The naphthalene-degrading activity of a Pseudomonas sp. strain isolated from a creosote-contaminated soil was shown to be encoded by the IncP9 plasmid pNF142 by transfer to Pseudomonas putida KT2442. The effects of the inoculant strain KT2442 (pNF142) and of naphthalene contamination on the soil bacterial community were studied in microcosms with the following treatments: (I) soil, (II) soil with naphthalene, (III) soil with naphthalene and inoculated with KT2442 (pNF142). The inoculant became the dominant bacterial population in treatment (III) as evidenced by cultivation and denaturing gradient gel electrophoresis (DGGE) analysis. The bacterial DGGE profiles revealed drastically reduced complexity due to the numerical dominance of the inoculant. However, group-specific fingerprints (β-proteobacteria, actinobacteria) that excluded KT2442 (pNF142) showed less severe changes in the bacterial community patterns. A major effect of naphthalene on the soil bacterial community was observed in treatment (II) after 21 days. Two dominant bands appeared whose sequences showed the highest similarity to those of Burkholderia sp. RP007 and Nocardia vinaceae based on 16S rRNA gene sequencing. These bands were less intense in treatment (III). The increased abundance of RP007-like populations due to naphthalene contamination was also confirmed by PCR amplification of the phnAc gene. The nahAc and nahH genes were detected in DNA and cDNA only in treatment III. Although the inoculant strain KT2442 (pNF142) showed good survival and expression of genes involved in naphthalene degradation, this study suggests that KT2442 (pNF142) suppressed the enrichment of indigenous naphthalene degraders.
Pseudomonas putida KT2442 (pNF142)
Denaturing gradient gel electrophoresis
The exploitation of microbes for cleaning contaminated sites is still one of the big challenges to environmental microbiologists. The efficient and reliable use of microbes carrying degradative pathways often suffers from a lack of information on their survival and metabolic activity under different environmental conditions, their potential effects on the indigenous microbial community and knowledge of the bioavailability of the pollutants. Because microbial degradation processes under environmental conditions result from enzyme, cell and community-based activities, a better understanding of these processes requires a polyphasic approach. DNA- and RNA-based methods enable researchers not only to follow the fate and activity of inoculant strains but also to evaluate potential shifts in the composition of indigenous bacteria during bioremediation processes. These shifts might result directly from the introduction of inoculant strains or consortia, or indirectly from changes in the concentration of pollutants, their degradation metabolites and their bioavailability. Recently developed molecular fingerprinting techniques based on the analysis of ribosomal RNA fragments amplified from cDNA or RNA are particularly suited to follow changes in the relative abundance of microbial populations during bioremediation processes or in response to the exposure to changing concentrations of pollutants . The detection of mRNA transcripts of degradative genes, although still a methodological challenge, offers the possibility of measuring in situ activity. This approach has recently been successfully applied to follow in situ, real #-time gene expression by characterising naphthalene dioxygenase mRNA transcripts from groundwater .
Because the success of bioremediation processes facilitated by microbial inoculant strains does not only depend on the ability of the inoculant strain to actively colonise the site but also on the expression of the degradative genes, most studies focussed on these aspects. The major aim of this study was to assess the response of indigenous soil bacteria to naphthalene either in the presence or absence of the inoculant strain KT2442 (pNF142).
The naphthalene-degrading activity of Pseudomonas sp. NF142 isolated from creosote-contaminated soil samples collected from a chemical plant in the Nizhnii Novgorod region (Russia) was shown to be encoded by the plasmid pNF142 by transferring it to Pseudomonas putida KT2442. Soil microcosm experiments were performed to provide information on the ability of P. putida KT2442 (pNF142) to survive in naphthalene-contaminated soil and to study the effect of its inoculation and of artificial naphthalene contamination on the soil bacterial communities. Soil samples were taken at five time points and analysed by selective cultivation and DNA/RNA-based methods. Community-level responses were mainly followed by analysing 16S rRNA gene fragments PCR-amplified from soil DNA or RNA by means of denaturing gradient gel electrophoresis (DGGE). Functional analysis focussed on the presence and expression of dioxygenase genes followed by PCR amplification from soil DNA or cDNA.
2 Materials and methods
2.1 Isolation of the naphthalene-degrading strain Pseudomonas sp. NF142 and transfer of plasmid pNF142 to P. putida KT2442
The naphthalene-degrading isolate NF142 was obtained after selective enrichment with naphthalene from a creosote-contaminated soil collected from a territory of the chemical plant Metoxil in the Nizhnii Novgorod region in 2001. Partial 16S rRNA gene sequencing of strain NF142 indicated that its 16S rRNA gene was identical with that of Pseudomonas sp. Cam-1 (AF098464), a polychlorinated biphenyl-degrading strain from arctic soil . The naphthalene-degrading activity of Pseudomonas sp. NF142 was shown to be carried on plasmid pNF142 by transferring it to P. putida KT2442. Plasmid pNF142 was assigned to the IncP9-δ group by PCR as described by Krasowiak et al. .
2.2 Setup of microcosms
The soil used in the experiments was composed of a sieved (<2 mm) mixture of sand (20%, v/v) and commercially available potting soil (clay substrate of Klasmann-Deilmann GmbH –http://www.klasmann-deilmann.de/ with the following characteristics (w/v): 160–260 mg l−1 N, 180–280 mg l−1 P2O5, 200–350 mg l−1 K2O, 80–150 mg l−1 MgO, organic matter 125 g l−1 and pH 5.5–6.0). Three different soil treatments were analysed: (I) control soil, (II) soil spiked with naphthalene (2 mg g−1), and (III) soil spiked with naphthalene (2 mg g−1) and inoculated with KT2442 (pNF142). The soil contamination was done by thoroughly mixing naphthalene crystals (2 mg g−1) into the soil. Before inoculation KT2442 (pNF142) was grown in Luria-Bertani (LB) broth for 24 h and then re-suspended in sterile saline (0.85%). The soil inoculation was performed by adding 1 ml of cell suspension per each 10 g of soil followed by careful mixing to a resulting concentration of 1 × 108 cfu g−1 of dry soil. The same volume of sterile saline was added to non-inoculated soils to standardise the final moisture content (34%) in each microcosm. The microcosms consisted of Falcon tubes of 50 ml containing 20 g (wet weight) of each soil. The flasks were kept partially open to avoid oxygen deprivation. Three entire microcosms were sacrificed per treatment after 1 h (referred to as day 0) as well as 1, 3, 11 and 21 days post-inoculation (dpi) for analysis. Five grams of soil were re-suspended in 45 ml sterile saline and treated mechanically by means of a Stomacher laboratory blender at maximum speed for 2 min. The remaining soil samples were kept frozen at −80 °C until processed for DNA/RNA extraction and naphthalene analysis by high-performance liquid chromatography (HPLC).
2.3 HPLC analysis
The concentration of naphthalene in the soil microcosms from each time interval was determined after hexane extraction via HPLC analysis (Waters HPLC dual gradients system, equipped with diode array and fluorescence detectors), fitted with a Varian ChromSpher 3PAH column (100 × 4.6 mm, 3 μm). Acetonitrile and water were used as the mobile phase (acetonitrile:water 40:60 to 100:0 in 20 min, then 5 min acetonitrile).
2.4 Cfu counts of P. putida KT2442 (pNF142) in naphthalene-contaminated soil
Serial dilutions of the cell suspension obtained by Stomacher treatment were plated onto King’ s B agar supplemented with rifampicin at 50 μg ml−1. The cfu counts were determined after 48 h incubation at 28 °C.
2.5 Extraction of RNA and DNA
Two-ml Eppendorf tubes were filled with 0.5 g of microcosm soil sample and 0.2 g of baked sterile glass beads and kept on ice until simultaneous DNA and RNA extraction as described by Hurt et al.  with the modifications made by Costa et al. . The separation of RNA from DNA was performed using the Qiagen® RNA/DNA Mini Kit – 25 (Qiagen GmbH) according to the manufacturer’ s protocol. The kits UltraClean™ 15 (Mobio) and the RNeasy Mini Kit (Qiagen GmbH) were used to purify the recovered DNA and RNA, respectively, according to the manufacturer’ s instructions.
2.6 Preparation of cDNA from soil RNA
The reverse transcriptase (RT) reaction mixtures (20 μl) containing 50 mM Tris–HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, 10 mM dithiothreitol, deoxynucleoside triphosphates, each at a concentration of 0.5 mM, 250 ng of random primers, and 50 U of Superscript II RNase H-reverse transcriptase (Life Technologies) were used to reverse transcribe purified RNA extracted from soil samples. The reactions were performed for 50 min at 42 °C, followed by incubation at 70 °C for 15 min to inactivate the reverse transcriptase. RT products were kept frozen at −20 °C until use. The suitability of the extracted RNA for RT-PCR amplification was checked by performing RT-PCR control experiments with 16S rRNA. To check any contamination with DNA in the RNA extracts for RT-PCR, controls were prepared consisting of PCR mixtures that contained RNA and were not previously subjected to RT reaction.
2.7 Naphthalene dioxygenase (nahAc, phnAc) and catechol 2,3-dioxygenase (nahH) PCR
PCR amplification of the genes nahAc, phnAc and nahH were performed with primer sets described by Wilson et al. , Lloyd-Jones et al.  and Wikström et al. , respectively (Table 1). The reaction mixtures contained: 1 μl template DNA (approx. 20 ng) or 2 μl of the cDNA reaction mixture, Stoffel buffer II (Applied Biosystems), 0.2 mM dNTPs, 3.75 mM MgCl2, 0.25 μg bovine serum albumin, 100 nM of the forward primer, 100 nM of reverse primer and 1 U AmpliTaq GoldTM Polymerase (Applied Biosystem) in each 25 μl reaction. A touchdown PCR programme was used for amplification of nahAC gene homologues as described by Wilson et al. . For phnAc amplification, an initial 5 min denaturation at 94 °C was followed by 35 thermal cycles of 45 s at 94 °C, 30 s at 65 °C and 60 s at 72 °C, followed by a final extension step at 72 °C for 4 min according to Lloyd-Jones et al. . The nahH PCR consisted of an initial 5 min of denaturation at 94 °C followed by 35 cycles of 30 s at 94 °C, 30 s at 52 °C and 45 s at 72 °C, followed by a final extension step at 72 °C for 4 min.
Probes were generated from specific nahAc and nahH PCR of P. putida KT2442 (pNF142) by labelling the PCR products, excised from the agarose gel, using DIG-labelled dUTP as recommended by the manufacturer (Roche). Southern blotting was done according to Sambrook et al. . PCR products were electrophoresed on agarose gels and subsequently transferred onto HYBOND N nylon membranes (Amersham-Pharmacia Biotech). Hybridisation was performed under conditions of medium stringency following the protocol published by Fulthorpe et al. . Hybridised probes were detected by using a DIG luminescent detection kit (Roche) as specified by the manufacturer and exposure to X-ray film (Roche).
2.9 PCR amplification of 16S rRNA gene fragments for DGGE analysis
Primers F984-GC and R1378 described by Heuer et al.  (Table 1) were used for amplification of bacterial 16S rRNA gene fragments (473 bp). The PCR mixture was: 1 μl template DNA (ca. 20 ng), Stoffel buffer (Applied Biosystems), 0.2 mM dNTPs, 3.75 mM MgCl2, 4% (w/v) acetamide, 100 nM F984-GC, 100 nM R1378, and 1 U Taq DNA polymerase (Stoffel fragment, Applied Biosystems) in each 25 μl reaction. After 5 min of denaturation at 94 °C, 35 thermal cycles of 1 min at 94 °C, 1 min at 53 °C, and 2 min at 72 °C, the PCR was finished by an extension step at 72 °C for 10 min.
2.10 Group-specific PCR of 16S rRNA gene fragments
The forward primers for β-proteobacteria F948β and actinobacteria F243HGC were used in combination with the reverse universal bacterial primers R1494 and R1378, respectively (Table 1). The reaction mixture was: 1 μl template DNA (ca. 20 ng), Stoffel buffer II (Applied Biosystems), 0.2 mM dNTPs, 3.75 mM MgCl2, 100 nM of the forward group-specific primer, 100 nM of reverse bacterial primer and 1 U AmpliTaq Gold™ Polymerase (Applied Biosystem) in each 25 μl reaction. After 5 min of denaturation at 94 °C, 20 thermal cycles of 1 min at 94 °C, 2 min at 64 °C (β-proteobacteria) or 1 min at 63 °C (actinobacteria) and 2 min at 72 °C, the PCR was followed by a final extension step at 72 °C for 10 min. Bovine serum albumin (0.1 μg μl−1) was added to all group-specific PCRs. The amplicons obtained with group-specific primers without GC clamp were diluted 1:20 and used as template for a second PCR with bacterial primers F984-GC and R1378  (Table 1).
2.11 DGGE analysis
DGGE of the amplified rRNA gene sequences was performed using the Dcode System (Universal Mutation Detection System, Bio-Rad). Amplified bacterial 16S rRNA gene fragments were applied to a double gradient polyacrylamide gel containing 6–9% acrylamide  with a gradient of 26–59% of denaturant. The run was performed in 1 × Tris–acetate–EDTA buffer at 58 °C at a constant voltage of 220 V for 6.5 h. The DGGE gels were silver-stained according to Heuer et al. .
2.12 Cloning and sequencing
Selected DGGE bands were excised with a scalpel and transferred into 1.5 ml Eppendorf tubes containing 500 μl of destaining solution (6 mM K3Fe(CN]6) and incubated at 37 °C in a heating block (Thermomixer, Eppendorf) for 10 min. The destaining solution was discarded and the gel slices washed twice with MilliQ water at 37 °C for 5 min each time. The isolation of DNA fragments from destained gel slices was performed as described by Schwieger and Tebbe . The recovered DNA was re-amplified and the PCR product was again subjected to DGGE analysis to ensure correct electrophoretic mobility within the gels. The amplification of 16S rRNA gene fragments from excised DGGE bands, cloning of the PCR products and clone screening were performed as described by Smalla et al. .
Cloning of PCR products amplified with primers targeting the phnAc gene was performed by using the pGEM-T Easy Vector System I (Promega Corporation) according to the manufacturer’ s instructions. The phnAc-PCR was performed again directly from the selected white colonies, and the PCR products were analysed by agarose gel electrophoresis to confirm the presence of the insert in the transformants. PCR products were digested with MspI (New England Biolabs) and analysed in 4% agarose gels (4% NuSieve® 3:1 agarose, Cambrex). Clones representing each group of MspI restriction enzyme patterns were sequenced.
Cloned DNA fragments were sequenced with the standard primers SP6 and T7 (IIT GmbH). 16S rRNA gene sequences were aligned and analysed with ARB software (Department of Microbiology, Technical University of Munich, München, Germany (http://www.arb-home.de]). The phnAc sequences were aligned and the similarities were determined by using the unweighted pair group method using arithmetic averages (UPGMA) included in the GeneCompar software (version 1.3, Applied Maths).
2.13 Cluster analysis
Analysis of bacterial community profiles in the soil microcosms was performed with the software package Gelcompar 4.0 program (Applied Maths). Background was subtracted by using a rolling disk method with an intensity of 10 (relative units), and the lanes were normalised. A dendrogram was constructed by the Pearson correlation index for each pair of lanes within a gel and cluster analysis by UPGMA.
2.14 Nucleotide sequence accession numbers
The 16S rRNA gene nucleotide sequence accession numbers are listed in Table 3. The phnAc sequences are available under GenBank Accession Nos. AY540615–AY540620.
Results of partial sequence analysis and tentative phylogenetic affiliations of bands
Sequenced clone and accession numbera
Most closely related bacterial 16S rRNA gene sequence
A1/A2 – AY376449c
Burkholderia sp. RP007
A3 – AY376448
Burkholderia sp. RP007
A4 – AY376447
Burkholderia sp. RP007
B1/B2/B3/B4 – AY376450
aGenBank accession number of the sequence from excised bands.
bGenBank sequence accession number of most closely related bacterial sequence.
cSequences with 100% of similarity. Only one sequence was submitted to the GenBank.
3.1 Survival of P. putida KT2442 (pNF142) in naphthalene-contaminated soil
Survival of KT2442 (pNF142) in bulk soil artificially contaminated with naphthalene was followed by plating of serial dilutions of soil samples onto King’ s B supplemented with rifampicin. During the first few days the cfu numbers of KT2442 (pNF142) increased by more than 1 order of magnitude reaching a maximal titre of 1.7 × 109 cfu g−1 of soil 3 dpi (see Table 2). At the end of the experiment a slightly lower titre than that determined at 0 dpi was found. The total nucleic acids (DNA/RNA) recovered from the soil samples were of sufficient purity for successful reverse transcription and PCR amplification. The DGGE fingerprints of the bacterial populations obtained from DNA and RNA samples showed little variation between replicates suggesting good reproducibility of the nucleic acid extraction protocol. DGGE analysis of 16S rRNA gene fragments amplified from total community DNA showed a remarkable numerical dominance of KT2442 (pNF142) in treatment (III) at 3 dpi (see Fig. 1). In contrast to the community patterns obtained for treatment (II) in which approximately 30 bands were detected, the bacterial patterns of treatment (III) showed a drastically reduced number of bands, and the bands corresponding to those of KT2442 (pNF142) were mainly visible. DNA- and RNA-based bacterial patterns of treatment (III) revealed that the inoculated strain remained a numerically dominant population during the time course of the experiment and, as indicated by the RNA-based profiles, seemed to be metabolically active (data not shown).
PCR-DGGE analyses of bacterial populations in soil samples amended with naphthalene with no inoculation and after inoculation of P. putida KT2442 (pNF142) (3 dpi). Bacterial standard (ST) , P. putida KT2442 (pNF142) (KT).
3.2 Indigenous soil bacterial communities affected by inoculation of KT2442 (pNF142) in naphthalene-contaminated soil
Due to the numerical dominance of strain KT2442 (pNF142), it was actually impossible to evaluate by DGGE whether or not inoculation affected the composition of the indigenous bacterial community. Therefore, group-specific primers which allow amplification of the actinobacterial or β-proteobacterial 16S rRNA gene fragments before amplification of the GC-clamped PCR-fragment in a second, nested PCR step [11,16] were applied. Since the inoculant strain belongs to the γ-proteobacteria, its 16S rRNA gene fragments were not amplified by the group-specific primers used. Comparison of the group-specific patterns observed for naphthalene-amended soil with or without inoculation of KT2442 (pNF142) revealed rather similar patterns for samples taken at day 0 or 3 dpi (data of the β-proteobacterial patterns in Fig. 2(a); actinobacterial patterns in Fig. 3(a)). Patterns differed clearly from the other samples only for samples of treatment (II) taken 21 dpi, mainly due to the appearance of a strong band in each of the replicates, and formed a separate cluster when Pearson correlation indices were analysed by UPGMA. While all replicates of the β-proteobacterial patterns (Fig. 2(b)) shared more than 80% similarity, the replicates of treatment II (a–c) and replicate b of treatment III of 21 dpi showed only 60% similarity with the main cluster. A similar trend was seen in the UPGMA analysis of the actinobacterial patterns (Fig. 3(b)). All replicates of treatment II (a–c) taken 21 dpi formed a separate cluster with less than 60% similarity with the main cluster. The patterns of replicates of the main cluster were highly similar to each other (more than 80%) except for two replicates of treatment III taken 21 dpi.
Analysis of β-proteobacterial populations in soil samples amended with naphthalene with no inoculation (II) and with inoculation (III) of P. putida KT2442 (pNF142) (0, 3 and 21 dpi). (a) DGGE community fingerprints of β-proteobacterial populations. The relative band position of the band corresponding to Burkholderia sp. is shown (A). Bacterial standard (ST) . (b) Dendrogram constructed with the β-proteobacterial DGGE profiles. The differences between profiles are indicated by percentage similarity.
Analysis of actinobacterial populations in soil samples amended with naphthalene with no inoculation (II) and after inoculation (III) of P. putida KT2442 (pNF142) (0, 3 and 21 dpi). (a) DGGE community fingerprints of actinobacterial populations. The relative position of the band corresponding to Nocardia vinaceae is shown (B). Bacterial standard (ST) . (b) Dendrogram constructed with the actinobacterial DGGE profiles. The differences between the profiles are indicated by percentage of similarity.
3.3 Effect of naphthalene on the soil bacterial community
To evaluate the effect of the naphthalene contamination on the structural composition of the soil bacterial community, 16S rRNA gene fragments amplified from total DNA and cDNA of soil samples taken at 0, 3 and 21 dpi from treatments I and II were analysed by DGGE (DNA: Fig. 4(a); cDNA: data not shown). Three independent replicates per treatment and per sampling time analysed were run in blocks next to each other.
Analysis of bacterial populations of non-inoculated soil microcosms comparing samples from microcosms not contaminated (I) and those artificially contaminated with naphthalene (II). (a) DGGE community fingerprints of bacterial populations. The band position of populations enhanced by naphthalene contamination is shown (A and B). Bacterial standard (ST) . (b) Dendrogram constructed with the bacterial DGGE profiles. Differences between the profiles are indicated by percentage of similarity.
The comparison of DNA-based DGGE profiles of samples taken at 0 dpi was only possible for the more intense bands because PCR amplification of DNA from naphthalene-amended soils seemed to be inhibited, in particular for replicate C (Fig. 4(a)). No differences were seen for the dominant bands. Comparison of DGGE profiles of samples taken 3 dpi also allowed the conclusion that the structural composition of the bacterial community seemed not to be affected by amendment with naphthalene. However, this picture changed 21 days after the start of the experiment. While the profiles of all other samples shared more than 70% similarity, the profiles of samples from naphthalene-amended soil taken 21 dpi were strikingly different and less than 40% similar to the other profiles (Fig. 4(b)). The same trend was found for RNA-based profiles (data not shown). Two new dominant bands, A and B, appeared in all three replicates of DNA- and RNA-based patterns from soils amended with naphthalene 21 dpi (Figs. 4(a) and 5).
PCR-DGGE analysis of 16S ribosomal RNA gene fragments amplified from DNA and RNA (cDNA) from bacterial populations of non-inoculated soil microcosms comparing samples from microcosms not contaminated (I) and artificially contaminated with naphthalene (II). Concomitantly, arrows indicate bands present in the DNA or RNA profiles which cannot be seen or have a reduced intensity when both profiles are compared with each other.
These bands were excised and re-amplified. The re-amplified DNA from band A resulted in one dominant band. The corresponding GC-less PCR product was cloned and four clones were sequenced (Table 3). The sequences of all four clones derived from band A in the patterns of soil amended with naphthalene were most similar to Burkholderia sp. RP007. While two of the clones had approximately 99% similarity to the RP007 sequences of the other clones had 98% and 97% similarity. The re-amplified DNA of band B from the bacterial DGGE patterns resulted in several bands when checked by DGGE and thus band B was extracted from the actinobacterial patterns. The sequences of the four clones derived from band B were identical and showed 98.85% similarity to Nocardia vinacea. The DGGE patterns of actinobacterial communities revealed that the Nocardia sp. population appearing in naphthalene-contaminated soil 21 dpi was hardly visible in the soil without naphthalene (data not shown). In contrast, a band with the identical electrophoretic mobility to Burkholderia sp. was detectable in the control soil (I) but only after enrichment for β-proteobacterial 16S rRNA gene fragments (data not shown). This population was obviously enriched in the presence of naphthalene because this band was not detectable in the bacterial patterns of the control soil (I) at day 21 (Fig. 4(a)). Overall, the DNA and RNA-based patterns were remarkably similar as exemplified in Fig. 5 for treatment II. When comparing DNA and RNA-based patterns, differences in the relative band intensity were observed only for a few bands. In the upper part of the gel (low denaturant concentrations) some bands were detected in the DNA-based patterns which were not detected or much fainter in the RNA-based patterns.
3.4 Detection of naphthalene-degrading genes and their expression in soil
Published primer systems targeting two genes coding for key enzymes in the naphthalene degradation pathway, the nahAc gene encoding the naphthalene-1,2-dioxygenase large subunit  and the nahH gene coding for the catechol-2,3-dioxygenase , were used to monitor the presence of the gene (DNA) and its expression (RNA). Both genes are localised on plasmid pNF142. Using the primer pair Ac114F and Ac596R (Table 1) the nahAc gene was detected in DNA from all soil samples inoculated with KT2442 (pNF142) (III) but not in soils from treatments (I) and (II). The nahAc gene could also be amplified from reverse transcribed RNA extracted from soil samples of treatment (III), indicating that this gene was expressed under soil conditions at all time points analysed (Fig. 6). The nahH gene was amplified from DNA of all soil samples of treatment (III) by means of primers 23DOF and 23DOR (Table 1) but a weak signal was also obtained from one sample of the treatment (II) at 11 dpi (data not shown). However, PCR amplicons were obtained from reverse transcribed RNA only for samples taken 3 dpi with KT2442 (pNF142) but not from cDNA of samples taken after 0 or 21 dpi (data not shown).
Southern blot hybridisation to nahAc of PCR products from DNA or RNA (cDNA) extracted from soil samples amended with naphthalene with no inoculation (II) and after inoculation (III) of P. putida KT2442 (pNF142) (KT) (0, 3 and 21 dpi). DIG-labelled DNA molecular weight marker VI (Roche).
Because the 16S rRNA gene sequence of one of the enriched populations in treatment (II) was almost identical to the sequence of Burkholderia sp. RP007, which carries the naphthalene dioxygenase gene phnAc, a primer system described by Lloyd-Jones et al. , targeting this gene, was used. Indeed, strong phnAc amplicons could be detected from DNA of all replicates of treatment (II) and from replicate b of treatment III 21 dpi (see Fig. 7). The detection of phnAc correlated with the appearance of the Burkholderia band in the bacterial DGGE patterns. However, the expression of this gene could not be detected in any sample from treatment II but curiously it was detected in replicate b of treatment (III) at 21 dpi (data not shown). The phnAc PCR product obtained from DNA of treatment (II) (mix of all three replicates) was cloned and sequenced. The phnAc gene was re-amplified from all 25 clones and PCR products were digested with MspI. The restriction patterns obtained showed that most of the clones had identical patterns (18/24). The remaining clone fragments had one strong band missing or had additional bands (data not shown). However, sequencing showed that all sequences were highly related to the phnAc sequence of Burkholderia sp. RP007 (more than 99% similarity).
Naphthalene dioxygenase like phnAc gene detection by PCR amplification using the primer sets developed for Burkholderia sp. strain RP007 . The PCR products were obtained from DNA samples extracted from soil microcosms not contaminated (I), artificially contaminated with naphthalene with no inoculation (II) and microcosms artificially contaminated with naphthalene inoculated with P. putida KT2442 (pNF142) (III).
3.5 Chemical analysis
The addition of naphthalene as crystals to the soil led to rather large variations in naphthalene concentrations at the beginning of the experiments (195 ± 142 mg kg−1 treatment II and 225.3 ± 156.7 mg kg−1 treatment III) and after 3 days (335 ± 185.3 mg kg−1 treatment II and 40 ± 37.3 mg kg−1 treatment III). These variations were much smaller after 21 days (1 ± 1 mg kg−1 treatment II and 68 ± 8.5 mg kg−1 treatment III) probably due to sublimation of the substrate in the soil leading to a much more uniform distribution of the substrate in the soil pores.
Despite the low recovery rate and the variations of naphthalene in the samples during the first 3 days, very clear differences could be seen between the soil samples of treatments III and II. The data demonstrate that no further decrease in the naphthalene concentration occurred in treatment III because naphthalene concentrations at 21 dpi were similar to those at 3 dpi. One interesting exception from this observation is microcosm b of treatment III where no naphthalene could be found after 3 weeks (data not shown). This finding corroborates with the observation of a deviating microbial community structure in this sample after 3 weeks. In soils from treatment II hardly any reduction in substrate concentration could be found after 3 days. However, in clear contrast to treatment III, all three replicates of treatment II contained naphthalene concentrations near the detection limit of 1 mg kg−1 soil at the last sampling point.
In contrast to previously published studies on the survival of inoculant strains, the emphasis of this study was on the microbial community response to naphthalene and to the inoculation of naphthalene-contaminated soil with the naphthalene-degrading strain P. putida KT2442 (pNF142). The ability of P. putida KT2442 (pNF142) to colonise bulk soil artificially contaminated with naphthalene and to express genes coding for key enzymes for naphthalene degradation was primarily assessed in a community context using DNA and RNA-based methods.
In the soil microcosm experiments performed in this study, the population density of the inoculant strain KT2442 (pNF142) increased during the first days by more than 1 order of magnitude. A similar increase in the population density of P. putida KT2442 carrying an IncPα plasmid was observed in manure-treated soil under field conditions . This kind of increase was not observed in soil plots which were not treated with manure. Although in this study the survival of the inoculant strain was not followed in soil which was not amended with naphthalene, it is tempting to speculate that nutrient availability is a key factor for the survival of this strain in bulk soil. In contrast to the good survival in bulk soil observed in the presence of manure  or in the presence of high concentrations of naphthalene, rather poor survival was observed in pot experiments for P. putida KT2442 in bulk soil poor in organic matter . Similarly, the cfu numbers of P. putida KT2442 introduced into non-vegetated field soil decreased constantly and the strain showed good colonisation of the rhizosphere when introduced as seed inoculum of corn and beans. The levels detected in the bulk soil were 1–2 orders of magnitude lower . Due to the release of root exudates and rhizodeposits, the rhizosphere is considered to be a nutrient-rich environment compared to bulk soil and the availability of nutrients seems crucial for the colonisation ability of KT2442. In this study, KT2442 pNF142 showed good colonisation of soil, most likely due to the availability of a substrate degradable by the inoculant strain.
The DGGE analysis of 16S rRNA gene fragments amplified with bacterial primers from DNA or RNA showed only minor differences between the DNA- and RNA-based patterns for all samples analysed. This observation does not seem to be unusual for bulk soils. Nogales et al.  found good correspondence of sequences in clone libraries derived from DNA or RNA extracted from contaminated soils. DGGE patterns are a display of the numerically dominant ribotypes . Due to the numerical dominance of KT2442 (pNF142), the number of bands observed in the bacterial DGGE patterns was drastically reduced, suggesting tremendous changes in the bacterial community composition. Only after using group-specific primers which did not amplify the 16S rRNA gene of the inoculant strain did it become clear that at least the groups analysed were less severely affected by inoculation. Other studies have used DGGE analysis of 16S rRNA gene fragments amplified from DNA to follow inoculant strains and their effect on bacterial community composition [22,23]. While inoculant strains can only be detected in the community patterns when they belong to the numerically dominant populations, high abundance of the inoculant might disturb evaluation of potential effects on the composition of the bacterial community. This study showed that the use of group-specific primers can help overcome this problem. The use of group-specific primers to generate 16S rRNA gene fingerprints offers not only the advantage of reducing the complexity and detecting populations which are less abundant [11,16], but also of evaluation of the effects of inoculant strains on the structural composition of the bacterial community, which would otherwise have been masked by numerically dominant inoculant strains.
One finding of this study was that at least two bacterial populations were enriched from soil artificially contaminated with naphthalene which was not previously exposed to this pollutant. Twenty-one days after naphthalene contamination (II), two dominant bands appeared in all three replicates of the DGGE patterns which were not detected in patterns of non-contaminated soil samples (I). Based on sequence analysis, the enriched bacterial populations showed highest similarity with genera Burkholderia and Nocardia. Isolates belonging to these genera have previously been reported as naphthalene or phenanthrene degraders [24–26]. Interestingly, in a recent study by Wilson et al.  most of the naphthalene-degrading soil isolates from a hillside adjacent to a seep belonged to different Burkholderia species, while all naphthalene-degrading bacteria isolated from sediments taken from the coal tar deposit showed highest similarity to different Pseudomonas species . The sequence of band A (Fig. 2 and 4(a)) showed highest similarity to the 16S rRNA gene sequence of Burkholderia sp. RP007, which was isolated from a PAH-contaminated site in New Zealand . It is unclear whether the slight differences in 16S rRNA gene sequences of the different clones sequenced result from 16S rRNA operon heterogeneities or from different but phylogenetically closely related populations. While in non-inoculated contaminated soil (II) a band co-migrating with the Burkholderia band could be detected in the β-proteobacterial fingerprints, a band corresponding to the enriched Nocardia population was barely detectable in the actinobacterial patterns. Thus it seems that this population was abundant at cell densities below the limit of detection. As in this study the enrichment of Burkholderia and Nocardia populations in naphthalene-contaminated soils was shown only by nucleic acid-based techniques, it remains speculative whether or not the enriched populations are indeed naphthalene degraders. However, it is most likely that they were enriched in response to the naphthalene contamination because these populations were either not observed or only present as weaker bands in the DGGE patterns in non-contaminated soil (treatment I). The relative abundance of the populations enriched in the presence of naphthalene was much lower when the soil was inoculated with KT2442 (pNF142) (treatment III; except replicate b). Furthermore, the intensity of band A correlates with degradation of naphthalene shown by HPLC analysis.
The nahAc and nahH genes were amplified from DNA extracted from soil samples of treatment III at all sampling times but not from soils which did not receive the inoculant strains. Obviously, nahAc- and nahH-containing bacterial populations were not enriched in soils contaminated with naphthalene (treatment II). This confirms findings by Laurie and Lloyd-Jones  who reported that the Burkholderia-derived naphthalene dioxygenase phnAc gene was present in different pristine soils from Siberia, the Antarctic and New Zealand but not the nahAc gene. Because the sequence of the enriched Burkholderia sp. populations showed highest similarity to the 16S rRNA gene sequence of Burkholderia sp. strain RP007, carrying a divergent set of PAH catabolic genes (phn genes), PCR amplification was used to test whether the phnAc gene was also enriched in our experiments. The detection of phnAc correlated with the appearance of the Burkholderia band in the bacterial DGGE patterns, supporting the idea that populations capable to degrade naphthalene were indeed enriched from soil with no previous history of naphthalene contamination.
By using PCR amplification of cDNA combined with Southern blot hybridisation we could furthermore demonstrate nahAc gene expression in soils of treatment III. In contrast, nahH expression was detected in one of the three replicates only 3 and 21 dpi, respectively. There may be several reasons for the fact that nahH gene expression was not detected at all sampling times and in particular not at 0 dpi. Gene expression of nahH might be induced after exposure to naphthalene, or mRNA might be less stable compared to the transcript of nahAc.
After 3 weeks two of the three microcosms (a, c) containing strain KT2442 (pNF142) run in parallel (treatment III) showed only moderate degradation of the hydrocarbon, while the third and all three microcosms amended only with naphthalene (treatment II) showed a naphthalene concentration below the detection limit. A comparison of the β-proteobacterial community patterns at 21 dpi showed that all microcosms with excellent naphthalene degradation fell into one separate cluster (IIa–c, IIIb) that shared only 60% similarity with the patterns of the remaining profiles.
The finding that a rapid initial decrease in the naphthalene concentration occurred in all three replicates of treatment III and the subsequent cessation of its removal might be explained by decreased bioavailability of naphthalene. Naphthalene is a hydrophobic organic contaminant and its bioavailability can be reduced due to its sequestration by the mineral and organic matter fraction of soil . The longer the time of contact the lower is its bioavailability. In contrast to treatment III, the microbial community of treatment II showed a higher ability to reduce naphthalene concentrations near to 1 mg kg−1 (21 dpi). The two populations related to bacteria belonging to the genus Burkholderia and Nocardia were enriched, and these may have specific physiological properties that could have contributed to enhanced bioavailability of naphthalene. The hydrophobic cell surfaces and the production of biosurfactants, which facilitate the degradation of less available PAH, have been described at least for members of the Nocardia–Rhodococcus–Mycobacterium group [29,30]. Wagner-Döbler et al.  showed that biphenyl-degrading bacteria could be enriched from soils and sediments of both polluted and non-polluted sites after 6 months. The majority of the biphenyl-degrading isolates belonged to the genus Rhodococcus.
A link between structural diversity and function could be made because the observed enrichment of a population with highest similarity of the 16S rRNA gene to Burkholderia sp. RP007 was confirmed by detection of a functional gene (phnAc). Although KT2442 (pNF142) showed excellent survival and expression of genes involved in naphthalene degradation, its introduction prevented enrichment of indigenous naphthalene degraders. Thus we suggest that in soils affected by PAH contamination (e.g. after an oil spill) degrading potential of the indigenous soil bacterial community in the first phase should be exploited before the need to introduce a degrading strain or microbial consortium in a second phase is considered.
This work has been supported by a DAAD postdoctoral fellowship and by INTAS Grants 99-01487 and 01-2383.
Application of denaturing gradient gel electrophoresis and temperature gradient gel electrophoresis for studying soil microbial communitiesElsasJ.D. van, WellingtonE.M.H., TrevorsJ.T.Modern soil microbiology1997Marcel DekkerNew York353373.
(2003) Horizontal transfer of phnAc dioxygenase genes within one of two phenotypically and genotypically distinctive naphthalene-degrading guilds from adjacent soil environments. Appl. Environ. Microbiol. 69, 2172–2181.