Abstract

Introduction

Sulfolane is an anthropogenic organo-sulfur molecule used in some oil and natural gas refineries, resulting in contamination of groundwater at industrial sites around the world^1–3, including in North Pole, Alaska, where it has contaminated hundreds of private drinking water wells. Despite its emerging importance as a groundwater contaminant, little is known about the environmental fate of sulfolane. Sulfolane biodegradation potential exists in activated sludge, contaminated aquifer substrate, and pristine soil, but the identity of the microorganism(s) responsible remains largely unknown^4–6. Three sulfolane degraders have been previously identified through pure-culture-based studies, with mixed enrichment cultures reportedly degrading sulfolane more quickly than pure cultures^7–9. It remains unknown how diverse, widespread, or abundant sulfolane degraders are in the environment, particularly in contaminated aquifers, where this information is valuable in assessing plume longevity and identifying remediation strategies, including natural attenuation and accelerated bioremediation.

Molecular techniques like stable isotope probing (SIP) are powerful tools for examining the active members from environmental microbial communities involved in the biodegradation of emerging organic contaminants such as sulfolane. ¹³C-based SIP is the process of exposing a microbial community to a chemical compound highly enriched in ¹³C, which otherwise accounts for roughly 1% of all carbon. The microorganisms that metabolize the ¹³C-labeled substrate will incorporate the heavy isotope into their biomolecules¹⁰. Analyzing the ¹³C-enriched DNA enables the identification of functionally relevant community members through DNA sequencing approaches including 16S rRNA gene amplicon and shotgun-metagenomic sequencing. Shotgun metagenomic sequencing can also shed light on the functional capabilities of the active organisms and identify metabolic pathways potentially being utilized^11,12.

To date, the environmental microorganisms known to be capable of degrading sulfolane to date have all been isolated from temperate regions. They include a Shinella sp. from Okinawa Japan, a Variovorax sp. from Alberta Canada, and a Pseudomonas sp. from Illinois USA^7–9. To our knowledge, the identity of sulfolane degrading microbes in a subarctic aquifer, like that found in North Pole, Alaska, have not previously been reported. We performed DNA-SIP with ¹³C-labeled sulfolane in contaminated North Pole, Alaska, aquifer substrate to elucidate the identity of subarctic sulfolane degraders while circumventing culture bias. We combined DNA-SIP with shotgun metagenomics to taxonomically identify microbes involved in sulfolane degradation as well as to gain insights into their genetic potential and possible degradation pathways that may be used to process sulfolane. This study demonstrates that, by combining these techniques, it is possible to not only identify but also obtain high-quality draft genomes of unknown emerging contaminant degraders from environmental samples.

Results

Isolation of ¹³C-labeled DNA

Quantitive PCR results showed a clear separation between the heavy and light DNA in the density gradient (Fig. 1A). In addition, there was an increase in the relative abundance of labeled DNA over the course of the incubation in the microcosms amended with labeled sulfolane. That indicated the labeled carbon was being assimilated by members of the microbial community. As expected, there was no quantifiable heavy (¹³C-labeled) DNA in any of the control microcosms amended with ¹²C-sulfolane (Supplementary Dataset ¹).

Open in a separate window

Figure 1

(A) Quantitative PCR results showing relative abundance of 16S rRNA gene copies in density gradient fractions after separation of the labeled and unlabeled DNA in representative microcosms amended with ¹³C-labeled sulfolane. The ¹³C-labeled fractions increase in relative abundance over time demonstrating the incorporation of ¹³C into prokaryotic DNA as sulfolane is biodegraded. (B) Sulfolane loss over time in SIP microcosms. Dashed arrows indicate when triplicate microcosms were destructively harvested. Error bars represent standard deviation.

Microbial Community Analysis

Statistical analyses showed that the microbial communities in the SIP “heavy” fractions, SIP “light” fractions, and ¹²C-sulfolane controls were all significantly different from each other (MRPP, p<0.05). These groups were also significantly different from the time-zero total community samples with the largest difference between T0 and the ¹³C-labeled communities (A=0.66, p0.05). However, once a community shift occurred and sulfolane degradation was initiated, which occurred by day 28, there was no significant change in the microbial community profile over the course of the incubation within the SIP “heavy” fractions, SIP “light” fractions, and ¹²C-sulfolane controls. Since there was no difference between T1, T2, and T3 communities (Fig. 1B), we combined the three timepoints from each treatment group into one representative sample for community analysis.

The microbial communities associated with the ¹³C-labeled DNA fractions were very low in richness (Chao1 65.6±27.8) and diversity (Inverse Simpson index 1.4±0.3) with only one OTU (OTU1) comprising 85.7±8.7% of the total labeled microbial communities (Fig. 2). A BLAST comparison of the 253-bp partial 16S rRNA gene sequence for OTU1 showed this gene fragment was 99% identical to five different species from three genera of the Comamonadaceae family (Rhodoferax ferrireducens, Rhodoferax saidenbachensis, Limnohabitans parvus, Acidovorax facilis and Acidovorax radicis). Analysis of the full 16S rRNA gene obtained via shotgun sequencing uniquely identified OTU1 as a Rhodoferax sp., as detailed below. The next two most abundant community members were Lysobacter sp. and Bacteriovorax sp., which, when combined, comprised less than 3% of the total labeled community (1.3±2.0% and 1.1±0.9% respectively). The dominant phylotype in the unlabeled or “light” fractions of the ¹³C-sulfolane incubation was OTU15 from the Sphingomonadaceae family, which represented 7.4% of the unlabeled community. OTU1 comprised 4.9% of the light fraction’s total community.

Open in a separate window

Figure 2

Abundance of OTU1 in all 16S rRNA gene amplicon community types. Since no differences were found between community structure over the course of incubation, all replicates from each timepoint (28, 32, and 36 days) were averaged together for the control (n=7), unlabeled fraction (n=9), and labeled fraction (n=9) communities. Time zero represents the subarctic aquifer substrate prior to sulfolane exposure (n=3). “Control” refers to the community incubated with unlabeled sulfolane. Error bars represent standard deviation from the mean.

Although OTU1 (presumably Rhodoferax sp.) was uniquely dominant in the ¹³C-labeled microbial community, it was not the only dominant organism detected in the total community analysis of the control microcosms that were exposed to ¹²C-sulfolane. In the ¹²C-control cultures, OTU1 was co-dominant with OTU3 (identified as Arthrobacter sp.), which represented 23.2±3.4% of the total microbial community, while OTU1 constituted 21.3±6.0%. OTU3 was also significantly more abundant than OTU1 in the starting community (paired t-test, df=2, t=−8.28, p=0.014). Even though OTU3 was a dominant member of the total community in ¹²C-sulfolane control cultures, it was ruled out as a sulfolane degrader due to its lack of incorporation of ¹³C from sulfolane into DNA.

As expected, the microbial community showed more OTU richness and diversity prior to sulfolane exposure (Chao1 2209.0±89.1, Inverse Simpson 140.2±12.7). The dominant ¹³C-enriched OTU1 was the 18^th most abundant phylotype in the T0 total community and comprised less than 0.1% of the microbial community prior to sulfolane exposure and incubation. The two most abundant T0 total community phylotypes were identified as Geobacter spp. and represented 4.2% and 3% of the total community, respectively. These OTUs were not found in the SIP “heavy” community. Although there was a small subset of archaea in the T0 microbial community, no labeled archaeal DNA was detected in the sequence data.

Metagenomic analyses

By shotgun sequencing the low-diversity heavy SIP fractions we were able to obtain a high-quality draft metagenome assembled genome (MAG) of the putative sulfolane-degrading microorganism. A total of 2.9M paired-end 2×250bp reads (1.5 Gbp) remained after quality control and trimming, which produced an assembly containing 12,437 nodes and a total length of 14 Mbp. A single large connected component made up 49% of the total size (6.9 Mbp) with a mean sequencing depth of 53x (Supplementary Fig. ¹). This component was composed of 395 contigs with an N50 of 146 kbp and a longest contig of 614 kbp. An additional 262 kbp were contained in 41 contigs on 5 additional connected components, of which two were closed circular plasmids and one was complete but not circular (Supplementary Fig. ²). The average length of the 11,998 unconnected contigs was 572bp; the mean depth of these contigs was 0.7x. The 395 contigs from the large connected component were defined as a MAG and annotated in JGI-IMG/ER. A total of 6537 protein coding genes were identified; the GC content was 60.84%. CheckM analysis found the genome to be 99.8% complete with only 0.50% contamination (defined as redundancy of putative single-copy genes).

The 16S rRNA gene for the MAG was aligned to the representative sequence of OTU1, the sulfolane-degrading Comamonadacaea sp. identified in the community dataset, using BLAST¹³. The 16S rRNA gene assembled from the metagenome was 100% identical to the amplicon, matching 253 of 253 bases and identifying this draft genome as the genome of the labeled sulfolane degrader. BLAST comparisons of the full 16S rRNA gene sequence from the metagenome-assembled genome identified the sulfolane degrader as a Rhodoferax sp., being 99% identical to Rhodoferax ferrireducens strain T118. A phylogenetic tree based on the 16S rRNA gene is also provided (Fig. 3).

Open in a separate window

Figure 3

Phylogenetic tree of Rhodoferax sp. OTU1 compared to closest type strain bacteria. Tree is based on full-length 16S rRNA gene sequence similarity. SH-aLRT node confidence values≥0.70 are shown.

It has been suggested that sulfolane biodegradation may follow the 4S-pathway described for diobenzothiophene desulfurization⁸. This pathway employs the genes dszA, dszB, and dszC to oxidize the sulfur to sulfite, which abiotically oxidizes to sulfate in aerobic conditions. We searched the MAG for these genes that may be involved with sulfolane degradation. A protein BLAST search found a gene that is homologous to dszA (42% identical, 59% positive, e-value of 10⁻¹⁰⁹) in the genome and a second, distantly related homolog (Supplementary Fig. ³) repeated 5 times on an associated plasmid (Supplementary Fig. ⁴). The MAG does not contain dszB, which cleaves the sulfite moiety to complete the remineralization of sulfur in the 4S-desulfurization pathway^14,15. However, the genome of the Rhodoferax sp. does have 84 genes involved with sulfur metabolism, including complete sulfur oxidation (sox) and alkanesulfonate utilization (ssu) pathways (Supplementary Fig. ⁵). To our knowledge there has only been one other gene, identified using Escherichia coli mutants, proposed to be associated with sulfolane metabolism^16,17. This gene is known as thdF and the MAG of the sulfolane assimilating Rhodoferax sp. does not contain this gene.

Discussion

Using DNA-SIP combined with metagenomics, we have identified a single OTU as the primary sulfolane-degrading organism in subarctic aquifer substrate and provided DNA evidence that strongly suggests it is a member of the Rhodoferax genus (Fig. 3). Of the 178 OTUs detected in the ¹³C-labled SIP fractions, OTU1 was identified as the dominant microorganism incorporating carbon from sulfolane in this subarctic aquifer substrate. Although initial analysis of the 253-bp amplicon data revealed OTU1 as a member of the Comamonadaceae family, the full 16S rRNA gene provided more a more rigorous genus-level taxonomic resolution. Phylogenetic analysis showed that OTU1 was 99% similar to type strain Rhodoferax ferrireducens T118 (Fig. 3), which was isolated from subsurface sediments collected in Oyster Bay, VA, USA. This relationship was also supported by whole genome database comparisons. When non-type strains were included in the phylogenetic analysis, most of the closest relatives were found in contaminated soil or groundwater with the top 20 originating from freshwater or terrestrial environments (Fig. 4). Although this is the first report of a Rhodoferax sp. degrading sulfolane, it is not surprising since members of the Rhodoferax genus are commonly found in contaminated freshwater environments and have been implicated in the degradation of other contaminants including herbicides, naphthalene, and benzene^18–20.

Open in a separate window

Figure 4

Phylogenetic tree of Rhodoferax sp. OTU1 and closest related non-type strain bacteria based on full-length 16S rRNA gene similarity. All non-type sequences represented here are >99% similar to Rhodoferax sp. OTU1. SH-aLRT node confidence values≥0.70 are shown. Rhodoferax ferrireducens is in bold for reference.

Prior to identifying Rhodoferax sp. OTU1 as a sulfolane degrader, the only environmental microorganisms known to degrade sulfolane originated from subtropical, humid subtropical, or continental climates. In Western Canada a sulfolane degrader was isolated from a contaminated aquifer and identified as a Variovorax sp. which is in the same family (Comamonadaceae) as Rhodoferax sp. OTU1⁸. Two other environmental isolates have been reported to degrade sulfolane. Pseudomonas maltophilia was isolated from the soil of an abandoned strip mine near Cambria, Illinois; USA and grew on sulfolane as the sole carbon source⁷. A novel Shinella sp. was isolated from soil in the Yambaru area of Okinawa Main Island; Japan and grew on sulfolane as the sole sulfur source⁹. To our knowledge, the only other report of a sulfolane-degrading bacterial species was a mutated strain of Escherichia coli, which was not isolated from the environment but did yield insights into sulfone degradation pathways¹⁶. A mutation in the thdA gene allowed this E. coli strain to degrade sulfolane via a novel sulfone oxidase enzyme. The authors proposed thdA to be a regulator gene for several genes involved in the metabolism of organo-sulfur compounds, including thdF thiophene oxidase¹⁶. The novel sulfone oxidase was never identified and thdA has not been sequenced, but the genetic sequence for thdF has been published¹⁷. We did not find thdF homologs in the MAG of the degrader, but that does not rule out the unidentified sulfone oxidase as being involved in sulfolane metabolism.

It has been proposed that the biodegradation of sulfolane followed the 4S-desulfurization pathway for dibenzothiophene due to structural relatedness of the compounds and the production of sulfate as a mineralization product⁸. This pathway involves the use of the dsz operon involving genes dszA, dszB, and dszC²¹. The gene product of dszA opens the ring structure after dibenzothiophene is converted to dibenzothiophene-5,5-dioxide by dszC¹⁵. We found no evidence of dszB being present in the MAG for the degrader, suggesting that sulfolane biodegradation in this strain does not utilize the 4S-desulfurization pathway. However, the MAG does encode a homolog to dszA at the end of an alkanesulfonate utilization pathway (Fig. 5). In addition, 5 copies of another dszA homolog are present on an IncP-family plasmid in the metagenomic co-assembly (Supplementary Figs ²–⁴). This plasmid has a copy number of about 3 relative to the chromosome, suggesting that the plasmid-borne dszA homolog is present at 15 times the copy number of the genomic dszA homolog. This is perhaps not surprising considering that the dibenzothiophene pathway is also generally plasmid-borne (Denis-Larose et al. 1997, Denome et al. 1994). If either of these DszA homologs can act on sulfolane as DszA does to dibenzothiophene, the remaining compound would be 4-hydroxy-butane sulfinic acid. Under aerobic conditions sulfinic acids can oxidize to sulfonic acids, with aliphatic sulfinic acids being more reactive than aromatic ones²². Although speculative, it is feasible that the resulting product (4-hydroxy-butane-sulfonic acid) is then degraded in a similar fashion to other alkanesulfonates and/or taurine, which this species is genetically equipped to process (Fig. 6, Supplementary Fig. ⁵).

Open in a separate window

Figure 5

Order of alkanesulfonate metabolism genes in Rhodoferax sp. OTU1 and Rhodoferax ferrireducens strain T118. Open arrows indicate shared ssu genes. The checkered arrow is the monooxygenase ssuD which is present in this operon in strain T118 but not the sulfolane assimilating species. However, a homolog to SsuD is elsewhere in the genome of the sulfolane-assimilating Rhodoferax sp. The dszA homolog is present in both species at the end of the ssu operon.

Open in a separate window

Figure 6

Schematic of proposed sulfolane biodegradation pathway. Protein identifications and Kegg Orthology values are as follows: Sulfonate transport system substrate-binding protein (SsuA; K15553), Sulfonate transport system ATP-binding protein (SsuB; K15555), Sulfonate transport system permease protein (SsuC; K15554), Dibenzothiophene sulfone monooxygenase (DszA; K22220). The flavin reductase refers to either SsuE (flavin reductase; K00299) or DszD (Genbank Accession AB051429.1; no KO or E.C. values available), which have coding regions in close proximity to dszA on the chromosome and plasmid respectively.

Greene et al.⁸ previously found that mixed cultures are more efficient at degrading sulfolane than isolates, but it was unclear if this was due to the exchange of specific nutrients between organisms, the removal of growth-inhibiting products, the combined metabolic attack on the substrate, or some combination. We were surprised to find no compelling evidence of sulfolane assimilation in any species other than Rhodoferax sp. OTU1 throughout the course of this labeling study. Although there were other labeled microbes in this community, they were in extremely low abundance and were likely labeled due to the scavenging of labeled biomolecules²³. In addition, the ¹³C-labeled community structure did not change substantially over the entire SIP incubation period, which might be expected if sulfolane metabolites were being degraded in series. Although Rhodoferax sp. OTU1 represented ~25% of the community in the control (unlabeled sulfolane) microcosms, it was co-dominant with an Arthrobacter sp. Despite the abundance of the Arthrobacter sp. in the controls, we found no labeled Arthrobacter DNA, which strongly suggests it is not involved in sulfolane biodegradation. It is likely this species is an opportunistic, heterotrophic, soil-bacterium that can tolerate the experimental sulfolane concentrations allowing it to become co-dominant with the Rhodoferax sp. However, if the Arthrobacter sp. was degrading sulfolane as a method of detoxification or energy production and not assimilating the carbon, we would not identify it as a degrader using these methods. We are currently working to isolate Rhodoferax sp. OTU1 into pure culture in order to enable definitive characterization of the sulfolane degradation pathway.

Conclusion

We have identified a Rhodoferax sp. as being the dominant and likely the exclusive sulfolane-degrading microorganism in enrichment cultures from contaminated subarctic aquifer substrate. By combining ¹³C-DNA-SIP with 16S rRNA gene amplicon sequencing and shotgun metagenomics, we were able to not only resolve the taxonomy of this degrader, but also gain insights into how it may be metabolizing sulfolane. We also suggest that sulfolane degradation by this Rhodoferax strain does not proceed following the previously proposed model⁸. Although an Arthrobacter sp. comprised ~25% of the total community in the ¹²C-sulfolane controls during active sulfolane biodegradation, it was not assimilating sulfolane carbon into its DNA and therefore likely not involved in the biodegradation process. We caution that simply analyzing changes in the microbial community profile during sulfolane biodegradation studies is not sufficient to determine species function and may be misleading. Although this study cannot confirm that Rhodoferax sp. OTU1 actively degrades sulfolane within the contaminated aquifer, it enables further studies of biodegradation potential in situ, such as through characterizing the environmental distribution of this organism and identifying environmental factors driving its abundance.

Methods

Subsurface samples used as inoculum for SIP studies were collected from Flint Hills Resources property located in North Pole, Alaska (64.7511°N, 147.3519°W) where groundwater sulfolane levels ranged from 0 to 34.8mgL⁻¹ at the time of sample collection²⁴. The sediment and groundwater samples used in this study came from a sulfolane-contaminated subarctic aquifer and a detailed site characterization and history of sulfolane use can be found in Kasanke and Leigh³. The aquifer sediment used in this study consisted of augured material from the installation of a new monitoring well at depths between 3 and 9m below ground surface. All sediment was sieved through a 2-mm screen prior to use. Groundwater used in this study came from an existing monitoring well approximately 30m from where sediment was collected. The well was screened 18.25m below the ground surface and has stable historical sulfolane concentrations of approximately 125µgL⁻¹²⁴. Groundwater was collected in September 2012 and sediment in March 2013. Both were stored at 4°C (the approximate average aquifer temperature) until use in February 2016. The top of the water table at time of sampling was 3m below ground surface and the aquifer has an average temperature of 3.4°C²⁴.

Supplementary information

Acknowledgements

Author Contributions

Conceptualization, C.P.K., E.R.C. and M.B.L.; Methodology, C.P.K. and E.R.C.; Resources, M.B.L. and E.R.C.; Writing, C.P.K., E.R.C. and M.B.L.

Notes

Footnotes

References