https://orcid.org/0000-0002-8702-9329
ABSTRACT
Many aquatic environments are at risk for oil contamination and alkanes are one of the primary constituents of oil. The alkane hydroxylase (AlkB) is a common enzyme used by microorganisms to initiate the process of alkane-degradation. While many aspects of alkane bioremediation have been studied, the diversity and evolution of genes involved in hydrocarbon degradation from environmental settings is relatively understudied. The majority of work done to-date has focused on the marine environment. Here we sought to better understand the phylogenetic diversity of alkB genes across marine and freshwater settings using culture-independent methods. We hypothesized that there would be distinct phylogenetic diversity of alkB genes in freshwater relative to the marine environment. Our results confirm that alkB has distinct variants based on environment while our diversity analyses demonstrate that freshwater and marine alkB communities have unique responses to oil amendments. Our results also demonstrate that in the marine environment, depth is a key factor impacting diversity of alkB genes.
INTRODUCTION
Oil contamination is a major concern in both marine and freshwater settings. As oil prospecting and transport has increased, there is increased risk of oil spills in aquatic settings. Many microorganisms have the ability to biodegrade oil as part of their metabolism and as such have been used in bioremediation of the spilled oil (Hazen, Prince and Mahmoudi 2016: 2121–9; Hazen and Techtmann 2018: 1–18). The diversity of oil-degrading taxa has been most extensively studied in the marine environment due to the scale of recent oil spills in that setting (Prince, Gramain and McGenity 2010; Hazen, Prince and Mahmoudi 2016: 2121–9). Many of these recent studies have shown a great diversity of microbes that respond to released oil, both in marine as well as freshwater environments (Dubinsky et al. 2013: 10860–7; Liu et al. 2017; Techtmann et al. 2017; Deshpande et al. 2018: 111–20; McFarlin et al. 2018; Ribicic et al. 2018: 370–8). Much of this previous work has focused on assessing diversity through the 16S rRNA gene, which provides information about the taxonomic affiliation of an organism, but often leaves functional information to be inferred. In this current study, we are interested in profiling microbial communities using a functional gene marker for alkane biodegradation.
Alkane degradation in the environment plays a crucial role in carbon cycling and bioremediation. Alkanes are saturated hydrocarbon chains that constitute anywhere from 20–50% of crude oil (Rojo 2009: 2477–90). One of the primary genes involved in microbial alkane degradation is the alkB gene, which encodes an alkane hydroxylase (Ji et al. 2013). Sequences for alkB have been found in diverse environmental metagenomes and isolates, from soil, marine, and freshwater microorganisms (Smith et al. 2013; Nie et al. 2014). Alkanes pose a threat in that short carbon chains can permeabilize lipid membranes and other biologically relevant hydrophobic molecules, while long chain alkanes form slicks which present physical barriers for nutrient uptake (Sikkema, Debont and Poolman 1995: 201–22). While there are other alkane hydroxylase gene families such as CYP153, the alkB family is more diverse and is the most commonly found across bacteria (Nie et al. 2014). The AlkB protein is a monooxygenase that catalyzes the addition of a hydroxyl group to the terminal methyl group of an n-alkane hydrocarbon chain under aerobic conditions. Further oxidation leads to the formation of a fatty acid which can be fed into the β-oxidation metabolic pathway (Watkinson and Morgan 1990: 79–92).
Previous work has characterized marine or terrestrial alkB diversity and prevalence (Kloos, Munch and Schloter 2006: 486–96; Wasmund et al. 2009: 7391–8; Smith et al. 2013); however, a limited number of reports exist with respect to alkB gene diversity and distribution in freshwater environments. This is a gap which this work seeks to fill. More than ever, expanded oil pipeline development, aging underwater oil pipelines, and shipping transport all have the potential to release oil into surface freshwater and adversely impact culturally and economically important regions. It has been suggested that the prevalence of oil-degrading bacteria in the marine environment is due to previous exposure of the community to hydrocarbons via natural oil seeps and production by algae (Rojo 2009: 2477–90; Hazen and Techtmann 2018: 1–18). Freshwater settings have fewer natural seeps and therefore may have a limited exposure history to alkanes. Many of these freshwater environments have relatively low ambient levels of n-alkanes when compared with the oceans (Parrish et al. 1988: 1–15). The Gulf of Mexico provides an example of the impact of natural seepage on a marine environment, with an estimated 1500–3800 barrels of crude oil per day entering the basin from natural seepage (Council 2003). The impact of previous exposure on hydrocarbon-related gene diversity was further demonstrated through the finding of higher diversity of alkB genes in environments with chronic hydrocarbon seepage when compared to historically pristine regions (Wasmund et al. 2009: 7391–8). Similarly, it has been long documented that areas with previous exposure to hydrocarbons have significantly higher abundances of hydrocarbon degrading organisms, compared to those with no exposure (Atlas 1981: 180–209; Head, Jones and Röling 2006: 173–82). These findings combined with the limited inputs of oil to freshwater systems suggest that a microbial metabolic response to a sudden increase in hydrocarbon concentration in pristine freshwater communities could be slower and less robust due to limited prior exposures (Deshpande et al. 2018: 111–20). Understanding the distribution and phylogenetic diversity of a crucial hydrocarbon-degrading gene such as the alkane hydroxylases in disparate environments could help inform decision making, both before and after an incident involving the release of large quantities of alkanes.
To better understand the diversity of alkB genes, we examined the complement of alkB genes in select freshwater and marine settings. Here we employed next-generation sequencing to profile the alkB diversity in these sites. As there is great concern over the potential for an oil spill from pipelines in and near the Laurentian Great Lakes, we focused our characterization of surface freshwater samples here. These freshwater sites, while historically free from large-scale anthropogenic hydrocarbon inputs, were selected due to the potential for such accidents from nearby pipelines. The marine samples were collected at various depths from the Sargasso Sea between Barbados and Bermuda, an area known for having few seeps and low basal hydrocarbon levels.
This study represents one of the first to examine the diversity the alkB gene family from surface freshwater microbial communities. We hypothesized in doing this: (i) If a separate set of alkB genes have evolved for freshwater function, then we would observe distinct clades of operational taxonomic units (OTUs) from freshwater samples and distinct OTU clades from marine samples, (ii) Oil would increase the diversity of the alkane hydroxylase genes in marine samples, (iii) Oil would have less of an impact on the community composition of surface freshwater gene diversity compared to marine systems, based on preliminary 16S rRNA analysis of the microbial community response to crude oil in the Great Lakes (Timothy Butler unpublished data) and (iv) AlkB diversity would decrease with depth, consistent with Wasmund et al. 2009 (Wasmund et al. 2009: 7391–8).
MATERIALS AND METHODS
Sample preparation
Of the 78 samples used in this experiment, 43 were from freshwater and 35 were from the ocean. The freshwater samples were collected in mid-August of 2016 from Lake Superior, Lake Michigan and Lake Huron aboard the NOAA 5501. Environmental conditions at the sampling locations are shown in Table S2 (Supporting Information). Samples were collected from surface water from three sites in the Straits of Mackinac and two sites in Lake Superior. Upon collection, one liter of ambient water was filtered through a 0.2 µm PES filter and frozen at −20°C. Marine microcosms were set up immediately after the water was collected on the ship. Due to limited storage and incubation space on ship, the samples for freshwater microcosm experiments were stored at approximately the water's temperature (approximately 25°C) for five days before being placed at 4°C for seven days until microcosm setup. Details on environmental conditions of ambient water samples are showing in Table S1 (Supporting Information).
Marine samples were collected in the Sargasso Sea from depths of 40, 1200, 3200 and 5000 m. These samples were taken as part of the UNOLS Chief Scientist Training Cruise in 2014 aboard the R/V Atlantic Explorer. Samples were collected using a CTD rosette with water collected at various depths throughout the water column. Four liters of water were immediately filtered onto a 0.2 µm PES membrane and stored at −20°C until DNA extraction.
Microcosms were set up for both freshwater and marine samples. Both freshwater and marine microcosms were part of larger projects that are described in Butler et al. in preparation and Miller et al. 2020 (Miller et al. 2020: e01448–20), respectively. The analyses presented here represent experiments of opportunity using the samples available. Freshwater microcosms were set up to examine the long-term impact of oil on microbial communities in surface fresh water. The marine microcosms were set up on ship to examine the short-term impacts of oil exposure on deep ocean communities. The oil used in each microcosm was chosen based on the oil types that could be released in those environments. Freshwater microcosms were amended with Bakken crude oil (API gravity 40.6°) and Cold Lake Diluted Bitumen (API gravity 21.7°) both of which are transported through pipelines that cross freshwater bodies. The marine microcosms were amended with Norne Blend (API gravity 29.6°), which is a light sweet crude similar to those found in adjacent marine hydrocarbon basins. Freshwater microcosms were set up using surface water from three sites in the Straits of Mackinac between Lake Michigan and Lake Huron as well as two stations in Lake Superior. Marine microcosms were set up using deep ocean water (>1000 m depth) from three sites. More details about microcosm set up are shown in Table S2 (Supporting Information).
Freshwater microcosms were set up by adding 100 ml of water into a serum bottle. Triplicate microcosms were set up for three treatments. One treatment was 25 ppm of Bakken crude oil, another treatment was 25 ppm of Cold Lake Diluted Bitumen (Dilbit) crude oil, and the third treatment was no-oil, control microcosms. Microcosms were incubated in the dark at room temperature (20°C) without shaking. These microcosms were sacrificially sampled at five time points once per week for five weeks. Samples were collected by filtering the water through a 47 mm diameter PES filter with a pore size of 0.2 µm. DNA was extracted from these filters using the modified Miller technique (Techtmann et al. 2015: e0120605). 16S rRNA gene sequencing was performed and results are presented in another manuscript (Butler et al. in preparation). DNA from these same microcosms were used to profile the alkB diversity throughout the Laurentian Great Lakes of North America. For samples from the Great Lakes, control microcosms were considered no-oil and oil-amended samples were considered with oil.
Oil amended microcosms were set up for select marine samples (Table S2, Supporting Information) by adding 10 ppm of crude oil to raw seawater and incubated at 4°C on ship with no shaking. Samples were collected at 12, 24 and 72 hours after addition of oil. For the marine samples, ambient water samples and control microcosm were considered no-oil, while oil-amended microcosms were considered oiled. Filters were extracted using the modified Miller technique. 16S rRNA data was analyzed and is presented in Miller et al. 2020 (Miller et al. 2020: e01448–20).
PCR, sequencing prep and sequencing
The alkB gene from environmental samples was amplified using the alkB-f1 and alkb-r1 primers from Kloos, Munch and Schloter 2006 (Kloos, Munch and Schloter 2006: 486–96). An initial PCR was performed with the gene specific primers to amplify the alkB gene from the environmental samples with 5 µl of environmental DNA used as a template. This consisted of an initial denaturation step of 98°C for 5 minutes, followed by a cycle of 10 seconds at 98°C, 30 seconds at 55°C, and 60 seconds at 72°C. This cycle was repeated 35 times, followed by 72°C for five minutes and a hold at 4°C before the plate was transferred and stored in at −20°C. The PCR products were purified using the AxyPrep magnetic bead PCR purification procedure. A second, eight cycle, round of PCR was used to add Illumina sequencing adaptors and sample-specific indices. Golay barcodes were used for indexing purposes (Caporaso et al. 2012: 1621–4). The final libraries were purified with the AxyPrep magnetic bead clean up procedure. Next, a Thermo Scientific NanoDrop was used to determine DNA concentration of the amplicons and an Agilent Bioanalyzer was used on several samples to verify the presence of a single amplicon. Once this was confirmed, the samples were multiplexed into pools of similar concentration. Concentration of the pools were determined using Qubit and were normalized to an equimolar concentration so that a final pool of 4 nM could be used for sequencing on the Illumina MiSeq using a v3–600 cycle kit. Raw sequencing reads are deposited in the SRA under Bioproject accession number PRJNA631991
Sequence processing and data analysis
Raw fastq files were demultiplexed and the forward reads were quality filtered in QIIME 1 (Caporaso et al. 2010: 335–6) to remove reads with average phred scores of less than 20. The quality filtered forward reads ranged from 290 to 301 bases in length. After filtering, 277 076 sequences were retained. The number of reads in each of the samples used in shown in Table S6 (Supporting Information). Due to the short nature of the sequences, all subsequent analyses were done using nucleic acid sequences in order to preserve as much information as possible. OTUs were picked using de novo OTU picking using UCLUST at a similarity of 97% (Edgar 2010: 2460–1). Chimeras were determined using VSEARCH (Rognes et al. 2016). Chimeric sequences were removed. Representative sequences were determined for each OTU. Initially, there were 30 000 OTUs present, and this was eventually trimmed down to 885 as follows; samples with less than 300 reads were removed, leaving 41 of the initial 78. OTUs were subsequently removed, leaving only OTUs present in three or more samples and OTUs with 20 or more reads. These trimming steps were designed to ensure that only OTUs with high prevalence and abundance in samples from which DNA was successfully extracted and amplified would be included in the analysis. After these successive rounds of trimming samples and OTUs, we ended with 41 samples containing a total of 885 OTUs. Differential abundance between samples for these OTUs was calculated using DESeq2 (Anders 2010: 1–17). OTUs were classified as ‘Freshwater’, ‘Marine’ or ‘Neither’ based on the log 2-fold change and the adjusted P value of less than 0.05 from DESeq2. We considered an OTU to be enriched in one category if it was present in that category with a log 2-fold change of greater than 2 or less than −2 and an adjusted P value of less than 0.05. OTUs enriched in Marine samples had a negative log 2-fold change while OTUs enriched in Freshwater samples had a positive log 2-fold change. If an OTU did not match our prevalence and abundance cutoffs and was not significantly different enriched in either the ‘Freshwater’ or ‘Marine’ categories based on DESeq2 analysis, it was considered enriched in neither sample type and labeled as ‘Neither’.
Similarly, OTUs enriched in oil and no-oil freshwater microcosms were determined using DESeq2. The full OTU table was subset to only include samples from freshwater mesocosms. OTUs were considered enriched in a category if they showed a log 2-fold change of greater than 2 for oil enriched OTUs and less than −2 for no-oil enriched OTUs and an adjusted P value of less than 0.05.
Representative sequences of the 885 OTUs along with 76 reference alkB gene sequences (Table S3, Supporting Information) were used construct a phylogenetic tree. 70 of these reference sequences were chosen representatives from each of the AlkB clades identified in Nie et al. 2014 (Nie et al. 2014). An additional 6 reference sequences were chosen from Smith et al 2013 (Smith et al. 2013). The 300 bp nucleic acid sequences were aligned using MAFFT with maxiter of 1000 (Katoh and Standley 2013: 772–80). Modeltest-ng was used to determine an appropriate model for maximum likelihood phylogenetic analysis as well as if the gamma distribution should be included in the model (Darriba et al. 2019: 291–4). A phylogenetic tree was constructed using FastTree with the General Time Reversible (GTR) model with the gamma distribution (Price, Dehal and Arkin 2010). Reliability of branch points was determined using the Shimodiara-Hasegawa test. The phylogenetic tree was rooted at the midpoint. A phylogenetic tree was visualized using the Interactive Tree of Life website (Letunic and Bork 2019). All other figures were produced using the ggplot2 package in R (R Core Team 2013). Diversity analysis was performed using phyloseq (McMurdie and Holmes 2013: e61217). Prior to diversity analysis, reads were rarefied to 270 reads per sample, the lowest number of reads present in any sample. Alpha diversity was calculated by rarifying the OTU table to a depth of 270 and repeating the rarifying step 100 times to obtain 100 rarified tables. These tables were then used to determine the mean Shannon diversity for each sample. Beta diversity was visualized using principal coordinates analysis (PCoA) plots, created from Bray–Curtis dissimilarity of the rarified OTU table. The Shapiro-Wilke test indicated that the Shannon diversity our data was not normally distributed (P ≤ 0.0001). Hence, Kruskal–Wallis tests were used to test the hypothesis that there were differences in the Shannon diversity between various comparisons throughout our analyses. Taxonomic analysis of representative sequences for each OTU was determined using kraken2 (Wood, Lu and Langmead 2019: 257). Taxonomy was determined using the MiniKraken pre-built 8 GB database. This was constructed from complete bacterial, archaeal, and viral genomes in RefSeq (as of Oct. 18, 2017). Taxonomy was extracted from the NCBI taxid generated for each representative sequence from the kraken output.
RESULTS AND DISCUSSION
Distinct alkB diversity in marine and freshwater environments
Diversity of functional genes may provide more in-depth insights into the diversity of organisms that perform a function than 16S rRNA sequencing alone. In this study we sought to characterize the diversity of alkB genes in freshwater and marine settings. Richness and evenness of the alkB gene in these samples were assessed by a number of metrics (Shannon, Inverse Simpson, total richness). The diversity of alkB genes was higher in the freshwater settings compared to the marine settings for every metric tested (Shannon (Fig. 1), Inverse-Simpson and total richness (Figs S1 and S2, Supporting Information)). Kruskal–Wallis tests were used to determine if the Shannon diversity was significantly different between freshwater and marine. There was a highly significant difference in diversity between freshwater and marine samples (P = 1.62 × 10−7). This test compared samples regardless of treatment state (oil or non-oiled). Therefore, a Kruskal-Wallis test was also performed between freshwater no-oil and marine no-oil samples. The ambient freshwater and marine samples remained significantly distinct (P = 2.02 × 10−4), confirming a divide between freshwater and marine alkB communities.
Shannon index of diversity of samples organized by environment and oil amendment. Boxes encapsulate the middle 50% of samples, the line indicates the median.
Not only were the richness and evenness different between freshwater and marine samples, but the composition of the alkB community was different. PCoA analysis using Bray–Curtis dissimilarities indicates that there was a distinct set of alkB genes present in the freshwater samples relative to the marine samples (Fig. 2). This PCoA plot explains 40.8% of the variance in the microbial community. There is a clear clustering of the samples by environment type (freshwater versus marine). PERMANOVA analysis indicates that the differences visualized in the PCoA ordination were significant (P = 0.001, R² = 0.2853). These findings suggest little overlap between the alkB genes found in freshwater and marine environments.
Principal coordinate analysis of samples, color indicates environment and oil condition.
To further understand the evolution and diversity of the distinct alkB genes in freshwater settings, we constructed a phylogenetic tree (Fig. 3) using representative sequences for the OTUs from our data set and select OTUs from previously published work (12) (Smith et al. 2013). Of our 885 OTUs, 525 out of the 885 OTUs were significantly enriched in either ‘freshwater’ or ‘marine’ based on their log2-fold changes and adjusted P values from DESeq2. The rest were assigned as ‘neither’, since they were not significantly enriched in one of the two environments. Our results indicate that the freshwater alkB genes generally form distinct clades from the marine alkB genes (Fig. 3). There were very few clades of alkB genes that had representatives from both marine and freshwater settings. This finding indicates that the evolution of freshwater alkB genes has been independent from the alkB genes found in the oceans.
Phylogenetic tree containing the OTUs of alkB genes found in our study. Color around the exterior indicate metadata related to each OTU. The outer ring represents the environment in which that OTU was enriched with light green representing OTUs enriched in marine sample and light blue enriched in freshwater samples. OTUs that were enriched in neither marine or freshwater are shown in black and reference sequences are shown in red. The middle ring indicates the taxonomic affiliation of an OTU, if any, as assigned by Kraken. The inner ring is only present for reference sequences pulled from the Nie et al. 2014 paper and the color indicates where they cluster in that publication. SH branch supports are shown as circles on the branch with the size of the circle corresponding to the branch support. Support value are only shown for branches with values of greater than 0.75. For a larger, rooted version of the tree, see Fig. S4 (Supporting Information).
This phylogenetic tree showed that many of the clades are dominated by a single taxonomic group. While there are some cases were our tree shows phylogenetic groupings similar to Nie et al. There are also some instances where phylogenetic groups from Nie et al are split into different clades in our tree. This incongruity could be due to the Nie et al tree being constructed from full length amino acid sequences and our tree being constructed from shorter nucleic acid sequences. Our tree also shows large clades of unclassified proteins that do not contain reference sequences from the eight clades identified in the Nie et al. These unclassified sequences may represent novel groups of alkB from uncultivated groups.
Insight into how AlkB variants have diverged in different environments could help to increase understanding of alkane degradation potential as well as forge a more general relationship between rRNA taxonomy and functional genes. Previous studies have indicated distinct functional gene community composition between marine and freshwater environments as well as distinct taxonomic communities (Wang et al. 2012: 8264–71). In particular, work by Nie et al. (2014) have shown that alkB genes from freshwater and marine environments appear to have been derived from organisms of distinct taxonomy. The Nie et al study demonstrates close ties between taxa and functional genes in phyla such as Actinobacteria, and Betaproteobacteria, but inconsistencies arise between 16S rRNA and the alkB genes of Gammaproteobacteria (Nie et al. 2014). However, Nie et al were using alkB genes derived from previous studies including isolates. Since isolates only represent a small proportion of the environmental diversity, these percentages may be skewed relative to the true environmental abundance and taxonomic distribution of alkB genes. Our results confirm this finding as there are clades that are almost exclusively composed either Betaproteobacteria or Actinobacteria (Fig. 3). For example, one clade that includes alkB sequences from Ralstonia spp. is dominated by alkB sequences classified as other Betaproteobacteria. Furthermore, another clade is primarily composed of Actinobacteria including reference sequences of alkB genes from Rhodococcus, Mycobacteria, and Prauserella. All of these Actinobacter reference sequences were from clade 1 described in Nie et al. Additionally, Gammaproteobacterial alkB sequences in clades with representative sequences from groups 1 and 2 were split into different clades.
If it is assumed that alkB functional genes across all phyla closely follow 16S rRNA phylogeny, then it is logical that alkB genes would exhibit separate clustering between freshwater and marine samples based on the distinct microbial community composition between those environments. Gammaproteobacterial alkB variants would be assumed to be enriched in marine samples amended by oil, based on previous work that demonstrated Gammaproteobacteria as some of the dominant responders to oil (King et al. 2015: 377–401). Our results indicate that the vast majority of alkB sequences from freshwater samples were unclassified (84.8% ± 8.4% of recovered reads) (Fig. 4). Overall, Betaproteobacteria and Gammaproteobacteria were more highly abundant in the marine samples compared to the freshwater samples. While Actinobacteria were present at similar abundances in marine and freshwater samples, phylogenetic analysis shows distinct clades of freshwater and marine Actinobacterial alkB sequences, including many freshwater alkB sequences being included in an unclassified class of Actinobacteria. This phylogenetic contrast between freshwater and marine alkBs suggests that there is a distinct evolutionary history of marine and freshwater alkB genes.
Taxa plot of alkB sequences with taxonomy assigned by kraken.
Impacts of Oil on alkB diversity
Previous studies have indicated that the addition of oil in the marine environment selects for a distinct population of oil-degrading microbes (Hazen et al. 2010: 204–8; Redmond and Valentine 2012: 20292–7; Liu et al. 2017; Hazen and Techtmann 2018: 1–18). To clarify the impact of added oil on phylogenetic diversity of alkB genes, we then performed alkB diversity analysis on oil-amended samples. In the marine samples, a decrease in Shannon diversity of the alkB genes was observed in the oil-amended samples (Fig. 1). However, there were very few oil-amended marine samples examined, which prohibited our ability to test if this trend was significant. On average the oil-amended microcosms had a higher alkB Shannon diversity than the controls (Bakken average = 3.603, Dilbit average = 3.655, Control average = 3.40). The increase in diversity upon oil addition to freshwater samples observed in this study is in contrast to the impact of oil on taxonomic community diversity in freshwater, which showed no significant impact of 16S rRNA diversity in these same samples (Butler et al. in preparation). In the marine environment previous studies have shown that taxonomic diversity often decreases upon oil addition as there is a selection for and dominance by a subset of microbes specialized in oil biodegradation (Mason et al. 2012: 1715–27; Liu et al. 2017).
The difference between no-oil and oiled community composition is clearly evident in the PCoA analysis of freshwater communities, which indicated that oil-amended samples cluster distinctly from the no-oil controls. While there was a significant difference in the community composition between no-oil and oil-amended samples, there was no significant difference in the alkB community composition between the two oil types (Pairwise PERMANOVA P value = 0.130) (Table S4, Supporting Information).
These results combine to demonstrate that there is a high level of diversity of alkB in uncontaminated freshwater and marine systems. When oil was amended to the freshwater samples, the composition of alkB genes changed from the baseline alkB gene composition. To better understand the impact of oil on alkB diversity in freshwater samples, we performed differential abundance analysis. We found that 75 alkB OTUs were significantly enriched in the oil amended treatments compared to five alkB OTUs being more highly abundant in the control treatment (Fig. S3, Supporting Information and supplementary dataset). The majority of OTUs enriched in oil were unclassified, however a number of Actinobacterial OTUs and Betaproteobacterial OTUs from the order Burkholderiales were enriched in the presence of oil. While the composition of the marine OTUs appeared not to change based on PCoA analysis, the alpha diversity was different between oil and non-oil samples. These findings are in line with previous work that shows that in both the marine and freshwater settings there is a distinct set of oil-degrading bacteria that were present at low levels in uncontaminated water and bloom in response to released oil (Brakstad et al. 2008: 540–52). Previous studies have shown a distinction in the diversity and abundance of alkB genes detected in plume and non-plume samples from the Deepwater Horizon oil spill (Lu et al. 2012: 451–60). Furthermore, studies performed in uncontaminated water have shown diverse sets of alkB genes in marine and brackish environments (Smith et al. 2013; Viggor et al. 2015: 507–16). Only two of the oil-amended marine samples resulted in sufficient number of sequences for analysis. This small sample size limited our ability to draw conclusions about the impact of oil on alkB diversity in the marine environment. The limited replication and the difference in microcosm set up also hindered our ability to compare the response to oil between freshwater and marine oil-amended samples.
Variation in alkB diversity by depth
Previous studies have indicated a stratification of taxonomic diversity with depth in the marine environment (Costello and Chaudhary 2017: R511-R27). 16S rRNA gene diversity has typically been reported highest in the shallower depths and decreasing with depth (Sanz-Sáez et al. 2019: 774992). Furthermore, the biological pump concept suggests that more labile carbon will be used in the surface regions and more recalcitrant carbon being enriched in the depths (Cavan et al. 2019). Based on these findings, we hypothesized that depth would impact the diversity of alkB genes in oceans. However, our analysis of Shannon diversity indicated little change in the Shannon diversity with depth. There was a minor, but non-significant increase in alkB Shannon diversity with depth (P = 0.412) (Fig. 5A). However, principal coordinate analysis (Fig. 5B) showed that samples from the same depths cluster together, with a large distinction between the epipelagic and the deeper depths. While less substantial, there is clustering of the upper bathypelagic samples from the deeper bathypelagic samples. PERMANOVA analysis indicates that there is a significant difference in alkB composition between the depth zones (P-value = 0.02, R2 0.3224, degrees of freedom: 2). Pairwise comparison indicates significant differences between the Epipelagic and Mesopelagic (FDR adjusted P value = 0.003) as well as the Epipelagic and the Bathypelagic (FDR adjusted P value = 0.0465).
(A), Plot of Shannon diversity of marine samples with available depth data. (B), Principal coordinate analysis of marine samples with available depth data. Color shade indicates depth (m).
The finding of stratification in alkB diversity in the Sargasso Sea is in contrast to the findings of Smith et al. (2013) (Smith et al. 2013). Smith et al. found that there was not a distinction in alkB diversity as a function of depth in the northern Gulf of Mexico. Their finding was in contrast to the stratification of the taxonomic diversity of microbial community in the northern Gulf of Mexico (Tolar, King and Hollibaugh 2013). The difference between our data and Smith et al. may be due to differences in the depth of samples collected. In Smith et al. all of the samples were collected in the epipelagic and mesopelagic zones at depths of less than 900 m. Whereas all of the deep-water samples collected from this study were at greater than 1000 m depths. Our data exemplifies alkB communities that organize more strongly by depth than sampling location in these marine samples (Fig. 5B). In particular our data indicate that the alkB diversity in the surface water samples (40 m) is dominated by Betaproteobacteria (average relative abundance 74% ± 20.5%) whereas Betaproteobacteria were a relatively low proportion of the deep-water (>1000 m) marine communities (6.9% ± 5.8%) (Fig. 4). Alternatively, the Gammaproteobacteria were more highly abundant in deep-water samples relative to the shallow samples (30% ± 26% in deep-water compared to 9.8% ± 9.1%).
CONCLUSION
The potential for microbes in the environment to assist in the bioremediation of oil has sparked much research into the diversity of oil degrading microbes. Here we compare the diversity of alkB genes in freshwater and marine environments. Our principal hypothesis in this study was that the complement of alkB genes present in freshwater would be distinct from those found in marine environments and that freshwater alkB display a distinct evolutionary history. Our results indicate there is a diverse set of alkB genes that are present in these distinct environments. There is a clear phylogenetic distinction in the complement of alkB genes present in freshwater and marine environments, which suggests a different evolutionary history of these genes between these environments. Furthermore, our results indicate that in freshwater systems the addition of oil selects for a distinct set of alkB gene compared to the non-contaminated water (pairwise PERMANOVA P value = 0.002) (Table S5, Supporting Information). This finding suggests that the there is a subset of alkB-encoding microbes that are most impacted by oil addition and may be specialized for oil degradation relative to the ambient set of alkB-encoding microbes. Finally, our results suggest that in the Sargasso Sea there is a stratification in the diversity of alkBs genes by depth. This suggests that a distinct set of alkB-encoding microbes are present at different depths in the water column. The diversity of hydrocarbon degrading genes has the potential to serve as a marker for oil biodegradation potential. This work helps to expand our understanding of the diversity of alkB genes and the diversity and distribution of hydrocarbon degrading microbes in natural settings.
Conflicts of Interest
None declared.
