Correspondence: William M. Moe, 3418G CEBA Building, Department of Civil and Environmental Engineering, Louisiana State University, Baton Rouge, LA 70803, USA. Tel.: +225 578 9174; fax: +225 578 8652; e-mail: firstname.lastname@example.org
Bacterial concentration and diversity was assessed in a moderately acidic (pH 5.1) anaerobic groundwater contaminated by chlorosolvent-containing DNAPL at a Superfund site located near Baton Rouge, Louisiana. Groundwater analysis revealed a total aqueous-phase chlorosolvent concentration exceeding 1000 mg L−1, including chloroethanes, vinyl chloride, 1,2-dichloropropane, and hexachloro-1,3-butadiene as the primary contaminants. Direct counting of stained cells revealed more than 3 × 107 cells mL−1 in the groundwater, with 58% intact and potentially viable. Universal and ‘Dehalococcoides’-specific 16S rRNA gene libraries were created and analyzed. Universal clones were grouped into 18 operational taxonomic units (OTUs), which were dominated by low-G+C Gram-positive bacteria (62%) and included several as yet uncultured or undescribed organisms. Several unique 16S rRNA gene sequences closely related to Dehalococcoides ethenogenes were detected. Anaerobically grown isolates (168 in total) were also sequenced. These were phylogenetically grouped into 18 OTUs, of which only three were represented in the clone library. Phylogenetic analysis of isolates and the clone sequences revealed close relationships with dechlorinators, fermenters, and hydrogen producers. Despite acidic conditions and saturation or near-saturation chlorosolvent concentrations, the data presented here demonstrate that large numbers of novel bacteria are present in groundwater within the DNAPL source zone, and the population appears to contain bacterial components necessary to carry out reductive dechlorination.
Biotransformation has been widely studied and applied for in situ remediation of chloroethanes and chloroethenes in cases where contaminants are present at relatively low concentrations in groundwater plumes (Lorah & Olsen, 1999; Hendrickson et al., 2002). Under anaerobic conditions, biotransformation of chloroethenes occurs through dehalorespiration, whereby the chlorinated ethenes serve as electron acceptors, resulting in successive reductive dechlorination from perchloroethene to trichloroethene, dichloroethene, vinyl chloride, and finally the nontoxic endproduct ethene. Chlorinated ethanes can also undergo successive reductive dechlorination reactions, although with more diverse pathways (Chen et al., 1996; Lorah & Olsen, 1999). Because chloroethenes and chloroethanes serve as terminal electron acceptors, an electron donor such as molecular hydrogen (H2) is required in dehalogenating microorganisms' energy metabolism (Cupples et al., 2003). In recent years, several bacteria capable of reductive dechlorination have been isolated in pure culture. These include Dehalobacter restrictus (Hollinger et al., 1998), Sulfurospirillum multivorans (formerly Desulfuromonas multivorans) and Sulfurospirillum halorespirans (Luijten et al., 2003), Desulfuromonas chloroethenica (Krumholz, 1997), and ‘Dehalococcoides’ sp. (Maymó-Gatell et al., 1999; He et al., 2003), the only group of microorganisms isolated in pure culture that can reductively dechlorinate vinyl chloride to ethene.
During in situ anaerobic bioremediation of chlorinated aliphatic compounds, it is hypothesized that syntrophic interactions such as interspecies hydrogen transfer among microbial community members play a critical role in contaminant biotransformation. For example, growth of ‘Dehalococcoides ethenogenes’ strain 195 apparently requires unknown growth factors contained in anaerobic sludge supernatant as well as molecular hydrogen (Maymó-Gatell et al., 1997). Although microbial community structure has been reported for several enrichment cultures capable of reductively dechlorinating chloroethenes (Duhamel et al., 2002; Richardson et al., 2002; Dennis et al., 2003; Rossetti et al., 2003; Gu et al., 2004), chloropropanes (Schlötelburg et al., 2002) or chlorobenzenes (von Wintzingerode et al., 1999), structure–function relationships remain poorly understood. With a few exceptions (e.g. Macbeth et al., 2004), microbial populations in reductively dechlorinating systems reported to date have been enriched over long time periods in laboratory systems with chlorinated solvent concentrations far less than saturation levels (i.e. in the absence of DNAPL) and in the presence of a readily available supply of electron donors (e.g. H2) (Duhamel et al., 2002; Richardson et al., 2002; Dennis et al., 2003; Rossetti et al., 2003; Gu et al., 2004). Because of these selective pressures, the microbial populations are not necessarily representative of in situ populations (Macbeth et al., 2004). Little research has been published on the characterization of in situ microbial populations at chlorosolvent-contaminated sites in general, and even less information is available regarding microbial populations from areas contaminated by chloroethanes and solvent mixtures (as opposed to only chloroethenes).
Because traditional methods (i.e. pump and treat) to remove DNAPL contamination in aquifers have historically been expensive, slow, and often ineffective, alternative approaches, such as monitored natural attenuation (MNA), are increasingly being investigated as remediation technologies (Pankow & Cherry, 1996). Microcosm studies of sediment and groundwater at the Petro Processors of Louisiana, Inc. Superfund Site indicate that microbially mediated reductive dechlorination is occurring in areas outside of the DNAPL source zone (i.e. in the contaminant plume) (Truex et al., 2001; Clement et al., 2002). The research described here was conducted using culture-dependent and culture-independent techniques to characterize the bacterial community within the DNAPL source zone in support of an effort to assess whether MNA is also feasible for the DNAPL source zone.
Materials and methods
Groundwater samples were collected from a waste recovery well (W-1024-1) located in the DNAPL source zone at the Brooklawn site, one of two areas collectively known as the Petro Processors of Louisiana, Inc. (PPI) Superfund Site, located approximately 10 miles north of Baton Rouge, LA. Operations at the Brooklawn site, opened in 1969 and operated until 1980, involved disposal of petrochemical waste, including free-phase chlorinated organics, by direct discharge to earthen ponds. Portions of the Brooklawn area were capped in the early 1990s, and starting in 1994, an array of recovery wells was installed to recover free-phase organic contaminants. Recovery of DNAPL in the source zone area is ongoing. Well W-1024-1 has a screened interval extending from 16.5 to 76.5 feet below ground surface in an area containing alternating layers of clay, silt, and sand. Additional details regarding contaminant hydrology for the site has been reported elsewhere (Clement et al., 2002). For microbial analyses, sterile 1.0 L glass bottles with Teflon-lined lids were filled with groundwater, leaving little or no headspace, and then placed on ice during transport to the laboratory (approximately 1 h).
Concentrations of volatile organic compounds were measured using US EPA method 624. Dissolved ethene, ethane and methane were measured using method RSK 175. Nitrate and nitrite were measured using US EPA method 353.2. Chloride was measured using US EPA method 325.2. Sulfate was measured by ion chromatography using US EPA method 300.0. Sulfide was measured using US EPA method 376.2. Ferrous iron was measured using US EPA method 3500-Fe D. Total organic carbon and total inorganic carbon were measured using US EPA method 5310B. Detailed descriptions of the US EPA analytical methods referenced above are available elsewhere (National Environmental Index, http://www.nemi.gov/). The pH of groundwater samples was measured using an Orion model 290A pH meter.
Total microbial numbers were determined by direct counting following staining with 4′,6-diamidino-2-phenylindole (DAPI). In this procedure, groundwater samples were first preserved by addition of 4% glutaraldehyde in 0.1 M cacodylate buffer (pH 7) to an equal volume of groundwater. Samples were preserved within 2 h after sample collection and were subsequently stored at −20°C until further processing. Samples were sonicated for 45 s (Branson Ultrasonic 170 V 50/60 Hz), stained for 1 h in the dark at a final DAPI concentration of 5 μg mL−1, and then collected on 0.20-μm-pore black Nuclepore polycarbonate filters. Filters were mounted on microscope slides and imaged using a Nikon Microphot-FXA epifluorescent microscope at 80 × magnification and a Nikon DM400 filter set (365 nm BP excitation, 400 nm LP dichroic, 400 nm LP barrier). Cells were counted in two separate groundwater samples each with 20 fields of view.
The fraction of live cells was estimated using a LIVE/DEAD BacLight Bacteria Viability Kit (Molecular Probes, Eugene, OR) according to the manufacturer's recommended protocol. In this protocol, a green-fluorescent nucleic acid stain, SYTO 9, is added to label all bacterial cells, and red-fluorescent propidium iodide to stain only those cells with compromised cell membranes. Viable cells fluoresce bright green, while dead or compromised cells fluoresce red. Samples were imaged using a Nikon Microphot-FXA epifluorescent microscope at 80 × magnification and a Nikon multiband DAPI-FITC-Rhodamine cube filter set (Boulos et al., 1999). Red and green cells were counted in 30 fields of view. The fraction of live cells was calculated as the average ratio of green cells to total (red+green) cells.
Groundwater was serially diluted, and 500 μL aliquots were spread on R2A plates. Plates incubated under aerobic conditions were buffered at pH 5.0 with 20 mM acetate buffer prior to solidification with agar (EMD, Gibbstown, NJ). Plates incubated under anaerobic conditions were buffered at pH 5.0 with 20 mM acetate buffer prepared in the same manner but were also supplemented with 1.0 mg L−1 resazurin (as a redox indicator) and 0.25 g L−1 cysteine hydrochloride (as a reducing agent) prior to solidification with agar. All plates were incubated at 30°C. Anaerobic plates were prepared and incubated in an anaerobic chamber (Coy, Grass Lake, MI) containing approximately 2% H2, 10% CO2, balance N2. To estimate the number of organisms represented by spores in the original sample, a pasteurization technique was employed in the anaerobic plate counting procedure (Rosencrantz et al., 1999). In this method, following serial dilution, groundwater samples were pasteurized by heating in a water bath at 80°C for 15 min prior to plating. The number of colonies on the plates was counted at 7, 14, 21 and 28 days.
For clone 16S rRNA gene library analysis, groundwater samples were transferred to sterile 15-mL polypropylene tubes, which were centrifuged at 4000 g for 10 min. Supernatant was decanted, leaving a sediment pellet that was resuspended in 0.5 mL of TE buffer (10 mM Tris, 1 mM EDTA, pH 8) and transferred to sterile 1.5-mL tubes, which were centrifuged at 16 000 g for 15 min. The supernatant was decanted, and the pellet (approximately 0.3 mL) was resuspended in fresh TE buffer (to a total volume of 2 mL) prior to freezing at −20°C until DNA extraction.
DNA was extracted using modifications of the method described by Herrick et al. (1993). In this process, 100 mg of polyvinylpolypyrilidone (Agros Organics, Geel, Belgium) was added to 2 mL of resuspended groundwater sediment and 10 mL of TE buffer, vortexed, and centrifuged at 11 500 g for 10 min. Supernatant was decanted, the pellet was resuspended in 5 mL of SET buffer (200 g L−1 sucrose, 0.05 M EDTA, 0.05 M Tris-HCl, pH 7.6), and 500 μL of lysozyme solution (c. 15 mg mL−1 lysozyme in water) was added. The sample was incubated at 37°C for 2 h with mixing by shaking at 5–15 min intervals. Then, 500 μL of 10% sodium dodecyl sulfate (filter sterilized with a 0.2-μm filter) was added, and three freeze–thaw cycles were performed (5 min at −80°C, then 5 min at 70°C). The sample was centrifuged at 11 500 g for 10 min. The supernatant was stored at 4°C. The pellet was resuspended in 5 mL of 0.12 M Na2HPO4 and 50 μL of proteinase K. The resuspended pellet was incubated in a 37°C water bath for 30 min, and then at 65°C for 1 h, with periodic mixing by shaking (every 5–15 min). The sample was centrifuged at 11 500 g for 30 min. The new supernatant was added to refrigerated supernatant from the previous step. The combined supernatant was centrifuged at 8000 g for 30 min.
Polyethylene glycol 8000 (PEG) (10 mL; 30% w/v in deionized water) was added and mixed by shaking. NaCl solution (1.5 mL, 5 M) was added and mixed well, and the sample was refrigerated overnight at 4°C. The sample was centrifuged at 22 500 g for 30 min. Supernatant was discarded and the pellet was resuspended in 4 mL of TE buffer and vortexed. Four 1 mL aliquots were each extracted with 500 μL of phenol, and this was followed by centrifugation for 10 min at 14 000 g and removal of the organic layer. An identical extraction procedure was carried out using chloroform in place of phenol. Clean-up was performed using a MoBio UltraClean PCR Clean-up Kit (Carlsbad, CA).
16S rRNA gene PCR, cloning, and sequencing
Two sets of oligonucleotide primers, one set consisting of 27f (5′-GAGTTTGATCCTGGCTCA-3′) and 1525r (5′-AGAAAGGAGGTGATCCAGCC-3′), universal to the 16S rRNA gene of all bacteria (Lane, 1991), and the other, DHC1f (5′-GATGAACGCTAGCGGCG-3′) and DHC 1377r (5′-GGTTGGCACATCGACTTCAA-3′), specific to variable regions of the 16S rRNA gene of ‘Dehalococcoides’ group bacteria (Hendrickson et al., 2002), were used in separate PCR reactions. For ‘Dehalococcoides’-specific primers, PCR reactions were performed under the conditions reported by Hendrickson et al. (2002). For universal bacterial primers, a hot start protocol was utilized, using minor modifications of the method described by Rainey et al. (1996). Each reaction (100 μL) contained 1 × Taq buffer with Mg2+, 2.5 U of Taq DNA Polymerase and 0.75 × TaqMaster PCR Enhancer (Brinkmann, Westbury, NY), as well as 200 μM each dNTP (Applied Biosystems, Forster City, CA) and 1 μL of purified community DNA. For the universal bacterial primers, 0.5 μg of each primer was added. PCR products were verified by gel electrophoresis prior to cloning. For the ‘Dehalococcoides’-specific primers, 0.3 μg of each primer was added.
16S rRNA gene products were cloned using a TOPO TA Cloning Kit for Sequencing (Invitrogen, Carlsbad, CA). PCR-amplified inserts were sequenced using a BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA). Sequencing reactions associated with inserts corresponding to the universal primers were performed as described by Rainey et al. (1996). Sequencing reactions associated with inserts corresponding to ‘Dehalococcoides’-specific primers employed the DHC1f primer for one reaction and the DHC774f primer (5′-GGGAGTATCGACCCTCTC-3′) (Hendrickson et al., 2002) in a second reaction. The temperature program for both reactions was as described by Hendrickson et al. (2002). Sequencing was performed using an ABI 377 Automated DNA Sequencer.
Genomic DNA was extracted using a MoBio UltraClean Microbial DNA kit (Carlsbad, CA) from 168 colonies isolated under anaerobic conditions on a variety of agar media, including: R2A (pH 5), Columbia anaerobic sheep blood agar (BD), plate count agar (Difco), nutrient agar (Difco), peptone–yeast extract–fructose (PYF) medium (Engelmann and Weiss, 1985) solidified with 15 g L−1 agar, and SRB medium (Widdel & Bak, 1992) supplemented with fatty acids (lactate, acetate and pyruvate, 10 mM each) modified by replacing bicarbonate buffer with phosphate buffer (30 mM) and sulfide with l-cysteine. PCR was performed using universal bacterial primers, and amplified 16S rRNA genes were then sequenced and analyzed as described above.
Rarefaction curves, diversity indices, and LIBSHUFF analyses
Estimates of microbial diversity within the clone library were further investigated using rarefaction analysis conducted using the analytical approximation algorithm aRarefactWin (Analytic Rarefaction, version 1.3, S. Holland, http://www.uga.edu/~strata/software/) with 95% confidence limits. Clone coverage and Shannon diversity index values were calculated as described by Good (1953) and Müller et al. (2002), respectively.
Comparisons between a partial section of the bacterial 16S rRNA gene corresponding to bases 134–596 in Escherichia coli numbering (c. 500 bp) were tested using the webLIBSHUFF version 0.96 program (Singleton et al., 2001) (http://LIBSHUFF.mib.uga.edu), which incorporates the coverage formula of Good (1953) to generate homologous and heterologous coverage curves. Sequence libraries between samples were shuffled 999 times. The measured distance between curves was calculated using the Cramér–von Mises test statistic (Pettitt, 1982). Distance matrices submitted to webLIBSHUFF were generated using the DNADIST program of Phylip, utilizing the Jukes–Cantor model for establishing nucleotide substitution rates.
Nucleotide sequence accession numbers
Representative sequences from the clone library constructed using universal bacterial primers and ‘Dehalococcoides’-specific primers have been deposited in GenBank under accession numbers DQ196601–DQ196618 and DQ196591–DQ196600, respectively. Representative sequences from cultured isolates have been deposited in Genbank under accession numbers DQ176646 and DQ196619–DQ196635.
Contaminant concentrations and geochemistry
The results of analysis of aqueous-phase organic contaminants in the groundwater are summarized in Table 1, and geochemical constituents are summarized in Table 2. As shown in the tables, groundwater from the DNAPL source zone contained high aqueous-phase concentrations of a variety of chlorinated compounds, including chloroethanes (average of 57.3 mg L−1 1,1,2,2-tetrachloroethane, 367 mg L−1 1,1,2-trichloroethane, and 540 mg L−1 1,2-dichloroethane), chloroethenes (52.2 mg L−1 vinyl chloride), chloropropanes (67.3 mg L−1 1,2-dichloropropane) and hexachloro-1,3-butadiene (6.58 mg L−1). Other chlorinated compounds may also have been been present; however, because of the high concentrations, detection levels were high (10–25 mg L−1 for most chlorinated volatile organic compounds) due to the large dilution factor necessary for analysis. Historical data from production wells indicate that additional chloroethenes, aromatics and polycyclic aromatic hydrocarbons are also constituents in the DNAPL at the site (Clement et al., 2002).
Table 1. Chlorosolvent concentrations measured in groundwater from well W-1024-1 at the PPI site
The presence of vinyl chloride (average concentration 52.2 mg L−1), ethene (9.57 mg L−1), ethane (0.302 mg L−1), and chloride (4510 mg L−1), all products of reductive dechlorination, suggests that dechlorination is occurring in the source zone. The fact that these concentrations are substantially higher than background levels previously reported for wells hydraulically upgradient of the DNAPL source zone (Clement et al., 2002) further supports this notion. Microcosm studies conducted using groundwater from within the DNAPL source zone indicate that the microbial population is able to reductively dechlorinate 1,1,2-trichloroethane, 1,2-dichloroethane and vinyl chloride to ethene (unpublished data).
Direct counting of DAPI-stained cells in two independent groundwater samples collected 17 days apart revealed 3.2 (±0.4) × 107 cells mL−1 and 3.7 (±0.7) × 107 cells mL−1. LIVE/DEAD BacLight microscopy counts revealed that 58% of the total cells were intact and potentially viable.
Plate count results are from 28 days of incubation, corresponding to the highest counts observed. No growth was observed on pH 5 R2A plates incubated under aerobic conditions. On pH 5 R2A plates incubated under anaerobic conditions, 1.3 (±0.2) × 104 CFU mL−1 were observed. Spores made up a minor fraction (2.5%) of the population enumerated on plates incubated under anaerobic conditions. These results confirm that viable bacteria were present within the groundwater.
Phylogeny of clone libraries and cultured isolates
Nearly complete (c. 1200 bp) 16S rRNA gene sequences of 223 clones created using PCR amplicons produced using universal bacterial primers were sequenced. Of these, 28 clones (12.6%) were identified as putative chimeric sequences and excluded from further analysis. Based on phylogenetic analysis of 16S rRNA gene sequences from the remaining 195 clones, the environmental 16S rRNA genes were tentatively grouped into 18 OTUs (Table 3). The 18 OTUs in the universal clone library spanned four phyla, with low-G+C Gram-positive bacteria (62%) and Actinobacteria (36%) having the largest representation, distantly followed by Proteobacteria (1.5%) and Chloroflexi (0.5%).
Table 3. Phylogenetic summary of bacterial community based on 16S rRNA gene sequences amplified using universal bacterial primers
Eighty-five clones derived from PCR products produced using primers specific to ‘Dehalococcoides’ 16S rRNA gene fragments were sequenced (c. 1300 bp). Of these, 23 clones had identical sequences, and 62 clone sequences were unique. Clones were closely related to ‘Dehalococcoides ethenogenes’ strain 195, with similarity ranging from 99.5% (7 bp difference) to 99.8% (3 bp difference). Thus, while ‘Dehalococcoides’ was represented by a single sequence in the universal clone library, there appears to be an appreciable amount of minor variation within the ‘Dehalococcoides’ population at the PPI site.
The 168 16S rRNA gene sequences determined for cultured isolates grouped into 18 OTUs (Table 4). The isolates were distributed among three phyla: Actinobacteria (73%), low-G+C Gram positive bacteria (20%), and Proteobacteria (7%). The phylogenetic positions of the universal clone OTUs and the isolate OTUs based on 16S rRNA gene sequences are shown in Fig. 1. Only three of the 18 isolate OTUs were represented in the clone library. The isolate OTU designated as BLI-A was 99.8% similar to the clone OTU designated as BLUC-C (where the prefix ‘BLI’ (Brooklawn Isolate) denotes isolate sequences and the prefix ‘BLUC’ (Brooklawn Universal Clone) denotes clone library sequences). These are most closely related to Actinomyces georgiae (X80413) and Actinomyces meyeri (X82451), respectively. Isolate OTU BLI-C was 98.3% similar to clone OTU BLUC-Q, which were both most closely related to ‘Desulfovibrio ferrireducens’ strain CY1 (AJ582755). Additionally, isolate OTU BLI-Q was 98.1% similar to clone OTU BLUC-P, both of which were most similar to Mogibacterium pumilum (AB021701) in terms of previously cultured isolates.
Table 4. Phylogenetic summary of bacterial community based on 16S rRNA gene sequences from anaerobic isolates
A comparison of phyla represented in the universal clone library from the PPI site to three previously published clone libraries of reductively dechlorinating microbial consortia is shown in Fig. 2. The first comparison library (Fig. 2b) is from a trichloroethene DNAPL-contaminated aquifer at Test Area North (TAN) at the US Department of Energy's Idaho National Engineering and Environmental Laboratory, where lactate addition has been used to stimulate in situ reductive dechlorination of trichloroethene to ethene (Macbeth et al., 2004). The second and third comparison libraries (Fig. 2c and d) are from functionally stable reductively dechlorinating enrichment cultures in which trichloroethene (Richardson et al., 2002) and dichloroethene (Gu et al., 2004) were biodegraded to ethene. At the PPI and TAN sites, both in situ environments, only a small percentage of the universal bacterial 16S rRNA clone libraries (1% at the PPI site and 3% at the TAN site) were closely related to known dechlorinators (‘Desulfuromonas michiganensis’ and ‘Dehalococcoides’ sp. strain VS at the PPI site and Sulfurospirillum multivorans and Trichlorobacter thiogenes at the TAN site). Additionally, the clone libraries at each site were dominated by low-G+C Gram-positive bacteria, with 62% and 65% representation at the PPI site and TAN site respectively. In contrast, clone libraries constructed from enrichment cultures reported by Richardson et al. (2002) and Gu et al. (2004) both contained a high proportion of clones (34–37%) closely related to ‘Dehalococcoides ethenogenes’, a known dechlorinator phylogenetically affiliated with a portion of the Chloroflexi phylum that has been informally proposed as class ‘Dehalococcoidetes’ (Hugenholtz & Stackebrandt, 2004). The enrichment culture clone libraries were composed of 33–36% low-G+C Gram-positive bacteria, a lower proportion than in the PPI (62%) and TAN (65%) sites. Another similarity between the bacterial community compositions at the PPI and the TAN sites was the low representation of Proteobacteria, present at levels of 2% and 3%, respectively. Other than the previously described similarities between the PPI site and the TAN site bacterial phyla, the phyla at each site were different: the PPI site contained 36%Actinobacteria, and the TAN site contained 14%Bacteriodetes, 13% OP.11, 4%Spirochaetes, and 1% OP.3.
To further compare the PPI and TAN clone libraries, sequence data and OTU designations from the universal bacterial clone libraries determined in both the PPI site and the TAN site were subjected to rarefaction analysis. Rarefaction curves for the libraries approached, but did not reach, a clear saturation, suggesting that analysis of additional clones would probably reveal further diversity (Fig. 3). Comparison of rarefaction curves suggests that the bacterial population at the PPI site is less diverse than that at the TAN site, even though there was a more extensive clone-sampling effort for the PPI site (195 clones) than for the TAN site (93 clones). The null hypothesis that there is no difference between the species richness at the two sites was rejected because the 95% confidence interval for each site did not overlap at high sample size. Clone coverage was calculated to be 91% at the PPI site and 76% at the TAN site, indicating that a larger portion of diversity was captured at the PPI site. In addition, there was less diversity at the PPI site as measured by the Shannon diversity index (3.59 with an evenness of 0.80 reported by Macbeth et al. (2004) for the TAN site vs. 2.19 with an evenness of 0.76 for the PPI site) and fewer universal clone OTUs (22 OTUs for the TAN site vs. 18 OTUs for the PPI site reported here).
LIBSHUFF comparisons of PPI and TAN data indicate that the communities were derived from significantly different bacterial populations (Table 5). The ΔC test statistic, a measure of the overall distance between the homologous (X) and heterologous (XY) curves, was highest when the PPI clone or isolate libraries were compared to the TAN site clone libraries (comparisons 1 and 2). A significant portion of this difference is due to low heterologous coverage estimates of D≤0.20 (data not shown), indicating that the clones generated from the PPI site are composed of many different higher-order taxonomic groups than the TAN aquifer site. LIBSHUFF comparisons of the PPI clone and isolate libraries were significantly different, although ΔC values were lower (comparison 3). Furthermore, poor heterologous coverage, D≤0.15, indicates that these libraries do share deep-branching taxa, but overall differences are driven by the dominance of clones related to the Acidaminococcaceae family (clones in OTUs BLUC-A, -E, -I, and -J) and the Clostridium isolates. It is likely that poor heterologous values in this comparison are the result of known biases in using culture-dependent and culture-independent techniques (Amann et al., 1995).
Table 5. LIBSHUFF comparisons of isolate and clone libraries
Homologous (X) coverage data
Heterologous (Y) coverage data
Collectively, the enumeration results indicate that a large number of microorganisms (>3 × 107 cells mL−1) are present in the groundwater, and 58% of the cells are apparently viable in spite of the high chlorosolvent concentrations and acidic pH observed at the sampling location. For comparison purposes, the total cell concentration observed in groundwater from the PPI site DNAPL source zone is roughly one-third of that observed in the laboratory-grown enrichment culture employed in a recent bioaugmentation effort (Lendvay et al., 2003). The fact that direct cell counts via microscopy were four orders of magnitude higher than plate counts demonstrates that a relatively low percentage (<0.1%) of the microbial community can be cultured using the techniques employed in this study. This is further supported by the fact that of the 18 OTUs in the universal clone library, only three were represented by cultured isolates, and it is consistent with previous reports that the majority of bacteria are not enumerated using standard plate count techniques (Amann et al., 1995).
Of the 18 OTUs represented in the universal clone library, only three (BLUC-M, -N, and -O), each consisting of just one clone, were greater than 97% similar to previously cultured microorganisms. Two of these, BLUC-M and BLUC-O, had high similarity to previously isolated dechlorinating microorganisms. The first, BLUC-M, was 99.8% similar to ‘Dehalococcoides’ sp. strain VS, which is able to reductively dechlorinate dichloroethene and vinyl chloride to ethene as part of its energy metabolism (Cupples et al., 2003). In a study of 24 chloroethene-dechlorinating sites throughout North America and Europe, at locations where ‘Dehalococcoides’ was detected, complete degradation of chloroethenes was observed. At locations where ‘Dehalococcoides’ was not detected, incomplete dechlorination was observed (Hendrickson et al., 2002). This suggests that the presence of ‘Dehalococcoides’ is a necessary prerequisite for complete degradation of chloroethenes (i.e. formation of ethene). The results described herein further expand the pH habitat and geographic range over which ‘Dehalococcoides’ sp. have been detected.
The second clone OTU most similar to a previously cultured dechlorinating microorganism, BLUC-O, was 97.2% similar to ‘Desulfuromonas michiganensis’, an acetotrophic anaerobe isolated from freshwater sediment that reductively dechlorinates 1,1,2,2-tetrachloroethane (Sung et al., 2003), a contaminant detected at high concentrations in the PPI groundwater measured in this study, to an endproduct of cis-dichloroethene. Laboratory studies have also demonstrated that ‘Desulfuromonas michiganensis’ can grow in the presence of free-phase perchloroethene (reductively dechlorinating perchloroethene to an endproduct of cis-dichloroethene), demonstrating that it is tolerant of high chlorosolvent concentrations (Sung et al., 2003).
Detection of sequences closely related to ‘Dehalococcoides’ and ‘Desulfuromonas michiganensis’ at the PPI site where DNAPL is present supports the notion that contaminant degradation is probably occurring in groundwater near the DNAPL source, a phenomenon previously demonstrated in laboratory studies (Nielsen & Keasling, 1999; Carr et al., 2000; Yang & McCarty, 2000, 2002; Cope & Hughes, 2001; Dennis et al., 2003). This is particularly true in the case of ‘Dehalococcoides‘ sp., because chlorosolvents are the only class of compounds known to serve as their terminal electron acceptors. Provided that this can be further substantiated, it may broaden the possibility for implementation of monitored natural attenuation as a scientifically defensible remediation strategy for the DNAPL source zone areas at this and other sites.
The third OTU with greater than 97% similarity to a previously cultured bacterium, BLUC-N, grouped within the β-Proteobacteria subdivision with 99.6% similarity to ‘Brachymonas petroleovorans’ strain CHX, an aerobic degrader of light hydrocarbons, including cyclohexane and toluene, which was isolated from oil refinery wastewater sludge (Rouvière & Chen, 2003). All of the universal bacterial clones except for this one were most closely related to facultative or obligate anaerobes.
The remaining 15 OTUs, covering 192 of the 195 universal bacterial clone library sequences (98% of the total), were distantly related to previously cultured bacteria (less than 97% similar). Thus, it appears that the composition of the microbial population in the DNAPL source zone at the PPI site is relatively novel and probably includes several new species and even new genera. Consequently, it is impossible to fully elucidate what function bacteria represented by these OTUs may serve within the microbial population and what contribution, if any, they may play in biotransforming contaminants found at the PPI site. It is interesting to note, however, that of these 15 OTUs less than 97% similar to previously cultured organisms, four (BLUC-B, -K, -L, and -G), comprising a total of 44 clones (23% of the library), phylogenetically grouped somewhat closely with uncultured or uncharacterized bacteria from dechlorinating populations (Fig. 1).
Of the cultured isolates, one OTU (BLI-M) was 99.6% similar to Clostridium bifermentans strain ATCC 638 (X75906), and 99.5% similar to Clostridium bifermentans strain DPH-1 (Y18787), a known dechlorinator. Strain DPH-1 is able to dechlorinate a number of chloroethanes, including 1,1,2-trichloroethane, dichloropropane and dichloromethane, along with several chloroethenes, including perchloroethene, at high concentrations (0.9 mM) (Chang et al., 2000). Like the OTUs from the clone library, most of the isolate OTUs are most closely related to genera with fermentative capabilities. All but two isolate OTUs (BLI-Q and BLI-R), or 166 of the 168 isolates (99% of the total), were most closely related to fermentative organisms, with eight OTUs (BLI-C, -F, -H, -I, -K, -M, -N and -O), covering 25 of the 168 isolates (15% of the total), representing organisms most similar to previously described bacteria that produce hydrogen during fermentative processes.
Representatives from the isolate OTU designated as BLI-B were recently characterized and described as Propionicicella superfundia gen. nov., sp. nov. (Bae et al., 2006). This new bacterial genus grows at pH levels as low as 4.5, tolerates high chlorosolvent concentrations, and produces propionate and acetate as fermentation products both in the presence and absence of chlorosolvents. The closest cultured phylogenetic relatives of several additional OTUs in both isolate and universal clone libraries are also known to be acid tolerant. For example, members of the Megasphaera (Haikara & Helander, 2002) and the genera Actinomyces (Takahashi & Yamada, 1999) and Clostridium (Kuhner et al., 2000; Flythe & Russell, 2005) have been reported to grow at the pH level observed in groundwater at the PPI site.
As statistically shown by LIBSHUFF comparisons, the PPI site clone and isolate libraries were significantly different. Owing to limitations imposed by culturing conditions (media, pH, carbon source, etc.) that do not closely replicate in situ environments, isolates that grow on commonly used culture media such as those employed in this study are not likely to represent dominant species from in situ populations (Amann et al., 1995).
Drawing definitive conclusions regarding the novelty of the microbial population at the PPI site in comparison to populations at other sites is difficult because few clone libraries have been constructed for similar sites (i.e. those with chloroethanes as dominant contaminants, acidic pH, or presence of DNAPL). However, current data suggest that the bacterial phylotypes collected at the PPI site contrast with those from the few comparable environments that have been studied. The PPI site and TAN site populations were statistically different based on LIBSHUFF analysis, partly due to the abundance of Actinobacteria (36%) in the PPI clone library, compared to no Actinobacteria in the TAN clone library. Based on the high representation of Actinobacteria in both the PPI clone and isolate libraries, it is assumed that they play a large role in the in situ bacterial community function; however, few or no Actinobacteria representatives have been reported for clone libraries from perchloroethene-, trichloroethene-, vinyl chloride- and dichloropropane-dechlorinating enrichment cultures (Richardson et al., 2002; Schlötelburg et al., 2002; Dennis et al, 2003; Rossetti et al., 2003; Gu et al., 2004).
Potential reasons why the PPI bacterial community may be novel and less diverse than the TAN bacterial community include the unique mixture of chlorinated solvents, low-pH conditions and high sulfate concentrations at the PPI site. In a study of soil bacterial communities at numerous locations worldwide, Fierer & Jackson (2006) found that environmental pH was the best predictor of bacterial diversity and richness, with peak diversity and richness values at a pH of approximately 7, and decreased values at higher and lower pH. Community differences, however, may also result from differences in geographic location, differences in geochemistry, or the fact that there was addition of an exogenous electron donor (lactate) at the TAN site. Addition of electron donors has previously been shown to affect microbial community structure (North et al., 2004). This latter effect may be the reason for the higher portion of dechlorinators observed in some enrichment cultures (Richardson et al., 2002; Gu et al., 2004) compared to in situ systems (Macbeth et al., 2004; this study). It should be noted, however, that not all enrichment cultures follow this trend, and many clone libraries contain sequences from as yet uncultured bacteria with unknown functions (Dennis et al., 2003; Rossetti et al., 2003). In the study described here, 98% of the sequences belong to taxonomic groups that are distinct at the species level or higher (less than 97% similar) and have not been described previously. It is anticipated that as the body of knowledge increases, the data presented herein will contribute to a better understanding of bacterial structure–function relationships in situ at this and other sites.
The significance of the fact that the clone library constructed using ‘Dehalococcoides’-specific PCR primers revealed many closely related but unique 16S rRNA gene sequences is unclear at this point. There is some evidence that multiple strains closely related to ‘Dehalococcoides ethenogenes’ may be responsible for the degradation of chlorinated compounds in some locations. For example, Hendrickson et al. (2002) reported that the population in a microcosm constructed from sludge collected from the bottom of a pond at an industrial site in Beaumont, TX contained three unique ‘Dehalococcoides’ 16S rRNA gene sequences (AF388531, AF388532, and AF388533). They also reported that groundwater from an industrial site in Niagara Falls, NY contained two unique 16S rRNA gene sequences (AF388544 and AF388545). Duhamel et al. (2002) reported that an enrichment culture, referred to as KB-1, derived from soil and groundwater from a contaminated site in southern Ontario, Canada, contained five distinct sequences closely related to ‘Dehalococcoides ethenogenes’. Richardson et al. (2002) reported 10 unique 16S rRNA gene sequences closely related to ‘Dehalococcoides’ in an anaerobic enrichment culture derived from contaminated soil at the Alameda Naval Air Station, CA.
Members of the ‘Dehalococcoides’ group can have a high degree of similarity in the 16S rRNA gene but use different chlorinated compounds as terminal electron acceptors. For example, three isolated ‘Dehalococcoides’ strains, strain VS, strain 195, and strain CBDB1, are phylogenetically similar on the basis of 16S rRNA gene sequences (>98.8% similar), but each meets its energy needs using different chlorinated compounds as terminal electron acceptors (Maymó-Gatell et al., 1999; Adrian et al., 2000; Cupples et al., 2003; Hölscher et al., 2004). Because of the relatively small number of ‘Dehalococcoides’ strains isolated and characterized to date, genetic variability within the ‘Dehalococcoides’ 16S rRNA clone library from the PPI site cannot yet be correlated with specific functional roles in contaminant transformation. Future study of specific functional genes may provide an indication of metabolic diversity.
Despite the acidity and the saturation or near-saturation chlorosolvent concentrations observed at the PPI sampling location, the data presented here suggest that a number of bacterial types, many of them novel, can grow in this environment. Among these OTUs were microorganisms closely related to known dechlorinators, fermenters, and hydrogen producers. These field data, supported by microcosm data and chemical analysis that revealed degradation products, suggest that dechlorination is probably occurring in spatial locations in close proximity to DNAPL.
The authors gratefully acknowledge the Hazardous Substance Research Center South and Southwest (through the HSRC Environmental Biotechnology Initiative) and NPC Services for financial support. The authors thank Ms Cindy Henk of the LSU Socolofsky Microscopy Center for assistance with microscopy.