1 Introduction

High molecular weight polycyclic aromatic hydrocarbons (HMW-PAHs), comprised of at least three fused aromatic rings, are of special concern due to their toxicity, mutagenicity, carcinogenicity, and association with inflammatory vascular responses (Kim et al. 2005). Pyrene, arising from incomplete combustion of hydrocarbons, is a widespread HMW-PAH in the soil environment. Microbial degradation is the main process to remove PAHs from soil environment. Thus, a comprehensive understanding of the microbial degradation of pyrene will facilitate the development of new strategies to improve the remediation of PAH-polluted soils.

Many bacteria using pyrene as the sole carbon source have been isolated from various environments, most of them belong to genera Mycobacterium (Heitkamp et al. 1988) and Rhodococcus (Walter et al. 1991), which are Gram-positive. Recently, several Gram-negative strains have also been identified as pyrene degrader, such as Sphingomonas sp. (Zhong et al. 2011) and Pseudomonas aeruginosa RS1 (Ghosh et al. 2014). However, knowledge of pyrene-degrading bacterial populations is still limited and is mainly based on the studies of cultured bacteria. Therefore, the characterization of indigenous microbial populations would be necessary for a better understanding of in situ biodegradation processes of pyrene in complex environments, such as soil.

The initial step of PAH biodegradation commonly occurs via the incorporation of molecular oxygen into the aromatic nucleus by a multi-component terminal aromatic ring hydroxylating dioxygenase (RHD) enzyme system (Kauppi et al. 1998). RHD is composed of a large α subunit (PAH-RHDα) and small β subunits (Mason and Cammack 1992). Functional markers encoding PAH-RHDα can be used to specifically target functional microbial populations involved in PAH biodegradation (Dionisi et al. 2004). For example, PAH-RHDα from Gram-positive bacteria (PAH-RHDα GP) were shown to be involved and specialized in the degradation of HMW-PAHs (Baldwin et al. 2003; Zhou et al. 2006; Cebron et al. 2009). Two biochemical pathways of pyrene degradation have been established in pyrene-degrading mycobacterial strains (Kim et al. 2007). nidA, which was highly conserved in Mycobacterium (Heitkamp et al. 1988), encodes the α-subunits of RHD that adds hydroxyl groups at C-4 and C-5 positions of pyrene in the degradation pathway (Kim et al. 2012). More recently, an alternative pyrene RHDα gene, nidA3, has been identified in Mycobacterium vanbaalenii PYR-1 (Kim et al. 2006), which was shown to be involved in the initial dihydroxylation step in an alternate detoxification pathway (Kim et al. 2007). Although the biochemical pathways of pyrene biodegradation have been identified in cultivated species, little is known about the in situ biodegradation pathways of pyrene in the soil environment. Moreover, how the abundance and diversity of potential pyrene degraders was influenced by soil environmental factors are not yet fully explored.

Therefore, the aim of the current study was to investigate the functional microbial populations and process involved in pyrene biodegradation in complex soil environment, including lower PAH-containing agricultural soil (LS) and higher PAH-containing industrial soil (HS). Total bacterial 16S rDNA gene, PAH-RHDα GP gene, and Mycobacteria 16S rDNA gene copies were regularly measured by real-time PCR to identify the potential pyrene degraders in soils. The abundance of functional genes, nidA and nidA3, was also monitored to characterize the likely mechanisms of pyrene biodegradation pathways in two soils.

2 Materials and methods

2.1 Soil property and preparing

Surface soils (0–20 cm) were collected from the suburb of Beijing, China. One was from an agricultural field with less PAH contamination (40° 17.54 N; 116° 06.81 E) (LS). This field is perennially planted with vegetables and corn. When we sampled soil, corn was grown at the field. The other soil was from an industrial green land. This land was long-term contaminated with PAHs, which is near a steel-manufacturing factory (39° 54.57 N; 116° 08.54 E) (HS). When we sampled soil, the land was covered with grass. Both soils are traditional cinnamon soil, which are classified as alfsol according to the US Department of Agriculture soil classification system. After being transported to laboratory, soil samples were air-dried, sieved (2 mm), and stored in the dark at 4 °C. The soil properties were determined following standard procedures and are shown in Table 1. The LS is silty clay loam, and the HS is clay loam. The pH and content of silt, clay, and organic matter of two soils were comparable, while the content of sand of agricultural soil was lower than that of industrial soil. Significantly, the concentration of total PAHs of industrial soil (HS) was 6.3-fold higher than that of agricultural soil (LS), and the pyrene concentration was almost 10-fold higher than that of agricultural soil.

Table 1 Properties of soils from lower PAH-containing soil (LS) and higher PAH-containingsoil (HS)
Full size table

2.2 Pyrene incubation experiment

In order to re-activate the soil microorganisms, dry soil samples (800 g) in duplicates were placed in separate glass beakers, moistened with sterile distilled water to the level at 50 % of water holding capacity, and incubated at room temperature (25 °C) for 2 weeks. The soils were then spiked with 1000 mg l−1 solution of pyrene in acetone to a final concentration at 60 mg kg−1, following the methods recommended by Brinch et al. (2002) with some adaptations. Briefly, a subsample of 10 % soil was spiked with the pyrene solution and mixed throughout. After evaporation of the acetone (overnight at 25 °C), the spiked subsample (10 %) was mixed with the rest (90 %) of the soil and shaken thoroughly. The moisture content of the samples was adjusted again to 50 % of water holding capacity. For unspiked control, soils were amended with the same amount of acetone as the pyrene-spiked soils. For the abiotic controls, the soils were autoclaved before spiking with pyrene solution (Park and Crowley 2006).

After the treated soils were homogenized, 10-g aliquots were placed into serum bottle and incubated in the dark at 25 °C. During incubation, triplicate samples were collected on days 0, 3, 7, 14, and 35, frozen immediately, and stored at −80 °C until analyses.

2.3 Pyrene quantification

Pyrene was extracted from 2 g soils which were dried by lyophilization and measured by accelerated solvent extractor (Dionex ASE-300). Analysis of pyrene followed the procedure described earlier (Zhang et al. 2011). Briefly, 100 mg kg−1 of fluoranthene-d10 and phenantherene-d10 were used as substitute and internal standards, respectively. Concentrated (1 ml) soil extracts were analyzed using an Agilent gas chromatograph (7890A) equipped with a mass-selective detector (model 5897): Helium was used as a carrier gas (constant velocity of 30 cm s−1) at 250 °C. The GC oven conditions are as follows: 50 °C for 1 min, a 10 °C min−1 ramp to 250 °C, and then a ramp of 5 °C min−1 to final temperature of 280 °C (hold for 10 min). Mass spectrometer detection was used with the selected ion monitoring system. Each treatment was carried out in triplicate. Analytical procedural blanks and fluorene-d10 (surrogate for pyrene) were used for quality control. Pyrene concentrations of analytical procedural blanks were very low (<0.01 mg kg−1), which were about 4 orders of magnitude lower than that of the soil samples. The average recovery efficiency of fluorene-d10 (surrogate for pyrene) was 89.1 ± 7.2 %. The final sample concentrations were blank corrected, but not surrogate recovery corrected. In addition, the recovery of the pyrene from the freshly spiked soils (on day 0) was 104.6 ± 4.4 %, indicating that the spiking procedures were validated.

2.4 Primer development

To design nidA3 targeted primer, the DNA sequences of nidA3 gene from several PAH-degrading Mycobacterium strains—PYR-1 (AF249301), 6PYR (AJ494745), JLS (AY330098), KMS (AY330100), and MCS (AY33010)—were obtained from NCBI GenBank (http://www.ncbi.nlm.nih.gov/) and aligned by MEGA5 software (Tamura et al. 2011). Conserved regions of nidA3 were selected from the aligned data, and three sets of real-time PCR primers targeting different regions of the nidA3 gene were designed by Primer Premier 5.0 software (Premier Biosoft International, USA). Among these, one primer set was chosen for real-time PCR analysis. To ensure specificity of the designed primer, ten cloned products from one of our DNA samples were sequenced, all showing more than 98 % amino acid similarity with PAH-degrading Mycobacterium sp. NidA3 (data not shown).

2.5 DNA extraction and real-time PCR analysis

DNA was extracted from 0.5 g of soil (dry weight) with a FastDNA Spin kit for soil (Bio101, USA) according to the manufacturer’s instructions. The concentration and quality of the extracted DNA were determined by spectrophotometer analysis (NanoDrop ND-1000, NanoDrop Technologies, Willmington, DE) and agarose gel electrophoresis.

Total bacterial (16S rDNA), Gram-positive PAH-RHDα gene (PAH-RHDα GP), Gram-negative PAH-RHDα gene (PAH-RHDα GN), Mycobacteria 16S rDNA, and nidA and nidA3 gene copies were evaluated by a Taqman or SYBR Green based real-time PCR quantification using an iCycler iQ5 themocycler (Bio-Rad, USA), and their primes and amplification protocols were summarized in Table 2. Tenfold serial dilutions of a known copy number of the plasmid DNA were generated to produce the standard curve. Each sample DNA was diluted for minimizing threshold cycle suppression by inhibitors for that sample type (typically 1:10). The real-time PCR efficiencies (87.9∼112.5 %) were examined to test for inhibition. R 2 values were more than 0.99 for all calibration curves. All real-time PCR analyses were performed using iCycler software.

Table 2 Quantitative real-time PCR primers in this study
Full size table

2.6 PAH-RHDα GP gene clone library construction and sequencing

Two clone libraries were constructed from LS and HS at day 35. The PCR amplification using the primers GP-F/GP-R is described in Table 2. PCR products were purified and ligated into the pMD19-T easy vector (Promega, USA) according to the manufacturer’s instructions. Plasmids were transformed into Escherichia coli JM109 cells (Takara), and more than 50 clones were randomly selected and sequenced for each clone library. All the sequences and their relatives obtained from the NCBI were aligned by using the MEGA5 software. The sequences with more than 97 % identity were grouped into one operational taxonomic unit (OTU) using the Mothur program (version 1.31.2, USA) by the furthest neighbor approach. Phylogenetic tree was created using the MEGA5 software by neighbor joining (NJ) method. The relative confidence of the tree topologies was evaluated by performing 1000 bootstrap replicates.

3 Results

3.1 Pyrene dissipation

The kinetics of pyrene dissipation in the two soils was monitored during incubation on days 0, 3, 7, 14, and 35. In the unautoclaved soils, pyrene dissipation exhibited similar trends over time in both soils, with an initial lag phase followed by a rapid disappearance (Fig. 1). From day 7, the dissipation of pyrene was noticeable, while no significant difference was observed between two soils. At day 35, more than 80 % of the spiked pyrene was dissipated in both soils. The pyrene dissipation rate was 1.30 μg day−1 g−1 dry soil in LS; however, the pyrene dissipation rate was significantly higher in HS, about 1.60 μg day−1 g−1 dry soil. In contrast, for the sterilized soils, there was little pyrene disappearance observed during the whole incubation. Moreover, a slight increase of pyrene levels was observed at the initial incubation phase (at day 3 and day 7), although the increase was not significant (ANOVA, P > 0.05). This result suggests that abiotic dissipation of pyrene is not a main route for pyrene removal in soils.

Fig. 1
figure 1

Pyrene dissipation in the soils during incubation with pyrene. Soils were spiked with 60 mg kg−1 pyrene; LS less PAH-contaminated soil, HS higher PAH-containing soil; data are means ± SE; n = 3

Full size image

3.2 Quantification of the potential pyrene degraders and functional genes

Total bacterial 16S rDNA gene copies were determined by real-time PCR. For two soils, the amount of bacterial 16S rDNA was relatively constant over time, either spiked with or without pyrene, with the gene copies ranging from 8.9 × 109 to 1.9 × 1010 copies g−1 dry weight soil (Fig. 2).

Fig. 2
figure 2

Quantification of bacterial 16S rDNA, PAH-RHDα GP, Mycobacterium 16S rDNA, nidA, and nidA3 gene copies during incubation; + pyrene, spiked with 60 mg kg−1 pyrene; − pyrene, unspiked with pyrene; data are means ± SE; n = 3

Full size image

Across the entire incubation times, the quantities of PAH-RHDα GP gene, Mycobacteria 16S rDNA gene, nidA (from Mycobacteria), and nidA3 (from Mycobacteria) increased significantly in both soils spiked with pyrene. For both soils, PAH-RHDα GP gene copies significantly (P < 0.01) increased with incubation time (Fig. 2). The copies of Mycobacteria 16S rDNA, measured by an assay specific to fast-growing Mycobacteria, increased by 5 and 20 times in LS and HS, respectively, after 35 days of pyrene incubation (Fig. 2). Copy numbers of genes, nidA and nidA3, encoding alpha subunits of two different ring hydroxylating dioxygenases, showed approximately 2 orders of magnitude higher at the end of experiment (35 days) in both soils compared with those at the beginning (Fig. 2). In contrast to pyrene-spiked soils, none of these detected genes showed significant changes in both soils without pyrene (Fig. 2). These results suggested that pyrene incubation increased the abundance of Mycobacteria and the expression of nidA and nidA3.

Between two pyrene-spiked soils sampled at day 35, there was a significant difference in abundances of PAH-RHDα GP gene, Mycobacteria 16S rDNA, nidA, and nidA3: all exhibited significantly higher copies at HS than at LS (Fig. 2). In addition, at the initial time of pyrene incubation (day 0), there was no significant difference in quantities of Mycobacteria 16S rDNA or PAH-RHDα GP genes between two soils, while HS had significantly higher nidA and nidA3 copy numbers than LS (Fig. 2). These observations were consistent with the fact that soil HS had experienced the historical contamination and typically exhibited higher concentration of PAHs than LS (Table 1).

3.3 Correlations among catabolic genotypes and pyrene degradation

To determine the relationship between catabolic gene copies which may be related to the degradation potential of the microbial communities, a correlation matrix of Spearman rank correlation coefficients was developed (Table 3). As shown in Table 3, the extent of pyrene degradation was positively correlated to all the target gene copies except bacterial 16S rDNA. In addition, Mycobacteria 16S rDNA, nidA, and nidA3 gene copies had significant positive correlations with each other. This is expected as Mycobacterium spp. are the primary carriers of nidA or nidA3 genotypes in the contaminated environment (DeBruyn et al. 2012).

Table 3 Spearman rank correlation coefficient (r s) matrix (n = 10)
Full size table

3.4 Diversity of PAH-RHDα GP genes

To investigate the diversity of PAH-RHDα GP genes, two clone libraries were constructed from pyrene-spiked LS and HS at day 35, respectively. A total of 10 OTUs were identified in the clone library of the LS (46 sequences analyzed), and 4 OTUs were identified in the clone library of the HS (43 sequences analyzed). The richness indices (Table 4), including Chao 1 estimator, Shannon index, and reciprocal of Simpson index, revealed that a higher diversity of PAH-RHDα GP gene was found in the pyrene-spiked LS than in the HS, and this was consistent with the result of the OTU richness curves (Fig. 3), which showed representative sampling of each clone library.

Table 4 Diversity and richness indices of PAH-RHDα GP gene sequences from two pyrene spiked soils
Full size table
Fig. 3
figure 3

OTU richness curves of PAH-RHDα GP gene libraries. The curves were obtained by rarefaction calculations of OTU abundance data and they represent libraries from two pyrene-spiked soils. Curves were calculated for 97 % sequence identity

Full size image

The phylogenetic analysis indicated that all of the obtained sequences of PAH-RHDα GP gene from LS and HS could be grouped into three distinct gene clusters (Fig. 4). Cluster I was the largest group and included a total of 80 PAH-RHDα GP gene sequences. Among cluster I, five and four sequences from soil LS and HS, respectively, were closely related to nidA3 of Mycobacterium. However, most sequence from cluster I could not be closely clustered with any functional gene characterized in cultivated microorganisms. They were affiliated with uncultured PAH-RHDα genes recovered from HMW-PAH-contaminated environment and were related to nidA3 from Mycobacterium with approximately 90 % similarity. Cluster II included three sequences from LS and five sequences from HS. These sequences were closely related to an uncultured Actinobacteria PAH-RHDα gene and could be clustered with pdoA2 derived from two cultivated Mycobacterium species. Cluster III contained only 1 sequence from LS, which was not closely related (with less than 87 % similarity) to any reported sequences available in public database. This suggested the existence of novel PAH-RHDα GP gene or microorganism in LS.

Fig. 4
figure 4

Neighbor-joining tree based on the multiple alignments of PAH-RHDα GP gene sequences. Bootstrap values represent 1000 replicates and only values above 50 % are shown. The scale bar represents the expected number of changes per nucleotide position (5 % nucleotide substitution). Numbers in parentheses (in bold) after the clone names correspond to the analyzed clone numbers

Full size image

4 Discussion

4.1 Dissipation of pyrene

Results from this study showed that the majority of added pyrene was dissipated in two soils after 35 days of incubation (Fig. 1). Compared with previous findings (Chen and Ding 2012; Peng et al. 2010; Xu et al. 2006), pyrene dissipation in both soils was relatively more rapid in this study. It was likely due to the rapid enrichment of activated indigenous degraders in response to the presence of pyrene, which would result in higher pyrene degradation (>96 %) in 1-month incubation (Hamdi et al. 2007). However, the rapid pyrene dissipation was not observed in the sterilized soils; this result suggests that pyrene dissipation in soil environment mainly relies on microbial degradation biotic process, rather than abiotic process. It was concerned that, under high content of organic matter and clay condition, abiotic PAH loss might also contribute to pyrene dissipation (e.g., sequestrated by soil matrix). However, in our study, no significant (ANOVA, p > 0.05) pyrene disappearance was observed in both sterilized soils over the incubation time (Fig. 1). We assumed that it was possible due to the time of aging being relatively short. This observation was consistent with the previously published data obtained from other PAH aging studies using sterilized soils (Luo et al. 2012; Macleod and Semple 2000; Cheung and Kinkle 2001; Mueller 2006).

4.2 Enrichment of functional microbes during pyrene degradation

To unravel the response of microbial populations in soils to pyrene pollution, real-time PCR based assay for potential functional microbes and genes was conducted under different soil environments. Specially, we quantified the bacterial 16S rDNA, PAH-RHDα GP, Mycobacteria 16S rDNA, nidA, and nidA3 gene abundances in two pyrene-spiked soils. The amount of bacterial 16S rDNA genes, a proxy for bacterial biomass (Suzuki et al. 2000), was relatively stable over time in both soils regardless of the addition of pyrene or not (Fig. 2). This result indicates that there was no significant change in the overall population density of the total bacteria in soils exposed to pyrene. This observation is similar to previous studies showing that 16S rDNA gene copy numbers were not affected by the addition of PAHs including pyrene, phenanthrene, and naphthalene (Park and Crowley 2006; Peng et al. 2010; Niepceron et al. 2014). However, the relatively constant density for the bacterial biomass did not rule out variations in the species composition of the bacterial community.

In order to further study the functional microbial population that might be involved in pyrene biodegradation in two soils, we monitored the bacterial PAH degradation community using primers targeting PAH-RHDα GP genes from a wide range of bacteria taxa (Gram-positive) (Fig. 2). The increasing abundance of PAH-RHDα GP gene over the incubation times with pyrene indicated that the Gram-positive bacteria harboring PAH-RHDα genes might be an important group of pyrene degraders in the test soils. Previous studies have implied that Gram-positive PAH degrader was the predominant PAH degrading microbial group in an industrial soil, where high concentrations of pyrene were detected (Meng and Zhu 2011). Gram-positive degraders might experience advantages in pyrene biodegradation by increasing pyrene bioavailability through forming biofilm directly on hydrophobic pollutants (Johnsen and Karlson 2004).

Sequencing of PAH-RHDα GP clone library was done to examine the phylogenetic distribution of the potential pyrene degraders in two tested soils (Fig. 4). Several sequences detected in LS and HS were closely related to nidA3 of Mycobacterium with more than 98 % nucleotide sequence identity, indicating the involvement of fast-growing Mycobacterium in pyrene biodegradation. However, the majority of retrieved sequences (40 of 46 clones in LS, 39 of 43 clones in HS) were closely related to PAH-RHDα genes derived from uncultured bacteria (Fig. 4). These genes might be a potential pyrene dioxygenase, as they showed more than 90 % nucleotide acid similarity with either nidA3 or pdoA2 from Mycobacterium (Fig. 4), which were considered as the key catabolic genes to degrade pyrene (Krivobok et al. 2003; Kim et al. 2006). These results suggested that Mycobacterium as well as those uncultured bacteria might play an important role in pyrene degradation. The identity of these uncultured bacteria should be further investigated by cultivation-independent approaches such as stable-isotope probing (SIP) (Peng et al. 2013).

To investigate the response of Mycobacteria, detected in PAH-RHDα GP clone library, to the addition of pyrene, Mycobacteria 16S rDNA gene was quantified in two pyrene-spiked soils during incubation times (Fig. 2). Results showed that the fast-growing Mycobacteria was presented and enriched during the pyrene biodegradation in both soils. This result indicates that Mycobacteria were induced to grow and were likely playing an important role in pyrene degradation. Previous studies also showed that fast-growing Mycobacteria populations were an important group of HMW-PAH (e.g., pyrene) degrader in PAH-contaminated soils (DeBruyn et al. 2007; Debruyn et al. 2009). Coincident with the detection of fast-growing Mycobacteria, the copy numbers of pyrene dioxygenase genes including Mycobacteria nidA and nidA3 increased significantly during incubation time (Fig. 2), which further suggested that fast-growing Mycobacteria were enriched in the test soils and had function genotypes for pyrene conversion. However, we could not exclude the possibility of other nidA or nidA3 carriers, such as Terrabacter spp., Phycicoccus spp., and Leifsonia spp. (DeBruyn et al. 2012), which might also be involved in pyrene biodegradation.

4.3 Enrichment of functional gene during pyrene degradation

The positive correlation between pyrene catabolic gene copies (nidA and nidA3) and pyrene degradation extent indicated that two pyrene catabolic pathways, targeted by nidA (degradation pathway) and nidA3 (detoxification pathway), respectively, might both potentially contribute to the pyrene biodegradation in the tested soils (Table 3). Although the in situ metabolic activity of nidA and nidA3 genes (detected as mRNA transcripts) remains to be established, it is likely that presence of catabolic genes could serve as markers of their actual function. For example, in the case of PAH-degrading communities, strong positive correlations have previously been found between nahAc (α subunit of naphthalene dioxygenase) gene copies and transcripts, indicating that the presence of catabolic genes in a community is related to their activity (Fleming et al. 1993; Yagi and Madsen 2009). Thus, our observations implied that both nidA and nidA3 genes might play metabolic roles in pyrene biodegradation in contaminated soils, and they might provide useful targets for monitoring HMW-PAH degradation. As this is, to our knowledge, the first quantitative detection of nidA3 in PAH-degrading communities, this also presents the first demonstration of a link between nidA3 abundance to function (pyrene biodegradation) in indigenous microbial communities.

4.4 Different response patterns of functional microbial community between two soils

Notably, functional microbial community potentially involved in pyrene biodegradation in two soils showed different abundance and diversity after exposure to pyrene, indicating a soil condition dependent response of the functional soil bacteria to pyrene contamination. The abundances of PAH-RHDα GP gene, Mycobacterium 16S rDNA, nidA, and nidA3 were approximately 1 order of magnitude higher in HS after spiking with pyrene than in the LS detected at day 35 (Fig. 2). These results might be due to (1) the initial abundance of these functional genes in HS being already higher than in LS before artificial pyrene spiking, and (2) the more rapidly growing of microbial populations carrying functional genes in HS (Ding et al. 2010). The diversity of the PAH-RHDα GP genes was lower in HS than in LS after 35 days of pyrene incubation based on the value of the Shannon index and reciprocal of Simpson (Table 4). These differences in functional microbial community structure between two soils might attribute to lower initial microbial diversity in HS, which was heavily contaminated by PAHs. Previous studies showed that indigenous microbial communities in heavily PAH-contaminated soils were less diverse than those in less contaminated areas due to the toxicity of PAHs (Smalla et al. 2007).

5 Conclusions

In conclusion, this study revealed that Mycobacterium as well as uncultured Gram-positive PAH-RHDα genotypes may be the important group of pyrene degraders in soils, and two pyrene catabolic pathways, targeted by nidA and nidA3, might potentially contribute to in situ biodegradation of pyrene. Moreover, the soil environment was indicated to significantly affect the response of functional microbial populations and biological process to pyrene spiking. These findings provide new insights into in situ pyrene biodegradation in soils and help to identify functional microbes from different soils.