Nanoliter qPCR Platform for Highly Parallel, Quantitative Assessment ...

Article pubs.acs.org/est

Nanoliter qPCR Platform for Highly Parallel, Quantitative Assessment of Reductive Dehalogenase Genes and Populations of Dehalogenating Microorganisms in Complex Environments Koshlan Mayer-Blackwell,† Mohammad F. Azizian,‡ Christina Machak,§ Elena Vitale,∥ Giovanna Carpani,∥ Francesca de Ferra,∥ Lewis Semprini,‡ and Alfred M. Spormann*,†,⊥ †

Civil and Environmental Engineering, §Geological and Environmental Sciences, and ⊥Chemical Engineering, Stanford University, Stanford, California 94305, United States ‡ Chemical, Biological and Environmental Engineering, Oregon State University, Corvallis, Oregon 97331, United States ∥ Environmental Technologies - Istituto Eni Donegani - ENI, 20097 San Donato Milanese, Italy S Supporting Information *

ABSTRACT: Idiosyncratic combinations of reductive dehalogenase (rdh) genes are a distinguishing genomic feature of closely related organohalogen-respiring bacteria. This feature can be used to deconvolute the population structure of organohalogen-respiring bacteria in complex environments and to identify relevant subpopulations, which is important for tracking interspecies dynamics needed for successful site remediation. Here we report the development of a nanoliter qPCR platform to identify organohalogen-respiring bacteria and populations by quantifying major orthologous reductive dehalogenase gene groups. The qPCR assays can be operated in parallel within a 5184-well nanoliter qPCR (nL-qPCR) chip at a single annealing temperature and buffer condition. We developed a robust bioinformatics approach to select from thousands of computationally proposed primer pairs those that are specific to individual rdh gene groups and compatible with a single amplification condition. We validated hundreds of the most selective qPCR assays and examined their performance in a trichloroethene-degrading bioreactor, revealing population structures as well as their unexpected shifts in abundance and community dynamics.

1. INTRODUCTION Bioremediation of groundwater aquifers and sediments contaminated with chlorinated aliphatic hydrocarbons (CAHs) depends on the activities of reductive dehalogenases that are present in some anaerobic microorganisms.1,2 Of particular importance are obligate organohalogen-respiring bacteria, such as Dehalococcoides or Dehalogenimonas sp., because reductive dehalogenation is the only known mode of metabolic energy conservation in these microorganisms, and each microorganism can carry up to 36 different nonredundant rdh genes.3−5 While organohalogen-respiring bacteria have been key for decontaminating polluted sites via biostimulation and bioaugmentation, there are many instances where such treatments have been hindered by the absence of key microorganisms and/ or genes, enzymatic inhibition,6−9 hydrological complications,10 or insufficient management of microbial ecology and associated biogeochemistry.11,12 Remediation of common groundwater contaminants such as tetrachloroethene (PCE), trichloroethene (TCE), 1,1,2-trichloroethane (1,1,2-TCA), and 1,2-dichloroethane (1,2-DCA) poses additional challenges since an appropriate assemblage of organohalogen-respiring bacteria plus their supporting microbial communities is required for © 2014 American Chemical Society

complete dechlorination of these compounds to a harmless end product. Furthermore, it is unclear whether faithful representatives of the well-studied laboratory isolates are dominant organohalogen-respiring bacteria in sediments and groundwater and to what extent their laboratory-studied phenotypes are relevant in the field. Given this uncertainty, managing bioremediation of CAHs requires (i) gauging the structure of the microbial community, in particular the organohalogenrespiring bacteria, and (ii) being able to identify and differentiate between closely related but functionally distinct subpopulations. Such information is crucial for predicting and controlling the ecological responses of the microbial communities to natural or engineered perturbations during bioremediation. Metagenomics,13 transcriptomics,14 proteomics,15 pan-genomemicroarrays,16,17 and functional-gene tiling microarrays18,19 have been used to study the eco-physiology of organohalogen-respiring bacteria. However, these approaches have not been widely applied Received: Revised: Accepted: Published: 9659

February 26, 2014 June 5, 2014 June 27, 2014 July 21, 2014 dx.doi.org/10.1021/es500918w | Environ. Sci. Technol. 2014, 48, 9659−9667

Environmental Science & Technology

Article

as tools in full-scale field studies due to the requirement of large amounts of DNA as input, bioinformatic complexity, cost constraints, and inadequate sensitivity of the assays for detecting low-abundance genes in complex genomic backgrounds. A number of single quantitative PCR (qPCR) assays targeting a few of the best-understood rdh genes have been shown capable of overcoming these obstacles. Single-gene qPCR is widespread in research laboratories and in the remediation industry, but there is significant need and interest in using simple but comprehensive PCR-based methods to accurately monitor rdh functional genes.20 The number of uncharacterized rdh genes continues to expand rapidly.21 As of January 2014, there are more than 690 nonredundant Rdh large subunit protein sequences in the NCBI database. To this end, a recently developed suite of 43 degenerate primers22 can be used to amplify a broader collection of rdh genes than was possible with previously used degenerate primer sets.23 In that method, short-read high throughput DNA sequencing is applied to PCRamplified DNA, pooled from all reactions, allowing amplified rdh gene sequences to be assembled by bioinformatics methods. While this degenerate primer suite, effectively amplifying a broad array of reductive dehalogenase genes, will likely be a valuable tool for gene discovery, these assays are not directly suitable for quantitative PCR detection of target genes because they are prone to false positive amplifications, produce large amplicons, and have a wide-range of predicted melting temperatures due to their high degree of degeneracy.22 Moreover, the library preparation, sequencing, and downstream bioinformatics expertise needed to use this method may limit its rapid use in engineering field studies. Given such and other constraints of existing molecular tools, we explored a microfluidics-based, massively parallel qPCR approach for covering much of the known rdh gene sequence space. The novel aspect of this approach is the design and validation of nearly 200 unique nondegenerate qPCR primer sets, which can be run in parallel in nanoliter volumes at a single stringent annealing temperature and buffer chemistry. Nanoliter qPCR (nL-qPCR) has been applied for detecting biomarker genes associated with bacterial pathogenesis,24 but so far, to our knowledge, this approach has not yet been applied to the field of bioremediation nor has it been used to quantitatively detect and distinguish between members of a diverse bacterial protein family. Here, we report the development and assessment of this parallel nL-qPCR platform as a tool for the quantitative analysis of rdh gene repertoires and identification of closely related populations of organohalogen-respiring bacteria. We show in a pilot study of a lab-scale bioreactor that this platform translates well to engineering applications. Quantitative data are obtained economically and rapidly from very modest amounts of DNA input material.

is limited to sequences of genes from known microorganisms. RD-OG sequence similarity does not guarantee shared substrate specificity, and members of distinct orthologue groups can have biochemical activity for a common substrate.21 A suite of novel qPCR primer pairs were designed to detect and distinguish between full-length and near-full-length reductive dehalogenase (rdh) gene groups. We used the Dehalogenase protein family (Pfam v 26.0) PF13486 as a database of nonredundant Rdh protein sequences.26 The Pfam database included sequences obtained from both microbial isolates and environmental samples, which we incorporated into the RD-OG framework. We limited consideration to sequences 350 to 700 amino acids in length. Corresponding rdh nucleotide sequences were downloaded from NCBI. Pfam amino acid sequences were clustered based on percent pairwise identity (PID) using blastp.27 Assays were designed for 54 reference sequences (Table S1, Supporting Information), most with at least one known high-PID homologue (>90% amino acid level). Figure S1, Supporting Information, shows the Pfam PF13468 in the context of a sequence similarity network.28 A Python script directed the software primer329 to generate thousands of candidate primer pairs, which we ranked based on oligonucleotide complementarity to high-PID sequences with the primary reference sequence (see bioinformatics programs and parameters in Figures S2 and S3). Where possible, we produced two assay types for each reference sequence. The first was to be “specific” to a reference sequence and those homologues with high PID. The second was meant to “extend” the number of sequences matched by the primers to include the reference sequence as well as many homologue sequences as possible. Supporting Information File 1 describes primers used in tabular form. We further designed assays for those rdh genes in Dehalococcoides mccartyi sp. that were not yet assigned to an orthologue group but had been putatively identified in our experimental systems by previous tiling-microarrays.19 Our computational pipeline also enabled the design of primers that differentiate among three closely related nucleotide sequence types of an important uptake hydrogenase (hupL) in Dehalococcoides mccartyi sp. Design of hupL primer sequences is discussed in the Supporting Information (Tables S2, S3, and S4). We also evaluated a number of commonly used primers targeting the 16S rRNA gene found in Dehalococcoides, Dehaligenimonas, Dehalobacter, Desulfitobacturium, and Geobacter for compatibility with our designed rdh primer set. To those assays developed uniquely for this study, we used four 16S rRNA marker gene assays developed in previous studies30−32 that tested compatible in our nL-qPCR conditions (Supporting Table S5). 2.2. Initial Screening of Candidate Assays. Primer performance data were collected using Wafergen Biosystems’ SmartChip MyDesign platform. Chips were prepared by robotically dispensing oligonucleotide primers (Integrated DNA Technology) at a final concentration of 1 μM into 100 nL wells. Assays were tested using both (20 Sample × 248 Assay) and (12 Sample × 384 Assay) formats. For each sample, data were collected from two separate chip runs using a standard Wafergen protocol: 95C for 3 min, then 40 cycles of (95C for 60 s, 60C for 70 s). Candidate assays were physically tested against a collection of 500bp synthesized linear DNA standards (Integrated DNA Technology) diluted to concentrations of 20,000, 2000, 200, and 20 copies per 100 nL reaction well. We also tested most assays at 5 copies per reaction well. The rdh candidate assays that failed to amplify standards at 20 copies per well or reproducibly amplified the negative control before cycle 28 were excluded from further

2. MATERIALS AND METHODS 2.1. Development of a rdh qPCR Assay Suite. Reductive dehalogenases (rdh) enzymes contain two 4Fe-4S clusters and one corrinoid cofactor per catalytic subunit.25 Rdhs are identified by the presence of amino acid sequence motifs for binding these cofactors as well as by pairwise amino acid sequence identity to biochemically characterized Rdh enzymes. A sequence-identitybased naming system for Rdhs was proposed wherein the protein family is divided into orthologue groups.21 ‘Reductive Dehalogenase Orthologue Groups’ (RD-OGs) are sets of two or more distinct Rdh sequences where all members share 90% amino acid identity with another member. RD-OG membership 9660

dx.doi.org/10.1021/es500918w | Environ. Sci. Technol. 2014, 48, 9659−9667


■

consideration. Candidate assays with PCR efficiencies less than 85% were also excluded. 168 assays passed this initial QC phase. The results from these serial dilution experiments at the three highest dilutions were used to construct assay-specific standard curves. A set of three Mus musculus genes were spiked into the master mix of both calibration and experimental chips to test for PCR inhibitors and ensure roughly similar amplification performance. Negative controls consisted of a complex genomic mixture absent of rdh genes. The mixture was constructed from DNA isolated from the following archaea and bacteria: Methanococcus maripaludis 109, Methanothermococcus thermolithotrophicus DSM 2095, Sporomusa ovata DSM 2662, Shewanella oneidensis MR-1, Geobacter metallireducens GS-15, Clostridium sporogenes, Sinorhizobium meliloti, and Bacteroides thetaiotaomicron. 2.3. Assay Specificity Tests. To validate the selectivity of newly designed assays to distinguish among rdh groups, total DNA was isolated by MoBio PowerSoil kit or by methods described by Behrens et al.33 from cultures highly enriched for Dehalococcoides mccartyi strains VS, GT, CBDB1, and 195. Samples were prepared at various bulk concentrations varying from 10 to 0.01 ng/μL. These were further diluted in Roche LightCycler 480 SYBR Green I Master Mix to final concentrations of 25 to 0.1 pg per well (Qubit dsDNA BR fluorometric quantification, Life Technologies). Additionally, a separate sample was amended with the above-mentioned genomic negative control mixture (50 pg) such that the Dehalococcoides DNA represented a minority fraction of total complex DNA mixture in each reaction. We next sought to validate our approach by examining for consistent amplification of high percent identity (PID) homologues across Dehalococcoides isolates, while also determining the frequency of false positives due to off-target amplifications. We predicted that primer sets containing three or more cumulative mismatches with a target gene would not amplify efficiently. If it did, it was classified as a false positive. By comparing this expectation with the amplification result, we designated each assay/isolate combination as true-positive, truenegative, false-positive, or false-negative if confirmed by duplicate chip results. We intended that the final rdh PCR suite include multiple assays for each reference group. On average, each group was represented by three unique assays. Individual assay results were aggregated in accordance with the recently developed Reductive Dehalogenase Orthologue Group naming system (see Methods 2.1).21 Each group presence/absence classification (true-positive, false-positive, false-negative, or true-negative) was determined by the majority result of all assays targeting that group. In the event that half or more of the assays resulted in no detection the group was considered absent. 2.4. Sampling of the Continuous Bioreactor. Operation and sampling procedure for the Evanite two-liter (EV2L) TCE-degrading continuously fed reactor has been previously described.34 Briefly, cells are grown at a mean-cell residence time of 50 days and continuously supplied with formate as electron donor. Reactor liquid (50 mL) were removed and centrifuged (8000 rcf) at 4 °C for 30 min, and solids were transferred to a MoBio PowerSoil bead-beating tube followed by isolation with Phenol Chloroform Isoamyl Alcohol saturated with Tris-HCl pH 8.05 (Ambion). The resulting DNA per sample was diluted to 10 ng/μL (Qubit dsDNA BR fluorometric quantification, Life Technologies), corresponding to 25 pg DNA per 100 nL reaction well. Chloroethene and hydrogen concentrations within the reactor were quantified by gas chromatography as described previously.35

Article

RESULTS AND DISCUSSION

The central challenge in designing primers for use in a highly parallel nL-qPCR platform is that all primers must operate at a stringent annealing temperature and buffer composition. That is, optimization of reaction conditions to improve performance of a single assay is not possible without affecting the performance of the entire set. Therefore, we developed a computational pipeline for batch primer development where all candidate assays were predicted to satisfy the same set of thermodynamic criteria. We then searched for a set of unique, nondegenerate primer sets that could capture all sequences in each rdh-group of interest. We physically tested 490 rdh and hupL candidate assays and selected from there 35% that performed satisfactorily. We found that some previously published microliter scale qPCR assays with compatible reported Tm suffered from poor efficiency or inadequate selectivity in the nL-qPCR platform under the standard thermal and buffer conditions tested (Supplemental Table S5). This emphasizes that nL-qPCR assay design may require systematic assay design procedures, validation, and calibration in a given buffer and reaction volume. With gains in throughput from the miniaturization and parallelization possible in nL-qPCR comes a trade-off in terms of the flexibility of this approach. That is, assays suites must be tested in batch, and experimenters may not be able to simply “mix-and-match” from previous useful assays optimized as single-reactions. In the following section we describe the results of our batch assay selection and validation. 3.1. Assay Specificity. Given high sequence similarity among rdh homologues,21 we tested whether our candidate assays were specific to their intended target. This specificity was predicted via a bioinformatics search for conserved nucleic acid signatures that were distinguishing features among groups of highly related rdh genes. Experimentally, we tested assay specificity in two ways. First, assay specificity was examined by attempting to amplify dilute linear rdh gene standards in the presence of a concentrated mixture of total genomic DNA isolated from eight nontarget anaerobic archaea and bacteria (see Methods 2.2). Genes for rdh are absent in these eight anaerobic microorganisms, but these microbes contain iron−sulfur-cluster and corrinoid-containing enzymes that share motifs similar to regions conserved in Rdh proteins. Second, we tested assays against four distinct strains of Dehalococcoides mccartyi for which we had a priori knowledge of their rdh and hupL gene composition. We used DNA from Dehalococcoides strains isolated from contaminated and wastewater treatment sites: VS (Victoria, Texas, USA), 195 (Ithaca, NY), GT (Cottage Grove, WI, USA), and CBDB1 (Jena, Germany). The isolates3,4,36 represent all three known Dehalococcoides subgroups as defined by 16S rRNA gene differences: Cornell, Victoria, and Pinellas.32 We observed amplification across target DNA concentrations ranging from 25 pg to 0.1 pg per 100 nL reaction. The assays were sensitive at the lowest target-DNA inputs tested (Figure S5). In the presence of more concentrated nontarget genomic DNA (50 pg per well) from 8 nonorganohalogen-respiring anaerobic bacteria, selectivity and sensitivity remained (Figures 1A, 1B, and S5). For PCR assays to be included in the final nL-qPCR platform, each assay had to (i) amplify its target with a PCR efficiency greater than 85% (most were >90%), (ii) not amplify negative control DNA prior to thermal cycle 28, and (iii) not exhibit self-dimerization as evidenced by a melt curve analysis. Of 490 candidate rdh and hupL assays tested, 168 were fully tested and found to meet all of these 9661



Article

Figure 1. (A) Assay calibration results across two chips with DNA standards applied in a 10-fold dilution series shows the sensitivity of the assays. Proximity to the 45-degree line reflects replicability across duplicate chips. Bottom right: the addition of genomic bait from 8 nonorganohalogen respiring bacteria at 10 to 100 times the copy ratio of the target did not cause a loss of sensitivity. (B) RD-OG and rdh abundance estimates from highly enriched Dehalococcoides mccartyi CBDB1 and GT cultures. The red dashed lines indicate the median rdh group abundance estimate, and the dotted lines demarcate the interquartile range of the observed values. The amount of DNA input per 100 nL reaction volume is shown above each sample name. The designation (+gDNA) indicates those samples run in the presence of a more concentrated genomic DNA bait mixture from eight anaerobic archaeal and bacterial isolates lacking reductive dehalogenase genes. (See methods for description of this mixture.)

group, and 3 hydrogenase gene types found in Dehalococcoides mccartyi. We anticipate the completeness of this primer suite can be improved by incorporating newly discovered rdh gene variants. 3.2. Sensitivity. The sensitivities of the qPCR assays were tested against linear DNA standards in 10-fold dilution series. Figure 1A shows technical replicates across two separate chips at four dilutions, spanning approximately 20,000, 2,000, 200, and 20 linear gene copies per 100 nL-reaction volume. Proximity to the 45-degree line reflects replicability across chips. The dynamic range of most nL-qPCR assays spans over 6 orders of magnitude, as has been shown in another study of the technology.37 We focused here on the sensitivity of our assays against low starting gene copy numbers likely to be found in mixed microbial populations. When our assays were calibrated against DNA standards, the amplification Ct values were technically reproducible across duplicate chips at 20,000, 2000, and 200 starting copies per reaction. At higher dilution, the technical variability increased, and at 20 copies per reaction the mean absolute Ct difference between cross-chip replicate samples was 0.73, compared to 0.33, 0.16, 0.13 at the respective higher concentrations (Figure 1A). The statistically unbiased nature of the errors, as well as results from counting simulations, indicates that an increased Ct difference at low copy numbers is to be expected in nL-qPCR (Figures S9, S10, Supporting Information). This finding is in line with previous studies and reinforces the point that detection of single-digit starting copies common in small reaction volumes requires attention to small-number statistics.37,38 3.3. Accuracy. In traditional practice, a single qPCR assay is often used to estimate the abundance of a target gene. Despite strong technical reproducibility, such estimates are not necessarily accurate in environmental samples with unknown gene content, where an unanticipated mismatch between target and primer sequences can cause systematic and reproducible shifts in measured Ct. With the large number of parallel reaction

criteria. Using these assays, group-level absence/presence classification was >90% accurate against the four Dehalococcoides isolates tested (Figure S4, Supporting Information). False-positive amplification can arise from permissive primer binding conditions. We noted instances in which an individual assay produced a false-positive result, usually manifested as a delayed amplification for an orthologue group not expected in a given Dehalococcoides isolate (Figure S4). The resulting gene counts were 2 to 3 orders of magnitude lower than the gene counts for the true-positives, suggesting that partial complementarity between primer and nontarget sequences carries the potential of producing a delayed Ct. A delayed amplification could not be easily distinguished as a false positive by the slope of the amplification curve alone; however, the partial redundancy we designed in the form of multiple assays targeting different nucleotide positions on each target reference rdh sequence allowed us to improve detection accuracy. Across the isolates tested, 15 of the 168 assays produced delayed Ct false positive results. In 66% of these cases, the other assays for the same target sequence yielded a correct negative result. Figure S4 reveals these individual false-positive events, but we used the majority result from multiple assays to improve the classification of true negatives vs false positives at the target level. We report absence/ presence classification accuracy on the target rather than the individual assay level. It should be emphasized, however, that the beneficial effect of this designed assay redundancy might be diminished when this tool is used in poorly characterized environments, such as complex natural sediments, where not every assay is expected to perfectly match its target. For unknown environments, single positive microliter scale PCR assay results are useful but may require further analysis. For some orthologue groups, none of the candidate assays passed all the above-mentioned quality control requirements and, thus, were not included. The final assay suite presented in this study encompasses 30 orthologue groups, 12 Dehalococcoides reductive dehalogenase types not-yet assigned an orthologue 9662



Article

formate concentration was reduced from 45 mM to 25 mM. The high degree of sequence similarity at the 16S rRNA gene level among Dehalococcoides mccartyi strains has complicated the tracking of distinct Dehalococcoides subpopulations via conventional qPCR or 16S short-amplicon sequencing. However, the relative stoichiometry of different subtypes of Dehalococcoides is important for modeling the degradation kinetics and partitioning of TCE, cDCE, and VC electron acceptors among closely related bacterial strains. Eight DNA samples archived from the reactor’s five-year operation were brought to a standard concentration of 10 ng/μL. Twenty ng of DNA was applied to duplicate nL-qPCR chips, resulting in a final concentration of 25 pg of total community DNA per reaction well. Thus, normalized population data should be interpreted in the context of declining overall biomass observed after the concentration in the influent formate decreased (Figure S6, Supporting Information). Using a correlation-based clustering method similar to that described by Marshall et al.,19 we inferred the population structure of organohalogen-respiring bacteria in the reactor. Briefly, gene abundance profiles were hierarchically clustered by their timeseries correlation (Figure 2A). We grouped correlated genes into clusters representing hypothesized operational strains if they were at similar absolute abundance. 4.2. Operationally Identified Strains. The clustering pattern of rdh and hupL gene counts suggests the presence of at least four distinct subpopulations of organohalogen-respiring bacteria within the EV2L reactor. We refer to these operational strains as Ev1, Ev2, Ev3, and Ev4. Throughout the reactor’s operation, Ev1 appears to be most numerous (Figure 2B). Multiple lines of evidence suggest that Ev1 is a Dehalococcoideslike-bacterium. In Ev1 at least four orthologue groups (5, 10, 13, and 23) are linked, which have so far only been found in Dehalococcoides isolates or environmental samples. Orthologue group 5 is the best understood, since it contains the characterized trichloroethene reductase tceA first discovered in Dehalococcoides mccartyi strain 195.40 The rdh gene clustering supports the assignment of a detected vinyl chloride reductase vcrA gene to the numerically less abundant Ev2−a second Dehaloccoides-like operational strain. In addition to vcrA, Ev2 is predicted to contain at least 9 rdh genes from other orthologue groups (10,11,13,15,17,20,21,30,32). In contrast to Ev1, Ev2 declined precipitously in the final 500 days of the time course. At day 168−prior to the introduction of formate-limiting conditions−the mean estimate of rdh gene count in Ev2 was 1500 ± 600 copies per 25 pg of total community DNA. By the end of the experiment, the mean estimate of Ev2 was 90 ± 60 copies. Moreover, the decline in Ev2 correlates with changes in reactor’s chemical performance, where the percentage of TCE converted fully to ethene dropped from 90% to 30%, consistent with the observed loss of vcrA (Figure 2D). At day 600, Ev2 was estimated to constitute more than 10% of the total Dehalococcoides population, but, by the end, it constituted less than 1%. Because the dominant tceA-containing strain Ev1 remained in far greater abundance, the dramatic decline in the vcrA-containing-population was not obvious from 16S rRNA-based qPCR measurements alone (Figure 2C). Ev3−a third operationally defined Dehalococcoides-like strain containing RD-OG 1,12,28,38, 40, and 48−was even more rare. This strain appears to have gained a modest presence by day 900, reaching an estimated mean of 40 ± 20 copies per 25 pg of total community DNA. This strain was near the limit of detection at the experiment’s onset and was no longer detected in the last 300

wells available to the nL-qPCR approach, it is possible to estimate the abundance of a target gene based on the combined results of multiple unique assays. While the measurement variability from this approach will be greater than that for an estimate based on a single set of primers, it is a potentially more robust option for probing previously unsequenced bacterial communities. We first explored this multiassay-per-target approach in wellcharacterized samples by examining the mean rdh gene counts in single isolates of Dehalococcoides mccartyi. We calculated the median gene count from multiple distinct assays targeting the same reference group. The mean value of these group counts was used to estimate the mean number of rdh copies per sample, which we presumed was derived from a near-clonal Dehalococcoides mccartyi population. For example, when DNA from Dehalococcoides mccartyi CBDB1 was supplied at 1 pg, 10 pg, and 25 pg per well, the mean estimates−and 95% confidence intervals−of rdh gene abundance were 960 ± 200, 9800 ± 1800, and 22000 ± 4600, respectively. Based on the 1.39 Mb genome size of the Dehalococcoides mccartyi CBDB1, 700, 7000, and 17500 copies are expected per 1 pg, 10 pg, and 25 pg of DNA, respectively. The range of estimates for individual rdh genes within a single Dehalococcoides strain were larger than expected, with some estimates greater than double the median estimate. Possible error propagated from 3-point calibration curves, notwithstanding, one possible explanation for the range in gene counts may be gene duplications, a relatively common evolutionary process in bacteria.39 Within a near-clonal population, duplicated genes exist in some portion of the population, resulting in total population level DNA that may contain some genes in higher copy number than others. In fact, in the sequenced genome of Dehalococcoides mccartyi VS, there are two instances of near identical rdh genes. Proteins ACZ61269.1 and ACZ62435.1 are 100% identical at the amino acid level, and the proteins ACZ62362.1 and ACZ62526.1 are 97% identical. Dehalococcoides cultures are maintained through years of serial transfer where gene duplication may occur. Another probable cause for the range of estimates in rdh gene abundance could be the lack of perfect complementarity between primers and intended target sequences. All primers were based on the reference nucleic acid sequences published in the NCBI database. Even for these sequenced isolates, recently accumulated mutations that may be present in the DNA retrieved for this study could produce mismatches, resulting in systematic downward Ct shifts. The likelihood of Ct shifts increases when primers are applied to previously unsequenced populations. In this context, heightened variability among gene counts is expected.

4. PILOT APPLICATION After establishing the sensitivity and selectivity of the assays in a controlled experimental context, we sought to evaluate the performance of the new nL-qPCR platform to determine subpopulation-level responses of dehalogenating microbes to electron donor limitation in a continuously fed TCE bioreactor. 4.1. Determining Population Structure and Dynamics of Organohalgen-Respiring Bacteria in a Continuous Bioreactor. We made use of the nL-qPCR platform to understand the effect of electron donor limitation on the population structure of organohalogen-respiring bacteria in a continuous-flow bioreactor inoculated with aquifer material from the Evanite contaminated site in Corvallis, OR, USA. The EV2L reactor was operated as a chemostat with influent TCE at 10 mM. After 168 days of the reactor’s 5 year operation, the influent 9663



Article

Figure 2. Reductive dehalogenase types detected in a TCE-fed reactor over a 5-year time course. (A) Hierarchical clustering of RD-OG types and unique rdh genes based on time-series correlations. (B) Median gene counts for each RD-OG and rdh at each sampled time point. Lines represent unique RD-OG, rdh, or hupL sequence types, with colors indicating assignment to a hypothesized strain based on hierarchical clustering. Unique shapes in the figure legend emphasize hupL types and biochemically characterized RD-OG. Normalized population data should be interpreted in the context of declining overall biomass observed upon lowering influent formate concentration after day 168 (Figure S6, Supporting Information). (C) Mean 16S rRNA gene level counts for Dehalococcoides and Geobacter compared with mean gene counts of hypothesized vinyl-chloride respiring strains: Ev2 containing vinyl-chloride reductase (vcrA) and Ev3 containing putative vinyl-chloride reductase (bvcA). (D) Chloroethene/ethene/hydrogen concentrations in the EV2L reactor. 9664



Article

Since increases in Geobacter 16S gene copies were observed after two sampling events, we cannot rule out the possibility that these events may have influenced the selective conditions within the reactor. For instance, an unintended introduction of trace oxygen or a change in reactor pH may tip the ecological balance in the Geobacter-like organism’s favor. The reactor appears to return to a Dehalococcoides-favorable equilibrium between days 981−1281, but a similar re-equilibration does not occur after the subsequent Geobacter increase at day 1347, suggesting a new stable state was reached (Figure 2B and 2C). Geobacter and multiple closely related Dehalococcoides-like strains often coinhabit contaminated sediment environments. Thus, better understanding of the factors governing their competition or coexistence is important for remediation outcomes. The variance in rdh estimates among genes assigned to each operational strain in the EV2L reactor, especially the data assigned to putative strains Ev1 and Ev4, is larger than what is typically obtained by traditional extensively validated qPCR assays targeting genes with exact knowledge of their nucleic acid sequences. While some of the observed variance may be due to gene duplication or fine-scale interpopulation heterogeneity, additional variance may be caused by assays amplifying genes highly similar in sequence but not 100% identical to the primer sequences. More experience with calibration procedures are needed to compare amplification behavior of standards to amplification of environmental sequences. Furthermore, achieving absolute accuracy relies on the fidelity of fluorometric DNA quantification and sufficiently broad calibration ranges to avoid extrapolation error, which may have been introduced for estimates above 20,000 starting copies per reaction. However, one intended use of the nL-qPCR approach is to identify populations of closely related dehalogenating microbes based on gene level covariation over time or in response to chemical perturbations. For this purpose, the observed decrease in accuracy represents an acceptable trade-off given the need for the diagnostic scope sufficient to economically track 50 to 100 gene clusters of interest from limited amounts of DNA input material. We have shown that rdh gene profiling by nL-qPCR is valuable for revealing the nature of competition and distinct perturbation responses among closely related organohalogen-respiring bacteria. Future work to combine metagenomic or amplicon based-gene mining22 with nL-qPCR monitoring approaches should improve assay-target complementarity and enable better understanding of spatial heterogeneity and population biology of organohalogen-respiring bacteria. In order to explore the use of the nL-qPCR tool for measuring metabolic potential in a field-application, we tested the rdh panel with DNA samples acquired by incubating aquifer pore water amended with different electron donors. We observed significant heterogeneity in population structure and stimulation response among incubated pore water microcosms taken from neighboring wells (Supporting Information, Figure S11). With the detection limits sufficiently low to apply nL-qPCR to incubated pore water samples amended with biostimulants, work is ongoing to apply this approach for site-wide analysis of biostimulated sediments and pore water within contaminated sites.

days of the time-course. The detection of this strain at such low absolute copy numbers highlights the sensitivity of this nL-qPCR platform for tracking rare populations in mixed bacterial ecosystems. In fact, this rare Dehalococcoides-like strain was not detected when similar samples were studied using a less sensitive tiling DNA−DNA hybridization microarray approach.19 Ev4−a fourth operational non-Dehalococcoides-like strain−was also detected. It is predicted to contain genes from orthologue groups 6 and 9. The known substrate range of orthologue group 6 members−so far found in Dehalobacter and Desulf itobacterium isolates−includes PCE as well as 1,2-DCA,41−43 so this strain’s precise role in a TCE-fed reactor is not obvious. Nevertheless, a niche for this strain was apparently stably maintained. 4.3. Linkage of rdh to hupL Genes. The strong correlation among hupL and rdh gene counts allowed us to infer linkages between functional genes. For instance, the genome of vcrAcontaining operational strain Ev2 appears to contain a Pinellastype hupL hydrogenase. Similarly, the gene abundance profile of the Cornell type hupL hydrogenase indicates that it is present in the numerically dominant tceA-containing population Ev1 (Figure 2B). If applied to more systems, this approach may reveal whether particular rdh and hupL genes are in linkage disequilibrium. If hydrogenases have different kinetic characteristics that are phenotypically relevant, consistent linkage between particular rdh and specific hupL types may delineate niche boundaries between subpopulations. This ecological information may prove useful for managing community structure during bioremediation, since the ratio of Dehalococcoides types influences the kinetics of different degradation steps. 4.4. Diametric Ev2 and Geobacter Population Shifts. We observed diametric shifts as one strain’s expansion consistently co-occurred with the recession of another strain and vice versa (Figure 2B). These shifts in a constantly fed mixed reactor are revealing and are suggestive of fine-scale niche boundaries determining the outcome of direct competition. Despite automation, very subtle shifts in chemical composition in a reactor may be sufficient to shift populations. Diametric shifts could also reflect density-dependent fitness dynamics observed during phagepredation on a subpopulation.44 The stability of the dominant Dehalococcoides-type suggests that predation was not the dominant ecological process in these systems, although the frequency of sampling was inadequate to rule it out completely. The modest DNA input requirements associated with the nL-qPCR technique will enable more frequent sampling regimes in future experiments. We observed an unanticipated diametric relationship between the vcrA-containing Ev2 population and the 16S rRNA marker gene for Geobacter (Figure 2C). Geobacter is most often studied as an iron-respiring bacterium.45,46 One strain of Geobacter has been shown to carry rdh genes and the capacity for growthlinked PCE-reduction.47,48 The negative correlation between a presumed Geobacter strain and Ev2 is consistent with competition for a shared resource, such as hydrogen, acetate, or a CAH electron acceptor, although other explanations for the diametric relations cannot be ruled out. The two-order of magnitude predicted increase in Geobacter population between days 1281 and 1347 coincided with a large percentage decline in the vcrA-type Ev2, but only a small decline was observed in the tceA-containg Ev1 strain. It is interesting to note that hydrogen concentrations gradually decreased from 5 nM to 1−2 nM over the period of 600 to 1731 days, corresponding to the decrease of Ev2 strain. Competition for hydrogen might be a factor for this decrease, since previous studies of different VC-respiring Dehalococcoides strains reported hydrogen thresholds near 1 nM.7,49

■

ASSOCIATED CONTENT

S Supporting Information *

Supporting Information File 1 contains oligonucleotide sequences of and information about the primers used. Additional figures, tables, and supporting methods can be found in Supporting Information File 2. Figure S1 shows RD-OG group in a sequence similarity network graph. Figures S2 and S3 9665



Article

(8) Chan, W. W. M.; Grostern, A.; Löffler, F. E.; Edwards, E. A. Quantifying the effects of 1,1,1-trichloroethane and 1,1-dichloroethane on chlorinated ethene reductive dehalogenases. Environ. Sci. Technol. 2011, 45, 9693−9702. (9) Yu, R.; Peethambaram, H. S.; Falta, R. W.; Verce, M. F.; Henderson, J. K.; Bagwell, C. E.; Brigmon, R. L.; Freedman, D. L. Kinetics of 1,2-Dichloroethane and 1,2-Dibromoethane Biodegradation in Anaerobic Enrichment Cultures. Appl. Environ. Microbiol. 2013, 79, 1359−1367. (10) Dupin, H. J.; Kitanidis, P. K.; McCarty, P. L. Pore-scale modeling of biological clogging due to aggregate expansion: A material mechanics approach. Water Resour. Res. 2001, 37, 2965−2979. (11) Stroo, H. F.; Major, D. W.; Steffan, R. J.; Koenigsberg, S. S.; Ward, C. H. Bioaugmentation with Dehalococcoides: a Decision Guide. In Bioaugmentation for Groundwater Remediation; Springer New York: New York, NY, 2012; pp 117−140. (12) Shani, N.; Rossi, P.; Holliger, C. Correlations between environmental variables and bacterial community structures suggest Fe(III) and vinyl chloride reduction as antagonistic terminal electronaccepting processes. Environ. Sci. Technol. 2013, 47, 6836−6845. (13) Hug, L. A.; Beiko, R. G.; Rowe, A. R.; Richardson, R. E.; Edwards, E. A. Comparative metagenomics of three Dehalococcoides-containing enrichment cultures: the role of the non-dechlorinating community. BMC Genomics 2012, 13, 327. (14) Lee, P. K. H.; Dill, B. D.; Louie, T. S.; Shah, M.; VerBerkmoes, N. C.; Andersen, G. L.; Zinder, S. H.; Alvarez-Cohen, L. Global transcriptomic and proteomic responses of Dehalococcoides ethenogenes strain 195 to fixed nitrogen limitation. Appl. Environ. Microbiol. 2012, 78, 1424−1436. (15) Rowe, A. R.; Heavner, G. L.; Mansfeldt, C. B.; Werner, J. J.; Richardson, R. E. Relating chloroethene respiration rates in Dehalococcoides to protein and mRNA biomarkers. Environ. Sci. Technol. 2012, 46, 9388−9397. (16) Hug, L. A.; Salehi, M.; Nuin, P.; Tillier, E. R.; Edwards, E. A. Design and Verification of a Pangenome Microarray Oligonucleotide Probe Set for Dehalococcoides spp. Appl. Environ. Microbiol. 2011, 77, 5361−5369. (17) Men, Y.; Lee, P. K. H.; Harding, K. C.; Alvarez-Cohen, L. Characterization of four TCE-dechlorinating microbial enrichments grown with different cobalamin stress and methanogenic conditions. Appl. Microbiol. Biotechnol. 2013, 97, 6439−6450. (18) Marshall, I. P. G.; Berggren, D. R. V.; Azizian, M. F.; Burow, L. C.; Semprini, L.; Spormann, A. M. The Hydrogenase Chip: a tiling oligonucleotide DNA microarray technique for characterizing hydrogen-producing and -consuming microbes in microbial communities. ISME J. 2012, 6, 814−826. (19) Marshall, I. P. G.; Azizian, M. F.; Semprini, L.; Spormann, A. M. Inferring community dynamics of organohalide-respiring bacteria in chemostats by covariance of rdhA gene abundance. FEMS Microbiol. Ecol. 2014, 86, 428−440. (20) Maphosa, F.; de Vos, W. M.; Smidt, H. Exploiting the ecogenomics toolbox for environmental diagnostics of organohaliderespiring bacteria. Trends Biotechnol. 2010, 28, 308−316. (21) Hug, L. A.; Maphosa, F.; Leys, D.; Löffler, F. E.; Smidt, H.; Edwards, E. A.; Adrian, L. Overview of organohalide-respiring bacteria and a proposal for a classification system for reductive dehalogenases. Philos. Tran. R. Soc. London, Ser. B 2013, 368, 20120322−20120322. (22) Hug, L. A.; Edwards, E. A. Diversity of reductive dehalogenase genes from environmental samples and enrichment cultures identified with degenerate primer PCR screens. Front Microbiol. 2013, 4, 1−16. (23) Krajmalnik-Brown, R.; Hölscher, T.; Thomson, I. N.; Saunders, F. M.; Ritalahti, K. M.; Löffler, F. E. Genetic identification of a putative vinyl chloride reductase in Dehalococcoides sp. strain BAV1. Appl. Environ. Microbiol. 2004, 70, 6347−6351. (24) Stedtfeld, R. D.; Baushke, S. W.; Tourlousse, D. M.; Miller, S. M.; Stedtfeld, T. M.; Gulari, E.; Tiedje, J. M.; Hashsham, S. A. Development and experimental validation of a predictive threshold cycle equation for quantification of virulence and marker genes by high-throughput

provide important bioinformatics parameters. Figures S4 and S5 show assay validation results against four Dehalococcoides isolates. Figure S6 shows Total Suspended Solids in EV2L reactor. Figures S7 and S8 compare nL-performance of previously validated tceA and vcrA PCR assays to newly developed RD-OG 5 and RD-OG 8 assays in this study. Figures S9 and S10 show expected technical variance at low copy numbers. Figure S11 shows results from pore-water incubation studies. Table S1 reports the accessions, genus, and associated RD-OG of each reference sequences used. Tables S2, S3, S4, and S5 show assays and validation results for hupL and 16S rRNA gene primers. Tables S6 and S7 report practical concerns involved in nL-qPCR. Information on method of quantification, practical detection limits, and detection of rare community members is also provided. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: 650-723-3668. E-mail: [email protected]. Corresponding author address: Stanford University, Clark Center E250, 318 Campus Drive, Stanford, CA 94305. Funding

This work was funded by Eni S.p.A. and NSF Grant (MCB1330832) to A.M.S. and L.S. K.M.B. was additionally supported by a NSF Graduate Research Fellowship (2011103493). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Jennifer Dang and Kathryn Thompson (UCSF Genome Core) for meticulous handling of the microfluidics. The laboratories of Drs. Lorenz Adrian and Ruth Richardson made kind gifts of DNA from Dehalococcoides strains. Drs. Susan Holmes (Stanford Department of Statistics) and Ian Marshall (Aarhus University) provided helpful discussion.

■

REFERENCES

(1) Bouwer, E. J.; McCarty, P. L. Transformations of 1- and 2-carbon halogenated aliphatic organic compounds under methanogenic conditions. Appl. Environ. Microbiol. 1983, 45, 1286−1294. (2) DiStefano, T. D.; Gossett, J. M.; Zinder, S. H. Reductive dechlorination of high concentrations of tetrachloroethene to ethene by an anaerobic enrichment culture in the absence of methanogenesis. Appl. Environ. Microbiol. 1991, 57, 2287−2292. (3) Seshadri, R.; Adrian, L.; Fouts, D. E.; Eisen, J. A.; Phillippy, A. M.; Methe, B. A.; Ward, N. L.; Nelson, W. C.; Deboy, R. T.; Khouri, H. M.; et al. Genome sequence of the PCE-dechlorinating bacterium Dehalococcoides ethenogenes. Science 2005, 307, 105−108. (4) McMurdie, P. J.; Behrens, S. F.; Müller, J. A.; Göke, J.; Ritalahti, K. M.; Wagner, R.; Goltsman, E.; Lapidus, A.; Holmes, S.; Löffler, F. E.; et al. Localized plasticity in the streamlined genomes of vinyl chloride respiring Dehalococcoides. PLoS Genet. 2009, 5, e1000714. (5) Moe, W. M.; Yan, J.; Nobre, M. F.; da Costa, M. S.; Rainey, F. A. Dehalogenimonas lykanthroporepellens gen. nov., sp. nov., a reductively dehalogenating bacterium isolated from chlorinated solvent-contaminated groundwater. Int. J. Syst. Evol. Microbiol. 2009, 59, 2692−2697. (6) Yu, S.; Semprini, L. Kinetics and modeling of reductive dechlorination at high PCE and TCE concentrations. Biotechnol. Bioeng. 2004, 88, 451−464. (7) Cupples, A. M.; Spormann, A. M. Vinyl chloride and cisdichloroethene dechlorination kinetics and microorganism growth under substrate limiting conditions. Environ. Sci. Technol. 2004, 38, 1102−1107. 9666



Article

nanoliter-volume PCR on the OpenArray platform. Appl. Environ. Microbiol. 2008, 74, 3831−3838. (25) Müller, J. A.; Rosner, B. M.; von Abendroth, G.; MeshulamSimon, G.; McCarty, P. L.; Spormann, A. M. Molecular identification of the catabolic vinyl chloride reductase from Dehalococcoides sp. strain VS and its environmental distribution. Appl. Environ. Microbiol. 2004, 70, 4880−4888. (26) Punta, M.; Coggill, P. C.; Eberhardt, R. Y.; Mistry, J.; Tate, J.; Boursnell, C.; Pang, N.; Forslund, K.; Ceric, G.; Clements, J.; et al. The Pfam protein families database. Nucleic Acids Res. 2012, 40, D290− D301. (27) Altschul, S. F.; Gish, W.; Miller, W.; Myers, E. W.; Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 1990, 215, 403−410. (28) Song, N.; Joseph, J. M.; Davis, G. B.; Durand, D. Sequence Similarity Network Reveals Common Ancestry of Multidomain Proteins. PLoS Comput. Biol. 2008, 4, e1000063. (29) Rozen, S.; Skaletsky, H. Primer3 on the WWW for general users and for biologist programmers. Methods Mol. Biol. 2000, 132, 365−386. (30) Cummings, D. E.; Snoeyenbos-West, O. L.; Newby, D. T.; Niggemyer, A. M.; Lovley, D. R.; Achenbach, L. A.; Rosenzweig, R. F. Diversity of Geobacteraceae species inhabiting metal-polluted freshwater lake sediments ascertained by 16S rDNA analyses. Microb. Ecol. 2003, 46, 257−269. (31) Smits, T. H. M.; Devenoges, C.; Szynalski, K.; Maillard, J.; Holliger, C. Development of a real-time PCR method for quantification of the three genera Dehalobacter, Dehalococcoides, and Desulf itobacterium in microbial communities. J. Microbiol. Methods 2004, 57, 369−378. (32) Hendrickson, E. R.; Payne, J. A.; Young, R. M.; Starr, M. G.; Perry, M. P.; Fahnestock, S.; Ellis, D. E.; Ebersole, R. C. Molecular analysis of Dehalococcoides 16S ribosomal DNA from chloroethene-contaminated sites throughout North America and Europe. Appl. Environ. Microbiol. 2002, 68, 485−495. (33) Behrens, S.; Azizian, M. F.; McMurdie, P. J.; Sabalowsky, A.; Dolan, M. E.; Semprini, L.; Spormann, A. M. Monitoring abundance and expression of “Dehalococcoides” species chloroethene-reductive dehalogenases in a tetrachloroethene-dechlorinating flow column. Appl. Environ. Microbiol. 2008, 74, 5695−5703. (34) Berggren, D. R. V.; Marshall, I. P. G.; Azizian, M. F.; Spormann, A. M.; Semprini, L. Effects of sulfate reduction on the bacterial community and kinetic parameters of a dechlorinating culture under chemostat growth conditions. Environ. Sci. Technol. 2013, 47, 1879−1886. (35) Azizian, M. F.; Behrens, S.; Sabalowsky, A.; Dolan, M. E.; Spormann, A. M.; Semprini, L. Continuous-flow column study of reductive dehalogenation of PCE upon bioaugmentation with the Evanite enrichment culture. J. Contam. Hydrol. 2008, 100, 11−21. (36) Kube, M.; Beck, A.; Zinder, S. H.; Kuhl, H.; Reinhardt, R.; Adrian, L. Genome sequence of the chlorinated compound-respiring bacterium Dehalococcoides species strain CBDB1. Nat. Biotechnol. 2005, 23, 1269− 1273. (37) Morrison, T.; Hurley, J.; Garcia, J.; Yoder, K.; Katz, A.; Roberts, D.; Cho, J.; Kanigan, T.; Ilyin, S. E.; Horowitz, D.; et al. Nanoliter high throughput quantitative PCR. Nucleic Acids Res. 2006, 34, e123−e123. (38) Dixon, J. M.; Lubomirski, M.; Amaratunga, D.; Morrison, T. B.; Brenan, C. J. H.; Ilyin, S. E. Nanoliter high-throughput RT-qPCR: a statistical analysis and assessment. BioTechniques 2009, 46, ii−viii. (39) Andersson, D. I.; Hughes, D. Gene amplification and adaptive evolution in bacteria. Annu. Rev. Genet. 2009, 43, 167−195. (40) Maymó-Gatell, X.; Chien, Y.; Gossett, J. M.; Zinder, S. H. Isolation of a bacterium that reductively dechlorinates tetrachloroethene to ethene. Science 1997, 276, 1568−1571. (41) Grostern, A.; Edwards, E. A. Characterization of a Dehalobacter Coculture That Dechlorinates 1,2-Dichloroethane to Ethene and Identification of the Putative Reductive Dehalogenase Gene. Appl. Environ. Microbiol. 2009, 75, 2684−2693. (42) Suyama, A.; Yamashita, M.; Yoshino, S.; Furukawa, K. Molecular characterization of the PceA reductive dehalogenase of Desulfitobacterium sp. strain Y51. J. Bacteriol. 2002, 184, 3419−3425. (43) Marzorati, M.; De Ferra, F.; Van Raemdonck, H.; Borin, S.; Allifranchini, E.; Carpani, G.; Serbolisca, L.; Verstraete, W.; Boon, N.;

Daffonchio, D. A Novel Reductive Dehalogenase, Identified in a Contaminated Groundwater Enrichment Culture and in Desulfitobacterium dichloroeliminans Strain DCA1, Is Linked to Dehalogenation of 1,2Dichloroethane. Appl. Environ. Microbiol. 2007, 73, 2990−2999. (44) Rodriguez-Valera, F.; Martin-Cuadrado, A.-B.; Rodriguez-Brito, B.; Pasić, L.; Thingstad, T. F.; Rohwer, F.; Mira, A. Explaining microbial population genomics through phage predation. Nat. Rev. Microbiol. 2009, 7, 828−836. (45) Lovley, D. R.; Giovannoni, S. J.; White, D. C.; Champine, J. E.; Phillips, E. J.; Gorby, Y. A.; Goodwin, S. Geobacter metallireducens gen. nov. sp. nov., a microorganism capable of coupling the complete oxidation of organic compounds to the reduction of iron and other metals. Arch. Microbiol. 1993, 159, 336−344. (46) Caccavo, F.; Lonergan, D. J.; Lovley, D. R.; Davis, M.; Stolz, J. F.; McInerney, M. J. Geobacter sulfurreducens sp. nov., a hydrogen- and acetate-oxidizing dissimilatory metal-reducing microorganism. Appl. Environ. Microbiol. 1994, 60, 3752−3759. (47) Sung, Y.; Fletcher, K. E.; Ritalahti, K. M.; Apkarian, R. P.; RamosHernández, N.; Sanford, R. A.; Mesbah, N. M.; Löffler, F. E. Geobacter lovleyi sp. nov. strain SZ, a novel metal-reducing and tetrachloroethenedechlorinating bacterium. Appl. Environ. Microbiol. 2006, 72, 2775− 2782. (48) Wagner, D. D.; Hug, L. A.; Hatt, J. K.; Spitzmiller, M. R.; PadillaCrespo, E.; Ritalahti, K. M.; Edwards, E. A.; Konstantinidis, K. T.; Löffler, F. E. Genomic determinants of organohalide-respiration in Geobacter lovleyi, an unusual member of the Geobacteraceae. BMC Genomics 2012, 13, 200. (49) Sung, Y.; Ritalahti, K. M.; Apkarian, R. P.; Löffler, F. E. Quantitative PCR confirms purity of strain GT, a novel trichloroetheneto-ethene-respiring Dehalococcoides isolate. Appl. Environ. Microbiol. 2006, 72, 1980−1987.

9667
