abundant PLFAs in G. metallireducens (b) and enrichment cultures (c). The four colours in B and C represent the four samplings at days 20, 29, 57 and 90. Fatty acids are designated as A: BwC, where A is the total number of carbon atoms, B the number of double bonds and C the position of the double bond from the methyl end of the fatty acids. The suffixes c and t stand for cis and trans, respectively.

The PLFA profiles obtained were highly similar (rank correlation factor ρ=0.999, P>0.05) and in both cases the PLFAs C:16 1w7c, C:16 and C:18 1w7c represent more than 70% of the detected fatty acids, consistent with previous studies (Lovley et al., 1993; Ludvigsen et al., 1999; Zhang et al., 2003). Combined with the molecular data, this finding indicates that Geobacteraceae are dominantly present and active in the enrichment cultures.

After the first aliquot of toluene was degraded (day 20), 13C1-toluene was supplied to the cultures in order to trace the 13C-labelled substrate in the growing biomass. PLFAs were extracted and the increase in their concentrations over time, as depicted in Fig. 4b and c, indicated that biomass increased during the incubation.

In addition, at the end of the incubation period of 100 days the mass spectra of the most abundant PLFAs were analysed in order to evaluate whether 13C was assimilated into the cell membrane of toluene degraders. The analysis of the mass spectra revealed that after the addition of labelled toluene, the PLFA molecules were enriched in heavier carbon atoms (m/z+1, +2) whereas a decrease in relative abundance of the original unlabelled PLFAs was observed (Table 3).

Table 3.   Relative abundances of unlabelled and 13C-labelled PLFAs after growth on labelled toluene
PLFA (2) Difference (TendT0) in relative abundance
of PLFAs (1)
  1. (1) The comparison was calculated taking into account the relative ratios of the four most abundant fragments in each PLFA mass spectrum at the beginning of incubation (T0, day 20) with unlabelled toluene and after the addition and degradation of labelled substrate (Tend, day 90). Negative or positive values indicate a decrease or increase in relative abundance respectively.

  2. (2) Fatty acids designation as in Fig. 4.

C:14m/z=242−1.38 ± 0.31%−2.89 ± 1.05%
m/z+11.29 ± 0.35%0.58 ± 0.49%
m/z+20.27 ± 0.07%3.55 ± 1.08%
C:16 1w7m/z=268−5.68 ± 0.03%−4.51 ± 0.98%
m/z+14.85 ± 1.01%1.62 ± 1.00%
m/z+21.73 ± 1.04%2.89 ± 2.42%
C:16m/z=270−1.13 ± 0.89%−1.29 ± 0.57%
m/z+10.43 ± 0.33%0.40 ± 0.62%
m/z+20.57 ± 0.51%0.28 ± 0.09%
C:18 1w7cm/z=296−2.85 ± 0.87%−3.52 ± 0.57%
m/z+11.93 ± 0.54%3.39 ± 0.75%
m/z+21.18 ± 0.29%1.13 ± 0.25%

This finding proved the incorporation of 13C in the growing biomass and therefore shows that the removal of toluene is linked to a culture that showed high similarity in PLFA profiles to G. metallireducens.


16S rRNA gene-based analysis, supported by PFLA analysis, revealed that a member of the Geobacteraceae was dominant and active in most BTX-degrading consortia obtained from the polluted aquifer surrounding the Banisveld landfill. The strong dominance of this phylotype in the actively BTX-degrading incubations, as well as it consistent presence after transfer to sediment-free cultures, suggests a link between this species and the ability to oxidize monoaromatic hydrocarbons under iron-reducing conditions, as already reported for other Geobacteraceae species (Lovley et al., 1993; Coates et al., 1996; Anderson et al., 1998; Rooney-Varga et al., 1999). In addition, the sequence of the 16S rRNA gene of the dominant Geobacteraceae in the microcosms was identical to the Geobacter phylotype previously found to be dominant and wide spread in a culture-independent molecular study on the iron-reducing, polluted part of the Banisveld aquifer (Röling et al., 2001; Lin et al., 2005). The presence and distribution of this phylotype in the polluted aquifer correlated with the relatively high concentrations of organic matter and micropollutants that decreased with distance from the landfill (Lin et al., 2005). However, up to now, due to the low BTEX concentration in groundwater in comparison with the total DOC, the presence of Geobacteraceae could not be directly related to the occurrence of BTEX in situ degradation. Combined with these previous studies, the present research suggests that Geobacteraceae play a key role in natural attenuation of BTX compounds in the plume of pollution downstream of the Banisveld aquifer.

The question is whether this phylotype actually oxidizes BTX directly or facilitates BTX degradation by utilizing intermediate degradation metabolites. Phylogenetic analysis of the bssA gene indicates that the sequences derived from the toluene- and xylene-degrading enrichments were significantly different (60% similarity) from that of G. metallireducens. The bssA gene fragments were more closely related (75–80% similarity) to denitrifying bacteria, e.g. Thauera and Azoarcus species. Therefore, bssA gene-based analysis seems to indicate that Geobacteraceae, although dominant in all the enrichment cultures, are not involved in the first step of toluene and xylene oxidation pathways, namely the attack of the methyl group by addition of a fumarate molecule catalysed by benzylsuccinate synthase. However, it appears that the toluene degraders in the enrichments also did not belong to Thauera and Azoarcus, because the bssA sequences clustered differently from those of Thauera and Azoarcus, and phylotypes most closely related to Thauera and Azoarcus species were not observed for any of the enrichments by 16S rRNA gene-based analysis. Consistently, the fact that bssA could not be detected in the benzene-degrading enrichments despite the dominance of the Geobacteraceae in these enrichments further supports the indirect role of this bacterial family in benzene but also toluene and xylene removal. Either benzylsuccinate or mebenzylsuccinate was previously detected as degradation metabolite in toluene- and meta-xylene-degrading cultures, respectively (Botton & Parsons, 2007), indicating that the enzymatic reaction of addition of fumarate to form the correspondent benzylsuccinate was indeed active. In addition, in the same work (Botton & Parsons, 2007), (13C-labelled) phenol was identified as a putative metabolite of (13C-labelled) benzene in benzene-degrading enrichments, whereas benzylsuccinate was not detected. The fact that bssA could not be retrieved in a benzene-degrading enrichment would therefore support the metabolic pathway hypothesized previously.

Synthropic degradation of BTX in the enrichments may also explain why, despite several attempts, no toluene- or benzene-degrading isolates could be obtained in pure culture from the enrichments. Isolates capable of toluene oxidation under nitrate-, sulphate- and iron-reducing conditions have been described (Lovley et al., 1989; Lovley & Lonergan, 1990; Beller et al., 1996; Achong et al., 2001; Coates et al., 2001), but in some cases syntrophic interactions are required to degrade toluene. For instance, the characterization of a methanogenic consortium showed that none of its individual members was capable of complete mineralization of toluene (Ficker et al., 1999). Syntrophic toluene degradation with fumarate as the sole terminal electron acceptor has been demonstrated in the laboratory for a consortium consisting of the toluene-oxidizing, nonfumarate-reducing G. metallireducens and a non-toluene-oxidizing, fumarate-reducing bacterium (Meckenstock, 1999).

Based on thermodynamic calculations (data not shown), the occurrence of syntrophic interactions in which Geobacteraceae would utilize degradation by-products formed by other bacteria present in the toluene-degrading enrichments would be possible, given the low concentrations of intermediate metabolites, such as acetate and hydrogen.

An alternative explanation for the results might be the horizontal transfer of bssA genes to Geobacteraceae, in which case the dominant Geobacteraceae phylotype could still directly and completely degrade BTX compounds. Previously, Song & Ward (2005) have suggested the occurrence of a horizontal transfer for another gene involved in anaerobic degradation of monoaromatics: the bcrA gene-encoding benzyol-CoA reductase. More recently, (Winderl et al., 2007) have proposed lateral bssA gene transfer among anaerobic toluene degraders based on the phylogenetic comparison of their ribosomal and bssA marker genes. Also, the clustering of Azoarcus and Thauera bssA genes into two separate groups, each containing both Azoarcus- and Thauera-derived bssA genes, suggests horizontal transfer. Still, the sequences of the three currently described bssA genes of Geobacter species (Winderl et al., 2007) are more than 80% identical and do not indicate horizontal transfer of the bssA gene to Geobacter species. Also, the lack of the bssA gene in benzene-degrading enrichments, where Geobacteraceae are also dominant, cannot be explained by this mechanism but rather supports the hypothesis of an indirect role of Geobacteraceae in BTX removal in the enrichments. Alternatively, the Geobacteraceae might be functionally diverse, i.e. only some of its strains possess and express the genetic potential to degrade a particular BTX compound or their intermediate metabolites. Such observations have been made for other species; Azoarcus isolates harbouring identical 16S rRNA genes were recently shown to differ in their ability to oxidize benzene under denitrifying conditions (Kasai et al., 2006).

The assimilation of different carbon sources was found to induce changes in bacterial PLFA patterns of a pure culture only when chemically dissimilar compounds were utilized as a growth substrate (Wick et al., 2003). However, this appears not to be the case for toluene or acetate. In fact, the utilization of toluene or acetate (Lovley et al., 1993; Zhang et al., 2003) as a growth substrate for G. metallireducens does not lead to a detectable effect on the PLFA signature: either profiles are characterized by the same most abundant fatty acids as reported in Fig. 4a. Therefore, the precise role of Geobacteraceae in the enrichments cannot be ascertained on the basis of the occurrence of labelled PLFAs, because this does not exclude the possibility that Geobacteraceae are not directly degrading toluene but instead grow on metabolic products, such as acetate, generated by other toluene-oxidizing consortia members.

When all DGGE profiles of BTX-degrading cultures were subjected to cluster analysis (data not shown), profiles generally grouped together according to three characteristics: firstly, the origin of the culture, namely groundwater or sediment (the inoculum utilized in first-generation enrichments), secondly, the specific BTX compound being degraded by each enrichment culture and thirdly the generation of the enrichments. Previous culture-independent analysis conducted at the Banisveld aquifer revealed clear differences in bacterial community composition between sediment and groundwater samples (Röling et al., 2001) This observation may explain why BTX-degrading microbial consortia based on groundwater differ from those derived from sediment. In the second instance, the clustering patterns of DGGE profiles of amplified 16S rRNA genes seem to reflect the specific metabolic capacities of the enrichment cultures with respect to the ability to degrade benzene, toluene or the three xylenes isomers. However, it should be pointed out that when cultures were exposed to a second BTX compound, immediate removal was observed in most of the cases (Botton & Parsons, 2006). The capacity to utilize more than one BTX compound as a carbon source might be related to the general occurrence of some species in the BTX-degrading microcosms. Thirdly, 16S rRNA gene base analysis also indicated that primary enrichments form a separate cluster from the secondary enrichments and the influence of the generation of the enrichments, as well as of inocula used to establish the enrichments, on the DGGE profiles is evident. In previous studies, a gradual increase with time of BTX degradation rates within first-generation cultures (Botton & Parsons, 2006) was observed and this trend steadily continued to increase in subsequent generations (Botton & Parsons, 2007). The average degradation rates of these same primary and secondary enrichments were recalculated here applying a different time interval – from incubation until sampling for molecular analysis – and a similar increasing trend was observed. The clustering, reduced diversity and higher degradation rates suggest that continued exposure to benzene, toluene or xylenes have induced a selection towards a narrower, common community adapted to BTX pollutants.

Unpolluted sediment incubations showed 16S rRNA gene DGGE profiles different from the patterns obtained with polluted sediment cultures. The occurrence of different microbial communities in clean and polluted sediment of the Banisveld aquifer was also reported in a previous work (Röling et al., 2001). The exposure of unpolluted sediment inoculations to each of the BTX compounds did not induce significant changes compared with unamended cultures, in accordance with previous findings indicating that incubation of unpolluted sediment with BTEX for almost 3 years did not promote the ability to degrade any of the contaminants (Botton & Parsons, 2006).


This research is part of the TRIAS project ‘Resilience of the groundwater system in reaction to anthropogenic disturbances’ (project number 835.80.007). Wilfred Röling is supported by the Dutch BSIK-funded Ecogenomics Research Programme.

The authors thank Traian Brad and Bin Lin for help and support in molecular work at the Department of Molecular Cell Physiology, Free University, Amsterdam.