Dechlorinating Microorganisms in a Sedimentary Rock Matrix

Article pubs.acs.org/est

Dechlorinating Microorganisms in a Sedimentary Rock Matrix Contaminated with a Mixture of VOCs Gláucia Lima,* Beth Parker, and Jessica Meyer School of Engineering, University of Guelph, 50 Stone Road East, Guelph, Ontario, N1G 2W1, Canada S Supporting Information *

ABSTRACT: Microbiological characterizations of contaminant biodegradation in fractured sedimentary rock have primarily focused on the biomass suspended in groundwater samples and disregarded the biomass attached to fractures and in matrix pores. In fractured sedimentary rock, diffusion causes nearly all contaminant mass to reside in porous, lowpermeability matrix. Microorganisms capable of contaminant degradation can grow in the matrix pores if the pores and pore throats are sufficiently large. In this study, the presence of dechlorinating microorganisms in rock matrices was investigated at a site where a fractured, flat-lying, sandstone− dolostone sequence has been contaminated with a mixture of chlorinated and aromatic hydrocarbons for over 40 years. The profile of organic contaminants as well as the distribution and characterization of the microbial community spatial variability was obtained through depth-discrete, high-frequency sampling along a 98-m continuous rock core. Dechlorinating microorganisms, such as Dehalococcoides and Dehalobacter, were detected in the rock matrices away from fracture surfaces, indicating that biodegradation within the rock matrix blocks should be considered as an important component of the system when evaluating the potential for natural attenuation or remediation at similar sedimentary rock sites.

■

INTRODUCTION Chlorinated solvents are one of the most common contaminants found in fractured sedimentary rock aquifers due to their persistence and mobility.1 However, more than 55% of the sedimentary fractured rock sites registered by the U.S.EPA have detectable biodegradation products, e.g., cis-dichloroethene (cisDCE), in groundwater samples.1 Characterization of biodegradation processes of chlorinated solvents in sedimentary rocks have exclusively focused on the analysis of suspended biomass in groundwater samples.2−6 Suspended biomass is expected to have different species richness (diversity and abundance) and physiological capabilities7 than biomass attached to fractures and rock matrix pores, if present. Furthermore, most of the contaminant mass in fractured sedimentary rocks resides in the low-permeability matrix blocks between fractures due to diffusion-driven mass transport.8 For this study, the rock matrix was considered a suitable location for contaminant biodegradation; this is a valid hypothesis, considering that matrix pores are large enough to harbor microorganisms. The presence of dechlorinating microorganisms (dechlorinators) in the primary matrix porosity of sedimentary rock has yet to be investigated. Only two publications have used sedimentary rock core samples in biodegradation studies at sites contaminated with chlorinated solvents. Bradley et al.5 deployed beakers filled with crushed rock samples inside a borehole for a year. DNA was extracted from the crushed rock, © 2012 American Chemical Society

which had been colonized with suspended (groundwater) biomass. Four of their five samples had Dehalococcoides species, the only bacteria reported capable of reductive dechlorination of trichloroethene (TCE) and tetrachloroethene (PCE) to ethene.9 Darlington et al.4 investigated the influence of sandstone minerals on biotic and abiotic degradation of TCE by adding air-dried and pulverized autoclaved rock to live groundwater enrichments. Neither of these studies investigated any link between dechlorination and the presence of microorganisms resident within the sedimentary rock matrices. Specifically, the methodologies adopted would not distinguish between microorganisms on rock fragments from mobile groundwater versus within fragment pores. The purpose of this study was to investigate whether dechlorinating microorganisms can be detected within the rock matrix itself. Rock core samples were taken from a vertical continuous core through a sequence of flat-lying sandstones and dolostones with known groundwater contamination involving a mixture of chlorinated solvents, ketones, and aromatics. Microbial DNA extracted from these rock core samples was analyzed by molecular methods. The field site has well characterized lithology and hydrogeological/hygeodroReceived: Revised: Accepted: Published: 5756

January 17, 2012 May 10, 2012 May 11, 2012 May 21, 2012 dx.doi.org/10.1021/es300214f | Environ. Sci. Technol. 2012, 46, 5756−5763

Environmental Science & Technology

Article

chemical conditions,10 and the contaminants have had over 40 years to diffuse from fractures into the matrix. Sampling at high frequency and at different distances away from fractures was adopted to represent the range of conditions (e.g., porosity, mineralogy, hydrochemistry, redox zonation) observed in the rock core that could influence the development of a microbial community capable of volatile organic compound (VOC) biodegradation. Due to expected low DNA concentrations in most samples, end-point nested polymerase chain reaction (EPPCR) targeted Dehalococcoides, Dehalobacter, Geobacter, and Sulf urospirillum as indicators of microbial reductive dechlorination,9,11−13 as these microorganisms have been detected in groundwater samples from other contaminated fractured rock sites.2,3,6,14 Denaturing gradient gel electrophoresis (DGGE) fingerprints were obtained from selected rock core DNA samples to compare microbial community structures and their variation in the rock matrix at different depths, lithologies, and distances to fractures. Many have speculated that matrix pore sizes are a limiting factor for microbial growth in sedimentary rock matrices. However, microbial growth in the matrix pores of sandstones is evident in both field samples and laboratory experiments. Sandstone matrix samples from a sandstone−shale sequence contained viable cells that grew at the expense of volatile fatty acids from an adjacent shale unit.15 In the laboratory, both nonmotile and motile microorganisms (Klebsiella pneumonia and Enterobacter aerogenes, respectively) were able to fully penetrate the matrix of the very porous Berea sandstone.16 Thus, some sandstones have adequate pore sizes and interconnectivity to harbor cells and allow their migration and growth.15,16 However, the contribution to contaminant degradation of dechlorinating microorganisms in the primary porosity of sedimentary rocks is unclear. This paper aims to show, through direct application of molecular methods to depth-discrete rock core samples, the presence of microorganisms responsible for organic contaminant biodegradation. The location of this microbial activity, be it in the fractures or in the matrix, is critical for informed decision-making regarding transport and fate of contaminant plumes and prospects for application of in situ degradation technologies to achieve desired end points within reasonable time or distance scales. If dechlorinators exist within and throughout the rock matrices, the contaminant mass contained in the matrix can be more easily accessed by biodegrading microbial consortia, as long as redox and substrate conditions are not limiting. This in situ degradation will reduce long-term effects of back-diffusion, which causes rebound of contamination in dual porosity systems,17,18 such as fractured sedimentary rocks.

Formation. The dissolved phase plume in the Lone Rock Formation has expanded to encompass an area of 4.24 km2 but, since 2003, a groundwater pump-and-treat system for hydraulic containment of the source zone mass discharge has caused the VOC concentrations in monitoring wells at and near the plume front to decline and the downgradient plume extent to diminish. Rock Core Sampling. Continuous coring was conducted in a vertical borehole (MP-19D) from the top of rock (12.1 m bgs) to 97.4 m bgs, using an HQ3 wireline coring system and tap water as the primary drilling fluid. The HQ3 system includes a split, stainless steel core sleeve, an inner core barrel, and an outer tube to which the coring bit is attached. Each run resulted in a 1.5-m long, 6-cm diameter core, which was retrieved and placed on a clean, aluminum foil-lined PVC tray where lithology and fractures were logged in detail. Sampling locations in the core were selected in accordance with the observed variability in hydrogeologic featuressuch as lithology, mineralogy, matrix porosity, and fracture occurrencewith the intent of representing observed and anticipated variability with increasing distances from fractures, where groundwater and contaminants migrate preferentially by advection. Rock matrix samples were taken at various distances from fractures, including above and below these features, as well as where lithology and mineralogy represented the variety of matrix porosity conditions and distances from preferred contaminant migration pathways, i.e., fractures and fracture network (Supporting Information, SI, Figure S1). Microbial sampling frequency was higher in the Lone Rock Formation because this zone represented the position of the aerially extensive groundwater plume. Rock core samples were collected in ∼5 cm lengths using a chisel and hammer, then split into two halves along the core axis: one half for DNA extractions and the other half for VOC analyses. Before each new sample was collected, the chisel was rinsed with methanol, then with deionized water, and dried with a clean laboratory wipe. Samples for DNA extractions were tightly wrapped in a clean sheet of aluminum foil, followed by clear plastic wrap and plastic paraffin film. Each sample was placed in a ziptop bag and shipped on ice to the University of Guelph. Samples for VOC analyses were immediately trimmed, crushed with a hydraulic rock crusher in the field, and preserved with purge and trap grade methanol (Sigma-Aldrich). VOC samples were preserved in 40-mL VOA bottles with PTFE-lined septa with screw-caps (VWR) and shipped overnight to the University of Guelph for extraction. This method for core collection and sampling is considered to be appropriate for the goals of this paper because assessment of the core barrel construction, pressure differentials surrounding the coring during/after drilling, and data from QA/QC samples and supporting complementary data sets indicate negligible invasion of drilling fluids into the core.19 The validity of the method is supported by other published studies.19,20 In the laboratory, VOC extraction was performed with a shake-flask method.21 Aliquots from the shake-flask method were screened for 36 VOCs by Stone Environmental laboratory (Montpelier, VT) according to EPA method SW8260B.22 Comprehensive QA/QC procedures were used to validate rock VOC concentrations through application of a modified EPA validation process.23 To support interpretation of rock formation properties, a total of 64 rock core samples were taken for physical property measurements from cored locations within 2 km of MP-19D

■

MATERIALS AND METHODS Study Site. The study site is located in Dane County, Wisconsin and comprises unconsolidated Pleistocene glacial deposits overlying Cambrian−Ordovician sandstones and dolostones. The site location, hydrogeological units, and groundwater flow are presented by Meyer et al.10 Between 1950 and 1970, a mixture of organic solvents in the nonaqueous phase, composed mainly of PCE, TCE, 1,1,1trichloroethane (1,1,1-TCA), BTEX (benzene, toluene, ethyl benzene, and xylenes), ketones, halomethanes, and chlorofluorocarbon compounds (CFCs) at low concentrations, was released into the overburden and underlying bedrock beneath the site, and preferentially accumulated and persisted between 41 and 57 m below ground surface (m bgs) in the Lone Rock 5757

dx.doi.org/10.1021/es300214f | Environ. Sci. Technol. 2012, 46, 5756−5763

Environmental Science & Technology

Article

that represented all the dominant lithologies associated with the stratigraphic units encountered at the site (SI). The stratigraphic units are generally flat lying, of nearly constant thickness, and have fairly uniform mineralogical and physical rock matrix properties across the site.10 Wet bulk density and imbibition porosity were determined gravimetrically.24 Permeability was determined with a nitrogen gas permeameter according to ASTM Standard Method D 4525-04.25 Total and interparticle porosities and pore radii were determined by mercury intrusion porosimetry26 on 22 samples. Backscattered scanning electron microscopy (BSEM) images were obtained from the same samples.26 Fraction of organic carbon (foc) was determined for 20 samples.27 DNA Extraction. To avoid sample contamination due to airborne microorganisms, samples were opened close to a flame in the laboratory. All materials (e.g., chisel, spatulas, and crushing cells) were washed with Eliminase (Decon Laboratories), rinsed with nanopure water, and flame-sterilized between samples. A minimum of 1 cm of the outer edge of each rock sample was trimmed off with a chisel to remove any drilling fluid or presence of ex situ microorganisms that might be present due to drilling, collection, and/or handling at surface. This has been shown effective by Kimura et al.19 The trimmed piece (about 1 cm3) was transferred with a flame sterilized spatula into a flame sterilized crushing cell and crushed using a hydraulic rock crusher. The crushed material was transferred to a sterile 7 oz. (207 mL) Whril Pack bag (Nasco, VWR) and stored at −20 °C until processed. DNA was extracted from 1 g of crushed rock using the Mo Bio PowerSoil DNA extraction kit (Carlsbad, CA) with final elution with 50 μL of Ultra Pure DNase/RNase-free sterile water (Invitrogen). DNA was extracted only from matrix samples to avoid parts of the core that were most likely affected by contamination during drilling and handling, with exception of three samples (locations II, III, and IV in Figure 1d, explained below). DNA templates from 66 samples were obtained: 39 from the Lone Rock Formation, 22 from sandstones above and below the Lone Rock Formation, and 5 from dolostone units. DNA concentrations varied from 3 to 98 ng/μL (50 μL template), which translates to 0.1 to 4.9 μg/g of crushed wet rock. Samples were taken at different distances from fracture surfaces, recognizing that fractures represent advective and primary transport pathways for the contaminant plume. Most of the fractures observed in the Lone Rock Formation at MP19D occur along bedding plane partings between three different sublithologies. These three sublithologies are28 (i) a flat pebble conglomerate in a glauconite matrix, (ii) a laminated quartzglauconite feldspathic sandstone, and (iii) a bioturbated quartzglauconitic sandstone. To investigate variation in the microbial community at the centimeter scale level, samples I, II, III, and IV (positions indicated on the right of Figure 1d) were selected and subsampled at 1 cm intervals, at increasing distances from observed fractures. Core sample I (42.85−42.88 m bgs; 2 subsamples) was from the bioturbated sandstone; sample II (43.02−43.07 m bgs; 3 subsamples) was from the glauconite rich conglomerate; and samples III (42.43−43.48 m bgs; 4 subsamples) and IV (45.18−45.28 m bgs; 3 subsamples) were from the laminated sandstone. The top and bottom centimeter of each sample was trimmed off and each subsample represented, on average, 1 cm depth intervals (SI, Figure S1). End Point Nested PCR and DGGE on 16S rRNA Genes. Five EP-PCR protocols targeting 16S rRNA genes were performed with specific primers (SI, Table S1) for

Figure 1. (a) VOC concentration profile in the Lone Rock Formation (μg/g of rock); (b) Lone Rock Formation sublithologies; (c) microorganisms detected by nested EP-PCR and DGGE, (+) indicate positive detections by PCR and ( × ) by DGGGE; (d) number of DGGE bands (m bgs) scored with Gene Tools (Syngene). Horizontal red lines between data columns represent visible fracture positions. Horizontal black arrow on the right of (d) indicates the position of a matrix sample 64 cm below a visible fracture. Subsampled sections I, II, III, and IV are detailed in Figure 3. Whenever symbols may look “bold” it is the result of superimposition due to tight sampling interval. Also, symbols that may look like a star are superimposition of + and × , meaning that a given genera was detected by both EP-PCR and DGGE.

Dehalococcoides,29 Dehalobacter,30 Sulf urospirillum,31 and Geobacter.31 To improve detection of dechlorinators possibly in low abundance and to reduce biases toward the most abundant groups,32 nested PCR was applied to all amplifications with these specific primers. The initial amplification was performed with primers 8f/1541r33 and 0.1−2 μL of DNA template. Following this step, 1.0 μL of 8f/1541r PCR products (from undiluted, and 1:10 and 1:50 dilutions) were used as templates for the second PCR with specific primers. Reaction volumes were 20 μL, except for PCR-DGGE when 25 μL was used. Archaea was detected with universal primers 1Af/1100Ar.34 Success of PCR was verified in agarose gels (1% ethidium bromide). DGGE fingerprints were obtained for 29 samples from the Lone Rock Formation according to Muyzer et al.35 with modifications.31 All visible DGGE bands were excised, soaked overnight in 100 μL of RNase/DNase-free water, and reamplified with the DGGE primers and 2 μL of the soaking water as the template. PCR products were then purified with 5758

dx.doi.org/10.1021/es300214f | Environ. Sci. Technol. 2012, 46, 5756−5763

Environmental Science & Technology

Article

the Mo Bio PCR Purification kit (Carlsbad, CA) and sequenced at the Genomics Facility at the University of Guelph. EP-PCR products obtained with specific primers were also purified and sequenced to confirm detections. Sequences were checked for close similarities (>95%) using BLAST (http://blast.ncbi.nlm. nih.gov). The number of bands in DGGE gels were scored with Gene Tools software v. 4.01 (Syngene), considering the lowest band intensity as 2% of the brightest band in each gel. Percent similarities among microbial communities were determined for DGGE fingerprints using clustering analysis with Gene Directory software v. 2.01a (Syngene), with the Jacquard similarity coefficient and unweighted pair group method with arithmetic mean (UPGMA). Relative band positions (1% tolerance) were also used for band identification.

■

RESULTS AND DISCUSSION VOC Concentration Profile. The complete VOC concentration profile (Figure S2) obtained from the analysis of numerous depth-discrete samples of the porous rock matrix indicates that most of the detectable contaminant mass (98%) is found in the Lone Rock Formation, between 39 and 45.2 m bgs (Figure 1). Contaminant concentrations can vary several orders of magnitude over short distances. Above the Lone Rock Formation, the majority of the VOCs in the rock matrix samples were below method detection limits (MDL), with the exception of BTEX concentrations which were between MDL and the limit of quantitation (LOQ). Below the Lone Rock Formation, a few samples had dichloromethane (DCM), 1,1,1TCA, and cis-DCE at concentrations below LOQ. The rock core VOC results for MP-19D (Figure 1) demonstrate that the majority of the detectable contaminant mass resides in the Lone Rock Formation, which has a low-permeability matrix. Comparison of contaminant composition and concentrations between the source area and the core location (MP-19D) suggests that contaminant degradation occurs along the groundwater flow path. The relative abundance of each contaminant group (halothanes, haloethe(a)nes, BTEX) in the rock core at MP-19D is different from the DNAPL composition at the source located 672 m upgradient. The DNAPL is dominated by 1,1,1-TCA, PCE, TCE, carbon tetrachloride (CT), and BTEX.36 At MP-19D, the relative abundance of BTEX drops from 21% in the DNAPL to 4% in the rock matrix samples when dissolved and sorbed phases are considered on a mole fraction basis. However, the abundance of chlorinated ethanes and ethenes increases from 66% in the source DNAPL to 69% in the rock matrix at MP-19D, with 44% of the total being 1,2-dichloroethane (1,2-DCA) and cisDCE, common degradation products of the predominant source constituents. Increases in cis-DCE concentrations over time have also been observed in groundwater samples from wells at the site.36,37 The abundance of halomethanes also increased from 5% in the source to 16% in the rock core profile, with DCM as the main halomethane detected. Overall, the contaminant composition found in matrix samples from the Lone Rock Formation reveals a mixture of organic compounds that includes electron acceptors (chlorinated compounds), donors, and sources of carbon (aromatics and ketones), which can sustain biodegradation. Cell Sizes Compared to Matrix Pore Sizes. The pore radii of the sandstones at the study site determined using mercury intrusion porosimetry vary between 0.8 and 18.9 μm. However, pore radii close to 50 μm were observed in BSEM images of sandstones (Figure 2). Besides pore size distribution

Figure 2. BSEM images: (a) laminated sandstone (Lone Rock Formation), (b) bioturbated sandstone (Lone Rock Formation), (c) Prairie du Chien Group (dolostone), and (d) St. Lawrence Formation (dolostone). Green arrow indicates a pore; red arrow indicates a pore throat. Bar scale equals 100 μm.

and interconnectivity, pore throats impose bottleneck constraints (red and green arrows in Figure 2b) to microbial growth and migration in rock matrices as their geometry and dimensions vary dramatically depending on grain sizes, sorting, and cementation. Based on pore-to-throat size ratios of about 3 reported for sandstones similar to those from the study site,26 pore throat radii for the Lone Rock Formation sandstones are estimated to vary from 0.025 to 6.3 μm. For comparison, Dehalococcoides cell diameters are about 0.5 μm.38 Microbial DNA was detected with at least one of the applied PCR methods in all 61 sandstone samples analyzed. Therefore, growth and migration of some dechlorinating microorganisms into the rock matrix of the Lone Rock Formation is physically possible. Three of five dolostone samples (two from the Prairie du Chien Group and one from the St. Lawrence Formation) had no amplifiable DNA (DNA concentrations 95%) to 16S rRNA gene fragments of Acetobacterium, Polaromonas, Methylophilus, and others (Table 1), all of which have been associated with dechlorination of a wide range of contaminants, including chloroethenes, chloroethanes, and chloromethanes. For instance, a species from the genus Polaromonas can grow solely on cis-DCE,40 while some Methylophilus species can use DCM as their sole carbon source.41 Moreover, species of Pseudomonas and Acetobacterium are capable of cometabolic degradation of chloromethanes and chloroethenes. DNA fragments of Sulf urospirillum (ε-Proteobacteria) and Dehalococcoides (Cloroflexi) were detected both by EP-PCR and DGGE band sequences (100% similarities in both casesFigures 1 and 3). In some cases, 16S rRNA fragments of both aerobes and anaerobes were detected in the same sample. For instance, sample II from 43 m bgs (Figure 1) had 16S rRNA fragments with high (>95%) similarities to Dehalococcoides and Sulf urospirillum, both obligate anaerobes, as well as to Polaromonas, which is aerobic (subsamples in Figure 3II). Hydrochemistry conditions in groundwater, based on conventional and depthdiscrete multilevel wells installed throughout the site, show a full range of redox conditions, but in general groundwater is aerobic, except where high concentrations of contaminants are present. So, redox zonation at micro- or nanoscales42 may exist in the sedimentary rock matrix at the study site. These results, based on amplification of 16S rRNA gene fragments, clearly show potential for dechlorination in the Lone

Figure 3. (a) Variation in detected microorganisms with distance to fractures (red lines) in Lone Rock Formation for subsamples at depth intervals I, II, III, and IV for EP-PCR (+) and DGGE ( × ). DNA from fracture surfaces was analyzed in samples II, III, and IV. Subsampling interval was 1 cm. (b) Number of DGGE bands in each subsample. (c) Sublithologies (photos are only illustrations of sample configurations): Sample I − bioturbated sandstone; Sample II − lithology change laminated sandstone to glauconitic conglomerate sandstone; Sample III − lithology change bioturbated sandstone to glauconitic conglomerate sandstone; and Sample IV − laminated sandstone and lithology change to glauconitic conglomerate sandstone at the bottom. (d) Total VOC per gram of wet rock in samples I to IV. Fracture positions in samples I to IV were 42.88, 43.02, 43.43, and 45.18 m bgs, respectively.

Rock Formation rock matrix, but cannot be used to ascertain if the microorganisms are active or not. However, compoundspecific isotope analysis (CSIA) performed in groundwater samples from the Lone Rock Formation demonstrated contaminant degradation along the flow path, with 80% of the TCE and 1,1,1-TCA degrading by midplume, which is close to the cored location, and persistence of cis-DCE.37 In addition, daughter products are measured throughout matrix samples 5760

dx.doi.org/10.1021/es300214f | Environ. Sci. Technol. 2012, 46, 5756−5763

Environmental Science & Technology

Article

may be less diverse than in groundwater samples, but their presence can significantly impact contaminant attenuation and this aspect of these fractured rock systems should not be overlooked or ignored.

with lack of parent compounds. Even though CSIA was performed in groundwater samples and acknowledging that redox conditions in fast flowing fractures can differ from matrix porewater, CSIA results demonstrate that contaminant degradation is occurring at the site. However, low cell numberssuggested by low DNA concentrations in most samplesand evidence of aerobic to mildly anaerobic conditions in groundwater throughout many portions of the contaminant plume37 suggest slow degradation rates and spatial variability of redox conditions from aerobic in the fractures to more reducing in the matrix, where most of the contaminant mass resides. Heterogeneities and Dechlorinators with Distance to Fractures. Heterogeneities in microbial communities throughout the matrix are demonstrated by variations detected with distance to fractures, as observed in all subsamples from the four different depth intervals analyzed in the Lone Rock Formation (Figure 3). Cluster analysis in DGGE fingerprints (Figure S5) of microbial communities along the rock core profile confirms this heterogeneity as fingerprint similarities were below 30% for most samples and subsamples. Dechlorinators (Polaromonas, Methylophilus, Sulf urospirillum, Dehalococcoides, and Dehalobacter) were detected in the rock matrix, both close to fractures (