OUP user menu

The role of a groundwater bacterial community in the degradation of the herbicide terbuthylazine

Anna Barra Caracciolo , Carmen Fajardo , Paola Grenni , Maria Ludovica Saccà , Stefano Amalfitano , Roberto Ciccoli , Margarita Martin , Alicia Gibello
DOI: http://dx.doi.org/10.1111/j.1574-6941.2009.00787.x 127-136 First published online: 10 December 2009


A bacterial community in an aquifer contaminated by s-triazines was studied. Groundwater microcosms were treated with terbuthylazine at a concentration of 100 μg L−1 and degradation of the herbicide was assessed. The bacterial community structure (abundance and phylogenetic composition) and function (carbon production and cell viability) were analysed. The bacterial community was able to degrade the terbuthylazine; in particular, Betaproteobacteria were involved in the herbicide biotransformation. Identification of some bacterial isolates by PCR amplification of the 16S rRNA gene revealed the presence of two Betaproteobacteria species able to degrade the herbicide: Advenella incenata and Janthinobacterium lividum. PCR detection of the genes encoding s-triazine-degrading enzymes indicated the presence of the atzA and atzB genes in A. incenata and the atzB and atzC genes in J. lividum. The nucleotide sequences of the PCR fragments of the atz genes from these strains were 100% identical to the homologous genes of the Pseudomonas sp. strain ADP. In conclusion, the results show the potential for the use of a natural attenuation strategy in the treatment of aquifers polluted with the terbuthylazine. The two bacteria isolated could facilitate the implementation of effective bioremediation protocols, especially in the case of the significant amounts of herbicide that can be found in groundwater as a result of accidental spills.

  • terbuthylazine
  • groundwater
  • bacterial degradation
  • Betaproteobacteria
  • Advenella incenata
  • Janthinobacterium lividum


s-Triazines are among the most commonly used herbicides in the world. In recent years, concern has been growing about the persistence, mobility and toxicity of triazines and their metabolites, owing to the residual concentrations of these compounds detected in aquifers (Tappe et al., 2002). In many European countries, a significant proportion of monitored groundwater is contaminated by triazines, primarily atrazine, terbuthylazine (TBA) and their desethyl-degradation metabolites; the concentration of these contaminants is often >0.1 μg L−1, the maximum admissible concentration under the EC legislation (EC 98/83EEC) for potable water (Tappe et al., 2002; Guzzella et al., 2003, Guzzella et al., 2006; Hildebrandt et al., 2008). There may also be nonsealed wells located in agricultural areas that become significantly (in the agricultural dose range) contaminated by accidental spills from the devices used to apply the herbicide (Fait et al., 2007).

Among s-triazines, terbuthylazine is the most persistent in surface environments (Guzzella et al., 2006; Carafa et al., 2007); in groundwater, its half-life ranges from 263 to 366 days (Navarro et al., 2004a). Toxic pollutant contamination of groundwater is a very serious environmental problem and is a risk to human health, especially because many communities depend on groundwater as the sole or as a major source of drinking water. In fact, >65% of the drinking water produced in Europe is sourced from groundwater. As the number of aquifers that cannot provide potable water is increasing, it is becoming necessary to study methods that will allow the removal of herbicides from groundwater, including physical, chemical and biological approaches.

The majority of studies addressing herbicide contamination of groundwater have focused on the rate of herbicide degradation. Only in the past few years has microbial acclimation to herbicides in subsurface aquifer environments been investigated (De Lipthay et al., 2003). Natural attenuation has been found in contaminated groundwater in situ (Tuxen et al., 2002; Williams et al., 2003) and in laboratory experiments using indigenous bacteria from contaminated sites (Mirigain et al., 1995; Franzmann et al., 2000; Johnson et al., 2000; Pucarevic et al., 2002; Harrison et al., 2003). To date, however, there have been few studies on natural attenuation in s-triazine-contaminated groundwater (Grenni et al., 2009b). The natural attenuation of other herbicides has only been found when the concentrations exceed 40 μg L−1 and when it is associated with acclimated bacterial communities displaying positive degradation of the compound (Johnson et al., 2000; Broholm et al., 2001; Tuxen et al., 2002). In fact, the recalcitrant behaviour of the herbicide in groundwater could be a result of concentrations that are too low to induce bacterial degradation (Tappe et al., 2002). Biodegradation and mineralization of s-triazines have been shown to be carried out by bacterial consortia and by strains isolated from contaminated sites (Aislabie et al., 2005; Grenni et al., 2009b). Most commonly, the bacterial metabolism of s-triazines has been reported to occur via two different upper pathways. The first pathway involves the enzymes encoded by the atzA-atzB-atzC genes. The second utilizes the initiating enzymes of the hydrolytic reactions encoded by the trz gene family, such as trzN, and the atzB-atzC genes (Shapir et al., 2007). These genes are widespread, highly conserved in bacteria and are often associated with transposable elements on plasmids (Devers et al., 2007; Shapir et al., 2007). Both of these pathways lead to the formation of cyanuric acid and alkylamines as common intermediates, which are ultimately degraded by the enzymes encoded by the atzDEF genes and by amine oxidases, respectively (Shapir et al., 2007).

Although there have been several studies showing biotic and abiotic triazine degradation in soil (Di Corcia et al., 1999; Barra Caracciolo et al., 2005a, Barra Caracciolo et al., 2005b), to our knowledge, degradation of these compounds in groundwater remains to be investigated. The fact that bacterial strains with the potential capability to degrade these compounds can be found in groundwater has important implications for planning remediation strategies and, in particular, for assessing the natural attenuation time and the implementation of bacterial strains for bioaugmentation purposes. In this study, we assessed the capacity of the autochthonous bacterial community of a shallow aquifer that is chronically contaminated with residual concentrations of s-triazines to degrade terbuthylazine in microbiologically active and sterile groundwater microcosms. Specifically, we studied the bacterial community of the groundwater microcosms, analysing their structure (bacterial abundance and phylogenetic composition) and function [cell viability and bacterial carbon production (BCP)] throughout the experimental period. We also investigated the putative herbicide degradation pathway in the bacteria isolated from this aquifer.

Materials and methods

Groundwater collection and characteristics of the aquifer

The criteria used for selection of the site were as follows: an intensive agricultural area in which the use of the herbicide terbuthylazine was a common practice, and intrinsic aquifer vulnerability (Daly et al., 2000). We selected an alluvial aquifer located near Assisi (PG), Central Italy, on the Petrignano Plane, at 216 m a.s.l. The water table was surficial at 12 m depth. According to the Umbria Regional Environmental Agency's monitoring surveys (2000–2008), the herbicide terbuthylazine and its metabolite desethyl-terbuthylazine are always found in the groundwater at this site in concentrations >0.1 μg L−1. It is also common to find significant nitrate contamination at this site (>100 mg L−1). Some parameters (pH, O2, redox potential, depth and conductivity) were analysed on site, and others were examined in the laboratory. Groundwater samples were collected by a sterile bailer from a well and placed in sterile polyethylene bottles to avoid any contamination. Subsamples were fixed or treated immediately for different purposes. The collected groundwater was stored (24 h maximum) at 4 °C before use. Following a 0.45-μm filtration, dissolved organic carbon was measured using a Shimadzu ASI-5000A Total Organic Carbon Analyser aqueous carbon analyser with a detection limit of 0.050 mg L−1. Moreover, the total cell abundance (no. of bacteria mL−1 water) and cell viability (% live cells/live+dead) were assessed on the same day as the sampling in three replicates, as described below.

Microcosm setup for degradation studies

The experimental set-up consisted of 80 closed glasses (100 mL capacity). Twenty microbiologically active microcosms (TBA) were set up with 50 mL of groundwater and the herbicide terbuthylazine (Dr Ehrenstorfer, Augsburg, Germany) at a concentration of 100 μg L−1. In order to compare the degradation without bacteria, 20 additional microcosms (sterile) were set up with previously sterilized groundwater (120 °C, 20 min) and treated with terbuthylazine (100 μg L−1). Moreover, 20 additional microcosms without terbuthylazine (control) were used as microbiological controls, to assess the effect of the herbicide on the bacterial community. The pH and oxygen content were constantly monitored throughout the experimental period in another 20 herbicide-treated microcosms used only for this purpose. All microcosms were maintained in the dark, gently shaken and incubated at 15 °C (the same temperature as that recorded in the aquifer). Two sacrificial microcosms for each experimental condition (TBA, sterile, control) were collected and subsampled for chemical and/or microbiological analyses at selected times (0, 7, 14, 28, 40, 60, 80, 124 and 175 days). For each analysis, we collected two subsamples from each microcosm (four replicates in total).

Chemical analysis

Terbuthylazine and desethyl-terbuthylazine concentrations were measured immediately after treatment and at various times until day 175. Cyanazine was added to the subsamples (four replicates for each condition) as an internal standard and each sample was extracted twice with methylene chloride. The extract was passed through a layer of anhydrous sodium sulphate to remove any residue of water and dried under a stream of nitrogen at room temperature. The extract was then reconstituted with 150–200 μL of methylene chloride and injected into a GC. The analyses were performed using a Thermo Finnegan Trace 2000 GC/MS (Waltham, MA) equipped with a model AS 2000 autosampler, operating with an electronic impact at 70 eV. The mean recovery was >80%.

Total cell number and viability

The total cell abundance (no. of bacteria mL−1) was determined in four replicates of ethanol-fixed subsamples (four replicates of 5 mL each) by direct count, using 4′-6-diamidino-2-phenylindole (DAPI) as a DNA fluorescence agent (Barra Caracciolo et al., 2005a, Barra Caracciolo et al., 2005b). Cell viability (% live cells/live+dead) was assessed in fresh replicate subsamples (5 mL each) using a two-dye fluorescent bacterial viability kit (Kit Live/Dead® Bacterial Viability Kit, BacLight) that distinguishes between viable (green) and dead (red) cells under a fluorescence microscope (Alonso et al., 2002). We calculated the live cell abundance (no. of live bacteria mL−1) from the total cell abundance, obtained by DAPI counts, multiplied by viability (expressed as % live cells/live+dead).


BCP was estimated by [3H]leucine (NEN Life Science Products, Boston) incorporation measurements, using the microcentrifugation method (Smith & Azam, 1992). Briefly, subsamples of groundwater (1.7 mL each) were amended with 20 nM radiotracer (saturation value) and incubated for 1 h at 20 °C. Zero-time controls were run by killing samples with 100% trichloroacetic acid (TCA, 5% final concentration) 15 min before leucine addition. The extraction of labelled macromolecules was carried out by washing with 5% TCA and 80% ethanol. Each washing step was performed by centrifugation at 15 800 g for 10 min at room temperature. The supernatant was discarded and 1 mL of liquid scintillation cocktail was added to all samples. Radioactivity was detected using a TRICARB 4430 (Packard Bioscience) scintillation counter. The rates of leucine incorporation were converted into units of C per millilitre (ng C mL−1 h−1) by applying the conversion factor of 3.1 kg C produced per mole of incorporated leucine (Kirchman, 2001).

Analysis of the bacterial community composition by FISH

The phylogenetic composition of the bacterioplancton was analysed in four replicates of fixed subsamples (5 mL each) by FISH, using Cy3-labelled commercially synthesized oligonucleotide probes (Biomers.net, Ulm, Germany).

The probes used were ARCH915 (Archaea domain), EUB338I-III (Bacteria domain), ALF1b (Alphaproteobacteria), BET42a (Betaproteobacteria), GAM42a (Gammaproteobacteria), HGC69A (Gram-positive with a high DNA G+C content), Pla46 (Planctomycetes), CF319a (Cytophaga–Flaviobacterium cluster phylum CFB), LGC354a (Firmicutes, Gram-positive bacteria with a low G+C content), EPS710 (Epsilonproteobacteria) and SRB385 (sulphate-reducing Deltaproteobacteria). Further details of these probes are available at http://www.microbial-ecology.net/probebase (Loy et al., 2003, Loy et al., 2007). Each groundwater subsample was filtered through a 0.2-μm polycarbonate membrane using a gentle vacuum (<0.2 bar), followed by 70%, 90% and 95% (v/v) ethanol series for 10 min each at room temperature, and then air dried. FISH of the harvested cells, counterstained with DAPI, was performed according to published protocols (Pernthaler et al., 2001; Barra Caracciolo et al., 2005c).

The averages of the number of cells binding each of the probes were calculated as a proportion of the total DAPI-positive cells from 10 to 20 randomly selected fields on each filter section (corresponding to 500–1000 stained cells). The slides were mounted with a drop of Vecta-Shield and the preparations were examined and counted on a Leica DM 4000B epifluorescence microscope at × 1000 magnification.

Isolation and characterization of bacterial strains from terbuthylazine-treated groundwater microcosms

In order to test for the presence of cultivable bacterial strains able to grow on the terbuthylazine herbicide as the sole carbon source, aliquots (10 μL) of TBA microcosm groundwater (collected at 40 and 60 days) were used as inoculum. The aliquots were plated on minimal medium MB (K2HPO4, 1.6 g L−1; KH2PO4, 0.4 g L−1; CaSO4·2H2O, 0.1 g L−1; MgSO4·7H2O, 1.0 g L−1; FeSO4·7H2O, 0.02 g L−1; (NH4)2SO4, 2 g L−1; agar, 15 g L−1), supplemented with 100 μg L−1 terbuthylazine as the carbon source and were incubated at two different temperatures (15 and 28 °C). Once some bacterial colonies appeared, they were subcultured on plates of the same media to obtain pure cultures. The isolates were subjected to morphological and Gram-staining characterization and subsequently to phylogenetic analysis by FISH.

Terbuthylazine degradation of bacterial isolates and FISH

Bacterial isolates were cultivated aerobically (in duplicate) at 28 °C in 100-mL flasks containing 30 mL of Luria–Bertani (LB) medium, supplemented with 10 mg L−1 of terbuthylazine and 0.03% casaminoacid. Chemical analysis of terbuthylazine was performed, as reported above, in order to assess the biodegradation capability of the two isolated strains.

Moreover, cells growing in the exponential phase were harvested by centrifugation and resuspended in phosphate-buffered saline. A volume of 30 μL of this cell suspension was adjusted to a concentration of 105–107 cells cm−2 and filtered onto 0.2-μm pore size polycarbonate filters (47 mm diameter, Isopore GTTP, Millipore, Germany). Samples were fixed using 70%, 90% and 95% (v/v) ethanol series for 10 min each at room temperature. Filters were stored at −20 °C until FISH was performed using the specific probes described above.

Species-level identification of the bacterial isolates was carried out by biochemical characterization (determined using the API 20NE and, when necessary, the API 50CH systems of BioMérieux SA), and by sequencing of the 16S rRNA genes.

DNA extraction and PCR amplification of the 16S rRNA gene for isolate identification

Identification of the environmental isolates was performed by comparative 16S rRNA gene sequence analysis. DNA from each isolate was extracted using the method described in Casas et al. (1995). The amplification of a 1500-bp fragment corresponding to the 16S rRNA gene was performed in a Mastercycler Personal (Eppendorf), using the universal primers and conditions described by Willems & Collins (1996). Amplicons were purified using the Qiaquick PCR Purification kit (Quiagen GmbH, Hilden, Germany), and both strands of the 16S rRNA gene were sequenced using the DyeDeoxy (dRhodamine) Terminator Cycle Sequencing kit in an automatic ABI Prism 373 DNA sequencer (Applied Biosystems) from SECUGEN facilities (Centro de Investigaciones Biologicas, CSIC, Spain). DNA sequences corresponding to the 16S rRNA gene were compared with those available in the GenBank/EMBL databases using the blast software (http://www.ncbi.nlm.nih.gov/BLAST).

Analysis of the genes encoding s-triazine-degrading enzymes in bacterial isolates

The pathway of terbuthylazine degradation in bacterial isolates was studied by detection of the atz genes, which encode the s-triazine-degrading enzymes, using two different methods: FISH and PCR.

AtzB-FISH analyses

Before filtering, isolated bacteria were incubated with lysozyme (12 mg mL−1) for 20 min at 37 °C to permeabilize the cells. Filters with cells fixed as described previously were analysed by FISH using the specific probe 5′FAM-AtzB1 (5′-GGAGAGCACCGATACTTTTCTT-3′), under the conditions described previously (Martin et al., 2008). Percentages of atzB-harbouring cells were calculated based on the total number of cells stained with DAPI.

PCR amplification of atz genes

PCR was performed using the primers designed (de Souza et al., 1998; Mulbry et al., 2002; Devers et al., 2004) to amplify the conserved DNA regions of the s-triazine catabolic genes atzA (atrazine chlorohydrolase), atzB (hydroxyatrazine ethylaminohydrolase), atzC (N-isopropylammelide isopropylamidohydrolase), atzD (cyanuric acid amidohydrolase) and trzN (triazine hydrolase). Reactions were conducted in a final volume of 100 μL, containing a DNA template (30–50 ng of bacterial DNA), 0.2 mM of dNTPs, 0.4 μM of both primers and 0.05 U μL−1 of Taq DNA polymerase (Biotools B & M laboratories SA). Following an initial denaturation step of 95 °C for 1 min, the amplifications were carried out in a Mastercycler gradient (Eppendorf) as follows: 40 cycles of 1 min at 94 °C, 1 min at 55 °C and 1 min at 72 °C, plus an additional 10-min cycle at 72 °C. PCR products were separated by electrophoresis on 1% agarose gels. Negative (no template DNA) and positive (50 ng of purified DNA from Pseudomonas sp. ADP) PCR controls were performed in the case of all atz genes. Amplification products were purified from the agarose gels or PCR reactions using the Geneclean Turbo Kit (MP Biomedicals LLC).

Statistical analysis

Results were expressed as mean±SE and were statistically analysed using Student's t-test (at P<0.05).


The main characteristics of the sampled aquifer and of the groundwater samples are reported in Table 1. The detection of s-triazines, together with the high nitrate concentration, demonstrates that this groundwater represents a contaminated ecosystem. Moreover, the bacterial abundance and the high percentage of cell viability, similar to or greater than that found in some surface soils studied (Martin et al., 2008; Grenni et al., 2009a), indicated the presence of a natural microbial community quite active in this aquifer.

View this table:

Main characteristics of the aquifer and of the groundwater samples

LithologyAlluvial sands, gravels
Geochemical faciesAlkaline bicarbonate
Temperature (°C)15
Eh (mV)210
Conducibility (μS cm−1)930
O2 (mg L−1)9.01
DOC (mg L−1)0.56
Terbuthylazine (μg L−1)0.149 ± 0.02
Desethyl-terbuthylazine (μg L−1)0.129 ± 0.01
Nitrate contamination (mg L−1)102 ± 7
Total cell abundance (no. of bacteria mL−1)2.0 E+04
Cell viability (%)72
  • DOC, dissolved organic carbon.

Terbuthylazine degradation in groundwater microcosms

Figure 1 shows the terbuthylazine concentration as a percentage of original concentration vs. time (days) under the two different experimental conditions (TBA and sterile). The disappearance time of 50% (DT50) values (expressed in days) calculated from the regression curve between the detected concentrations (Ct) and the sampling times (t) show that there was a halving of the initial concentration of terbuthylazine in 151±0.9 days (r=0.97) in the microbiologically active microcosms (TBA). Under sterile conditions, about 70% of the initial herbicide concentration remained at the end of the experiment and the theoretical DT50 value was 224±3 days (r=0.93). These results indicate that the autochthonous bacterial community played an active role in degradation of the herbicide. The decrease in herbicide in the sterile condition was presumably due to terbuthylazine dehalogenation to its hydroxylated form, which, as is well known, occurs through hydrolysis (Mandelbaum et al., 1993; Di Corcia et al., 1999). The metabolite desethyl-terbuthylazine was detected at increasing concentrations, starting from day 80 and reaching about 10 μg L−1 at day 175. This compound was never found in the sterile microcosms.


Decrease (%) of terbuthylazine concentrations in microbiologically active (TBA) and sterile (sterile) groundwater microcosms. The DT50 of the herbicide was calculated by the exponential equation obtained from the regression between concentrations (Ct) vs. time (days) in each experimental condition.

Oxygen and pH did not change significantly during the entire experimental period (t-tests not significant) under any of the conditions, with average values of 7.8±0.1 mg L−1 and 7.0±0.2, respectively. Consequently, we can exclude the effect of these parameters on the herbicide degradation pattern.

Live cell abundance

The live cell abundance in the treated (TBA) and the untreated (control) microcosms was similar, increasing to 105 live bacteria mL−1 until day 60. However, in the TBA microcosms, a significant increase was observed at both day 80 and day 124 (t-test, P<0.05), with values of 2.8–3 × 105 and 2.5–2.9 × 105 live bacteria mL−1, respectively. At the end of the experiment (175 days), the live cell abundance was again comparable between the two conditions, reaching a value of approximately 4.0 × 104 live bacteria mL−1.


The BCP, expressed as ng C mL−1 h−1, is reported in Fig. 2 (on the left y-axis). The BCP values were higher in the control than in the TBA microcosms at days 14 and 40. A reversal of this trend was observed at days 80 and 124, when BCP was higher in the TBA microcosm (t-test, P<0.05). Using the live cell abundance and BCP values, the bacterial growth rate (μ) and doubling time or turnover time [T2=(ln2)/μ] can be calculated using the exponential growth model, μ=ln((live cell abundance+BCP)/live cell abundance)/time (Koch, 1994; Amalfitano et al., 2008). The doubling time values obtained by this analysis (Fig. 2, on the right y-axis) show that the cell turnover time was higher in the control than in the TBA microcosms at both day 80 and day 124, indicating a faster growth rate in the presence of the herbicide.


BCP in columns and active cell-doubling time in dots [T2=(ln2)/growth rate; see text] at different incubation times. The error bars indicate SEs of four independent values.

Analysis of bacterial community composition by FISH

The use of 16S rRNA gene-targeted oligonucleotide probes made it possible to determine the structure of the autochthonous bacterial community at the phylogenetic level. At day 0, about 80% of the cells detected by DAPI belonged to the Bacteria domain (Fig. 3a), and only 1% belonged to the Archaea domain.


Bacterial community structure detected by FISH in natural groundwater. Probes used for Bacteria subgroups detection: ALF1b, Alphaproteobacteria; BET42a, Betaproteobacteria; GAM42a, Gammaproteobacteria; PLA46, Planctomycetes; HGC69A, Gram-positive with a high DNA G+C content; LGC354a Firmicutes, Gram-positive bacteria with a low G+C content; CF319a, Cytophaga–Flaviobacterium cluster phylum CFB; EPS710, Epsilonbacteria and SRB385, sulphate-reducing bacteria (a). Dynamics over time of the Betaproteobacteria subclass (% of the Bacteria domain) detected with the probe BET42a during microcosm incubation (b). The values are means of four analyses. The SE for each value is shown (±).

In all microcosms, the percentage of Archaea did not display a significant variation in time or between the different conditions, remaining at approximately 1% in all the samplings, with a transient increase to 3% at day 124 under both conditions.

In the TBA microcosms, the percent of cells positive to the general EUB probes diminished drastically from day 0 (about 80%) to days 80 (35%), 124 (20%), and 175 (35%). A decrease over time was also observed in the control microcosms, although the values were always higher than in the TBA samples and never <50% of DAPI-positive cells (data not shown). The lower percentages of Bacteria detected in the TBA samples indicated that most Bacteria were negatively affected by the herbicide, except the Betaproteobacteria subclass. In fact, the Betaproteobacteria were quite dominant, representing 23%, 95% and 45% of the Bacteria domain at days 80, 124 and 175, respectively (Fig. 3b).

Bacterial strain isolation from terbuthylazine-treated microcosms

We obtained four different bacterial isolates from the terbuthylazine-treated microcosms. All four isolates were Gram-negative and their phylogenetic characterization by FISH, using all the available probes (listed previously), showed that they were Bacteria belonging to the Betaproteobacteria subclass (98–100% of DAPI-stained cells). Further identification of these groundwater isolates was carried out by PCR amplification and sequencing of their 16S rRNA genes. Based on 16S rRNA gene alignment, strain 2-GA-2008 showed a close phylogenetic relationship with Janthinobacterium lividum, displaying 99.6% similarity in a 1380-bp overlap with the type strain of this bacterium (accession number Y08846). With the other species of the genus, Janthinobacterium agaricidamnosum (accession number Y08845), the percentage of 16S rRNA gene sequence similarity was 99.0%. The biochemical characteristics of the strain 2-GA-2008 determined using the API 20NE system were in accordance with those described for J. lividum DSM 1522 T (Lincoln et al., 1999). Therefore, the phylogenetic and biochemical results indicate that the isolate 2-GA-2008 belongs to this species. The other three environmental isolates, 4GA-2008, 6GA-2008 and 7GA-2008, exhibited the highest level of 16S rRNA gene sequence similarity (99.4%) to the strain Advenella incenata (accession number AM944734), and, using a polyphasic approach that included phenotypic, genetic and phylogenetic studies, were identified as A. incenata (Gibello et al., 2009).

Biodegradation of terbuthylazine by the environmental isolates

Both J. lividum and A. incenata were able to grow in liquid culture (MB) supplemented with terbuthylazine as the sole carbon source, displaying the capacity to degrade the herbicide with a DT50 of 121±10 days (r=0.94) and 88±6 days (r=0.98), respectively.

Using the AtzB1 probe, FISH analysis allowed us to detect the presence of atzB in the gene pool of both J. lividum and A. incenata. In both cases, the percentage of AtzB1-positive bacteria was >70%. The fact that hybridization signals were obtained in the pure cultures of these bacteria may suggest that they possess the genetic potential for s-triazine degradation.

Additional data on the presence of terbuthylazine-degrading genes in these bacteria were obtained by PCR analysis. Figure 4 shows the PCR fragments of the DNA amplicons generated with the primer sets used for the detection of each catabolic s-triazine gene (200 nt for atzA, 204 nt for atzB and 228 nt for atzC). The results show that A. incenata contained the atzA and atzB genes (Fig. 4a), while only the atzB and atzC genes were found in J. lividum (Fig. 4b). However, neither atzD nor trzN could be detected using this standard PCR method. Nucleotide sequences of the PCR fragments of the atz genes and from the low-molecular-weight DNA band from A. incenata atzB (Fig. 4a) were 100% identical to the homologous genes of the Pseudomonas sp. strain ADP.


Amplification products obtained in the PCR assays for screening the atz genes in Advenella incenata (a) and Janthinobacterium lividum (b). Lane 1, 100-bp DNA ladder (Biotools B & M laboratories SA); lane 2, negative control of atzA PCR; lane 3, positive control of atzA PCR (DNA from Pseudomonas sp. strain ADP); lane 4, atzA-PCR amplification of the DNA from a bacterial isolate; lane 5, negative control of atzB PCR; lane 6, positive control of atzB PCR (DNA from Pseudomonas sp. strain ADP); lane 7, atzB PCR amplification of the DNA from a bacterial isolate; lane 8, negative control of atzC PCR; lane 9, positive control of atzC PCR (DNA from Pseudomonas sp. strain ADP); lane 10, atzC-PCR amplification of the DNA from a bacterial isolate; lane 11, negative control of atzD PCR; lane12, positive control of atzD PCR (DNA from Pseudomonas sp. strain ADP); and lane 13, atzD-PCR amplification of the DNA from a bacterial isolate.


Terbuthylazine degradation has been reported to be quite variable (Barra Caracciolo et al., 2005a), and depends on bacterial activity and on abiotic factors (such as organic matter content, pH and, above all, temperature) that directly or indirectly influence the degradation rate. In previous studies on terbuthylazine degradation in soil at two different depths (Di Corcia et al., 1999; Barra Caracciolo et al., 2001), it was found that at 15 °C the DT50 was 180 days at the surface and 200 days in the subsoil. Our result of 151 days, found in TBA microcosms, agrees with the conclusions of these findings that terbuthylazine is highly persistent at relatively low temperatures. Temperature may be one environmental factor that affects the persistence of terbuthylazine in groundwater, where its half-life was found by (Navarro et al. 2004a, Navarro et al., 2004b) to range from 263 to 366 days. Groundwater is very different from surface ecosystems (absence of light, low carbon and oxygen availability and relatively low temperatures) for a community mainly composed of microorganisms adapted to these physical and chemical characteristics (de Lipthay et al., 2003). Thus, the prolonged persistence of terbuthylazine in groundwater may be an indication of lower microbial activity or the absence of herbicide-degrading organisms.

In this study, the autochthonous bacterial community found in the groundwater was able to degrade the herbicide terbuthylazine at the relatively high concentration used. This capability was presumably acquired through chronic exposure to contamination with the herbicide due to intensive agriculture and the high permeability of the aquifer studied (Daly et al., 2000). The microbial community present in the groundwater degraded the herbicide at a significant rate (Fig. 1). The fact that the bacterial community played a role in terbuthylazine degradation was confirmed by the changes in the bacterial community function (viability, carbon production) and structure (abundance and phylogenetic structure determined by FISH) observed in the TBA microcosms. The significant increase in live cell abundance observed in the TBA microcosms at days 80 and 124, in correspondence with the simultaneous increase in BCP and decrease of the corresponding cell-doubling time (Fig. 2), can be ascribable to the activity of specific degrading populations of the bacterial community. The lower values of T2 in the TBA compared with control microcosms, both at 80 and 124 days, showed that the bacterial populations favoured by the presence of the herbicide were more active. In fact, 124 days represented the closest sampling time to the half-life of the initial herbicide concentration.

The FISH phylogenetic results may further support the fact that the herbicide exerted selective pressure on the bacterial community, promoting the dominance of the Betaproteobacteria group. These data suggest that the Betaproteobacteria were better able to survive under this condition or that they could be involved in the degradation of the herbicide. This last hypothesis was confirmed by the isolation of two bacterial species (J. lividum and A. incenata), belonging to this bacterial group, that are able to degrade terbuthylazine (DT50 of 121±10 and 88±6 days). Both these isolates contain some atz-degradation genes. Detection of the atzB gene was first screened by FISH (Fig. 4c and d) and, in addition, s-triazine genes encoding the degrading enzymes were further analysed by PCR in both bacterial isolates. The results showed that A. incenata contained the atzA and atzB genes, while the atzB and atzC genes were found in J. lividum (Fig. 4). Curiously, although atzA was not detected in J. lividum, this bacterium was able to degrade terbuthylazine (DT50 of 121±10 days), presumably through the catabolic activity of the enzyme encoded by the atzB gene, which might catalyse both deamination and dechlorination reactions at different rates, as was suggested previously (Seffernick et al., 2007). None of these strains appeared to contain the atzD gene, suggesting that these bacteria were able to degrade terbuthylazine, transforming it to desethyl-terbuthylazine and cyanuric acid, but they were unable to complete the herbicide mineralization. Thus, in contrast to Pseudomonas ADP and other s-triazine degraders (Sadowsky et al., 1998; Wackett et al., 2002), the growth of these bacteria on terbuthylazine as the sole carbon source is only supported by the production of the putative alkylamidohydrolase products (alkylamines).

On the other hand, the dominant presence of Betaproteobacteria in the TBA microcosm at day 175 did not ensure that the residual terbuthylazine (about 40 μgL−1, corresponding to 30% of the initial concentration) was further removed. In fact, at the end of the experiment, a residual concentration of 30 μg L−1 terbuthylazine was still detected.

Bacterial degradation of s-triazines in groundwater has not been observed by other authors (Johnson et al., 2000; Pearson et al., 2006); however, we recently isolated a bacterial strain capable of using s-triazines as its sole carbon source in a pure liquid culture (Grenni et al., 2009b). To the best of our knowledge, the results of this study are the first to show that terbuthylazine (at 100 μg L−1) can be effectively biodegraded by the autochthonous bacterial community in groundwater microcosms. Interestingly, the variable distribution of the atz genes in these bacteria suggests the hypothesis that their terbuthylazine degradation pathways may result from different genetic elements in the microbial communities, rather than from the acquisition through horizontal transfer of a genetic ‘cassette’ encoding the entire set of atz genes. A similar situation has been described in the atrazine-degrading bacteria of genus Pseudaminobacter (Topp et al., 2000). However, the sequences of the atz genes were identical to those reported in every atrazine-degrading species examined to date (de Souza et al., 1998; Aislabie et al., 2005; Shapir et al., 2007), indicating that these genes are highly conserved and widely distributed between the different bacteria genera.

In conclusion, our study shows the potential for the use of a natural attenuation strategy in the treatment of aquifers polluted with the terbuthylazine herbicide. Moreover, the two bacteria isolated (A. incenata and J. lividum) could facilitate the implementation of effective bioremediation protocols in the case of the significant amounts that can be found in groundwater as a result of accidental herbicide spills.

However, at the end of the experiment, a residual concentration of 30 μg L−1 was still detected and whether the biodegradation may occur at even lower concentrations still has to be investigated. The complete removal of residual concentrations is in fact a crucial factor in the success of recovery strategies. Our results encourage further studies on the role of the bacterial populations in groundwater contaminated at lower concentrations.


Funding for this study was provided by the ‘Italian Environmental Ministry’ (Project: The Exposure Scenario and Risk Potential of Biocides) and by Comunidad de Madrid (Project: Evaluacion de Impacto Ambiental y Recuperacion del medio natural en emplazamientos contaminados, EIADES). The authors thank Francesca Falconi for her technical assistance, and the Umbria Regional Environmental Agency (dott. A. Martinelli) for help during sample collection and field measurements and for supplying s-triazine monitoring data.


  • Editor: Max Häggblom


View Abstract