Hydrocarbon-degrading potential of microbial communities from Arctic plants



Lyle G. Whyte, Department of Natural Resource Sciences, McGill University, Macdonald Stewart Building, 21111 Lakeshore Road, Ste. Anne de Bellevue, Montreal, QC H9X 3V9, Canada. E-mail: lyle.whyte@mcgill.ca



To explore rhizospheric microbial communities from Arctic native plant species evaluating their bacterial hydrocarbon-degrading capacities.

Methods and Results

Eriophorum scheuchzeri, Potentilla cf. rubricaulis, Oxyria digyna, Salix arctica and Puccinellia angustata plant species were collected at Eureka (Canadian high Arctic) along with their rhizospheric soil samples. Their bacterial community fingerprints (16S rRNA gene, DGGE) were distinctive encompassing members from the phyla: Bacteroidetes, Firmicutes, Actinobacteria and Proteobacteria. Isolated diesel-degrading bacteria belonged to the phyla Actinobacteria and Proteobacteria. Strains of Mycobacterium, Nocardia, Rhodococcus, Intrasporangiaceae, Leifsoni and Arthrobacter possessed alkB and Pseudomonas possessed both ndoB and xylE gene sequences. Two Rhodococcus strains mineralized [1-14C] hexadecane at 5 and −5°C. From the rhizosphere of P. angustata, larger numbers of hydrocarbon-degrading bacteria were isolated than from other plant rhizosphere samples and all three genes alkB (from Actinobacteria), ndoB and xylE (from Pseudomonas) were detected by PCR.


(i) Arctic plants have unique rhizospheric bacterial communities. (ii) P. angustata has potential for phytoremediation research at high Arctic soils. (iii) Isolated bacteria mineralized hydrocarbons at ambient low temperatures.

Significance and Impact of the Study

To the best of our knowledge, this is the first rhizospheric exploration examining the phytoremediation potential of five Arctic plants and evaluating their microbial hydrocarbon-degrading capacities.


Oil spills, caused by shipping accidents, fuel tank and pipeline leaks in addition to inappropriate disposal practices, have polluted Arctic soils and waters (Atlas and Cerniglia 1995; Burgherr 2007). Furthermore, hydrocarbon contamination may increase due to projected increased human activities in northern areas (Midkhatovich and Fenton 2007; Yergeau et al. 2012) augmented by climate change. Phytoremediation is a biological treatment for soil and water based on the use of plants and their rhizospheric micro-organisms to remediate heavy metals and organic pollution (Badri et al. 2009; Wenzel 2009). Phytoremediation of hydrocarbon-contaminated soils relies heavily on rhizospheric degradation, also known as rhizodegradation because the microbial activity occurring in the proximity of the plant roots significantly contributes to the removal of organic pollutants (Gerhardt et al. 2009). Phytoremediation success is driven principally by the interactions and effects of plant species, the diversity and activity of the indigenous soil microbial community and environmental factors that affect hydrocarbon removal (Nie et al. 2011).

Phytoremediation has been assessed in subarctic systems. For example, Palmroth et al. (2002) applied a mixture of white clover (Trifolium repens) and pea (Pisum sativum) to phytoremediate a subarctic soil from Tampere, Finland, removing 68% of the diesel after 30 days in a growth chamber. In Saskatchewan, Canada, over 50% of the total petroleum hydrocarbons (TPH) was removed from a flare pit soil treating it with Altai wild rye (Elymus angustus Trin) and over 40% of the TPH were removed treating it with alfalfa (Medicago sativa L.) and tall wheat grass (Agropyron elongatum) (Phillips et al. 2009). Tolerance to hydrocarbons was determined for some cold-tolerant plant species (Rogers et al. 1996; Robson et al. 2003).

Native plant species inhabiting polluted areas are the most appropriate for in situ phytoremediation (Sangabriel et al. 2006; Lee et al. 2008); yet, there is little information regarding phytoremediation using plants native to Arctic environments. Oil-contaminated polar soils have indigenous hydrocarbon-degrading bacteria, typically Rhodococcus, Sphingomonas or Pseudomonas (Aislabie et al. 2006) at densities ranging from 105 to 106 colony forming units per gram of soil (CFU g−1) (Greer et al. 2010).

In the present study, rhizospheric soil from five different plant species plus four bulk soil samples from the high Arctic were analysed to compare their in situ microbial communities, using the information as screening criteria to detect plant species potentially applicable in phytoremediation treatments. To achieve this goal, we determined microbial abundance and microbial community profiles, detected hydrocarbon-degrading genes in the rhizosphere soils, and isolated and characterized hydrocarbon-degrading bacteria and assessed their mineralization activities.

Materials and methods

Study site description and sampling strategy

Samples were collected near the Eureka high Arctic weather station on Ellesmere Island, Nunavut, Canada (79° 58·800′N and 85° 55·800′W) where a diesel fuel was spilled (approx. 37 000 l) contaminating approx. 3200 m3 of soil in 1990 (Whyte et al. 2001). The Eureka area is classified as a polar desert. During the summer months (June–August), there is little rainfall and the daily average temperatures are typically between +2·3 and +5·7°C, rarely reaching a maximum of 20°C (Canadian Climate Normals 1971–2000; http://climate.weatheroffice.gc.ca). From the beginning of April to the end of August, there is almost continuous sunlight. Virtually, no sunlight occurs between mid-October and late February. Over 9 months every year, the daily average temperature ranges from −7·7 to −38·4°C (Canadian Climate Normals 1971–2000). In August 2004, five different plant species with their rhizospheric soil plus three bulk soil samples (collected 1·5 m from the plant) were aseptically collected from a diesel-impacted area. An additional rhizospheric sample plus a bulk sample were collected from a noncontaminated (pristine) area located c. 500 m away from the spill area. In total, six rhizospheric samples as well as four bulk soil samples were collected and stored individually in sterile plastic bags and transported to the laboratory at 4°C (McGill University, Montreal, Canada). Microbial enumerations were performed starting in September 2004. Portions of the samples were also frozen at −20°C for further analysis.

Physicochemical soil characterization and plant identification

Soil subsamples were oven-dried (60°C for 48 h) and finely ground (<1-mm mesh) prior to soil physicochemical analyses. Soil pH was measured (Hendershot et al. 1993). Soil texture was determined with the hydrometer method described by Sheldrick and Wang (1993). Soil organic carbon and total nitrogen were determined by combustion at 900°C with a Carlo-Erba NC Soils Analyzer (Milan, Italy). TPH (C10–C50) were extracted from 5 g of soil and quantified by GC-MS essentially as described in Centre-d'expertise-en-analyse-environnementale-du-Québec (1997). The six plants were identified based on their morphological characteristics at the Department of Plant Science, McGill University (Montreal, Canada) by Dr Laurie Consaul.

Microbial enumeration

To estimate the abundance of micro-organisms present in the soil samples, microscopic counts of microbial cells were carried out using 5-(4,6-dichlorotriazinyl) aminofluorescein (DTAF) essentially as described by Kepner and Pratt (1994) but adapted to 1 g of soil, counting cells from 102 and 103 dilutions filtered onto 25-mm-diameter black polycarbonate 0·22 μm pore filters (Osmonics Inc., Minnetonka, MN) with an epifluorescence microscope (Eclipse E600W; Nikon, Melville, NY). An average cell count from twenty counted fields per filter is reported. Aerobic heterotrophic bacteria and aerobic hydrocarbon-degrading bacteria were enumerated by the spread plate technique in R2A agar media (Becton, Dickson and Co., Mississauga, Canada) and in Rennie media (Rennie 1981) with diesel as the sole carbon source (modified by Rivera-Cruz et al. (2004)) by placing a filter square of Whatman (1·5 cm2) impregnated with approx. 200 μl of diesel onto the inner lid of the Petri plate. The first serial dilution consisted of 5 g of soil sample mixed with 15 ml of 0·1% w/v cold sodium pyrophosphate (Na4P2O7 10·H2O, pH 7·0). Selected dilutions were spread onto triplicate Petri plates and then incubated at 24°C for 14 days or at 5°C for 28 days. Averages of the CFU g−1 soil were calculated. The analysis of variance (anova) of microscopic and spread plate counts was determined using sas/stat software, ver. 9.2 (SAS Institute Inc., Cary, NC, USA).

Total community DNA extraction

Total DNA was extracted in duplicate from one gram of soil per sample via the MoBio Laboratories Ultraclean Soil DNA Kit (Carlsbad, CA, USA), following manufacturer′s instructions with a reduction in the bead-beating time to 2 min and the use of the alternative lysis protocol. A polyvinylpolypyrrolidone (PVPP) spin column filtration step was included (Berthelet et al., 1996) to eliminate PCR inhibitors, such as hydrocarbons, humic and fulvic acids. DNA was quantified with a Nano Drop Spectrophotometer ND-1000.

PCR amplification of bacterial 16S rRNA gene

General bacteria primers 341(F) 5′-CCTACGGGAGGCAGCAG-3′ and 758(R) 5′-CTACCAGGGTATCTAATCC-3′ (MWG Operon, Huntsville, AL, USA) were used in the PCR amplification of partial 16S rRNA genes for phylogenetic analysis, and primer 341F with GC clamp (5′-GCGGGCGGGGCGGGGGCACGGGGGGCGCGGCGGGCGGGGCGGGGGCCTACGGGAGGCAGCAG-3′) and 758(R) were used to amplify fragments of 16S rRNA genes for denaturing gradient gel electrophoresis (DGGE) analysis. PCR reactions (25 μl) contained 0·5 mmol l−1 of primers 341F and 758R, 1× PCR buffer and 1·5 mmol l−1 of MgCl2, 0·2 mmol l−1 of each deoxynucleoside triphosphate, 0·6 μl of 10 mg ml−1 bovine serum albumin, 1 U of Taq polymerase (Invitrogen, Life Technlogies, Burlington, Canada) plus 2–10 ng of DNA. Thermocycling conditions consisted of an initial denaturing 96°C for 5 min, 10 cycles of 96°C for 1 min, touchdown annealing temperature of 60°C for 45 s (decreasing 1°C each cycle to 55°C), extension 72°C for 1·5 min, and then, 15 cycles of 96°C for 1 min, 55°C for 45 s, 72°C for 1·5 min and a final extension of 72°C for 5 min.

PCR amplification of alkB, ndoB and xylE genes

Three catabolic genes from the aerobic hydrocarbon degradation pathway were analysed: alkB (encoding the alkane hydroxylase), ndoB (encoding the α-subunit of the iron sulfur protein of naphthalene dioxygenase) and xylE (encoding the 2-3-catechol dioxygenase) (Whyte et al. 2001; Luz et al. 2004). The primers used to amplify gene fragments of alkB (approx. 550 bp), ndoB (approx. 642 bp) and xylE (approx. 834 bp), respectively, were alkB(F) 5′-CIGIICACGAIITIGGICACAAGAAGG-3′ and alkB(R) 5′-IICGITGITGATCIIIGTGICGCTGIAG-3′; ndoB(F) 5′-CACTCATGATAGCCTGATTCCTGCCCCCGGCG-3′ and ndoB(R) 5′-CCGTCCCACAACACACCCATGCCGCTGCCG-3′; xylE(F) 5′-GTGCAGCTGCGTGTACTGGACATGAGCAAG-3′ and xylE(R) 5′-GCCCAGCTGGTCGGTGGTCCAGGTCACCGG-3′. PCR reagents were prepared as previously described (testing the direct DNA extracts as well as 1 : 5 and 1 : 10 dilutions). The following thermocycling conditions were applied: 96°C for 5 min, 10 cycles of 96°C for 1 min, touchdown annealing temperature starting at 58°C for 1 min finishing at 53°C for 1 min and 72°C for 1·5 min, followed by 20 cycles of 94°C for 1 min, 55°C for 1 min, 72°C for 1·5 min and a final extension of 72°C for 10 min. The strains Rhodococcus sp. Q15 (Whyte et al. 2002b), Pseudomonas putida 17484 and Pseudomonas putida 33015 (Whyte et al. 1996) were the positive controls for the amplification of alkB, ndoB and xylE, respectively. The PCR-amplified DNA fragments were visualized by gel electrophoresis (0·8% agarose TAE gels, ran at 85 V for 45 min) with ethidium bromide staining (Sambrook and Russell 2001) observed on a Bio Imaging System (Syngene, Frederick, MD, USA).

Phylogenetic analyses of gene sequences

The Nucleotide Basic Local Alignment Search Tool (BlastN) (Altschul et al. 1990) and the EzTaxon-e server (http://eztaxon-e.ezbiocloud.net/; Kim et al. 2012) were used to compare the 16S rRNA gene sequences using GenBank database. alkB, ndoB and xylE gene sequences were compared against GenBank databases via BlastN (Altschul et al. 1990; Zhang et al. 2000) and TBlastX (Altschul et al. 1997).

Denaturing gradient gel electrophoresis

Approximately 1000 ng of DNA (PCR amplicons) was loaded per lane. DGGE was performed according to Steven et al. (2007) on an 8% polyacrylamide/bis-acrylamide gel with 45–65% denaturant gradient. Electrophoresis was performed at 60°C and 80 V for 16 h. Gel banding patterns were visualized by staining for 35 min with 1 : 10 000 (v/v) Vistra Green (GE Lifesciences, Quebec, Canada) and destained with 1× TAE buffer for 20 min, and then observed on a Bio Imaging System (Syngene). DGGE gel images were analysed with GelCompar II software, and the neighbour-joining algorithm was used to construct dendrograms. Band numbers (N), intensity and distribution in the DGGE gels were used to estimate the Range-weighted richness (Rr) as well as the Functional organization (Fo) following the procedure shown by Marzorati et al. (2008).

Isolation and characterization of hydrocarbon-degrading bacteria

Representative colonies isolated from Rennie modified Petri plates (from the 103 dilution) were differentiated based on colony morphology, pigmentation and time of colony appearance. These colonies were then restreaked onto new Petri plates with Rennie modified and R2A media supplemented with diesel and then incubated at 24 and 5°C to obtain pure colonies. Genomic DNA of each isolated strain was obtained by boiling lysis (Sambrook and Russell 2001), and a 400-bp fragment of the 16S rRNA gene was PCR amplified and sequenced as described previously. Colonies having aerobic catabolism of hydrocarbons were subsequently screened by PCR amplifying the alkB, ndoB and xylE genes as described above. Resulting amplicons were sequenced at Genome Québec Innovation Centre (McGill University). The ability of selected strains to mineralize 14C-labelled hexadecane at 5 and −5°C was determined as described by Whyte et al. (1996). Briefly, for each tested strain, three 30 ml serum bottles (Supelco; Sigma-Aldrich, Ontario, Canada) containing 10 ml of solid MSM (Greer et al. 1990) combined with 50 ppm of yeast extract and 5% NaCl were supplemented with 100 000 dpm [1-14C] hexadecane (specific activity, 59 mCi/mmol) (Sigma, St Louis, MO, USA), plus nonradioactive hexadecane [100 μLg−1]; a CO2 trap consisting of a 1 ml glass tube filled with 0·5 ml of 1 mol l−1 KOH. The amount of trapped 14CO2 was determined by liquid scintillation spectrometry (LS6500 Multipurpose Scintillation Counter; Beckman Coulter, Mississauga, Canada) after retrieving the KOH solution and mixing it with 20 ml Scintiverse BD Cocktail (Fisher Scientific, Ottawa, Canada). Microcosms were incubated in the dark at 5 and −5°C, and 14CO2 evolution was monitored for a total of 93 days.

Nucleotide sequence accession numbers

The 16S rRNA gene sequences obtained in this study were deposited in GenBank database under the following accession numbers: HQ654251HQ654260 corresponding to DGGE bands, and from JF339990 to JF340035 corresponding to isolated strains. The alkB, ndoB and xylE gene sequences have the accession numbers from JF520628 to JF520637.


Plant species and physicochemical characteristics of soil samples

Collected plants (Fig. S1) were identified as Eriophorum scheuchzeri Hoppe (Arctic cotton), Potentilla rubricaulis Lehm (Cinquefoil), Oxyria digyna L. Hill (Mountain sorrel), Salix arctica Pall (Arctic willow), Puccinellia angustata (R. Br.) Rand and Redfield (Narrow alkali grass). The specimen collected from the noncontaminated area was also identified as P. angustata.

Soil physicochemical characteristics are shown in Table 1. In general, rhizospheric samples had higher percentages of total nitrogen and total carbon than bulk samples. C/N ratios were between 10 and 20 with the exception of V-Pr (1·36). Generally, rhizospheric samples tended to have more moisture (4–41·33%) than bulk samples (2–10%). TPH were below 100 mg kg−1 in all tested samples.

Table 1. Characteristics of soil samples collected in the summer of 2004
Sample IDSample characteristicsTotal N (%)Total C (%)C/NMoisture (%)
  1. Total petroleum hydrocarbons were below 100 mg kg−1 in all samples.

V-EsEriophorum scheuchzeri rhizospheric soil contaminated with diesel drainage pond0·363·9110·8641·33
V-PrPotentilla cf. rubricaulis Lehm rhizospheric soil contaminated with diesel0·741·011·366·67
V-OdOxyria digyna L. Hill rhizospheric soil contaminated with diesel0·182·1212·0923·33
VP-PaPuccinellia angustata rhizospheric soil ‘pristine’0·182·6715·224·00
V-PaP. angustata rhizospheric soil contaminated with diesel0·101·9519·928·00
V-SaSalix arctica rhizospheric soil contaminated with diesel0·131·7313·198·67
BaBulk soil from slope, contaminated with diesel0·060·6611·363·33
BbBulk soil, contaminated with diesel close to P. rubricaulis0·030·6620·702·00
BcBulk soil, contaminated with diesel close to Salix arctica0·050·5711·392·67
BPBulk soil ‘pristine’0·142·3116·7310·00

Microbial enumeration

Microscopic counts and colony forming units are shown in Table 2. Microscopic counts of rhizospheric samples were numerically higher and significantly different (P ≤ 0·05) from those of bulk samples. The highest microscopic bacterial count (5·9 × 109 cell g−1) was detected in rhizospheric sample V-Es and the lowest (1·5 × 109 cell g−1) in bulk sample Bb. CFUs of heterotrophic and hydrocarbon-degrading bacteria obtained in cultures at 24°C were not significantly different (P ≤ 0·05) from those obtained at 5°C. Numbers of cultivable heterotrophic and diesel-degrading bacteria were one or two orders of magnitude higher in rhizospheric than in bulk samples at both incubation temperatures (24 and 5°C), and differences were statistically significant. Puccinellia angustata rhizosphere from the pristine site had statistically significant (P ≤ 0·05) higher populations of heterotrophic (1·8 × 108 CFU g−1 at 24°C and 6·9 × 107 CFU g−1 at 5°C) and diesel-degrading bacteria (7 × 107 CFU g−1 at 24°C) than the other samples.

Table 2. Microbial enumeration
Sample IDMicroscopic DTAF (107 cell g−1)24°C5°C
Heterotrophic (107 CFU g−1 soil)Diesel degraders (107 CFU g−1 soil)Heterotrophic (107 CFU g−1 soil)Diesel degraders (107 CFU g−1 soil)
  1. CFU g−1, colony forming units per gram of dry soil.

  2. Data are presented as means (n = 6) ±SD. For each column, means with super indices of different letters are significantly different at P ≤ 0·05; means shearing super indices of the same letters are not significantly different according to Tukey test results.

V-Es590 ± 90a6·6 ± 0·39bb1·8 ± 0·14c5·7 ± 0·77b0·96 ± 0·01ab
V-Pr280 ± 30e4·6 ± 0·75c1·8 ± 0·05c2·4 ± 0·27d1·1 ± 0·09a
V-Od320 ± 64d7·2 ± 0·46b3·6 ± 0·2b2·7 ± 0·47c0·74 ± 0·02b
VP-Pa430 ± 96b18 ± 0·19a7·0 ± 0·13a6·9 ± 0·01a1·3 ± 0·11a
V-Pa380 ± 30c2·6 ± 0·13d1·5 ± 0·11c2·6 ± 0·13c1·1 ± 0·14a
V-Sa310 ± 79d1·1 ± 0·09e0·27 ± 0·02d0·15 ± 0·04d0·06 ± 0·001d
Ba190 ± 63f0·27 ± 0·01fg0·14 ± 0·03de1·5 ± 0·11d0·04 ± 0·006de
Bb150 ± 63g0·25 ± 0·01fg0·09 ± 0·01e0·23 ± 0·10d0·02 ± 0·01de
Bc160 ± 71g0·8 ± 0·03ef0·20 ± 0·01d0·5 ± 0·02d0·2 ± 0·02c
BP180 ± 77fg0·09 ± 0·01g0·06 ± 0·01e0·05 ± 0·007d0·02 ± 0·008e

16S rRNA gene DGGE fingerprints and phylogenetic analysis of soil bacterial communities

Figure 1a shows the bacterial 16S rRNA gene DGGE fingerprints of the rhizosphere and bulk Arctic soil samples. The bands number, intensity and position revealed a unique structure of bacterial communities in the rhizospheres of E. scheuchzeri (Arctic cotton), P. rubricaulis (Cinquefoil), O. digyna (Mountain sorrel), S. arctica (Arctic willow) and P. angustata (Narrow alkali grass), as well as in the bulk soils collected from the pristine and diesel-contaminated sites. Figure S2 illustrates the Rr and Fo calculated from each sample DGGE fingerprint. Although rhizospheric samples had an average Rr value of 22, compared with an average Rr value of 18 from bulk samples, the Rr values of the rhizosphere and bulk soils (either diesel contaminated or pristine) were above 10 and below 30, corresponding to a medium range-weighted richness and reflecting the carrying capacity of healthy but not exuberant environments. The Fo values of the samples ranged from 37 to 46%. A dendrogram based on the bacterial DGGE profiles is presented in Fig. 1b. A high significance (>98% boot strap values) of clustering occurred, forming three major branches: two branches grouped mainly rhizospheric samples and another branch clustered bulk samples. According to these analyses, bacterial communities from the rhizosphere of P. rubricaulis (V-Pr) were more similar to communities from E. scheuchzeri (V-Es) and O. digyna than to bulk soil (Bb) communities. Rhizospheric communities from P. angustata in the pristine soil were found to cluster with the same plant rhizosphere from the diesel-contaminated soil rather than with the pristine bulk soil (BP).

Figure 1.

DGGE of 16S rRNA gene fragments of rhizospheric and bulk Arctic soils. (a) Image of DGGE gel (45–65% denaturant gradient); labels at the top refer to rhizospheric samples: [V-Es] Eriophorum scheuchzeri, [V-Pr] Potentilla rubricaulis, [V-Od] Oxyria digyna, [VP-Pa] Puccinellia angustata (pristine), [V-Pa] P. angustata (polluted) and [V-Sa] Salix arctica; bulk samples: [Ba] from slope, [Bb] nearby sample [V-Pr], [Bc] nearby sample [V-Sa], [BP] pristine. Excised bands: (1) VEs2 (HQ654251), (2) Ba11 (HQ654252), (3) VPr13 (HQ654253), (4) VOd15 (HQ654254), (5) VOd16 (HQ654255), (6) VPa31 (HQ654256), (7) VPa32 (HQ654257), (8) VSa37 (HQ654258), (9) BP18 (HQ654259), (10) Bb14 (HQ654260). (b) Dendrogram analysis of DGGE image (constructed with Gelcompare II using the neighbour-joining method). Numbers on the nodes are the bootstrap values based on 1000 replicates. The scale bar indicates the estimated number of base changes per nucleotide sequence position.

Phylogenetic analysis of partial 16S rRNA gene sequences from selected DGGE bands revealed that Arctic soil bacteria were closely related to bacteria from the following phyla: Bacteroidetes, Firmicutes, Actinobacteria and Proteobacteria and specifically to Belliella, Flavobacterium, Fusibacter, Arthrobacter, Rhodococcus, Sphingomonas and Pseudomonas. The most closely related bacteria were retrieved from contaminated soil, agricultural soil, sediments, permafrost, ice or water from cold or frozen environments and/or from environments polluted with organic compounds in different regions (Table 3).

Table 3. Nucleotide sequence similarity of 16S rRNA gene DGGE bands
NoBand (accession number)SampleClosest sequence match (accession number)aSimilarity (%)Environment of match sequence
  1. a

    Superscript indicates the database where the closest sequence match was obtained: 1, EzTaxon-e; 2, BLAST.

1VEs2 (HQ654251)V-Es1Fusibacter tunisiensis strain BELH1(T) (FR851323)96·6Anaerobic reactor treating phosphogypsum and olive mill wastewaters in Tunisia
2Uncultured bacterium clone LVBR10aG04 (GQ167327). Fusibacter (genus)99Brine of an Ice-Sealed Antarctic Lake
2Ba11 (HQ654252)Ba1Sphingomonas jaspsi strain TDMA-16(T) (AB264131)98·4Carotenoid-producing bacteria isolated from Misasa, a radioactive site in Japan
2Uncultured Kaistobacter sp. strain AMPA9 (AM935082). Sphingomonas (genus)100Pilot-scale bioremediation process of a hydrocarbon-contaminated soil
3VPr13 (HQ654253)V-Pr1Pseudomonas mandelii strain CIP 105273(T) (AF058286)100Natural mineral waters
2Pseudomonas sp. strain J35(2008) (EU375656) Pseudomonas (genus)99Bacteria degrading various organic pollutants isolated from agricultural soil
4VOd15 (HQ654254)V-Od1,2Flavobacterium limicola strain ST-82(T) (AB075230). Flavobacterium (genus)99·51Organic-polymer-degrading bacterium isolated from freshwater sediments
5VOd16 (HQ654255)V-Od1Stakelama pacifica strain JLT832(T) (EU581829)98·4Pacific Ocean
2Uncultured bacterium clone AC14C2BF12 (JQ428474). Sphingomonas (genus)100Alkaline saline soil from Mexico spiked with anthracene
6VPa31 (HQ654256)V-Pa1Arthrobacter phenanthrenivorans strain Sphe3(T) (CP002379)99·67Phenanthrene-degrading bacterium, isolated from a creosote-contaminated soil in Greece
2Arthrobacter sp. strain KFC-94 (EF459530). Arthrobacter (genus)99Soil sample from Kafni Glacier in the Himalayas
7VPa32 (HQ654257)V-Pa1Sideroxydans lithotrophicus strain ES-1 (CP001965)94·97Iron-contaminated groundwater in Michigan, USA
2Uncultured bacterium clone A50Sp-15 (AJ965911); Beta-proteobacteria (class)99Alpine lakes
8VSa37 (HQ654258)V-Sa1Demequina oxidasica strain YM05-1041(T) (AB522640)86·91Marine environments
2Unidentified bacterium clone CN-1_SL2_G08 (EF219906); Actinobacteria (class)98Antarctic terrestrial habitats
9BP18 (HQ654259)PB1Belliella baltica strain BA134(T) (AJ564643)94·55Surface water of the central Baltic Sea
2Uncultured bacterium clone 33MIC111 (JF341252) Cyclobacteriaceae (family)100Concrete sewer biofilm
10Bb14 (HQ654260)Bb1Rhodococcus qingshengii strain djl-6(T) (DQ090961)97·39Carbendazim-degrading bacterium isolated from carbendazim-contaminated soil from Jiangsu province
2Uncultured bacterium clone Eur3BacAL.33 (EU218663).Rhodococcus (genus)99Permafrost/ground ice core profile from the Canadian high Arctic

alkB, ndoB and xylE genes

alkB was more broadly distributed among rhizospheric and bulk samples (5 of 10) than ndoB (2 of 10 samples) and xylE (1 of 10 samples). More specifically, alkB was consistently amplified in the rhizosphere of E. scheuchzeri (V-Es), P. rubricaulis (V-Pr) and P. angustata (VP-Pa and V-Pa) as well as in one bulk sample (Bc). Puccinellia angustata rhizosphere from contaminated soil was the only sample where alkB, ndoB and xylE were detected in all PCR amplifications. Similarly both alkB and ndoB were also detected in the P. angustata rhizosphere from pristine soil.

Hydrocarbon-degrading bacterial strains

Forty-six putative diesel-degrading bacteria were isolated from both rhizosphere and bulk samples in N-free Rennie medium Petri plates with diesel as the sole carbon source (Fig. 2). Genomic DNA of 15 representative strains, being frequently detected in rhizosphere samples but having different colony morphologies, was screened for the presence of alkB, ndoB and xylE genes. The alkB gene was PCR amplified in eight isolates from diverse genera: Mycobacterium, Nocardia, Rhodococcus, Intrasporangiaceae, Leifsonia and Arthrobacter. Both ndoB and xylE genes were identified in Pseudomonas strain 5K-VPa. The rhizosphere of P. angustata had colonies with the same morphologic characteristics (size, colour, borders, etc.) of the strains listed in Table 4.

Figure 2.

Phylotype composition of hydrocarbon-degrading isolates from Eureka samples. Determined by sequencing a fragment of the 16S rRNA gene (100% corresponds to 46 strains isolated from modified Rennie media Petri plates).

Table 4. Closest matches for the alkB, ndoB and xylE genes from the isolated strains
Strain (accession number)alkB, ndoB or xylE Gene accession numberStrain classification based on 16S rDNA genePositive for geneClosest match from GenBank hydrocarbon degradation-related geneSimilarityEnvironment of origin from GenBank sequence
alkB ndoB xylE nt %a.a %
  1. nt, nucleotide, a.a, amino acid.

1·1-VEs (JF340027) JF520628 Mycobacterium +  Alkane rubredoxin-dependent monooxygenase of clone alkB21mpn_ingol (GU184266)8486Hydrocarbon-contaminated German soils
1·3-VEs (JF340004) JF520629 Rhodococcus +  alkB gene fragment of Rhodococcus sp. strain 11/8p (DQ376002)9392Antarctic
1·5-VEs (JF340017) JF520630 Leifsonia +  alkB gene fragment of Rhodococcus sp. strain 11/8p (DQ376002)9594Antarctic
1·12-VEs (JF340026) JF520631 Mycobacterium +  Alkane rubredoxin-dependent monooxygenase of clone alkB21mpn_ingol (GU184266)8587Hydrocarbon-contaminated German soils
3·2-VPr (JF340009) JF520632 Nocardia +  Putative alkane hydroxylase fromstrain alkW34 (DQ287995)9095German grassland soil
3·3-VPr (JF340019) JF520633 Rhodococcus +  alkB gene from Rhodococcus sp. strain H1 (FJ435353)9398Shorelines after the Prestige oil spill in Spain
7·19-VPa (JF339999) JF520634 Arthrobacter +  Hydroxylase rubredoxin of Nocardioides sp. strain CF8 (AF350429)9397Alkane-utilizing bacterium
7·31-VPa (JF340005) JF520635 Unclassified Intrasporangiaceae+  Putative alkane monooxygenase from clone alkW2-3 (DQ288027)8689German grassland soil
5K-VPa (JF339990) JF520636 Pseudomonas  + Fragment from the plasmid pNAH20 (AY887963.3)100100River polluted with phenolic compounds
JF520637    +Fragment from the plasmid pNAH20 (AY887963.3)100100River polluted with phenolic compounds

Table 4 shows the closest GenBank matches to the alkB, ndoB and xylE sequences from the isolated strains. The alkB genes from strains 1·1-VEs (JF340027), 1·12-VEs (JF340026) and 7·31-VPa (JF340005) did not have a high similarity with previously sequenced alkB genes and may represent novel alkane monooxygenases. The mineralization assay showed that Rhodococcus sp. strain 1·3-VEs transformed 18·6% of [1-14C] hexadecane into 14CO2 at 5°C and 1·8% at −5°C after 93 days of incubation (data not shown); similarly, Rhodococcus sp. strain 3·3-VPr mineralized 19·3% of [1-14C] hexadecane incubated at 5°C and 1·2% at −5°C for the same period of time (data not shown). Hexadecane mineralization in sterile control microcosms was <1% under all incubation conditions.


In this study, we examined the soil microbial communities inhabiting Arctic rhizospheric soils. The physicochemical characteristics of the soils were similar to those from previous studies of high Arctic soils (González et al. 2000), for example, low carbon (<5%) and nitrogen (<1%) contents. Extreme environments such as the Arctic represent a challenging habitat due to the prevailing cold temperatures reducing biochemical reaction rates, cell membrane fluidity and nutrient availability; however, micro-organisms and plants have adapted to survive in these conditions (Margesin et al. 2007a,b). Plants such as E. scheuchzeri, P. rubricaulis, O. digyna, S. arctica and P. angustata belonging to the Cyperaceae, Rosaceae, Polygonaceae, Salicaceae and Poaceae families inhabit sandy or silty soils with low organic content or soils with oxygen depletion associated with soil flooding like in imperfectly drained areas and margins of ponds (Aiken et al. 1996). E. scheuchzeri, P. rubricaulis, O. digyna, S. arctica and P. angustata are widely distributed in the Canadian Arctic Archipelago including Baffin, Devon, Ellesmere, Axel Heiberg, Parry, Cornwallis, Banks and Victoria Islands (Aiken et al. 1996). It may be that the same mechanism allowing them to grow in soils with limited oxygen exchange also allows them to grow on hydrocarbon-contaminated oily soils. Previous studies support the fact that the presence of plants modifies soil microbial populations in relation to relative abundances (Rivera-Cruz et al. 2004), community structure (Siciliano et al. 2003), relative abundance of catabolic genes (Da Silva et al. 2006) and hydrocarbon degradation activities (Kim et al. 2006; Lee et al. 2008).

The plants stimulated microbial communities strongly and permanently enough to produce a significant increase in microbial abundance detectable by direct microscopic counts (109) and by cultivable heterotrophic bacteria (107). There was no significant difference between viable plate counts incubated at 24 or 5°C indicating that the cultivable heterotrophic bacteria were primarily psychrotolerant rather than psychrophilic, similar to previous studies on Arctic soil communities (Whyte et al. 1996, 2001). Samples were collected 14 years after a diesel contamination event and were preserved frozen for months prior to laboratory analyses. Therefore, results obtained during this study revealed a temporal rhizospheric effect over soil microbial community and/or a residual effect of a diesel contamination event. Plant presence not only modified bacterial numbers but also modified the community structure as shown by dendrogram analyses of DGGE banding patterns. It has been recognized that niche-based processes and contingencies are likely to play an important role in determining patterns of microbial species richness and that the taxonomic diversity of a microbial community partially reflects its metabolic diversity (Fierer and Lennon 2011).

Community profiles from samples from environments sharing characteristics with our samples had Rr values intermediate between previously studied oil-contaminated soil (Rr value of 8·1) and garden soil (Rr of 220) and were similar to those of a legume rhizosphere (Rr of 78), and an Arctic sea ice sample had an Rr of 26 (Marzorati et al. 2008). In our study, plants had a stronger impact on microbial community composition than diesel contamination; however, Rr values ranged between 10 and 30, indicating that neither the presence of plants nor hydrocarbons appeared to drastically affect the amount of bacteria, which the Arctic soil is naturally able to sustain. Furthermore, the Fo values (37–48) of rhizosphere and bulk samples indicated that microbial communities had an adequate distribution of dominant and resilient bacteria, where 20% of the total diversity corresponded to a range from 37 to 48% of the total bacteria in the soil. This kind of functional organization allows the microbial populations to overcome adverse conditions such as drastic environmental changes or strong contamination events (Fernandez et al. 2000).

The phylogeny of 16S rRNA gene fragments further supports Rr and Fo findings, as four different phyla (Bacteroidetes, Firmicutes, Actinobacteria and Proteobacteria) were detected by sequencing only 10 DGGE bands randomly selected from all 10 samples. Therefore, the results of the present work are not an exhaustive description of the bacterial diversity or community structure from this environment but are an insight to the bacterial community structure and composition of Eureka high Arctic soils influenced by different plant species and diesel contamination. Nevertheless, the community composition detected in this study showed bacterial genera similar to those reported in studies of frozen, cold or polluted environments.

The potential to biodegrade hydrocarbons in these soil samples was determined by both culture-independent and culture-dependent methods. The gene encoding an alkane hydroxylase (alkB) was detected in four of six rhizospheric samples but only in one of four bulk samples. Moreover, genes associated with the synthesis of naphthalene dioxygenase (ndoB) (Kurkela et al. 1988) and 2-3-catechol dioxygenase (xylE) ( et al. 1991) were only amplified in P. angustata rhizospheric samples. The fact that genes associated with alkane oxidation were detected more frequently than genes related to aromatic oxidation could be due to larger availability of alkanes due to the diesel composition (c. 75% alkanes and 25% aromatic hydrocarbons) or simply by a methodological bias. Prevalence of various genes involved in the catabolism of hydrocarbons is modified by the presence of plants, and it can either increase or decrease depending on the plant species (Siciliano et al. 2003). Also, the prevalence of alkB and ndoB was found to increase in contaminated soils as compared with pristine Eureka soils (Whyte et al. 2002b). The proportion of alkB containing micro-organisms was correlated to the concentration of n-alkanes in a contaminated Antarctic soil (Powell et al. 2006).

The potential to biodegrade hydrocarbons in these soil samples was determined by both culture-independent and culture-dependent methods. The gene encoding an alkane hydroxylase (alkB) was detected in four of six rhizospheric samples but only in one of four bulk samples. Moreover, genes associated with the synthesis of naphthalene dioxygenase (ndoB) (Kurkela et al. 1988) and 2-3-catechol dioxygenase (xylE) (Benjamin et al. 1991) were only amplified in P. angustata rhizospheric samples. The fact that genes associated with alkane oxidation were detected more frequently than genes related to aromatic oxidation could be due to larger availability of alkanes due to the diesel composition (c. 75% alkanes and 25% aromatic hydrocarbons) or simply by a methodological bias. Prevalence of various genes involved in the catabolism of hydrocarbons is modified by the presence of plants, and it can either increase or decrease depending on the plant species (Siciliano et al. 2003). Also, the prevalence of alkB and ndoB was found to increase in contaminated soils as compared with pristine Eureka soils (Whyte et al. 2002b). The proportion of alkB containing micro-organisms was correlated to the concentration of n-alkanes in a contaminated Antarctic soil (Powell et al. 2006).

The number of bacteria forming colonies on culturing media is a parameter with ecological relevance even though it does not represent the real number of active bacteria in the original environment (Nichols 2007). Based on the plate count method, cultivable diesel-degrading bacteria were approx. 10-fold greater in rhizospheric (107) than in bulk (106) soils. Pristine bulk soil samples had the lowest values, being one order of magnitude lower (105) than diesel-contaminated samples (106). Among rhizospheric samples, those from P. angustata had the highest numbers of diesel-degrading bacteria (7·0 ± 0·13 × 107 at 24°C and 1·3 ± 0·11 × 107 at 5°C). Other studies show that hydrocarbon contamination (Aislabie et al. 2001), plant species presence (Liste and Prutz 2006) and nutrient addition (Margesin et al. 2007a) increase the abundance of hydrocarbon-degrading bacteria in soils.

Hydrocarbon-degrading isolates were mainly classified into Actinobacteria (mostly Arthrobacter, Rhodococcus and Sanguibacter) and Proteobacteria (mostly Pseudomonas) as shown in Fig. 2. Similarly, Yergeau et al. (2012) in a metagenomic analysis of diesel-contaminated Alert soils (Canadian Arctic) treated in biopiles (bioremediation technology in which excavated soils are mixed with soil amendments) determined that the Proteobacteria Pseudomonas, Caulobacter and Sphingomonas and the Actinobacteria Rhodococcus were enriched. xylE and ndoB were only detected in Pseudomonas strain (5K-VPa), and their nucleotide and amino acid sequences were closely related to sequences of the pNAH20 plasmid from Pseudomonas fluorescens strain PC20 (Merimaa et al. 2006).

The bulk sample Bb was negative for the PCR amplification of alkB, but the sequence of band (Bb14) isolated from it was 95% similar to the 16S rRNA gene fragment from the hexadecane degrading isolated strains 1·3-VEs and 3·3-VPr, which were both alkB positive, suggesting that the alkB gene of 1·3-VEs and 3·3-VPr may be in a plasmid and its copy number may have increased in the rhizospheric samples. These two Rhodococcus sp. (strains 1·3-VEs and 3·3-VPr) were shown to mineralize hexadecane at low (5°C) and subzero (−5°C) temperatures, suggesting that there may be slow microbial degradation activity in the soil even at subzero temperatures.

Our results suggest that high G+C Gram-positive bacteria are principally responsible for the biodegradation of alkanes as alkB was only detected in Actinobacteria, and their alkB nucleotide and amino acid sequences were closely related to alkB sequences from Rhodococcus sp. and Nocardioides sp. strains. These results may indicate that Actinobacteria are the predominant n-alkane degraders in the studied soils, which is consistent with a previous study (Whyte et al. 2002b) showing a higher prevalence of alkB genes from Rhodococcus spp. than Pseudomonas putida in high Arctic diesel-contaminated soils. Yergeau et al. (2012) reported fluctuations in the expression of Pseudomonas alkB (higher in the first month of treatment) and Rhodococcus alkB (higher after a year of treatment). Our results may be biased not only by the selection of strains tested for alkB, ndoB and xylE PCR amplification but also by the culturing media, which could be selective for the growth of Actinobacteria and Proteobacteria and by the primers used, which could be more efficient to amplify sequences from bacteria closely related to the reference strains, mainly Proteobacteria and Actinobacteria.

The present study is an insight to the bacterial community structure and composition of Eureka high Arctic soils influenced by different plant species and diesel contamination. E. scheuchzeri, P. rubricaulis, O. digyna, S. arctica and P. angustata not only modified bacterial abundance but also the bacterial community structure of high Arctic soils. The rhizosphere of P. angustata had the highest abundance of hydrocarbon-degrading bacteria and seems to have the highest prevalence of genes encoding hydrocarbon oxidizing enzymes from all samples. Therefore, it appears that P. angustata has potential application in phytoremediation treatments for high Arctic hydrocarbon-contaminated areas. Our research group is further investigating the effect of P. angustata on the microbial communities and the rates of hydrocarbon removal in different soils from the Canadian high Arctic.


During this research, the first author was supported by grant 204563 from Consejo Nacional de Ciencia y Tecnología de México (CONACYT, México). The authors of this paper have no conflict of interest to declare. We want to thank Dr Joann Whalen (Department of Soil Science, McGill University, Macdonald Campus) for the physicochemical analyses of soil samples and Diane Labbé (NRC, Montreal, Canada) for the determination of TPH. Dr Blaire Stevens, Dr Thomas Daniel Niederberger, Roland Wilhelm, Dr David Meek, Marianne and Claire-Louise Michelle Poilly who provided important feedback to the present research and manuscript edition.