PCR-DGGE method to assess the diversity of BTEX mono-oxygenase genes at contaminated sites

Authors


  • Editor: Kornelia Smalla

  • Present address: Eva M. Top, Department of Biological Sciences, University of Idaho, Moscow, ID, USA.

Dirk Springael, Division of Soil and Water Management, Catholic University of Leuven (KULeuven), Kasteelpark Arenberg 20, 3001 Heverlee, Belgium. Tel: +32 0 16 321604; fax: +32 0 16 321997; e-mail: dirk.springael@biw.kuleuven.be

Abstract

tmoA and related genes encode the α-subunit of the hydroxylase component of the major group (subgroup 1 of subfamily 2) of bacterial multicomponent mono-oxygenase enzyme complexes involved in aerobic benzene, toluene, ethylbenzene and xylene (BTEX) degradation. A PCR-denaturing gradient gel electrophoresis (DGGE) method was developed to assess the diversity of tmoA-like gene sequences in environmental samples using a newly designed moderately degenerate primer set suitable for that purpose. In 35 BTEX-degrading bacterial strains isolated from a hydrocarbon polluted aquifer, tmoA-like genes were only detected in two o-xylene degraders and were identical to the touA gene of Pseudomonas stutzeri OX1. The diversity of tmoA-like genes was examined in DNA extracts from contaminated and non-contaminated subsurface samples at a site containing a BTEX-contaminated groundwater plume. Differences in DGGE patterns were observed between strongly contaminated, less contaminated and non-contaminated samples and between different depths, suggesting that the diversity of tmoA-like genes was determined by environmental conditions including the contamination level. Phylogenetic analysis of the protein sequences deduced from the amplified amplicons showed that the diversity of TmoA-analogues in the environment is larger than suggested from described TmoA-analogues from cultured isolates, which was translated in the DGGE patterns. Although different positions on the DGGE gel can correspond to closely related TmoA-proteins, relationships could be noticed between the position of tmoA-like amplicons in the DGGE profile and the phylogenetic position of the deduced protein sequence.

Introduction

Sites containing groundwater plumes contaminated with benzene, toluene, ethylbenzene and xylenes (BTEX) form a worldwide problem (Atlas & Cerniglia, 1995; Swoboda-Colberg et al., 1995). Microbial degradation of BTEX can occur under both aerobic and anaerobic conditions, and in situ bioremediation is being applied increasingly for aquifer treatment at BTEX-contaminated sites (Barker et al., 1987; Lovley, 2001). In addition, there is a growing interest in monitored natural attenuation of contaminated aquifers as a long-term clean-up strategy (Barker et al., 1987). An adequate site characterization and a long-term monitoring plan are essential for both in situ bioremediation and monitored natural attenuation. As such, there is an increasing requirement for tools enabling rapid assessment of the presence and activity of the biological catalysts involved in BTEX degradation at a contaminated site before and during treatment (Baldwin et al., 2003). A powerful tool is the detection of genes involved in degradation of xenobiotic compounds by PCR, and of their activity by means of RT-PCR, in DNA/mRNA extracted from aquifer samples (Baldwin et al., 2003). However, the application of (RT-)PCR provides information on whether or not a specific catabolic function is present and/or active but does not provide information on the diversity of the (transcribed) genes, unless a clone library constructed from the amplicons is analyzed by sequencing. The availability of a method combining PCR detection and direct diversity analysis of genes encoding a catabolic function is useful for a variety of reasons, namely: (1) to have a rapid picture of the genetic diversity corresponding to a catabolic function of microbial communities in a given environment; and (2) to evaluate rapidly the stability/fitness/dynamics of catabolic genes, operons and their respective hosts in relation to (man-induced) environmental perturbations over space and time (Wawer et al., 1997; Rosado et al., 1998; Henckel et al., 1999; Duarte et al., 2001; Fjellbirkeland et al., 2001; Kahng et al., 2001; Watanabe et al., 2002; Junca & Pieper, 2003a, b). Recently, Junca & Pieper (2003a, b) applied amplified functional DNA restriction and single-strand conformation polymorphism DNA analysis to determine catechol 2,3-dioxygenase gene diversity in BTEX-contaminated aquifers. Catechol 2,3-dioxygenase catalyzes meta-cleavage of the aromatic ring, a central step in degradation of BTEX. However, no method has been described allowing the rapid assessment of the diversity of genes/enzymes involved in the initial attack of BTEX.

In this study, we report on a PCR-denaturing gradient gel electrophoresis (DGGE) method to analyze the diversity of genes, designated as tmoA-like genes, encoding proteins belonging to the first subgroup of subfamily 2 α-subunits of the hydroxylase component of bacterial multicomponent mono-oxygenases involved in the initial attack of aerobic BTEX degradation, in environmental samples. This subgroup includes the α-subunits of most of the BTEX mono-oxygenase complexes identified up to now (Byrne et al., 1995; Kitayama et al., 1996; Bertoni et al., 1998; Ma & Herson, 2000; Kahng et al., 2001; Baldwin et al., 2003). The significance of tmoA-like genes in aquifers has been shown previously by using PCR or DNA-DNA hybridization (Ogram et al., 1995; Guo et al., 1997; Baldwin et al., 2003; Cavalca et al., 2003). Although showing significant polypeptide similarity, the proteins belonging to this group display different BTEX substrate specificities and can attack the aromatic ring at different positions. A new minor degenerate primer set, designated TMOA-F/TMOA-R, was designed, which allowed the generation of tmoA-like gene fragments suitable for the DGGE analysis approach.

Materials and methods

Bacterial strains, media and growth conditions

Bacterial strains used in this study are described in Table 1. They were routinely grown at 30°C on 869-rich medium or on Tris minimal medium supplemented with the appropriate carbon source (Mergeay et al., 1985). For growth on BTEX, Tris minimal medium agar plates or liquid cultures were incubated in closed containers in which the BTEX compounds were distributed via the gas phase from a vial containing cotton wool spiked with 200–600 μL of pure product of the appropriate BTEX compound (Janssen Chemica, Beerse, Belgium). The containers were incubated in the dark on an orbital horizontal shaker at 200 r.p.m. at a constant temperature of 30°C.

Table 1.   Description of bacterial strains used in this study
OrganismBTEX degrading capacity and
relevant catabolic genotypes*
ReferencePCR results with
primer set
TMOA-/TMOA-R
  • B, benzene; T, toluene; EB, ethylbenzene; o-X, o-xylene; m-X, m-xylene; p-X, p-xylene;+, PCR signal;−, no PCR signal.

  • *

    Genes encoding the hydroxylase α-subunit of BTEX mono-oxygenase are underlined. Genes encoding hydroxylase α-subunit proteins belonging to the first subgroup of subfamily 2 are given in bold.

Sphingomonas yanoikuyae B1T, m-X, p-X, xylMAB, xylCXYFEGJQKIHTKim & Zylstra (1999)
Burkholderia cepacia G4B, T, tomA012345, tomBShields et al. (1995)
Burkholderia sp. strain JS150B, T, EB, tbmABCDEF, tbc2ABCDEFJohnson & Olsen (1997)+
Ralstonia pickettii PKO1B, T, EB, o-X, m-X, p-X, tbuA1UBVA2C (tbuT), tbuD (tbuR), tbuEFGKIHJ (tbuS), tbuXBertoni et al. (1998)+
Ralstonia eutropha JMP134Phenol, phlKLMNOPX (phlR)Ayoubi & Harker (1998)+
Ralstonia metallidurans CH34T, B, o-XMergeay et al. (1985), Springael, (unpublished data)+
Pseudomonas aeruginosa JI104B, bmoABCDEFKitayama et al. (1996)+
Pseudomonas mendocina KR1B, T, tmoXABCDEF (tmoST)Yen et al. (1991)+
Pseudomonas putida F1B, T, EB, todFC1C2BADE, todGIH, (todST)Zylstra & Gibson (1989)
Pseudomonas putida mt-2 (PaW1)T, m-X, p-X, xylUWCMABN, xylXYZLTEGFJQKIH, (xylS, xylR)Burlage et al. (1989)
Pseudomonas putida MT53T, m-X, p-X, xylUWCMABN, xylXYZLTEGFJQKIH, cdoKok et al. (1999)
Pseudomonas putida MT15T, m-X, p-X, xylUWCMABN, xylXYZLTEGFJQKIH, cdoKeil et al., 1985)
Pseudomonas stutzeri OX1B, T, o-X, touABCDEF (touR), xylAMBertoni et al. (1998)+

Soil samples used and their physical-chemical properties

Subsurface samples were derived in March 2001 from a BTEX (mainly benzene)-contaminated oil-refinery site, situated in Northern Bohemia (Czech Republic) (Fig. 1). The contamination plume is spreading in a south-west direction. The groundwater flow velocity is in the range of 20 to 30 cm day−1. Subsurface soil samples were taken from different locations in the contaminated aquifer (A1, A2, A3 and A4) and from different depths designated as X, Y and Z, in which X represents the vadose zone (depth 2.58 m), Y the capillary fringe (depth 2.91 m) and Z the saturated zone (depth 3.84 m). The sampling was performed by percussion core drilling with minimal heating of the core. In an anaerobic tent, the aquifer was taken out of the cores (2 g per Falcon tube) using a sterile spatula and stored at −80°C prior to DNA extraction. The tested soil samples were accordingly designated as A1X, A1Y, A1Z, A2X, A2Y, A2Z, A3X, A3Y, A3Z and A4Z. All were mixtures of three soil samples taken at the same place and depth within a distance of ±1 m. For practical reasons, samples could not be taken from the vadose and capillary fringe zone of location A4. Groundwater at sampling points A1 and A2 was strongly contaminated with all BTEX compounds. Groundwater at A3 contained only a minor benzene concentration, whereas the groundwater at A4 was not contaminated (Table 2). At all locations, groundwater oxygen concentrations were below 1 mg L−1 (Table 2). Groundwater sampling and analysis was performed as described previously (Hendrickx et al., 2005a).

Figure 1.

 Map of the studied benzene, toluene, ethylbenzene and xylenes (BTEX)-contaminated site in the Czech Republic showing the benzene contaminated groundwater plume and indicating relevant sampling locations. Based on measured benzene concentrations in appropriate groundwater monitoring wells in March, 2001, model calculations using the processing modflow program (version 7, US Geological Survey, Reston, VA) demonstrated that the site contains two plumes. Locations A1, A2, A3 and A4 are indicated. Location A3 is probably influenced by two plumes (M. Cernik, personal communication). The arrow indicates the direction of the groundwater flow. Numbers prefixed with PV, D and HJ indicate the positions of groundwater monitoring wells, with the number below indicating the measured benzene concentration (in μg mL−1) at that well in March 2001.

Table 2.   Characteristics of the tested soil samples and groundwater as measured in March 2001 and results of the PCR with primer set TMOA-F/TMOA-R on soil template DNA
Soil
sample
LocationZoneConcentration groundwater (mg L−1)TOC soil
(%w)
PCR detection of
tmoA-like genes
Benzene
(mg L−1)
Toluene
(mg L−1)
Ethylbenzene
(mg L−1)
Xylenes
(mg  L−1)
O2
(mg L−1)
NO3
(mg L−1)
NH4+
(mg L−1)
Redox
(mV)
  1. NA, not applicable; ND, not determined;+, strong PCR signal;++, very strong PCR signal.

A1XA1VadoseNANANANANANANANA2.36+
A1YA1Capillary fringeNANANANANANANANA1.55+
A1ZA1Saturated>100>10>1>100.666.620.81161.21++
A2XA2VadoseNANANANANANANANA2.50+
A2YA2Capillary fringeNANANANANANANANA3.57+
A2ZA2Saturated>100>10>1>100.381.31.52711.51++
A3XA3VadoseNANANANANANANANA2.36+
A3YA3Capillary fringeNANANANANANANANA2.56+
A3ZA3Saturated0.01-0.1<0.0002<0.0002<0.00020.1611.70.2ND2.22++
A4ZA4Saturated<0.0002<0.0002<0.0002<0.00020.160.412.841682.01+

Design of primer set TMOA-F/TMOA-R

A degenerate primer set consisting of primers TMOA-F (5′-CGAAACCGGCTT(C/T)ACCAA(C/T)ATG-3′, Pseudomonasmendocina KR1 tmoA position 588 to 609) and TMOA-R (5′-ACCGGGATATTT(C/T)TCTTC(C/G)AGCCA-3′, P. mendocina KR1 tmoA position 1069 to 1092) was designed based on the alignments of the multicomponent mono-oxygenase hydroxylase large α-subunit proteins and the corresponding gene sequences. This alignment was obtained with the gcg wisconsin protein and dna analysis program (version 7.0) (Genetics Computer Group, Madison, WI, USA), which is available on the BEN (Belgian EMBnet Node) Homepage (http://ben.vub.ac.be/) and with the bionumerics software (version 2.50) (Applied Maths, Kortrijk, Belgium). The selectivity of the primer set was evaluated by visual analysis of the primer region within the constructed alignment, by the Advanced BLAST Search program available from GenBank (GenBank, National Centre for Biotechnology Information, Bethesda, MD) and by PCR application of the primers on the DNA from a variety of relevant BTEX-degrading bacterial strains listed in Table 2. A 40 bp long GC-clamp (Muyzer et al., 1993) was attached to the 5′ end of the TMOA-F primer to allow DGGE analysis of the amplicons. The primer couple GC-TMOA-F/TMOA-R amplified a 505 bp sequence of tmoA, resulting in a PCR product of 545 bp.

DNA extraction from cell cultures and soil

Total genomic DNA was extracted from purified bacterial strains grown in 5 mL Tris minimal medium exposed to BTEX vapors or in 5 mL 869 medium as described by Vanbroekhoven et al. (2004). Before use in PCR, the DNA concentration in the extracts was adjusted to 100 ng μL−1.

DNA was extracted from 2 g of mixed soil using a protocol modified from El-Fantroussi et al. (1997). The soil was suspended in 4 mL Tris–glycerol buffer (10 mM Tris, 15% glycerol, pH 7) and the cells in the soil were mechanically lysed by beating with glass beads (0.10–0.11 mm) in a bead beater apparatus (B. Braun Biotech International GmbH, Melsungen, Germany) for 30 s (0°C) followed by an enzymatic lysis with lysozyme (final concentration 2 mg mL−1; 30 min at 37°C). A 120 μL quantity of a 20% sodium dodecyl sulfate solution and 32 μL of a Proteinase K solution of 32 mg mL−1 were added and vials were incubated for 30 min at 50°C. This was followed by an addition of 2 mL of a phosphate buffer consisting of 5.34 g Na2HPO4·2H2O and 4.08 g KH2PO4 in 500 mL H2O-adjusted to pH 8.0. The vials were placed on ice and subjected to a second bead-beating step (30 s). Glass beads, soil and cell debris were removed through centrifugation (3 min at 18 000 g) and the crude DNA extract was further purified by a single extraction with phenol/chloroform/isoamylalcohol (25 : 24 : 1). A 1 g quantity of acid-washed polyvinylpolypyrrolidone (Sigma-Aldrich NV/SA, Bornem, Belgium) was added to the DNA solution to remove copurified humic acids, and vials were incubated for 30 min on ice while being shaken every 5 min. Polyvinylpolypyrrolidone was removed by centrifugation (3 min at 1800 g). The DNA was precipitated from the supernatant with ethanol 100% (2 × extract volume; Merck KGaA, Darmstadt, Germany) and 3 M sodium acetate, pH 5.2 (0.1 × extract volume; Merck KGaA, Darmstadt, Germany) overnight at −20°C and subsequently washed with ethanol 70%. The DNA pellet was resuspended in 400 μL TE-buffer and further cleaned from coprecipitated impurities over a Wizard column (Wizard DNA Clean-Up System; Promega Corporation, Madison, WI). The DNA concentration in the 50 μL soil extract was measured spectrophotometrically (Ultrospec 3000 UV/Visible spectrophotometer; Pharmacia Biotech, Cambridge, UK); c. 6 to 31 μg DNA g−1 soil was obtained.

PCR amplification of tmoA-like genes from pure strain and soil DNA

The final PCR profile used with the TMOA-F/GC-TMOA-F and TMOA-R primer pairs consisted of an initial denaturation of 5 min at 95°C, followed by 35 cycles of one denaturation step for 1 min at 94°C, one annealing step for 1 min at 61.2°C and one elongation step for 2 min at 72°C. The last step consisted of an extension for 10 min at 72°C. PCR was performed on Biometra (Biometra, Göttingen, Germany) or Perkin Elmer (Perkin Elmer, Norwalk, CT) PCR machines. The PCR mixture contained 100 ng of pure strain DNA or soil DNA as templates, 1.25 U exTaq Polymerase, 10 pmol of the forward primer, 10 pmol of the reverse primer, 200 μM of each dNTP and 5 μL of 10 ×exTaq reaction buffer (20 mM MgCl2) in a final volume of 50 μL. All primers were synthesized by Westburg (Westburg BV, Leusden, the Netherlands). The exTaq Polymerase, dNTPs and PCR buffer were purchased from TaKaRa (TaKaRa Ex Taq™; TaKaRa Shuzo Co. Biomedical Group, Japan). To check PCR product sizes, 10 μL of the PCR reactions were analyzed by 2% agarose gel electrophoresis in TAE (Tris acetic acid EDTA buffer) as described by Sambrook et al. (1989).

DGGE analysis

Denaturing gradient gel electrophoresis analysis of the PCR products was performed as described by Muyzer et al., (1993) on 6% polyacrylamide gels with a denaturing gradient of 40–70% (where 100% denaturant gels contain 7 M urea and 40% formamide). Electrophoresis was performed at a constant voltage of 120 V for 15 h in 1 × TAE running buffer at 60°C in the DGGE machine (INGENYphorU-2; INGENY International BV, Goes, the Netherlands). After electrophoresis, the gels were stained with 1 × SYBR Gold nucleic acid gel stain (Molecular Probes Europe BV, Goes, Leiden, the Netherlands) and photographed under UV light using a Pharmacia digital camera system with liscap image capture software (version 1.0) (Image Master VDS; Pharmacia Biotech). The processing of the DGGE gels was done with the bionumerics software. Duplicate PCRs on the same DNA extract always resulted into identical DGGE profiles.

Sensitivity of the PCR based method for tmoA-like gene detection

To examine the detection limit of PCR, a known amount of viable cells of Ralstonia pickettii PKO1 containing the tmoA-like gene tbuA1 was added to sterilized soil A2X at different final cell concentrations (approximately 108, 106, 104 and 102 CFU g−1). Sterile soil was obtained by autoclaving the soil twice at 120°C for 50 min in a Melag autoclave type 23 (Melag, Berlin, Germany). Inoculum cells were harvested from liquid cultures grown in 869-rich medium, washed twice and added in 100 μL aqueous suspension of the appropriate cell density to 1 g of sterilized soil. The exact quantity of added cells was calculated by plating 100 μL of serial dilutions of the used bacterial culture on 869 rich medium or on Tris minimal medium supplemented with toluene. DNA was extracted from the soils directly after inoculation and used as template in PCR with primer sets TMOA-F/TMOA-R and GC-TMOA-F/TMOA-R. No PCR product was obtained from DNA extracted from soil without inoculation, proving that indigenous tmoA-like gene sequences were destroyed by sterilization.

Sequence analysis of amplified BTEX mono-oxygenase α-subunit gene fragments

Bands were excised from the DGGE gel and incubated overnight at 4°C in 50 μL of sterile H2O. PCR was performed on 1 μL of this eluted DNA using primer set TMOA-F/TMOA-R. After agarose gel electrophoresis, the resulting 505 bp DNA fragment was extracted from the agarose using the Millipore Ultrafree®-DA centrifugal filter device (Millipore, Bedford, MA) and cloned into plasmid vector pCR®2.1-TOPO® using the TOPO TA Cloning® Kit with the TOP10 One Shot® Chemically Competent cells (N.V. Invitrogen SA, Merelbeke, Belgium) as described by the manufacturer. The DGGE patterns of the cloned fragments were compared with each other and with those of the tmoA-like gene fingerprints obtained from the soil DNA to relate cloned fragments to DGGE bands. From all cloned fragments showing different migration on DGGE, a 500 bp long fragment was sequenced by Westburg. Nucleotide sequencing of plasmids with cloned inserts was carried out on both strands using the TMOA-F and TMOA-R primers. A similarity analysis of the DNA sequences was obtained by applying the Advanced BLAST Search program available from GenBank. The DNA sequences were submitted to the ‘transeq’ algorithm of the emboss program (version 1.9.1) (BEN, Brussels, Belgium) to translate them in protein sequences. The protein sequences were imported into the alignment of the hydroxylase large α-subunit protein sequences using the bionumerics software 2.5 and edited manually to remove amino acid positions of ambiguous alignment and gaps. Sequence similarities were calculated and a distance-based evolutionary tree was constructed using the neighbor-joining algorithm of Saitou & Nei (1987). Distances were generated using the Kimura matrix (Kimura, 1980). The topography of the branching order within the dendrogram was evaluated by using the maximum-parsimony character-based algorithm in parallel with bootstrap analysis with a round of 1000 reassemblings. An outgroup of the closely related XamoA and IsoA proteins (subgroup 2 of subfamily 2 in phylogenetic tree described by Kahng et al. (2001) was included to root the tree.

Nucleotide sequence accession numbers

The nucleotide sequences of the tmoA-like gene fragment of the clones and isolates VM883 and VM886 reported in this study are available from GenBank under accession numbers AY450309–AY450335, AY450336 and AY450337, respectively.

Results and discussion

Design and specificity of a new degenerate primer set for PCR detection of tmoA-like genes and subsequent DGGE analysis

Because other primer sets for detection of the corresponding genes showed too large a degeneracy for DGGE applications (Baldwin et al., 2003), a new moderately degenerate primer set (each primer with twofold degeneracy at two positions), consisting of primers TMOA-F and TMOA-R, was designed. The designed primer pair TMOA-F/TMOA-R detects the corresponding genes of all the members of subgroup 1 of the second subfamily based on the phylogenetic tree described by Kahng et al. (2001). Because of too large a number of mismatches (five to 19 mismatches with the forward primer TMOA-F and 15 to 20 mismatches with the reverse primer TMOA-R), genes encoding α-subunits belonging to subgroup 2 of subfamily 2 and to the more distantly related subfamilies 3 and 4 should not be amplified by the designed primer set. The theoretical assumptions concerning the TMOA-F/TMOA-R primer set selectivity were evaluated by performing PCR on DNA extracted from a variety of BTEX-degrading bacterial strains producing hydroxylase α-subunit proteins more or less related to TmoA, as listed in Table 1. Fragments of the expected size were obtained with the DNA from all bacteria containing tmoA-like genes. No PCR products were obtained with DNA from strains which do not contain tmoA-like genes (Table 1).

PCR detection of tmoA-like gene sequences in BTEX-degrading isolates

Thirty-five BTEX-degrading bacterial strains were examined for the presence of tmoA-like genes by PCR with primer set TMOA-F/TMOA-R. They were isolated from the same hydrocarbon-polluted aquifer sample under different conditions (aerobic/microaerophilic) (Bastiaens, unpublished data). Only two strains, namely the o-xylene degrading strains VM883 and VM886, both tentatively identified as Pseudomonas syringae, showed the expected 505 bp PCR product. The DNA sequence of the PCR product showed 100% similarity with the tmoA-like gene touA of Pseudomonasstutzeri OX1 (Bertoni et al., 1998), confirming the specificity of the new primer set. These results might indicate that the tmoA-like genes are relatively rare in BTEX-degrading isolates from the examined site. Using other primer sets targeting other genes involved in BTEX catabolism, we found that most of the tested BTEX degraders isolated from that site carry the xylM gene, which encodes XylM catalysing the initial attack of TEX via the alkyl side chain, rather than genes encoding enzymes catalysing the direct oxidation of the ring (Hendrickx, unpublished data). The site from which the strains were isolated contained mainly m- and p-xylene, which are degraded by mono-oxygenases attacking the methyl side chain of the compound. The hydrolase component of such mono-oxygenases is encoded by xylM-like genes. This might explain the occurrence of xylM sequences instead of tmoA-like sequences in the BTEX-degrading isolates from that site.

Differentiation of PCR amplified tmoA-like gene sequences by DGGE

The potential of the PCR-DGGE method to differentiate between different tmoA-like gene sequences from different BTEX-degrading bacteria was investigated. Therefore, tmoA-like gene fragments of P. mendocina KR1 (tmoA), P. stutzeri OX1 (touA), Ralstonia pickettii PKO1 (tbuA1), Pseudomonas aeruginosa JI104 (bmoA), Ralstonia metallidurans CH34 (reut2189), Burkholderia sp. strain JS150 (tbc2A) and those from the two isolates P. syringae VM883 and VM886 were amplified with primer set GC-TMOA-F/TMOA-R and analyzed on DGGE. The amplicons obtained from the reference strains showed single bands, but at different positions on the DGGE gel (Fig. 2b). This shows that the degeneracy of the primer set did not affect the banding profile. As expected from their sequences, the PCR fragments from strains VM883 and VM886 demonstrated a migration pattern identical to the pattern of the amplified touA gene fragment of P. stutzeri OX1 (data not shown). tmoA-like gene amplicons, amplified using a primer set with a GC-clamp attached to the reverse primer instead of the forward primer, showed a less clear difference in migration distance between the amplified touA and tmoA gene fragments (data not shown). These results show that application of the designed primer set GC-TMOA-F/TMOA-R followed by DGGE analysis of the PCR amplicons allows discrimination between the amplified tmoA-like gene sequences from different BTEX-degrading bacteria, and that the method can be used for diversity analysis of these genes.

Figure 2.

 (a) tmoA-like gene denaturing gradient gel electrophoresis (DGGE) fingerprint of soil samples. Bands A, B, C, D, E, F, G, H, I, J and K from which clones were derived are indicated; roman numbers between brackets indicate corresponding phylogenetic TmoA-protein analogue groups based on Fig. 3. (b) DGGE analysis of cloned fragments from bands indicated in (a) and their positioning in DGGE relative to tmoA-like gene fragments amplified from the reference strains. DGGE bands from which the clones originated are indicated between brackets.

Sensitivity of the PCR detection method using primer sets TMOA-F/TMOA-R and GC-TMOA-F/TMOA-R

To examine the sensitivity of PCR to detect tmoA-like gene sequences in soil, a decreasing concentration of R. pickettii PKO1 cells containing the tmoA-like gene tbuA1 was added to sterile soil A2X. PCR was performed on DNA extracts using primer sets TMOA-F/TMOA-R and GC-TMOA-F/TMOA-R. A detection limit of around 103–104 gene copies g−1 soil was found, assuming one copy of the gene per cell (data not shown), with primer set TMOA-F/TMOA-R, a commonly found PCR detection limit in environmental samples such as soil (Kowalchuk et al., 1999), compost (Kowalchuk et al., 1999) or seawater (Sinigalliano et al., 1995). However, with primer set GC-TMOA-F/TMOA-R, a detection limit of 106 gene copies g−1 soil was obtained (data not shown), suggesting that problems might be encountered when using the PCR-DGGE method for exploration of the diversity of tmoA-like genes in samples where those genes are present at low concentrations. GC-clamp effects were also reported by others (Rosado et al., 1998; Vanbroekhoven et al., 2004; Leys et al., 2005). This problem was solved by performing a semi-nested PCR with primer set GC-TMOA-F/TMOA-R on the PCR products obtained with primer set TMOA-F/TMOA-R, which lowered the detection limit to 103–104 gene copies g−1 soil, assuming one copy of the gene per cell (data not shown).

PCR detection and DGGE profiling of tmoA-like gene sequences in contaminated subsurface soil samples from a BTEX polluted site

The PCR-DGGE method using the new primer set was applied for detection and diversity profiling of tmoA-like gene sequences in different subsurface soil samples, namely A1X, A1Y, A1Z, A2X, A2Y, A2Z, A3X, A3Y, A3Z and A4Z (Table 2), derived from a site that contained a BTEX-contaminated groundwater plume. tmoA-like gene sequences were amplified from the DNA extracts using the seminested PCR approach, and the resulting amplicons were subjected to DGGE analysis. As shown in Table 2, PCR products of the expected size were obtained for all soil samples. Strong signals were especially recovered from the samples from the saturated aquifer zone, including noncontaminated sampling point A4. As such, despite the low dissolved oxygen concentrations in the groundwater, tmoA-like gene sequences were easily recovered from the aquifer. Similar results were reported by others (Ogram et al., 1995; Guo et al., 1997; Baldwin et al., 2003; Cavalca et al., 2003). Recent analysis of 16S rRNA gene libraries indicates that the aquifer at both contaminated and non-contaminated locations is dominated by aerobic and facultative anaerobic organisms such as Pseudomonas, while no strict anaerobic organisms such as Sulphate Reducing Bacteria and Geobacter which both contain anaerobic BTEX-degrading strains could be identified (Hendrickx et al., 2005a). Upstream of the contamination plume, oxygen was present (M. Cernik, Aquatest). Possibly a rapid consumption of the incoming oxygen occurs, sustaining aerobic BTEX degradation and bacteria carrying BTEX mono-oxygenase and dioxygenase genes. Furthermore, at the site, the concentration of nitrate was low at the contaminated area but high upstream, indicating that rapid consumption of incoming nitrate occurs, which might have sustained oxygen-requiring but nitrate-enhanced degradation of BTEX under hypoxic conditions (in other words, conditions with low concentrations of dissolved oxygen), as described for R. pickettii PKO1 and other aerobic BTEX degraders carrying tmoA-like genes (Kukor & Olsen, 1996). In a parallel study, we targeted the same samples for the presence of other genotypes involved in initial attack of BTEX. This study showed that tmoA-like genes along with xylM-like genes encoding attack of toluene, ethylbenzene and xylene at the alkyl-side chain were the main BTEX initial oxidation genes present at the site (Hendrickx et al., 2005b).

Denaturing gradient gel electophoresis analysis of the amplicons (Fig. 2a) showed that at all contaminated locations (A1, A2 and A3), the different zones displayed different tmoA-like gene compositions, with the capillary fringe showing the lowest diversity. In addition, the DGGE profile also showed differences between subsurface samples taken at the same depth, but at different locations at the site. DGGE profiles obtained from the highly contaminated A2 sample and the less contaminated A3 sample were very similar for each zone, despite the difference in contamination level and history. Because A3 is located at the fringe of the expanding BTEX plume, it might suggest that, upon contamination, the downstream community rapidly evolves towards or is becoming dominated by the upstream community. Unexpectedly, the tmoA gene profiles differed between A1 and A2, although both locations are affected by similar BTEX contamination levels. This indicates that another reason caused the diversification of the A1 profile, which was dominated by one band (band B), from the A2 and A3 profiles. Interestingly, the multiple tmoA-like sequences in A2Z and A3Z seem to become dominated in A1Z by sequences corresponding to band B, because this band appears in the DGGE profile from the A2Z sample but not in the fingerprint of sample A3Z. These observations indicate that the tmoA-like gene composition and most probably the bacterial community carrying those genes at A2Z is changing to a tmoA-like gene composition/community similar to that of A1Z or vice versa.

To determine the sequence diversity of the recovered tmoA-like genes corresponding to the DGGE bands, the bands from the DGGE fingerprints of the Z-layer from A1, A3 and A4 locations were cloned and sequenced. We concentrated on the Z layer because samples were available for both the contaminated and non-contaminated locations (Fig. 3a). DGGE analysis of cloned fragments from bands A and D showed that those bands consisted of tmoA-like genes with slightly different mobility (Fig. 2b). One clone was sequenced from each set of fragments showing identical migration on DGGE, except for bands A and D. Table 3 shows the nearest protein match based on BLASTX analysis of the cloned sequences and the G+C content, whereas Fig. 3 shows the phylogenetic tree of the deduced amino acid sequences. Clear differences in G+C content existed between nucleotide sequences of tmoA-like gene PCR fragments obtained from sample A1Z and PCR fragments from A4Z, which reflected their different migration pattern. Therefore, it seemed that a shift from higher G+C content tmoA-like gene sequences towards lower G+C content tmoA-like gene sequences took place as the BTEX contamination was increasing, which probably reflects differences in the host community structure along the BTEX gradient. This is in congruence with recent observations based on the analysis of 16S rRNA gene libraries, which showed that, at least in the aquifer, the contaminated and non-contaminated locations are colonized by different bacterial communities (Hendrickx et al., 2005a).

Figure 3.

 Phylogenetic tree of benzene, toluene, ethylbenzene and xylenes (BTEX) mono-oxygenase α-subunit protein sequences detected in the examined subsurface soils as deduced from the tmoA-like gene clones. The evolutionary tree was generated by the neighbour-joining method. Distances were generated using the Kimura matrix. The maximum-parsimony algorithm was used to evaluate branching orders and the topology was evaluated by bootstrap analysis (1000 reassemblings). Percentages of bootstrap support are indicated at the branch points. An outgroup of the closely related XamoA/xamoA (Xanthobacter sp. Py2) and IsoA/isoA (Rhodococcus sp. AD45) protein sequences belonging to subgroup 2 of subfamily 2 was included to root the tree. The bar at the top indicates the estimated evolutionary distance. The α-subunit of the hydroxylase component of toluene/benzene mono-oxygenase from Burkholderia sp. strain JS150 (Tbc2A), toluene 3-mono-oxygenase from R. pickettii PKO1 (TbuA1), phenol hydroxylase from R. eutropha JMP134 (PhlK), benzene mono-oxygenase from P. aeruginosa JI104 (BmoA), hypothetical protein from R. metallidurans CH34 (Reut2189), toluene 3-mono-oxygenase (T3MO) from B. cepacia AA1 (TbhA), toluene/o-xylene mono-oxygenase from P. stutzeri OX1 (TouA) and toluene 4-mono-oxygenase from P. mendocina KR1 (TmoA) is indicated. DGGE bands (see Fig. 3) from which the clones were derived and accession numbers of the corresponding gene sequences are indicated between brackets. Roman numbers indicate the grouping of the TmoA protein analogues as reported in the text.

Table 3.   BLASTX analysis of sequenced cloned denaturing gradient gel electrophoresis bands obtained in PCR with primers GC-TMOA-F and TMOA-R from the benzene, toluene, ethylbenzene and xylenes-polluted soil samples
BandClone designation
(Nucleotide sequence
accession no.)
OriginG+C
content
(%)
Nearest match in BLASTX analysis
(Protein accession no.) (Host)
Protein
identity
(%)
AA1Z/5 (AY450313)A1Z51.55TmoA (Q00456) (P. mendocina KR1)98
BA1Z/1 (AY450309)A1Z58.17TbhA (AAB58740) (B. cepacia AA1)89
A1Z/2 (AY450310) 57.94TbhA (AAB58740) (B. cepacia AA1)91
A1Z/3 (AY450311) 58.17TbhA (AAB58740) (B. cepacia AA1)89
A1Z/4 (AY450312) 57.94TbhA (AAB58740) (B. cepacia AA1)89
A1Z/6 (AY450314) 57.53TbhA (AAB58740) (B. cepacia AA1)92
A1Z/7 (AY450315) 57.72TbhA (AAB58740) (B. cepacia AA1)89
A1Z/8 (AY450316) 57.49Reut2189 (ZP_00023244) (R. metallidurans CH34)88
A1Z/9 (AY450317) 57.94TbhA (AAB58740) (B. cepacia AA1)91
A1Z/10 (AY450318) 57.72TbhA (AAB58740) (B. cepacia AA1)89
A1Z/11 (AY450319) 57.94TbhA (AAB58740) (B. cepacia AA1)91
A1Z/12 (AY450320) 57.94TbhA (AAB58740) (B. cepacia AA1)89
CA3Z/10 (AY450330)A3Z61.04TbhA (AAB58740) (B. cepacia AA1)88
DA3Z/1 (AY450321)A3Z51.12TbhA (AAB58740) (B. cepacia AA1)80
EA3Z/4 (AY450324) 52.89TbhA (AAB58740) (B. cepacia AA1)82
FA3Z/2 (AY450322) 55.53Reut2189 (ZP_00023244) (R. metallidurans CH34)83
GA3Z/7 (AY450327) 53.73TouA (CAA06654) (P. stutzeri OX1)66
HA3Z/8 (AY450328) 56.14Reut2189 (ZP_00023244) (R. metallidurans CH34)79
IA3Z/6 (AY450326) 57.44TouA (CAA06654) (P. stutzeri OX1)63
JA3Z/9 (AY450329) 56.10Reut2189 (ZP_00023244) (R. metallidurans CH34)78
KA4Z/1 (AY450331)A4Z60.47TbhA (AAB58740) (B. cepacia AA1)90
A4Z/3 (AY450333) 65.65Tbc2A (AAG40794) (B. cepacia JS150)86
A4Z/4 (AY450334) 59.96TbhA (AAB58740) (B. cepacia AA1)88

The sequence results show further that BTEX mono-oxygenases are much more diverse than can be deduced from the current knowledge on BTEX mono-oxygenases characterized in cultured BTEX-degrading isolates. Six major groups could be recognized, with groups I, II and IV containing only deduced sequences from tmoA-like genes originating from direct cloning from environmental samples. Bands corresponding to those sequences occupied the middle part of the DGGE gel, filling up a blank space between tmoA-like gene fragments from reference strains. In addition, bands showing a migration on DGGE similar to reference tmoA-like bands resulted in phylogenetically related protein sequences. For example, bands A, B and K migrated similar to the fragment amplified from KR1, JS150 and CH34, respectively, and provided deduced proteins grouping with TmoA in group III, TcbA2 in group VI and Reut2189 in group V, respectively. However, in other cases, bands showing a different mobility on DGGE corresponded to closely related protein sequences in the TmoA phylogenetic tree. For example, protein sequences deduced from fragments corresponding to bands B and L and some to band K were phylogenetically strongly related and all grouped in group V. In addition, one protein sequence, grouping in group VI, deduced from an amplicon corresponding to band K was found to be phylogenetically diverse from protein sequences, grouping in group V, deduced from the other band K amplicons. Those data indicate that there exists a certain degree of relationship between the position of amplified tmoA-like gene fragments on the DGGE gel and the phylogenetic position of the deduced protein sequences but that different positions on the DGGE gel can correspond to closely related TmoA-proteins.

To explain the difference in composition of mono-oxygenases and probably bacterial community composition between different zones and locations along the BTEX-contamination plume, different hypotheses can be suggested. Different zones often have different oxygen concentrations, which might affect community composition. These differences could be caused by the fact that populations with different oxygen demands for BTEX degradation are developing in the different zones, and that those have different α-subunit hydroxylase gene sequences (Futamata et al., 2001a). Another hypothesis could be the fact that bacteria carrying enzymes with a high affinity towards BTEX (mainly benzene) concentrations (in other words, bacteria which can grow on very low BTEX concentrations) became overgrown by TmoA-like mono-oxygenase carrying bacteria with a low affinity for BTEX in the highly contaminated parts of the site (Futamata et al., 2001a, b). In addition, because of the high concentration of toxic BTEX at locations A1, A2 and even A3, bacterial populations with solvent resistance characteristics might have become selected at those locations, leading to a different tmoA-like gene composition between contaminated and non-contaminated locations (Ramos et al., 2002).

In conclusion, PCR-DGGE was shown to be an appropriate tool for determining sequence diversity of tmoA-like genes in environmental samples and assessing the difference and dynamics of the tmoA-like gene composition in response to (changing) different environmental parameters at different locations of a contaminated site. Along with PCR targeting other catabolic genotypes involved in either aerobic or anaerobic BTEX degradation eventually combined with a fingerprinting method and 16S rRNA gene-based community analysis, a detailed picture of the biodegradative features of a contaminated sample and their hosts can be generated. In the future, it would be interesting to look for the active tmoA-like gene diversity upon BTEX stress over space and time by applying an RT-PCR-DGGE variant of the method on mRNA extracts from environmental samples. This knowledge will significantly increase our understanding of the in situ degradation processes taking place and the organisms involved at contaminated sites and can be useful in the design and evaluation (monitoring) of new bioremediation strategies.

Acknowledgements

This work was supported by EC-project QLK3-CT2000-00731. We thank M. Černík and T. Lederer from AQUATEST Inc., Czech Republic, for supplying soil samples and for the information on the examined site, Tatiana Vallaeys for help in primer design, and P. A. Williams, J. J. Kukor, M. S. Shields, G. J. Zylstra, P. Barbieri and D. T. Gibson for kindly providing the reference strains used in this study. B. H. was supported by a PhD fellowship of Vito.

Ancillary