Applied Microbiology and Biotechnology

, Volume 64, Issue 4, pp 576–587

Intrinsic bioremediability of an aromatic hydrocarbon-polluted groundwater: diversity of bacterial population and toluene monoxygenase genes

Authors

  • L. Cavalca
    • Dipartimento di Scienze e Tecnologie Alimentari e MicrobiologicheUniversità degli Studi di Milano
  • E. Dell’Amico
    • Dipartimento di Scienze e Tecnologie Alimentari e MicrobiologicheUniversità degli Studi di Milano
    • Dipartimento di Scienze e Tecnologie Alimentari e MicrobiologicheUniversità degli Studi di Milano
Original Paper

DOI: 10.1007/s00253-003-1449-6

Cite this article as:
Cavalca, L., Dell’Amico, E. & Andreoni, V. Appl Microbiol Biotechnol (2004) 64: 576. doi:10.1007/s00253-003-1449-6

Abstract

The functional and phylogenetic biodiversity of bacterial communities in a benzene, toluene, ethylbenzene and xylene (BTEX)-polluted groundwater was analysed. To evaluate the feasibility of using an air sparging treatment to enhance bacterial degradative capabilities, the presence of degrading microorganisms was monitored. The amplification of gene fragments corresponding to toluene monooxygenase (tmo), catechol 1,2-dioxygenase, catechol 2,3-dioxygenase and toluene dioxygenase genes in DNA extracted directly from the groundwater samples was associated with the presence of indigenous degrading bacteria. Five months of air injection reduced species diversity in the cultivable community (as calculated by the Shannon-Weaver index), while little change was noted in the degree of biodiversity in the total bacterial community, as characterised by denaturing gradient gel electrophoresis (DGGE) analysis. BTEX-degrading strains belonged to the genera Pseudomonas, Microbacterium, Azoarcus, Mycobacterium and Bradyrhizobium. The degrading capacities of three strains in batch liquid cultures were also studied. In some of these microorganisms different pathways for toluene degradation seemed to operate simultaneously. Pseudomonas strains of the P24 operational taxonomic unit, able to grow only on catechol and not on BTEX, were the most abundant, and were present in the groundwater community at all stages of treatment, as evidenced both by cultivation approaches and by DGGE profiles. The presence of different tmo-like genes in phylogenetically distant strains of Pseudomonas, Mycobacterium and Bradyrhizobium suggested recent horizontal gene transfer in the groundwater.

Introduction

Aromatic hydrocarbons are common groundwater contaminants associated with petroleum product releases. Microbial degradation of these priority pollutants is regarded as a promising approach with which to clean up contaminated aquifers. Benzene, toluene, ethylbenzene, and xylene (BTEX), components of petroleum products, are of particular concern because of their toxicity and low solubility, representing a potential long-term source of groundwater contamination. All BTEX components can be easily degraded by aerobic microorganisms and the addition of oxygen to contaminated aquifers to stimulate aerobic degradation is a common bioremediation practice (Lovley and Lloyd 2000). The genetic and biochemical phenomena associated with BTEX degradation have been elucidated with a few model strains (Smith 1990; Zylstra 1994), but the diversity of degrading strains in the environment is still relatively unexplored (Gülensoy and Alvarez 1999). A well known example of microbial catabolic diversity can be seen in the different oxidation pathways for toluene catabolism in strains that also grow on other mono-aromatic substrates (Newman and Wackett 1995; Shields et al. 1995; Kukor and Olsen 1990; Whited and Gibson 1991): (1) strains growing on m- and p-xylene normally hydroxylate the methyl group (TOL pathway), (2) strains growing on ethylbenzene normally hydroxylate the aromatic ring (TOM and TOD pathways).

When assessing bioremediation strategy and its outcome in groundwater it is important to have some knowledge of the diversity of the autochthonous microbial communities and their degradative potential. Methods such as ribosomal intergenic spacer analysis (RISA), and denaturing gradient gel electrophoresis (DGGE), which involves the separation of PCR-amplified fragments by differential electrophoretic migration on polyacrylamide gels, have been used to describe bacterial community structures in soil and water (Gray and Herwing 1996; Ranjard et al. 2000).

In this study, the phylogenetic and catabolic diversity of aerobic degrading bacterial communities has been analysed. The cultivable and uncultivable fraction of a BTEX-polluted groundwater sampled from a monitoring well of an air-sparging pilot plant has been studied in order to determine the feasibility of an air sparging approach as remediation strategy for the site.

Materials and methods

Site description and groundwater sampling

The soil and groundwater underneath a paint factory located in the Southern part of Milan have been contaminated, probably since 1963, with aromatic hydrocarbons. The contaminated aquifer consists of silty sand and gravel to a depth of 30 m. Groundwater temperatures are between 18.3°C and 19°C. The groundwater table lies 3–4 m below the ground surface and the water in most of the area flows in a southerly direction. The hydraulic gradient is approximately 0.0033 m m−1 and the hydraulic conductivity for the phreatic aquifer is 80 m day−1; groundwater velocity is approximately 280 m year−1.

Four areas were identified as being responsible for soil and groundwater contamination by benzene, toluene, ethyl benzene, xylenes and naphtha solvent. The contamination, discovered in 1995, was caused by the leakage of underground and surface hydrocarbon storage tanks located in areas A1 and A2, respectively, of the washing plant located in area A3, and by the leakage of underground pipes in area A5, connecting the storage tanks of area A1 to manufacturing departments. In these areas, a soil atmosphere survey was performed in 1996 at 14 points of the site to assess the extent of contamination in the unsaturated soil with a gas vapour Geoprobe. Based on this survey, 14 inspection wells (Z1–Z13) were drilled to a depth of between 9 and 20 m inside the factory, downgradient of the sources and in the background zone, which was hydrogeologically upgradient (Fig. 1).
Fig. 1

Map of the contaminated site. A1, A2, A3, A5 sources of contamination, Z1Z13 groundwater inspection wells. Expanded region Schematic overview of the experimental air-sparging plant. S1 Air-sparging filter; S1–5, S7, S9 groundwater monitoring wells at 4.5 m depth; S6, S8 groundwater monitoring wells at 7.5 m depth; AM1, AM2, AM3 air monitoring wells. Arrow Direction of groundwater flow

In September 1997 an experimental air-sparging plant was installed in area A1 to gain data about the radius of air influence for the design of a full-scale air-sparging plant to form a barrier, placed across the flow path of the contaminated plume, for the reclamation of the groundwater and to prevent spread of the hydrocarbon plume outside the factory. The pilot plant was built in area A1 and consisted of an injection air-sparging filter (S1) screened from 5.7 to 6.7 m from groundwater level, of eight groundwater monitoring wells (S2–S9) [∅ 2 inches (5.08 cm) with a screened portion of 1 m] set 2 m apart, six at 4.5 m depth and two at 7.5 m, and of three air monitoring filters (Fig. 1).

For chemical analyses, groundwater samples were drawn up manually from the eight monitoring wells, for microbiological analyses only from well S2. The samples were put into 1 l sterile serum bottles and kept at 4°C before analysis. Groundwater hydrocarbon concentration was determined according to Standard methods for the examination of water and wastewater (APHA 1992); dissolved oxygen concentration (DO) and temperature were measured weekly using a Microprocessor Oximeter OX1 196 (WTW) in water samples from the eight monitoring wells.

Total cultivable bacteria

Three groundwater samples (1,000 ml) were collected for microbiological analysis: one before the start of the air-sparging, the second after 3 months of air-sparging treatment, and the third after 5 months of treatment. The samples were transported to the laboratory where they were serially diluted in a sterile NaCl solution (9 g l−1). Duplicates (1 ml) of appropriate suspensions were seeded onto solid tryptone yeast extract medium (TYEM) (Mikesell et al. 1991) for heterotrophic bacteria, and into M9 mineral medium (Kunz and Chapman 1981) diluted 10-fold (0.1× M9), with 400 mg l−1 toluene, 1,2,4-trimethylbenzene or a mixture of o-, m-, p-xylene added separately (Origgi et al. 1997), for hydrocarbon-degrading bacteria. The heterotrophic and hydrocarbon-degrading bacteria were counted after 5- or 14-day incubations at 25°C. Controls without aromatic compounds were prepared by inoculating sterile medium with dilutions of the microbial suspensions.

PCR amplification of ribosomal regions

Enzymatic amplification of the almost complete 16S rDNA was performed on DNA extracted from the bacterial isolates and from 40-ml groundwater samples as specified elsewhere (Cavalca et al. 2000), by using eubacterial universal primers P27f and P1495r specific to the nucleotide sequence of an Escherichia coli 16S rDNA gene. A nested PCR reaction for V3 amplification was carried out according to Muyzer and Smalla (1998).

Internal transcribed spacers (ITS) between the small (16S) and the large (23S) ribosomal subunit sequences was amplified by PCR using primers for universal 16S 1406f (5′-TGYACACACCGCCCG-3′) and bacterial-specific 23Sr (5′-GGGTTBCCCCATTCRG-3′) according to Ranjard et al. (2000).

PCR amplification of oxygenase gene fragments

Based on the consensus of known sequences of toluene o-xylene monooxygenase touA (Bertoni et al. 1998), toluene 2-, 3- and 4-monooxygenases (Shields et al. 1995; Kukor and Olsen 1990; Whited and Gibson 1991) and benzene monooxygenase bmoB (Kitayama et al. 1996), degenerate primers were developed for PCR amplification of homologous monoxygenase genes in water isolates. The primers were located at positions 376 and 1,569 of the reference nucleotide sequence of the touA gene from strain Pseudomonas stutzeri OX1. Amplifications were performed in a total volume of 25 μl, containing: 2.5 μl 10× buffer (Bioline, London, UK); 2.5 mM MgCl2; 200 μM dNTPs (Invitrogen, Renfrew, UK); 0.25 μM forward primer 5′-AAGACCTATCCSGARTACGT-3′ and 0.25 μM reverse primer 5′-GGCTGG ATCWGRCCTGCSAGGAA-3′ (Pharmacia, Uppsala, Sweden), and 1 U Taq polymerase (Bioline). Amplifications of different oxygenase gene sequences were obtained by two different thermal profiles: the first started with 94°C for 3 min followed by the addition of 1 U Taq polymerase, then 35 cycles (94°C for 1 min, 55°C for 1 min, 72°C for 1 min) and a final extension step at 72°C for 10 min; in the second, DNA denaturation and Taq polymerase addition was followed by two series of 35 cycles, one with an annealing temperature of 60°C and the second at 50°C.

Amplification of catechol 1,2 dioxygenase (catA) was performed in 50 μl final volume containing: 5 μl 10× buffer (Bioline), 1.25 mM MgCl2, 200 μM dNTPs (Invitrogen), 0.25 μM forward primer 5′-GTC ACC TAC GAC GAA TAC AAC GGC CTC AAG-3′ and 0.25 μM reverse primer 5′-AAG TAG AGC TGG GTG GTG AT-3′ (Invitrogen), 1 U Taq polymerase (Bioline), and 1 μl template DNA. CatA primers corresponded to positions 631 and 1,205 of the reference sequence of the catA gene in strain Rhodococcus erythropolis ATCC4277 (Murakami et al. 1997). The thermal profile was: 94°C for 2 min, Taq polymerase was then added to the PCR mixture and this was followed by 35 cycles (94°C for 40 s, 60°C for 50 s, 72°C for 1 min), then 7 min at 72°C.

Amplification of toluene dioxygenase (todC1) was carried out according to Whyte et al. (1996), that of xylene monoxygenase (xylA,M) according to Cavalca et al. (2000) and catechol 2,3-dioxygenase (xylE) by the method described by Joshi and Walia (1995).

All amplification products were checked on 2% agarose gels in 1× TAE buffer and stained by standard procedures (Sambrook et al. 1989).

Denaturing gradient gel electrophoresis

V3 PCR products were run on a vertical acrylamide gel in a DCODE Universal Mutation Detection System (Bio-Rad, Hercules, Calif.). Denaturing gradient gel electrophoresis (DGGE) was performed with 8% (w/v) polyacrylamide gels in 1× TAE buffer (20 mM Tris acetate pH 7.5, 10 mM sodium acetate, 0.5 mM Na2 EDTA) with a linear chemical gradient ranging from 40 to 80% denaturant. The solutions were prepared by mixing appropriate volumes of two 0–100% denaturant [7 M urea, 40% v/v formamide (Amersham Biosciences, Uppsala, Sweden)] stock solutions. Gels were run at a constant voltage of 70 V for 16 h at 55°C. After completion of electrophoresis, the gels were stained in an ethidium bromide solution (0.5 mg l−1) and documented with a GelDoc system (Bio-Rad).

Sequence analysis and sequence accession numbers

All the strains presented in this work were identified by 16S rDNA nucleotide sequence analysis. The nucleotide sequences of 16S rDNA and of monoxygenase internal fragments were determined according to the Perkin Elmer ABI Prism protocol (Applied Biosystems, Foster City, Calif.). Primers used in the PCR reaction for sequencing products were the same as those in normal PCR reactions. The forward and reverse samples were run on an Applied Biosystems 310A sequence analyser.

Monooxygenase nucleotide sequences obtained in the present study are available on EMBL-GenBank databases under the following accession numbers: AJ294751, for toluene monooxygenase (tmo) of Mycobacterium sp. C3; AJ294752, for tmo of Pseudomonas sp. Z22; AJ295227, for tmo of Bradyrhizobium sp. C9.

Statistical methods

RISA profiles were used to characterise the isolated strains. Fragment size was estimated using a linear regression equation between the molecular mass of the DNA ladder and the log of the distance covered by fragments within the same gel run. A distance matrix was constructed and the UPGA method was used to build a similarity tree by the Jaccard coefficient using an NTSYS software package.

Similarity trees among monoxygenase gene sequences were constructed on the basis of nucleotide sequence alignments obtained by the ClustalW 1.7 program, using the Jalview program of the European Bioinformatic Institute (http://www.ebi.ac.uk/clustalw/) (Thompson et al. 1997).

Growth characteristics of isolates

One hundred colonies grown on TYEM plates were randomly isolated from each water sample and their growth characteristics were studied according to Cavalca et al. (2000). Strains were maintained as glycerol stocks at −20°C.

Degradation experiments

The degradation of toluene, benzene and m-xylene was studied with Pseudomonas sp. Z22, Mycobacterium sp. C3 and Azoarcus sp. H54 strains. They were grown as liquid cultures in 100 ml vials each containing 15 ml 0.1× M9 medium supplemented with 250 mg l−1of the appropriate aromatic compound. Each vial was inoculated with 1 ml cell suspension to an OD600 of 0.15 measured using a Beckman model DU 640 spectrophotometer, after growth on the same medium. The vials were then sealed with Teflon-coated grey butyl rubber stoppers and aluminium crimps and incubated at 30°C on a rotary shaker. At defined incubation times three replicates and three non-inoculated controls were sacrificed to determine the extent of degradation and the production of metabolites by HPLC analysis according to Cavalca et al. (2000). The compounds were detected by their UV absorbance at 254 nm using a UV-975 Intelligent Jasco detector, and identified by comparing the HPLC retention times with those of authentic standards. The aromatic compound concentration in the samples was calculated using standards of known concentration.

Chemicals

Benzene, toluene, o-, m-, p- isomers of xylene and 1,2,4-trimethylbenzene (purity>99%) were purchased from Merck (Darmstadt, Germany). Ethylbenzene, o-, m-, p- isomers of cresol, benzoic acid, 3-methylcatechol, 4-methylcatechol, catechol, 3,4-dihydroxybenzoic acid, 1- and 2-phenylethanol and phenol (purity>99%) were purchased from Sigma-Aldrich (Steinheim, Germany).

Results

Site and groundwater characterisation

Substantial differences in the unsaturated atmosphere of the contaminated location were found during the survey phase: hydrocarbon concentrations in unsaturated soil interstitial air of areas A1 and A2 ranged from undetectable levels to 18,000 mg m−3, 5,200 mg m−3, 2,100 mg m−3 and 20,000 mg m−3 for toluene, xylenes, ethylbenzene and naphtha solvent, respectively.

Groundwater contamination extended to a depth of 14 m. The higher hydrocarbon concentrations were found in some monitoring wells of areas A1 and A2, where up to 160, 62,500, 6,200, 16,730 and 3,400 μg l–1 benzene, toluene, xylenes, ethylbenzene and naphtha solvent, respectively, were found.

Hydrocarbon concentrations in water from the eight monitoring wells in the study pilot plant area are reported in Table 1. Values decreased over the 5 weeks monitored and only S2 samples still showed high contaminant levels. Table 2 summarises the chemical and microbiological data for S2 samplings monitored for a longer time. At the beginning of the treatment, the water was characterised by anoxic conditions (DO <1 mg l−1; United States Environmental Protection Agency 2002). After 5 months of application, the air-sparging led to an increase in DO from 0.8 to 2.8 mg l−1, and a concomitant reduction of more than 96% of the hydrocarbon level was observed
Table 1

Hydrocarbon concentrations in water from the eight groundwater monitoring wells of air-sparging pilot plant area A1

Well

Sampling in 1997

Contaminant concentration (μg l−1)

Inspection well

Sample

Toluene

Xylenes

Ethylbenzene

Naphtha solvent

S2

I (31 October)

18,239

21,230

3,500

9,692

21 November

17,735

41,058

3,530

6,460

II (5 December)

5,733

17,337

2,829

5,613

S3

I (31 October)

288

695

945

525

21 November

274

443

5

214

II (5 December)

23

85

5

196

S4

I (31 October)

5

2,531

670

631

21 November

5

78

5

257

II (5 December)

5

5

5

245

S5

I (31 October)

26

1,785

5

940

21 November

5

5

5

5

II (5 December)

5

5

5

5

S6

I (31 October)

5

5

5

21

21 November

5

5

5

5

II (5 December)

5

28

5

19

S7

I (31 October)

10,160

18,747

7,240

5,863

21 November

130

79

418

443

II (5 December)

5

5

5

178

S8

I (31 October)

58

245

42

146

21 November

5

87

8

136

II (5 December)

5

5

5

5

S9

I (31 October)

4,604

7,794

3,109

2,603

21 November

37

190

45

193

II (5 December)

5

21

5

168

Table 2

Chemical and microbial characteristics of groundwater samples from well S2. Values are the mean of two determinations. DO Dissolved oxygen;T toluene; X mixture of o, m, and p-xylene;EtB ethylbenzene; AHB aerobic heterotrophic bacteria; T-DB toluene-degrading bacteria; X-DB xylene-degrading bacteria;TMB-DB 1,2,4-trimethylbenzene-degrading bacteria; cfu colony forming unit; MPN most probable number

Sample

DO (mg l−1)

Contaminant concentration (μg l−1)

Bacterial concentration

T

X

EtB

AHB (cfu ml−1)

T-DB (MPN ml−1)

X-DB (MPN ml−1)

TMB-DB (MPN ml−1)

I (31 October 1997)

0.8

18,239

21,230

3,502

1.5×106

2.3×104

2.3×104

6.2×103

II (5 December 1997)

1.9

5,733

17,337

2,829

1.1×106

6.2×104

2.3×103

2.3×103

III (24 March 1998)

2.8

88

160

139

1.5×105

6.2×103

2.3×102

2.3×103

At the first groundwater sampling, aerobic heterotrophic bacteria, toluene- and xylene-degrading bacteria, were in the order of 106 cfu ml−1 and 104 MPN ml−1, respectively. After 5 months of air-sparging treatment the number of aerobic heterotrophic and toluene-degrading bacteria had decreased by one order of magnitude (Table 2). A plausible explanation for the low change in the hydrocarbon degrader population could be their presence at maximum field capacity already at the first sampling or to an artificial constancy of population numbers due to a dilution effect. On the basis of the measured DO of groundwater samples from the eight monitoring wells of the air-sparging pilot plant (data not shown), the radius of air influence was determined to be 4 m.

Retrieval of catabolic genes in nucleic acids extracted from groundwater

PCR amplification with primers specific for tmo, todC1, xylA,M, xylE and catA gene fragments was used to assess the presence of BTEX degradation genes.

Expected PCR fragments corresponding to tmo, todC1, xylE and catA were found in the DNA directly extracted from the three groundwater samples (Fig. 2a). The xylA,M gene was never amplified. xylE and catA genes of lower pathways were also retrieved in the environmental DNA samples.
Fig. 2a, b

Detection of catabolic genes in DNA extracted from water at the three samplings. a PCR amplification of toluene monooxygenase (tmo), toluene dioxygenase (todC1), catechol 2,3-dioxygenase (xylE), catechol 1,2 dioxygenase (catA). Lanes: M 1 kb ladder (Bioline, London, UK); I, II, III DNA from water from the first, second and third samplings, respectively; tmo C+ Mycobacterium sp. C3 positive control; todC1 C+ Escherichia coli JM109 pDTG601 positive control; xylE C+ E. coli JM109 pGSH2960 positive control; catA C+ Rhodococcus erythropolis ATCC4277 positive control. b HhaI restriction fragment length polymorphism (RFLP) analysis of tmo amplified from degrading isolates Pseudomonas sp. Z24 and Mycobacterium sp. C3, and from the three groundwater samples

Amplification of tmo fragment from DNA extracted from groundwater

PCR amplification with tmo-specific primers was used to assess the presence of the different toluene degradation genes. The expected mixture of 1,200 bp tmo amplicons was found in the three groundwater communities tested. No additional non-specific bands were present in the amplification products (Fig. 2a). The expected 1,200 bp tmo amplicons from groundwater DNA and isolated strains were purified from the gel and subjected to restriction fragment length polymorphism (RFLP) characterisation with three different enzymes: EcoRI, PvuII, HhaI. Examination of monooxygenase gene RFLP entries in the GenBank and EMBL genetic databases did not reveal patterns similar to those observed in our isolates or in the groundwater DNA, evidencing a previously undescribed diversity (polymorphisms) of monooxygenase genes. The tmo RFLPs from water were in accord with band profiles obtained from the isolated strains (Fig. 2b): in the first two samplings the tmo of Pseudomonas sp. was the predominant profile in groundwater DNA, while in the third sample the tmo of Mycobacterium sp. was predominant. This indicated that all the monooxygenase alleles present in the groundwater communities matched those of the isolated bacteria.

Biodiversity of total cultivable bacteria

One hundred colonies were isolated from the heterotrophic bacteria count plates of groundwater samples at the outset and after 5 months of air-sparging, according to their different morphologies. The two communities were subjected to RISA analysis of the intergenic spacers of the ribosomal operon for phylogenetic characterisation.

According to statistical analysis of RISA patterns (Fig. 3), populations of 27 and 16 RISA operational taxonomic units (OTU) were present at the first and third samplings, respectively. Of 27 RISA OTUs, only four (P1, P2, P4 and P24) were present in both populations, with P24 being always predominant, accounting for 12 and 18 isolates, respectively.
Fig. 3

Distribution of operational taxonomic units of isolates based on ribosomal intergenic spacer analysis (RISA) genotypes at the first (I) and third (III) samplings. The number of occurrences of each RISA-OTU in the corresponding sample is shown in brackets. The tree was constructed using the NTSYS software package with Jaccard coefficient. Distances are marked at junctions: the branch lengths are the percentage mismatch between two nodes. Species identifications were obtained by 16S rDNA nucleotide sequence analysis of representative strains

The distribution of different RISA patterns was homogeneous in the two bacterial populations, there being no distinct spatial separation evident among the branches of strains on the distance tree.

Biodiversity of uncultivable bacteria during 5 months of air-sparging treatment

The biodiversity of the total bacterial population in the groundwater samples, analysed by DGGE profiles, revealed slightly different patterns (Fig. 4). The overall degree of biodiversity did not change significantly during the 5 months of air-sparging treatment.
Fig. 4

Denaturing gradient gel electrophoresis (DGGE) analysis of PCR-amplified V3 16S rDNA region from degrading isolates and from the three groundwater DNA samples. Retrieved sample and numerical occurrence of strains are in brackets. Pseudomonas sp. strain-specific bands are connected to groundwater microbial community bands by dotted lines

The DGGE profiles of some isolated degrading bacteria matched with bands present in the community profiles. In particular, bands from Pseudomonas sp. P24 and P34 RISA OTUs were always found in the DNA profiles of the groundwater samples. These data are in accordance with analysis of cultivable bacteria, showing that strains of the P24 RISA OTU were the most abundant isolates in both the first and third samplings. Bands belonging to other OTUs were not visible in the water community profiles, and probably belonged to low cell number bacteria. Two intense water profile bands did not match with any of the DGGE bands of the isolated strains, suggesting the presence of well adapted strains in both the anoxic (0.8 mg l−1) and aerobic (2.8 mg l−1) groundwater environments that were not cultivable under our experimental conditions.

Oxygenase systems in degrading strains

The catabolic capability of isolated strains was evaluated by recording their aerobic growth response on seven different hydrocarbons, with PCR amplification of a set of catabolic genes for oxygenase systems also being performed (Table 3).
Table 3

Growth characteristics of degrading strains and PCR amplifications of catabolic genes. TToluene; B benzene; EtBethylbenzene; m-, p-X m-,p-xylene; TMB1,2,4-trimethylbenzene; C catechol; RISA ribosomal intergenic spacer analysis

Sample

RISA group

Strains

Identification

Aromatic compoundsa

PCR amplifications

T

B

EtB

m-X

p-X

TMB

C

todC1

tmoA

xylA,M

xylE

catA

I (October 1997)

P5

H10, H33

Microbacterium sp.

+

+

+

P19

H54

Azoarcus sp.

+

+

+

P21

Z12, Z17, Z22, Z23, Z24,

Pseudomonas sp.

+

+

+

+

+

+

+

P24

Z1, Z2, Z3, Z4, Z5, Z6, Z7, Z8, Z9, Z13, Z14, Z18

Pseudomonas sp.

+

+

III (March 1998)

P24

C4, C6, C7, C8, C13, C14, C25, C28, C32, C34, C36, V6, V10, V12, V16, V20, V32, V34

Pseudomonas sp.

+

+

P39

C3

Mycobacterium sp.

+

+

+

+

+

+

P35

C9

Bradyrhizobium sp.

+

+

+

+

+

+

P34

C27

P. veronii

+

+

+

a+Growth after 3 day incubation (OD600nm>0.5) in two subsequent transplants, absence of growth

Eight strains isolated from the first sampling, belonging to three distinct RISA (P5, P19, P21), related to Microbacterium sp., Azoarcus sp., and Pseudomonas sp., respectively, by 16S rDNA sequence analysis, were able to grow on mono-alkylbenzenes and catechol; however, only Pseudomonas sp. strains of P21 were able to also grow on benzene. Monooxygenase degenerate primers amplified a tmo gene fragment corresponding to the expected size in the five Pseudomonas sp. strains of P21 RISA OTU. These isolates were also characterised by the presence of both catA and xylE gene fragments, coding for enzymes for ortho and meta aromatic ring cleavages. None of the genes sought were amplified in the Microbacterium sp., Azoarcus sp. and P. veronii strains, probably because of their low gene sequence homology with the sequences used to design the primers. Further work will be required for the characterisation of these strains.

Three strains isolated from the groundwater at the third sampling, belonging to P39, P35 and P34 OTUs, identified as Mycobacterium sp., Bradyrhizobium sp. and Pseudomonas veronii, respectively, grew on mono-, di- and tri-alkyl benzenes but not on benzene. The Mycobacterium sp. and Bradyrhizobium sp. strains presented a gene fragment using monooxygenase primers for tmo, but no catA or xylE genes were found in these strains.

All the Pseudomonas sp. strains of the most abundant P24 RISA OTU were able to grow on catechol, the central metabolite of different aromatic degradation pathways, but not on any of the hydrocarbons tested. In the PCR experiments these strains presented only xylE gene fragments. None of the toluene-degrading isolates gave positive signals for todC1 or xylA,M genes of the TOD and TOL upper oxygenative pathways.

Sequence analyses of tmo-like genes

Besides a different growth substrate range, the degrading strains also possessed different tmo-like genes. On the basis of the different restriction profiles of the amplicons (Fig. 2b), tmo fragments of Pseudomonas sp., Mycobacterium sp. and Bradyrhizobium sp. were sequenced.

A similarity tree was constructed using archetypal oxygenase genes present in databases (Fig. 5). The tmo gene of strain Pseudomonas sp. Z22 showed 78% homology to benzene monooxygenase (bmoB) of P. aeruginosa JI104 (acc. no. D83068) and 75% to toluene-3-monoxygenase tbhA of Burkholderia cepacia AA1 (acc. no. AF001356). The tmo gene of Mycobacterium sp. C3 had 66% homology to both phenol hydroxylase (pheB) of Ralstonia eutropha JMP134 (acc. no. AF065891) and toluene-3-monooxygenase (tbuA) of Pseudomonas pickettii PK01 (acc. no. U04052). The tmo gene of Bradyrhizobium sp. C9 had 60% homology with both pheB and with tbuA. No homology was found with the methyl monooxygenase xylA,M gene of the TOL pathway.
Fig. 5

Phylogenetic tree of tmo gene fragments generated from CLUSTALW 1.7 alignment of translation products of tmo fragments of Pseudomonas putida Z22, Mycobacterium sp. C3, Bradyrhizobium sp. C9, and published sequences using the neighbour joining method. Sequences obtained in this work have been deposited in GenBank/EMBL databases. Distances are marked at junctions: the branch lengths are the percentage mismatch between two nodes

Degradation pathways in Pseudomonas sp. Z22, Mycobacterium sp. C3 and Azoarcus sp. H54

Degradation experiments were carried out using three strains: Pseudomonas sp. Z22 and Mycobacterium sp. C3, as representative strains possessing the two different tmo genes, and Azoarcus sp. H54, as model of a degrading group whose catabolic genes were not amplified under our experimental conditions.

During the degradation experiments of toluene with the growing cells of Pseudomonas sp. Z22, HPLC analysis identified p-cresol, 4-methylcatechol and 3,4-dihydroxybenzoic acid (protocatechuic acid) in the culture broths (Fig. 6a). The formation of these metabolites suggested that toluene was hydroxylated at the para position, yielding p-cresol, which was further oxidised to 3,4-dihydroxybenzoic acid, according to the Pseudomonas mendocina KR1 pathway (Whited and Gibson 1991), and to 4-methylcatechol, as reported for Burkholderia sp. strain JS150 (Johnson and Olsen 1997). Benzene was degraded by Z22 (Fig. 6d), through the formation of phenol and catechol by means of monoxygenation reactions, as reported for P. aeruginosa JI104 (Kitayama et al. 1996).
Fig. 6

Time course of toluene (ac), benzene (d) and m-xylene (e) degradation by growing cells of Pseudomonas sp. Z22 (a, d), Mycobacterium sp. C3 (b, e) and Azoarcus sp. H54 (c): ◆ toluene, ○ p-cresol, x 4-methylcatechol, ▲ 3-methylcatechol, − 3,4-dihydroxybenzoate, ■ cresol, ◇ benzene, + phenol, △ catechol, ● m-xylene, □ 3-methylbenzoic acid, * 3-methylsalicylic acid

Growing cells of Mycobacterium sp. C3 degraded 250 mg l−1 toluene slower than Pseudomonas sp. Z22. HPLC analyses of the culture broths revealed three transient peaks: one peak corresponded to either benzoic acid or m-cresol, since both compounds have the same retention time, the other two to catechol and 3-methylcatechol (Fig. 6b). The degradation of m-xylene by C3 (Fig. 6e) occurred through the formation of 3-methylbenzoic acid and 3-methylsalicylic acid, suggesting that the monooxygenase could oxidise both the methyl group and the aromatic ring.

Azoarcus sp. H54 degraded 250 mg l−1 toluene in 12 h, through the formation of cresol (either o- or m-) and 3-methylcatechol (Fig. 6c). H54 also degraded ethylbenzene at the same rate (data not shown). Although we did not succeed in amplifying any monooxygenase-like genes in this strain, these intermediates indicated the presence in the strain of a monoxygenase activity directed to the aromatic ring only of alkyl-substituted benzenes as the strain did not grow on benzene.

Discussion

The presence of hydrocarbon-degrading bacteria and the decrease over time in hydrocarbon concentrations in groundwater samples from the eight monitoring wells (Tables 1, 2) were the presupposition for the groundwater bioremediation by building a series of air-sparging wells to design a biobarrier. The increased DO (>1.9) of the water following air-injection and the concomitant reduction in hydrocarbon concentration are consistent with a predominantly biodegradation-driven mass reduction in the groundwater, supporting the growth of bacteria. This is in accordance with the data of Chiang et al. 1989, by which BTEX aerobic degradation rates can take place when DO is above 2 mg l−1. On the basis of these results, however, we cannot conclude that biological processes alone contributed to the attenuation measured. Abiotic processes such as dilution by groundwater flow or volatilisation by air-sparging could also explain the observed loss.

The amplification of the genes for todC1, tmo, and meta and ortho cleavage dioxygenases (xylE and catA) in groundwater DNA confirmed the presence of different degradative abilities in the bacterial community. Under limiting oxygen concentration, strains with tmo were found to be more efficient than those with todC1, while at high oxygen concentration, strains carrying the dioxygenases degraded toluene at significantly higher rates (Leahy and Olsen 1997). The presence of a low DO content in the aquifer before the start of the air-sparging treatment (0.8 mg l−1) might have favoured microorganisms with TOM-like pathways, and only monooxygenase-carrying strains were isolated. In any case, the presence of todC1 genes in groundwater DNA did not exclude the presence of this system in uncultivable or low cell number bacteria. In fact the different catabolic pathways encountered in the DNA of water samples indicated that BTEX degradation could have occurred at different oxygen concentrations (anaerobic and/or aerobic), and at different rates. Under our PCR conditions, the tmo genes detected in the cultivable degrading isolates were identical to those amplified in the contaminated groundwater, as evidenced by the presence of the same tmo RFLP patterns.

The cultivable bacterial populations analysed at two sampling times had different levels of diversity, as revealed by the Shannon-Weaver diversity index (H’). H’ was equal to 1.26 and 0.99 for the first and the third sampling, respectively, and the currently most-contaminated sample (Table 2) showed the highest diversity, as reported by Bakermans and Madsen (2002). The RISA pattern analysis revealed that 74% of the isolates from the first sampling and 37.5% of the isolates from the third sampling possessed unique RISA patterns, in accordance with the hypothesis that aquifer microbial communities often present unique microbial species, although they are less diverse than soil habitats (Pederson 2001). The different taxonomic units were distributed homogeneously between the two samplings, indicating that the cultivable bacterial population did not change significantly during the 5 month air-sparging treatment. The limited number of DGGE bands that differed in the samplings supported the small change of diversity in the uncultivable population.

Cluster analyses indicated the occurrence of α, β and γ Proteobacteria and of a low G-C content Gram-positive group. Five main genus-related groups were present within the groundwater community, with Pseudomonas sp. and Azoarcus sp. being the most represented, comprising both degrading and non-degrading strains, as observed in other polluted aquifer bacterial populations (Ridgway et al. 1990; Pelz et al. 2001). The DGGE bands of the total bacterial population confirmed the consistent dominance of the Pseudomonas group.

Hydrocarbon-degrading bacteria were not dominant. Instead, Pseudomonas P24 RISA-OTU was always dominant in the cultivable population of all the samplings, probably due to the ability of all the strains to grow on catechol, the central metabolite of many aromatic compound degradation pathways, and to adapt well to both highly polluted anoxic conditions and to low polluted aerobic conditions (Commission of the European Union 1980).

Degradative ability was spread over different genera, reflecting the importance of functional biodiversity in polluted environments to decontaminate mixtures of compounds. After the initial monoxygenation of toluene on the aromatic nucleus to yield p-cresol, Pseudomonas sp. Z22 was able to degrade this intermediate via protocatechuic acid and 4-methylcatechol. This suggests the occurrence of diverse pathways operating simultaneously or that the monooxygenase-like system carried by Z22 could perform two different oxygenative attacks: one towards the aromatic nucleus and one towards the methyl group, as demonstrated for other strains during the degradation of p-cresol (Heinaru et al. 1999). These pathways were found separately in Burkholderia sp. JS150 (Johnson and Olsen 1997) and P. mendocina KR1 (Whited and Gibson 1991), both carrying monooxygenase-like genes resembling that of strain Z22, even if not completely homologous. The presence of xylE and catA genes in Pseudomonas sp. strain Z22, and in all the Pseudomonas sp. strains of the P21 group, indicated the capability of these strains to metabolise p-cresol via meta and ortho cleavage, according to Pseudomonas putida NCIB 9869 pathways (Hopper and Taylor 1975).

Different pathways operating simultaneously to degrade toluene seemed to occur also in Mycobacterium sp. C3: benzoic acid and 3-methylcatechol, the first deriving from monoxygenation towards the methyl group and the second from attack on the aromatic nucleus, were identified. This behaviour has been documented for other Gram-positive strains that metabolise o-xylene through methyl group and ring oxidation (Vanderberg et al. 2000).

The monooxygenase systems of Pseudomonas sp. Z22 and of Mycobacterium sp. C3 had different substrate ranges. The former recognised benzene and mono-alkylbenzene, the latter only mono-, di- and tri-alkylbenzene. The effective role of monooxygenase in the oxidation of toluene will be the subject of further work.

In Azoarcus sp. strain H54, which was able to degrade toluene and ethylbenzene but not benzene, the monooxygenase activity was directed only towards the aromatic nucleus.

Deduced aminoacid sequence analysis grouped the tmo genes into two separate clusters: the first comprising monooxygenases of the Gram-positive high G-C content Mycobacterium sp. C3 and of α Proteobacterium Bradyrhizobium sp. C9, the second comprising sequences of published soil γ Proteobacteria strains and of Pseudomonas sp. Z22. Homologies encountered were low, ranging from 60% to 78%, suggesting that the monooxygenases retrieved from the groundwater strains might have evolved differently from those of soil bacteria published in GenBank. The presence of the same system in Mycobacterium sp. C3 and Bradyrhizobium sp. C9 suggests convergent adaptation of different microorganisms to degrade a multiplicity of substrates. The clustering together of the two sequences, despite their being retrieved from phylogenetically distant strains, suggests that a horizontal gene transfer of existing genes could have occurred between indigenous microorganisms. Dissemination of catabolic plasmids and transposons is a possible mechanism for genes to adapt to the degradation of xenobiotic compounds in natural environments (Van Der Meer et al. 1998; Hollowell et al. 1999), and complementary catabolic pathways possibly play an important role in the evolution of new degradative capabilities (Top et al. 1998).

Following this laboratory and pilot plant investigation to determine the outcome of the experimental plant, a full scale air-sparging plant for reclamation of the site has been installed.

Acknowledgements

This work was supported by a CNR research grant and by Piano Nazionale Biotecnologie Vegetali, Area di Ricerca “Microrganismi Utili”, MIPA, Italy.

Copyright information

© Springer-Verlag 2003