Nitrate-driven anaerobic oxidation of ethane and butane by bacteria

Abstract The short-chain gaseous alkanes (ethane, propane, and butane; SCGAs) are important components of natural gas, yet their fate in environmental systems is poorly understood. Microbially mediated anaerobic oxidation of SCGAs coupled to nitrate reduction has been demonstrated for propane, but is yet to be shown for ethane or butane—despite being energetically feasible. Here we report two independent bacterial enrichments performing anaerobic ethane and butane oxidation, respectively, coupled to nitrate reduction to dinitrogen gas and ammonium. Isotopic 13C- and 15N-labelling experiments, mass and electron balance tests, and metabolite and meta-omics analyses collectively reveal that the recently described propane-oxidizing “Candidatus Alkanivorans nitratireducens” was also responsible for nitrate-dependent anaerobic oxidation of the SCGAs in both these enrichments. The complete genome of this species encodes alkylsuccinate synthase genes for the activation of ethane/butane via fumarate addition. Further substrate range tests confirm that “Ca. A. nitratireducens” is metabolically versatile, being able to degrade ethane, propane, and butane under anoxic conditions. Moreover, our study proves nitrate as an additional electron sink for ethane and butane in anaerobic environments, and for the first time demonstrates the use of the fumarate addition pathway in anaerobic ethane oxidation. These findings contribute to our understanding of microbial metabolism of SCGAs in anaerobic environments.

Although the microbiology of aerobic oxidation of SCGAs has been well studied [10], the microorganisms and metabolic pathways involved in the anaerobic oxidation of these gases have only been identified in recent years.The archaeal species "Candidatus Argoarchaeum ethanivorans" and "Candidatus Syntrophoarchaeum" oxidize ethane and butane via the formation of ethyl-or butylcoenzyme M, respectively, in syntrophic consortia with sulphatereducing bacteria (SRB) [11,12].In contrast, the deltaproteobacterial isolate Desulfosarcina aeriophaga BuS5 oxidizes propane and butane via a reaction with fumarate, generating propyland butyl-succinates (the fumarate addition pathway), coupled with the direct reduction of sulphate to sulphide [13,14].The bacterium "Candidatus Methylomirabilis oxyfera" was also shown to be able to degrade ethane and propane, although it remains unknown whether these carbon sources support continuous growth [15].Moreover, our recent study described a bacterial species "Candidatus Alkanivorans nitratireducens" belonging to the Class of Symbiobacteriia that can oxidize propane via the fumarate addition pathway coupled to the reduction of nitrate to nitrite [16].The oxidation of ethane and butane coupled to nitrate reduction is yet to be shown, but would also be thermodynamically feasible (Equations ( 1) and ( 2)) and potentially important given the prevalence of nitrate in natural environments [17,18].
Anaerobic ethane oxidation remains poorly understood, with direct evidence for this metabolic process limited to archaea [11,19].Indeed, ethane activation mediated by bacteria has not been proven, in clear contrast to the multiple discoveries of SRBmediated anaerobic propane and butane degradation [13,20,21].The fumarate addition pathway is considered the most common mechanism for anaerobic degradation of hydrocarbons including propane, butane, and various other n-alkanes ranging from C 6 (nhexane) to C 16 (n-hexadecane) [22][23][24][25].The oxidation of ethane via this mechanism is also likely to occur in the environment, given ethyl-succinate, the signature metabolite generated by ethane activation via reaction with fumarate, is frequently detected in hydrocarbon-rich environments, such as crude oil production wells, coal beds, and oilfields [26][27][28].However, physiological evidence for anaerobic ethane oxidation via the fumarate addition pathway is lacking.
In this study, we address knowledge gaps by enriching microbial consortia able to couple anaerobic ethane and butane oxidation to nitrate reduction, and characterizing the key metabolic pathways via a multi-omics approach (metagenomics, metatranscriptomics, and metaproteomics).The alkane-oxidizing population in both enrichments is the same species as the anaerobic propane-degrading bacteria "Ca. A. nitratireducens" identified previously [16], and is suggested to mediate ethane and butane oxidation via reactions with fumarate.

Enrichment cultures able to mediate nitrate-dependent anaerobic oxidation of ethane and butane
Two anaerobic bioreactors seeded with activated sludge and anaerobic digestion sludge from a wastewater treatment plant were operated for more than 1000 days.One was fed with ethane (C 2 H 6 ) and nitrate, while the other with butane (C 4 H 10 ) and nitrate.The C 2 H 6 -fed bioreactor showed simultaneous consumption of C 2 H 6 and nitrate, resulting in the production of dinitrogen gas and ammonium, along with transient accumulation of nitrite (Supplementary Fig. 1A).Similarly, nitrate consumption and ammonium production were observed in the C 4 H 10 -fed reactor (Supplementary Fig. 1B).No nitrate consumption was observed in the control incubations without the addition of either C 2 H 6 or C 4 H 10 or enrichment culture biomass (Supplementary Fig. 2), indicating that nitrate reduction (to nitrite, dinitrogen gas, and ammonium) was a biological process and coupled to the consumption of these alkanes.
Stoichiometric experiments were conducted directly in the parent C 2 H 6 -fed reactor or with subcultures from the parent C 4 H 10 -fed reactor to establish nitrogen and electron balances.The reduction of NO 3 − proceeded in two distinct phases for both C 2 H 6 -and C 4 H 10 -fed systems (Fig. 1A and B, Supplementary Fig. 3).In Phase 1, NO   4, Supplementary Fig. 6) bioreactors.These metagenome-assembled genomes (MAGs) had average nucleotide identities of 99.96% and 99.55%, and average amino acid identities (AAI) of 99.96% and 99.57% (Supplementary Fig. 7) to the "Ca. A. nitratireducens" genome previously recovered from the C 3 H 8 -fed culture [16], confirming that the three genomes likely represent the same species [29].Other dominant species include Patescibacteria and Fimbriimonadaceae in the C 2 H 6 -fed reactor, and Promineofilaceae, Phycisphaerales, and Anaerolineales in the C 4 H 10 -fed reactor (Supplementary Tables 3 and 4).Further annotation of MAGs for these bacteria suggest that they do not contain genes known to facilitate anaerobic SCGA oxidation, including genes for alkylsuccinate synthase (AssA) and alkylcoenzyme M reductase, indicating that they unlikely play a direct role in SCGA metabolism.The metabolic activity and potential of these bacteria require further investigation.

Analyses of metabolic pathways of "Ca. A. nitratireducens"
Consistent with "Ca. A. nitratireducens" originating from the C 3 H 8fed system (referred to as MAG/population P), the closed genomes of "Ca. A. nitratireducens" in C 2 H 6 -and C 4 H 10 -fed bioreactors (referred to as MAGs/population E and B) both contain three alkylsuccinate synthase catalytic subunits (Supplementary Fig. 8A), which are phylogenetically distant from other available fumarate addition enzymes in the UniProt database (Supplementary Fig. 8B).A search of the metagenome libraries confirmed "Ca. A. nitratireducens" as the only microorganism harbouring the key AssA gene.To support the role of these AssA complexes in ethane/butane oxidation, key metabolites from the active cultures were analysed by ultra-high-sensitivity triple quadrupole mass spectrometry.A mass peak (m/z: 275 > 73.1) at the retention time of 9.940 min was detected for the C 2 H 6 -fed bioreactor, corresponding to the ethyl-succinate standard (Fig. 2A).Also, a mass peak (m/z: 303.0 > 147.1) at the retention time of 12.245 min was detected for the C 4 H 10 -fed bioreactor, corresponding to the butyl-succinate standard (Fig. 2B).These findings support that ethane/butane were activated by addition of fumarate, thus generating ethyl/butyl-succinate, which is consistent with the action of AssA.MAGs E and B also harbour other key genes involved in the further degradation of ethyl/butyl-succinate, including the methylmalonyl-CoA mutase genes (mcmA) for carbon-skeleton rearrangement, the propionyl-CoA carboxylase genes (pccB) for decarboxylation, and the genes for beta-oxidation (Supplementary Data 2 and 3, Fig. 2C).The propionyl-CoA generated from beta-oxidation could enter the methylmalonyl-CoA pathway to regenerate fumarate for subsequent rounds of ethane/butane activation.The acetyl-CoA may be completely oxidized to CO 2 or used for fumarate regeneration via the oxidative tricarboxylic acid (TCA) cycle.CO 2 can also be generated by the oxidation of acetyl-CoA through the reverse Wood-Ljungdahl (WL) pathway for MAGs E and B (Supplementary Data 2 and 3, Fig. 2C), consistent with that proposed for MAG P and the sulphate-dependent propane oxidizer-D.aeriophaga BuS5 [14,16].The metatranscriptomic and metaproteomic data indicated that MAGs E and B had associated expression of the proposed fumarate addition pathway for complete ethane/butane oxidation to CO 2 after alkane additions (Supplementary Data 2 and 3, Fig. 2C).The "Ca. A. nitratireducens" dominated the transcriptome profile of both the C 2 H 6 -(61.5% of the total transcriptome reads, Supplementary Table 5) and C 4 H 10 -fed bioreactors (84.5% the total transcriptome reads, Supplementary Table 6), while the relative activities of other co-existing microbial population were substantially lower.This indicates that "Ca. A. nitratireducens" are the key drivers of anaerobic alkane oxidation in these systems.
Similar to MAG P, the MAGs E and B both encode genes encoding nitrate reductase (napAB) and cytochrome c nitrite reductases (nrfAH) required for DNRA process, which are all expressed (Supplementary Data 2 and 3, Fig. 2C).The expression of nrfAH was much higher for populations B than E in Phase 2, consistent with the significantly higher DNRA rates (P < 0.05) in the C 4 H 10fed bioreactor (0.77 ± 0.27 mmol/l/day) compared to the C 2 H 6fed bioreactor (0.11 ± 0.08 mmol/l/day).The NapAB and NrfA were also identified in protein extracts from both the C 2 H 6 -and C 4 H 10 -fed cultures (Supplementary Data 2 and 3, Fig. 2C), further supporting that populations E and B were performing DNRA in these systems.Other members of the communities also expressed genes for DNRA but at much lower levels compared to E and B MAGs (Supplementary Data 4).The closed E and B MAGs both lack nitric oxide-producing nitrite reductase (nirS/K) but encode nitric oxide reductase (norB) and nitrous oxide reductase (nosZD), which is consistent with MAG P. The norB and nosD genes were expressed and detected in the protein extracts for both the E and B populations (Supplementary Data 2 and 3, Fig. 2C), suggesting the active roles of these populations in the reduction of nitric oxide to dinitrogen gas.The phenomena that dinitrogen gas was generated without the apparent involvement of nirS/K for the dominant "Ca. A. nitratireducens" in all three systems indicates that this species may indeed utilize a novel gene or novel pathway to reduce nitrite to nitric oxide [16].However, we cannot completely rule out the possibity that other microorganims like Fimbriimonadaceae, Burkholderiales, and Promineofilum in the ethane reactor, and Phycisphaerales, Anaerolineales, and Promineofilaceae in the butane reactor, may also contribute to nitrite reduction to nitric oxide or denitrification to dinitrogen gas, given that they express nirS/K and other denitrification genes (Supplementary Data 4).To further confirm the exact metabolic pathways of "Ca. A. nitratireducens" for nitrogen and carbon transformations, pure culture isolation is likely needed.

Short-chain gaseous alkane metabolic versatility of "Ca. A. nitratireducens"
Structural modelling and molecular dynamics (MD) simulations were conducted to understand the potential functions of different AssAs in "Ca. A. nitratireducens."MAG E encodes three AssAs that are 852aa in length with differing AAI between them (90.96% and 96.60%, Supplementary Table 7).The shorter AssA genes identified in the P and B MAGs for "Ca. A. nitratireducens" were found to be due to open reading frame calling issues [30] via full length alignments of E MAG AssA genes to the AssA regions of the P and B MAGs, along with manual identification of the start and stop codon.Further analyses of the AssAs in three MAGs show that full length alignment to the conserved domain (cd01677) for pyruvate formate lyase 2 and related enzymes is only found in the 852aa AssAs [31], suggesting these AssAs are more likely to be complete.Given the overall high AssA gene similarities between MAGs (Supplementary Table 8), the three complete AssA genes in MAG E were used for structural modelling and MD simulations.
The MD results suggest that AssA1 cannot stably bind to the key substrate-fumarate (Supplementary Movie S1), while AssA2 and AssA3 can form stable binding complexes with fumarate and ethane/propane/butane (Supplementary Movie S2-S7, Fig. 3A-F, Supplementary Fig. 9).Hydrogen bonding networks were found to be critical for the SCGA and fumarate bindings (Fig. 3A-F, Supplementary Fig. 10).In addition, the putative radical sites Cys489 and Gly828 are situated at the core of AssA2/AssA3 and close to each other in all binding complexes (Fig. 3A-F).These characteristics were suggested to be important for radical transfers in glycyl radical enzymes [32,33], indicating that the radical transfer pathway may govern fumarate addition in AssA2/AssA3.
Metatranscriptomic profiles of the ethane, propane, and butane systems were mapped onto E MAG to ensure consistency of the AssA gene lengths.In support of the MD results, AssA1 are relatively lowly expressed in all systems (Fig. 3G).However, the expression levels of AssA2 and AssA3 are relatively high in all systems (Fig. 3G), suggesting these proteins are more likely responsible for SCGA activation by "Ca. A. nitratireducens."[16] (Cy3, red; targeting "Ca. A. nitratireducens") and EUBmix probe set [37] (f luorescein isothiocyanate label, green; all bacteria); "Ca. A. nitratireducens" cells appear yellow (red + green) and other bacterial cells appear green; the scale bar indicates 20 μm; the representative image was selected based on the visual assessment of >3 separate hybridization experiments; FISH was performed as detailed in our previous study [16].
To further validate if "Ca. A. nitratireducens" is indeed able to oxidize all three SCGAs, substrate range tests were conducted for the C 2 H 6 -, C 3 H 8 -, and C 4 H 10 -fed cultures.Incubation of subcultures from the C 2 H 6 -, C 3 H 8 -, and C 4 H 10 -fed bioreactors with the other two SCGAs showed obvious ethane/propane/butane oxidation coupled to nitrate reduction to dinitrogen gas and ammonium (Fig. 4A-F, Supplementary Fig. 11).These results provide evidence that "Ca. A. nitratireducens" has the metabolic versatility to oxidize the three tested SCGAs using nitrate as a terminal electron acceptor.

Implications
This study identified "Ca. A. nitratireducens" as a metabolically diverse anaerobic SCGA oxidizer able to utilize ethane, propane, and butane.In previous studies, SRB affiliated with the Desulfosarcina-Desulfococcus cluster and the archaeon Candidatus "Syntrophoarchaeum" were suggested to be only capable of oxidizing propane and butane but not ethane [12,13,20].Conversely, the archaeon Candidatus "Ethanoperedens thermophilum" could only oxidize ethane [19].Importantly, this study identified a bacterium performing anaerobic ethane oxidation, previously known for archaea only.This study also provides the first physiological evidence for the involvement of the fumarate addition pathway in anaerobic ethane oxidation, closing a key knowledge gap in our understanding of anaerobic SCGA oxidation.Furthermore, the newly discovered nitrate-dependent anaerobic ethane and butane oxidation (n-DAEO/n-DABO) indicate that nitrate is an additional electron sink for C 2 H 6 and C 4 H 10 , potentially contributing to reducing the negative impacts of C 2 H 6 and C 4 H 10 on air quality and on climate.C 2 H 6 and C 4 H 10 are recognized as indirect greenhouse gases with net global warming potentials of 10 and 7 times, respectively, that of CO 2 (100-year horizon) [34].Moreover, they also contribute to the production of hazardous substances including carbon monoxide and peroxyacetyl nitrate [35], which are significant air pollutants.This research advances our understanding of the role of microorganisms in constraining SCGA emissions by identifying another microbially mediated link between the global carbon and nitrogen cycles.Considering the widespread presence of nitrate and rising emissions of nonmethane SCGAs caused by oil and natural gas exploitation [9], "Ca. A. nitratireducens" may play an important role in global carbon and nitrogen cycling.

Bioreactor setup and operation
Activated sludge (50 ml) and anaerobic digestion sludge (100 ml) from a full-scale wastewater treatment plant (Luggage Point, Brisbane, Australia) were used as inoculum for the ethane and n-butane (hereafter butane) bioreactor enrichment.This choice of inoculum was based on previous successful enrichment of anaerobic propane degradation bacteria from this source and the small quantities of ethane and butane detected in anaerobic digestion systems [36].The incubations with ethane or butane as a sole carbon source were set up in a lab bioreactor with a volume of 1.12 and 2.3 l, respectively.An anoxic mineral medium [16] of 0.67 and 1.69 l was initially added to the ethane and butane reactor (∼1:4.5 and 1:11.3 of sludge to medium ratios, respectively), leaving a headspace of 0.3 and 0.46 l, respectively.showed simultaneous nitrate and propane/butane consumption with production of dinitrogen gas and ammonium; the same was observed for subcultures from the propane bioreactor supplemented with ethane (C) and butane (D), and the butane bioreactor provided with ethane (E) and propane (F); each test was conducted in triplicate (results of other tests were included in Supplementary Fig. 11).
The ethane/butane reactors were periodically f lushed with pure ethane/butane gas (99.99%,Coregas, Australia) to maintain the ethane/butane partial pressure in the headspace between 0.9 and 1.2 atm.A concentrated stock solution (80 g NO 3 was manually pulse-fed to the reactors to replenish NO 3 − to 20-30 mg N l −1 .The bioreactors were continuously mixed using a magnetic stirrer (IKA, Labtek, Australia) at 650 rpm and operated in a thermostatic chamber (35 ± 1 • C).Every 1-4 months, the stirrers were stopped for 24 h to allow biomass to settle, and the supernatant of 0.2-0.8l was then replaced with fresh medium.
The pH was manually adjusted to 6.8-7.5 using a 1 M anoxic HCl solution.Liquid samples (0.4-0.6 ml each) were collected periodically (2-5 samples per week) and filtered immediately using a 0.22 μm membrane filter (polyethersulfone filter, Millex, USA) for the analysis of NO 3 − , NO 2 − , and NH 4 + .A gas sample (100 μl) from the headspace was withdrawn regularly (three to five times per week) using a gas-tight syringe (1710 SLSYR, Hamilton) for the determination of C 2 H 6 and N 2 .

Batch tests for nitrogen and electron balances
Stoichiometric tests were carried out in situ for the biomass of the 1.12 l ethane parent reactor on Days 490, 522, and 559 to investigate nitrogen and electron balances.For stoichiometry determination of nitrate reduction coupled to anaerobic butane oxidation, triplicate batch tests were conducted in 650 ml glass vessels with a subsample of 500 ml biomass anaerobically transferred from the 2.3 l butane parent bioreactor.Total amounts of ethane/butane and N 2 were calculated by considering ethane/butane/N 2 in both the headspace (monitored) and liquid phase (calculated with Henry's law).Two negative control groups were set up in 600 ml bottles: (i) control groups containing only enriched cultures and nitrate (ethane/butane was removed by f lushing the bottles with pure argon gas for 20 min); (ii) abiotic control groups without enriched cultures (only synthetic medium containing ethane/butane and nitrate was provided).

Isotope labelling experiment
A 480 ml subculture from the ethane/butane bioreactor was transferred to a 600 ml glass vessel.The ethane culture was f lushed with pure C 2 H 6 for 10 min, and the 5 ml 13 C-labelled C 2 H 6 ( 13 CH 3 13 CH 2 ,99 atom % 13 C, Sigma) was injected into the headspace, followed by an introduction of 0.12 ml nitrate stock solution (40 g N l −1 ), which contained ∼1% 15 N-labelled NO 3 − (98 atom % 15 N, Sigma).The butane culture was f lushed with argon gas (99.99%,Coregas, Australia) for 20 min.Approximately 24 ml 13 C-labelled butane ( 13 CH 3 13 CH 2 13 CH 2 13 CH 3 , 99 atom% 13 C, Sigma) was injected into the headspace through the septum.Approximately 1 ml nitrate stock solution (10 g N l −1 ) containing ∼1% 15 N-labelled sodium nitrate (98 atom % 15 N, Sigma) was added to achieve a concentration of ∼20 mg N l −1 .Liquid samples were collected (two to five samples per week) and filtered through 0.22 μm filters for analyzing soluble nitrogen species and respective isotopic fractions.Gaseous samples were collected (four to seven samples in total) from the headspace using a gastight syringe (model 1710 SL SYR, Hamilton, USA) and injected into helium-f lushed exetainer vials (Labco, UK) for measuring total C 2 H 6 , C 4 H 10 , CO 2 , and N 2 in gas phases and their isotopic fractions.For the measurement of the dissolved CO 2 , ∼0.5 ml liquid samples were collected and injected into vacuum vials, followed by acidification with HCl stock solution (1 M), and settled for at least 0.5 h to achieve gas-liquid equilibrium before CO 2 quantification.

Substrate range tests for "Ca. A. nitratireducens"
To examine whether the C 2 H 6 -fed culture has the capability of oxidizing propane and butane, two batch tests were setup by mixing 200 ml culture from the C 2 H 6 -fed bioreactor with 280 ml anoxic mineral medium in 600 ml glass vessels.The two batch reactors were then f lushed with pure propane and butane gases, respectively, to remove dissolved ethane and provide propane and butane.The nitrate stock solution (10 g N l −1 ) was added to the reactors to achieve an initial concentration of ∼20 mg N l − .The batch tests were conducted in triplicate.Liquid and gas samples were collected as described above.Similarly, cultures from the parent C 3 H 8 -or C 4 H 10 -fed bioreactor were also transferred to new batch reactors and then incubated with ethane and butane, or ethane and propane.

Chemical analysis
Soluble nitrogen species (NO 3 − , NO 2 − , and NH 4 + ) and gas components including C 2 H 6 , CO 2 , and N 2 in the headspace were determined as described previously [16].The butane, 13 C-labelled butane, 13 C-labelled ethane, 13 CO 2 , 29 N 2 , and 30 N 2 in gaseous samples were quantified using a GC (7890A, Agilent, USA) coupled to a quadrupole mass spectrometer (MS, 5957C inert MSD, Agilent, USA).The GC-MS was operated as described in the supplementary text.

Functional annotation
Preliminary annotation across MAGs and unbin contigs was performed using Prokka 1.14.5 [58].Predicted protein sequences were then searched against KEGG (July 2021) using kofamscan 1.3.0[59], and the hit with an e-value <1e-10 and maximal F-measure was selected for each gene.UniRef100 [60] (March 2020) was searched against using diamond [61] v2.0.11.149 with "blastp-sensitive."The best hit with e-value <1e-5 and identity >30% was selected for each gene and mapped to the KEGG Orthology database.The eggNOG v5 [62] was searched against using emapper 2.1.5[63].Metabolic pathways were reconstructed using KEGG.Pathways identified to be >75% complete were considered as "present."Fulllength AssA genes from the P and B MAGs were identified based on blastn hits to the AssA genes from E MAG and translated using NCBI's Open Reading Frame (ORF) finder.

Metatranscriptomic sequencing and data analysis
Two distinct phases were observed for the nitrate reduction in both ethane and butane bioreactors (Fig. 1A and B).For total RNA extraction, the active enriched culture (10 ml) collected from each phase was mixed with 30 ml of RNAlater solution (Sigma-Aldrich) and left to stand for 1 h before extraction.Total RNA was then extracted and sequenced as described previously [16].
The metatranscriptomic paired-end reads were mapped to dereplicated genome sets and filtered using minimum cut-off values of 97% identity and 75% alignment.The Symbiobacteriia MAG generated from the ethane system was selected as the representative SymBio MAG due to the presence of full length AssA genes.TranscriptM (GitHub--sternp/transcriptm) was used to unambiguously mapp mRNA for each ORF and calculate the total transcripts per million (TPM).

Protein extraction and metaproteomics
For protein extraction, enrichment cultures collected from Phase 1 and 2 (10 ml each phase) were pelleted by centrifugation (18 000 g, 4 • C) and then washed with 1 × Phosphate buffered saline (PBS).The cell lysis and total protein digestion were performed as described previously [16].The digested peptides were analysed by liquid chromatography-tandem mass spectrometry using a Dionex Ultimate 3000 RSLCnano-LC system coupled to a Q-Exactive H-X Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Scientific).Mass spectra were searched against the annotated closed MAGs of E and B, respectively, in Thermo Proteome Discoverer.The identified proteins contained at least one unique peptide with a stringency cut-off of false discovery rate (q value) <0.05.

Structural modelling and molecular dynamics simulation
The amino acid sequences of three complete AssAs were acquired from the closed genome of "Ca. A. nitratireducens" in the ethanefed system.The tertiary structures of AssAs were modelled with Alphafold-2 [64].Fumarate and alkanes were bound to corresponding AssA by CB-dock-2 [65].AssA-Fumarate-Alkane complexes were solvated by CHARMM-GUI [66] with a thickness of 15 Å.Water type was TIP3P [67] and the force field was CHARMM36m [68].NaCl (200 mM) was used to ionize the systems [69].The final systems were then subjected to the MD simulations with NAMD 2.12 [70].Periodic boundary condition was applied to the simulating box, and particle mesh Ewald was used for the long-range electrostatic interactions.The pressure was set at 1 atm using a Langevin thermostat with a damping coefficient of 1/ps.A Nose− ´Hoover Langevin piston barostat with a decay period of 25 fs was applied.The temperature was reassigned every 500 steps.Simulations for each model include two steps.The first is 1 ns equilibration (NVP), and the second is 50 ns production run (NPT).

Root mean square f luctuation and deviation calculations
The root mean square f luctuation for α-carbons of the amino acid residues is calculated with Equation (1) [71,72].
where i represents the residue ID, T represents the total simulation time (Here is the number of frames), and r i (t j ) represents the position of residues i at time t j .The r i ref is the reference position of residue i, calculated by the time-average position.
To measure the average distance between two protein structures, the root mean square deviation is calculated with Equation (2).
where W = w i is the weighting factor, and N is the total number of atoms.The r i (t) is the position of atom i at time t after least square fitting the structure to the reference structure.The r i ref is the reference position of residue i defined by the reference structure (Here we used the initial structure as the reference).

Hydrogen bond analyses
The hydrogen bonds were analysed by VMD [72] based on 100 frames obtained from the last MD simulations of 40-50 ns.The cut-off distance and angle for hydrogen bond analyses were set as 3.5 Å and 20 • , respectively.
Innovation and Technology Commission of the Government of the Hong Kong Special Administrative Region.

Figure 1 .
Figure 1.Mass and electron balance batch tests, along with isotope labelling experiments, confirmed that anaerobic ethane/butane oxidation was coupled to nitrate reduction by the bioreactor enrichment cultures fed with C 2 H 6 /C 4 H 10 ; (A, B) typical biochemical profiles of the ethane (A, started on Day 490) and butane (B, started on Day 1100) systems showing simultaneous nitrate and ethane/butane consumption with transitory formation of nitrite, and production of dinitrogen gas and ammonium; there were two distinct phases for NO 3 − reduction; in Phase 1, NO 3 − was reduced to NO 2 −

Figure 2 .
Figure 2. Metabolic intermediates, inferred metabolic pathways, and a f luorescent micrograph of "Ca. A. nitratireducens"; (A) partial ion chromatograms (ion transition, m/z: 275>73.1) of culture extracts from the C 2 H 6 -fed bioreactor displayed a characteristic peak at a retention time of 9.940 min, matching the ethylsuccinate standard; (B) a characteristic peak at a retention time of 12.245 min (ion transition, m/z: 303.0>147.1),consistent with the peak from the butyl-succinate standard, was observed for the culture extracts from the C 4 H 10 -fed bioreactor (n = 4 at different sampling points); (C) cell cartoon illustrating "Ca. A. nitratireducens" in the C 2 H 6 -or C 4 H 10 -fed bioreactors (E or B) use AssA to activate ethane/butane to ethyl/butyl-succinate, which are further converted to acetyl-CoA and propionyl-CoA; fumarate could be regenerated by the methylmalonyl-CoA pathway or the TCA cycle; CO 2 is produced through the TCA cycle or the reverse WL pathway; the E and B both harbour genes that enable denitrification (except nirS/K) and dissimilatory nitrate reduction to ammonium; the colour of the square and circle symbols indicates the normalized gene expression values calculated as TPM (total TPM); blue bold text shows that the proteins were fully or partially detected in the protein extracts ( * Mthfd and * Fthd were only identified in B and E, respectively), while proteins in black text were not detected; (D) a composite f luorescence micrograph of the C 4 H 10 -fed enrichment culture hybridized with the SYMB-1018 probe[16] (Cy3, red; targeting "Ca. A. nitratireducens") and EUBmix probe set[37] (f luorescein isothiocyanate label, green; all bacteria); "Ca. A. nitratireducens" cells appear yellow (red + green) and other bacterial cells appear green; the scale bar indicates 20 μm; the representative image was selected based on the visual assessment of >3 separate hybridization experiments; FISH was performed as detailed in our previous study[16].

Figure 3 .
Figure 3.The MD simulations and gene expression of AssAs in "Ca. A. nitratireducens"; the structural representations of binding complexes of AssA2 with fumarate and ethane (A)/propane (B)/butane (C), and AssA3 with fumarate and ethane (D)/propane (E)/butane (F); key residues of Cys489 and Gly828 are close to each other in all systems; residues with occupancy of hydrogen bonds >50% were also included in the figures; (g) the normalized gene expression values of assA genes in "Ca. A. nitratireducens" from C 2 H 6 -, C 3 H 8 -, and C 4 H 10 -fed systems (calculated as total TPM).

Figure 4 .
Figure 4. Substrate range tests for the "Ca.A. nitratireducens" enriched in the C 2 H 6 -, C 3 H 8 -, and C 4 H 10 -fed bioreactors; subculture from the ethane bioreactor supplemented with propane (A) or butane (B) showed simultaneous nitrate and propane/butane consumption with production of dinitrogen gas and ammonium; the same was observed for subcultures from the propane bioreactor supplemented with ethane (C) and butane (D), and the butane bioreactor provided with ethane (E) and propane (F); each test was conducted in triplicate (results of other tests were included in Supplementary Fig.11).