Correspondence: Natsuko Hamamura, Center for Marine and Environmental Studies, Ehime University, Bunkyo-cho 3, Matsuyama, Ehime 790-8577 Japan. Tel./fax: +1 81 89 927 8551; e-mail: email@example.com
Soil bacterial population dynamics were examined to assess patterns in microbial response to contamination by different petroleum mixtures with variation in n-alkane profiles or toxic constituents such as pentachlorophenol (PCP). Three soil types from distinct areas of the United States (Montana, Oregon, and Arizona) were used in controlled perturbation experiments containing crude oil, kerosene, diesel, or diesel plus PCP spiked with 14C-hexadecane or 14C-tridecane. After a 50-day incubation, 30–70% of added 14C-alkanes were mineralized to 14CO2 in Montana and Oregon soils. In contrast, significantly lower mineralization was observed with diesel or kerosene (< 5%) compared to crude-oil treatment (~45%) in the Arizona soil. Different hydrocarbon mixtures selected both unique and common microbial populations across all three soils. Conversely, the contamination of different soils with the same mixture selected for distinct microbial populations. The most consistent genotype observed, a Rhodococcus-like population, was present in the Montana soil with all mixture types. The addition of PCP selected for PCP-tolerant alkane-degrading specialist populations. The results indicated that petroleum mixture type influenced hydrocarbon degradation rates and microbial population selection and that soil characteristics, especially organic content, could also be an important determinant of community responses to hydrocarbon perturbation.
Complex petroleum hydrocarbon mixtures including crude oil, diesel, and kerosene consist of varying concentrations of n- and branched alkanes, cycloalkanes, phenolics, aromatics, and polycyclic aromatic hydrocarbons (PAHs). Although these mixtures are comprised of similar constituents, the relative abundance of mixture components (e.g. aliphatics, aromatics) and toxic compounds (e.g. chlorophenols, heterocyclics) varies considerably, and these variations are potentially important in influencing the spatio-temporal dynamics of hydrocarbon-degrading microbial populations. For instance, the inhibition of specific populations may occur due to the toxicity of mixture components, such as pentachlorophenol (PCP), a common additive in petroleum products including creosote and wood preservatives (Lantz et al., 1997; Beaulieu et al., 2000). Successional changes in microbial populations may occur due to temporal compositional changes within mixtures resulting from degradation and/or transport phenomena (Colores et al., 2000; Röling et al., 2002). Temporal changes in complex mixtures occur due to abiotic processes, such as aging, volatilization and sequestration (Hatzinger & Alexander, 1995; Pignatello & Xing, 1996; Luthy et al., 1997), and microbial processes, such as preferential degradation of mixture components (Olson et al., 1999; Hamamura et al., 2006). Consequently, the types of complex mixtures and the original soil type are likely to both be important determinants of microbial population selection in contaminated environments.
Our overarching goal is to understand factors that influence microbial responses to environmental contamination. In a previous study, we examined the microbial response to crude-oil contamination in seven diverse soils from geographically distinct locations to assess the importance of soil type as a determinant for selecting specific hydrocarbon-degrading microbial populations (Hamamura et al., 2006). The results showed that soil type influenced the rate and pattern of hydrocarbon degradation as well as the specific microbial populations responding to crude-oil contamination. Specifically, Rhodococcus-like populations were involved in hydrocarbon degradation in the majority of soils, while other alkane-degrading populations unique to each soil were also identified. Different patterns of sequential n-alkane disappearance were observed in different soil types (Hamamura et al., 2006). Moreover, it was not clear whether the dominant mixture components were degraded by specialists whose populations may be selected by the different mixture types, or generalists with capabilities to degrade multiple mixture types. A better understanding of the effects of contaminant mixtures on microbial response will help provide a mechanistic basis for the interactions between specific mixture components and individual microbial populations.
In this study, we examined whether consistent patterns emerged due to mixture types within the same soil and/or across different soils. The effects of four different complex hydrocarbon mixtures on bacterial response were studied in three contrasting soil types. Crude oil, diesel, and kerosene were used to examine the effect of mixture composition on degradation rates of dominant alkane components. To evaluate the toxicity of specific mixture components (e.g. chlorophenolics) that may affect patterns in microbial population response, an additional treatment containing PCP plus diesel was also examined. Specific objectives were to assess the effects of different hydrocarbon mixtures on (1) microbial contaminant mineralization patterns; (2) microbial community response determined using 16S rRNA gene-targeted molecular analysis; and (3) selection of specific hydrocarbon-degrading microbial populations across three distinct soil types. These soils were selected based on our previous study (Hamamura et al., 2006) because they showed different crude-oil degradation patterns associated with the selection of different microbial populations and exhibited a range in pH (7.6, 5.4, and 8.8) and organic carbon content (3.9, 5.7, and 0.2%). Changes in bacterial populations following contamination with hydrocarbon mixtures were monitored using 16S rRNA gene-targeted denaturing gradient gel electrophoresis (DGGE). The toxic effect of PCP was examined in diesel treatments and on hydrocarbon-degrading bacterial isolates cultivated from these same samples and shown to be representative of indigenous populations. The combination of genotypic and phenotypic characterization of microbial populations selected using different contaminant mixtures revealed distinct effects of mixture type on microbial response across the different soils.
Materials and methods
Hydrocarbon mixture amendments were conducted using the surface horizon of three contrasting soils: Beaverton loam soil from Montana (MT), Jory clay soil from Oregon (OR), and Casa Grande sandy loam soil from Arizona (AZ). Soils were collected from native sites with no history of crude-oil contamination, as confirmed by gas chromatography–mass spectrometry (GC-MS) analysis, which showed no detectable hydrocarbons. Soils were passed through a 2-mm sieve and stored field moist at 4 °C. The pH, organic carbon content, and gravimetric water content at 33 kPa were determined to be 7.6, 3.9% and 29.9% for MT; 5.4, 5.7%, and 38.9% for OR; and 8.8, 0.2%, and 12.0% for AZ (Hamamura et al., 2006). Organic carbon and extractable NO3-N, K, and P contents for each soil were reported previously (Hamamura et al., 2006).
Hydrocarbon mixtures and amendment experiments
The crude oil used in this study consisted of n-alkanes with chain lengths of C9–C31: 80% of the total n-alkanes ranged from C12 to C24 (30%: C12–C15, 26%: C16–C19, and 26%: C20–C24), while the diesel and kerosene consisted mainly of n-alkanes with chain length of C10–C22 and C10–C16, respectively (Supporting Information, Fig. S1). Hydrocarbon mixture amendment assays were conducted at least in duplicate at 25 ± 2 °C in the dark without shaking, using 150-mL serum bottles containing 30 g (dry wt) soil as described previously (Hamamura et al., 2006). The reproducibility of the techniques was previously confirmed using triplicate bottles for same soil types (Hamamura et al., 2006) and was shown by small standard errors between duplicate bottles examined frequently during time course experiments in this study (Figs 1 and 2). Uncontaminated control soils were prepared and treated identically. Sterile soils were prepared by autoclaving three times for 1 h within 24-h intervals and used as abiotic controls. Crude oil (Conoco Corp., Billings, MT), diesel (ExxonMobil Corp.), and kerosene (Ace Hardware Corp., Oak Brook, IL) were mixed with 300 000 dpm [1-14C]hexadecane (> 98% purity, specific activity = 2.6 mCi mmol−1, Sigma Chemical, St. Louis, MO) or [1-14C]tridecane (> 98% purity, specific activity = 10.46 mCi mmol−1, Sigma Chemical), then added to each bottle to achieve a final concentration of 2% (w/w). The effect of PCP addition to diesel fuel was examined with MT and OR soils. The PCP concentrations commonly detected in wood-treating plants ranged from hundreds to thousands mg L−1 (EPA/540/F-95/506G, Miller et al., 2004), and a final concentration of 600 mg L−1 was used for this study. PCP was dissolved in diesel, which was then mixed with 300 000 dpm [1-14C]hexadecane or [14C-UL]PCP (> 98% purity, specific activity = 10.4 mCi mmol−1, Sigma Chemical), and added to each bottle to achieve a final diesel concentration of 2% (w/w). To eliminate the possibility of nutrient and oxygen limitation, we supplemented soils with a nutrient solution [final nutrient concentrations: (NH4)2SO4 (1.1 mM), NH4NO3 (5.9 mM), KH2PO4 (0.7 mM), K2HPO4 (0.7mM), KOH (0.28 mM), H2SO4 (0.28 mM), MgCl2 (1.6 mM), CaCl2 (3.2 mM), and FeCl2 (0.02 mM)] (Grosser et al., 2000) and periodically purged the headspace with air (see below). The total volume of sterile deionized water and nutrient solution added to the soil was calculated to achieve an equivalent matric potential of 33 kPa.
The depletion of individual hydrocarbon components and PCP was assessed by monitoring changes in chemical composition by capillary GC-MS as described previously (Hamamura et al., 2005). The efficiency of the solvent extraction procedure varied among soil types (MT: 78.4 ± 13.6, OR: 77.3 ± 9.4, and AZ: 99.8 ± 16.9%), likely due to the varying degrees of sorption to natural organic matter. Loss of total n-alkane fractions from control treatments with autoclaved soils was 15.0 ± 8.6% for crude oil and 29.8. ± 9.3% for diesel and diesel+PCP after a 50-day incubation, and 52.2 ± 1.2% for kerosene after a 35-day incubation, likely due to enhanced volatilization in air-purged reaction vessels. Thus, the % of n-alkane and PCP depletion in nonsterile treatments was normalized to the extractable amounts and then calculated relative to the amount of each n-alkane and PCP fraction recovered from the control treatments at corresponding time points. Although the GC-MS quantification of recovered hydrocarbons from soils indicates the depletion of hydrocarbon components, this method does not differentiate the biodegradation from the loss of hydrocarbons due to other biological effects (e.g. incorporation to biomass) and abiological (e.g. sorption) effects. To confirm the biodegradation of hydrocarbon components, the mineralization of [1-14C]hexadecane, [1-14C]tridecane, or [14C-UL]PCP was determined by measuring the evolution of 14CO2 as described previously (Hamamura et al., 2006). The amendments were purged twice a week with humidified CO2-free air to trap 14CO2, which also served to keep the system aerobic. After completion of the experiment, mass balance was determined based on the sum of evolved 14CO2 plus residual soil 14C measured using total combustion (Hamamura et al., 2005); average recoveries were 97.3 ± 4.6, 92.8 ± 7.9, and 80.6 ± 7.4% for [1-14C]hexadecane, [1-14C]tridecane, and [14C-UL]PCP amendments, respectively.
Bacterial populations in the hydrocarbon-amended microcosms were monitored by DGGE analysis of PCR-amplified 16S rRNA gene fragments from each treatment. DNA was extracted from 0.5 g (dry wt) subsamples after bead-beating, and 16S rRNA gene fragments were PCR-amplified using Bacteria-specific primer 1070F and the universal primer 1392R containing a GC-clamp; the PCR products were then separated using DGGE as described previously (Ferris et al., 1996). The reproducibility of the techniques was previously confirmed using DNA extracts prepared from triplicate bottles for same soil types (Hamamura et al., 2006). Dominant bands identified using DGGE were purified and their nucleotide sequences were obtained as described previously (Ferris et al., 1996). Sequences were aligned using sequencher 4.1 (Gene Codes Corporation, Ann Arbor, MI) and compared to the GenBank database using blast (Altschul et al., 1990). The nucleotide sequences reported in this paper have been deposited in the GenBank database under accession numbers DQ983819–DQ983830.
Growth of alkane-degrading isolates in the presence of PCP
Alkane-degrading isolates obtained from crude oil-amended soils (Hamamura et al., 2006) were grown in liquid minimal (Xm) medium (Hamamura et al., 1999) containing 1% (v/v) n-hexadecane as a sole carbon source by inoculation from a single colony. A 50-μL inoculum of hexadecane-grown cells (early stationary phase culture) was transferred in triplicate to 60-mL vials containing 10-mL Xm medium with 1% (v/v) n-hexadecane (0.3 mM final concentration) with 6, 60, or 600 mg L−1 (final concentration) of PCP dissolved in acetone (100 μL). Control cells were grown under identical conditions with 1% n-hexadecane and 100 μL acetone. The cultures were incubated at 25 ± 2 °C with constant shaking (150 r.p.m.) in the dark, and optical density at 600 nm was determined periodically from aseptically collected subsamples.
Effects of mixture types on alkane biodegradation
The hydrocarbon mineralization capabilities of indigenous microbial populations in three soils were assessed by measuring 14CO2 production from hydrocarbon mixtures amended with 14C-hexadecane, an important component in all three mixtures (Fig. 1). Although the rates and extents of recovered 14CO2 varied among the mixture types, approximately 30–70% of added 14C-hexadecane was mineralized after a 50-day incubation in both MT and OR soils. In contrast, significantly lower 14C-hexadecane mineralization was observed with two of the three mixture types in the AZ soil. For example, the amount of 14CO2 produced from 14C-hexadecane in the presence of diesel or kerosene was < 5% after 50-day (Fig. 1) and increased to only 10.4 and 7.2%, respectively, after 75 days (data not shown). The rate of hexadecane mineralization was similar in MT soil amended with crude oil or diesel (0.13 and 0.11% 14CO2 day−1 g−1 soil, respectively), but was considerably lower in the presence of kerosene (0.03% 14CO2 day−1 g−1 soil), even though the specific activity of 14C-hexadecane was higher in kerosene due to the lower abundance of hexadecane (Fig. S1). The extent of hexadecane mineralization (i.e. ~35%) was also lower in the presence of kerosene. Similarly, rates of hexadecane mineralization were lower in OR soils in the presence of diesel or kerosene (0.067% 14CO2 day−1 g−1 soil in crude oil vs. 0.04–0.042% 14CO2 day−1 g−1 soil for diesel and kerosene). Moreover, longer lag phases were observed in the presence of diesel or kerosene (17–19 days) compared to the crude-oil amendment (7 days). To examine whether the observed differences in mineralization rate were due to the lower abundance of hexadecane in kerosene, the mineralization of 14C-tridecane, a major component in kerosene, was also examined. Mineralization patterns of 14C-tridecane were similar (MT and AZ soils) or slower (OR soil) than 14C-hexadecane (Fig. S2). Consequently, kerosene has a definite effect on microbial degradation of both hexadecane (a minor component) and tridecane (a major component).
Concomitant monitoring of n-alkanes by GC-MS confirmed the depletion of other hydrocarbons present in each mixture type (Fig. S3). For example, > 90% of all n-alkane components were depleted within 50 days in the MT soil independent of mixture type. However, the disappearance of n-alkanes in other soil types (OR and AZ) was considerably slower in the presence of kerosene. Specifically, 90, 85, and 56% of the total n-alkanes were depleted within 50 days in crude oil-, diesel-, and kerosene-amended OR soils, respectively. The slower disappearance of n-alkane was even more pronounced in AZ soils where only 25 and 18% of the total n-alkanes were depleted in diesel and kerosene amendments, respectively. The amount of n-alkane depletion was also lower in crude-oil treatments (80% loss) relative to other soils. Moreover, the AZ soil exhibited preferential depletion of C10–C12n-alkanes.
The effect of PCP on diesel biodegradation
The effect of PCP on hydrocarbon degradation and microbial response was examined in MT and OR soils where substantial diesel degradation was observed (Fig. 1). Different chemical and microbial responses to PCP addition were noted in these two soil types (Fig. 2). A substantial decrease in 14C-hexadecane mineralization was observed in the MT soil when PCP was added with diesel compared to treatments without PCP (~50% lower mineralization within 50 days). In contrast, no significant change in 14C-hexadecane mineralization was observed when PCP was added to diesel-amended OR soil. Concomitant monitoring of n-alkanes by GC-MS revealed similar trends to those observed using 14C-hexadecane where the addition of PCP resulted in less depletion of total n-alkanes in MT soil (over 50 days), but no difference in n-alkane depletion in OR soil (Figs S3b and S4).
The mineralization of PCP was also determined in MT and OR soils by monitoring the evolution of 14CO2 from separate treatments containing UL-14C PCP. The amount of PCP mineralized after 50 days was only 3.2 and 5.0% in MT and OR soils, respectively (Fig. 2), but continued to increase to 4.8 and 8.4% after 145 days (data not shown). Conversely, GC-MS measurements showed 42 and 82% PCP depletion in MT and OR soils, respectively (Fig. S4). The efficiency of the solvent extraction procedure did not vary substantially for n-alkanes vs. PCP (67 and 56% recovery in MT and 48 and 57% in OR soil, respectively). A much smaller fraction of 14C-PCP recovered as 14CO2 compared to the actual disappearance of parent PCP suggests that only partial mineralization of PCP occurred, even after other n-alkanes in diesel were depleted.
Molecular analysis of bacterial community response
To assess mechanisms responsible for changes in hydrocarbon degradation as a function of different hydrocarbon mixture type, the bacterial populations responding to different soil amendments were examined directly using PCR-amplified 16S rRNA gene sequences separated via DGGE (Fig. 3). Differences in microbial community response were noted in the same soil type treated with different mixture types as well as among soils treated with the same mixture type. Soils amended with complex hydrocarbon mixtures (i.e. crude oil, diesel, kerosene, and diesel+PCP) showed the emergence of prominent DGGE bands (Fig. 3a–c), compared to an uncontaminated control (MT soil) evaluated during the same incubation period (Fig. 3d). Thus, the detection of prominent 16S rRNA gene sequences in treated soils was clearly due to hydrocarbon perturbation. It is unclear, however, whether any decreases in intensity of bands corresponding to native populations were due to toxicity of added contaminants or a reduced ability to see these bands due to the dominance of bands whose intensities rose with contaminant addition. The distribution of genotypes across soils and mixture types is summarized in Table 1.
Table 1. Sequence analysis of 16S rRNA gene fragments detected in soils contaminated with hydrocarbon mixtures
Treatment of MT soil with different mixture types resulted in the emergence of common as well as distinct DGGE bands (Fig. 3a and Table 1). Band N1 was consistently detected in all four mixture-type amendments, and its sequence was 100% identical to the 16S rRNA gene of a known alkane-degrading actinobacterium, Rhodococcus erythropolis NRRL B-16531 (van Beilen et al., 2002). Other prominent DGGE bands such as N3 (Burkholderia-like) were observed in all mixture types except crude oil, while band N4 (Pseudomonas-like) was observed in diesel and diesel+PCP amendments. The band N3 sequence is identical to the 16S rRNA gene sequence of alkane-degrading Burkholderia isolates previously cultured from MT soil treated with crude oil (Table 1), consistent with the role of this population in hydrocarbon degradation in MT soil. In the diesel+PCP amendment, an additional Pseudomonas-like population (band N5) appeared on day 19 together with bands N1 and N4, followed by the emergence of band N3 at day 40. The population corresponding to band N4 (also Pseudomonas-like) became less prominent after day 40, while bands N1 and N5 persisted throughout the 72-day incubation. Alkane-degrading Pseudomonas-like isolates cultivated from crude oil-amended MT soil have 16S rRNA gene sequences identical to the band N5 sequence (Table 1). Amendment of MT soil with shorter chain-length alkane mixtures (i.e. diesel and kerosene) resulted in a shift to more gram-negative populations compared to the emergence of gram-positive populations in treatments with crude oil.
No consistent population types were observed across the four mixture types in the OR soil amendments (Fig. 3b and Table 1). Treatments with diesel showed banding patterns similar to those observed in crude-oil amendments, but an additional Burkholderia-like population (band R5) emerged on day 27. This Burkholderia-like population (Table 1), which is distinct from other Burkholderia-like organisms represented by bands R6 or N3 (MT soil), was also detected in kerosene-amended soil treatments. This observation suggests that this population may preferentially utilize shorter- to mid-chain-length n-alkanes (i.e. C10–C22). Two additional populations represented by bands R4 (Sphingomonas-like) and R6 (Burkholderia-like) were detected in kerosene and diesel+PCP amendments, respectively. Alkane-degrading isolates with 16S rRNA gene sequences identical to band R2, R5, and R6 sequences were also cultivated from crude oil-amended OR soil, consistent with their possible role in hydrocarbon degradation (Table 1).
Although hydrocarbon degradation rates were significantly lower in diesel- and kerosene- vs. crude oil-amended AZ soils, the emergence of distinct DGGE bands and successional patterns were observed during the incubation period (Fig. 3c). The alkane-degrading Burkholderia-like population (band N3) observed in MT soil was also detected in kerosene- and diesel-amended AZ soils. The DNA sequences of many DGGE bands detected in diesel- and kerosene-amended AZ soils were closely related to those of known hydrocarbon-degrading genera, such as Rhodococcus, Sphingomonas, and Ralstonia (Table 1). Alkane-degrading Nocardioides isolates (identical 16S rRNA gene sequence to band A5) obtained from kerosene-amended AZ soil were also previously cultivated from crude oil-amended AZ soil (Table 1). Consequently, although hexadecane mineralization to CO2 was not significant in kerosene- and diesel-amended AZ soils, results from molecular analysis indicate that hydrocarbon-degrading populations were enriched in these AZ-treated soils.
Pentachlorophenol effects on the growth of hydrocarbon-degrading isolates detected in soil microcosms
Isolates corresponding to several of the prominent populations identified using DGGE (i.e. bands N1, N3, and R5) were previously cultivated from crude oil-amended MT, OR, and AZ soils using hexadecane as a sole carbon source (Table 1). Those isolates include Rhodococcus sp. strain N1 and Burkholderia sp. strain N3 from MT soil, and Burkholderia sp. strain R5 from OR soil. In contrast, Pseudomonas sp. strain N5 (MT soil) and Burkholderia sp. strain R6 (OR soil) isolates, which have identical 16S rRNA gene sequences as the sequences of bands N5 and R6, respectively, were detected only in diesel+PCP amendments (Table 1 and Fig. 3). This provided an opportunity to test the hypothesis that PCP played an important role in selecting for specific alkane-degrading populations detected in the presence of PCP. None of these isolates mineralized [UL-14C]PCP to detectable levels of 14CO2 (data not shown); thus, the growth of each isolate was examined in the presence of 6, 60, and 600 mg PCP L−1 with hexadecane as a carbon source (Fig. 4). Among the MT soil isolates, Pseudomonas sp. strain N5 showed higher tolerance to PCP compared to Burkholderia sp. strain N3 and Rhodococcus sp. strain N1, which showed no growth at 6 mg PCP L−1. The two OR soil isolates (Burkholderia sp. strain R6 and strain R5) exhibited different tolerance to PCP (60 mg L−1), and this observation provides a basis for understanding effects of PCP addition on microbial selection in the presence of diesel fuel. Hydrocarbon-degrading populations associated with the most intense DGGE bands after PCP addition (strains N5 and R6) were more tolerant of elevated PCP compared to populations present in other mixture types. None of the tested isolates were able to grow in liquid media containing 600 mg PCP L−1, which was the final concentration used in diesel+PCP soil amendments. Because PCP can be sorbed to soil organic matter and become less bioavailable (Lee et al., 1990; Andreux et al., 1993), the concentration of bioavailable PCP in soil amendments was likely to be lower than 600 mg L−1.
We previously showed that soil type is an important determinant of specific microorganisms responding to crude-oil perturbation (Hamamura et al., 2006). The current study evaluated other common petroleum mixtures and a toxic additive (diesel, kerosene, and PCP) using three distinct soil types under conditions where other parameters (e.g. water potential, nutrients, temperature, aeration) were held constant. Different complex hydrocarbon mixtures selected both unique and common microbial populations in all three soil types. Conversely, treatment of different soils with the same mixture type selected for distinct microbial populations except in one case, where a Burkholderia-like N3 population was observed in kerosene- and diesel-amended MT and AZ soils. The most consistent genotype observed across different mixture types was the R. erythropolis-like N1 population, important in the MT soil after amendment with all mixture types. This observation supports the hypothesis that R. erythropolis-like organisms function as generalists capable of degrading a broad range of different chain-length n-alkanes. These findings are consistent with a previous report that R. erythropolis NRRL B-16531 is capable of degrading C6 to C36n-alkanes (van Beilen et al., 2002). In addition, this genotype was the most frequent population observed across seven different soil types contaminated with crude oil (Hamamura et al., 2006, 2008). Collectively, these results confirm the versatility of Rhodococcus spp. (Larkin et al., 2005; McLeod et al., 2006) and their importance in hydrocarbon bioremediation in diverse soil environments.
The microbial community profiles observed across different mixture types also suggest the importance of specialists associated with preferential utilization of specific n-alkane fractions (Table 1). For instance, Burkholderia-like populations represented by DGGE bands N3 and R5 were detected in MT and OR soils, respectively, in both diesel and kerosene treatments. These organisms may preferentially utilize shorter- to mid-chain-length alkanes (C10–C22) in diesel and kerosene. Conversely, a Rhodococcus-like population only detected in crude oil-amended MT soil (band N2) appeared to be associated with the disappearance of longer n-alkanes (i.e. >C22). Changes in microbial community composition observed during disappearance of hydrocarbons in the MT soil were also associated with the degradation of specific n-alkane fractions. For example, the Pseudomonas-like N4 population was important during 27–49 days in diesel-only treatments as well as at day 19 in diesel+PCP treatments (Fig. 3a) and corresponded with the preferential depletion of C11–C17n-alkanes (Figs S3b and S4).
Although some genotypes were detected consistently across different mixture types within the same soil, phenotypic variation within the same genotype may contribute to differences in observed hydrocarbon degradation rates. For instance, the R. erythropolis-like N1 population detected in MT soil across all mixture types may be comprised of very closely related populations that differ with respect to the range of n-alkane chain lengths they are able to degrade. For example, various R. erythropolis strains with identical 16S rRNA gene sequences have been shown to contain three to five different alkane hydroxylase genes (alkB) that may be involved in the oxidation of different n-alkanes (van Beilen et al., 2002). Consequently, the differences in degradation rates observed in the current study may reflect functional variations within R. erythropolis strains. It should be noted that alkane-degrading strains with identical 16S rRNA gene sequences could exhibit distinct physiological properties, and those strains were distinguishable by the distinct alkB genes they possess (Hamamura et al., 2008). Consequently, further characterization of alkB gene diversity may resolve the phenotypic variations (i.e. degradation rate) among the in situ Rhodococcus population.
Interactions of various components in hydrocarbon mixtures are expected to influence solubility and mobility and potentially lead to detrimental or beneficial effects on biodegradation. Auffret et al. (2009) examined the degradation of petroleum mixtures and additives by two Rhodococcus strains and emphasized that specific compounds present in various mixture types may have inhibitory effects on the biodegradation of other compounds. This could be the case with Burkholderia-like N3 population, which was detected in kerosene- and diesel-degrading MT amendments as well as AZ amendments with substantially lower kerosene and diesel degradation activities. Many of the 16S rRNA gene sequences identified in kerosene- or diesel-amended AZ soils were closely related to those of known hydrocarbon-degrading bacteria, including crude-oil degrader Nocardioides oleivorans (Schippers et al., 2005) and PAH degraders, Ralstonia sp. (Bodour et al., 2003) and Sphingomonas sp. (Eriksson et al., 2003). The diesel and kerosene used in this study contained ~5% PAHs, and these populations might be preferentially utilizing PAHs rather than n-alkanes in AZ soil amendments.
Additionally, soil types that exhibit distinctly different physiochemical properties are also expected to influence the sorption, solubility, and mobility of mixture components differently, thus possibly affecting the biodegradation capacity of the same population in different soils. For instance, the organic carbon content is an important factor affecting the sorption of nonpolar organic chemicals (Karickhoff, 1981, 1984; Doucette, 2003). The organic carbon content of the AZ soil (0.2%) is 20–30 times lower than those of MT and OR soils (3.9 and 5.7%, respectively); thus, hydrocarbon components are expected to be less prone to partition in the solid phase in the AZ soil. Indeed, the efficiency of hydrocarbon recovery from AZ soil was significantly higher (99.8 ± 16.9%, P <0.05 Student's t-test) than that of MT or OR soils (78.4 ± 13.6 and 77.3 ± 9.4%, respectively). Consequently, it is possible that the increase in bioavailable hydrocarbons, especially short-chain n-alkanes with higher aqueous solubility, caused greater toxic effects to microbiota in AZ soil and resulted in impaired biodegradation activities with diesel and kerosene amendments (Fig. 1). The organic matter content has also shown to be fundamental to the development of diversity in soil (Øvreås & Torsvik, 1998; Tiedje et al., 2000; Girvan et al., 2005). Comparison of two soils with differing organic matter showed that higher organic carbon content was associated with greater bacterial community diversity and that the more diverse soil exhibited greater resistance and resilience to perturbation stresses (Girvan et al., 2005). Our results support the hypothesis that soils with higher organic carbon contents (MT and OR) were more resistant to perturbation by various hydrocarbon mixtures.
The addition of PCP was shown to select for specific PCP-tolerant alkane-degrading populations in both MT and OR soils. Although no substantial PCP mineralization was detected in this study, PCP mineralization capability has been previously reported among phylogenetically diverse group of organisms including both gram-negative (e.g. Pseudomonas, Ralstonia, Burkholderia, Sphingomonas) and gram-positive (e.g. Mycobacterium, Nocardioides, Rhodococcus) bacteria (Häggblom & Valo, 1995; McAllister et al., 1996; Männistö et al., 1999), including the genotypes detected in PCP+diesel amendments in this study. In our study, 42 and 82% of PCP depletion was detected by GC-MS analysis in MT and OR soils after 50 days, respectively. Because only a small fraction of [14C-UL]PCP was mineralized to 14CO2, it is possible that PCP was dehalogenated, but only partially cleaved by dioxygenase and mineralized. Further work is necessary to identify metabolites formed during the partial degradation of PCP in soil amendments.
The alkane-degrading populations associated with the most intense DGGE bands after PCP treatment and only found in the presence of PCP (N5 and R6) showed higher tolerance to elevated concentrations of PCP compared to other isolates detected in these soils under other amendment conditions (Fig. 4). Rhodococcus sp. strain N1 was unable to grow in the presence of 6 mg L−1 PCP, although the corresponding DGGE band was persistently detected during the incubation period in diesel+PCP-amended MT soil (Fig 3a). Again, it should be noted that because we examined only one representative isolate corresponding to each DGGE band, any physiological variation (e.g. PCP tolerance) within a population with identical 16S rRNA gene sequences would have been missed. Further characterization using high-resolution analysis of alkB genes may enable the differentiation of distinct alkane-degrading populations responding to PCP among the closely related strains with different PCP tolerance.
It is clear from our study that petroleum mixture types can influence the kinetics of hydrocarbon degradation as well as the specific microbial populations responding to the perturbation. The microbial populations present in treatments containing a toxic mixture component (i.e. PCP) suggested that specialists with physiological adaptations (as indicated by the selection of PCP-tolerant alkane-degrading populations in diesel+PCP treatments) may provide increased resilience in microbial community response. Previous studies have shown that microbial diversity and community structures are influenced by soil characteristics rather than geographical distances (Johnson et al., 2003; Fierer & Jackson, 2006; Dequiedt et al., 2009). Our results demonstrated that soil type is also an important determinant of microbial responses to introduced petroleum mixtures.
We thank Dr. John Baham, Oregon State University, and Dr. Jeff Silvertooth, University of Arizona, for collecting soil samples. We also appreciate technical assistance from Sara Olson, Mary Bateson, Katherine Schultz, and Rich Macur. We are indebted to the Conoco refinery (Billings, MT) for providing us with crude oil. This work was supported by the USEPA (Project No. 829357-01-0) and the Montana Agricultural Experiment Station (Projects 911398 and 911352).