Environ. Sci. Technol. 2006, 40, 3293-3298
Biodegradation, Bioaccessibility, and Genotoxicity of Diffuse Polycyclic Aromatic Hydrocarbon (PAH) Pollution at a Motorway Site ,†
ANDERS R. JOHNSEN,* JULIA R. DE LIPTHAY,‡ FREDRIK REICHENBERG,§ SØREN J. SØRENSEN,‡ OLE ANDERSEN,| PETER CHRISTENSEN,| MONA-LISE BINDERUP,⊥ AND CARSTEN S. JACOBSEN† Department of Geochemistry, Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark, Department of Microbiology, University of Copenhagen, Sølvgade 83H, DK-1307 Copenhagen K, Denmark, Department of Environmental Chemistry and Microbiology, National Environmental Research Institute, Frederiksborgvej 399, PO Box 358, DK-4000 Roskilde, Denmark, Department of Life Sciences and Chemistry, Roskilde University, Postboks 260, DK-4000 Roskilde, Denmark, and Institute of Food Safety and Nutrition, Danish Veterinary and Food Administration, Mørkhøj Bygade 19, DK-2860 Søborg, Denmark
Diffuse pollution of surface soil with polycyclic aromatic hydrocarbons (PAHs) is problematic in terms of the large areas and volumes of polluted soil. The levels and effects of diffuse PAH pollution at a motorway site were investigated. Surface soil was sampled with increasing distance from the asphalt pavement and tested for total amounts of PAHs, amounts of bioaccessible PAHs, total bacterial populations, PAH degrader populations, the potential for mineralization of 14C-PAHs, and mutagenicity. Elevated PAH concentrations were found in the samples taken 1-8 m from the pavement. Soil sampled at greater distances (12-24 m) contained only background levels of PAHs. The total bacterial populations (CFU and numbers of 16S rDNA genes) were similar for all soil samples, whereas the microbial degrader populations (culturable PAH degraders and numbers of PAH dioxygenase genes) were most abundant in the most polluted samples close to the pavement. Hydroxypropyl-β-cyclodextrin extraction of soil PAHs, as a direct estimate of the bioaccessibility, indicated that only 1-5% of the PAHs were accessible to soil bacteria. This low bioaccessibility is suggested to be due to sorption to traffic soot particles. The increased PAH level close to the pavement was reflected in slightly increased mutagenic activity (1 m, 0.32 ( 0.08 revertants g-1 soil; background/ 24 m: 0.08 ( 0.04), determined by the Salmonella/ microsome assay of total extractable PAHs activated by liver enzymes. The potential for lighter molecular weight PAH * Corresponding author e-mail:
[email protected]; phone: +45 3814 2328. † Geological Survey of Denmark and Greenland. ‡ University of Copenhagen. § National Environmental Research Institute. | Roskilde University. ⊥ Danish Veterinary and Food Administration. 10.1021/es060008u CCC: $33.50 Published on Web 04/13/2006
2006 American Chemical Society
degradation in combination with low bioaccessibility of heavier PAHs is proposed to lead to a likely increase in concentration of heavier PAHs over time. These residues are, however, likely to be of low biological significance.
Introduction Large areas of Denmark (1) and other densely inhabited countries are diffusely polluted with polycyclic aromatic hydrocarbons (PAHs) inadvertently released by the incomplete combustion of organic matter and fuels. These diffusely contaminated areas are characterized by low pollutant concentrations that cannot be tracked to a single point source. Studies of microbial degradation of soil PAHs and genotoxic effects have focused mainly on creosote and gasworks soils of limited area, but with high contaminant concentrations. However, the diffusely polluted areas may pose a bigger problem in terms of square kilometers and tons of polluted soil. The biological effects of the low contaminant levels found in these areas have not previously been addressed in detail. In the present study, we investigated the levels and effects of diffuse PAH pollution of surface soil at a 30-year-old, highly trafficked motorway site. By using an interdiciplinary approach, we related the PAH levels, the potentials for microbial degradation of PAHs, and the mutagenic activity at increasing distances from the asphalt pavement. PAH levels were determined both by solvent extraction of total PAHs and by hydroxypropyl-β-cyclodextrin (HPCD) extraction of bioaccessible PAHs (2-5). The microbial potentials for degradation of PAHs were estimated by quantification of microbial PAHdioxygenase genes, by most probable number (MPN) counts of phenanthrene degraders, and by determination of potentials for mineralization of 14C-labeled model PAHs. These data were compared to in vitro mutagenicity data on solventextracted PAHs.
Materials and Methods Study Site. A motorway north of Copenhagen (Motorway 14, km 16.8 at Vejenbrød) was the chosen study site. This locality has been highly trafficked during the last 30 years, with an annual average daily traffic (in 2004) of 35 800 vehicles (Danish Road Directorate, person communication). The emission of benzo[a]pyrene was 5.2 g km-1 y-1 and the total PAH emission (naphthalene, fluoranthene, benzo[a]pyrene, benzo[b+j+k]fluoranthene, benzo[g,h,i]perylene, and indeno[1,2,3-cd]pyrene) was 94 g km-1 y-1 (Danish Road Directorate, unpublished data). Soil was sampled in three transects, separated by 3m, at the eastern side of the motorway, as the prevailing wind direction in Denmark is from the west. Soil was sampled from the upper 10 cm at 1, 3, 5, 8, 12, 16 and 24 m from the pavement. The 1-meter samples were coarse sand, and the rest were clay soil. Soil samples were homogenized by sieving through a 4-mm sieve followed by mixing. Soil PAH Analyses. The soil samples were analyzed for the following PAHs: naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benzo[c]phenanthrene, benzo[a]anthracene, triphenylene+chrysene, naphthacene, benzo[b+j+k]fluoranthene, benzo[a]pyrene, perylene, indeno[1,2,3-cd]pyrene, dibenzo[a,h]anthracene, benzo[g,h,i]perylene, dibenzo[a,l]pyrene, dibenzo[a,i]pyrene, and dibenzo[a,h]pyrene as described previously (6), except that the samples were not dried under vacuum prior to extraction. VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3293
Number of Colony-Forming Units (CFU). Triplicate estimates of heterotrophic CFU from each soil sample were counted on R2A agar as previously described (6). Chitinase Activity. The chitinase activity as an indicator of fungal biomass, was determined by the turnover of 4-methylumbelliferyl (MUF)-labeled N-acetyl-β-D-glucosaminide to fluorescent 4-methylumbelliferone (4-MU) (7). The substrate concentration was 10 µM and the assay time was 2 h. From each soil sample, the chitinase activity (nM h-1 g-1 soil dry weight) was measured for 10 replicate 100mg subsamples, 5 additional replicates served as blind controls, and 5 served as quenching controls. Phenanthrene Degraders. The sizes of the culturable PAH degrader populations were estimated as most probable numbers (MPN) of phenanthrene degraders according to a previously published microplate method (8) based on respiratory reduction of the tetrazolium compound WST-1 in growth-positive wells. The incubation time was increased from three to four weeks to increase the difference between positive and negative wells. Soil from parallel samples was mixed. Quantitative Real-Time PCR of PAH Dioxygenase (nah, phnA, pdo1) and 16S rDNA Genes. Whole-community DNA was extracted from three replicate 0.5-g samples of each soil by application of a bead-beating procedure using the FastDNA Spin Kit for Soil (BIO 101, Vista, CA) as described by de Lipthay et al. (9). Conventional PCR-detection of PAH dioxygenases was performed using primers targeting nahlike genes (10), phnA genes (11), and pdo1 genes (6, 12). Detection of 16S rDNA genes was performed using the 27f (13) and 518r (14) primers. Quantitative real-time PCR (qPCR) was performed with the four primer sets using a BioRad iCycler (Bio-Rad, Hercules, CA). All q-PCR reactions were performed in a total volume of 25 µL using the QIAGEN QuantiTect SYBR Green PCR Kit (QIAGEN, Valencia, CA). The applied primer concentration was 0.5 µM for the detection of NAH and 16S rDNA genes, and 1.0 µM for phnA and pdo1 genes. The conditions for the q-PCRs were as described for the conventional PCRs (6, 9) except that they were started with a 15-min activation step (95 °C), and that a total of 50 PCR cycles were performed. The q-PCR reactions included 2 µL of template DNA, i.e., soil DNA extract or standard DNA. The number of bacterial cells containing PAH degradation or 16S genes in the soil samples was determined from standard curves of genomic DNA derived from Pseudomonas putida OUS82 (NAH) (10), Burkholderia sp. RP007 (phnA) (15), Mycobacterium sp. 6PY1 (pdo1) (12), or Escherichia coli MC1061 (16S rDNA) (16). All standard curves were made from DNA extracts corresponding to approximately 101 to 108 cells per PCR, assuming that one bacterial cell contains 5 × 10-15 g of DNA (17). DNase treated (30 min at 37 °C with DNA-free (Ambion)) soil DNA extract (2 µL) was further added to q-PCRs with standard DNA to better mimic the PCR conditions of the soil DNA samples. Following all q-PCRs, melting curve analysis as well as conventional agarose gel electrophoresis was done to confirm the validity of the PCR products. PAH Degradation Potential. The potentials for mineralization of 9-[14C]-phenanthrene (Sigma-Aldrich, >98% purity), sidering-U-[14C]-anthracene (Amersham, 95% purity), 3-[14C]-fluoranthene (Sigma-Aldrich, >95% purity) and 4,5,9,10-[14C]-pyrene (Amersham, >95% purity) of the soil samples were tested as described previously (6). PAHBioaccessiblity.Hydroxypropyl-β-cyclodextrin(HPCD, >97% purity, Wacker-Chemie, Burghausen, Germany) extraction of the bioaccessible PAH-fractions was carried out according to Cuypers et al. (2). Extractions were done in 50-mL glass serum bottles with Teflon-lined butyl rubber lids. To each bottle was added 2.5 g soil (dry weight) and 50 3294
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 10, 2006
mL of HPCD in aqueous solution (140 g L-1) containing sodium azide (7.7 mM) as a biocidal agent. The bottles were placed horizontally on an orbital shaker (Unimax 2010 Heidolph, Schwabach Germany, 150 rev min-1) at 20 °C. After 2, 6, 12, 24, 48, 70, and 155 h, the bottles were centrifuged at 4000g for 15 min, the supernatants were decanted and fresh HPCD solution was added. The supernatants were stored at 5 °C in amber glass vials with Teflon lids and analyzed for PAHs within two weeks. Quantification of PAHs in the HPCD extracts was carried out using HPLC-fluorescence detection (Agilent 1100 system with G1321A FLD (Ex. 260 nm; Em. 350, 420, 440 and 500 nm) fitted with a Trident Guard in-line filter (Restek, Bellefonte, PA)). Separation column: Varian CP-Ecospher 4 PAH operated at 0.5 mL min-1 (28 °C, 100 µL injection); gradient t ) 0 min, 80% (w/w) methanol, t ) 5-30 min, linear gradient 80-100% (w/w) methanol, t ) 30-45 min 100% (w/w) methanol. Phenanthrene and pyrene could not be quantified due to analytical interference. The level of quantification was defined as 10 times the standard deviation of the signal noise level. A simple two-parameter exponential model (eq 1) was fitted to the time series (0-155 h) to estimate the maximum PAH release (bioaccessible concentration Cb). Ct is the cumulative concentration of released PAH at time t.
Ct ) Cb(1 - e-kt)
(1)
Mutagenicity. For mutagenicity testing, 30 g of each soil sample was extracted as for total soil PAH analysis. The extracts were evaporated to nearly dryness under a gentle stream of nitrogen, and transferred to dimethyl sulfoxide (DMSO) in a concentration corresponding to15 g of soil per mL. The extracts were tested in the Salmonella/microsome plate assay as described by Maron and Ames (18). Five concentrations (from 300 to 18.8 mg per plate) were tested in triplicate in the Salmonella test strain TA98 both with and without Aroclor 1254 induced rat liver S9-mix from Wistar male rats (18). For experiments with metabolic activation, 3 mg of S9 protein per mL of S9 mix was used. Negative controls (DMSO only) and positive controls were included in each experiment. The positive controls were 2-nitrofluorene without S9 mix and benzo[a]pyrene with S9 mix. After 48 h incubation, the colonies were counted using an automatic colony counter from Perspective Instruments. The mutagenic potency was calculated as the linear slope of the concentration-response curve.
Results and Discussion Total PAHs. PAHs in diffusely polluted soil originate from multiple sources. Along roads, the PAHs are emitted primarily from the combustion of fuel, but also by particles released from car brakes and the wear of tires (19), and possibly also from the wear of the pavement. The PAH content dropped rapidly with increasing distance from the motorway (Figure 1). The PAH concentrations of the 1-, 3-, 5- and 8-meter samples were significantly higher than those in the 24-meter samples (one-sided t-test, p < 0.05), however the PAH concentrations more than 3 m from the pavement (0.25-0.43 mg kg-1) were comparable to levels previously found in pristine Danish soils (20). The apparently high PAH content of one 12-meter sample (3.2 mg kg-1) was probably an artifact caused by a small asphalt lump or a tire particle in the subsample used for PAH analysis, as the sample did not show other characteristics of high PAH content (increased mutagenicity or high PAH degradation potential). For details and standard deviations, see the Supporting Information. A likely explanation for the observed decrease in PAH concentration with increasing distance is that PAHs may have
FIGURE 1. Average concentrations of PAHs in soil from three independent transects. been deposited at the roadside by splash from the road in periods with rain, whereas in dry periods the PAHs were probably transported over longer distances by the wind. The PAH profiles were for all samples dominated by the heavy molecular weight (HMW) PAHs with four or more fused benzene rings, constituting from 86 to 91% of the total PAHs. The content of the 6-ring PAHs, dibenzo[a,l]pyrene, dibenzo[a,i]pyrene, and dibenzo[a,h]pyrene was for all samples below the limit of detection. Microbial Populations. The total bacterial populations, determined by counts of colony forming units (CFU) and real-time PCR quantification of 16S rDNA genes, were similar for all the soil samples (Table 1, P > 0.05 when compared to 24-meter samples by two-sided t-tests). CFU were on average 180 times smaller than the 16S rDNA estimate, confirming the general notion that less than one percent of the total bacterial population is culturable on rich media (21). Neither CFU nor 16S rDNA genes were correlated with the PAH concentration of the soil samples (log(CFU) vs log(total PAHs), R2 ) 0.173; log(16S rDNA) vs log(total PAHs), R2 ) 0.012, see Supporting Information). The bacterial PAH degraders were, in contrast to the total bacterial populations, most abundant close to the motorway (Table 1). MPNs of culturable phenanthrene degraders were high within the first 3 meters, and then declined with increasing distance. Proper statistical evaluation was not possible as no particular density can be assigned to samples with all dilutions negative (12- and 24-meter). pdo1 genes were detected in all samples, but were most abundant in the 1- and 3-meter samples (Table 1, P < 0.05 when compared to 24-meter samples by one-sided t-tests). However, the density of pdo1 was only slightly correlated to the total PAH concentration (log(pdo1) vs log(total PAHs), R2 ) 0.375, see Supporting Information). nah genes were detected only in the 3-meter samples (data not shown); however, the densities could not be reliably quantified by real-time PCR. phnA genes were not detected in any of the soil samples. The dominance of pdo1genes, originally described for isolates belonging to Mycobacterium (12), suggests that mycobacteria and related genera were the most important PAH degraders in the motorway soil. PAH Mineralization Potentials. The PAH-degradation potentials were quantified by spiking the soil samples with 14C-labeled phenanthrene, anthracene, fluoranthene, or pyrene. PAH mineralization potentials were measured as the time needed for mineralization of 10% of the added label (t0.1, (20)). To estimate the t0.1 values for fluoranthene, it was
necessary to extrapolate the 8-, 12-, and 24-meter samples by linear regression. The four PAHs were mineralized by all soil samples, albeit at different rates (Table 2 and Supporting Information). The most easily degraded PAH was phenanthrene, and the least degraded was fluoranthene. Despite the differences in t0.1, the four PAHs showed similar patterns of highest PAH mineralization close to the pavement and decreasing t0.1 with increasing distance. High PAH-degradation potentials in the most polluted 1- and 3-meter samples suggest that accumulation of PAHs from the traffic was caused by low bioaccessibility of the traffic PAHs rather than limitations within the microbial populations. The t0.1 of the soil samples did not correlate with the size of the fungal populations measured as chitinase activity (phenanthrene, R2 ) 0.020; anthracene, R2 ) 0.033; pyrene, R2 ) 0.029, fluoranthene, R2 ) 0.032, see Supporting Information), indicating that fungal enzymes were less important in the PAH mineralization. Bioaccessibility. One way to quantify the bioavailability of a compound in environmental samples is to determine the quantity of the compound that can be mobilized and become available to biouptake or biodegradation (22). This bioaccessible quantity may be determined by depletive sampling. For PAHs, the bioaccessibility to microbes may be estimated by hydroxypropyl-β-cyclodextrin (HPCD) extraction of the desorbable PAHs (2-5). HPCD is a soluble, ringshaped hepta-saccharide with a hydrophobic center that complexes with PAH molecules. The presence of high concentrations of dissolved HPCD in the diffusive boundary layers around soil particles will rapidly and effectively increase the transport efficiency across this surface-to-bulk boundary layer by increasing the solubility of the PAH (23). The release of PAH to the aqueous bulk phase through the diffusive boundary layers, which often act as bottlenecks for diffusive mass transfer of PAHs in soil (24), is therefore significantly increased. The principle is to add HPCD in surplus so that the release of PAH from soil particles is the limiting factor. The first proposed method for HPCD assessment of bioaccessibility of soil PAHs was based on a single extraction with 6 or 20 h contact between soil and HPCD solution (4). However, this study was carried out with soil that was artificially polluted with “fresh” PAHs, a scenario very different from the aged PAHs found in our motorway soil. We therefore used a modification of the method in which the soil samples were repeatedly extracted with HPCD solution (2). The efficiency of this method was previously demonstrated for two naturally aged, PAH-contaminated sediments (2). In those sediments, bioaccessible PAHs were rapidly released within the first 20-50 h followed by slow release approaching constant residual PAH concentrations. In our study, the release of bioaccessible PAHs was slower (Figure 2), however 90% or more of the PAH predicted to be bioaccessible by eq 1 was released during the 155 h extraction. With increasing distance from the pavement, the bioaccessible PAH concentrations rapidly fell below the limits of quantification, hence, bioaccessibility could only be reliably estimated for the 1-meter samples (Table 3 and Supporting Information). As little as 1-5% of the PAHs were estimated to be accessible to the soil microbes. A likely explanation is that much PAH may be associated with black carbon particles (soot) in the motorway soil. These particles are released to the environment in combustion processes, especially in diesel engines, and they act as “super-sorbents” for PAHs (25, 26). A recent desorption study of PAHs associated with traffic soot estimated that only 0.5-10% of the PAH, depending on molecular mass, would desorb into aqueous media during 120 days (27). Estimated time scales for the release of the remaining heavy molecular weight (HMW) PAHs amounted to several millennia (27). Therefore, it is not surprising that HMW PAHs have accumulated in the motorway soil, even VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3295
TABLE 1. Microbial Populations Determined as Numbers of Heterotrophic Colony Forming Units (CFU) on R2A Plates, Cell Numbers Based on 16S rDNA, Most Probable Numbers (MPN) of Culturable Phenanthrene Degraders, Numbers of PAH Degraders Based on the PAH-dioxygenase Gene pdo1, and Fungal Populations Estimated by Chitinase Activitya distance from pavement (m)
CFU (cells g-1 × 108)
16S rDNA (cells g-1 × 108)
phenanthrene degraders (cells g-1)
pdo1 dioxygenase (cells g-1 × 102)
Chitinase activity (nmol g-1h-1)
1 3 5 8 12 16 24
0.8 ( 0.6 0.6 ( 0.3 1.1 ( 0.7 1.5 ( 0.6 1.0 ( 0.5 0.8 ( 0.1 1.2 ( 0.6
140 ( 20 180 ( 70 140 ( 60 320 ( 38 90 ( 70 150 ( 20 230 ( 90
5100 [2700-11000] 8900 [4100-18000] 1300 [610-2700] 260 [78-860]
