Anaerobic Microbial Metabolism of Dichloroacetate

Dichloroacetate (DCA) is ubiquitous in the environment due to natural formation via biological and abiotic chlorination processes and the turnover of chlorinated organic materials (e.g., humic substances). Additional sources include DCA usage as a chemical feedstock and cancer drug and its unintentional formation during drinking water disinfection by chlorination.

commenced after a lag phase of about 3 weeks, and consumption was complete within 2 weeks (Fig. 1A). In parallel incubations, DCM-grown cultures rapidly consumed additional DCM (;150 mmol per bottle) within 1 to 2 days (Fig. 1A). DCA concentrations remained constant in heat-inactivated and no-inoculum controls. Incubations using chloride-free medium demonstrated that the consumption of 43 6 4 mmol DCA resulted in concomitant formation of 90 6 7 mmol chloride, indicating that both chlorine substituents were released during DCA catabolism by mixed culture RM (Fig. 1B). During DCA utilization, formate was transiently produced, and up to 7 6 2 mmol per bottle was observed (Fig. 1B). The terminal product was acetate, and 22 6 2 mmolabout half the amount of the added DCA (i.e., 43 6 4 mmol)-was formed (Fig. 1B). In contrast to DCM-grown cultures, DCA-fed cultures did not produce methane. Following the consumption of 2 mM DCA, the cultures were visibly turbid, with optical density at 600 nm (OD 600 ) values less than 0.1 optical density unit. Quantitative realtime PCR (qPCR) targeting the 16S rRNA gene of strain RM revealed 141 6 32-fold increases of cell abundances from (1.04 6 0.56) Â 10 6 per ml (cells introduced with the inoculum) to (1.46 6 0.23) Â 10 8 per ml following DCA consumption (Fig. 1C), demonstrating DCA catabolism by strain RM. Monochloroacetate (CH 2 Cl-COO -[MCA]) was never detected in cultures growing with DCA, and MCA could not replace DCA as a growth substrate.
During growth with DCA (538. 8 6 4.8mmol of DCA per bottle), H 2 was intermittently produced, and a maximum amount of 7.00 6 0.06 mmol of H 2 was observed (Fig. 2). H 2 was slowly consumed to a threshold concentration of 1,500 6 180 ppmv corresponding to 3.75 6 0.45 mmol H 2 per bottle (Fig. 2). H 2 was previously identified as an intermediate of DCM metabolism in culture RM, which supported growth of hydrogenotrophic methanogens (e.g., Methanospirillum spp.) and homoacetogens (e.g., Acetobacterium spp.) to produce methane and acetate, respectively (43,44). In addition to H 2 , carbon monoxide (CO) was detected as a transient intermediate during DCA metabolism and increased from 0.03 6 0.003 mmol to a maximum of 0.24 6 0.01 mmol per bottle (Fig. 2). The transient formation of H 2 and CO only occurred in live cultures amended with DCA (see Fig. S1 in the supplemental material). CO formation was not observed in DCM-grown cultures (Fig. S1).
Microbial community response to enrichment with DCA. 16S rRNA gene amplicon sequencing revealed changes in microbial community structure in response to repeated transfers with DCA as the sole energy source. In the first transfer cultures with DCA, strain RM was the dominant population, accounting for approximately 65% of all sequences ( Fig. 3; see Table S1 in the supplemental material). Following six consecutive transfers with DCA, the relative sequence abundance of amplicons representing strain RM increased to 87%, implying that strain RM is responsible for DCA degradation (Fig. 3). In DCM-grown cultures, Methanospirillum contributed about 2% to the total 16S rRNA gene amplicons (Fig. 3) and was implicated in methane formation (42,43). Following repeated transfers in mineral salts medium with DCA, Methanospirillum sequences were no longer detected (Fig. 3), consistent with the loss of methane forma- FIG 3 Microbial community structure responses to consecutive transfers of mixed culture RM with DCA as the sole energy source, as revealed by 16S rRNA gene amplicon sequencing. Taxa with relative abundances below 1% were categorized as "Others." The operational taxonomic units (OTUs) representing bacteria and archaea are reported to the lowest taxonomic rank possible. "Ca. Dichloromethanomonas elyunquensis" was the dominant population in mixed culture RM, and continuous transfers with DCA resulted in further enrichment. tion in cultures grown with DCA. Populations belonging to the genus Anaerolineae were also eliminated during repeated transfers with DCA (Fig. 3). The operational taxonomic units (OTUs) representing the genera Acetobacterium and Treponema, both known to comprise species capable of H 2 /CO 2 reductive acetogenesis, were maintained at relative abundances of around 1% and 4%, respectively (Fig. 3). OTUs representing Bacteroidales were also maintained at a relative abundance of 4 to 5% during consecutive transfers with DCA. The abundances of OTUs representing Desulfovibrio, Dethiosulfovibrionaceae, Sulfuricurvum, and Veillonellaceae all declined after repeated transfers with DCA (Fig. 3).
To functionally characterize the putative HADs of the strict anaerobe "Ca. Dichloromethanomonas elyunquensis" strain RM, the genes encoding HAD1 and HAD2 were cloned and heterologously expressed in Escherichia coli. Assays with cell extracts of the E. coli transformant carrying the had1 gene revealed the stoichiometric conversion of DCA to glyoxylate (see Fig. S2 in the supplemental material). Extracts of E. coli cells expressing HAD2 did not convert DCA to glyoxylate (Fig. S2). Similarly, cell extracts of an E. coli strain carrying the empty vector without an had gene did not catalyze the conversion of DCA to glyoxylate (Fig. S2). Based on these findings, we concluded that HAD1 was responsible for the initial attack on DCA in strain RM. After purification of the His-tagged HAD1 protein using a HisTrap Ni Sepharose column, a single protein band with a size of approximately 25 kDa was observed in SDS-PAGE (Fig. 5A), matching the expected size of the HAD1 protein (i.e., 221 amino acids with a calculated  In vitro assays demonstrated that the purified HAD1 stoichiometrically converted DCA to glyoxylate at a rate of 90 6 4.6 nkat mg 21 protein (Fig. 5B). The purified HAD1 also converted MCA to glycolate, but at an approximately 70-fold lower rate of 1.3 nkat mg 21 protein (see Fig. S3 in the supplemental material). The purified HAD1 protein did not exhibit activity toward DCM, trichloroacetate, and mono-or difluoroacetate.
Comparative proteome analysis.  Table S2 in the supplemental material. HAD1 (prokka_14346) was detected in the proteomes of DCA-and DCM-grown cells; however, the expression was greater in DCA-grown cells at both time points, with log 2 fold changes of 12.38 and 11.69. HAD2 (prokka_14344) was not detected in cultures grown with either substrate (Fig. 6), an observation consistent with the in vitro enzyme activity results (Fig. S2) and indicating that HAD2 is not involved in DCA metabolism.
Fermentation of Dichloroacetate ® (POR; prokka_27155, _20890, and _11715), did not show statistically significant abundance fold changes in DCA-grown versus DCM-grown cells (Fig. 6); however, they were all detected in the proteome, with potential roles in the transformation of 2-phospho-D-glycerate to acetyl-CoA. The presence of phosphate acetyltransferase (PTA; prokka_76015) and acetate kinase (ACK; prokka_39944) enzymes, both of which are encoded on the genome, was confirmed in the analyses of both DCA-and DCM-grown cells. Interestingly, PTA was less abundant in DCA-grown cells, while the ACK abundance was higher than that in DCM-grown cells. Although more information is required to understand the regulatory controls of these enzymes, their detection in cultures growing with DCA suggests an acetyl transfer reaction followed by a dephosphorylation reaction to convert acetyl-CoA to acetate, which was measured as a terminal product in DCA-grown cultures (Fig. 6). In addition, all of the Wood-Ljungdahl pathway (WLP) proteins encoded on the genome of strain RM were detected in the proteome of DCA-grown cells, suggesting the involvement of WLP enzymes in DCA catabolism (Fig. 6).

DISCUSSION
Previous studies focused on DCA catabolism in aerobes; however, the fate of DCA under anoxic conditions remained obscure. The anaerobic mixed culture RM could be maintained with DCA as the sole source of energy, and the molecular analyses implicated the DCM-degrading bacterium "Ca. Dichloromethanomonas elyunquensis" strain RM in DCA catabolism. The findings demonstrate that specialized anaerobes metabolize DCA and illustrate that the range of substrates strain RM can utilize is not limited to DCM.
Initial enzymatic attack on DCA. Both chlorine substituents were released during DCA degradation, and acetate was an end product (Fig. 1B). The stepwise reductive dechlorination of trichloroacetate (CCl 3 -COO -[TCA]), DCA, and MCA via reductive dechlorination is thermodynamically favorable, with Gibbs free energy changes of -171.2, -154.0, and -152.0 kJ per reaction, respectively, under standard conditions with H 2 as electron donor (48). The transformation of TCA to DCA has been observed when TCA was incubated with mouse or rat gut microflora under anoxic conditions (49). Reductive dechlorination of TCA to DCA was explicitly demonstrated in an axenic culture of Geobacter thiogenes (formerly Trichlorobacter thiogenes) strain K1; however, a cryptic sulfur-sulfide redox cycle was involved in dechlorination, and the organism apparently does not perform organohalide respiration (50,51). The genome of strain RM encodes three putative reductive dehalogenases (RDases), and two of them were expressed during growth with DCM (47). One of these RDases (prokka_14638) was detected during growth with DCA, albeit at very low abundance. The experimental efforts did not generate any evidence for reductive dechlorination. MCA, the product of a single reductive dechlorination (hydrogenolysis) reaction, was neither detected as an intermediate nor supported growth of mixed culture RM. Furthermore, the utilization of DCA as an electron acceptor in organohalide respiration should result in the formation of stoichiometric amounts of acetate; however, only about 50% of the initial amount of DCA was recovered as acetate in RM cultures. The experimental data indicate that strain RM ferments DCA to acetate, H 2 , and CO 2 and does not utilize DCA as an electron acceptor in organohalide respiration (Fig. 6).
The integrated physiologic, proteogenomic, and enzymatic studies pinpoint a novel HAD involved in converting DCA to glyoxylate in a strictly anaerobic bacterium. HADs belong to a large superfamily of hydrolases with diverse substrate specificities and catalyze the hydrolytic dehalogenation of 2-haloalkanoic acid to the corresponding 2hydroxyalkanoic acids (38,52). Many aerobic bacteria, including members of the genera Pseudomonas, Xanthobacter, and Moraxella, possess HADs that convert DCA to glyoxylate to initiate DCA metabolism and growth (34)(35)(36)38). HADs are not sensitive to O 2 and do not require any cofactors such as O 2 or glutathione for activity. In contrast, mammalian liver cells employ glutathione S-transferases (GSTs) to convert DCA to glyoxylate (31-33). GSTs play central roles for detoxification of various groups of harmful compounds, such as halogenated nitrobenzenes, arene oxides, and quinones (53). DCM dehalogenases of aerobic and facultative aerobic methylotrophic bacteria, which catalyze the conversion of DCM to formaldehyde, also belong to GSTs (54,55). The active site of GSTs is the thiol group of the glutathione cofactor, which in its reduced form performs a nucleophilic attack on nonpolar compounds containing an electrophilic carbon, nitrogen, or sulfur atom (53). GSTs strictly require glutathione as a cofactor and have been found in eukaryotes, some aerobic and facultative aerobic methylotrophic bacteria, and recently in a cyanobacterium (33), but never in strict anaerobes (53). HADs, in contrast, employ the carboxyl group of an aspartate residue in the active center to carry out an SN 2 nucleophilic attack on the a-carbon atom of the halogenated carboxylate substrate to displace a halogen atom and produce an enzyme-bound ester intermediate. The ester bond is subsequently hydrolyzed to produce the corresponding D-2-hydroxyalkanoate and regenerate the aspartate residue (56,57). Although both HADs (EC 3.8.1.2) and GSTs (EC 2.5.1.18) are capable of removing chlorine substituents from DCA, yielding the same products (i.e., glyoxylate and inorganic chloride), their reaction mechanisms are fundamentally different and belong to distinct enzyme classes.
DCA fermentation pathway. Based on 16S rRNA gene sequence analysis, "Ca. Dichloromethanomonas elyunquensis" strain RM is phylogenetically related to Syntrophobotulus glycolicus (45), an isolate capable of fermenting glyoxylate to glycolate, CO 2 , and H 2 (58,59). Based on physiological and enzymatic evidence, S. glycolicus strain FlGlyR T (DSM 8271) was proposed to metabolize glyoxylate via malyl-CoA to glycolate, CO 2 , and H 2 under anoxic conditions without an external electron acceptor (58). Genes encoding HADs were not found on the genome of S. glycolicus (60), and strain FlGlyR T could not utilize DCA as a growth substrate. The HAD-catalyzed DCA dehalogenation leads to the formation of glyoxylate, which is subsequently fermented by strain RM, generating acetate, CO 2 , and H 2 . In addition to DCA, strain RM is also able to ferment glyoxylate (see Fig. S5 in the supplemental material), consistent with the observation that glyoxylate is an intermediate of DCA metabolism. In contrast to S. glycolicus, strain RM does not possess the canonical genes for malyl-CoA lyase and malate dehydrogenase, and we never detected glycolate in culture supernatant. Instead, the comparative proteome analysis revealed high expression of glyoxylate carboligase (Gcl; prokka_21461) in DCA-grown cells (Fig. 6). This enzyme catalyzes the condensation of two molecules of glyoxylate to form tartronate semialdehyde, suggesting this C 3 compound is a pathway intermediate. The comparative proteome analysis further revealed the abundance of proteins (i.e., GlxR, Hyi, Ghr, Gck, Eno, Pyk, and POR) potentially involved in converting tartronate semialdehyde to acetyl-CoA via glycerate and pyruvate (Fig. 6). Glyoxylate metabolism through tartronate semialdehyde and the glycerate pathway was proposed previously in the oxalate-degrading anaerobic bacterium Oxalobacter formigenes based on the detection of enzymatic activities in the cellfree crude extract, specifically the activities of glyoxylate carboligase (Gcl), tartronic semialdehyde reductase (GlxR), and glycerate kinase (Gck) (61).
The genome of strain RM encodes a complete WLP, and the corresponding proteins were detected in DCA-grown cells, indicating the involvement of the WLP in DCA metabolism. Based on gene content, the tricarboxylic acid (TCA) cycle is incomplete, and the oxidation of acetyl-CoA through the TCA cycle is not possible. Half of the acetyl-CoA formed during DCA metabolism is likely oxidized to CO 2 via the reverse WLP generating reducing equivalents (i.e., electrons and protons) (Fig. 6). The WLP has also been implicated in glyoxylate metabolism in the thermophilic homoacetogenic bacterium Moorella sp. strain HUC22-1, which ferments glyoxylate to acetate and CO 2 via malyl-CoA rather than tartronate semialdehyde and glycerate (62). Anaerobic DCM metabolism also proceeds via the WLP (43,63), and all WLP proteins were highly expressed in strain RM cells during growth with DCM (47). To facilitate direct comparisons, the pathway postulated for anaerobic DCM metabolism in strain RM is shown in Fig. S6 in the supplemental material. The detection of formate and CO as intermediates during DCA metabolism ( Fig. 1 and 2) lends further support for the involvement of the WLP in DCA metabolism. CO is an obligatory intermediate of the WLP, generated by the bifunctional enzyme CO dehydrogenase/acetyl-CoA synthase (CODH/ACS) during the reduction of CO 2 (64). CO was not detected in cultures grown with DCM, which may be explained by the direction of the CODH/ACS reaction during DCA versus DCM metabolism, viz., the oxidative direction during DCA degradation versus the reductive direction during DCM catabolism. Another possible explanation is the mineralization of DCM to CO 2 and H 2 via the oxidative WLP, with a small fraction of the DCM carbon being assimilated through anabolic reactions (i.e., via the reductive route of the WLP) with CO as an intermediate (Fig. S6). Although strain RM metabolizes both DCM and DCA via the WLP, the proteomic data indicate relative higher expression levels of WLP proteins (with the exception of MetF) in DCM-grown cells. Degradation of both DCA and DCM generated H 2 , but methanogenesis only occurred in DCM-grown cultures. Repeated transfers with DCA eliminated methanogens (Fig. 3), hinting at possible toxic effects of DCA on methanogens (65). H 2 consumption in DCA-grown mixed culture RM was attributed to bacteria performing H 2 / CO 2 reductive acetogenesis, e.g., Acetobacterium and Treponema (66, 67). H 2 was eventually scavenged to 1,500 6 180 ppmv during growth on DCA (Fig. 2), which is consistent with the H 2 consumption threshold concentration range reported for H 2 /CO 2 reductive acetogenesis as the terminal electron accepting process (68,69).
"Ca. Dichloromethanomonas elyunquensis" strain RM has resisted isolation-presumably due to the requirement for a hydrogenotrophic partner population to remove H 2 (Fig. 3). Elevated H 2 partial pressures inhibit DCM (43) and DCA degradation (see Fig. S7 in the supplemental material) indicative of strict syntrophy, and strain RM relies on H 2 -scavenging populations to metabolize DCM and DCA. Based on the physiological observations and proteomic data, DCA metabolism in strain RM generates acetate, CO 2 , H 2 , chloride (Cl -), and biomass ( Fig. 1) and proceeds according to equation 1: The generated H 2 is consumed in H 2 /CO 2 reductive acetogenesis leading to acetate formation according to equation 2: Therefore, DCA catabolism in mixed culture RM proceeds according to equation 3, with half of the DCA being reduced to acetate and the other half being oxidized to CO 2 , which is consistent with the experimentally measured stoichiometry, viz., the ratio of DCA degraded versus acetate generated was 2.15 6 0.05 to 1 (Fig. 1B). Based on these observations, half of the acetate formed is directly derived from DCA, and the other half is generated via reductive acetogenesis by utilizing H 2 generated during DCA catabolism: Implications. Natural and anthropogenic processes introduce DCA into the environment, and the prevalence of had genes in the genomes of aerobic bacteria can be Chen et al. ® viewed as a consequence of ubiquitously present DCA (6)(7)(8)21). DCA formation has been reported in various terrestrial environments, as well as marine and peat bog ecosystems (7,8). Information about DCA pool sizes is only available from coniferous forest soils, which contain approximately 300 ng g 21 soil (6). Based on this information, we calculate that the global DCA amount in coniferous forest soils alone exceeds 8 Â 10 9 kg (assuming an area of 40 Â 10 6 km 2 , a surface soil depth [A horizon] of 40 cm, and a soil bulk density of 1.6 g/cm 3 ) (70). Because information about DCA fluxes in environmental systems is lacking, DCA turnover may be substantial even in the absence of measurable DCA pools. The heretofore unrecognized anaerobic DCA degradation pathway via glyoxylate likely constitutes the dominant route of DCA catabolism under electron acceptor-depleted conditions, with implications for carbon and electron flow in anoxic environments. DCA fermentation generates acetate and H 2 , both of which are central intermediates during carbon cycling and can fuel anaerobic food webs. Therefore, the abiotic and biotic formation and subsequent fermentation of DCA may be relevant processes for sustaining microbial activity in energy-depleted environments such as the deep subsurface. The findings also have bearing on the clinical use of DCA as a drug and future studies should explore if members of the gut microbiota have the ability to ferment DCA and assess the responses of the gut microbiome to DCA treatment (71,72).

MATERIALS AND METHODS
Chemicals. DCA (purity, .99.8%) and DCM (.99.95%) were purchased from Sigma-Aldrich Co. (St. Louis, MO) and Acros Organics (Fair Lawn, NJ), respectively. Gas mixtures with H 2 partial pressures of 10, 50, and 100 ppmv were purchased from Airgas (Radnor, PA), and CO gas (.99.0%) was purchased from Sigma-Aldrich Co. and used for standard curve preparation. All other chemicals used were analytical reagent grade or higher.
Microorganisms and cultivation. Mixed culture RM was derived from pristine freshwater sediment and maintained with DCM as the sole energy source for 8 years (42,45). Culture RM was routinely grown in 160-ml glass serum bottles containing 100 ml of anoxic, bicarbonate-buffered (30 mM, pH 7.3) basal salts medium reduced with 0.2 mM sulfide and 0.2 mM L-cysteine (73). The vessels were sealed with black butyl rubber stoppers (Bellco Glass, Inc., Vineland, NJ) under a headspace of N 2 /CO 2 (80/20 [vol/ vol]), with 5 to 10 ml neat DCM (78 to 156 mmol) provided as the sole electron donor prior to inoculation from a DCM-grown culture (5% [vol/vol]). Cultures that had consumed the initial dose of DCM received 1 to 2 mM DCA to examine its potential utilization as an energy source. Following the consumption of DCA, RM cultures were repeatedly transferred (3% [vol/vol]) with DCA as the sole energy source before the experiments reported herein were initiated. All culture vessels were incubated at 30°C in the dark without agitation. To quantitatively measure inorganic chloride release during DCA degradation, incubations were conducted in chloride-free medium with bromide salts substituting for chloride salts.
Syntrophobotulus glycolicus strain FlGlyR T (DSM 8271) was purchased from DSMZ-German Collection of Microorganisms and Cell Cultures GmbH (Braunschweig, Germany) and was grown in the anoxic, bicarbonate-buffered (30 mM, pH 7.3) basal salts medium as described above, with glyoxylate (5 mM) as the sole energy source. To test DCA as a potential substrate for S. glycolicus strain FlGlyR T , 2 and 5 mM DCA replaced glyoxylate in medium inoculated from a glyoxylate-grown S. glycolicus culture (3% [vol/ vol]).
DNA extraction and quantitative real-time PCR. For DNA extraction, 5 ml of culture suspension was periodically collected during a growth cycle on DCA or DCM and filtered onto 0.22-mm-pore Durapore membranes (Millipore, Cork, Ireland). DNA was extracted using the DNeasy PowerSoil DNA isolation kit (Qiagen, Hilden, Germany) following the manufacturer's protocol. 16S rRNA gene-targeted qPCR was used to monitor growth of strain RM in cultures grown with DCA or DCM. qPCR assays used primers and a probe specifically targeting the 16S rRNA gene of strain RM (45) and were conducted using an ABI ViiA7 real-time PCR system (43,45).
16S rRNA gene amplicon sequencing. To monitor the microbial community response to repeated transfers with DCA, 16S rRNA gene-based amplicon sequencing targeting the V4 region of both bacterial and archaeal 16S rRNA genes was performed following established procedures (74,75). The amplicons were sequenced on the Illumina MiSeq platform (San Diego, CA), and data analysis was performed with the QIIME v.1.9.1 software package (76). Raw sequencing reads were jointly paired, demultiplexed, and trimmed to a length of 250 bp, and chimeric reads were removed. After quality control, over 100,000 individual sequences were obtained for each library generated with DNA samples collected from consecutive transfers with DCM or DCA. Operational taxonomic units (OTUs) were picked via the default UCLUST pipeline (77) and filtered at a 0.005% threshold. Taxonomic assignments were performed using the RDP classifier trained against the Greengenes 16S rRNA gene database (version 13.8) (78). Taxonomy, relative abundance, and sequences of representative OTUs are shown in Table S1. The most abundant sequence within each taxon was chosen as the representative sequence for each OTU.
Heterologous had gene expression, purification, and in vitro activity testing. The pET-28a(1) expression vector backbone was used to clone and express had genes carrying an N-terminal His tag.
E. coli strain BL21(DE3) carrying an had expression plasmid was grown in 300 ml of Terrific Broth (Thermo Fisher) with 50 mg ml 21 kanamycin at 37°C and 150 rpm to an optical density at 600 nm (OD 600 ) of 1.0. Cultures were then induced with 1 mM isopropyl b-D-1-thiogalactopyranoside (IPTG; Thermo Fisher) and incubated overnight at room temperature at 150 rpm. Cells were collected by centrifugation at 9,000 Â g for 25 min, suspended in 25 mM Tris-HCl buffer (pH 7.5), and sonicated at 50% amplitude (Branson Sonifier 250; Branson Sonifiers, Danbury, CT) in an ice bath for 8 min with a 50% duty cycle (30 s on and 30 s off). The lysate was centrifuged at 38,000 Â g for 20 min, and the supernatant was passed through a 5-ml HisTrap Ni Sepharose column (GE Healthcare) using an ÄKTA Prime fast protein liquid chromatography (FPLC) system (GE Healthcare, Pittsburgh, PA). Proteins were eluted with 25 mM Tris-HCl buffer (pH 7.5) containing 100 mM NaCl and 300 mM imidazole. Fractions containing protein were combined, and the elution buffer was exchanged with 4 ml of 100 mM Tris-HCl (pH 8.0) by using a 10-kDa-cutoff Amicon Ultra-4 filter unit (Millipore). Protein concentrations were quantified using the Bradford assay (80), and protein purity was examined by SDS-PAGE and Coomassie blue staining. A standard protein marker (Bio-Rad, Hercules, CA) with protein sizes ranging from 10 to 250 kDa allowed size estimations. The purified HAD1 protein stock solutions (0.6 mg ml 21 ) were frozen immediately and stored at 280°C. The N-terminal hexahistidine tag was retained for all experiments.
Enzyme assays were conducted in 100 mM Tris-HCl buffer (pH 8.0) in Eppendorf tubes in a total assay volume of 0.5 ml. DCA was added at a concentration of 4 mM, and the reactions were started by adding 3 mg of HAD protein. The tubes were agitated with 150 rpm at 30°C. Aliquots of 50 ml were collected over time, acidified with 1 ml of 1 M H 2 SO 4 to quench the reaction, centrifuged at 17,000 Â g for 5 min at room temperature, and analyzed by high-performance liquid chromatography (HPLC) to determine the concentrations of DCA and glyoxylate. HAD enzyme activity was calculated based on the formation of glyoxylate: 1 nkat is the amount of enzyme that generates 1 nmol of glyoxylate per second.
Global proteomics of RM cultures grown with DCA versus DCM. Prior to proteomic analysis, cultures were consecutively passaged at least three times on the same substrate (i.e., DCA or DCM). Following the consumption of 361.0 6 19.5 mmol of DCA and 373.2 6 5.8 mmol of DCM, the respective cultures received one additional feeding of the respective substrate. Samples for proteomic analysis were collected upon the consumption of the first (time point 1 [TP1]) and the second (TP2) substrate feedings (Fig. S4). Cells grown with DCA or DCM were collected from triplicate cultures by passing 100 ml of culture suspension through Sterivex 0.22-mm-pore filter units (EMD Millipore Corporation, Billerica, MA). The outlet of a filter unit was capped, and 1.5 ml of boiling SDS lysis buffer (4% SDS [wt/ wt] in 100 mM Tris-HCl buffer, pH 8.0) was added. Following gentle agitation on a shaker with threedimensional (3D) gyratory action for 1 h at room temperature, a 3-ml plastic syringe was connected to the cartridge's inlet, the unit was inverted, and as much lysate as possible was transferred into the syringe. The filter units were rinsed once with 500 ml of fresh lysis buffer at room temperature, and the recovered volumes were combined. The cell lysates were then subjected to trichloroacetic acid precipitation followed by urea denaturation, reduction, blocking of disulfide bonds, and tryptic digestion (trypsin/protein ratio of 1:50 [wt/wt]) as described previously (81). The protein contents in crude and peptide extracts were quantified using the bicinchoninic acid (BCA) assay (Pierce Biotechnology, Waltham, MA). Peptide extracts were stored at 280°C until liquid chromatography-tandem mass spectrometry (LC-MS/ MS) analysis. Proteomics data sets from culture RM were obtained with an Orbitrap Q Exactive Plus mass spectrometer (Thermo Fisher Scientific, Waltham, MA) equipped with an electrospray ionization (ESI) source and interfaced with a Proxeon EASY-nLC 1200 system. Peptides (2 mg) from each sample were suspended in solvent A (2% acetonitrile, 98% water, 0.1% formic acid) and injected onto a 75-mm-innerdiameter microcapillary column packed with 35 cm of Kinetex C 18 resin (1.7 mm, 100 Å; Phenomenex). Peptides were separated using a 90-min gradient from 2% to 30% solvent B (80% acetonitrile, 20% water, 0.1% formic acid), followed by an increase to 40% solvent B within 10 min and a 10-min wash with 98% solvent A. The flow rate was kept at 250 nl min 21 . MS data were acquired with the Thermo Xcalibur software version 4.27.19, with a topN method, where N was capped to 15. Other scanning and spectral data collection parameters were similar to those reported previously (82). All spectral data collected in this study have been deposited in the MASSIVE and ProteomeXchange repositories with identifiers MSV000086520 and PXD022742, respectively (ftp://massive.ucsd.edu/MSV000086520/). Peptide and protein identification by database searching. MS/MS raw data files from culture RM were searched against a database of sequences annotated from the draft genome of "Ca. Dichloromethanomonas elyunquensis" (accession no. LNDB00000000), to which common contaminant proteins were appended (www.thegpm.org/crap). The MyriMatch v2.2 algorithm was used for standard database searching and was set to the same parameters described previously (82,83). Confidently identified peptides at a false-discovery rate below 1% were assembled into proteins using the IDPicker v.3.1 software (84). Every protein in the data set was identified with at least two unique peptide sequences. For label-free quantification, the MS1-level peptide precursor intensities were extracted from IDPicker with IDPQuantify, summed by protein, and then divided by the sequence length of the protein to which they matched (85). Protein abundance values were log 2 transformed and then normalized by mean central tendency analysis with Inferno RDN (https://omics.pnl.gov/software/infernordn). The Perseus software (86) was then used to filter proteins with non-zero abundance values in two out of three biological replicates in at least one growth condition and time point. After data filtering, undetected proteins (i.e., proteins with missing abundance values) were imputed with a simulated Gaussian distribution of lowabundance values to provide non-zero abundance metrics at the detection threshold. This approach enables statistical analyses across the entire data set, as is commonly done in proteome measurements (87,88). Pairwise t test comparisons were conducted to identify proteins having statistically significant abundance changes (P , 0.05) between growth conditions at each respective substrate feeding time point.
Analytical methods. DCA, glyoxylate, formate, and acetate were analyzed on an Agilent 1200 series HPLC system equipped with an Aminex HPX-87H column (Bio-Rad, Hercules, CA) operated at 30°C and a multiple-wavelength detector set to 210 nm. Operation was isocratic using 4 mM H 2 SO 4 as the eluent at a flow rate of 0.6 ml min 21 . Aqueous samples (200 ml) were acidified with 4 ml 1 M H 2 SO 4 and filtered prior to HPLC analysis. The identification of peaks was based on retention times of authentic standards, and quantification was achieved using external calibration curves. DCM was measured by manually injecting 0.1-ml headspace samples into an Agilent 7890A gas chromatograph (Santa Clara, CA) equipped with a DB-624 column (60-m length, 0.32-mm inside diameter, 1.8-mm film thickness) and a flame ionization detector as described previously (89). Chloride ions were measured with an ion chromatograph using a Dionex ICS-2100 system equipped with a 4-by 250-mm IonPac AS18 hydroxide-selective anion-exchange column (Thermo Fisher Scientific, Waltham, MA) operated at 30°C. The 10 mM KOH eluent was delivered at a flow rate of 1 ml min 21 , and an ERS 500 suppressor (4 mm) was set at a current of 57 mA. To follow the formation of H 2 and CO, 0.5 ml of culture headspace samples was injected into a Peak Performer 1 gas chromatograph coupled with a reducing compound photometer (Peak Laboratories, Mountain View, CA) with detection limits for H 2 and CO below 8 ppb by volume (ppbv).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.