Dehalococcoides mccartyi gen. nov., sp. nov., obligately organohalide-respiring anaerobic bacteria relevant to halogen cycling and bioremediation, belong to a novel bacterial class, Dehalococcoidia classis nov., order Dehalococcoidales ord. nov. and family Dehalococcoidaceae fam. nov., within the phylum Chloroflexi

Enrichment and isolation efforts with chlorinated ethenes and chlorinated benzenes as electron acceptors yielded unique, organohalide-respiring bacterial isolates from geographically distinct freshwater locations that shared more than 98 % 16S rRNA gene sequence similarity. The first isolate obtained dechlorinated tetrachloroethene (PCE) to vinyl chloride, ethene and inorganic chloride and was designated ‘Dehalococcoides ethenogenes’ strain 195 (Maymó-Gatell et al., 1997). Members of the ‘Dehalococcoides’ group play key roles in the bioremediation of chlorinated solvent-contaminated sites, and consortia have been developed for bioaugmentation (Löffler & Edwards, 2006). Because of this successful industrial application, reductive dechlorination and organohalide respiration have been the subject of intense study for over a decade (Löffler et al., 2013). The name ‘Dehalococcoides’ has been used extensively in the literature and Hugenholtz & Stackebrandt (2004) have proposed the class ‘Dehalococcoidetes’; however, the genus ‘Dehalococcoides’ has yet to be described. The intent of this work is the formal taxonomic description of the genus Dehalococcoides, and we present phenotypic, phylogenetic and genotypic characteristics of organohalide-respiring Dehalococcoides isolates including strains 195, BAV1, CBDB1, FL2, GT and VS. All strains were enriched from anoxic microcosms established with river sediments, aquifer materials or digester sludge (Table 1) amended with fermentable substrates (e.g. benzoate, pyruvate, lactate) or hydrogen as electron donor and a chlorinated ethene or a chlorinated benzene as electron acceptor. The enrichment process included repeated transfers (1–10 % inoculum, v/v) in reduced, completely synthetic mineral salts medium amended with acetate as the carbon source, hydrogen as electron donor (supplied in 2- to 200-fold excess of the theoretical electron equivalents required for reduction of the chlorinated electron acceptor), vitamins (Wolin et al., 1964) and a chlorinated ethene (strains 195, BAV1, FL2, GT, VS) or a trichlorobenzene (strain CBDB1) as electron acceptor (Adrian et al., 2000; He et al., 2003, 2005; Maymó-Gatell et al., 1997; Müller et al., 2004; Sung et al., 2006). Methanogenic archaea were eliminated by the addition of high concentrations of chlorinated solvents (up to 0.5 mM aqueous concentration) (strains BAV1, GT, 195) or 2–4 mM 2-bromoethanesulfonate (strains CBDB1, FL2 and VS). Selection against H₂/CO₂-metabolizing homoacetogens was accomplished by repeated transfers to medium amended with antibiotics that interfere with peptidoglycan biosynthesis [i.e. vancomycin (0.1 g l⁻¹) or ampicillin (0.05–3 g l⁻¹)] or transfer to bicarbonate-free medium (strain VS) (Müller et al., 2004) or medium with acetate as the only source of reducing equivalents (strain BAV1) (He et al., 2002, 2003). The medium composition used for isolation and maintenance of the isolates varied slightly, and used different reductants including iron sulfide (FeS) (0.5 mM nominal concentration) (Ehrenreich & Widdel, 1994), a combination of sodium sulfide (0.2–0.5 mM), l-cysteine (0.2 mM) and dl-dithiothreitol (0.5–1 mM), or titanium(III) complexed with citrate or nitrilotriacetate (0.2–1.5 mM) (Moench & Zeikus, 1983; Zehnder & Wuhrmann, 1976). Isolation efforts used the dilution-to-extinction principle in liquid and semi-solid [0.5 % (w/v) low-melting-point agarose; gelling point <30 °C] medium (Adrian et al., 2000; Löffler et al., 2005). Dechlorinating activity was recovered from 10⁻⁵ to 10⁻⁷ dilution tubes and repeated dilution-to-extinction series yielded pure cultures of strains 195, VS, FL2 and GT. Tiny (<0.1 mm), opaque and barely visible colonies formed after 1.5 to 3 months in soft agar medium, and transfer of single colonies to fresh medium yielded pure cultures of strains BAV1 and CBDB1.

Culture purity was examined by multiple approaches including the generation of 16S rRNA gene clone libraries and sequencing of the inserts of multiple clones, quantitative enumeration of strain-specific reductive dehalogenase (RDase) genes (Ritalahti et al., 2006; Sung et al., 2006; Adrian et al., 2007) and 16S rRNA genes (Ritalahti et al., 2006), microscopic observation, growth experiments with alternate combinations of electron donor and acceptors and in rich media, as well as genome sequencing. The 16S rRNA genes of the isolates were highly similar, sharing more than 98 % sequence identity (Fig. 1). Phylogenetic analysis affiliated the sequences of the isolates with the phylum Chloroflexi, a deeply branching lineage within the domain . This poorly characterized bacterial phylum encompasses a phylogenetically diverse collection of 16S rRNA gene sequences and is separated into at least six major lineages including the Anaerolineae, Caldilineae, Dehalococcoidia (proposed in this study), the marine SAR202 cluster (without cultured representatives), ‘traditional’ Chloroflexi and the Thermomicrobia (Hugenholtz & Stackebrandt, 2004; Yamada et al., 2006). The class ‘Dehalococcoidetes’ was informally introduced to accommodate organohalide-respiring bacteria related to ‘Dehalococcoides ethenogenes’ strain 195 (Hugenholtz & Stackebrandt, 2004). The Dehalococcoides isolate 16S rRNA gene sequences, together with a large number of environmental and mixed-culture clone sequences, form a tight cluster, which Hendrickson et al. (2002) divided into three phylogenetic subgroups based on sequence signatures in the hypervariable V2 and V6 regions of the 16S rRNA gene. The ‘Cornell’ subgroup includes strains 195 and MB (Cheng & He, 2009), the ‘Victoria’ subgroup includes strain VS and the ‘Pinellas’ subgroup comprises the other available Dehalococcoides isolates, as well as most cultured Dehalococcoides strains identified in mixed cultures and the majority of the environmental clone sequences (Fig. 1). Members of the ‘Pinellas’ subgroup share identical or highly similar (1–3 nucleotide differences) 16S rRNA gene sequences, and 14–16 and 23–24 base differences distinguish the 16S rRNA genes of members of the Pinellas subgroup from members of the Victoria and Cornell subgroups, respectively. In addition to Dehalococcoides, the class Dehalococcoidia encompasses two other organohalide-respiring clusters, represented by the isolate ‘Dehalobium chlorocoercia’ DF-1 and Dehalogenimonas lykanthroporepellens BL-DC-8 and BL-DC-9^T (Fig. 1).

The Dehalococcoides isolates exhibited a highly restricted catabolic range, and only certain chlorinated and brominated compounds served as respiratory electron acceptors (Table 1). All isolates required hydrogen as the electron donor, which could not be replaced by formate. Acetate served as a carbon source. Fermentative growth was not observed, and no other growth-supporting electron donor/acceptor combinations were identified. Comparative genome analysis of strains 195, CBDB1, BAV1, GT and VS (Kube et al., 2005; Seshadri et al., 2005; McMurdie et al., 2009) identified no other respiratory or fermentative energy-conserving pathways, thus corroborating the results from physiological studies that members of this bacterial group are strictly organohalide-respiring bacteria. Cytochromes were absent. Dechlorination and growth occurred between pH 6 and 8, with highest activity measured between pH 6.9 and 7.5. All isolates were mesophilic, and dechlorination was measured at incubation temperatures ranging from 15 to 35 °C, with optimum growth occurring at 25–30 °C, except for strain 195, which was cultured from an anoxic digester and had an optimum near 35 °C. No dechlorination occurred at 4 or 40 °C, but cultures stored at 4 °C or at room temperature for several months recovered dechlorination activity after prolonged lag times. Spores were not found, and incubation at 45 °C or above resulted in loss of dechlorination activity and viability. The lag time before dechlorination started following 1–5 % (v/v) transfers under optimum temperature and pH conditions ranged from a few days to 2–6 weeks, depending on the reductant and the age of the inoculum. The shortest lag times were observed in medium reduced with FeS and/or titanium (III). Similar dechlorination performance and growth were observed when incubated under static conditions or with modest agitation. The doubling times and growth yields of the isolates ranged from 0.8 to 3 days and 6.3×10⁷ to 3.1×10⁸ cells per µmol chloride released, respectively (Table 2). Although densities of 10⁷–10⁸ cells per ml culture fluid were achieved, turbidity was too low to be monitored via optical density measurements. The low biomass yield of about 1 mg (wet wt) from 100 ml culture suspension has hampered detailed biochemical studies and prevented certain chemotaxonomic analyses (e.g. cell-wall composition) typically included in the description of a novel genus.

Microscopic analysis of cell suspensions revealed that the cells were small and shared unusual morphology features. Suspended cells were visible by phase-contrast microscopy for brief moments but then ‘disappeared’ from the field of view, apparently due to a disc-shaped morphology with a thickness of less than 0.2 µm, which is the limit of the light microscope’s resolving power. Transmission and scanning electron microscopy confirmed round, disc-shaped cells, no more than 1 µm wide and 0.1–0.2 µm thick, and characteristic biconcave indentations on opposite flat sides of the cell were visible (Fig. 2a–d). The estimated volume of a single Dehalococcoides cell is 0.02 µm³, which is roughly 30-fold lower than the mean Escherichia coli cell volume of 0.6–0.7 µm³ (Duhamel et al., 2004; Kubitschek, 1990; Löffler et al., 2013), or about twice the cell volume of ‘Pelagibacter ubique’ (SAR11) (Rappé et al. 2002). Scanning electron microscopy revealed unusual cell surface features such as filamentous appendages (Fig. 2a). The function of these appendages is unclear, but they may play a role in attachment of cells to each other or to surfaces, which has been observed in laboratory and field studies (Lendvay et al., 2003; Amos et al., 2009). Motility was never observed, and no growth occurred on agar surfaces. The addition of vancomycin and ampicillin did not prevent dechlorination or growth through several subsequent transfers, suggesting that the Dehalococcoides isolates were resistant to antibiotics that interfere with peptidoglycan biosynthesis. This finding suggested an unusual cell-wall structure, a hypothesis supported by the observations that the cell wall of strain 195 did not react with peptidoglycan-specific lectin (Maymó-Gatell et al., 1997). Transmission electron micrographs revealed an unusual cell-wall ultrastructure (Fig. 2e, f) that resembled the S-layer cell walls of archaea (Kandler, 1993), with no discernible peptidoglycan layer as typically seen in micrographs of bacteria. A 105–110 kDa protein (DET1407; GenBank accession number AAW39334) has been proposed to be an S-layer component (Morris et al., 2006).

The genome sequences of Dehalococcoides strains 195, BAV1, CBDB1, GT and VS are available (Table S1, available in IJSEM Online), and genome analyses corroborated the morphological and physiological observations. For example, genes encoding hydrogenase and reductive dehalogenase protein complexes were present, but genes implicated in sulfate, nitrate or fumarate reduction were absent. Genes encoding cytochrome-containing proteins and cytochrome biosynthesis were absent. Therefore, the genome analysis supported experimental evidence that Dehalococcoides strains are restricted to cytochrome-independent organohalide respiration as the only means of energy conservation. Motility genes, sporulation genes and genes for peptidoglycan biosynthesis were not identified.

Oxygen exposure irreversibly inhibited dechlorination, growth and viability (Amos et al., 2008), and strictly anoxic techniques for culture manipulations were vital. For example, plastic syringes were stored inside anoxic chambers for several days or filled with sterile solutions of reductants [i.e. 1 mM Na₂S or 1 mM titanium (III)] prior to flushing with oxygen-free N₂ and use. The redox potential of the medium also affected dechlorination and viability, and successful cultivation required the addition of reductants. Cultures with redox potentials above −110 mV (i.e. redox indicator resazurin pink), but without measurable oxygen, did not exhibit dechlorination activity, even after the addition of reductant(s). Interestingly, growth of strain CBDB1 in the titanium (III)-reduced medium devoid of any amended sulfur compound was sustained, suggesting that impurities of reagent-grade chemicals provide sufficient sulfur to fulfil the growth requirements of Dehalococcoides strains.

To achieve standardized growth conditions for the available Dehalococcoides isolates, a MOPS-buffered medium containing the following components (per litre) was devised: 1 g NaCl (17 mM), 0.41 g MgCl₂ . 6H₂O (2 mM), 0.27 g NH₄Cl (5 mM), 0.52 g KCl (7 mM), 0.15 g CaCl₂ . 2H₂O (1 mM), 0.2 g KH₂PO₄ (1.5 mM), 0.25–1 mg resazurin (1–4.4 µM), 0.41 g sodium acetate (5 mM), 1 ml trace element solution (1000× stock) and 1 ml selenite-tungstate solution (1000× stock) (Löffler et al., 1999). The medium was boiled and cooled to room temperature under a stream of N₂ before the following components were added (final concentrations per litre): 0.084 g NaHCO₃ (1 mM), 0.048 g Na₂S . 9H₂O (0.2 mM) and 0.031 g l-cysteine (0.2 mM). The medium was dispensed into N₂-flushed 20- to 160-ml glass vials or serum bottles and closed with thick butyl-rubber stoppers (Geo-Microbial Technologies or Bellco Glass) or Teflon-lined butyl-rubber septa (Ochs) and aluminium crimps. After autoclaving at 121 °C for 30 min, the following additions were made: 20 mM filter-sterilized sodium MOPS (1 M stock, adjusted to pH 7.4 with 1 M HCl), 0.5 mM titanium (III) nitrilotriacetate (25 mM stock, filter-sterilized), vitamins [0.5 ml three-vitamin 2000× stock, 1.0 ml six-vitamin 1000× stock and 1.0 ml cyanocobalamin 1000× stock; all filter-sterilized (Rosner et al., 1997)]. Sterile hydrogen gas (2–20 % of headspace, v/v) was added using a syringe filter (0.2 µm). The final medium had a pH of 7.1–7.2 and was reduced (i.e. resazurin colourless) and without precipitate. All Dehalococcoides strains grew in this medium, and cultures were transferred (2 % inoculum, v/v) to fresh medium every 2–10 weeks.

To generate biomass for lipid analysis, the Dehalococcoides isolates were grown in this MOPS-buffered medium and amended with two or three feedings of substrate [i.e. PCE for strain 195; trichloroethene (TCE) for strain GT and FL2; cis-dichloroethene (DCE) for strain BAV1; an equimolar 1,2,3-/1,2,4-trichlorobenzene mixture for strain CBDB1] to achieve dehalogenation of a total amount of about 100 µmol chlorinated electron acceptor. When about 80–90 % of the final electron acceptor amendment was consumed, the bottles were shipped via overnight carrier to Microbial Insights (). The total lipid dissolved in chloroform was fractionated, and the phospholipid fraction was analysed for phospholipid fatty acid (PLFA) profiles (White et al., 2005). The cellular fatty acids of the Dehalococcoides strains analysed included large proportions of even-carbon-number, straight-chain and mid-chain-branched, saturated fatty acids, which comprised 51.8–64.3 % and 17.4–43.5 %, respectively, of the total PLFA (Table S3). These values are similar to those reported previously for strain BAV1 (68.8 %) and strain FL2 (59.5 %) grown in a different mineral salts medium (White et al., 2005). The major PLFA were 14 : 0 (15.7 mol%), br15 : 0 (6.2 mol%), 16 : 0 (22.7 mol%), 10-methyl 16 : 0 (25.8 mol%) and 18 : 0 (16.6 mol%). Non-branched monounsaturated PLFA were present in low abundances in strains FL2, BAV1 and GT (<2.5 mol%), and larger amounts were detected in strain CBDB1 (<7.2 mol%) and strain 195 (10.4 mol%). In a previous analysis, no monounsaturated PLFA were detected in strains FL2 and BAV1; however, unusual furan fatty acids including 9-(5-pentyl-2-furyl)-nonanoate (Fu18 : 2ω6) and 8-(5-hexyl-2-furyl)-octanoate (Fu18 : 2ω6) were found (White et al., 2005). The current study corroborated the presence of two 18-carbon furan PLFA in strains FL2 (3.7 mol%), BAV1 (2.0 mol%) and GT (7.7 mol%), although these unusual furan fatty acids were not detected in strain CBDB1 or strain 195. Odd-carbon-number furan fatty acids were not detected.

Phenotypic properties including a non-phototrophic, strictly anaerobic and mesophilic metabolism restricted to organohalide respiration, G+C content and a unique cell morphology distinguish members of the Dehalococcoidia from cultured representatives of all other lineages currently affiliated with the phylum Chloroflexi (discussed by Moe et al., 2009). The Dehalococcoidia currently comprises three lineages: Dehalococcoides, represented by the six isolates described herein, strain MB (Cheng & He, 2009) and strains ANAS1 and ANAS2 recently obtained from a trichloroethene-to-ethene-dechlorinating consortium (Lee et al., 2011), the genus Dehalogenimonas (represented by Dehalogenimonas lykanthroporepellens strains BL-DC-8 and BL-DC-9^T), and a separate branch represented by a single isolate informally named ‘Dehalobium chlorocoercia’ strain DF-1 (Fig. 1). The cultured Dehalococcoidia depend strictly on organohalide respiration to conserve energy for maintenance and growth; however, key differences exist between the three currently known branches. The Dehalococcoides share ~90 and ~87.5 % 16S rRNA gene sequence identity, respectively, with Dehalogenimonas lykanthroporepellens and ‘Dehalobium chlorocoercia’. The Dehalococcoides strains share properties with the two described Dehalogenimonas strains, but genotypic, chemotaxonomic and phenotypic features distinguish the two groups. Aside from 10–11 % 16S rRNA gene sequence divergence, the G+C content of the Dehalococcoides genomes is 6–8 % lower than the 55 mol% G+C determined for the Dehalogenimonas lykanthroporepellens genome and 10–12 % lower than the G+C content of the ‘Dehalobium chlorocoercia’ genome (Table S1). While Dehalogenimonas lykanthroporepellens BL-DC-8 and BL-DC-9^T were reported to reduce electron acceptors via a dihaloelimination mechanism (i.e. two chlorine atoms are eliminated from adjacent, saturated carbon atoms to yield an alkene in a two-reducing-equivalent reduction), the six Dehalococcoides strains carry out hydrogenolysis and, in the case of some substrates like 1,2-dichloroethane, dihaloelimination (except strain CBDB1). The Dehalococcoides strains were irreversibly inhibited by brief oxygen exposure, whereas strains BL-DC-8 and BL-DC-9^T were more oxygen tolerant (Yan et al., 2009). ‘Dehalobium chlorocoercia’ is represented by a single strain, designated DF-1 (May et al., 2008), and the genus and species names have yet to be validly published. Distinguishing features of strain DF-1 from the Dehalococcoides strains include 10 % 16S rRNA gene sequence divergence, a smaller cell size of 137±51 nm thickness and 75–339 nm diameter, sheaths or capsules surrounding the cells, aggregate growth, the use of formate as an electron donor, and susceptibility to antibiotics that interfere with peptidoglycan biosynthesis. Further, strain DF-1 cannot be grown in defined medium and requires co-cultivation with a Desulfovibrio strain or the addition of a Desulfovibrio strain cell extract (May et al., 2008).

Results from whole-genome comparisons (Table S2) are consistent with the proposal to include all described Dehalococcoides strains in a single species of the new genus Dehalococcoides gen. nov. Strains BAV1, CBDB1 and GT show high genome-aggregate average nucleotide identities (ANI) (>98.7 %). These values are well above the 95 % ANI that corresponds to the 70 % DNA–DNA hybridization standard that is frequently used for species demarcation (Goris et al., 2007), revealing that these strains show a high degree of genomic relatedness. Strains BAV1, CBDB1 and GT show lower ANI values, 86–87 %, to strains VS and 195 (Table S2). However, the gene-content differences among the most phylogenetically divergent pairs of genomes are comparable to, if not smaller than (e.g. BAV1/VS vs BAV1/GT comparison), the gene-content differences seen among the most closely related genomes. The application of a Dehalococcoides pangenome microarray confirmed extensive gene-content similarities experimentally between the different strains (Hug et al., 2011). The five sequenced Dehalococcoides genomes carry a similar number of strain-specific genes, regardless of the degree of their evolutionary divergence, and these genes are restricted to the RDase and accessory genes, as well as hypothetical genes, encoded in genomic islands and mobile genetic elements (Kube et al., 2005; Seshadri et al., 2005; McMurdie et al., 2009). Although the genome sequences of strains FL2, MB, ANAS1 and ANAS2 are not available, microarray studies suggest that they contain the same characteristic Dehalococcoides core genes (Hug et al., 2011; Lee et al., 2011). The presence/absence of RDase operons reflect the main difference in gene content between the Dehalococcoides strains. The high gene-content similarity indicates that these Dehalococcoides strains share overlapping ecological niches and are under similar ecological constraints; otherwise, the strain-specific environmental settings would have selected for different gene contents. This contrasts with observations for members of other genera such as Escherichia, Shewanella and Burkholderia, which show similar evolutionary relatedness (e.g. share similar ANI values) to each other to that shown by the Dehalococcoides strains, but are more versatile and diverse, and therefore share fewer functional genes (Konstantinidis et al., 2006). Based on their high shared gene content and genomic relatedness, as well as unifying morphological and physiological characteristics, we propose that the Dehalococcoides strains should be included within the same species.

The sensitivity to oxygen and redox potential limits the distribution of Dehalococcoides to environments where strictly anoxic conditions prevail. Not surprisingly, all Dehalococcoides isolates were derived from anoxic environments, including digester sludge, sediments and aquifers. Except for strain FL2, the Dehalococcoides isolates were obtained from contaminated environments. The quest for microbes capable of chlorinated compound transformation has introduced a strong bias towards sampling polluted sites, but the isolation of strain FL2 from a river sediment with no reported contamination with halogenated compounds demonstrated that Dehalococcoides strains are members of natural microbial communities not impacted by anthropogenic contamination. The obvious question that arises is, how do strictly organohalide-respiring Dehalococcoides survive in pristine environments and conserve energy for cell maintenance and growth? A plausible explanation is the use of naturally occurring organohalides, which have been observed in a variety of habitats including marine ecosystems, sediments, subsurface environments and soils (Öberg, 2002; Gribble, 2003; Krzmarzick et al., 2012). The production of organohalogens in soils has been demonstrated (de Jong et al., 1994; Hoekstra et al., 1999), and anoxic microenvironments occur in unsaturated soils (Schink, 2006), suggesting that the habitat range of Dehalococcoides probably extends beyond anoxic sediments, sludges and saturated aquifers, which have been the source materials for enrichment efforts to date. Moreover, the presence of 12–36 sets of genes predicted to encode RDases indicates that this genus is highly adapted to use naturally occurring as well as anthropogenic organohalogens.

Phylogenetic, phenotypic and genomic characteristics distinguish the Dehalococcoides isolates from other genera in the phylum Chloroflexi, including other organohalide-respiring populations (i.e. Dehalogenimonas lykanthroporepellens and ‘Dehalobium chlorocoercia’; May et al., 2008; Moe et al., 2009). We propose the new genus, Dehalococcoides gen. nov., with Dehalococcoides mccartyi sp. nov. as the type species, to accommodate the new isolates. The type species of the genus, Dehalococcoides mccartyi, was formerly designated ‘Dehalococcoides ethenogenes’. Dehalococcoides mccartyi is embedded in the family Dehalococcoidaceae fam. nov., of the order Dehalococcoidales ord. nov., of the class Dehalococcoidia classis nov. Based on phylogenetic, physiological, morphological and genomic differences, we propose that Dehalogenimonas and ‘Dehalobium’ represent distinct genera within the Dehalococcoidia (Kittelmann & Friedrich, 2008; Hugenholtz & Stackebrandt, 2004).