Hydrocarbon biostimulation and bioaugmentation in organic carbon and clay-rich soils

Soil Biology & Biochemistry 99 (2016) 66e74

Contents lists available at ScienceDirect

Soil Biology & Biochemistry journal homepage: www.elsevier.com/locate/soilbio

Hydrocarbon biostimulation and bioaugmentation in organic carbon and clay-rich soils
che a, Olivier Tromme d, Philippe Thonart b, c, Thibaut Masy a, b, c, *, Sandrine Demane Philippe Jacques b, c, Serge Hiligsmann c, Timothy M. Vogel a University of Lyon, Ecole Centrale de Lyon, Laboratoire Amp
ere, UMR CNRS 5005, Environmental Microbial Genomics Group, Avenue Guy de Collonges 36, 69134 Ecully, France b University of Li
ege, Gembloux Agro-Bio Tech, Microbial Processes and Interactions (MiPI), Passage des D eport es 2, 5030 Gembloux, Belgium c University of Li
ege, Faculty of Life Sciences, Walloon Center of Industrial Biology, Chemin de la Vall ee 2, 4000 Li
ege, Belgium d Sanifox SPRL, Rue Enhet-Centre 47, 5590 Chevetogne, Belgium a

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 December 2015 Received in revised form 20 April 2016 Accepted 23 April 2016

Hydrocarbon-contaminated organic carbon-rich clayey soils are challenging for bioremediation stakeholders since the pollutant is heterogeneously distributed and poorly bioavailable due to its strong adsorption on clay and organic particles. In addition, biodegradation rates are restricted by limited diffusion of oxygen and nutrients to hydrocarbon-degrading aerobes. This study assessed the benefits of bioaugmentation with the strain Rhodococcus erythropolis T902.1 versus those from biostimulation and anaerobic natural attenuation in terms of hydrocarbon (HC) degradation efficiency and changes in the bacterial community structure in a diesel-polluted clay-rich soil. Three soil samples with a similar total organic content but with a different HC concentration (0.2, 1.0 and 6.5 g/kg) were compared in a microcosm experiment. Despite a limitation in oxygen transfer, R. erythropolis T902.1 enhanced a greater HC degradation compared to the biostimulation treatment. However, this advantage decreased with time as the proportion of Rhodococci declined from 25% initially to 1% of the global community after 80 days of treatment. Similarly, the alkB gene proportion in bioaugmented soils decreased to levels close to those of biostimulated soils. Consequently, further engineering was suggested to improve the resilience of the inoculum to ensure its long-term presence and activity in such polluted environments. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Bioremediation Rhodococcus erythropolis Hydrocarbons (HC) Oxitop qPCR 16S-rRNA gene sequencing

1. Introduction Hydrocarbons (HC, including mineral oil, BTEX and PAH) are widely spread in nature as a result of both long-term use and frequent release. In Europe, they occur on average in 45% of identified contaminated sites and this frequency reaches 85% in Belgium (Van Liedekerke et al., 2014). Furthermore, they are particularly difficult to remediate in carbon-rich clayey soils because of their lower bioavailability and the difficulty to provide oxygen for fast HC degrading aerobes. Several limiting factors can be considered: heterogenic distribution of organic matter and associated HC, restricted oxygen and nutrients transfer,


ge, Gembloux Agro-Bio Tech, Microbial * Corresponding author. University of Lie porte s 2, 5030 Gembloux, Processes and Interactions (MiPI), Passage des De Belgium. E-mail addresses: [email protected] (T. Masy), [email protected] (T.M. Vogel). http://dx.doi.org/10.1016/j.soilbio.2016.04.016 0038-0717/© 2016 Elsevier Ltd. All rights reserved.

competition between carbon sources and between microorganisms. Consequently, even linear HC degrade slowly or not at all in such soils while they are relatively rapidly biodegraded in other soils. Treatment technologies that rely on stimulating indigenous microorganisms (biostimulation) or adding specific degraders (bioaugmentation) might be stymied by these rate-limiting conditions. Bioaugmentation is considered to have the same dependence on physicochemical conditions as biostimulation even if the active microbial population should be higher. Although bioaugmentation is an ecologically- and cost-effective technique, its efficiency and usefulness is thus debatable and there is a need for improved understanding of the causes that can lead to its failure, such as diverse environmental constraints and poor adaptation ability of laboratory-cultivated microorganisms (Vogel, 1996; El Fantroussi and Agathos, 2005; Boon and Verstraete, 2010; Tyagi et al., 2011). Rhodococcus sp. appears to be a good candidate for use in bioaugmentation, since this genus is ubiquitous (Bell et al., 1998) and

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74

can degrade a broad range of aliphatic, branched, cyclic, (polycyclic) aromatic, sulfur-containing and chlorinated HC (Martinkova et al., 2009; Larkin et al., 2010a, 2010b). Rhodococcus sp. can also withstand environmental stresses, such as low temperature, low water content, high pH, high salinity, lack of nutrient, presence of toxic solvents or pollutants (Whyte et al., 1999; de Carvalho, 2010, 2012; de Carvalho et al., 2014). Rhodococci synthesize mycolic acids and trehalolipids to modify their membrane hydrophobicity in order to protect themselves from these adverse environments and to solubilize and assimilate hydrophobic substrates that are poorly bioavailable in soil (Lang and Philp, 1998; Kuyukina et al., 2005; Kuyukina and Ivshina, 2010). In this study, the desiccationtolerant strain Rhodococcus erythropolis T902.1 was tested. This strain is suitable for large production in bioreactors and conservation in powdered form before its use on the field (Weekers et al., 1999). The objective of this work was to assess the potential benefits of bioaugmentation with R. erythropolis T902.1 versus those for aerobic biostimulation and anaerobic natural attenuation for the effective degradation of poorly available hydrocarbons in carbonrich soils. Three levels of pollution ([HCC10eC40] ¼ 0.2, 1.0 and 6.5 g/kg dry matter) were tested in a microcosm study. HC degradation and oxygen uptake were monitored to evaluate performance differences between the treatments. In addition, microbiological parameters thought to confirm the degradation processes were measured. These include the concentrations of the alkane monooxygenase (alkB) gene and the different microbial populations such as R. erythropolis. The alkB gene was widely studied to assay the biodegradation potential of n-alkanes in soils (e.g. Kloos et al., 2006; Powell et al., 2006). It encodes an alkane monooxygenase, i.e. a membrane-bound hydroxylase involved in the assimilation and oxidation of linear and branched HC, which catalyzes this critical reaction step enabling HCs to be used in the general microbial metabolism. This gene has been observed in contaminated soil (Kloos et al., 2006) and was expressed during active bioremediation of diesel-contaminated soils (Yergeau et al., 2009, 2012). However, alkB gene expression (transcripts) was not taken into account since it is not necessarily correlated to HC level and degrading activity, according to a previous study on chronically polluted sediments (Paisse et al., 2011). Finally, changes in the microbial community structure as a function of the HC level, time and the type of treatment were evaluated using 16S-rRNA (rrs) gene analysis.

67

piezometers developed). In addition, peat is also present and heterogeneously mixed within the loam in the least and moderately polluted soils, which contributes to their high carbon content. 2.2. Microcosms: experimental conditions and oxygen consumption monitoring Each soil underwent three treatments: an anaerobic natural attenuation (An), a biostimulation with air (Ox) and a bioaugmentation with air and R. erythropolis T902.1 (Re). A sample (150 g) of soil was placed in either a 250 mL bottle that had been degassed with N2 and sealed with a rubber cap (An) or in a precalibrated Oxitop® bottle that measured O2 consumption once every day (Ox and Re treatments). One digital unit measured by the Oxitop system corresponded approximately to 1.9 mg of O2 consumed in the bottle i.e. 20 mg O2/kg dry soil. Inoculation consisted of spraying soil with 2 mL of a concentrated resuspension (0.85% NaCl) of a pure liquid culture of R. erythropolis T902.1. This corresponded to a concentration of approximately 8  107 CFU per gram of soil. Two mL of sterile 0.85% NaCl solution were sprayed on non-inoculated soils. Each treatment was run in triplicate and the 27 bottles were maintained at 22  C without agitation. Bottles were opened every 5 days, or sooner if ¾ of the maximum oxygen uptake (30 digits measured by the manometer) was reached, in order to prevent oxygen limitation. During each bottle opening, bottles were agitated to ensure complete reoxygenation of the gas phase and dissolved NaOH pellets in the CO2 trap were replaced. 2.3. Hydrocarbon monitoring A C10-C40 HC analysis by GC-FID was performed on 20 g of soil sampled from the microcosms after 0, 40, 80 and 100 days (ISO 16703, Wessling, Saint-Quentin-Fallavier, France). 2.4. DNA extraction

2. Materials and methods

DNA was extracted from 0.30 g of wet soil after 0, 20, 40 and 80 days of the experiment, with the NucleoSpin® Soil kit (Macherey Nagel, Düren, Germany) following the manufacturer's instructions but with a double lysis step (SL1þSX then SL2). DNA was finally resuspended in 100 mL of elution buffer (5 mM TriseHCl, pH 8.5) and gDNA concentration was measured with the Qubit® dsDNA HS Assay kit and a Qubit® fluorometer (ThermoFisher Scientific, Waltham, USA). Then, DNA was stored at 20  C and kept on ice in further manipulations.

2.1. Soil sampling and characterization

2.5. qPCR analyses

Soil was sampled from a site polluted with diesel and heating fuel. The site is an operating truck-fill station where several potential leaks were assumed. A transect of eight 3 m-long wells (piezometers 501 to 508) was realized at one of the site border to delimit the contaminant plume (details are provided in Appendix A). From core drilling, three soil samples (3 kg) were selected and characterized (Table 1). They originate from three different horizons (20e100, 100e200 and 200e300 cm, cf. Appendix A), possess a similar total organic carbon content (2.6%) but are differently contaminated ([HCC10eC40] ¼ 0.2, 1.0 and 6.5 g/kg dry matter). The monoaromatic HC content is quite low: 3.39 ± 0.20 mg/kg of dry matter for the most polluted soil and concentrations are under limit of detection (LOD) for lesser polluted soil samples. The soil structure mainly consists of a low permeable and clay-rich loam that þ limits O2 diffusion (no NO 3 , 1 ± 0.6 mg/L NH4 , pO2 of 0.9 ± 0.4 mg/L and negative redox potential in the groundwater from the

Quantitative PCR were performed on the alkB gene coding for alkane monooxygenase and on the 16S-rRNA gene with either specific primers for the species R. erythropolis or wide-range primers covering Eubacteria (Table 2) (Bell et al., 1999; Fierer et al., 2005; Powell et al., 2006). The amplification performance was first assessed by a PCR with a gradient temperature (T gradient €ttingen, Germany) on DNA isolated from thermocyler, Biometra, Go the pure strain T902.1 and from some soil samples to establish the optimal annealing temperature (Table 2). Amplicon specificity was checked on a 2% agarose gel and with a GeneRuler Low range DNA ladder (ThermoFisher Scientific, Waltham, USA). For all primer pairs, the crude extracts of soil gDNA did not exhibit PCR or qPCR inhibition. Two mL of crude-extracted gDNA, DNA-free water or of standard DNA were added to a premix containing 0.8 mL of each primer at 10 mM, 6.4 mL of DNA-free water and 10 mL of 2 SensiFast™ SYBR No-Rox mix (Bioline, London, UK). All the qPCR reactions were

68

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74 Table 1 Principal physicochemical characteristics of selected soils for the microcosm experiment. DMC: Dry Matter Content; Ret. Cap.: Retention Capacity; CEC: Cationic Exchange Capacity; TOM and TOC: Total Organic Matter and Carbon; VAC: Volatile Aromatic Compounds (BTEX and aromatics); HC: linear hydrocarbons.

performed in duplicates on the Rotor-Gene® 6000 apparatus (Corbett Life Science-Qiagen, Venlo, Netherlands) with the following steps: 1) 3 min of gDNA denaturation at 95  C; 2) 40 cycles with denaturation (95  C, 5 s), annealing, elongation (cf. Table 2) and 3) signal acquisition at the end of the elongation step; and finally 4) a melt curve from 50  C to 95  C with 1  C increments. Raw data were analyzed with the Rotor-Gene Q Software v2 (Qiagen, Venlo, Netherlands). Calibration was performed in triplicates with PCR amplicons of each target gene from R. erythropolis T902.1 as standards. It gave a linear range from 103 to 109 copies of each target gene per mL.

algorithm (Masella et al., 2012), removing primer sequences and keeping amplicons with a length from 395 to 450 bp (overlap from 114 to 169 bp). Next, joined sequences were chimera-checked with UCHIME (Edgar et al., 2011) and the RDP-gold database before their processing with the command pipeline QIIME (Caporaso et al., 2010) as follows: they were clustered and assigned through the command pick_open_reference_otus (Uclust algorithm (Edgar, 2010)) with the latest Greengenes database available. Then, the OTUs table obtained was analyzed with core_diversity_analyses, compare_categories and group_significance. 2.7. Statistical analyses

2.6. 16S-rRNA gene sequencing and data processing The workflow proposed by Illumina for the sequencing of 16SrRNA gene amplicons (V3-V4 region) on the MiSeq system (http:// web.uri.edu/gsc/files/16s-metagenomic-library-prep-guide15044223-b.pdf) was slightly adapted as follows: amplicon PCR was done with Titanium Taq DNA polymerase and buffer (Clontech, Maidenhead, UK), dNTP (Invitrogen, Carlsbad, USA), PCR grade water and 3 ng of gDNA. The thermocycler ran for 27 cycles and then the products were purified on 1.5% agarose gel and with the Illustra GFX kit (GE Healthcare, Buc, France) based to manufacturers' instructions. Once sequenced, amplicons were demultiplexed and trimmed by the MiSeq System and forward and reverse reads, containing the target sequence with its corresponding primer, were compiled as FASTQ files. These reads were first joined with the PANDAseq

Repeated measures ANOVA as well as pairwise ManneWhitney and Student t tests (a ¼ 0.05) were carried out to detect differences between HC concentrations over time and to assess if a significant degradation occurred. Ordinary least square regression was also applied to correlate oxygen consumption and HC degradation for the different types of soil (Fig. 2B) or treatment times (Fig. 3, Figs. C.1 and C.2). 3. Results 3.1. Hydrocarbon and oxygen monitoring First, a C10-C40 hydrocarbon (HC) analysis was performed on soils sampled from the microcosms during their incubation under the three conditions (anaerobic (An), aerobic (Ox), aerobic with

Table 2 Genes amplified by qPCR and experimental conditions. Gene name

Primer name

Primer sequence

alkB

alkbF alkbR Eub338 Eub518 Re1 f

AAC TAC ATC GAG CAC TAC GG 59  Ce5 sec TGA AGA TGT GGT TGC TGT TCC ACT CCT ACG GGA GGC AGC AG 60  Ce20 sec ATT ACC GCG GCT GCT GG CGT CTA ATA CCG GAT ATG ACC TCC 64  Ce10 sec TAT C GCA AGC TAG CAG TTG AGC TGC TGG T

16S-Eub 16S-Re

Re2 r

Annealing step

Elongation step

Amplified zone

Amplicon Specificity size (bp)

References

72  Ce10 sec

alkB

±100

Unknown

Powell et al., 2006

V3 region

±200

Eubacteria

Fierer et al., 2005

Rhodococcus erythropolis

Bell et al., 1999 Stackebrandt et al., 1988

72  Ce10 sec

V2eV3 region 449 165e192/609e633

HC concentration (mg/kg dry soil)

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74

8000

1500

7000

1200

6000

900

5000

600

T0

T40

T80

69

1200

T100

1000 800 600 400

4000

300

3000

200

0

High An

High Ox

High Re

0

Med. An

Med. Ox

Med. Re

Low An

Low Ox

Low Re

Fig. 1. Hydrocarbon concentration evolution with time in the three different soils (High, Medium and Low contamination) under the three treatments (An, Ox and Re). Error bars correspond to standard errors of the mean.

Oxygen consumption (g O2/kg of dry soil)

A

comparing treatments Re and Ox, the addition of R. erythropolis T902.1 seems to improve the use of oxygen towards a higher HC degradation, as illustrated in Fig. 3, Figs. C.1 and C.2 (see appendix C). Several correlations were calculated on the triplicates, separated or pooled together (Table C.1) and this trend was confirmed at each sampling time, especially at higher pollution. However, this improvement of degradation yield linked to the bioaugmentation is decreasing with time, as regression lines from treatments Re and Ox are converging with time (Figs. C.1 and C.2 in appendix C).

3.2. qPCR experiments In order to support the chemical results obtained in previous section, changes in HC degrading community through the three treatments were assessed. A qPCR of the alkB gene, coding for alkane hydroxylases, was performed on the DNA extracted from soils after 0, 20, 40 and 80 days. DNA extraction yields are provided in Appendix D (Fig. D.1C). They generally increased with time for the aerobic treatments (with fluctuations between sampling times), which is consistent with microbial growth, whereas they fluctuated or decreased for the anaerobic treatment. From these yields, the alkB gene concentration was calculated in two different ways: by the number of copies per gram of soil (Fig. 4A), or by the number of copies per nanogram of DNA (Fig. 4B). As the microbial community was developing, an increase in the total alkB gene content was expected under aerobic conditions. This is the case for the most polluted soil (Fig. 4A, red lines) and this trend was delayed for the moderately polluted soil with a final increase between 40 and 80 days (orange lines). However, for the least polluted soil, the

9

High Re

B

2.5

8

High Ox

Hydrocarbon degradation (g HC/kg dry soil)

added R. erythropolis T902.1 (Re)) after 0, 40, 80 and 100 days (respectively T0, T40, T80 and T100) (Fig. 1). GC-FID chromatograms (Fig. B.1) and HC fractions distribution (Table B.1) are provided in Appendix B for further details on the HC degradation. For highly and moderately contaminated soils, the mean trend showed HC degradation under each condition. The decrease of the mean concentration was the greatest for the bioaugmented (Re) soil then less for the biostimulated soil (Ox), and finally the least for the anaerobic attenuation (An) (pvalue ¼ 0.009). However, soil heterogeneity hampered most statistical interpretations: there was no significant difference between treatments and likewise no significant HC degradation was detected compared to initial HC amounts, except that HC values are lower after 100 days in the most polluted soil, based on ManneWhitney tests (Bonferroni pvalue ¼ 0.015). Furthermore, when the three treatments are pooled together (n ¼ 9), Student t-tests revealed a significative HC degradation after 100 days for the soils highly (pvalue ¼ 0.003) and moderately (pvalue ¼ 0.031) contaminated, compared to initial concentrations. By contrast, the least polluted soil HC content was too variable from one sampling to another (from <20 mg/kg to 1600 mg/kg of dry matter) and the overall mean HC concentration of 304 ± 358 mg/kg prevented any conclusion about the HC trend. While the three soils contain approximately the same total organic carbon content at the beginning of the experiment (Table 1), oxygen uptake is related to the HC level, independently from the type of aerobic treatment (Fig. 2A). Based on sampling times (T40, T80 and T100), a regression was calculated for each HC level between the HC degradation compared to the initial concentration and the respective oxygen uptake (Fig. 2B). When

2.0

7

Med. Ox

6

Med. Re

5

Low Re

4

Low Ox

3

Sampling

2 1

Low

Medium

1.5 1.0

y = 0.1729x - 0.4258 R² = 0.6835

0.5 0.0

y = 0.0007x - 0.1643 R² = 1E-05

-0.5

0 0

20

40

60

80

Time after inoculation (days)

100

High

y = 0.3514x - 1.1698 R² = 0.7021

1

2

3

4

5

6

7

8

9

Oxygen consumption (g O2/kg dry soil)

Fig. 2. A: Oxygen consumption evolution with time from the three different soils under the two aerobic treatments (Re, full symbols and Ox, empty symbols). B: Correlation between hydrocarbon degradation and oxygen consumption for each soil type. Red diamonds, orange dots, and green triangles: high, medium and low pollution. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

70

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74

A

B

Triplicates, after 100 days

Hydrocarbon degraded (g HC/kg of dry soil)

3.0

Re y = 0.436x - 1.745 R² = 0.712

2.5 2.0

Re

2.0

Ox

1.5

Ox y = 0.359x - 1.424 R² = 0.758

1.5

Means, all sampling times

2.5

y = 0.241x - 0.798 R² = 0.587

1.0

1.0

y = 0.389x - 1.292 R² = 0.835

0.5

0.5

0.0

0.0 -0.5

-0.5 1

2

3

4

5

6

7

8

9

Oxygen consumption (g O2/kg of dry soil)

10 11

1

2

3

4

5

6

7

8

Oxygen consumption (g O2/kg of dry soil)

9

Fig. 3. Correlations between hydrocarbon degradation and oxygen consumption for both aerobic treatments. A: after 100 days with triplicates separated; B: for all sampling times with triplicates pooled together (means). Black triangles: treatment Re; grey diamonds: treatment Ox.

B

1E+7

AlkB concentration (copies/ng of DNA)

AlkB concentration (copies/g of soil)

A

1E+6

High Ox High Re High An 1E+5 20

40

60

80

4000 3000 2000 1000 0

100

1E+7

20

40

60

80

100

6000

AlkB concentration (copies/ng of DNA)

AlkB concentration (copies/g of soil)

High Re High Ox High An

5000

0 0

1E+6

Medium Re Medium Ox Medium An 1E+5

Medium Re Medium Ox Medium An

5000 4000 3000 2000 1000 0

20

40

60

1E+7

80

Low Ox Low Re Low An

1E+6

0

100

1E+5

1E+4

20

40

60

8000

AlkB concentration (copies/ng of DNA)

0

AlkB concentration (copies/g of soil)

6000

80

100

Low Re Low Ox Low An

7000 6000 5000 4000 3000 2000 1000 0

0

20

40

60

80

Time after inoculation (days)

100

0

20

40

60

80

100

Time after inoculation (days)

Fig. 4. qPCR of the alkB gene for each HC level and the three treatments applied (Re, Ox and An). AlkB gene concentration is expressed as the number of copies per gram of soil (A) or per ng of gDNA (B).

total alkB gene content, in addition to being lower initially compared to the other soils, remained stable at 106 copies/g and then decreased (Fig. 4A, green lines), which indicated a possible loss of HC degrading potential.

The second set of graphs (Fig. 4B) depict the relative proportion of alkB genes to a fixed amount of DNA and highlight the decreasing effect of the bioaugmentation (Re, darker colors): the proportion of alkB genes supplied by the inoculated strain was probably

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74

71

Table 3 Proportion of R. erythropolis in the bacterial community detected by qPCR and of Rhodococcus sp. detected by 16S-rRNA gene sequencing after chimera removal, OTUs clustering and assignment.

dominant at the beginning of the experiment but became limited after 40 days. In order to link this trend with the proportion of R. erythropolis, two qPCR on the 16S-rRNA gene were implemented: a broad range qPCR for numerating all Bacteria and another specific qPCR that targeted the species R. erythropolis. Table 3 gives the ratio between both measurements for the samples corresponding to the treatment Re. For the treatments Ox and An, no R. erythropolis was detected, but a few non-specific amplification occurred (different melt curve), which may indicate a possible presence of other related species. Finally, the ratio between alkB and 16S-rRNA genes was also plotted (Fig. D.1D, see appendix D), which led to the same observations as those obtained from Fig. 4A and B: the proportion of aerobic HC-degraders in the community was mainly dependant of the HC level present in the soil and the treatment applied. Nevertheless, curves from Fig. 4B and Fig. D.1D differ slightly as the alkB/ 16S ratio does not take into account Eukarya. Even if the aerobic treatment led to bacterial growth (Fig D.1E), the proportion of 16S genes per ng of gDNA decrease, which might indicate a relative higher growth of eukarya (e.g. fungi) in the aerated soils, whereas anaerobic soils were proportionally enriching in bacteria (Fig. D.1F). In addition, the initial concentration of the inoculum was estimated at 8.0  107 CFU/g of dry soil. After 80 and 100 days, samples from the most polluted soil were diluted and plated onto mineral medium (Bushnell-Haas Broth) complemented with rifampicine (50 mg/L), cycloheximide (120 mg/L) and diesel (2 mL/L) to appraise inoculum residual concentrations. While no colony developed in the negative controls (samples from An and Ox treatments), characteristic colonies of the strain T902.1 were present at concentrations of 3.5  106 and 1.1  106 CFU/g of dry soil respectively after 80 and 100 days. Similarly, R. erythropolis-specific 16S gene copy number decreased from initially 9.9  107 to 1.3  107 copies/g of dry soil after 80 days in that soil, whereas it decreased from 7.7  107 to 3.7  106 and from 3.6  107 to 3.7  106 copies/g in moderately and low polluted-soils respectively. 3.3. 16S-rRNA gene sequencing Along with qPCR analyses, 16S-rRNA gene sequencing was performed to detect changes in the bacterial community for the three treatments applied on the three soils. The proportion of Rhodococcus sp. assigned sequences correlated well with the relative quantity measured by qPCR (Table 3). The quantity of endogenous Rhodococci detected in unamended soils (treatments Ox and An) never reached above 0.062%, while in the amended soil (Re) the inoculated strain decreased from approximately 25%e1% after 80 days. The predominance of some taxa under specific experimental conditions was assessed by principal component analysis (PCA) (Fig. 5) and by significance tests from the OTUs table (Bonferroni pvalue <0.05). The taxa significantly related to a specific treatment or a soil are displayed in Table 4. The highly polluted soil is

separated from the moderately and lowly polluted soils, which are more closely related (Fig. 5A). The third PC seems to be mainly dependent on the presence of Rhodococcus sequences (Fig. 5B), whose impact on the global community is decreasing with time. In addition, other algorithms were also tested, such as BrayeCurtis distances calculation (Fig. E.1, see appendix E), that avoids overrepresentation of abundant OTUs. Even if a shift of the points is observed with time for these graphs, it was not attributable to the appearance of specific taxa, but rather to a decrease of the alpha diversity partly linked to the aerobic treatments (data not shown). Proteobacteria constituted the main phylum representing on average two thirds of the bacterial community (Table E.1). Only slight changes were observed with time at the phylum level, such as an increase of the Acidobacteria proportion (Fig. E.2). 4. Discussion HC degradation and oxygen uptake seemed to be related to the ratio between the HC concentration and the total organic carbon content (Figs. 1 and 2, Table 1). The higher oxygen demand required to degrade hydrocarbons (about 3.5 g of O2 per g of HC) compared to other organic compounds that might be easier to metabolize as well as their relative abundance and availability in the soil could explain this relation. As described in previous studies (Weissenfels et al., 1992; Lashermes et al., 2010; Yang et al., 2011), the prevalence of organic matter probably masks the contamination due to sequestration and humification. These processes could also explain why HCs are much more heterogeneously distributed and form a persistent residual fraction in less contaminated soils, where peat was present. For example, the least polluted soil had a basal oxygen consumption that did not seem to lead to overall HC degradation, even if some hot spots of contamination were detected amongst the samples. In addition, the global coefficient of variation (from all the HC measurements taken together) was 30.5%, 87.2% and 118.1% for the most, moderately low and least polluted soils, respectively, whereas their initial HC/TOC ratios were of 21.3, 2.9 and 0.6%, respectively. Another factor explaining higher O2 consumption was the higher bacterial and alkB gene content linked to the higher HC level in the soil (Fig. 4A and Fig. D.1E). The existence of a correlation between degrading genes and HC concentrations in soils has been bron et al., 2008; reported previously (e.g. Powell et al., 2006; Ce Yergeau et al., 2012). In contrast, other studies did not observe any correlation when HC concentrations were very low (e.g. Wasmund et al., 2009). In spite of the lack of statistically significant differences in HC content between treatments, the general trend showed a higher efficiency in the oxygen usage towards HC degradation when R. erythropolis T902.1 was added, which is even more critical when the pollution concentration is higher (Fig. 3 and Appendix C). From correlations obtained in Fig. 3B, bioaugmented soils required 5.89 g of O2 to degrade 1 g of HC on average, whereas biostimulated soils required 7.47 g/g. This difference is important when considering the oxygen limitation due to soil structure and high organic carbon

72

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74 40

H-Re-T20

A

H-Ox-T40

30

H-An-T80 H-An-T20 H-Ox/An-T0/T40 H-Ox-T20 H-Ox-T80 H-Re-T40

80

H-Re-T80

0

g_Geothrix g_Methylibium L-Ox-T20 L-Ox-T40 g_Acidovorax

o_Ellin6067

o_Solibacterales

-10 -20 -30

f_Alteromonadaceae; g_HB2-32-21

f_Anaerolinaceae; g_WCHB1-05 o_Myxococcales o_Bacteroidales

g_Rhodococcus

c_Betaproteobacteria

L-Re-T40

g_Methylotenera

g_Parvibaculum

L-An-T20 M-An-T20 L-An-T80 M-An-T80 L-An-T40 g_Gallionella L-Re-T80 L-Ox-T80 o_Methylophilales M-Re-T0 f_Xanthomonadaceae M-Re-T20 f_Chitinophagaceae M-An-T40 L-Ox/An-T0 M-Ox-T20 o_SBla14 M-Ox/An-T0 M-Ox-T40 M-Re-T40 g_Rhodoplanes f_ACK-M1 L-Re-T0 f_Acidobacteriaceae

L-Re-T20

M-Ox-T80

f_Bradyrhizobiaceae f_Rhodospirillaceae

40 30

20

40

M-Re-T20

20 M-Re-T40

-10

50

g_Thiobacillus

0

g_Thiobacillus

60

70

PC1 (58.87%)

o_PYR10d3

f_Rhodocyclaceae

f_Bradyrhizobiaceae g_Methylibium f_Rhodospirillaceae o_Methylophilales

M-Re-T80

30

g_Geothrix L-Re-T40

-20 10

g_Rhodococcus

10

-40 0

H-Re-T0 L-Re-T20

50

o_PYR10d3

PC3 (9.30%)

PC2 (16.78%)

f_Comamonadaceae

10

L-Re-T0

M-Re-T0

60

H-Re-T0

20

B

70

f_Rhodocyclaceae

g_Gallionella L-Re-T80

o_SBla14 L-Ox/An-T0 M-Re-T80 f_Xanthomonadaceae

L-An-T80 L-An-T20 L-An-T40 M-Ox-T40 M-An-T80 M-Ox-T0/20 M-Ox-T80 M-An-T40 L-Ox-T80 M-An-T20

-40

-30

-20

-10

0

f_Anaerolinaceae; g_WCHB1-05 H-Re-T40 H-Re-T20 o_Bacteroidales H-Ox-T40 f_Comamonadaceae H-Re-T80 H-Ox-T20 f_Alteromonadaceae; L-Ox-T20 g_HB2-32-21 H-AnL-Ox-T40 H-Ox-T80 T0/20/40/80

10

20

30

40

PC2 (16.78%)

Fig. 5. Pearson PCA biplots of the experimental conditions (dots) and related major taxa assigned after OTUs clustering (diamonds). PC1, PC2 and PC3 represent the three first principal components which together explain 85% of the total variability. H: heavily polluted soil, M: moderately polluted soil and L: least polluted soil.

Table 4 Taxa significantly represented in a specific experimental variable (Bonferroni p-value <0.05). Taxa in bold correspond to the major drivers of the variability depicted in Fig. 5.

content (Table 1). The supply of supplementary degrading genes from the added Rhodococcus in the community (Fig. 4B) was probably the main factor for this more efficient treatment.

However, the effect of bioaugmentation was mostly limited to the first 40 days as shown by the converging curves between treatments Ox and Re (Figs. C1 and C2, Fig. 4B) and the decreasing

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74

proportion of Rhodococcus with time (Table 3 and Fig. 5). Based on qPCR analysis and plate counting, initial inoculum concentrations decreased on average by approximately 90% in 80 days to finally reach 5  106 cells/g of dry soil. Competition and predation from endogenous microorganisms should have regulated the initial high presence of this strain compared to the total flora. In addition, the initial proportion of Rhodococcus sp. (25% of the total bacterial community) could also be overestimated as the strain T902.1 was sprayed onto the soil and thus could be recovered with a higher efficiency than the endogenous flora during the DNA extraction. Nonetheless, the strain T902.1 still represents 1% of the bacterial community after 80 days (Table 3) and should be considered ecologically important since Rhodococcus contains many hydrocarbon degrading genes (McLeod et al., 2006; Martinkova et al., 2009). Furthermore, the similarity between qPCR and sequencing results in Table 3 suggests that any potential bias from the sequencing analysis was limited. The more polluted soil seemed to contain a higher diversity of HC-degrading endogenous bacteria. Most taxa significantly represented in the heavily polluted soil (Fig. 5 and Table 4) include bacteria either previously mentioned as degraders, such as Rhodocyclaceae (Oren, 2014; Singleton et al., 2015), the pyrenedegrading order PYR10d3 (Singleton et al., 2006), iron-reducing Geothrix sp. (Coates et al., 1999) and MTBE-degrading Methylibium sp. (Nakatsu et al., 2006); or detected in contaminated environments, such as the uncultured bacteria WCHB1-05 (Anaerolinaceae) originally coming from a jet-fuel-polluted aquifer (Dojka et al., 1998) and HB2-32-21 (Altermonadaceae) found in coastal sedi et al., 2010; Koo et al., 2014). In ments after an oil spill (Païsse moderately and weakly polluted soils, some taxa could also be related to hydrocarbon degraders, such as Xanthomonadaceae (Chang and Zylstra, 2010), or Thiobacillus and Methylophilales that can degrade dimethylsulfide (Eyice et al., 2015). Additionally, these latter two taxa and Gallionella sp. could be involved in iron and sulfur oxidation linked to reoxygenation of humified organic matter present in these soils. However, the presence of the other significant taxa is more difficult to interpret due to limited available information (order SBla14) or due to their high bacterial and functional diversity (e.g. Acidobacteriaceae, Bradyrhizobiaceae). Therefore, 16S-rRNA gene sequencing analysis was consistent with the higher degrading potential in the most polluted soil, while in moderately low and least polluted soils, the natural organic matter (peat) selected other taxa apparently not implicated in HC degradation. Furthermore, the persistence of several taxa including anaerobic bacteria (Table 4) highlights the limitation of oxygen transfer in this asphyxiating soil, especially since treatments Ox and An are not clearly separated by PCA (Fig. 5 and Fig. E.1). Passive aeration was not sufficient for such soil and the use of oxidants or air enriched in oxygen as well as a dynamic system to increase exchange surface area should increase degradation rates. The dominance of Proteobacteria and Actinobacteria in organic matter-rich HC-polluted soils has been reported previously (Bell et al., 2013). In that study, Betaproteobacteria abundance was also positively correlated to higher diesel degradation in biostimulated soils. Yergeau et al. (2012) also mentioned that Pseudomonas and Rhodococcus species were successively dominant in Artic biopile soils and actively expressed HC-degrading genes. In another study from Sun et al. (2015), Actinobacteria (8), Acidobacteria (8) and Proteobacteria (5) were the most abundant phyla in 21 out of 24 contaminated top soils sampled from 6 different oilfield sites in China. In addition, alkB gene quantification consisted in a first approach to describe functional changes during the three treatments applied but this is not the sole degrading-gene and selected primers might have a limited specificity. A metagenomic analysis could provide further

73

information on the whole degrading capacity of the endogenous flora and on the complementary genes brought by the inoculated strain. Even if soil heterogeneity hindered statistical interpretations, all the results are generally consistent with an effective degradation occurring in highly and moderately polluted soils whereas the least polluted soil had basal oxygen consumption with no visible HC degradation. Increasing the number of samples and/or of replicates might reduce uncertainties. However, the amount of soil in the microcosms (150 g) limited the number of samples for representative HC analysis (at least 25 g per sample were necessary). Nevertheless, bioaugmentation with R. erythropolis T902.1 supplied a supplementary degrading capacity and might have redirected oxygen uptake towards a higher HC degradation and away from the degradation of natural organic matter. As a consequence, this bioaugmentation might reduce treatment time in soils containing high concentrations of organic matter. Acknowledgment
la This work was supported by a Fonds pour la Formation a Recherche dans l’Industrie et dans l’Agriculture (F.R.I.A.) grant funded by the Belgian National Fund for Scientific Research (Fonds National de la Recherche Scientifique, FNRS) and attributed to T. Masy. Supplementary data As referenced above, complementary information is presented as appendices on (A) the site characterization, (B) hydrocarbon monitoring, (C) the correlation between oxygen uptake and hydrocarbon degradation, (D) the qPCR analyses and (E) handling of 16S-rRNA gene sequencing data. These supplementary data can be found at http://dx.doi.org/10.1016/j.soilbio.2016.04.016. References Bell, K.S., Philp, J.C., Aw, D.W.J., Christofi, N., 1998. The genus Rhodococcus. J. Appl. Microbiol. 85, 195e210. Bell, K.S., Kuyukina, M.S., Heidbrink, S., Philp, J.C., Aw, D.W.J., Ivshina, I.B., Christofi, A., 1999. Identification and environmental detection of Rhodococcus species by 16S rDNA-targeted PCR. J. Appl. Microbiol. 87, 472e480. Bell, T.H., Yergeau, E., Maynard, C., Juck, D., Whyte, L.G., Greer, C.W., 2013. Predictable bacterial composition and hydrocarbon degradation in Arctic soils following diesel and nutrient disturbance. ISME J. 7, 1200e1210. Boon, N., Verstraete, W., 2010. Bioaugmentation of hydrocarbons. In: Timmis, K.N. (Ed.), Handbook of Hydrocarbons and Lipid Microbiology. Springer-Verlag, Heidelberg, pp. 2531e2543. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., Fierer, N., Pena, A.G., Goodrich, J.K., Gordon, J.I., Huttley, G.A., Kelley, S.T., Knights, D., Koenig, J.E., Ley, R.E., Lozupone, C.A., McDonald, D., Muegge, B.D., Pirrung, M., Reeder, J., Sevinsky, J.R., Turnbaugh, P.J., Walters, W.A., Widmann, J., Yatsunenko, T., Zaneveld, J., Knight, R., 2010. QIIME allows analysis of highthroughput community sequencing data. Nat. Methods 7, 335e336. bron, A., Norini, M.-P., Beguiristain, T., Leyval, C., 2008. Real-Time PCR quantifiCe cation of PAH-ring hydroxylating dioxygenase (PAH-RHDa) genes from gram positive and gram negative bacteria in soil and sediment samples. J. Microbiol. Methods 73, 148e159. Chang, H.K., Zylstra, G.J., 2010. Xanthomonads. In: Timmis, K.N. (Ed.), Handbook of Hydrocarbon and Lipid Microbiology. Springer Berlin Heidelberg, pp. 1805e1811. Coates, J.D., Ellis, D.J., Gaw, C.V., Lovley, D.R., 1999. Geothrix fermentans gen. nov., sp. nov., a novel Fe(III)-reducing bacterium from a hydrocarbon-contaminated aquifer. Int. J. Syst. Evol. Microbiol. 49, 1615e1622. de Carvalho, C.C.C.R., 2010. Adaptation of Rhodococcus to organic solvents. In: Alvarez, H.M. (Ed.), Biology of Rhodococcus. Springer Berlin Heidelberg, pp. 109e131. de Carvalho, C.C.C.R., 2012. Adaptation of Rhodococcus erythropolis cells for growth and bioremediation under extreme conditions. Res. Microbiol. 163, 125e136. de Carvalho, C.C.C.R., Marques, M.P.C., Hachicho, N., Heipieper, H.J., 2014. Rapid adaptation of Rhodococcus erythropolis cells to salt stress by synthesizing polyunsaturated fatty acids. Appl. Microbiol. Biotechnol. 98, 5599e5606. Dojka, M.A., Hugenholtz, P., Haack, S.K., Pace, N.R., 1998. Microbial diversity in a

74

T. Masy et al. / Soil Biology & Biochemistry 99 (2016) 66e74

hydrocarbon- and chlorinated-solvent-contaminated aquifer undergoing intrinsic bioremediation. Appl. Environ. Microbiol. 64, 3869e3877. Edgar, R.C., 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26 (19), 2460e2461. Edgar, R.C., Haas, B.J., Clemente, J.C., Quince, C., Knight, R., 2011. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194e2200. El Fantroussi, S., Agathos, S.N., 2005. Is bioaugmentation a feasible strategy for pollutant removal and site remediation? Curr. Opin. Microbiol. 8, 268e275. Eyice, O., Namura, M., Chen, Y., Mead, A., Samavedam, S., Schafer, H., 2015. SIP metagenomics identifies uncultivated Methylophilaceae as dimethylsulphide degrading bacteria in soil and lake sediment. ISME J. 9, 2336e2348. Fierer, N., Jackson, J.A., Vilgalys, R., Jackson, R.B., 2005. Assessment of soil microbial community structure by use of taxon-specific quantitative PCR assays. Appl. Environ. Microbiol. 71, 4117e4120. Kloos, K., Munch, J.C., Schloter, M., 2006. A new method for the detection of alkanemonooxygenase homologous genes (alkB) in soils based on PCR-hybridization. J. Microbiol. Methods 66, 486e496. Koo, H., Mojib, N., Thacker, R., Bej, A., 2014. Comparative analysis of bacterial community-metagenomics in coastal Gulf of Mexico sediment microcosms following exposure to Macondo oil (MC252). Antonie van Leeuwenhoek 106, 993e1009. Kuyukina, M.S., Ivshina, I.B., Makarov, S.O., Litvinenko, L.V., Cunningham, C.J., Philp, J.C., 2005. Effect of biosurfactants on crude oil desorption and mobilization in a soil system. Environ. Int. 31, 155e161. Kuyukina, M.S., Ivshina, I.B., 2010. Rhodococcus biosurfactants: biosynthesis, properties, and potential applications. In: Alvarez, H.M. (Ed.), Biology of Rhodococcus. Springer Berlin Heidelberg, pp. 291e313. Lang, S., Philp, J.C., 1998. Surface-active lipids in rhodococci. Antonie van Leeuwenhoek 74, 59e70. Larkin, M.J., Kulakov, L.A., Allen, C.C.R., 2010a. Rhodococcus. In: Timmis, K.N. (Ed.), Handbook of Hydrocarbon and Lipid Microbiology. Springer-Verlag, Heidelberg, pp. 1839e1852. Larkin, M.J., Kulakov, L.A., Allen, C.C.R., 2010b. Rhodococcus: genetics and functional genomics. In: Timmis, K.N. (Ed.), Handbook of Hydrocarbon and Lipid Microbiology. Springer-Verlag, Heidelberg, pp. 1345e1353. Lashermes, G., Houot, S., Barriuso, E., 2010. Sorption and mineralization of organic pollutants during different stages of composting. Chemosphere 79, 455e462. Martinkova, L., Uhnakova, B., Patek, M., Nesvera, J., Kren, V., 2009. Biodegradation potential of the genus Rhodococcus. Environ. Int. 35, 162e177. Masella, A., Bartram, A., Truszkowski, J., Brown, D., Neufeld, J., 2012. PANDAseq: paired-end assembler for illumina sequences. BMC Bioinforma. 13, 31. McLeod, M.P., Warren, R.L., Hsiao, W.W.L., Araki, N., Myhre, M., Fernandes, C., Miyazawa, D., Wong, W., Lillquist, A.L., Wang, D., Dosanjh, M., Hara, H., Petrescu, A., Morin, R.D., Yang, G., Stott, J.M., Schein, J.E., Shin, H., Smailus, D., Siddiqui, A.S., Marra, M.A., Jones, S.J.M., Holt, R., Brinkman, F.S.L., Miyauchi, K., Fukuda, M., Davies, J.E., Mohn, W.W., Eltis, L.D., 2006. The complete genome of Rhodococcus sp. RHA1 provides insights into a catabolic powerhouse. Proc. Natl. Acad. Sci. 103, 15582e15587. Nakatsu, C.H., Hristova, K., Hanada, S., Meng, X.-Y., Hanson, J.R., Scow, K.M., Kamagata, Y., 2006. Methylibium petroleiphilum gen. nov., sp. nov., a novel methyl tert-butyl ether-degrading methylotroph of the Betaproteobacteria. Int. J. Syst. Evol. Microbiol. 56, 983e989. Oren, A., 2014. The family Rhodocyclaceae. In: Rosenberg, E., DeLong, E., Lory, S., Stackebrandt, E., Thompson, F. (Eds.), The Prokaryotes. Springer Berlin Heidelberg, pp. 975e998.

~ i-Urriza, M., 2011. Are alkane hydroxylase genes Paisse, S., Duran, R., Coulon, F., Gon (alkB) relevant to assess petroleum bioremediation processes in chronically polluted coastal sediments? Appl. Microbiol. Biotechnol. 92, 835e844. , S., Gon ~ i-Urriza, M., Coulon, F., Duran, R., 2010. How a bacterial community Païsse originating from a contaminated coastal sediment responds to an oil input. Microb. Ecol. 60, 394e405. Powell, S.M., Ferguson, S.H., Bowman, J.P., Snape, I., 2006. Using real-time PCR to assess changes in the hydrocarbon-degrading microbial community in Antarctic soil during bioremediation. Microb. Ecol. 52, 523e532. Singleton, D.R., Sangaiah, R., Gold, A., Ball, L.M., Aitken, M.D., 2006. Identification and quantification of uncultivated Proteobacteria associated with pyrene degradation in a bioreactor treating PAH-contaminated soil. Environ. Microbiol. 8, 1736e1745. Singleton, D.R., Dickey, A.N., Scholl, E.H., Wright, F.A., Aitken, M.D., 2015. Complete genome sequence of a novel bacterium within the family Rhodocyclaceae that degrades polycyclic aromatic hydrocarbons. Genome Announc. 3 e0025100215. Stackebrandt, E., Smida, J.A.N., Collins, M.D., 1988. Evidence of phylogenetic heterogeneity withing the genus Rhodococcus: revivial of the genus Gordona (Tsukamura). J. Gen. Appl. Microbiol. 34, 341e348. Sun, W., Li, J., Jiang, L., Sun, Z., Fu, M., Peng, X., 2015. Profiling microbial community structures across six large oilfields in China and the potential role of dominant microorganisms in bioremediation. Appl. Microbiol. Biotechnol. 99, 8751e8764. Tyagi, M., da Fonseca, M.M.R., de Carvalho, C.C.C.R., 2011. Bioaugmentation and biostimulation strategies to improve the effectiveness of bioremediation processes. Biodegradation 22, 231e241. Van Liedekerke, M., Prokop, G., Rabl-Berger, S., Kibblewhite, M., Louwagi, G., 2014. Progress in the Management of Contaminated Sites in Europe. European Commission - Joint Research Center, Luxembourg, p. 72. Vogel, T.M., 1996. Bioaugmentation as a soil bioremediation approach. Curr. Opin. Biotechnol. 7, 311e316. Wasmund, K., Burns, K.A., Kurtboke, D.I., Bourne, D.G., 2009. Novel alkane hydroxylase gene (alkB) diversity in sediments associated with hydrocarbon seeps in the Timor Sea, Australia. Appl. Environ. Microbiol. 75, 7391e7398. Weekers, F., Jacques, P., Springael, D., Mergeay, M., Diels, L., Thonart, P., 1999. Improving the catabolic functions of desiccation-tolerant soil bacteria. In: Davison, B.H., Finkelstein, M. (Eds.), Twentieth Symposium on Biotechnology for Fuels and Chemicals. Humana Press, Gatlinburg, Tennesee, pp. 251e266. Weissenfels, W., Klewer, H.-J., Langhoff, J., 1992. Adsorption of polycyclic aromatic hydrocarbons (PAHs) by soil particles: influence on biodegradability and biotoxicity. Appl. Microbiol. Biotechnol. 36, 689e696.
re, L., Koval, S.F., Whyte, L.G., Slagman, S.J., Pietrantonio, F., Bourbonnie Lawrence, J.R., Inniss, W.E., Greer, C.W., 1999. Physiological adaptations involved in alkane assimilation at a low temperature by Rhodococcus sp. strain Q15. Appl. Environ. Microbiol. 65, 2961e2968. Yang, Y., Zhang, N., Xue, M., Lu, S.T., Tao, S., 2011. Effects of soil organic matter on the development of the microbial polycyclic aromatic hydrocarbons (PAHs) degradation potentials. Environ. Pollut. 159, 591e595. Yergeau, E., Arbour, M., Brousseau, R., Juck, D., Lawrence, J.R., Masson, L., Whyte, L.G., Greer, C.W., 2009. Microarray and real-time PCR analyses of the responses of high-arctic soil bacteria to hydrocarbon pollution and bioremediation treatments. Appl. Environ. Microbiol. 75, 6258e6267. Yergeau, E., Sanschagrin, S., Beaumier, D., Greer, C.W., 2012. Metagenomic analysis of the bioremediation of diesel-contaminated Canadian high Arctic soils. PLoS One 7, e30058.