Microbial interactions with naturally occurring hydrophobic sediments: Influence on sediment and associated contaminant mobility

Water Research 92 (2016) 121e130

Contents lists available at ScienceDirect

Water Research journal homepage: www.elsevier.com/locate/watres

Microbial interactions with naturally occurring hydrophobic sediments: Influence on sediment and associated contaminant mobility I.G. Droppo a, *, B.G. Krishnappan a, J.R. Lawrence b a b

Environment Canada, 867 Lakeshore Road, Burlington, Ontario, Canada Environment Canada, 11 Innovation Blvd., Saskatoon, Saskatchewan, Canada

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 August 2015 Received in revised form 8 January 2016 Accepted 17 January 2016 Available online 22 January 2016

The erosion, transport and fate of sediments and associated contaminants are known to be influenced by both particle characteristics and the flow dynamics imparted onto the sediment. The influential role of bitumen containing hydrophobic sediments and the microbial community on sediment dynamics are however less understood. This study links an experimental evaluation of sediment erosion with measured sediment-associated contaminant concentrations and microbial community analysis to provide an estimate of the potential for sediment to control the erosion, transport and fate of contaminants. Specifically the paper addresses the unique behaviour of hydrophobic sediments and the role that the microbial community associated with hydrophobic sediment may play in the transport of contaminated sediment. Results demonstrate that the hydrophobic cohesive sediment demonstrates unique transport and particle characteristics (poor settling and small floc size). Biofilms were observed to increase with consolidation/biostabilization times and generated a unique microbial consortium relative to the eroded flocs. Natural oil associated with the flocs appeared to be preferentially associated with microbial derived extracellular polymeric substances. While PAHs and naphthenic acid increased with increasing shear (indicative of increasing loads), they tended to decrease with consolidation/biostabilization (CB) time at similar shears suggesting a chemical and/or biological degradation. PAH and napthenic acid degrading microbes decreased with time as well, which may suggest that there was a reduced pool of PAHs and naphthenic acids available resulting in their die off. This study emphasizes the importance that any management strategies and operational assessments for the protection of human and aquatic health incorporate the sediment (suspended and bed sediment) and biological (biofilm) compartments and the energy dynamics within the system in order to better predict contaminant transport. Crown Copyright © 2016 Published by Elsevier Ltd. All rights reserved.

Keywords: Sediment Deposition Erosion Microbiology Transport Flocculation Pollutants

1. Introduction The lower Athabasca River and many of its tributaries (e.g. Ells River) cut through the Fort McMurray Formation, the geological strata that constitute the Oil Sands deposit in northern Alberta, Canada. Visual observations of the lower Ells River shows occasional areas of oil sheens at the waterebank interface and the development of oil sheens when disturbing recently deposited sediment via samplers or footsteps. For the Ells River, Yergeau et al. (2012) showed a two order of magnitude increase in total petroleum hydrocarbons, total single-chain hydrocarbons, total aromatic

* Corresponding author. E-mail address: [email protected] (I.G. Droppo). http://dx.doi.org/10.1016/j.watres.2016.01.034 0043-1354/Crown Copyright © 2016 Published by Elsevier Ltd. All rights reserved.

hydrocarbons, and an order of magnitude increase in the sum of USEPA 16 priority PAHs moving from headwaters to mouth. Whether this was an accumulation of natural PAHs with progression downstream or a site specific change due to geological variations was not determined (active mining in this basin has not yet commenced). High natural petroleum hydrocarbon content has translated to very hydrophobic sediment within this basin. Hydrophobicity measurements on the Ells River range from 34% (17 km upstream from mouth) (this study) to as high as 96% (near confluence with Athabasca River) (Droppo et al., 2015). It is known that hydrophobic micropores associated with the sediment particles/flocs themselves can retain significant amounts of organic contaminants while at the same time allowing for slow release (Cheng et al., 2012). Nonspecific hydrophobic interactions are

122

I.G. Droppo et al. / Water Research 92 (2016) 121e130

acknowledged as an important mechanism for organic pollutant binding to natural organic matter (Chefetz et al., 2000). Further, it has been shown that biofilm productivity and biomass collected from the Athabasca River and tributaries (including the Ells River) can be reduced with exposure to natural bituminous (oil) compounds (Yergeau et al., 2013). The unique high oil content (relative to other rivers) of the sediment also provides for a selective community structure that can associate with such conditions (Droppo et al., 2015; Yergeau et al., 2012); often hydrocarbon consumers. While hydrophobic contaminant sorption/desorption with sediments (particularly soils) has been studied extensively, (e.g., Cheng et al., 2012; Karickhoff et al., 1979), the general hydrophobic characteristic of particle, regardless of the origin of the hydrophobicity, has received very little attention with respect to their influence on sediment dynamics and microbial interactions. Droppo et al. (2015) have shown that microbial interactions with bed sediments for the Ells River plays a role in the stabilization of the sediments from erosive forces and influences eroded floc size, shape and density. Microbial interactions can increase the energy required to erode bed sediment (Gerbersdorf and Wierpecht, 2015) with Droppo (2009) showing bed sediments to be up to 10 times stronger with biofilms than without. Further, the elastic/plastic nature of eroded flocculated particles (provided by the attached microbial community/biofilm integration) can generate considerable strength relative to electrochemically bound flocs (Liss et al., 1996). The level of microbial activity within the water column has been shown to be influenced by the sediment dynamics within the system (Walters et al., 2014a, 2014b). Microbial mediation of sediments (bed and/or suspended) can also play a strong role in the biogeochemistry of the sediment and overlying water column [e.g., influence on redox reactions (Elliott et al., 2014; Elliott and Warren, 2014; Saulnier and Mucci, 2000)]. While it is known that sediment dynamics (erosion, transport and fate) can play a significant ecological role within river systems, influencing all trophic levels, geomorphological processes, and with possible significant socioeconomic implications (Grabowski et al., 2011; Chapman, 1988) knowledge of the factors controlling these sediment processes [particularly those of hydrophobic (oil associated) sediments] is relatively limited. An improved understanding of hydrophobic sediment dynamics and the role that the microbial community may play in this, will help better predict the fate and effect of upstream sources of contaminants to downstream impact areas. As such, this study links an experimental evaluation of sediment erosion with measured sediment-associated contaminant concentrations and microbial community analysis to provide an estimate of the potential for sediment to control the erosion, transport and fate of contaminants in the Ells River, Alberta, Canada. The specific objectives of this paper are 1) to describe the unique behaviour of hydrophobic sediments within a river system, 2) to improve our understanding of the microbial community structure associated with hydrophobic sediment and how its development may influence associated contaminant transport. 2. Methods 2.1. Field program 2.1.1. Study river The Ells River has a mean annual flow of 9 m3 s
1 and drains an area of 2450 square km (Environment Canada, 2011; Carson, 1990). The River drains into a variety of different soil types with the headwaters beginning in the Birch Mountains. These soils range from non to exceedingly stony, loams to clay and are predominately formed on calcareous till with the exception of the many bogs surrounding the river (Alberta Environment, 1982). The lower

portion of the river cuts through the Fort McMurray formation (Conly et al., 2002). This basin is only beginning to be developed for in situ bitumen extraction and represented a relatively undisturbed basin at the time of sampling. 2.1.2. Bed sediment sampling for sediment dynamic assessment A sediment-water mixture (1000 L) from the river was collected from a mid-stream location (57140 43.4300 N; 111430 57.6200 W) on October 5, 2012 using a submersible pump with mild bed disturbance upstream of the pump to mobilize recently deposited sediment. The pump sampling was carried out at several locations within a 50 m length of the river. The sediment-water mixture was transported to the laboratory in a refrigerated (4  C) transport truck. Additional bulk bed sediment (100 Kg) was collected in polyethylene 20L buckets by using a stainless steel shovel cleaned with acetone to scrape the top 1 cm of deposited sediment within a backwater area at the site. This additional sediment was used to form a cohesive bed surface within the flume (described below) during erosion experiments. 2.2. Characterisation of fine sediment dynamics Erosion characteristics of fine sediments along with bioesediment interactions of the Ells River were studied experimentally using a laboratory rotating annular flume (Krishnappan, 1993). The flume consists of a 5.0 m in mean diameter, 0.30 m wide and 0.30 m deep channel with a counter rotating top cover (ring) that fits just inside the flume (~1.5 mm gap on either side) and makes contact with the water surface within the flume (Krishnappan, 1993). The flows generated in the flume are close to two dimensional with a bed shear stress distribution across the width of the flume that is relatively uniform (Krishnappan and Engel, 2004). The flume calibration results of Krishnappan and Engel (2004) were used to predict the relationship between the bed shear stress and the rotational speeds of the flume. 2.2.1. Erosion experiment Ells River bed sediment was added to the flume (sieved at 250 mm to remove organic plants and large material) which was then mixed at a high rate of speed for 30 min to thoroughly mix the sediment and water. The flume speed was then gradually reduced to a stop to allow the mixture to settle and consolidate/biostabilize (bed thickness approximately 2 cm). Three different consolidation/ biostabilization (CB) periods were used in these experiments (3-, 6and 9-days). Following a CB period, the flume and the lid were then set in motion and their speeds were incrementally increased in 60 min time intervals (bed shear intervals used were 0.028, 0.046, 0.072, 0.1, 0.134, 0.169 and 0.211 Pa). Suspended solid (SS) concentrations were estimated using a calibrated flush mount optical backscatter sensor (OBS) located on the outer side of the flume at mid depth. Additional sediment samples were collected every 10 min to measure the SS concentration variation as a function of time and for additional calibration of the OBS probe. The critical bed shear for erosion (tc) was defined as the bed shear stress at which the SS showed a well-defined increase in concentration. After 50 min of a given shear step applied, an additional 50 mL of water/ sediment suspension was collected for biological analysis as per Section 2.6 and 2.7. To maintain water contact with the lid of the flume all sample volumes removed were immediately replaced with additional river water. Sediment-water samples were analysed for concentration of solids by a gravimetric method that consisted of filtering the sample (0.45 mm pre-weighed Millipore filter), and drying and weighing the residue.

I.G. Droppo et al. / Water Research 92 (2016) 121e130

2.2.2. Biofilm development and sampling Bed sediments could not be sampled from the flume prior to experiments as this would influence the integrity of the bed and produce false erosion estimates. As such, microcosms were used in which the same depth of Ells River sediment and water were placed in glass tanks and allowed to consolidate/biostabilize for matching periods as within the flume (3-, 6-, 9-days). Twelve glass slides (Fisher Scientific, Pittsburgh, PA, 300  100  1 mm, rinsed in DI water, wrapped in aluminium foil and autoclaved) were placed on the surface of the sediment prior to complete SS settling (to allow a layer of Ells River sediment to coat the slides and facilitate the development of sediment-associated biofilms for the 3-, 6- and 9day CBs). In this way the biofilm could be sampled and analysed as per below without affecting the integrity of the bed surface within the flume. 2.3. Dry density analyses Samples were settled within three glass beakers forming a similar bed thickness to the experimental zone with 18 cm of water overlaying. Dry density (dry weight of sediment/bulk volume) profiles were measured for 3-, 6- and 9-day consolidation periods at 1 mm increments using an ultrasound Ultra High Concentration (UHC) meter manufactured by Delft Hydraulics, The Netherlands (Berkhout, 1994). Note that the UHC meter can only, however, assess density changes at 0.5 cm and below the surface due to the ultrasound probes (1 cm diameter) requiring full submersion in the sediment. 2.4. Relative hydrophobicity Using the microbial adherence to hydrocarbons method (MATH) of Rosenberg et al. (1980), the relative sediment hydrophobicity was analysed. The sediment was washed twice using Nanopure water and then centrifuged at 3000 g for 5 min, after each washing. The sample was then shaken to re-suspend the pellet and then sonicated (Fisher Scientific, Sonic Dismembrator 100) for 30 s. Initial absorbance of the dispersed suspension (Io) was adjusted as close to 1.50 ± 0.02 as possible, using Nanopure water to dilute (l ¼ 400 nm). Ten mL of the adjusted sediment suspension was mixed for 2 min with 1 mL hexadecane, using the Vortex mixer (VWR). After mixing, the phases were left to separate for 10 min in a separatory funnel. The aqueous phase was collected (l) and the absorbance measured at l ¼ 400 nm. The relative hydrophobicity was calculated using Eq. (2):

 % Hydrophobicity ¼



Io
l *100 Io

(2)

Where Io ¼ initial absorbance l ¼ post separation absorbance

2.5. Contaminant analysis One litre whole water samples collected 50 min into each shear level from the flume were run using AXYS Analytical Services Ltd, Sidney, BC, Canada method MLA-021. This method provides a spectrum of 74 PAH groups (51 parent and 23 alkylated). For this study Fluoranthene, Pyrene, Benzo(a)pyrene, Naphthalene and C4Naphthalenes are reported. The method is ISO/IEC 17025:2005 accredited for all parameters by CALA and is based on USEPA method 8270C/D modified by EPA 1625B (application of isotope dilution/recovery correction quantification to semi-volatile

123

analysis). The GCMS system employed is an Agilent 6890N GC with a Restek RTX-5 capillary column, coupled to an Agilent 5973N MS operated at a unit mass resolution in electron impact (EI) ionization mode. The method measures all target compounds in one run, using Multiple Ion Detection for 2 masses from each specific target and surrogate. Ratioing peak criteria for each specific target and surrogate must be met for positive identification. Alkylated groups are the sum of specified target masses over a defined retention window, calibrated from alklyated PAH technical mixtures. Sample processing commences with the addition of a wide variety of perdeuterated compounds, added prior to commencement of liquid/liquid extraction with dichloromethane. Clean-up and fractionation are performed on silica columns. Typical detection limits, based on multiple, ratioing masses at 3:1 signal to noise ratio, are in the 1e10 ng/L range, and are measured in each sample. Standard batch QC includes on OPR, method blank, and duplicate plus calibration verification samples in each. A 5 point calibration system for specific targets is used. Method QC acceptance criteria for surrogate and native recoveries, reproducibility, and accuracy are determined as per CFR 40 guidance. Method blank acceptance limits are established from the mean blank level plus two standard deviations. (Full details of the method may be found in Appendix A of supplementary data). 2.6. Confocal laser scanning microscopy and probes Confocal laser scanning microscope (CLSM) image sequences were collected using an MRC 1024 confocal laser scanning microscope (Zeiss, Jena, Germany) attached to a Microphot SA microscope (Nikon, Tokyo, Japan) equipped with a plan-apochromat, oil immersion 60X, 1.4 numerical aperture lens. Bacteria were stained with the nucleic acid stain SYTO9 (Invitrogen-Molecular Probes, Eugene, OR). Fluorescence of SYTO9 (Invitrogen-Molecular Probes) was recorded in the green (excitation/emission 488/522 nm) channel. The fluor-conjugated lectins Triticum vulgaris (N-acetyl glucosamine residues and oligomers), Helix pomatia and Concanavalin A (Invitrogen-Molecular Probes) were used to investigate compositional changes in extracellular polymeric substances (EPS). Lectin staining was performed as described in detail by Neu et al. (2001). In brief, for staining, the lectins were dissolved at 100 mg ml
1 in filter-sterilized (pore size, 0.20 mm; FisherBrand; Fisher Scientific) tap water. FUN-1 a fungal stain from (InvitrogenMolecular Probes) was used for visualization of fungal hyphae. 2.7. Microbial analysis Plate counts were carried out using 10% trypticase soy agar (TSA) for total aerobic plate counts (TAPC), and rose bengal agar (RBA) for total fungi. Most Probable Numbers (MPN's) were determined using Bushnell Haas Medium (Difco Products, Detroit, MI) as the basal medium with either Fluka Napthenic Acids, n-hexadecane, or a mixture of five PAH's anthracene (1 g L
1), phenanthrene (10 g L
1), dibenzothiophene (1 g L
1), fluorine (1 g L
1), dissolved in pentane as a source of carbon (Haines et al., 1996). In brief, samples sediment, water or sediment þ water were suspended by vortexing and serially diluted (1:10) for inoculation of agar media and MPN plates. The MPN's were done in 24 well plates. All incubations were carried out at 22C ± 2C in the dark for 28 days. Growth in MPN plates was detected using the metabolic indicator iodonitrotetrazolium chloride (Sigma, St Louis, MI). The data was reported as CFU's/ml or gram or MPN/ml or/gram for water or water þ sediment mixtures and sediment alone respectively. Carbon utilization spectra were determined for water, water þ sediment, sediment and biofilm samples using commercial Eco-plates (Biolog, Hayward, CA) as described by Lawrence et al.

124

I.G. Droppo et al. / Water Research 92 (2016) 121e130

(2004). Denaturing gradient gel electrophoresis analysis (DGGE) was performed as described in detail (Lawrence et al., 2015). In brief amplified DNA products were checked for specificity and size on agarose gels and the PCR product was separated by DGGE (Muyzer et al., 1993, Muyzer and Ramsing 1995) using an Ingeny phorU2 system (Ingeny, Leiden, The Netherlands). After electrophoresis, the gel was stained with SYBR Green I (1:10000 dilution; Molecular Probes, Eugene, OR.) for 15 min with gentle agitation and photographed using the AlphaImager 3300 gel documentation and image analysis system (Alpha Innotech Corporation, San Leandro, CA.). 2.8. Principal component analysis Band detecting, matching and processing of DGGE gels were completed with the GelCompare II software 4.6 (Applied Maths, Kotrijk, Belgium). Fingerprint data was processed by generating a band-matching table (Boon et al., 2002). The binary data was exported and compared by principal component analysis (PCA) with PRIMER v6 software (PrimerE-Ltd, Lutton, UK). Statistical analyses of PCA scores generated from the first two axes were run using an analysis of similarity (ANOSIM) with PRIMERr v6 software (Clarke, 1993). The inclusion of DGGE ladders allowed GelCompareII to normalize the position of bands in all of the lanes under examination. PRIMER v6 was also used to perform PCA on other data sets obtained from the analyses i.e., carbon utilization, plate counts, MPN, etc. 3. Results and discussion 3.1. Erosion experiments Fig. 1 provides the time-series plots for the erosion experiments that encompass the 3-, 6- and 9-day CB periods following the same one hour incremental increases in shear stress. The bed sediment is relatively stable at low shear conditions as the eroded SS concentrations remain low (below 175 mg L
1). Once erosion began for each shear step, SS concentration increased gradually and showed a tendency to approach a steady state value. As such, increasing shear stresses were required to erode deeper material and to continue to increase the SS concentration. This behaviour is indicative of cohesive sediment and suggests an increase in stability with depth of sediment (Amos et al., 2010). Indeed, using the UHC analysis, the sediment was found to consolidate as indicated by an increase in sediment density with depth (Figure S1). On average density increased by 0.8 g cm
3 from 5 to 25 mm depth regardless of CB [there was no significant difference between regression lines for the 3 CB periods suggesting an even rate of consolidation with

Fig. 1. Erosion plots (SS concentration) relative to change in shear stress level.

depth [modified t-test (Larkin, 1978), p ¼ 0.05]. As such, there is no doubt that physical consolidation and dewatering is occurring over time with this sediment. Such consolidation does not however, rule out biostabilization contributing to the overall increase in stability with time (Gerbersdorf and Wierpecht, 2015). Interestingly, while not statistically significant, the 6- and 9-day run started with the lowest density at 0.5 cm below the bed surface (0.6 g cm
3; See Figure S1) possibly suggesting a larger accumulation of biofilm with higher water trapping capacity for these CBs. The 3-day run was the most easily erodible as evidenced by the rapid increase in SS starting at approximately 0.07 Pa [the critical bed shear stress for erosion (tc-3day)] and generating a final SS concentration of 1200 mg L
1 at a shear of 0.21 Pa. Both the 6- and 9-day runs were relatively similar in their erosion sequence with identical tc-6day and tc-9day of 0.17 Pa. The 6- and 9-day runs plateaued at only 375 and 280 mg L
1 respectively at the highest shear. This further illustrates their significant strength relative to the 3day run. While marginal, the 9-day run showed some signs of being stronger with the lower SS concentrations at peek shear. From previous tests with multiple rivers and similar CB periods, it is unusual that the 6- and 9-day runs were similar in erosion characteristics. This observation, however, may suggest that a steady state condition (biologically and physically) has been reached after 6-days. Although direct comparison is not possible due to difference in methods, it is worth comparing these critical bed shear stress for erosion values to other rivers that also transport cohesive sediments. Table 1 illustrates that the 6- and 9-day Ells River sediment erodes at similar shears with many experiments with comparable days of consolidation, while the 3-day run tends to erode easier than comparable results from other rivers (Table 1). While not investigated in this paper, Droppo et al. (2015) have shown that once eroded, the sediment settles very poorly, remaining in suspension (i.e. critical shear stress for deposition was estimated at less than 0.01 Pa). 3.2. Hydrophobic sediment and bacterial interactions Sediment chemical evaluation confirmed “bitumen like oil” present within the Ells River sediment [Oil Characterization Method 607.1 (Environment Canada, 2013a) and Petroleum Hydrocarbons Method 612.1 (Environment Canada, 2013b)] and as stated above, oil sheens are often spotted on the water surface. It is hypothesised that the high natural oil content in the sediments (Droppo et al., 2015) may be partially related to the unusual transport behaviour stated above. Fig. 2 provides representative micrographs of flocs eroded after 6- and 9-day CB periods. Although direct micro-diagnostics is not possible to confirm the presence of natural oil with TEM analysis, Fig. 2 suggests the presence of natural oil within the eroded flocs of the Ells River. There appears to be a strong association of oils with the EPS fibrils secreted by bacteria (Fig. 2a), with Fig. 2c illustrating a good example of cell-fibril oil associations. Often as shown in Fig. 2d there are oil-like spheres trapped within flocs. Such spheres have never been observed by the authors from multiple TEM images for rivers across Canada. It is unclear from these images as to how strongly the possible oil is associated with the inorganic clay particles observed. While little is known about oil and fine sediment interactions, it is hypothesized that the hydrophobicity of this association may result in a faster rate of bed dewatering once the sediment deposits, thus making it more stable with time (i.e., water will dissociate from the hydrophobic sediment). It is possible that for the 3-day run (Fig. 1) this dewatering may not have occurred to any significant level and sediment connections (flocculation) were not occurring, thus making it easier to erode. Yallop et al. (2000) have

I.G. Droppo et al. / Water Research 92 (2016) 121e130

125

Table 1 Critical bed shear stress for erosion measurements from the literature at various CBs. Sediment type

CB periods (Days)

Critical bed shear stress for erosion (Pa)

Source

Ells River, Alberta Canada Hamilton Harbour, Ontario, Canada Hamilton Harbour, Ontario, Canada South Nation River, Ontario, Canada Stormwater pond, Ontario, Canada Fraser River Delta 5 intertidal sites Athabasca/Muskeg River, Alberta, Canada Hay River, NWT, Canada Agricultural tile-drain sediment, Ontario, Canada Castle River, Alberta, Canada York River, Verginia, USA

3, 6, 9 2, 4 5 2, 7, 14 2, 3, 7 in-situ In-situ (5 different measurement devices) 1, 3, 7 1, 2, 7 4.75 2, 7, 14 in-situ

0.07, 0.17, 0.06, 0.10 0.32 0.14, 0.19, 0.12, 0.13, 0.11e0.50 0.03e0.58 0.16, 0.21, 0.12, 0.16, 0.12 0.10, 0.14, 0.045

This paper Droppo (2009) Droppo et al. (2001) Droppo (2009) Droppo (2009) Amos et al. (1997) Widdows et al. (2007) Garcia-Aragon et al. (2011) Stone et al. (2008) Stone and Krishnappan (1997) Stone et al. (2011) Maa and Kim (2002)

0.17

0.23 0.23 (4.7 outlier) 0.26 0.21 0.16

Fig. 2. Representative micrographs illustrating (a) fibrils with apparent oil coating eroded after a 6-day CB, (b) an apparent oil droplet eroded after a 6-day CB, (c) bacteria with surrounding fibrils coated in apparent oil eroded after a 9-day CB and (d) apparent multiple oil droplets within a floc eroded after a 9-day CB.

reported that dewatering and concomitant changes in wet bulk density may result in some possible community changes as evidenced by changes in microbial carbohydrates and chlorophyll-a in intertidal sediments, However, it is unclear if dewatering influenced our experimental microbial community structure as discussed in Section 3.3. Further, the presence of oil associated with the suspended sediment may result in an increased repulsive force thus promoting smaller flocs and minimizing the flocculation

process (Droppo et al., 2015). Indeed the Ells River suspended flocs were generally smaller (floc d50s and d90s were around 3 mm and 12 mm respectively) than those associated with other Canadian rivers also transporting cohesive sediments (e.g., Stone et al., 2011). It is clear that the oil content of the suspended sediment will likely make a strong contribution to the unusual transport characteristics of the Ells River. Uncharacteristic of other rivers, the TEM images did not show a

126

I.G. Droppo et al. / Water Research 92 (2016) 121e130

strong association of bacteria with inorganic particles directly. This is unusual, as for reasons of protection and organic food sources, often bacteria will associate with fine sediments (Hartz et al., 2008; Bonilla et al., 2007). Brinkmeyer et al. (2015) have shown that in bed sediments of the southern USA greater than 75% of E. coli and enterococci are associated with fine sands while in suspension 67% of E. coli and 62% of enterococci were associated with silts and clays (<63 mm). Further, these bacteria can remain viable even at depth for extended periods of time and temperatures (Brinkmeyer et al., 2015; Droppo et al., 2009). Contrary to these observations, while clearly bacteria will be present in the Ells River system, the strength of their association with sediment is not clear. Using CLSM, Fig. 3 demonstrates that the 3-day flocs were less consolidated with significantly more loose material that was not associated with larger flocs. Contrary to non-oil sands sediments using similar CLSM staining techniques (Droppo et al., 2006) these flocs showed areas where they were not coated with polysaccharides (EPS) in all three CB runs. Clearly from Fig. 2 the EPS is more individualistic and not associated with inorganic compartments. Given the associations observed in Fig. 2 of oil to EPS, it may be possible that these non-organic coated areas do not have as high a degree of oil. Generally EPS is a perfuse component of the floc, representing the main architectural matrix that holds the floc together (Liss et al., 1996). The poor association of EPS to inorganic particles may reduce the potential for bioflocculation (Gerbersdorf and Wierpecht, 2015) to occur with flocs likely to be weaker due to bonds promoted primarily by electrochemical forces. Flocs with low EPS content will in general be less flexible and more prone to breakup (weaker flocs) (Droppo, 2001). This lack of EPS connectivity to inorganic particles, may partially explain the smaller than usual floc size for the Ells River. When viewing the total aerobic population however, Fig. 4a shows very little change in total aerobic plate counts between the 3-and 6-day runs but a substantial increase in the 9-day even though visual observations would suggest a continuous increase in biomass with CB times. The biological compartment of fungi (Fig. 4b) does however show an increase with both CB time and with erosion depth (change in shear). While fungal populations are well known to be ubiquitous within sediments performing an organic decomposition function (Gerbersdorf and Wierpecht, 2015; Barlocher and Murdock, 1989), the role of fungal dynamics with regards to influencing bed sediment stability is poorly understood. Fungi have been found to range from 2 to 39% of the total biomass,

however, this is proportional in relation to other community components (e.g., bacteria) and may be relatively stable in associations (i.e., relative change in biomass may be due to changes in other populations and not fungi) (Barlocher and Murdock, 1989). Regardless, although qualitative, Fig. 5 illustrates that fungus fibrils can be part of the eroded flocs from the Ells River. With the complex dendritic network of hyphae within most fungal structures, if present in large quantities, it would not be surprising if these organisms were substantive contributors to bed sediment stability. The observation of increasing fungal abundance with increasing shear (Fig. 4b) suggests that the fungal organisms are able to develop with depth in the sediment bed. Such a three-dimensional structure within the sediment lends support to their stabilizing nature.

3.3. Contaminant interactions within the hydrophobic sediments/ microbial consortia As many contaminants can be associated with sediment particles, including parent and alkylated PAHs, the sediment dynamics will often control the loading of contaminants downstream. As the critical bed shear stress for erosion is surpassed and erosion begins, contaminants associated with the sediment are entrained and transported. Table 2 illustrates this process regardless of CB (with the exception of Fluoranthene 9-day run and Naphthalene) for the erosion experiments with increasing PAHs once sediment is in suspension at high shear (low shear is pre-erosion). The lack of significant increases in Naphthalene with increasing shear and suspended sediment is influenced by a high blank value but may suggest that this compound is primarily associated with the dissolved phase. Parent Naphthalene can be water-soluble at neutral or alkaline pH and, as such, naphthenic acids should be quite mobile in surface waters (Clemente and Fedorak, 2005). Alkylated Naphthalenes (e.g. C4-Naphthalenes e Table 2) are however hydrophobic, with the level of hydrophobicity increasing with the number of alkyl carbon substituents and, as such, tend to be more particulate bound (Wang et al., 2014). This is substantiated by the strong elevation in C4-Naphthenates concentrations in the whole water samples with erosion (Table 2). It is interesting to note that the PAHs in Table 2 show a reduction in contaminant concentration for both the low shear (pre-erosion) and high shear (maximum erosion) conditions as the CB times continue. The reduction in PAH concentrations for the successive low shear results may be due in

Fig. 3. CLSM images of eroded flocs collected at a shear of 0.1 Pa for 3-, 6- and 9-day runs. Red ¼ extracellular polysaccharide substances (stain ¼ WGA-TRITC); Green ¼ Bacteria (stain ¼ Syto 9); Blue ¼ reflectance (e.g., mineral, clay, diatom reflectance). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

I.G. Droppo et al. / Water Research 92 (2016) 121e130

127

Fig. 4. Graphs illustrating change in a) total aerobic plate counts (CFU/ml), b) Fungi counts (CFU/ml), c) napthenic acid degraders (MPN/ml) and d) PAH degraders (MPN/ml), with CB times and increasing erosion shear. Black bars represent the shear at which the mass erosion began (critical bed shear stress for erosion).

Fig. 5. CLSM images of eroded flocs collected at a shear of 0.075 Pa for the 6-day run. Green ¼ Fungal cells (arrow); Red ¼ EPS (with cells); Blue ¼ reflectance (mineral, clay, diatom reflectance). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

part to the additional settling time although the 6- and 9-day

results were relatively similar suggesting possibly equilibrium in

128

I.G. Droppo et al. / Water Research 92 (2016) 121e130

Table 2 Example of average PAH (Parent and Alkylated) changes with shear (low and high) and time (n ¼ 2). Variable units ¼ ng l
1

Blank

3-day low shear

3-day high shear

6-day low shear

6-day high shear

9-day low shear

9-day high shear

Fluoranthene Pyrene Benzo(a)pyrene Naphthalene C4-Naphthalenes

0.29 0.24 0.19 1.57 13.3

2.71 10.9 4.75 10.0 330

11.9 36.9 19.6 9.81 1090

1.2 5.52 2.46 3.74 134

3.41 15.8 9.50 2.87 310

2.01 5.66 3.28 5.96 150

<2 16.2 9.48 3.38 233

the sediment carrying capacity (i.e. only particles not able to settle remain in suspension with associated PAHs). The reduced PAHs at the 6- and 9-day high shear stress will also be influenced by the lower SS concentration in these samples (see Fig. 1) (i.e., 6- and 9day runs were more stable and resulted in lower SS concentrations compared to the 3-day runs). Regardless, there is also the potential that this reduction in PAHs in the 6- and 9-day runs at low and high shear stress may be because the degradable PAHs have already been removed before the 6-day CB occurred. Although a direct correlation between Table 2's compound specific results and the general microbial class of “PAH degraders” cannot be made, Fig. 4d does illustrate that the number of PAH degraders is high for the 3day run but then drops off to similar levels for the 6- and 9-day runs (i.e. die-off due to loss of degradable PAHs or other available carbon sources is possible). The same trend is also observed for Naphthenic Acid (NA) with MPN NA degraders starting off high for the 3-day run but then dropping substantially for the 6- and 9-day runs (Fig. 4c). This effect was also evident in the Biolog data (Figure S2) where the 3-day run showed greater levels of activity (i.e. utilizing putrescine and glucose-1-phosphate as C-sources) than the 6-day or 9-day runs (6-day was also higher than the 9-day run). Further there is a trend for decreased utilization of these carbon sources over the course of each run, 3-, 6-, and 9-days. This corresponded to a possible reason for the drop in the concentration of Naphthalene and the alkylated PAH C4-Naphthalines for the 6- and 9-day runs (Table 2). Interestingly, for all organisms encompassed within Fig. 4, the point at which mass erosion occurs (critical bed shear stress for erosion) does not correspond to a large jump in counts. This may suggest that there is a gradual sloughing off of cells or a release of daughter cells which is a natural biofilm function as discussed by Bester et al. (Bester et al., 2013) prior to this point. With the loss of C-source e.g., PAH as well other hydrocarbons and available organic carbon, early in the experiment, these organisms may die off due to lack of carbon and energy resulting in the lower counts. While a biological phenomenon appears to provide some explanation for the change in PAHs, the role of direct chemical degradation should not be ruled out (e.g., volatile losses). Clearly both chemical and biological degradation will vary with compound; however, delineation of the relative contribution of each to contaminant degradation is difficult. Nonetheless, there is a clear role for microbial processes in the observed changes. Fig. 6 provides insight into how the microbial population changes over the 3 CB periods in relation to biofilm, water, sediment and eroded floc. With averaged triplicate samples, clear populations emerge in which the biofilms and sediment prior to introduction to the flume and water collected at time 0 (before flow initiated e T0) and at the end of the erosion (TE) are grouped in one population and are distinct from the eroded floc populations of each CB period. The commonality of both the biofilms and preerosion (T0) and end of erosion (TE) water samples i.e. within same component (43% similarity) is suggestive of a strong interaction between the bed and planktonic populations and that the microbial community did not change in the water (no floc) during erosion experiments. Focusing the principal component analysis however more specifically on the results of microbial community

structure analyses based on DGGE of DNA from individual water samples does suggest that there are differences between the microbial communities with CB times (similarity ¼ 48%) (Figure S3). The same is true for the PCA of the DGGE analyses of microbial community structure in biofilm samples (Figure S4) which show that the 6-day and 9-day samples were similar but significantly (ANOSIM p < 0.05) different from the initial 3-day biofilm (similarity e 60%). Part of these differences is believed to be due to the necessity [due to large volumes of sediments and water required (100 kg and 1000L respectively)] to use the same sediment and water for each CB period. As such, there was likely to be a cumulative effect/evolution on the microbial community between runs (i.e. while the 9 day consolidation/biostabilization run was settled for the required 9 days following disturbance, the water and sediment has been in the flume for 20 days by the time the 9 day run was completed). Regardless, the variation in microbial population does correlate to the observed differences in bed strength observed (Fig. 1). This is particularly true for the biofilm which shows similar microbial populations for the 6 and 9 day runs (Figure S4) which corresponded with similar bed sediment stability (Fig. 1). It is interesting to note that although the biofilms are similar in microbial composition to the water components, when accounting for the erosion of sediment, the microbial populations change as shown with the principal component Figure S4. Such a change would suggest that the microbial community associated with the eroded flocs is different from the substantive biofilm possibly due to differences in the physical substratum and/or differences in the substrates available for metabolic activity [bacteria to bacteria (and bacteria-products) vs. bacteria to sediment particle]. 4. Conclusions It is clear that shear levels, the microbial consortium and SS with associated contaminant concentrations have an integrated and interactive symbiotic relationship within riverine systems that mediates the transport of contaminants to downstream locations. One cannot be studied without accounting for the influence of the others. In this study, sediments with a hydrophobic property (made so by natural oil content) demonstrate unique transport and particle characteristics (small size, poor settling). As “oil and water do not mix”, it is not surprising to suggest that high hydrophobicity influence the erosion, transport and fate of sediment and associated contaminants. The microbial consortium varied between the four components of the river system (water, bed sediment, suspended floc and biofilms) with the fungal population appearing to be an integral aspect with possible influences on bed sediment stability. Biofilms were observed to increase with CB times and generated a unique microbial consortium relative to the eroded flocs. While PAHs and NAs increased with increasing shear (indicative of increasing loads), they tended to decrease with time (CB) suggesting a chemical and/or biological degradation. PAH and NA consuming microbes decreased with time as well which may suggest that there was a reduced pool of PAHs and NAs available resulting in their die off. This study emphasizes the importance that any management strategies and operational assessments for the

I.G. Droppo et al. / Water Research 92 (2016) 121e130

129

Fig. 6. Principal component analysis of all CB samples collected at each shear step including, initial bioflims, water collected before shear and after the completion of final shear steps, and pre (bulk river sample) and post sieved (<75 mm) sediments.

protection of human and aquatic health incorporate the sediment (SS and bed sediment) and biological (biofilm) compartments and the energy dynamics within the system in order to better predict contaminant transport. Acknowledgements The authors wish to acknowledge the technical support of Christina Jaskot, Julie Roy and George D.W. Swerhone of Environment Canada, Burlington and Saskatoon. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.watres.2016.01.034. References Alberta Environment, 1982. Soil Inventory of the Alberta Oil Sands Environmental Research Program Study Area (Report No. 122). Retrieved from. http://sis.agr.gc. ca/cansis/publications/surveys/ab/ab42/ab42_report.pdf. Amos, C.L., Umgiesser, G., Ferrarin, C., Thompson, C.E.L., Whitehouse, R.J.S., Sutherland, T.F., Bergamasco, A., 2010. The erosion rates of cohesive sediments in Venice lagoon, Italy. Cont. Shelf Res. 30 (8), 859e870. Amos, C.L., Fenney, T., Sutherland, T.F., Luternauer, J.L., 1997. The stabiilty of finegrained sediment from the Fraser River delta. Estuar. Coast. Shelf Sci. 45 (4), 507e524. Barlocher, F., Murdoch, J.H., 1989. Hyporheic biofilms - a potential food source for interstitial animals. Hydrobiologia 184, 61e67. Berkhout, N., 1994. Manual “UHCM”, Ultra High Concentration Meter. Delft Hydraulics, Rotterdam. Bester, E., Wolfaardt, G.M., Aznaveh, N.B., Greener, J., 2013. Biofilms' role in planktonic cell proliferation. Int. J. Mol. Sci. 14 (11), 21965e21982. Bonilla, T.D., Nowosielski, K., Cuveiler, M., Hartz, A., Green, M., Esiobu, N., McCorquodale, D.S., Fleisher, J.M., Rogerson, A., 2007. Prevalence and distribution of fecal indicator organisms in South Florida beach sand and preliminary assessment of health effects associated with beach sand exposure. Mar. Pollut. Bull. 54, 1472e1482. Boon, N., Windt, W.D., Verstraete, W., Top, E.M., 2002. Evaluation of nested PCR DGGE (denaturing gradient gel electrophoresis) with group-specific 16S rRNA primers for the analysis of bacterial communities from different wastewater treatment plants. FEMS Microbiol. Ecol. 39, 101e112. Brinkmeyer, R., Amon, R.M.W., Schwarz, J.R., Saxton, T., Roberts, D., Harrison, S., Ellis, N., Fox, J., DiGuardi, K., Hochman, M., Duan, S., Stein, R., Elliott, C., 2015. Distribution and persistence of Escherichia coli and Enterococci in stream bed and bank sediments from two urban streams in Houston, TX. Sci. Tot. Environ. 502 (01), 650e658. Clarke, K.R., 1993. Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol 18, 117e143. Carson, M.A., 1990. Evaluation of Sediment Data for the Lower Athabasca River Basin Alberta. Calgary, Alberta: Water Resources Branch Inland Water

Directorate. Environment Canada, Ottawa. Chapman, D.W., 1988. Critical review of variables used to define effects of fines in redds and large salmonids. Trans. Am. Fish. Soc. 117, 1e21. Chefetz, B., Deshmukh, A.P., Hatcher, P.G., 2000. Pyrene sorption by natural organic matter. Environ. Sci. Technol. 34, 2925e2930. Cheng, H., Hu, E., Hu, Y., 2012. Impact of mineral micropores on transport and fate of organic contaminants: a review. J. Contam. Hydrol. 129e130, 80e90. Clemente, J.S., Fedorak, P.M., 2005. A review of the occurrence, analyses, toxicity, and biodegradation of naphthenic acids. Chemosphere 60, 585e600. Conly, F.M., Crosley, R.W., Headley, J.V., 2002. Characterizing sediment sources and natural hydrocarbon inputs in the lower Athabasca River, Canada. J. Environ. Eng. Sci. 1, 187e199. Droppo, I.G., 2009. Biofilm structure and bed stability of five contrasting freshwater sediments. Mar. Freshw. Research 60, 690e699. Droppo, I.G., 2001. Rethinking what constitutes suspended sediment. Hydrol. Proc. 15, 1551e1564. Droppo, I.G., Andrea, L.D., Krishnappan, B.G., Jaskot, C., Trapp, B., Basuvaraj, M., Liss, S.N., 2015. Understanding sediment dynamics of the ells river: towards improved modelling of contaminant source, fate and effect. J. Soils Sed. 15, 467e479. Droppo, I.G., Liss, S.N., Williams, D., Nelson, T., Jaskot, J., Trapp, B., 2009. The dynamic existence of waterborne pathogens within river sediment compartments e implications for water quality regulatory affairs. Env. Sci. Technol. 43 (6), 1737e1743. Droppo, I.G., Liss, S.N., Williams, D., Leppard, G.G., 2006. River sediment/pathogen interactions: importance for policy development of safe water practices. In: Rowan, J.S., Duck, R.W., Werritty, A. (Eds.), Sediment Dynamics and the Hydromorphology of Fluvial Systems, pp. 314e321. IAHS Pub. No. 306. Droppo, I.G., Lau, Y.L., Mitchell, C., 2001. The effect of depositional history on contaminated bed sediment stability. Sci. Total Environ. 266 (1e3), 7e13. Elliott, A.V.C., Warren, L.A., 2014. Microbial engineering of floc Fe and trace element geochemistry in a circumneutral, Remote Lake. Environ. Sci. Technol. 48 (12), 6578e6587. Elliott, A.V.C., Plach, J.M., Droppo, I.G., Warren, L.A., 2014. Collaborative microbial Fe-redox cycling by pelagic floc bacteria across wide ranging oxygenated aquatic systems. Chem. Geol. 336, 90e102. Environment Canada, 2013a. Method 607.1, Oil Characterization in Environmental Samples, Prairies and Northern Laboratory for Environmental Testing. Revision No. 5, July 12, Edmonton, Alberta, Canada. Environment Canada, 2013b. Method 612.1, Oil Characterization in Environmental Samples, Prairies and Northern Laboratory for Environmental Testing. Revision No. 5, July 18, Edmonton, Alberta, Canada. Environment Canada, 2011. Lower Athabasca Water Quality Monitoring Plan: Phase 1 (Catalogue No. En14e42/2011E-PDF). ISBN 978-1-100-18471-5, Retrieved from. http://www.ec.gc.ca/Content/8/A/1/8A1AB11A-1AA6-4E12-937360CF8CF98C76/WQMP_ENG.pdf. Garcia-Aragon, J., Droppo, I.G., Krishnappan, B.G., Trapp, B., Jaskot, C., 2011. Erosion characteristics and floc strength of Athabasca River cohesive sediments: towards managing sediment related issues. J. Soils Sed. 11, 679e689. Gerbersdorf, S.U., Wierpecht, S., 2015. Biostabilization of cohesive sediments: revisiting the role of abiotic conditions, physiology and diversity of microbes, polymeric secretion, and biofilm architecture. Geobiology 13, 68e97. Grabowski, R.C., Droppo, I.G., Wharton, G., 2011. Erodibility of cohesive sediment: the importance of sediment properties. Earth Sci. Rev. 105, 101e120. Haines, J.R., Wrenn, B.A., Holder, E.L., Strohmeier, K.L., Herrington, R.T., Venosa, A.D., 1996. Measurement of hydrocarbon-degrading microbial populations by a 96-

130

I.G. Droppo et al. / Water Research 92 (2016) 121e130

well plate most-probable-number procedure. J. Ind. Microb. 16 (1), 36e41. Hartz, A., Cuvelier, M., Nowosielski, K., Bonilla, T.D., Green, M., Esiobu, N., McCorquodale, D.S., Rogerson, A., 2008. Survival potential of Escherichia coli and enterococci in subtropical beach sand: implications for water quality managers. J. Environ. Qual. 37, 898e905. Karickhoff, S.W., Brown, D.S., Scott, T.A., 1979. Sorption of hydrophobic pollutants on natural sediments. Water Research 13, 241e248. Krishnappan, B.G., 1993. Rotating circular flume. J. Hydraul. Eng. ASCE 119 (6), 758e767. Krishnappan, B.G., Engel, P., 2004. Distribution of bed shear stress in rotating circular flume. J. Hydraul. Eng. ASCE 130 (4), 324e331. Larkin, P.A., 1978. BiometricsdA Handbook of Elementary Statistical Tests. University of British Columbia, Vancouver, British Columbia, Canada. Lawrence, J.R., Topp, E., Waiser, M.J., Tumber, V., Roy, J., Swerhone, G.D.W., Leavitt, P., Paule, A., Korber, D.R., 2015. Resilience and recovery: the effect of triclosan exposure timing during development, on the structure and function of river biofilm communities. Aquat. Toxicol. 161, 253e266. Lawrence, J.R., Chenier, M., Roy, R., Beaumier, D., Fortin, N., Swerhone, G.D.W., Neu, T.R., Greer, C.W., 2004. Microscale and molecular assessment of the impacts of nickel, nutrients and oxygen level on river biofilm communities. Appl. Environ. Microb. 70, 4326e4339. Liss, S.N., Droppo, I.G., Flannigan, D., Leppard, G.G., 1996. Floc architecture in wastewater and natural riverine systems. Environ. Sci. Technol. 30 (2), 680e686. Maa, J.P.-Y., Kim, S.-C., 2002. A constant erosion rate model for fine sediment in the York River, Virginia. Environ. Fluid Mechan. 1, 343e360. Muyzer, G., Ramsing, N.B., 1995. Molecular methods to study the organization of microbial communities. Wat. Sci. Technol. 32, 1e9. Muyzer, G., De Waal, E.C., Uitterlinden, A.G., 1993. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl. Environ. Microbiol. 59, 695e700. Neu, T.,R., Swerhone, G.D.W., Lawrence, J.R., 2001. Assessment of lectin-bindinganalysis for in situ detection of glycoconjugates in biofilm systems. Microbiology 147, 299e313. Rosenberg, M., Gutnick, D., Rosenberg, E., 1980. Adherence of bacteria to

hydrocarbons: a simple method for measuring cell-surface hydrophobicity. FEMS Microb. Lett. 9 (1), 29e33. Saulnier, I., Mucci, A., 2000. Trace metal remobilization following the resuspension of estuarine sediments: Saguenay Fjord. Can. Appl. Geochem. 12, 203e222. Stone, M., Emelko, M.B., Droppo, I.G., Silins, U., 2011. Biostabilization and erodibility of cohesive sediment deposits in wildfire-affected streams. Water Research 45 (2), 521e534. Stone, M., Krishnappan, B.G., Emelko, M.B., 2008. The effect of bed age and shear stress on the size distribution and particle morphology of eroded cohesive sediment deposits in an annular flume. Water Research 42 (15), 4179e4187. Stone, M., Krishnappan, B.G., 1997. Transport characteristics of Tile-Drain sediments from an agricultural watershed. Water Air. Soil Poll. 99, 88e103. Walters, E., Schwarzwalder, K., Rutschmann, P., Muller, E., Horn, H., 2014a. Influence of resuspension on the fate of fecal indicator bacteria in large-scale flumes mimicking an oligotrophic river. Water Research 48, 466e477. Walters, E., Schwarzwalder, K., Rutschmann, P., Muller, E., Horn, H., 2014b. Persistence of fecal indictor bacteria in sediment of an oligotrophic river: comparing large and lab-scale flume systems. Water Research 61, 276e287. Wang, Z., Yang, C., Parrott, J.L., Frank, R.A., Yang, Z., Brown, C.E., Hollebone, B.P., Landriault, M., Fieldhouse, B., Liu, Y., Zhang, G., Hewitt, L.M., 2014. Forensic source differentiation of petrogenic, pyrogenic, biogenic hydrocarbons in Canadian oil sands environmental samples. J. Hazard. Mater. 271, 166e177. Widdows, J., Friend, P.L., Bale, A.J., Brinsley, M.D., Pope, N.D., Thompson, C.E.L., 2007. Inter-comparison between five devices for determining erodability of intertidal sediments. Contin. Shelf Res. 27, 1174e1189. Yallop, M.L., Paterson, D.M., Wellsbury, P., 2000. Interrelationships between rates of microbial production, exopolymer production, microbial biomass, and sediment stability in biofilms of intertidal sediments. Microb. Ecol. 39 (2), 116e127. Yergeau, E., Lawrence, J.R., Sanschagrin, S., Roy, J.L., Swerhone, G.D.W., Korber, D.R., Greer, C.W., 2013. Aerobic biofilms grown from Athabasca watershed sediments are inhibited by increasing bituminous compounds concentrations. Appl. Environ. Microb. 79 (23), 7398e7412. Yergeau, E., Lawrence, J.R., Sanschagrin, S., Waiser, M.J., Korber, D.R., Greer, C.W., 2012. Next-generation sequencing ofmicrobial communities in the Athabasca River and its tributaries in relation to oil sands mining activities. Appl. Environ. Microbiol. 78, 7626e7637.