Long-term survival of Dehalococcoides mccartyi strains in mixed cultures under electron acceptor and ammonium limitation

Abstract Few strains of Dehalococcoides mccartyi harbour and express the vinyl chloride reductase (VcrA) that catalyzes the dechlorination of vinyl chloride (VC), a carcinogenic soil and groundwater contaminant. The vcrA operon is found on a Genomic Island (GI) and, therefore, believed to participate in horizontal gene transfer (HGT). To try to induce HGT of the vcrA-GI, we blended two enrichment cultures in medium without ammonium while providing VC. We hypothesized that these conditions would select for a mutant strain of D. mccartyi that could both fix nitrogen and respire VC. However, after more than 4 years of incubation, we found no evidence for HGT of the vcrA-GI. Rather, we observed VC-dechlorinating activity attributed to the trichloroethene reductase TceA. Sequencing and protein modelling revealed a mutation in the predicted active site of TceA, which may have influenced substrate specificity. We also identified two nitrogen-fixing D. mccartyi strains in the KB-1 culture. The presence of multiple strains of D. mccartyi with distinct phenotypes is a feature of natural environments and certain enrichment cultures (such as KB-1), and may enhance bioaugmentation success. The fact that multiple distinct strains persist in the culture for decades and that we could not induce HGT of the vcrA-GI suggests that it is not as mobile as predicted, or that mobility is restricted in ways yet to be discovered to specific subclades of Dehalococcoides.


Introduction
Chlorinated solvents are among the most prevalent and persistent soil and groundwater contaminants in industrialized countries (Moran et al. 2007). Tetra-or perchloroethene (PCE) and trichloroethene (TCE) contamination originates from the on-going use of dry-cleaning solvents and metal degreasing agents, respectively (Doherty 2000). These compounds, and their transformation intermediates such as vinyl chloride (VC), are known to have toxic or carcinogenic effects so their widespread soil and groundwater contamination poses a great concern to human health (Müller et al. 2004). In the late 1980s, following the discovery that microorganisms could completely dechlorinate these solvents under anaerobic conditions (Freedman and Gossett 1989), bioremediation and bioaugmentation emerged as highly successful treatment options for these problematic pollutants (Löffler and Edwards 2006). A microbial culture enriched from soil contaminated with chlorinated solvents in Southern Ontario, Canada, resulted in the KB-1™ mixed microbial consortium (Duhamel et al. 2002). KB-1™ has been used commercially for bioaugmentation for more than 20 years to remediate chlorinated compounds (Major et al. 2002, Löffler and. In the KB-1™ culture, chlorinated ethenes, such as PCE and TCE, are sequentially dechlorinated via cis-dichloroethene (cDCE) and VC to ethene (Duhamel et al. 2002, Perez-de-Mora et al. 2017, Molenda et al. 2020. Reduc-tive dechlorination in KB-1™ is primarily performed by multiple Dehalococcoides mccartyi strains in a growth-linked process called organohalide respiration (Duhamel et al. 2002, Perez-de-Mora et al. 2017, Molenda et al. 2020. Based on 16S rRNA sequence similarity, all D. mccartyi strains are categorized into three phylogenetic clades or subgroups: Pinellas, Victoria, and Cornell. The Pinellas clade is represented by strain CBDB1, the Victoria clade is represented by strain VS, and the Cornell clade is represented by the type strain, strain 195 (Hendrickson et al. 2002, Löffler et al. 2013. Lab-grown KB-1 cultures enriched on different chlorinated substrates select for different strains of D. mccartyi that express different reductive dehalogenases (Perez-de-Mora et al. 2017). Reductive dehalogenases belong to a broad protein family (PF13486) and have been classified into Ortholog Groups (OGs) on the basis of > 90% amino acid pairwise identity (Hug et al. 2013, Molenda et al. 2020). In the TCE-enriched KB-1 culture, strains belonging to the Pinellas clade are dominant, expressing the VC reductases VcrA (OG 8) and BvcA (OG 28; Molenda et al. 2020). Recently, the TCEenriched KB-1 culture was also found to contain low abundance Cornell clade strains, which express the TCE reductase TceA (OG 5; Molenda et al. 2020). VC reductases are the most critical since VC is the most toxic dechlorination intermediate and following VC dechlorination, complete dechlorination is achieved.
Interestingly, the operon which encodes VcrA was found on a mobile genetic element, called the vcrA-Genomic Island (GI; McMurdie et al. 2011). In the KB-1 TCE-enrichment culture, the vcrA-GI was identified in a circularized and extrachromosomal state within the cell, which could theoretically be transferred between D. mccartyi strains through horizontal gene transfer (HGT; Regeard et al. 2005, McMurdie et al. 2011, Molenda et al. 2019. However, there is no direct evidence of HGT of the vcrA-GI, and the mechanism of transfer between D. mccartyi remains unknown. To try to induce HGT of the vcrA-GI, we blended the KB-1 TCEenrichment culture with another mixed culture, called Donna II (Fennell et al. 1997). We called the blend of the two enrichment cultures DKB (Donna + KB-1). The KB-1 TCE-enriched consortium is the ideal vcrA-GI donor because the vcrA-containing D. mccartyi strain is highly abundant in this culture. The Donna II culture is a mixed microbial consortium that contains only one D. mccartyi strain, strain 195, that dechlorinates PCE to VC via organohalide respiration utilizing the PCE dehalogenase PceA (OG 30) and TceA (Fennell et al. 1997). Subsequently, VC is dechlorinated to ethene slowly via cometabolism since D. mccartyi strain 195 does not contain the vcrA-GI or any other VC reductases and, therefore, cannot grow on VC (Regeard et al. 2005). Of particular interest for experimental design, D. mccartyi strain 195 is capable of nitrogen fixation in the absence of available nitrogen sources via the nif operon (Lee et al. 2009). Whereas the vcrA-GI-containing D. mccartyi strain from TCE-enriched KB-1 does not have a nitrogen fixation operon, and thus an incapability for nitrogen fixation. In a theoretical HGT event, the KB-1 TCE-enrichment culture would contain the donor strain of the vcrA-GI and D. mccartyi strain 195 in Donna II would be the recipient strain. By applying the selective pressure of providing only VC as the sole energy source (electron acceptor) in medium without ammonium, we thought that a hybrid D. mccartyi strain which could fix nitrogen and respire VC would emerge.
After 4 years of observation, we found no evidence to support HGT of the vcrA-GI. Instead, in two of three replicates of the DKB culture, we observed VC dechlorination activity even though vcrA gene copies were low and tceA and D. mccartyi 16S rRNA copies were high. We then sequenced and modelled the TceA of the DKB culture and found a mutation in the predicted active site, i.e. hypothesized to influence substrate specificity. Additionally, over the course of the study, two previously unknown Cornell strains of D. mccartyi were identified in KB-1 that contain nitrogen-fixing genes. We determined that these genes were expressed and active at low ammonium concentrations, a beneficial feature of the KB-1 culture for bioremediation.

Enrichment cultures
The KB-1 culture originated from microcosms prepared with aquifer materials from a TCE-contaminated site in southern Ontario in 1996 (Duhamel et al. 2002). The KB-1 TCE-enrichment culture is maintained biweekly with 0.76 mM TCE as the electron acceptor and 5 × electron equivalents of methanol (MeOH) as the electron donor, referred to as KB-1/TCE-MeOH, as previously described (Duhamel et al. 2004, Duhamel and Edwards 2006. The Donna II culture originated from an enrichment culture seeded with digester sludge from a wastewater treatment plant in Ithaca, NY, United States. The Donna II culture was maintained batch-style at Cornell University with 0.11 mM of PCE as the electron acceptor and butyrate as the electron donor, as previously described (Smatlak et al. 1996, Fennell et al. 1997). The 'DKB' cul-ture was formed at Cornell University in 2012 when 1000 ml of KB-1/TCE-MeOH and 700 ml of Donna II were combined. The DKB culture was maintained with PCE and butyrate, similar to the Donna II culture. In 2015, the DKB culture was shipped to the University of Toronto and was subsequently used as inoculum to create DKB subcultures grown under different conditions of varying selective pressures, as described below.

Experimental set-up and monitoring
DKB subcultures were fed either PCE or VC as electron acceptor, and butyrate as electron donor. Additionally, DKB subcultures were either cultured in medium with ammonium ( ), or without ammonium ( ) requiring nitrogen fixation. Therefore, four experimental conditions were prepared: (i) DKB PCE , (ii) DKB PCE , (iii) DKB VC , and (iv) DKB VC ( Figure S1, Supporting Information). Additionally, a KB-1 control culture, KB-1/TCE-MeOH, was used as inoculum for subculture triplicates into medium without ammonium and amended with VC, referred to as KB-1 VC ; this condition was expected to be a negative control for KB-1 growth under VC-degrading, ammonium-limiting conditions (Figure S1, Supporting Information). To create each subculture, each 2 ml sample of culture (KB-1/TCE-MeOH or DKB) was centrifuged at 13 000 × g for 15 minutes at room temperature, the supernatant was discarded, and the pellet was resuspended by flicking and pipetting in 2 ml of anaerobic, autoclaved distilled H 2 O. This process was repeated three times, to wash ammonium from the pellet. Following, the 2 ml volume of washed inoculum was transferred into 198 ml of anaerobic mineral medium (Duhamel and Edwards 2007) with (10 mM) or without NH 4 Cl in 250 ml serum bottles sealed with butyl rubber stoppers (Geo-Microbial Technologies Inc.). An additional 500 μl of vitamin stock (Edwards and Grbic-Galic 1994) was added to each bottle during set-up, and 50 μl was added for each 20 ml of medium added during medium amendments. The culture was purged with N 2 : CO 2 (80:20) gas, and the headspace was over-pressurized with N 2 :CO 2 for 3 seconds. The DKB subcultures were amended periodically with 310 μelectron equivalents (μeeq) of gaseous VC (5 ml) or neat PCE (4 μl) as electron acceptor, and 4 × eeq sodium butyrate stock as electron donor. Electron equivalents were used to calculate the mass of electron donor or acceptor fed to the cultures as a way to establish a ratio of electrons required for organohalide respiration given that each hydrogenolysis reaction requires 2 electrons for the removal of a chlorine atom. Therefore, the KB-1/TCE-MeOH subcultures, KB-1 VC , were maintained with 310 μeeq VC as electron acceptor, and 5 × eeq of MeOH and ethanol mixture (50:50 on an eeq basis) as electron donor. Cumulative electron acceptor consumed was monitored over time. Acceptor and donor were reamended as needed when depleted in each subculture, and medium was amended following large volume sampling for nucleic acid extraction in batch style maintenance. Triplicate bottles were prepared for each of the four test conditions. In 2016, to reduce number of bottles monitored, only one of the triplicate cultures from the DKB subcultures with ammonium was maintained; prior to this time, these triplicates were behaving similarly. Therefore, from 2016 onward, eight DKB subcultures and three KB-1 VC cultures were maintained and analyzed.

DNA extraction
Samples were taken from KB-1/TCE-MeOH, Donna II, DKB, and DKB subcultures at various times throughout the experiment for DNA extraction. For each DNA sample, a volume of 20 ml of liquid culture was removed anaerobically. The 20 ml aliquot was cen-trifuged in a sealed Falcon conical centrifuge tube at 6870 × g for 20 minutes at room temperature, using a swinging-bucket rotor. Aerobically, the supernatant was decanted, and the pellet was resuspended immediately using 60 μl of Solution C1 and 500 μl of liquid from the PowerBead tube, from the DNeasy PowerSoil Kit (Qiagen). All subsequent kit instructions were followed. Final DNA elution volume was 50 μl of Solution C6 (elution buffer). DNA was quantified by NanoDrop or using the High-Sensitivity DNA kit with the Qubit® 3.0 Fluorometer (ThermoFisher Scientific). All DNA extracts were stored at −80 • C.

RNA extraction and cDNA synthesis
RNA was extracted from 10 ml of the KB-1 VC cultures and the KB-1/TCE-MeOH culture. Cells were pelleted by centrifuging at 5000 × g at 4 • C for 15 minutes. Pellets were then resuspended in 500 μl of supernatant and stabilized with 1 ml of RNAprotect Bacteria Reagent (Qiagen). The RNeasy® Protect Bacteria Mini Kit (Qiagen) was used to extract RNA from the cells by physical disruption via bead-beating. The kit procedure was followed, and RNA was eluted in 30 μl of RNase-free water. Following, RNA was DNase-treated using 10 μl of DNase A and 70 μl of DNase buffer for 30 seconds, then checked for the absence of DNA. To purify, the RNA was cleaned up using the RNeasy® MinElute® Cleanup kit (Qiagen) and eluted in 15 μl of RNase-free water. Following, cDNA was synthesized from the RNA using the Invitrogen Superscript® VILO cDNA synthesis kit (Invitrogen), and procedure was followed according to kit instructions. cDNA was quantified using the High-Sensitivity DNA kit with the Qubit® 3.0 Fluorometer. All cDNA extracts were stored at −80 • C.

Dehalococcoides mccartyi strain biomarker selection for quantitative PCR of 16S rRNA and functional genes
To track the growth of the different D. mccartyi strains in the DKB subcultures, we designed strain-specific biomarkers for quantitative PCR (qPCR). The abundance of all D. mccartyi strains was quantified by the 16S rRNA gene, using primers Dhc1f and Dhc264r (Hendrickson et al. 2002), since there is a single 16S rRNA gene copy per genome. To quantify the vcrA-containing strain from KB-1/TCE-MeOH and all extrachromosomal vcrA-GIs, we used primers vcrA670f and vcrA440r targeting the vcrA gene (Molenda et al. 2016). The tceA500f and tceA795r (Fung et al. 2007) primers were used for tracking tceA-containing D. mccartyi strains in KB-1/TCE-MeOH and D. mccartyi strain 195. To quantify strain 195, we designed primers for a unique rdhA (DET_RS00960), i.e. truncated (truncated rdhA, 'trdhA') and predicted to encode a nonfunctional reductive dehalogenase. It is only 1200-bp long compared to functional rdhA with lengths of approximately 1500 bp. As well, this gene contains two iron-sulfur cluster binding domains (CX 2 CX 2 CX 3 CP) 2 but does not contain commonly conserved motifs: a twin-arginine TAT membrane export sequence (RRXFXK) nor a cobalamin binding domain. Primers for trdhA were designed for the D. mccartyi strain 195 genome, using Primer2 (Untergasser et al. 2012) in Geneious 8.1.9 (Kearse et al. 2012). To track nitrogen fixation (nif) genes, we quantified the nifD gene, which encodes the nitrogenase molybdenum-iron protein α-chain. The nifD primers used in this study were used to characterize nitrogen fixation in D. mccartyi strain 195 (Lee et al. 2009). These nifD primers were also used for transcription analysis of nitrogen fixation genes in KB-1 VC and KB-1/TCE-MeOH, by reverse transcription (RT)-qPCR. Lastly, we went back to archived DNA samples to quantify the vinyl chloride reductase , bvcA ( Figure S2, Supporting Informa-tion), using primers bvcA318f and bvcA555r (Waller et al. 2012). More information on primers can be found in (Table S1, Supporting Information).
As a qPCR standard for absolute quantification, a plasmid with concatenated sequences corresponding to the D. mccartyi 16S rRNA gene, vcrA, tceA, and bvcA was used, as previously described (Molenda et al. 2019). This concatenated plasmid allowed us to calculate accurate ratios of these rdhA genes to 16S rRNA gene copies. For nifD and trdhA qPCR standards, regions were PCR amplified, purified and cloned into Escherichia coli (TOPO™ TA Cloning™ Kit for Sequencing, Invitrogen; additional methods in the Supplemental Information).
All qPCRs were prepared in a UV PCR cabinet and each qPCR was run in duplicate or triplicate, using a CFX96 real-time PCR detection system. Each 20 μl qPCR reaction was prepared in UV-treated UltraPure nuclease-free water containing 10 μl of EvaGreen® Supermix (Bio-Rad Laboratories, Hercules, CA), 0.5 μl of each primer (forward and reverse, from 10 μM stock solutions), and 2 μl of DNA template or standard plasmid dilution series, from 10 1 to 10 7 copies of plasmid per μl. Thermocycler program: initial denaturation at 95 • C for 2 minutes, followed by 40 cycles of denaturation at 98 • C for 5 seconds, varied annealing temperatures (Table S1, Supporting Information), followed by extension for 10 seconds at 72 • C. All qPCR standards and quality metrics (efficiency, standard curve details, and so on) can be found in Table S2 (Supporting Information), with calculations of absolute gene abundances per ml of culture (Table S3, Supporting Information). The quantification limit (corresponding to lowest calibration standard) was ∼1 × 10 3 copies/ml, but values as low as 1 × 10 1 copies/ml could be detected. To compare nifD expression among different experimental conditions, the absolute nifD gene abundances of genomic DNA and cDNA were determined. Using this data, we calculated nifD transcript per gene (TPG) ratios, as the ratio of transcript copies per ml of culture to gene copies per ml of culture.

Clone library preparation and sequencing of tceA
A clone library of tceA was generated from one of the DKB VC cultures. As control, a tceA clone library was generated in parallel using the KB-1/TCE-MeOH culture. The two tceA clone libraries were generated using the TOPO™ XL-2 Complete PCR Cloning Kit (Invitrogen). PCR primers were designed to amplify the 2200 bp tceAB region (Table S1, Supporting Information). Blunt end tceA PCR products were produced using the Platinum SuperFi polymerase, bands were extracted from an agarose gel, and ligated into pUC19 and transformed into One Shot™ OmniMAX™ 2 T1R Chemically Competent E. coli cells. A volume of 50 or 100 μl of transformed E. coli were spread on 50 μg/ml kanamycin LB plates, with 40 μl of 40 mg/ml X-gal in dimethylformamide (DMF) solution and 40 μl of filter-sterilized 0.1 M IPTG for blue-white selection. Following, 10 colonies were selected each from the DKB subculture and from the KB-1/TCE-MeOH culture and transferred cultures onto a patch plate for further analysis. From the patch plate, colonies were incubated in to 10 ml of 50 μg/ml kanamycin LB broth, and plasmids were extracted using a QIAprep Spin Miniprep Kit (Qiagen). Each plasmid was PCR amplified using the T3 and T7 cloning primers to confirm successful transformation. For the first round of Sanger sequencing, plasmids were sequenced using the T3 and T7 cloning primers, at the SickKids Center for Applied Genomics (TCAG) sequencing/synthesis facility (Toronto, Canada). A total of 10 sequences each from the forward and reverse were then aligned, and two consensus se-quences were generated at a 99% nucleotide identity threshold. Using the consensus sequences, primers were designed (Table S1, Supporting Information) for the remaining 845 bp. For the second round of sequencing, the final 99% nucleotide identical consensus sequences were generated by aligning these 20 sequences. The cloned tceA sequences from this study, the tceA published on NCBI (AAW39060), and the tceA from the Donna II metagenome (IMG-M taxon ID: 2032320001) were aligned using the MUSCLE aligner (Edgar 2004) in Geneious 8.1.9.

TceA protein modelling
Protein models were independently produced for TceA from the D. mccartyi 195 isolate and the TceA from one of the DKB VC cultures. Each sequence was first trimmed to remove the TAT signal peptide sequence predicted by SignalP-5.0 (Almagro Armenteros et al. 2019). The models were produced from four different predictive modelling servers, AlphaFold2, Robetta, I-TASSER, and Phyre2, and assessed based on quality scores determined by the MolProbity assessment tool (Chen et al. 2010, Roy et al. 2010, Song et al. 2013, Kelley et al. 2015, Studer et al. 2020, Jumper et al. 2021. The AlphaFold2 and Robetta models had the highest quality scores and were used for further analysis; however, all models were used to validate the predicted spatial locations of the mutated residues. The cobalamin cofactor was modelled into the active site using AutoDock Vina which was consistent with the docking position of known crystal structures (Trott and Olson 2009). The iron-sulfur clusters were superimposed from the Sulfurospirillum multivorans PceA crystal structure (PDB ID: 4UR2; Bommer et al. 2014). The substrate access channels were predicted using the CAVER 3.0 plugin in PyMol v2.3.4, only the channels predicted on the catalytic face of the cobalamin cofactor were kept (Chovancova et al. 2012). Models were imaged and the molecular interactions were predicted using PyMol v2.3.4 (The PyMOL Molecular Graphics System, Version 2.3.4 Schrödinger).

Long-term growth of DKB subcultures under varying degrees of selective pressure
To analyze the effects of experimental conditions on growth and dechlorination, we monitored the cumulative electron acceptor consumed over 4 years (Table S3, Supporting Information). Overall, cultures grown without ammonium ( ) did not consume as much electron acceptor as cultures with ammonium ( ) in each series (Fig. 1A). The DKB PCE culture consumed twice as much PCE as the DKB PCE cultures, while the DKB VC culture consumed nearly six times more VC than the DKB VC cultures over the same time. The DKB VC cultures, grown under the highest selective pressure conditions of VC without ammonium, dechlorinated the least amount of electron acceptor of all DKB subcultures (Fig. 1A). These findings confirmed that the selective pressures on the DKB subcultures impacted the metabolism of D. mccartyi to varying degrees.
Trends in the abundance of key strain biomarkers helped decipher what occurred in the DKB (blend of KB-1 and Donna II) and KB-1 control cultures over nearly 4 years of incubation. The KB-1/vcrA strain was highly abundant (10 7 -10 9 copies/ml of culture) in all cultures except the DKB VC (no ammonium) replicate cultures ( Fig. 1D-F). In DKB VC 1 and 2 cultures ( Fig. 1D and E), the KB-1/vcrA strain decreased in abundance over time, whereas in the DKB VC 3 culture, KB-1/vcrA abundance was maintained (Fig. 1F). Conversely, an increase in the 195/trdhA strain was observed in DKB VC 1 and 2 cultures, whereas it was near the quantification limit of 10 3 copies/ml in the DKB VC 3 culture throughout the experiment (Fig. 1D-F). All together, these results suggest that the dominant strain in DKB VC 1 and 2 cultures was D. mccartyi strain 195 from the Donna II culture, and the dominant strain in the DKB VC 3 culture was D. mccartyi strain KBTCE1 from the KB-1 culture ( Figure S3, Supporting Information). An interesting and complementary result is that the nondominant D. mccartyi strains, such as KB-1/bvcA and KB-1/tceA, persisted in low abundances throughout the 4 years of incubation. If the vcrA island were to be transferred to strain 195, the 'hybrid' D. mccartyi strain would have both vcrA and trdhA, and thus would appear in Fig. 1(D)-(F) as an equal abundance of KB-1/vcrA and 195/trdhA, yet this result was never observed. Therefore, these results refute the existence of a predicted hybrid strain that acquired the vcrA-GI.
However, early in the experiment, we noticed that KB-1 VC cultures were able to dechlorinate VC even without ammonium provided in the growth medium ( Fig. 1A; Table S4, Supporting Information). This condition was intended to be a control where dechlorination and growth were not anticipated. As dechlorination continued, it was suspected that the KB-1 VC culture may contain nitrogen-fixing bacteria, later justified by the high abundance of D. mccartyi nifD copies/ml of culture (Fig. 1F). We, therefore, analyzed nitrogen fixation activity in the KB-1/TCE-MeOH and KB-1 VC cultures.

Complete nitrogenase operons identified in two KB-1 strains
In 2012, when this experiment was conceived, the only D. mccartyi strain that was known to fix nitrogen was the D. mccartyi strain 195 isolate (Lee et al. 2009). After the DKB experiment had begun, metagenomic sequencing of the KB-1/TCE-MeOH culture revealed the presence of more than three strains of D. mccartyi (Molenda et al. 2019). The vcrA-GI containing strain, KBTCE1, was the most abundant strain and belonged to the Pinellas clade. However, the other two strains, KBTCE2 and KBTCE3, were both tceA-containing Cornell strains that were not previously detected in the culture. Genome analysis revealed that strains KBTCE2 and KTBCE3 contained complete nitrogenase operons, encoded by nif genes, and   (Lee et al. 2009) and mixed PW4 culture (Kaya et al. 2019), grown with ( ) and without ( ) ammonium.
were, therefore, predicted to fix nitrogen. The nitrogenase operons in strains KBTCE2 and KBTCE3 share 99.0% nucleotide pairwise identity with the Cornell strain 195 isolate nitrogenase operon (DET_RS05950-RS05990; Fig. 2A; Table S8, Supporting Information). The KBTCE2 and KBTCE3 nitrogenase operons were almost identical to each other, with one single nucleotide polymorphism (SNP) in the nifK gene ( Fig. 2A), which is the α-subunit of the iron-molybdenum protein in the nitrogenase complex (Raymond et al. 2004).
To determine if KBTCE2 and KBTCE3 strains could utilize the nif operon to fix nitrogen, we quantified the nifD gene copies and transcripts in the KB-1 VC cultures and the KB-1/TCE-MeOH culture, referred to here as KB-1 TCE . In the KB-1 VC cultures, the absolute abundance of nifD transcripts was 2.3 (± 1.7) times more than the KB-1 TCE culture, even though the absolute abundance of the nifD gene was 24 (± 9.6) times less (Fig. 2B). Therefore, D. mccartyi strains KBTCE2 and KBTCE3 in the KB-1 VC cultures were actively transcribing the nif operon indicative of nitrogen fixation. However, the KBTCE2 and KBTCE3 strains are tceAcontaining strains and do not contain vcrA, so these strains are not able to obtain energy from VC in these experimental conditions. This is different from the KB-1 TCE culture in which strains KBTCE2 and KBTCE3 can obtain energy through the dechlorination of TCE. Therefore, it is a conundrum as to how these strains obtain energy when grown in the presence of VC as the sole electron acceptor in the KB-1 VC cultures, especially since nitrogen fixation is an energetically expensive process (Leigh and Dodsworth 2007). This finding might suggest syntrophic relationships between D. mccartyi strains, in the exchange of energy for available nitrogen sources. Alternatively, this activity may point towards the function of the constitutively expressed reductive dehalogenase (OG 15) observed under 'starvation' conditions in D. mccartyi strain 195 (DET_RS07915; Johnson et al. 2008, Rahm andRichardson 2008) and in KB-1 (DQ177510; Waller et al. 2012, Liang et al. 2015. It may be that RdhA from OG 15 can sustain the population with an unknown electron acceptor in what is perceived as 'starvation'. Further work is required to determine if a reductive dehalogenase from OG 15 is expressed in the KB-1 VC cultures and its role in providing energy to D. mccartyi strains. The nifD transcripts-per-gene (TPG) ratio was calculated for the KB-1 VC cultures and for KB-1 TCE (Fig. 2C). The TPG ratios of KB-1 cultures were comparable to approximate TPG ratios of the D. mccartyi strain 195 isolate (Lee et al. 2009) and the PW4 groundwater aquifer-derived enrichment culture dominated by Cornell strains (Kaya et al. 2019;Fig. 2C; Table S6, Supporting Information). Therefore, we identified two more strains of D. mccartyi that can fix nitrogen, both originating from the KB-1/TCE-MeOH culture. Dehalococcoides mccartyi strains KBTCE2, KBTCE3, and 195 are the only strains currently known to fix nitrogen, all belonging to the Cornell clade (Molenda et al. 2020). Therefore, the nitrogen fixation characteristic of D. mccartyi appears to be specific to the Cornell clade.

Dehalococcoides mccartyi strain 195 TceA predicted VC-dechlorinating activity
Dehalococcoides mccartyi strain 195 in the Donna II culture contains tceA and D. mccartyi strains KBTCE2 and KBTCE3 in the KB-1/TCE-MeOH also contain tceA (Molenda et al. 2020). The tceA gene encodes the TCE reductive dehalogenase, TceA, which catalyzes the reductive dechlorination of TCE to cDCE and cDCE to VC (Tang et al. 2013). TceA can also dechlorinate 1,2-dichloroethane (1,2-DCA) to ethene and trace amounts of VC (Duhamel and Edwards 2007). However, it was recently discovered that TceA can dechlorinate VC to ethene coupled to organohalide respiration in the presence of sufficient vitamin B 12 (Yan et al. 2021). Ethene formation occurred when vitamin B 12 concentrations were 10 μg/l or greater and dechlorination rates were positively correlated to B 12 concentrations (Yan et al. 2021). In the DKB experiment, vitamin B 12 , in the form of cyanocobalamin, was maintained at a concentration of 6 μg/l in the culture medium, which would suggest that VC would not be consumed. However, our data revealed slow VC dechlorination where ethene was formed at a rate of 0.88 (± 0.58) μmol/day in the DKB VC 1 culture ( Figure S4, Supporting Information), which was comparable to the rate of Clreleased when vitamin B 12 concentrations were 10 μg/l (Yan et al. 2021). Furthermore, the observed increase in tceA and D. mccartyi with the simultaneous decrease in vcrA (Fig. 1D-F) suggested that D. mccartyi were able to sustain growth on VC dechlorination with tceA with 6 μg/l of vitamin B 12 ( Figure S4, Supporting Information).
To determine if the active TceA was from the Donna II or KB-1 culture, we generated a tceA clone library from the DKB VC 1 culture (Table S7, Supporting Information) and the KB-1/TCE-MeOH culture as a control. The cloned tceA sequence from the KB-1/TCE-MeOH culture was identical (100% nucleotide pairwise identity) to the tceA in published KBTCE2 and KBTCE3 genomes ( Figure  S5A, Supporting Information). The cloned tceA sequence from the DKB VC 1 culture was 100% identical to the tceA in D. mccartyi strain 195 from the Donna II metagenome sequenced in 2010 (Figure S5B, Supporting Information), but not the original sequence of isolated strain 195, sequenced in 2005. Therefore, the observed VC-dechlorinating activity was not the result of a mutation during this experiment. TceA from the Donna II metagenome was aligned to TceA from the isolated D. mccartyi strain 195 and other TceA from OG 5 ( Figure S6, Supporting Information). In this alignment, two nonsynonymous mutations were identified between the strain 195 TceA and Donna II TceA: V261A and I481T. Interestingly, the I481T mutation was also observed in the recently sequenced TceA of strain FL2, which was shown to have higher rates of VC respiratory activity compared to strain 195, even at high vitamin B 12 concentrations (Yan et al. 2021).
To explore whether these two mutations could have an impact on the substrate specificity and activity of the protein, we generated models of the TceA structure using the sequence from D. mccartyi strain 195 isolate and from the Donna II metagenome. Models were generated independently through several prediction servers to increase confidence in the predicted locations of each mutation (Table S9, Supporting Information), with the highest scoring models produced by the Robetta and AlphaFold2 servers (Song et al. 2013, Jumper et al. 2021). In each model the I481T mutation was consistently near the active site in both the D. mccartyi strain 195 isolate and the Donna II strain 195 TceA models (Fig. 3B). Whereas the location of the V261A mutation had some variability but was always predicted to be removed from the active site. It is unlikely that the V261A mutation plays a role in altering the enzymatic activity (Fig. 3A). Visualization of the predicted I481T location led to two potential hypotheses that could contribute to enhanced VC respiration: alteration of the substrate access channel, and hydrogen bonding between the residue and an [4Fe-4S] cluster to shift its reduction potential. These are described below.
The crystal structure of PceA, a PCE reductive dehalogenase, from the organohalide-respiring bacterium S. multivorans, depicts a narrow substrate channel leading to the buried hydrophobic binding pocket (Bommer et al. 2014). To visualize where the sub-  strate may enter the active site in our models, we predicted the structure all of the tunnels with access to the catalytic face of the cobalamin cofactor using Caver 3.0 (Chovancova et al. 2012). The models predicted that the I481T mutation is situated at the mouth of the major substrate access channel in each model (Fig. 3B). Access channels can have a drastic impact on the substrate scope of an enzyme as they act as a filter for the active site. It has been established in directed evolution research that mutations in the substrate channel can lead to a change in enzyme activity (Kokkonen et al. 2019). Thus, it is speculated that the I481T mutation in TceA could alter substrate preferences. The I481T mutation is also situated in a position where it can form interactions with one of the [4Fe-4S] clusters. While the level of influence that the [4Fe-4S] clusters and their redox potential have on RDase activity is unstudied, it is well-known that the environment surrounding the clusters affects their redox potential (Langen et al. 1992, Birrell et al. 2016. The presence of hydrogen-bonding residues has been suggested to have an impact on the redox potential though there are other factors at play and a direct correlation with hydrogen bonding is not clear (Stephens et al. 1996). The I481T mutation introduces the potential for a hydrogen bonding interaction and is predicted to be in close proximity (2.8Å) of the [4Fe-4S] binding position (Fig. 3C).
As the redox potential provides a driving force for respiration and reduction of the electron acceptor, changes to the environment around the [4Fe-4S] clusters should be considered as a possible mechanism of fine-tuning the RDase activity towards a certain substrate.
Predictive modelling is a powerful tool in gaining insight as to the potential impact of these mutations; however, biochemical characterization and mutagenesis experiments need to be done to unequivocally determine the substrate range of the Donna II D. mccartyi strain 195 TceA. Furthermore, it would be beneficial to determine if mutations in TceA enhance VC respiration when combined with vitamin B 12 supplementation, similarly to strain FL2 (Yan et al. 2021), or if TceA mutants can completely respire VC with lower concentrations of vitamin B 12 .

Implications for HGT of the vcrA-GI
One objective of this experiment was to try to promote HGT of the vcrA-GI from D. mccartyi strain KBTCE1 into strain 195. As previously mentioned, HGT would be manifested by high/equal abundance of both KB-1/vcrA and 195/tceA biomarkers with VC dechlorination (Table 1). These criteria were not met in any DKB subculture. To further confirm that the vcrA-GI had not integrated into the D. mccartyi strain 195 genome, we PCR-amplified the in-tegration site at the ssrA gene locus (McMurdie et al. 2011). We did not observe vcrA-GI integration in any of the DKB subcultures at the ssrA locus ( Figure S7, Supporting Information). As a result, we definitively concluded that HGT of the vcrA-GI did not occur in under the experimental conditions of the DKB experiment. In the family of KB-1 cultures enriched on different substrates, eight D. mccartyi genomes were closed (Molenda et al. 2020). These genomes revealed that clade-specific, and even strain-specific, characteristics provide D. mccartyi populations with ecological advantages to persist within mixed communities and to prevent gene transfer of mobile genetic elements between dissimilar strains. With the benefit of many more genome sequences, we now have evidence that D. mccartyi clade speciation plays a role in strain compatibility for HGT. To date, there are no Cornell strains that contain the vcrA-GI, only Pinellas and Victoria clades. There are several other clade-specific characteristics, including nitrogen fixation, which is found only in the Cornell clade, as previously described. HGT between clades may also be prevented by cellular defence mechanisms such as CRISPR-Cas systems (Molenda et al. 2019(Molenda et al. , 2020. With improved appreciation of the diversity of strains even within the KB-1/TCE-MeOH culture itself, there is also no evidence of vcrA-GI HGT between the Pinellas and Cornell strains now known to exist within this culture.

Starvation and long-term survival of D. mccartyi strains
These clade-specific and strain-specific characteristics may also play a role in the observed persistence of D. mccartyi strains and their long-term survival under starvation conditions. As previously mentioned, this finding may point towards RdhA of unknown function, such as the starvation RdhA (OG 15), which has not been functionally characterized but commonly observed in transcriptomic analyses (Johnson et al. 2008, Rahm and Richardson 2008, Waller et al. 2012, Tang et al. 2013. Furthermore, the long-term survival of D. mccartyi strains may explain why at least eight strains of D. mccartyi were identified in the KB-1 culture that has been maintained in the laboratory for more than 20 years (Molenda et al. 2020). The observations of the DKB experiment provide insights into the growth and survival of D. mccartyi in environments with low flow rates or where cells are attached to a surface, such as in the natural environment, that may also experience long-term starvation conditions. Organohalogens naturally exist in the environment, typically at low concentrations, and were the natural substrates for organohalide respiring bacteria before exposure to high concentrations of anthropogenic sources of organohalogens (Field 2016). It is likely that D. mccartyi have always been capable of long-term survival in the absence of organohalogens and these conditions have promoted strain variation among D. mccartyi communities.
While this experiment started off as a simple attempt to promote HGT, the breadth and depth of knowledge of D. mccartyi has developed immensely since the experiment began in 2012. In the time since the DKB culture was prepared at Cornell University, D. mccartyi was renamed (Löffler et al. 2013), the number of closed genomes of D. mccartyi in NCBI quintupled from 5 to 25 (Kube et al. 2005, Sung et al. 2006, Pöritz et al. 2013, Wang et al. 2014, Molenda et al. 2016, prophages (Waller et al. 2012), mobile genetic elements (Molenda et al. 2019), and CRISPR-Cas systems (Molenda et al. 2019) were identified in KB-1, and the genomes in the KB-1/TCE-MeOH (Molenda et al. 2020) and Donna II cultures (IMG-M taxon ID: 2032320001) were sequenced. Considering these advancements, the findings of the DKB experiment contribute to this knowledge and will inform attempts to induce HGT in the future.

Supplementary data
Supplementary data are available at FEMSMC online.