Climate change, the opening of the northwest passage, the production and transportation of oil reserves in addition to the large size and number of ocean-going vessels, are putting all Canada's oceans at an elevated risk for an oil spill. Responses to marine oil spills include physical (skimming and recovery), chemical (dispersants, herders) and biological processes (biodegradation). Natural attenuation, a weathering process that includes physical, chemical and biological action on spilled oil, is a potential remediation strategy that needs to be explored and exploited. In the Canadian context, we are using genomics approaches to better understand the natural populations of oil degrading microorganisms in our oceans, their diversity, spatial and temporal dynamics, and locations that may be more vulnerable to oil spills.
The purpose of this study was to perform an evaluation of the effectiveness of an in situ microcosm experimental system to study indigenous microbial communities that have oil degrading potential and to determine whether this experimental system could have an impact on acute toxicity to various marine organisms. In situ microcosms are slitted columns that contain support matrices such as clay beads or river rocks, with or without an oil coating. Columns can be incubated in different locations, at different depths and different time periods, during which microbial biofilm develops on the support materials. By using oil coated and uncoated matrices, comparative microbial community data that demonstrates the response of the microbial community to the presence of oil can be obtained. Long-term incubations (1 year) conducted at CFS-Alert showed that known oil-degrading bacteria (Colwellia, Oleibacter, Thalassolituus, Cycloclasticus, Oceanobacter and Alcanivorax) became dominant only on the oil coated matrices, confirming their presence in the local seawater.
Acute toxicity tests were performed in aquaria on a variety of test organisms to evaluate the possible effects of oil components leaching into the water from the in situ microcosms. Limited and transient toxicity to only two tested organisms (green sea urchin fertilization and green algal growth). Considering the analyses were conducted in a closed circulation system, it is highly likely that in an open ocean environment, toxicity would be negligible. Data from these studies will be valuable to support guidelines for the exploitation of natural attenuation as an alternative response measure (ARM) to address oil spills in Canadian waters.
Climate change, the opening of the northwest passage, the exploitation of offshore eastern oil reserves and the transportation of bitumen products to the west of Canada for distribution are putting all Canada's oceans at an elevated risk for an oil spill. The question that needs to be addressed is whether we are adequately prepared should such an event take place. One of the objectives of this project is to ensure that we have in place the appropriate measures to deal with an oil spill of any magnitude, no matter where it occurs in our oceans. For example, a recent study compiling data on shipping traffic in the Canadian Arctic Ocean has shown an increase of 75% over the past decade (Dawson et al. 2018).
Natural attenuation comprises a number of processes that impact oil spills in the marine environment. These include evaporation, dissolution, dispersion, emulsification, transformation and biodegradation. Because oil is naturally and frequently observed in the ocean, many microbes, primarily bacteria, have the ability to use these compounds as carbon and energy sources. Some of these microbes are so specialized that they are obligate oil consumers (Head et al., 2006). Natural attenuation of hydrocarbons played a significant role in the disappearance of oil in the Deepwater Horizon accident (Mason et al, 2010). This work will expand the knowledge base of the occurrence and extent of naturally-occurring oil-degrading microorganisms in our oceans with a particular emphasis on closing the gap in our ability to understand and potentially exploit their use should an oil spill occur. Microbes associated with a variety of oil spill remediation scenarios, including in situ burning and oil skimming, will be examined in different experimental settings to determine what factors could limit or increase biodegradation potential. The overall objective will be to develop a better understanding of natural oil degradation potential in our oceans, to exploit this potential and to produce key baseline data as a reference for environmental impact and recovery.
The objective of this study was to validate the use of an in situ microcosm system to study oil biodegradation under natural conditions and to use acute toxicity tests to evaluate the potential environmental impacts of using this experimental system. Matrices (clay beads, natural river rocks) with or without an oil coating are incubated in situ in different marine environments for different time periods. These surfaces provide a support for the enrichment of aquatic microorganisms that favour the material surfaces and/or the substrates on the surfaces. In this study we present results from incubations conducted at CFS-Alert (northern tip of Ellesmere Island), as an example. The results generated to date have demonstrated that natural microbial populations in the various locations where in situ microcosms have been installed do have the potential for oil biodegradation. There has been a clear enrichment of oil degrading microbial communities on oiled beads where the in situ microcosms have been deployed, such that a clear distinction can be made between microbes colonizing oiled matrices verses microbes colonizing non-oiled matrices and those that were detected initially in the water column. One of the concerns in using these experimental systems has been the potential toxicity of the in situ microcosms, which was addressed through assessment of several acute toxicity assays to determine whether there are any impacts related to using these experimental systems to evaluate oil biodegradation potential. These analyses are important to validate the use of in situ microcosms for future experimental work to evaluate natural attenuation potential. Other studies using the in situ microcosm system are underway to study the kinetics of biodegradation in more detail and to examine conditions that could improve rates.
Preparation of river rocks for in situ microcosms deployed at CFS Alert
River rocks (commercially available sized rocks) were washed thoroughly with tap water then soaked overnight in 5% HCl, then rinsed several times with distilled water, dried and sterilized in an oven at 200°C for 4 h. Alaska North Slope crude oil (ANS) oil was artificially weathered by purging with air for 48 h (modified after Li et al., 2009), The river rocks (100 g) were soaked in 250 mL beakers containing weathered ANS oil for 2 hours (mixing every 30 min) and subsequently transferred onto trays in a fume hood and weathered for another 2 days. The oil-coated rocks were placed into pre-weighed mesh cloth sacks then added to the middle of the PVC columns (6 cm OD, 32 cm long) with threaded caps at each end and slotted over ca. 15 cm of its mid area (Figure 1). Sterile uncoated river rocks were placed in the columns below and above the rocks to fill the columns entirely.
Installation and recovery of in situ microcosms at CFS Alert (moorings)
The in situ microcosms were deployed in inner Dumbell Bay at CFS Alert (82°29′58″N, 62°20′5″W), NU on August 23, 2018. Seawater (SW) samples were collected adjacent to both moorings. SW samples (2 L) were filtered at 10 psi, in triplicates onto Millipore 0.22 uM polyethersulfone membranes (Fisher Scientific, Ottawa, ON) for T=0 collection of genomic material for total nucleic acid extraction. The filters were transferred into a 50 ml Falcon tube and immediately transferred to the freezer.
The in situ microcosms that were deployed in inner Dumbell Bay at CFS Alert in 2018 were recovered on August 31, 2019, after 1 year. The two moorings that were deployed each contained 3 microcosms with uncoated river rocks and 3 microcosms with ANS oil coated river rocks. In situ microcosms that were recovered were quickly processed for chemistry and genomics analyses. River rocks dedicated for chemistry were transferred to 250 mL amber glass bottles with PTFE lined caps and immediately stored at −20°C.
Total nucleic acid extraction, sequencing libraries and bioinformatics of in situ microcosms (moorings)
Total nucleic acids were recovered from 10 g aliquots of river rocks using a modified version of the hexadecyl trimethyl ammonium bromide (CTAB) method of Ausubel (2002) as described in Tremblay et al (2017). DNA samples were treated with RNase If (New England Biolabs, Whitby, ON), purified using 1 volume of magnetic beads from the Macherey-Nagel NucleoMag NGS Clean-up and Size Select kit (D-MARK Biosciences, Toronto, ON), then quantified using the Quant-iT PicoGreen assay from Fisher Scientific Ltd. (Edmonton, AB).
16S rRNA gene amplicon libraries, sequencing and bioinformatics
Total DNA was amplified using the 515F-Y/926R primer pair covering the V4 and V5 hyper-variable regions of the 16S rRNA gene was used to identify the microbial composition and diversity in each sample. These primers can detect both archaea and eubacteria, having the following sequences: 515F-Y (5′-GTGYCAGCMGCCGCGGTAA-3′) and 926R (5′-CCGYCAATTYMTTTRAGTTT-3′). The 16S rRNA gene libraries for sequencing were prepared according to Illumina's “16S Metagenomic Sequencing Library Preparation” guide (Part # 15044223 Rev. B). Equal amounts of each indexed PCR product were pooled and diluted pooled samples were loaded on an Illumina MiSeq and sequenced using a 500-cycle MiSeq Reagent Kit v2 Illumina (San Diego, CA, USA). Sequencing data were analyzed using an AmpliconTagger workflow (Tremblay and Yergeau 2019).
Metagenomic libraries, sequencing and bioinformatics
Metagenomic sequencing libraries were prepared with 10 ng of DNA using the QIAseq FX DNA Library Kit from Qiagen following the instructions of the manufacturer. Normalization was performed by pooling equimolar amounts of libraries after quantification on the Agilent 4200 TapeStation System using the TapeStation High Sensitivity D5000 assay (Agilent Technologies). The quality of the pooled library was assessed using the Agilent 4200 TapeStation System as described above. The pool was sequenced using Illumina HiSeq4000 technology at the Centre d'expertise et de services Génome Québec, in a rapid mode 2 × 100 bp configuration. A total of 15 samples were submitted for metagenome sequencing and the resulting data (9.8 Giga-bases) were processed through our metagenomic bioinformatics pipeline (Tremblay et al. 2017). Coverage profiles of each sample were merged to generate an abundance matrix (rows = contig, columns = samples) and a corresponding CPM (Counts Per Million–normalized using the TMM method; edgeR v3.10.2; (Robinson, McCarthy, and Smyth 2010)). Taxonomic summaries were performed using a combination of in-house Perl and R scripts and Qiime v.1.9.1 (Caporaso et al. 2010).
Toxicity Test Procedure
Preparation of river rocks and clay beads for toxicology experiments.
The clay beads were rinsed thoroughly with tap water twice for 1 min then once with distilled water in a stainless steel sieve with a mesh opening size of 4.75 mm. The beads were transferred to a tray with paper towels and allowed to dry. Once dried, they were washed again with distilled water to remove traces of fine powder. The beads were dried as before and autoclaved in a glass bottle for 20 min.
ANS crude oil and Cold Lake Blend dilbit (CLB), were artificially weathered by purging with air for 48 h (modified from Li et al., 2009), and used to coat 65 g of clay beads and 100 g of river rocks. The clay beads were soaked in a 250 mL beaker containing ANS, CLB oils or Marine Diesel (MD) for 3 h (mixing every 30 min) and subsequently transferred onto trays in a fume hood and weathered for another 2–4 days (ANS and CLB oils) and for 24 h (MD). The river rocks were prepared as described previously and soaked with ANS and CLB oils for 2 h (mixing every 30 min).
Clay beads in the fall of 2019 and winter of 2020 contained 3.62 g and 6.23 g of ANS oil, respectively. Similarly, river rocks contained between 1.01 to 1.38 g of ANS oil and 1.73 to 2.25 g of CLB oil for both the fall and winter experiments. Clay beads coated with MD, in the fall and winter lost 6 and 4.4 % of their mass after weathering in a fumehood for 24 h which amounted to 5.71 g and 6.32 g of remaining MD on the beads.
Uncoated clay beads or river rocks (negative controls), and the oil coated clay beads or river rocks were individually placed in microcosms (middle-slotted PVC tubes, 6 cm OD, 32 cm long) (Fig. 1). Each column was then individually deployed in a 20-L glass tank containing recirculating (5 L/min) natural seawater from the Bay of Fundy, NB (50 μm filtered) for 14 days. Oil concentrations were monitored daily by fluorometry (Horiba Aqualog®; excitation-emission range: 240–800 nm) and water samples (1 L) were collected from each tank (n = 7) on days 0, 3, 7, and 14, and used as exposure media for toxicity tests at 100, 50, and 25% strength. Test solutions were characterized by fluorometry at pre- and post-exposure and select samples were retained for subsequent genomic and chemical analysis (data not shown). Acute toxicity tests were conducted using green algae (Dunaliella tertiolecta), ice-algae (Nitzschia frigida), green sea urchin (Strongylocentrotus droebachiensis), American lobster (Homarus americanus), and Atlantic cod (Gadus morhua). The toxicity tests were conducted at Huntsman Marine Science Center (St. Andrews, NB, Canada) in Fall 2019 and Winter 2020, using waters collected from the tanks on days 0, 3, 7, and 14 post-deployment as exposure media in addition to, a negative control (seawater tank), a positive control series (3 copper sulfate concentrations covering the species specific LC50), and a substrate control (uncoated clay beads or river rocks). The test methods for each season and species are briefly described below and summarized in Table 1.
The green algae Dunaliella tertiolecta and American lobster larvae (Homarus americanus) were used in the Fall 2019 toxicity tests. The acute toxicity bioassay of green algae was evaluated based on the algal growth rate at 72 h exposure to each treatment. The test solution was prepared from dilutions of the tank water and was allowed to warm at 22°C (optimal growth conditions for this species) before adding to the sterile 25-ml borosilicate glass test tubes (three replicates per treatment; 9 ml of test solution and 1 ml of algal inoculation). Environmental conditions (e.g., lighting, pH, and temperature) were monitored daily during the exposure, with each test tube manually swirled twice per day. At the end of the 72 h exposure period, 1 ml of 1% glutaraldehyde preservative was added to each experimental unit. The larval lobster toxicity test was conducted in the environmental chamber at 15 ± 2°C with a 16 h light and 8 h dark photoperiod. Tests solutions were transferred to 20-mL scintillation vials each containing one lobster larvae, with 10 replicates per treatment. Lobster larvae were assessed for toxic effects (i.e., immobility and mortality) at 24 h and 48 h. Water quality parameters including dissolved oxygen (%DO), pH, salinity (mg/L), and temperature (°C) were measured in the test solutions at the beginning and end (48 h) of the exposure.
An Arctic algae (Nitzschia frigida), Atlantic cod (Gadus morhua), and the green sea urchin (Strongylocentrotus droebachiensis) were used for toxicity tests in Winter 2020. The 72 h algal growth bioassay with the arctic algae was conducted with the same culture preparation and methods as described for the green algae, expect the test temperature was 5 ± 2°C and the bioassay conducted with the day 14 water had an extended exposure duration of 7 days to accommodate the tislow growth rate of this species. The 24 h toxicity tests with the Atlantic cod embryos were conducted in the environmental chambers at 5 ± 2°C with 16:8 light/dark photoperiod. Each treatment had 3 replicates in 20-mL scintillation vials, each containing 20 mL of test solution and 20 embryos. After 24 h exposure, each scintillation vial was assessed for hatched larvae and unhatched embryos, with the unhatched embryos transferred to the clean seawater and monitored daily for mortality or and hatching. In a parallel cod exposure, post-hatch larvae were also exposed to the treatments (3 replicates per treatment, 10 cod larvae replicate), with survival assessed at the end of the 24h exposure period. The green-sea urchin fertilization bioassay followed the Environment and Climate Change Canada Standard Methods (EPS 1/RM/27; 2011).
Toxicity Statistical Analysis
All analyses were conducted in R (R Core Team 2012) with a significance level of 0.05. Analysis of the test endpoints (e.g., survival, growth, fertilization) followed the data handling methods prescribed for each specific method. All data were assessed for normality and homogeneity of variance using Shapiro-Wilk and Levene's tests, respectively. The analysis of variance (ANOVA) was performed followed by Dunnett's test when significant differences were identified. Controls (e.g., control seawater and matrix controls) were compared for significant differences (p < 0.05) and where significant difference did not exist between controls, they were pooled. In case of significant difference between controls, the appropriate paired matrix control was used for further data analysis. To provide a common metric for comparison across seasons and species, the responses of the exposed organisms were normalized to the response of the control exposures.
RESULTS AND DISCUSSION
In Situ Microcosms
Microcosms deployed at CFS-Alert in 2018 were recovered after one year of in situ incubation. The oiled and unoiled rocks were recovered, and total DNA was extracted for microbial community analysis and samples were also submitted for residual hydrocarbon determination. The total bacterial community structures were analyzed using principal coordinate analysis of Bray-Curtis Beta diversity dissimilarity. The communities differed considerably between the oil and unoiled rocks (Fig. 2). The initial water contained a microbial population that was also considerably different from the populations that colonized the rocks, showing that the rocks had enriched specific communities from the initial seawater that were different, depending on whether they were exposed to oil or not. Enrichment of oil degraders from the initial seawater showed that the oil degraders were not initially part of the dominant microbial populations. The tight replicate clustering of the populations for each sample type (initial seawater, unoiled rocks, oiled rocks) were separated into distinct zones in the PCoA plots. The clustering demonstrated that the transition from the initial seawater was distinct and differently directed depending on whether oil was present or not, so that the microbial communities that developed on the support material were influenced by the presence of the oil substrate.
To further characterize the differences between the initial seawater and the oiled and unoiled rocks, the dominant bacterial taxa were plotted (Fig. 3). The initial seawater was dominated by Polaribacter (Flavobacteriales), Pelagibacter ubique (Pelagibacterales) and Ascidiaceihabitans (Rhodobacterales) as examples, which comprised more than 35% of the community.
These bacteria were not dominant on the oiled or unoiled rocks, which were dominated by distinctly different bacteria. Although there were several bacteria that dominated both rock types, the oiled rocks had a unique dominance of Oceanobacter, as well as a unique Cycloclasticus that was not part of the dominant bacteria on the unoiled rocks. These two genera are well-known hydrocarbon degrading bacteria (Teramoto et al., 2009; Dombrowski et al., 2016).
In addition to the taxonomically dominant bacteria, we also examined various functional genes involved in hydrocarbon degradation using the metagenomic sequencing data (Fig. 4). There was an enrichment of genes associated with hydrocarbon degradation in the oil-coated river rocks in comparison to the initial seawater and the uncoated river rocks. In particular, the alkane monooxygenase (alkB, alkM) gene involved in initial alkane oxidation and the benzoate dioxygenase genes (benABC) were more dominant in the oil coated river rocks in comparison to the initial seawater and the uncoated rocks. These data support the selective enrichment of hydrocarbon degrading bacteria on the oil-coated river rocks.
Toxicity: Exposure Assessment
The loading of product onto the substrates varied with product and substrate type, but was fairly consistent between the seasons. For the Fall 2019 testing, the loadings were 0.84 g (ANS on river rocks), 1.80 (CLB on river rocks), 2.96 g (ANS on clay beads), and 5.71 g (MD on clay beads). For the Winter 2020 testing, the loadings were 1.37 (ANS on river rocks), 2.03 (CLB on river rocks), 5.99 (ANS on clay beads), and 6.32 g (MD on clay beads). Based on the fluorometry profiles, the tank water treated with MD and ANS coated clay beads showed a concentration increase in the first two days after deployment, which was greater than that observed with the ANS and CLB coated river rocks. Irrespective of type of petroleum product, the tank water from treated clay beads demonstrated a gradual increase in concentration over the course of experiment (14 d), while treated river rocks showed a decline in concentration over the same period. These release dynamics may be influenced by the variable loadings on substrates, with the clay beads having 3.5 - 4 times greater mass loaded than the paired river rock column. The relative ranking of concentration from each tank was the ANS and MD coated clay beads followed by ANS and CLB coated river rocks.
A consistent relationship was identified between the nominal strength of the test solution (i.e., 25, 50, and 100%) and concentration response, with the relative fluorescence unit (RFU) increasing with increasing oil strength of the test solutions, while the signals of the matrix controls (uncoated beads or rocks) did not show a significant difference from the seawater control at each time point (p < 0.05). The relative concentration of the 100% test solutions from each treatment varied at each sampling point, with the overall increase of signals for the ANS and MD treated clay beads followed by the ANS and CLB treated river rocks with lower RFU at day 14. Upon receipt of analytical results (e.g., TPH, TPAH etc.), the RFUs will be converted to estimated concentrations to provide greater context for the results. Based on previous work, the estimated maximum concentration observed in the CLB treatment was 4.5 μg/L PAH (sum 31 analytes), observed 5 days after deployment.
Toxicity: Effects Assessment
No significant difference was identified between control seawater and control matrices (i.e., clay beads or river rocks; p < 0.05). The failure to meet test validity criteria (e.g., control performance) resulted in the exclusion of the ice-algae data (days 0, 3, 7, and 14), the lobster bioassay (days 3, 7, and 14) and the day 3 exposure from the green sea urchin fertilization assay. Due to the biological constraints, the cod embryo was not available for the toxicity test at day 0. The data from the valid toxicity tests are presented in Figure 5, with responses normalized to the pooled control (full data summary in Table 2). The majority of the responses from the exposures represent less than a 20% difference from the response of the control organisms, highlighting a lack of biologically significant differences.
There were few statistically significant effects of exposure observed in this study, with the majority of effects being observed in the green sea urchin fertilization assay. The fertilization of green sea urchin was reduced significantly in the 100% ANS coated clay beads (p = 0.002) and 100% CLB coated river rocks (p = 0.01) solutions in the exposure conducted on day 0. Statistically significant reductions in fertilization were observed in the exposures conducted on day 7 in the 100% ANS coated clay beads (p = 0.008), the 25 and 100% MD coated clay beads (p = 0.04 and 0.001), and the 25 and 100% CLB coated river rocks (p = 0.004 and 0.001). In the cases of the MD coated clay beads and the CLB coated river rocks, there was not a consistent concentration response, as evidenced by the lack of significant difference in the 50% treatments despite greater RFU and estimated concentration than the 25% solutions. Results of the analytical characterization of the solutions may provide further insight. The survival of larval cod was significantly different from the pooled controls only for the exposure conducted on day 3 in the 100% ANS coated clay beads exposure (p = 0.01). The growth of green algae was significantly different from the pooled controls only for the exposure conducted on day 3 in the 100% ANS coated river rocks exposure (p = 0.003).
The in-situ microcosm system was used to evaluate the natural attenuation potential for oil in an Arctic marine environment at CFS-Alert. The results demonstrated the presence of specific hydrocarbon-degrading natural bacteria in this environment that were enriched on oil coated matrices (river rocks) in comparison to non-coated matrices. Key genes involved in aliphatic and aromatic hydrocarbon-degradation were also enriched on the oil coated river rocks. The in-situ microcosm system demonstrated that this concept for the analysis of natural oil degrading activity is a valid approach to assess natural attenuation capability. Acute toxicity analyses of the in situ microcosm system under controlled laboratory conditions showed that there was intermittent and low toxicity to only two of the tested organisms (green sea urchin fertilization and green algae growth), from the leaching of some oil off the matrix. As toxicity tests were conducted in a closed-recirculated system, it is likely that in an open ocean environment, any toxicity would be minimal to non-detectable.
The authors gratefully acknowledge funding for this project from Fisheries and Oceans Canada's Multi-Partner Research Initiative, a program under Canada's Oceans Protection Plan. We also thank the Environment Office of the Department of National Defense at CFB-Trenton for logistical support for research work conducted at CFS-Alert.