Based on the transient exposure of Chesapeake Bay sediments to hydrocarbons and the metabolic versatility of known anaerobic alkane-degrading microorganisms, it was hypothesized that distinct Bay sediment communities, governed by geochemical gradients, would have intrinsic alkane-utilizing potential under sulfate-reducing and/or methanogenic conditions. Sediment cores were collected along a transect of the Bay. Community DNA was interrogated via pyrosequencing of 16S rRNA genes, PCR of anaerobic hydrocarbon activation genes, and qPCR of 16S rRNA genes and genes involved in sulfate reduction/methanogenesis. Site sediments were used to establish microcosms amended with n-hexadecane under sulfate-reducing and methanogenic conditions. Sequencing of 16S rRNA genes indicated that sediments associated with hypoxic water columns contained significantly greater proportions of Bacteria and Archaea consistent with syntrophic degradation of organic matter and methanogenesis compared to less reduced sediments. Microbial taxa frequently associated with hydrocarbon-degrading communities were found throughout the Bay, and the genetic potential for hydrocarbon metabolism was demonstrated via the detection of benzyl-(bssA) and alkylsuccinate synthase (assA) genes. Although microcosm studies did not indicate sulfidogenic alkane degradation, the data suggested that methanogenic conversion of alkanes was occurring. These findings highlight the potential role that anaerobic microorganisms could play in the bioremediation of hydrocarbons in the Bay.
INTRODUCTION
Petroleum hydrocarbons are frequently released into marine environments via natural seeps, as well as anthropogenic activities including crude oil extraction, transport, storage and refining processes (National Research Council 2003). An estimated 1.3 × 106 metric tons of petroleum enter marine systems each year, of which approximately 55% are attributable to anthropogenic sources (National Research Council 2003). The scales of different pollution events can vary dramatically, resulting in variable impacts on marine ecosystems. This was well illustrated in the Gulf of Mexico by the blowout of the Macondo 252 well and the subsequent sinking of the Deepwater Horizon, which resulted in an unprecedented amount of crude oil being released (∼4.1 to 4.4 million barrels) (Crone and Tolstoy 2010; Operational Science Advisory Team 2010). Although a significant proportion of the Macondo 252 oil was removed through human intervention or physical processes (78%), the remainder had a fate classified as ‘other’, suggesting that some of the oil and gas may have been removed via microbially mediated processes (Ramseur 2010). Subsequent studies investigating microbial communities in the Gulf of Mexico water column, deep-sea sediments and coastal sediments have provided overwhelming evidence that the microbial community played an important role in the removal of the oil (for reviews, see Joye, Teske and Kostka 2014; Kimes et al., 2014; King et al., 2015). These events and the initial devastation of the Deepwater Horizon spill prompted immediate discussion about oil spill assessment and preparedness, especially for delicate and economically important ecosystems, such as the Chesapeake Bay (Behn 2010).
The Chesapeake Bay is the largest estuary in the United States, with a watershed encompassing 165 000 km2 of forest and woodland (64%), agricultural land (24%) and urban areas (8%) (Paolisso et al., 2013). More than 100 000 rivers and streams drain into the Chesapeake Bay (Chesapeake Bay Program 2014a). The Bay has a larger land-to-water ratio than any other coastal body in the world (Chesapeake Bay Program 2014a). This, along with the extensive dendritic shoreline (18 800 km) (Kemp et al., 2005) and low flushing rates (i.e. flushing time is approximately 200 days) (Fisher et al., 1988), makes the Bay vulnerable to high nutrient loading and other types of contamination. As a result of nitrogen (N) and phosphorous (P) loading, the Bay has suffered from increased phytoplankton abundance, declining water clarity, depletion of bottom-water oxygen, redox changes in sediment biogeochemistry, decreases in benthic microalgal primary production and loss of benthic macroinfauna, loss of oyster beds and benthic filtration, major shifts in fish populations, loss of seagrasses and other submersed vascular plants, and loss of tidal marshes as nutrient buffers (for review, see Kemp et al., 2005). The Bay has also suffered from pollution with metals, polychlorinated biphenyls (PCBs) and hydrocarbons (Chesapeake Bay Program 2014b). From both an ecological and economic perspective, the Chesapeake Bay is of significant value. Beyond commercial fishing, it was estimated in 2001 that for persons living in parts of Virginia, Maryland and the District of Columbia, the annual benefits of the Bay ranged from $357.9 million to $1.8 billion based on (1) recreation (fishing, boating and swimming); (2) health; (3) property values; (4) regional economic impacts; and (5) non-use value (Morgan and Owens 2001). Despite restoration and mitigation efforts, the Bay is still at a continual risk for hydrocarbon contamination (and other types of pollution) via commercial shipping, recreational boating activity and urban inputs.
Typically, the major hydrocarbon inputs to the Bay are urban runoff (Foster et al., 2000) and atmospheric deposition (Webber 1983). Concern about a large hydrocarbon spill event emerged in the 1970s due to the proposed construction of superports for oil tankers. This prompted several investigations of the Bay's microbial potential for degradation of petroleum and petroleum compounds (Walker and Colwell 1973; Walker, Colwell and Petrakis 1976a; Walker, Petrakis and Colwell 1976b; Okpokwasili et al., 1984; West et al., 1984). By the 1990s, the importance of this research was self-evident. There were 3651 oil spill events (each spill >75 gallons) in Chesapeake Bay between 1985 and 1994, which led to an estimated release of more than 1.3 × 106 gallons of oil (Balch 1997). In 2010, the Bay suffered one of its worst oil spills when 140 000 gallons of oil were spilt into the Patuxent River as a result of a ruptured underground pipeline (Michel et al., 2009). Due to the transient, but continual, exposure to hydrocarbons over several decades, hydrocarbons are measureable in Bay bulk water and the aquatic surface microlayer (e.g. alkane concentrations ranging from 3.16 ± 0.77 to > 200 μg L−1) (Hardy et al., 1990), as well as sediments (Walker, Colwell and Petrakis 1976a; Walker, Petrakis and Colwell 1976b; Arzayus, Dickhut and Canuel 2001). Accordingly, microbial studies have demonstrated the enrichment of petroleum hydrocarbon and PAH-degrading bacteria from Chesapeake Bay water and sediment (Walker, Colwell and Petrakis 1976a; West et al., 1984), the impact of different refined fuels and crude oils on the growth of microbial populations enriched from Chesapeake Bay water (Walker et al., 1976b), and the effect that prior oil exposure has on the number of cultivable petroleum-degrading microorganisms enriched from Chesapeake Bay water and sediment (Walker and Colwell 1973). All of these prior studies, however, were conducted under aerobic conditions because anaerobic degradation of hydrocarbons was not well described or very well understood at the time. However, research during the last 25 years has unveiled novel microbial strategies for the anaerobic activation and degradation of hydrocarbons (for review, see Heider and Schühle 2013), which are particularly important in sediments impacted by petroleum compounds where oxygen can be rapidly depleted. Among these strategies is the addition of aliphatic and aromatic hydrocarbons to the double bond of fumarate (i.e. ‘fumarate addition’), which is catalyzed by the glycyl radical enzymes alkylsuccinate synthase (ASS)/methylalkylsuccinate synthase (MAS) (Callaghan et al., 2008; Grundmann et al., 2008) and benzylsuccinate synthase (BSS) (Leuthner et al., 1998), respectively. As such, genes encoding the catalytic subunits of BSS and ASS (bssA and assA) serve as potential biomarkers for ‘fumarate addition’ in anaerobic hydrocarbon-impacted environments (Callaghan et al., 2010; Agrawal and Gieg 2013; Callaghan 2013).
To our knowledge, the intrinsic capacity of Bay sediment microbial communities to mediate anaerobic hydrocarbon transformation has not been investigated. In the event of an oil spill, the shallow depth of the Chesapeake Bay would likely play an important role in the transport of hydrocarbons to Bay sediments. Therefore, in the wake of the Deepwater Horizon oil spill, we took advantage of a cruise of opportunity to assess the potential for anaerobic hydrocarbon degradation in Chesapeake Bay sediments via next-generation sequencing of 16S ribosomal RNA genes, molecular surveys of functional genes for anaerobic degradation pathways and microcosm experiments. Specifically, we focused on the anaerobic conversion of n-hexadecane due to the relevance of alkanes as crude oil pollutants. Based on the transient exposure of Chesapeake Bay sediments to hydrocarbons and the metabolic versatility of known anaerobic alkane-degrading microorganisms, it was hypothesized that distinct Bay sediment communities, governed by geochemical gradients, would have the potential for alkane-degrading activity under sulfate-reducing and/or methanogenic conditions.
MATERIALS AND METHODS
Sampling sites and sample collection
Samples were collected aboard the R/V Hugh R. Sharp during a transect cruise of Chesapeake Bay in August 2010. Bay oxygen concentration data for 2009 were obtained from the Chesapeake Bay Program (CBP) Water Quality Database and used as an a priori guide for site selection. Four sites were then chosen based on the presence or absence of bottom anoxia during the 2010 sampling (Table S1 and Figs S1–S3, Supporting Information). Water column oxygen concentrations during our cruise were monitored using the onboard CTD device (Note: cruise track and CTD data are available via the Biological and Chemical Oceanography Data Management Office via dataset number: HRS100808BW). Sediments were obtained by gravity coring, and core liners were immediately sectioned (1-ft intervals), capped and moved to the on-board lab. Each 1-ft section is referred to as a ‘horizon’ hereafter. A piece of the core liner was removed from the middle of each horizon, and core material was immediately sampled for enrichment studies, pore water analysis and community DNA extraction.
Sediment pore water analysis
Sediment pore water from each station horizon was obtained using a titanium pore water squeeze cell (GEOTEK, Daventry, UK) in a 5-ton manual hydraulic press. Several cubic centimeters of core material were removed from the center of horizon core for this analysis. A total of 5 mL of pore water was collected from each station horizon, placed in cryovials, immediately frozen in liquid nitrogen and stored at −80°C. Sulfate concentrations were determined in triplicate using a Dionex ICS-1000 ion chromatograph equipped with an IonPac AS4A-SC anion exchange column and a conductivity detector (Dionex, Sunnydale, CA, USA). Samples were pre-filtered through a 0.22 μm membrane disk filter to remove particulate matter and diluted 10-fold in deionized water prior to analysis. The eluent contained 1.8 mM Na2CO3 and 1.7 mM NaHCO3, and the flow rate was 2 mL min−1.
For methane analysis, triplicate sediment samples (ca. 3 cm3) were collected from each station horizon and immediately placed into 10-mL serum bottles containing 6 mL of 3.7% filter-sterilized formaldehyde to halt microbial activity. Bottles were immediately capped with butyl rubber stoppers and stored at 4°C for transport back to the laboratory. Methane was analyzed using a Varian 3300 gas chromatograph equipped with a Porapak Q 80/100 column and a flame ionization detector using helium as the carrier at a flow rate of 20 mL min−1. The injector, column and detector temperatures were held at 100, 100 and 125°C, respectively. Methane concentrations were determined using the ideal gas law equation (PV = nRT) and measuring the amount of methane in the headspace, the culture volume, the headspace volume and the headspace pressure.
DNA extraction
Triplicate sediment samples (ca. 1 cm3) were collected from each station horizon, placed into MoBio Powersoil® Bead Tubes (MoBio, Carlsbad, CA, USA), frozen in liquid nitrogen and stored at −80°C until extraction. Total genomic DNA was extracted using the MoBio Powersoil® Kit (MoBio, Carlsbad, CA, USA) according to the manufacturer's instructions, and DNA concentrations were quantified using a Qubit 2.0 Fluorometer and Quant-iT dsDNA BR Assay Kit (Life Technologies, Carlsbad, CA, USA).
Quantitative PCR
The numbers of bacterial and archaeal 16S ribosomal RNA gene copies per gram of wet sediment were quantified in triplicate via SYBR Green-based quantitative PCR (qPCR). Bacterial primers 27F (5′-AGAGTTTGATCMTGGCTCAG-3′ (Nakatsu and Marsh 2007) and 519R (5′-GWATTACCGCGGCKGCTG-3′) (Turner et al., 1999) and archaeal primers A344F (5′-ACGGGGIGCAGCAGGCGCGA-3′) (Nakatsu and Marsh 2007) and A533R (5′-ATTACCGCGGCTGCTGG-3′) (Weisburg et al., 1991) were used for amplification. Reactions were performed in 30-μL volumes containing 15 μL of 2X Power SYBR Green PCR Master Mix (Life Technologies, Carlsbad, CA, USA), 125 nM of each primer and 2 μL of template DNA (1:15 dilution). Thermocycler conditions were as follows: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 55°C for 1 min and 72°C for 1 min. Reactions were carried out in a 7300 Real Time PCR Machine (Life Technologies, Carlsbad, CA, USA). DNA from Desulfococcus oleovorans strain Hxd3 and Methanospirillum hungatei strain JF-1 served as standards.
The abundances of sulfate-reducing microorganisms and methanogens were estimated in triplicate by determining the number of gene copies per gram of wet sediment for each horizon by quantification of dsrA (dissimilatory sulfite reductase) and mcrA (methyl-coenzyme M reductase) genes, respectively. Amplification of dsrA genes was carried out using dsr1F (5′-ACSCACTGGAAGCACG-3′) and dsrQ2r (5′-GTTGAYACGCATGGTRTG-3′) primers (Chin et al., 2008) [Note: a recent study by Müller et al. (2014) established a publically available dsrAB/DsrAB database and a set of recommended primers for ecological investigations]. Amplification of mcrA genes was conducted using forward primers ME3MF (5′-ATGTCNGGTGGHGTMGGSTTYAC-3′) and ME3MFe’ (5′-ATGAGCGGTGGTGTCGGTTTCAC-3′) and reverse primer Me2r’ (5′-TCATBGCRTAGTTDGGRTAGT-3′) as described by Nunoura et al. (2008). Reaction volumes were 30 μL and contained 15 μL of 2X Power SYBR Green PCR Master Mix (Applied Biosystems, Carlsbad, CA, USA), 250 nM of each primer and 2 μL of template DNA (1:15 dilution). Thermocycler conditions were as follows: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 52°C for 1 min and 72°C for 1 min. Reactions were carried out in a 7300 Real Time PCR Machine (Life Technologies, Carlsbad, CA, USA). Plasmid DNA obtained from Chesapeake Bay dsrA and mcrA clone libraries was used to generate standards in qPCR reactions. These clones were generated from respective dsrA and mcrA PCR products using the TOPO TA Cloning Kit with pCR®4 TOPO vector (Life Technologies, Carlsbad, CA, USA) as recommended by the manufacturer. Inserts were sequenced to confirm their identities.
Detection of assA/bssA genes
Community DNA from the surface horizons and horizon 6 at station 908 was surveyed via PCR for the presence of genes encoding the catalytic subunits of glycyl radical enzymes associated with the anaerobic activation of alkanes (assA) (Callaghan et al., 2008; Grundmann et al., 2008) and aromatic hydrocarbons (bssA) (Leuthner et al., 1998). Surface horizons were chosen for this analysis based on the hypothesis that the microbial communities in surface sediments would serve as the sediment's ‘first responders’ in the event of an oil spill. Horizon 6 at station 908 was selected for further investigation due to its high methane concentration. Nine primer pairs were employed as previously described (Callaghan et al., 2010) (Table S2, Supporting Information) (Note: these primers primarily target assA and have a more limited capacity to detect bssA or nmsA homologs). A touchdown PCR protocol was conducted for 50-μL reaction volumes containing 25 μL of 2X DreamTaq Master Mix (Thermo Fisher Scientific, Waltham, MA, USA), 400 nM of each primer, 5 μL of betaine (5M) and 2 μL of template DNA (1:15 dilution). Thermocycler conditions were as follows: 95°C for 4 min followed by 2 cycles at each annealing temperature (i.e. 95°C for 1 min, 63 to 54°C for 1 min, 72°C for 2 min), 19 cycles at the plateau annealing temperature (53°C) and a final extension step at 72°C for 10 min. For samples that did not yield amplification via the touchdown method, the PCR protocol was conducted under less stringent parameters via gradient PCR (annealing temperatures ranging from 55 to 65°C). Reactions were performed in volumes of 50 μL containing 25 μL of 2X DreamTaq Master Mix (Thermo Fisher Scientific, Waltham, MA, USA), 2 μM of the forward and reverse primer, 1 μL (5 units μL−1) of DreamTaq polymerase (Thermo Fisher Scientific, Waltham, MA, USA) and 2 μL of template (1:15 dilution, 1:5 dilution for station 818). PCR products were cleaned with a QIAquick PCR Purification Kit (Qiagen, Valencia, CA, USA) and cloned into the pCR™-II vector using a Dual Promoter TA Cloning Kit (Life Technologies, Carlsbad, CA, USA) following the manufacturer's instructions, and inserts of the expected size were sequenced. Reads were assembled into OTUs at 97% similarity, and nearest matches for each OTU were determined using BlastX of the NCBI NR database. Resulting OTUs and their closest NCBI matches were translated into protein sequences and aligned with representative AssA and BssA sequences from several well-described strains using Megalign Software (DNASTAR Inc., Madison, WI, USA) and the ClustalW alignment method. Neighbor-joining trees were constructed with pairwise deletion and performing 10 000 bootstrap replicates. Pyruvate formate-lyase served as the outgroup for phylogenetic analysis.
Microbial community analysis
The surface horizons for each of the stations and horizon 6 at station 908 were chosen for further analysis for the reasons stated above. The diversity of 16S rRNA genes was assayed in triplicate for each of the selected horizons via pyrosequencing of multiplexed PCR products. Bacterial 16S rRNA genes were amplified using the forward primer 27F (see above) and the reverse primer 338R (5′-TGCTGCCTCCCGTAGGAGT-3′) (Nakatsu and Marsh 2007), producing a 311 bp amplicon. The PCR primers contained 5′ Titanium Fusion adapter sequences (forward primer A-tag: CCATCTCATCCCTGCGTGTCTCCGACTCAG; reverse primer B-tag: CCTATCCCCTGTGTGCCTTGGCAGTCTCAG) as well as a unique 8-nucleotide barcode tag in the reverse primer (Hamady et al., 2008) to allow direct 454 sequencing. Reactions were performed in 50-μL volumes. Reaction mixtures included 0.2 μM of the ‘tagged’ forward primer, 0.25 μM of the reverse primer, 0.25 μL of DreamTaq (5 units μL−1) (Thermo Fisher Scientific, Waltham, MA, USA), PCR Supermix (Life Technologies, Carlsbad, CA, USA) and 2 μL of template DNA (1:15 dilution). Thermocycler conditions for bacterial 16S rRNA genes were as follows: 95°C for 7 min and 30 cycles of 95°C for 20 s, 55°C for 20 s and 72°C for 40 s. Archaeal amplification conditions were identical except that the extension step at 55°C lasted for 60 s. Archaeal 16S rRNA genes were initially amplified using primers A8F (5′-TCCGGTTGATCCTGCC-3′) and A344R (5′-TCGCGCCTGCTGCICCCCGT-3′) producing a 336 bp amplicon that was tagged with Titanium adaptors described above (Nakatsu and Marsh 2007). However, due to inefficient amplification, the protocol was modified, and the archaeal 16S rRNA genes were amplified using A8F and A344R primers without the adaptors and then ‘tagged’ via a six-cycle secondary PCR reaction as previously described (Wawrik et al., 2012). PCR products were purified using a QIAquick PCR Purification Kit (Qiagen, Valencia, CA, USA), and concentrations were quantified using a Qubit 2.0 and Quant-iT dsDNA BR Assay Kit (Life Technologies, Carlsbad, CA, USA). Equimolar amounts of bacterial and archaeal PCR products were combined and sequenced using 454 GS FLX Titanium sequencing.
Sequence analysis
Reads were denoised to remove sequence errors via the denoise_wrapper.py script in QIIME (Version 1.8.0), and primer/adaptor sequences were trimmed. Chimeric sequences were detected via the reference-based chimera detection algorithm, USEARCH, in QIIME and removed (Caporaso et al., 2010). No primer mismatches were allowed, and the remaining high-quality sequence reads were grouped into OTUs at 97% similarity for both Archaea and Bacteria. Sequences were aligned to the SILVA reference alignment database (Pruesse et al., 2007) using PYNAST. Taxa that accounted for ≥1% reads in any of the 15 libraries (i.e. five sediment locations sequenced in triplicate) were defined as ‘core taxa’, which were further analyzed to assess similarities among sites using PC-ORD (Version 6, MjM Software). To test for similarities and/or differences among sites, taxa frequency data were arcsine-square-root transformed, and a multi-response permutational procedure and a one-way permutational multivariate analysis of variance (PerMANOVA) (McCune, Grace and Urban 2002) were performed using a Bray–Curtis distance measure and 5000 permutations. Non-metric multidimensional scaling (NMDS) was used to visualize grouping patterns of the community in each pyrosequenced library. A scree plot was first conducted in order to determine the appropriate number of dimensions for ordination, and both archaeal and bacterial data sets were analyzed several times using identical parameters to ensure that consistent results were obtained. Parameters for NMDS included: Bray–Curtis distance measure, 1000 runs with real data, 1000 runs of Monte Carlo test with randomized versions of the data, data plotted using two axes and rotated with orthogonal principal axes and starting configurations were chosen randomly. In addition, community richness, diversity (Shannon and Simpson indices) and evenness were assessed using PC-ORD (Version 6, MjM Software).
Microcosm experiments
Sediment samples (ca. 2 cm3) were collected from each horizon and immediately placed into sterile serum bottles under N2 while aboard the R/V Hugh R. Sharp. Bottles were sealed with butyl rubber stoppers and flushed with syringe-filtered N2 gas to maintain anaerobic conditions. Bottles were stored at 4°C during transport and during laboratory storage until microcosms were established.
The surface horizons for each of the stations, as well as horizon 6 at station 908, which had a very high concentration of methane in the pore water, were chosen for microcosm experiments for the same reasons stated above for assA/bssA gene surveys and pyrosequencing. Microcosms were established under sulfate-reducing and methanogenic conditions using basal mineral medium (NaCl, 20 g L−1; MgCl2·6H2O, 3 g L−1; CaCl2·2H2O, 0.15 g L−1; NH4Cl, 0.25 g L−1; KH2PO4, 0.2 g L−1; and KCl, 0.5 g L−1) (pH 7.2) (Widdel and Bak 1992) and strictly anaerobic technique. Sodium sulfate (25 mM) was included in media for sulfate-reducing cultures. Mineral medium was supplemented with trace elements (10 mL L−1) (Tanner 1997) and 0.1 mL of resazurin (1 g L−1 stock). Media was degassed for 45 min under a stream of N2:CO2, and aliquots (45.5 mL) were distributed into 160-mL serum bottles using anaerobic technique, sealed with butyl rubber stoppers and secured with aluminum crimp seals. After sterilization, each bottle was supplemented with 0.5 mL of filter-sterilized RST vitamins (Tanner 1997) modified to include 50 mg L−1 nicotinamide, 5 mg L−1 pyridoxine·HCl, 5 mg L−1 thiamine·HCl, 5 mg L−1 riboflavin, 5 mg L−1 vitamin B12, 5 mg L−1 biotin, 5 mg L−1 folic acid, 5 mg L−1 calcium pantothenate, 5 mg L−1 thioctic acid, 5 mg L−1p-aminobenzoic acid, 0.4 mL cysteine-sulfide (12.5 g L−1 of each) and 1.5 mL NaHCO3 from a 10% stock (w/v). Mercaptoethanesulfonic acid was included in the vitamin solution at a concentration of 5 mg L−1 for methanogenic incubations. Filter-sterilized hexadecane (0.1 mL) (Sigma, St. Louis, MO, USA) was added as an overlay to appropriate bottles.
Sediment inoculation was performed in an anaerobic chamber under N2:H2 (95:5). A sediment slurry was established with 2 g of core sediment and 50 mL of sulfate-free basal mineral medium. From the sediment slurry, 2 mL were syringe-injected into the appropriate treatment bottles. The amount of the sediment inoculum was selected to introduce sufficient biomass and to minimize the amount of endogenous carbon, which would make it more difficult to discern sulfate loss and/or methane production over background levels. Bottles were removed from the anaerobic chamber, and the headspace was flushed three times with filter-sterilized N2:CO2 (80:20). Five treatment conditions were established in triplicate for each horizon tested (Table S3, Supporting Information) and included active cultures (amended with an overlay of hexadecane and the sediment inoculum), abiotic media controls (amended with hexadecane but no sediment inoculum), background controls (sediment inoculum with no hexadecane), sterile controls (amended with hexadecane and sediment, autoclaved on three consecutive days) and positive controls [amended with an overlay of hexadecane, sediment inoculum and a 10% (v/v) inoculum of Desulfatibacillum alkenivorans strain AK-01 (approximately 105 cells)]. Desulfatibacillum alkenivorans strain AK-01 is a known alkane-utilizing sulfate reducer originally isolated from the Arthur Kill waterway (So and Young 1999). AK-01 was used as a positive control because this organism can utilize a range of alkanes (C13–C18) under sulfate-reducing conditions. Additionally, AK-01 has been shown to utilize n-hexadecane syntrophically with the methanogen M. hungatei strain JF-1 in the absence of sulfate (Callaghan et al., 2012). AK-01 therefore served as a potential positive control under methanogenic conditions (i.e. methane production in these incubations above background levels would indicate that the sediments contained methanogenic archaea with the ability to couple with a known hexadecane utilizer, whereas absence of methane in these incubations would suggest that AK-01 could not couple syntrophically with the indigenous methanogens). Microcosms were incubated at room temperature (∼24–25°C) (in situ water temperatures above the sediment averaged 27.5°C; see Table S1, Supporting Information) in the dark for 672 days. Microcosm activity was monitored via sulfate loss on a Dionex ICS-1100 (Dionex, Sunnydale, CA, USA) equipped with an IonPac AG23 anion exchange column using eluent of 4.5 mM Na2CO3 and 0.8 mM NaHCO3 at a flow rate of 1 mL min−1. Methane production was monitored as described above.
Accession numbers
Sequences of assA and bssA were deposited in GenBank under the following accession numbers: KM096832-KM096849. The 16S rRNA gene sequence data were deposited in NCBI's Short Read Archive under the following accession number: SRP044028.
RESULTS
August 2010 cruise CTD data confirmed hypoxic conditions in near bottom waters of the upper Bay (stations 908 and 858) as observed in 2009 (Fig. S2, Supporting Information) and 2010 (Fig. S3, Supporting Information). Sediment gravity cores were therefore collected at four sites along the salinity gradient that spanned the hypoxic and oxic zones (Fig. S1 and Table S1, Supporting Information). The upper Bay cores (stations 908 and 858) collected within the area of seasonal hypoxia were dominated by silty clay that appeared to be sulfidogenic. Lower Bay sites (stations 818 and 707) yielded gray sandy cores that contained carbonate shell debris. Qualitatively, these cores appeared to contain less organic matter than upper Bay sediments.
Sediment pore water analysis
Overall, pore water sulfate concentrations in cores collected from the upper Bay were ca. two orders of magnitude lower than in cores from the lower Bay stations (stations 818 and 707) (Fig. 1A), and concentrations in the upper Bay declined rapidly with depth (i.e. < 0.1 mM). Methane was detected in all sampled horizons in the upper Bay, but concentrations were negligible in lower Bay sediments (Fig. 1B). Pore water methane concentrations in the upper Bay increased with depth, ranging from 0.47 ± 0.02 to 2.07 ± 0.32 mM at station 908 and between 0.25 ± 0.03 and 0.54 ± 0.05 mM at station 858. Alternative terminal electron acceptors, such as nitrate and nitrate, were below the limits of detection via ion chromatography at all stations and depths (data not shown).
Depth profiles of (A) sulfate and (B) methane concentrations in Chesapeake Bay sediment pore water. Methane measurements were obtained for triplicate sediment samples from each horizon via gas chromatography. Sulfate concentrations were determined by analyzing triplicate pore water samples via ion chromatography.
Microbial community analysis
A total of 57 633 bacterial and 17 901 archaeal sequence reads were obtained via 454-sequencing. Proteobacteria contributed to the largest proportion of the bacterial communities at each of the locations, ranging from 24 to 51% of the 16S rRNA reads. Proteobacteria were significantly more abundant in upper Bay sediments (averaging stations 908 and 858) compared to lower Bay sediments (averaging stations 818 and 707) (P = 1.32E–04) (Fig. 2A). Delta- and Gammaproteobacteria made up the largest proportions of the Proteobacteria, accounting for 67–90 and 5–22% of proteobacterial reads, respectively (Fig. 2A). Both Delta- and Gammaproteobacteria accounted for significantly greater proportions of libraries in upper Bay sediments compared to lower Bay sediments (P = 1.62E–03 and 5.61E–05, respectively). Detected gammaproteobacterial lineages included the Chromatiales, Thiohalophilus, Xanthomonadales, Sedimenticola, Xanthomonadales, Oceanospirillales, Legionellales, Methylococcales and Alteromonadales (Table S4, Supporting Information). Dominant within the Gammaproteobacteria were unclassified lineages (55–79% of reads), as well as the Chromatiales, which accounted for 5–30% of gammaproteobacterial reads. Among the Deltaproteobacteria, the Desulfobacterales (8–20% of all reads) and the Syntrophobacterales (2–13% of all reads) were the most prominent orders, with both being significantly more abundant in upper Bay sediments than lower Bay sediments (P = 0.03 and 7.90E–04, respectively). Within Syntrophobacterales, Syntrophus was the dominant genus, representing 23 and 76% of Syntrophobacterales reads at stations 908 and 858, respectively, whereas Smithella accounted for 4 and 16% of requisite reads, respectively. Chloroflexi were detected in high proportional abundances at all sites, accounting for 10–38% of bacterial 16S rRNA reads and making up a significantly greater proportion of the community in the lower Bay (P = 6.94E–05). The majority of Chloroflexi-like sequences were classified within the class Dehalococcoidetes (7–37% of all reads) and the genus Dehalogenimonas (7–33% of all reads), with both taxonomic groups being more abundant in lower Bay sediments (P = 1.33E–04 and 6.60E–05, respectively). With respect to depth, Dehalococcoidetes were proportionally more abundant in horizon 6 compared to the surface horizon at station 908 (P = 1.74E–05), whereas Deltaproteobacteria were less abundant with depth at this station (P = 8.90E–04). Also detected in Bay sediments were a diverse group of Firmicutes, accounting for 5–14% of all sequences. A large proportion of these reads were attributed to the Clostridia (55–88% of Firmicute reads). Both the Firmicutes (phylum) and the Clostridia (class within the Firmicutes) were proportionally more prevalent in upper Bay sediments compared to lower Bay sediments (P = 5.71E–03 and 0.04, respectively). At the family level, Firmicute lineages in Bay sediments included several Bacillales, including Bacillaceae, Paenibacillaceae, Staphylococcaceae, Thermoactinomycetaceae, Enterococcaceae, Lactobacillaceae and Streptococcaceae. Detected, classifiable Clostridia families included Clostridiaceae, Eubacteriaceae, Peptococcaceae, Peptostreptococcaceae, Veillonellaceae, Natranaerobiaceae and Thermoanaerobacteraceae (Table S4, Supporting Information). None of the Firmicute OTUs assigned beyond the order level accounted for more than 1% of reads in any of the samples, and 42–65% of reads attributed to Firmicutes were either annotated as unclassified Clostridia or unclassified Firmicute lineages.
Microbial community composition in Chesapeake Bay sediments as determined by 454-pyrosequencing of partial 16S rRNA gene PCR products. (A) Bacterial and (B) archaeal 16S rRNA genes were amplified separately, and reads were analyzed using QIIME (version 1.8.0) (Caporaso et al.2010). Community composition data are shown at the class taxonomic level. Minor phylogenetic groups, which could not be visually resolved in the bar graphs, are not included in the legend.
With respect to the archaeal communities (Fig. 2B), the upper Bay stations were dominated by Euryarchaeota (81–88% of archaeal reads), whereas Crenarchaeota were significantly more prevalent at lower Bay stations (P = 1.88E–04). At the class level, the euryarchaeal sequences were primarily attributed to the Methanomicrobia, Thermoplasmata or were unclassified Euryarchaeota. Methanomicrobia and Thermoplasmata were proportionally more abundant in upper Bay sediments (P = 1.51E–08 and 5.63E–05, respectively) (Table S5, Supporting Information).
Clustering of the 16S rRNA reads produced 2086 bacterial and 861 archaeal OTUs. ‘Core taxa’ within libraries were defined as taxa that occurred at ≥1% frequency in at least one of the libraries. The frequencies for these dominant (core) groups at the class level were used for ordination using NMDS. NMDS indicated that the bacterial communities in the upper Bay are distinct from those in the lower Bay (Fig. S4A, Supporting Information), whereas upper Bay archaeal communities clustered more tightly than those for the lower Bay (Fig. S4B, Supporting Information). PerMANOVA analysis indicated that replicates from each of the five horizons were more similar to each other than to other sites. Analysis of the core bacterial and archaeal communities through a one-way PerMANOVA using Bray–Curtis as a distance measure, with groups defined by site and 5000 randomizations, indicated an observed test statistic of F = 67.24 (P = 2.0E–04) for Archaea and an observed test statistic of F = 35.56 (P = 2.0E–04) for Bacteria (Table S6, Supporting Information). The Shannon diversity index ranged from 2.48 to 2.86 for Bacteria and from 1.37 to 1.69 for Archaea, and evenness ranged from 0.84 to 0.94 for Bacteria and 0.77 to 0.94 for Archaea (Table S7, Supporting Information).
Quantitative PCR
Total bacterial abundances in Chesapeake Bay sediment, as determined by the quantification of rRNA genes (and assuming one 16S rRNA gene per genome), ranged between 4.30 × 106 and 5.63 × 107 per gram of wet sediment. The 16S rRNA gene abundances declined with depth in the sediment and were greater in the upper Bay compared to the lower Bay sediments (Table S8, Supporting Information). Copy numbers of dsrA genes ranged between 3.78 × 104 and 2.98 × 106 per gram of wet sediment (Table S8, Supporting Information). At stations 908 and 818, dsrA copy numbers decreased with depth, with relative frequencies (based on the ratio of dsrA copies to 16S gene copy numbers) of 5.28–0.96% and 3.47–0.23%, respectively (Fig. S5A and Table S8, Supporting Information). Station 707 exhibited the highest relative frequency of sulfate reducers at approximately 2 ft (0.6 m) below the surface. The relative frequencies of sulfate reducers were fairly constant with depth at station 858, averaging 1.65% (Fig. S5A and Table S8, Supporting Information).
The number of archaeal rRNA genes ranged from 107 to 108 per gram of wet sediment for stations 908, 858 and 707, whereas abundances at station 818 were an order of magnitude lower (Table S9, Supporting Information). Quantification of mcrA genes indicated at least an order of magnitude difference between stations 908 and 858 and stations 818 and 707 (105 to 106 and 104 to 105 per gram of wet sediment, respectively). On average, the relative frequencies of methanogens accounted for ∼4.9% of the archaeal community at stations 908 and 858, whereas they accounted for less than 1% (0.97%) of archaeal populations at stations 818 and 707 (Fig. S5B, Supporting Information). These data are consistent with greater proportions of reads classified within the Methanomicrobia in stations 908 and 858 vs 818 and 707 (see above). The estimated proportional abundances of methanogens among the Archaea, as measured via qPCR, are lower than in 16S rRNA gene sequence data, which likely reflects a limitation of the mcrA PCR primers used herein to quantitatively capture the full diversity of this gene in the environment.
Detection of assA/bssA genes
Bay sediments were surveyed for assA and bssA genes via PCR. Using touchdown PCR, bssA genes were detected in surface horizons at upper Bay stations 908 and 858 with primer set no. 2 (Table S2, Supporting Information). A gradient PCR protocol was carried out under less stringent parameters on the remaining samples, and assA gene PCR products were obtained from DNA from surface horizons at all four stations using primer set no. 7 and at depth at station 908 (horizon 6) using primer set no. 1. The gradient PCR protocol did not yield bssA gene products from the surface horizons at stations 818 or 707, or from the depth horizon at station 908. Overall, sequencing of cloned PCR products allowed the identification of one bssA genotype (stations 908 and 858) and seventeen assA genotypes in Chesapeake Bay sediments (Fig. 3, Table S10, Supporting Information). Among the observed assA genotypes, several were most similar to sequences previously obtained from hydrocarbon-impacted North Atlantic coastal sites (e.g. Arthur Kill NJ/NY and Gowanus Canal, NJ). Additionally, assA OTUs 1, 2 and 15 were closely related to assA genes recently reported in the draft genomes of Smithella sp. ME-1 and Smithella SCADC (Tan, Nesbø and Foght 2014), which were derived from different methanogenic alkane-degrading enrichment cultures (Embree et al., 2013; Tan et al., 2013). Ten out of seventeen Chesapeake Bay assA OTUs formed a clade with a clone obtained from Gulf of Mexico sediment potentially exposed to oil originating from the Deepwater Horizon oil spill (Kimes et al., 2013). The bssA sequences detected here all assembled into a single OTU at 97% similarity (Fig. 3) and were found to be similar to bssA in Desulfobacula toluolica DSM 7467/Tol2, a sulfate-reducing bacterium originally isolated from anoxic marine sediment (Eel Pond, Woods Hole, MA, USA) (Rabus et al., 1993).
Neighbor-joining dendrogram of translated assA and bssA gene sequences detected in Chesapeake Bay sediments. Sequence reads were assembled into OTUs at 97% similarity, and closest matches for each OTU were determined using BlastX of the NCBI NR database. Resulting OTUs and closest matches were translated into protein sequences and aligned with representative AssA and BssA sequences from several well-described strains using Megalign Software (DNASTAR Inc., Madison, WI, USA) and the ClustalW alignment method. Neighbor-joining trees were constructed with pairwise deletion and performing 10 000 bootstrap replicates. Bootstrap values below 65 are not shown. Pyruvate formate-lyase served as the outgroup for phylogenetic analysis. Abbreviations: Ass (alkylsuccinate synthase), Mas (methylalkylsuccinate synthase), Bss (benzylsuccinate synthase), Nms (naphthylmethylsuccinate synthase) and Pfl (pyruvate formate-lyase). GenBank accession numbers are indicated in parentheses. Stations where OTUs were detected are indicated on the right.
Microcosm experiments
Sulfate-reducing and methanogenic microcosms were established using Bay sediments. For all stations, the positive controls containing sediment, hexadecane and D. alkenivorans strain AK-01 exhibited significant sulfate loss compared to the background controls (P-values ranged between 1.86E–04 and 1.58E–02) (Fig. S6, Supporting Information). The time for complete sulfate depletion in positive controls varied among stations, but was statistically significant (compared to initial concentrations) for all stations by 40 weeks of incubation. After additional sulfate amendments (∼25 mM), the AK-01-amended cultures continued to demonstrate sulfate loss (Table S11, Supporting Information). After 672 days, the active treatments and background controls at each of the stations exhibited small, but significant sulfate loss (P < 0.05) compared to the time-zero concentrations, but they were not statistically different from each other (Table S11, Supporting Information).
Microcosms established under sulfate-reducing conditions from surface horizon sediments collected at stations 858 and 908 produced significantly more methane compared to the background controls after 672 days of incubation (P = 3.32E–03 and 8.77E–03, respectively) (Fig. 4 and Table S12, Supporting Information). With respect to the AK-01 positive controls (under sulfate-reducing conditions), significantly more methane was observed in the surface horizons at stations 908, 858, 818 and 707 than in the background controls, whereas a significant difference was not seen in the positive control at depth at station 908 (horizon 6) (P = 0.07). Additionally, methane production in the AK-01 positive control was significantly higher than in active treatments only at Station 707 (Fig. 4 and Table S12, Supporting Information).
Methane production was monitored in microcosms established from Chesapeake Bay sediments under sulfate-reducing and methanogenic conditions. Microcosms were established with sediments from the surface horizons at each station, as well as the deepest horizon (horizon 6) at station 908. Five treatments were established in triplicate, including (active) enrichments that included media, sediment and a hexadecane overlay; (positive) control enrichments that included media, sediment, hexadecane and D. alkenivorans strain AK-01; (background) controls that contained media and sediment; (media) controls containing only medium and a hexadecane overlay; and (sterile) controls containing media, sediment and hexadecane, which were autoclaved on three consecutive days for sterilization. An asterisk (*) indicates methane production significantly above background controls after 672 days of incubation. A (†) indicates AK-01-amended microcosms with significantly higher methane production than the active treatments.
Microcosms established under methanogenic conditions for stations 908 (surface horizon and horizon 6), 858 and 707 produced significantly higher levels of methane (P < 0.05) than background controls after the 672-day incubation period. A small amount of methane was observed in the killed controls for station 858 as well as station 908 (horizon 6), with observed quantities being significantly less than those observed in background controls (P = 0.02–1.1E–05) (Fig. 4 and Table S12, Supporting Information). No methane production occurred in media-only controls under sulfate-reducing or methanogenic conditions. The AK-01 positive control established under methanogenic conditions did result in significant (P < 0.05) methane production in comparison to the background control at the surface horizons at stations 858 and 707, as well as the depth horizon at station 908 (Fig. 4). Overall, significantly more methane was produced in methanogenic microcosms established from upper Bay sediments as compared to sediments collected from lower Bay cores (all pairwise P-values < 8.27E–05).
DISCUSSION
The Chesapeake Bay is a seasonally stratified estuary that experiences summer bottom anoxia, which has become increasingly widespread since its initial identification in the 1930s (Newcombe, Horne and Shepherd 1939; Officer et al., 1984). Anoxia initiates in the spring when increased freshwater and nutrient loading lead to halocline-dependent stratification and increased phytoplankton productivity. The anoxia is then driven by benthic decay of organic matter from sinking phytoplankton and from the previous summer's and fall's seasonal phytoplankton blooms (Taft et al., 1980; Officer et al., 1984; Boesch, Brinsfield and Magnien 2001). A hydrocarbon spill in the Chesapeake Bay therefore has the potential to impact both oxic and anoxic water masses as well as their underlying sediments. Therefore, one aim of the work presented here was to characterize and compare the microbial communities associated with sediments located in areas of frequent hypoxia with those that are less frequently affected by hypoxic waters. Cores were collected across the Bay's salinity gradient, which encompasses both hypoxic and oxic areas, to assess the potential for anaerobic alkane degradation, as a proxy for natural attenuation in the event that an oil spill should occur.
Assuming conservative mixing of seawater (∼28 mM sulfate at a salinity of 35) and given the salinities at stations in the upper Bay (908 and 858; Table S1, Supporting Information) where hypoxia was observed, it can be estimated that the overlying water could contain up to ∼8–9 mM sulfate. Pore water sulfate concentrations, however, were substantially lower (< 0.3 mM), indicating consumption of terminal electron acceptors including sulfate, yielding methanogenic conditions. These data coincide with pore water methane concentrations (Fig. 1B), which indicated high levels of methane throughout upper Bay sediment cores, reaching supersaturated levels in horizon 6 of station 908. These observations are consistent with the important role that sulfate reducers play in the conversion of organic matter in coastal ecosystems, particularly near-shore (Jørgensen 1982). Conversely, at the lower Bay stations (stations 818 and 707), water column salinities would indicate sulfate concentrations of ca. 12 and 22 mM, respectively, which are mirrored by similarly high sulfate concentrations observed in the sediment pore water (Fig. 1A). High sulfate concentrations in lower Bay sediments may reflect less intense input of organic matter via sedimentation and/or input of organic matter that is at a later stage of decomposition and more refractory to oxidation, resulting in the incomplete depletion of terminal electron acceptors (Jørgensen 1982). Alternatively, given the apparent higher porosities of core materials (based on visual inspection) at stations 818 and 707, sufficient pore water exchange with overlaying water might allow for continuous replenishment of sulfate, at least to the depths sampled in this study. Despite the large differences in sulfate concentrations between the upper and lower Bay sediments, no clear trend was observed with respect to differences in the abundances of sulfate-reducing organisms among the different stations (based on qPCR of dsrA and the primers used herein) (Fig. S5A, Supporting Information). Conversely, methanogenic archaea accounted for a 2- to 8-fold greater proportion of the archaeal populations in upper Bay sediments (Fig. S5B, Supporting Information), consistent with overall trends in pore water methane concentrations and the notion that sediment communities associated with the Bay's hypoxic zone are predominantly methanogenic.
High methane concentrations in upper Bay sediments coincided with a greater abundance of sequences classified within the deltaproteobacterial order Syntrophobacterales. Syntrophobacterales, specifically Syntrophus and Smithella spp., are common in methanogenic hydrocarbon-degrading communities, including methanogenic oil sands tailings, oil sands tailings enrichment cultures, hydrocarbon-contaminated sediments and aquifers, methanogenic hexadecane-degrading consortia, oil field production water, methanogenic coal seam groundwater and coal-impacted wetlands (see Gray et al., 2011 and references therein; Siddique et al., 2011; Wawrik et al., 2012; Cheng et al., 2013; Tan et al., 2013). Furthermore, a greater proportion of Firmicutes was detected in the upper Bay sediments. These bacteria are well known for their ability to process and ferment complex organic matter and are often detected in hydrocarbon-amended enrichment cultures and hydrocarbon-impacted environments (Gieg, Duncan and Suflita 2008; Penner, Foght and Budwill 2010; Wawrik et al., 2012). More recently, it has also been reported that some members of the Firmicutes, such as Clostridiales, may play an important role in the activation of hydrocarbons under methanogenic conditions (Fowler et al., 2012). Among the archaeal communities, upper Bay sediment libraries contained large proportions of Euryarchaeota, particularly the methanogenic class Methanomicrobia consistent with both the measured pore water methane concentrations and mcrA data (Fig. 1B and Fig. S5B, Supporting Information). Methanomicrobia are often detected in methanogenic hydrocarbon-amended enrichment cultures and hydrocarbon-contaminated systems, and it has been hypothesized that this group of methanogens plays a key role in the conversion of hydrocarbons via coupling with the requisite syntrophs (for review, see Gray et al., 2010).
Compared to the upper Bay, the lower Bay sediment 16S rRNA libraries contained proportionally fewer sequences within groups traditionally associated with organic matter fermentation and methanogenesis. Specifically, significantly greater proportions of sequences classified as Dehalococcoidetes (Chloroflexi) were detected in these sediments. Dehalococcoidetes and closely related groups are known to be involved in organohalide respiration and have potential roles in bioremediation of chlorinated compounds that have been used for decades as industrial solvents (Richardson 2013). The latter is relevant to the Bay, because of a history of PCB pollution (Ashley and Baker 1999; Walker, McNutt and Maslanka 1999; Foster et al., 2000; King et al., 2004). Archaeal communities included large proportions of sequences within the Thermoprotei, which have been detected in methanogenic alkane-degrading enrichment cultures, albeit at low levels (Gray et al., 2011).
Given large genome variability within species and high rates of lateral gene transfer, 16S rRNA gene sequences are a poor indicator for microbial functional traits. To obtain a clearer picture of a community's potential ability to degrade specific pollutants, functional gene markers are frequently used. As previously discussed, aliphatic and aromatic hydrocarbon addition to fumarate (i.e. ‘fumarate addition’) is one of several mechanisms of anaerobic hydrocarbon activation (for review see, Heider and Schühle 2013). It is catalyzed by the glycyl radical enzymes ASS/MAS (Callaghan et al., 2008; Grundmann et al., 2008) and BSS (Leuthner et al., 1998), respectively. The genes encoding the catalytic subunits of ASS and BSS (assA and bssA) are considered useful biomarkers in this regard (for review, see Callaghan et al., 2010; Agrawal and Gieg 2013; Callaghan 2013). More recently, intense efforts have been focused on elucidating pathways of methanogenic conversion of hydrocarbons. To date, there have been several studies that have detected bssA (for review see, Callaghan 2013) and/or assA in methanogenic enrichment cultures and/or methanogenic hydrocarbon-impacted environments (Davidova et al., 2011; Li et al., 2012; Mbadinga et al., 2012; Wang et al., 2012; Wawrik et al., 2012; Zhou et al., 2012; Aitken et al., 2013; Cheng et al., 2013), providing evidence that fumarate addition may play an important role in the hydrocarbon activation step. A recent study of an n-hexadecane-degrading methanogenic enrichment culture aimed at identifying requisite alkane-degrading bacteria resulted in a draft genome of Smithella sp. ME-1 (Embree et al., 2013), which was subsequently reported to contain a nearly full-length assA gene to which metatranscriptomic reads were mapped (Tan et al., 2014). These observations are consistent with data from another methanogenic alkane-degrading enrichment culture (SCADC) (Tan et al., 2013), in which a single copy of assA (GenBank accession KF824850) was recovered from a partial Smithella sp. genome (Tan et al., 2014). The genus Smithella is a member of the family Syntrophaceae, and assA genotypes closely related to this gene from Smithella sp. were found in both upper Bay stations (908 and 858) and station 707 (Fig. 3). Genotypes of assA closely related to the sulfate-reducing, alkane-degrading strains D. alkenivorans AK-01 and Desulfoglaeba alkanexedens ALDC were also detected at all four stations. Moreover, assA genotypes similar to those detected in the Gulf of Mexico sediments near the Deepwater Horizon oil spill were also detected. These data are consistent with the presence of bacteria capable of alkane utilization under methanogenic (i.e. syntrophic) and sulfate-reducing conditions throughout Chesapeake Bay in both surface sediments and at depth.
In contrast, bssA-like sequences were only observed in the surface horizons of the upper Bay stations (Fig. 3). Given the limited number of samples analyzed here, our ability to derive conclusions regarding the biogeography of ass and bss genes in the Bay is limited. However, the substrate range of BSS includes toluene, ethylbenzene and xylene isomers (i.e. TEX) (for review, see Heider and Schühle 2013), which are far more soluble than aliphatic compounds such as the mid- to longer-chain alkanes. It is possible that the shorter residence times of these more soluble compounds in the water column and sediments may influence the lack of enrichment and/or biogeography of TEX-degrading microorganisms in the Bay. Alternatively, primer specificity may hinder the ability to detect bssA-type genes at some sites. To date, PCR primers that capture the full range of known bssA genotypes have not been reported (Acosta-González, Rosselló-Móra and Marqués 2013; von Netzer et al., 2013), and it is therefore possible that bacteria potentially capable of TEX degradation are more widely distributed in Bay sediments than observed here.
In an effort to further investigate the potential for hydrocarbon degradation by microbial communities in Chesapeake Bay sediments, microcosm experiments were conducted using hexadecane as a substrate under sulfate-reducing and methanogenic conditions. Cultures were maintained for >600 days. The long incubation time is not atypical of other studies, in which lag times associated with methanogenic degradation of long-chain alkanes have been observed to be as long as 280 days (Siddique et al., 2011). Despite the lengthy incubation, these experiments resulted in several observations. First, the addition of hexadecane did not significantly stimulate sulfate loss in the absence of D. alkenivorans strain AK-01 as a positive control (Fig. S6, Supporting Information). This observation suggests that the detected assA genotypes are potentially not affiliated with the indigenous and ‘strict’ sulfate reducers (i.e. they may be affiliated with the indigenous syntrophs). Alternatively, the absence of sulfate reduction may simply be an issue of substrate specificity. For example, known sulfate-reducing bacteria that utilize alkanes have broad, but variable, substrate ranges: D. alkenivorans AK-01 can utilize C13–C18 alkanes (So and Young 1999); D. alkanexedens ALDC utilizes C6–C12 alkanes (Davidova et al., 2006); and D. oleovorans Hxd3 utilizes C12–C20 alkanes (Aeckersberg, Bak and Widdel 1991). The second observation was the production of methane under both sulfate-reducing and methanogenic conditions. Under sulfate-reducing conditions, the active treatments and positive controls produced significantly more methane than the background controls for microcosms established with upper Bay surface sediments compared to the microcosms established with lower Bay sediments (Fig. 4). Moreover, the addition of hexadecane under methanogenic conditions appeared to stimulate methanogenesis at stations 908 (surface and at depth), 858 and 707. Significant methane production was not observed under sulfate-reducing or methanogenic conditions in incubations established with station 818 sediment. Together, the higher levels of methane production in hypoxia-influenced sites (i.e. upper Bay sediments) are consistent with the higher abundances of Syntrophaceae and acetoclastic and hydrogenotrophic methanogens observed in upper Bay sediments.
CONCLUSION
Research addressing the fate and transport of the oil associated with the Deepwater Horizon oil spill demonstrated that the microbial community played an important role in remediation via natural attenuation mechanisms (for reviews, see Joye, Teske and Kostka 2014; Kimes et al., 2014; King et al., 2015). With respect to the Gulf of Mexico, the microbial community demonstrated a rapid and robust response (for reviews, see Joye, Teske and Kostka 2014; Kimes et al., 2014; King et al., 2015). Molecular analyses of plume water and ocean and coastal sediments via microarrays, targeted gene surveys, metagenomics and metatranscriptomics highlighted the importance of both aerobic and anaerobic hydrocarbon degradation (for reviews, see Joye, Teske and Kostka 2014; Kimes et al., 2014; King et al., 2015). In contrast to the Gulf of Mexico, the Chesapeake Bay is a much smaller, shallower and more dynamic ecosystem, driven by different physical and chemical processes, and the predicted response to a large oil spill would also be very different. Realistically, physical remediation would in all likelihood be the most exploited tactic in an oil spill response for a system like the Bay. Unlike the Deepwater Horizon oil spill, which elicited a fast, ‘aerobic response’ of microbial communities, the Bay's bioremediation capacity in the water column and in the sediments would likely be influenced by its periods of seasonal hypoxia. Long term, this could dictate an increased dependence on the anaerobic microbial communities to metabolize the hydrocarbons that partition to sediments. Although hydrocarbons are probably not a selective pressure on Bay sediments, past investigations have demonstrated the ability of Bay microbial communities to utilize hydrocarbons aerobically (Walker and Colwell 1973; Walker, Colwell and Petrakis 1976a, Walker, Petrakis and Colwell 1976b; Okpokwasili et al., 1984; West et al., 1984). Here, we report that the microbial communities of Bay sediments include microbial taxa frequently associated with the anaerobic conversion of hydrocarbons. The potential for anaerobic aromatic and aliphatic hydrocarbon transformation is further supported by the detection of bssA and assA genotypes at different locations throughout the Bay and the ability to stimulate methane production in the presence of hexadecane under sulfate-reducing and methanogenic conditions. The occurrence of natural attenuation of hydrocarbons under anaerobic conditions can therefore be taken into account when considering a remediation strategy in response to a major spill in the Chesapeake Bay ecosystem.
We wish to thank the crew of the R/V Hugh R. Sharp for outstanding support in the field. We would also like to thank Steven E. Baer at Bigelow Laboratory for Ocean Sciences for his assistance with oxygen data analysis.
FUNDING
This work was supported in part by National Science Foundation grants OCE-0961900 and MCB-091265 and by BP (The Gulf of Mexico Research Initiative, Project No. 130206).
Conflict of interest statement. None declared.
REFERENCES
