Microbial Community Structures in Petroleum Contaminated Soils at an Oil Field, Hebei, China

1 Introduction

Petroleum contamination, a widespread environmental problem in soils, sediments, and groundwater, which is mostly caused by anthropogenic activities, leakage of underground storage tanks and accidental spills during extraction, transportation, maintenance, and storage, occurs quite often and has been causing ecological impacts 1, 2. Petroleum hydrocarbons that consist of complicated mixtures of non-aqueous and hydrophobic components such as n-alkanes, aromatics, resins and asphaltenes are now raising a huge public concern, and threatening the soil ecosystems due to their toxicity, mutagenicity and carcinogenicity 3, 4.

Bioremediation of petroleum and derivatives in contaminated soils has been claimed to be the novel, environmental and cost-effective alternative compared to physicochemical treatment 5, 6. The efficiency of remediation highly depends on the degradation ability of the microorganisms 6, 7, the environmental conditions and the chemical structure of the petroleum hydrocarbons 8. Soil is considered to be the most microbially diverse environment on earth and 1 g of soil can contain up to 10⁹ microbial cells, representing more than 10 000 genomes 9, 10. Oil-contaminated soil contains microbes that are capable of utilizing hydrocarbons as carbon source under both aerobic and anaerobic conditions 6. Traditional cultivation methods have been used to isolate bacteria involved in the degradation of petroleum hydrocarbons, but only the minority of petroleum degrading microorganisms in soil can be cultivated and a large number of microbes in oil-contaminated ecosystems remain unknown 11-14.

Several studies have focused on the analysis of microbial communities in the petroleum hydrocarbons or organic compounds contaminated soils, water, and sediments using DNA based technologies 1, 2, 14-16. At an oil-contaminated lagoon in France, sediment bacterial community structures were obviously associated with the gradient of oil contamination 1. A selection pressure was forced on these bacterial communities by the complex pollution constitutes, and oil degrading bacteria did not increase as oil contaminants increased. Conversely, in a study of microbial communities at an oil-contaminated site in China, more known hydrocarbon-degrading bacteria (e.g., Alcanivorax spp., Marinobacter spp., and Halomonas spp.) occurred in the contaminated soil as compared to the pristine soil 15. Although the culture-independent molecular studies provide a better understanding of microbial community composition than pure culture ones 16, how the structure varies under complex petroleum contamination gradients is still not fully unveiled, particularly in the subsurface soil ecosystems.

The bacterial diversities and communities of soil samples with different petroleum contaminants located at an oil field in Hebei, China, were investigated. After phylogenetic analysis of 16S rRNA gene sequences, taxonomic groups (e.g., phylum, family, and genus) of each sample exerted a huge difference. Multivariate statistical analysis between microbial communities and petroleum contaminated soil properties was conducted. These results may provide insights into the distribution of microbial communities with potential degrading capacities under complex petroleum hydrocarbon contamination.

2 Materials and methods

2.4 Phylogenetic and statistical analysis

The clone sequences obtained in this research were checked for chimeric artifacts by Bellerophon. The obtained sequences were aligned and calculated to generate distance matrix, which was used as the input file to the DOTUR program that can assign sequences to phylotypes or operational taxonomic units (OTUs) for every possible distance. In this study, any 16S rRNA gene sequences with ≥97% similarity were assigned to the same OTU. Shannon–Weaver, Simpson, and Chao-1 diversity indexes were calculated. Dominance in clone libraries was calculated as 1–Shannon evenness index (1–E), where is represented as follows:

(1)

where H is the Shannon–Weaver diversity index, S is the total richness in a sample. These sequences were also compared with 16S rRNA gene sequences deposited in public databases using the BLAST search program. The 16S rRNA gene sequences of various bacteria obtained from the GenBank database were aligned with the new sequences using BioEdit 7.0. The phylogenetic tree was constructed by the neighbor joining method with robustness of 1000 bootstrapping value in MEGA 4.0.

To determine the extent to which the separated samples share the common bacterial species, the species co-occurrence pattern was investigated among samples by Venn diagram using Mothur Program 20. Canonical correspondence analysis (CCA) was used to demonstrate the relation of contaminated soil properties versus each sample, and soil properties versus relative abundance of bacterial phyla. Unweighted pair group method with arithmetic mean (UPGMA) dendrograms using Bray–Curtis distance and Euclidean distance matrices were built within PAST software 3.07, to investigate if closer geochemical properties exerted closer bacterial community compositions.

3 Results

3.3 Phylogenetic analysis of bacterial sequences

The 16S rRNA gene sequences of all three bacterial libraries fell into 16 phylogenetic divisions (Fig. 1): Gammaproteobacteria, Alphaproteobacteria, Betaprotebacteria, Deltaproteobacteria, Bacteroidetes, Acidobacteria, Actinobateria, Deinococcus-Thermus, Verrucomicrobia, Planctomycetes, TM7, OD1, Nitrospira, Armatimonades, Gemmatimonadetes and Chloroflexi, and WS_2, WS_4, WS_6 fell into 14, 12, and 9 bacterial divisions, respectively. Gammaproteobacteria, Alphaproteobacteria, Bacteroidetes, and Verrucomicrobia were ubiquitous in all three samples. Betaprotebacteria, Acidobacteria, and Actinobateria also took up huge proportions in some libraries.

Gammaproteobacteria was the most abundant group in all three bacterial 16S rRNA gene libraries, which accounted for 48.5, 55.0, and 32.6%, respectively (Fig. 1). Most of the sequences in Gammaproteobacteria were affiliated to the clones that were isolated from similar contaminated ecosystems, and they were affiliated to Alcanivoracaceae, Xanthomonadaceae, and Pseudomonadaceae-like organisms, whereas remarkable difference could be observed among samples (Fig. 2A). Alphaproteobacteria was the second dominant group in WS_2 (10.0%) and WS_4 (20.2%), and the third predominant group in WS-6 (11.9%) (Fig. 1). Clones in Alphaproteobacteria were mainly affiliated to Rhodospirillaceae, Caulobacteraceae, Sphingomonadaceae, and Erythrobacteraceae, and their abundances in each sample were also dramatically different. Betaproteobacteria was the second abundant group in WS_6 (17.8%) and the fourth predominant group in WS_2 (9.2%), while only 3.9% of WS_4 were from this subphylum (Fig. 1). Clones clustered to Rhodocyclaceae and Comamonadaceae were the main components of Betaproteobacteria (Fig. 2B). The OTUs from Deltaproteobacteria were not abundantly observed (3.8, 1.6, and 3.7% for WS_2, WS_4, and WS_6, respectively). Clones related to the family Nannocystineae composed the biggest family group in this subphylum. Acidobacteria was one of the most significant bacterial groups in WS_2 and WS_6, representing 10.0 and 10.4%, respectively, whereas barely occurring in WS_4 (1.6%) (Fig. 1). Actinobacteria (10.4%) and Bacteroidetes (5.2%) were two of the main components in WS_6. However, both of them were taking minor proportions in WS_2 (0.0 and 3.8%) and WS_4 (3.1 and 3.9%) (Fig. 1). Verrucomicrobia was ubiquitous in all three libraries, representing 5.4, 6.2, and 6.7% for WS_2, WS_4, and WS_6, respectively (Fig. 1). Isolation sources of clones from other minor phyla were mostly similar to hydrocarbon contaminated sites (Fig. 2D).

3.4 Statistical analysis

The soil samples had different physicochemical characteristics (Table 1). WS_2 was mainly contaminated by CHCs and PAHs (Fig. 3A). WS_4 was primarily polluted by BTEX. TPH and TOC were highest in WS_6. For each dominant bacterial group (phylum-level), the main controlling factors were different. For instance, Betaproteobacteria was better controlled by pH than other parameters, while Actinobateria was controlled by TPH and TOC (Fig. 3B). UPGMA cluster analysis was conducted to compare the community structures by calculating Bray–Curtis distance. It was found that WS_2 was closer to WS_4 than WS_6 which contained the most TPH and TOC (Fig. 4A). Similar results could be obtained through comparison of their physicochemical characteristics by calculating Euclidean distance (Fig. 4B). Total shared richness (OTUs) of all samples was 4, and were 5, 10, and 11 between WS_2 and WS_4, WS_2 and WS_6, and WS_4 and WS_6, respectively.

4 Discussion

In this study, microbial communities in soils and their relationships with the petroleum contaminants at an oil field in Hebei, China, were investigated. The potential indicators for soil bioremediation, particularly in oil spills and similar crises were explored. The microbial diversity indexes (Shannon (H)) of WS_2 (3.66), WS_4 (3.60) and WS_6 (4.0) of studied samples were close to those of the oil hydrocarbon contaminated soils at Thuringia, Germany (3.782 to 3.925) 21, but aged petroleum contaminated soil showed lower index (1.45) 22. Although the spatial distance between each sampled locations was close (<20 m), their physicochemical characteristics were comparably different (Table 1). Microbial diversity was correlated with pH in many soil types 23. pH generally reached its peak in soils near neutral, and was negatively correlated with diversity index when pH>8 24, 25. However, highest diversity occurred in the sample with highest pH (9.62) in the present study. Biodiversity was reported to be negatively affected by petroleum contamination in many environments 26, 27, whereas these results showed a completely opposite trend. Shannon index and microbial abundance reached highest as TPH increased. The abundance of petroleum hydrocarbons can enhance the diversity of heterotrophic microorganisms which use petroleum hydrocarbon as carbon source.

A great number of 16S rRNA gene sequences of the clone libraries exhibited high similarity (>95%) to some unexplored microorganisms originating from several similar polluted ecosystems, including groundwater contaminated by trichloroethene and cis-dichloroethylene, soils and aquifer contaminated by hydrocarbons, and soils contaminated by heavy oil, petroleum, gasoline, and creosote (Fig. 2). This phenomenon is commonly observed in many petroleum contaminated ecosystems 1, 13, 14, 16. These microbes could adapt to various petroleum and organic contaminated ecosystems, and had the potential capacity of degrading such contaminants. In addition to unexplored potential degraders, well-studied petroleum degrading bacteria diversely occurred in all samples, but exerted different distributions. In a study of oil polluted site, microbial communities were well driven by the hydrocarbon contamination level 1. Interestingly, heavier hydrocarbon contamination did not induce an increase of oil-degrading bacteria. The distributions of microbial community structures could mainly result from the environmental heterogeneity (Fig. 3), which was confirmed by the UPGMA cluster analysis (Fig. 4). Microbial communities in two samples with closer geochemical characteristics clustered together.

Proteobacteria, which was commonly associated with hydrocarbon degradation, was always abundantly observed in petroleum contaminated soil and sediment environments 15, 26, 28, 29. Gammaproteobacteria was commonly reported to be the predominant group in hydrocarbon-contaminated soils 13, 21. In the present study, Gammaproteobacteria constituted the largest group in all soil samples, but was less dominant in the sample with highest TPH (9538.34 mg/kg) (Fig. 1), which is consistent to Liu et al. 15. Two of the most abundant sequence types in WS_2 library, representing 16.9 and 10.8%, formed subdivisions with Alcanivorax species and the clones from family Alcanivoracaceae isolated from hydrocarbon contaminated soil. In WS_4, whose TPH was between WS_2 and WS_6, the dominant groups were Alcanivorax and Stenotrophomonas. However, both of them were rarely found in WS_6. Alcanivorax, a marine hydrocarbonoclastic bacterium, could become the predominant microbe in oil-contaminated open oceans, coastal waters 30, or some halophilic environments 31. The growth of Alcanivorax could be promoted by the addition of nitrogen and phosphorus fertilizers 32, 33. Several field studies have demonstrated the immense importance of Alcanivorax in oil-spill bioremediation 30, 32, 33. Alcanivorax borkumensis is able to degrade alkanes up to C32, branched aliphatic hydrocarbons, isoprenoid hydrocarbons such as phytane, as well as alkylarenes and alkylcycloalkanes, and can grow on oligotrophic environments 34. These findings highlight the critical role of Alcanivorax playing in natural remediation of oil-polluted marine systems. In a hydrocarbon contaminated soil, Alcanivorax was also abundantly observed, along with Marinobacter and Halomonas 15. Because WS_2 was mainly contaminated by PAHs and CHCs, the abundance of Alcanivorax could serve as a potential degrader for such contaminants. Nonetheless, as the relative abundance of Alcanivorax decreased as TPH increased, its abundance might be restricted by high concentrations of petroleum hydrocarbons. Stenotrophomonas was only found in WS_4 which had the highest BTEX concentration (9.36 µg/kg). Stenotrophomonas spp. were also able to degrade hydrocarbons like PAHs 35, 36. Besides, Pseudoxanthomonas, which was capable of degrading petroleum hydrocarbon compounds like BTEX and PAHs 37, 38, did not exist in the lowest petroleum contaminated sample (TPH = 473.5 mg/kg). OTUs formed clusters to well-studied hydrocarbon degrading bacteria Pseudomonas spp. 21, 29 and Hydrocarboniphaga daqingensis 40 were widespread in all three samples, and their highest abundance occurred in WS_2. Species from Pseudomonas spp. that was able to degrade alkanes, PAHs, polychlorinated biphenyls and other oil hydrocarbons had already been widely applied for the bioremediation of oil-contaminated sites 21, 39.

Alphaproteobacteria was supposed to be correlated with BTEX concentration in soil samples (Fig. 3B). Few researches have reported the degradation ability of Alphaproteobacteria, the major bacteria found in our samples including Dongia. Nevertheless, Sphingomonas from Sphingomonadaceae, which was observed in all libraries, has been reported to be a degrader of organic contamination 41, 42. The most dominant clone from Betaprotebacteria was isolated from gasoline contaminated ecosystem. This subphylum constituted more parts in WS_6 than in the other ones with lower TPH (representing 6.7, 2.3, and 1.6% for WS_6, WS_2, and WS_4, respectively). In addition to Proteobacteria, bacteria from other phyla also constituted huge components of the soil samples. Acidobacterium spp. occurring in WS_2 and WS_6 were able to utilize several organic matters as carbon sources such as methyl-tert-butyl ether 43. Kasai et al. reported that the amount of clones related to Bacteroidetes in petroleum contaminated soil (TPH = 12 000 mg/kg) was higher than that in the uncontaminated soil 44, which was confirmed in our result. The Bacteroidetes and Actinobacteria were two of the most significant groups in the sample with highest TPH, but rarely found in less contaminated samples. The distributions of these two phyla were mainly affected by TPH and TOC (Fig. 3B). Some bacteria genera found in these two phyla (e.g., Arthrobacter, Rhodococcus, Flavobacterium, Cytophaga) had been reported to be able to degrade petroleum hydrocarbons 45-48.

5 Concluding remarks

Microbial diversities, abundances and community structures differed with the gradients of petroleum contamination. Gammaproteobacteria ubiquitously dominated in all samples, while the relative abundance was lowest in the highest TPH contaminated sample. The most dominant OTU in this sample was affiliated to an uncultured bacteria isolated from similar contaminated ecosystem, demonstrating that several significant petroleum degrading microbes were still unknown. Alcanivorax was predominant in the PAHs and CHCs-contaminated soil but TPH was the lowest. Both Alcanivorax and Stenotrophomonas were predominant in highest BTEX-contaminated soil. The potential biodegraders of petroleum contaminated subsurface soils were expected to be found within genera Alcanivorax, Stenotrophomonas, Pseudomonas, Pseudoxanthomonas, Sphingomonas, Hydrocarboniphaga, Acidobacterium, etc. There is a potential relationship between petroleum constituents (e.g., TPH, CHC, BTEX, and PAH) and specific bacterial taxa. Although more biological replicates and sufficient samples will be required, the respond of bacterial community profiles to the complex petroleum contaminants was performed in the present study. Future work should emphasize on exploring the degrading capacities of these heterotrophic bacteria for the bioremediation of petroleum contamination.