The enriched archaea activate alkanes with highly transcribed Acrs
Neither the two abundant archaeal MAGs nor the entire coassembly contain alkylsuccinate synthases, which excludes the activation of alkanes through fumarate addition. Instead, both MAG 1 and MAG 4 encode three Mcrs consisting of the three subunits McrA, McrB, and McrG. Comparably high numbers of Mcr-related genes are only known for the sister group Ca. Syntrophoarchaeum20. The six mcrA sequences, which code for the catalytic subunit of the enzyme35, form three distinct clusters of two mutually similar sequences of each species (≥ 89% identity) in the clade of multi-carbon alkane-activating acrAs (Fig. 3a, Extended Data Table 2)16–18, 24. These clusters are highly similar to acrAs of Ca. Syntrophoarchaeum (Extended Data Table 2). Notably, neither MAG 1 nor MAG 4 contain a canonical Mcr for methane oxidation like the one present in ANME-1. None of the other retrieved MAGs contains acr genes.
In MAG 4, two Acrs form a continuous operon, whereas the third acrA gene is separated from the acrB and acrG subunits. In MAG 1, two acrA genes are isolated while one is part of an operon (Extended Data Fig. 3). A separation of the acrA from the other subunits was previously reported for one operon in Ca. Syntrophoarchaeum18, further highlighting the similarities between the two groups.
Both candidate taxa highly expressed the genomically isolated acrA subunit of the third cluster (Fig. 3b, Supplementary Table 4), placing it among the top 19 (C8) to top 4 (C5) expressed genes. These acrA sequences are located close to a cluster of Acrs that presumably activate long-chain alkanes, for instance by oil-associated Ca. Methanoliparum archaea which consume C16 + alkanes 21,22.
As demonstrated for other alkane-degrading cultures18,19, a selective inhibitor of the Mcr/Acr, the CoM analogue 2-bromoethanosulfonate (BES)36,37, suppressed sulfide production in the C6 and C14 cultures (Extended Data Fig. 4a,b), consistent with an Acr-dependent alkane activation mechanism. Indeed, metabolite extracts of the individual alkane-degrading cultures contained the corresponding alkyl-CoM as indicative activation product (Fig. 3c, Extended Data Fig. 5). Extracts from cultures incubated with longer alkanes contained more sub-terminal than terminally activated substrates (2-alkyl-CoM > 1-alkyl-CoM), whereas the incubations with shorter alkanes produced both variants in similar ratios (Fig. 3c, Extended Data Fig. 5). An activation at terminal and subterminal positions was previously observed for Ca. Syntrophoarchaeum18. The comparatively high activation rate at the terminal position for shorter alkanes is surprising, because the C-H bonds at terminal positions are stronger than those at subterminal positions particularly in short alkanes38. A further degradation of the non-terminally activated compounds would likely require a prior rearrangement to the 1-alkyl-CoM similar as described for bacterial alkane degradation39.
We conclude that the archaea represented by MAG 1 and MAG 4 grow on the supplied alkanes. We propose the genus name Candidatus Alkanophaga, consistent with the recently suggested name Candidatus Alkanophagales23 for this clade, and analogous to the closely related methane-oxidizing Ca. Methanophagales (ANME-1)40.The species that was abundant in cultures supplied with shorter, volatile alkanes (C5-C7) represented by MAG 4 is termed Ca. Alkanophaga volatiphilum and the species abundant in cultures with longer alkanes (MAG 1) Ca. Alkanophaga liquidiphilum, according to their preferred substrate properties. Based on amino acid identities (AAI) of 55–59% between the Ca. Alkanophaga MAGs and the ANME-1 MAGs included in the phylogenomic analysis (Supplementary Table 5), we propose that Ca. Alkanophaga form the family Ca. Alkanophagaceae next to the family ANME-141. This is supported by the GTDB classification of the ANME-1 B39_G2 MAG.
Substrate tests indicated that Ca. A. volatiphilum cannot degrade alkanes larger than C9, while Ca. A. liquidiphilum can degrade all alkanes between C6 and C15 (Extended Data Fig. 6). Neither species grew on C3 or C4, despite coding highly similar Acrs to Ca. Syntrophoarchaeum. Highly resolved crystal structures of the active Acrs could help elucidate these differences in substrate preference. Other factors such as specific membrane transporters might modulate substrate selectivity and require further investigation.
Candidatus Alkanophaga completely oxidize the alkanes to CO2
The degradation of the alkyl-CoM units generated by the Acr requires a conversion to acyl-CoA (Fig. 4b). The underlying reactions for this transformation are unknown, but for other alkane-degrading archaea some candidate enzymes have been proposed. The C2-oxidizing Ca. E. thermophilum may catalyze these reactions with tungstate-containing aldehyde:ferredoxin reductases (Aors). This archaeon encodes three aor copies located closely to genes of the (reverse) methanogenesis pathway, and expresses them during ethane oxidation. While both Ca. Alkanophaga MAGs encode complete aor gene sets and several additional aor domains, those genes were only moderately expressed (Supplementary Table 4). This is similar to other alkane oxidizers16,18 and contradicts a significant role of the Aor for this conversion. Ca. Syntrophoarchaeum might employ a highly transcribed methyltransferase for this step18; however, this enzyme appears mismatched for interactions with larger alkane substrates. The conversion of alkyl-CoM to acyl-CoA requires further investigation.
Like Ca. Syntrophoarchaeum18, we expect Ca. Alkanophaga to process the acyl-CoA units via the β-oxidation pathway, therewith cleaving acetyl units from the activated substrate42. Ca. Alkanophaga encode several copies of the four genes of even-chain β-oxidation, which were expressed during alkane oxidation (Figs. 4 and 5, Extended Data Fig. 7, Supplementary Table 4). For a complete oxidation of odd-chain alkanes, three additional enzymes are required to degrade the potentially toxic three carbon compound propionyl-CoA42,43. Two of these enzymes, propionyl-CoA carboxylase (pcc) and methylmalonyl-CoA epimerase (mce), are absent in both Ca. Alkanophaga MAGs, making it unlikely that propionyl-CoA is degraded this way. An alternative pathway, the conversion to pyruvate and succinate via the methylcitrate cycle43, is equally unlikely because five of the six necessary genes are absent in both MAGs. Alternatively, propionyl-CoA could be attached to proteins (propionylation)44.
The acetyl-CoA units produced by β-oxidation are either shuttled into biomass production or are completely oxidized. For the latter, the acetyl-CoA decarbonylase/synthase (Acds) splits the acetyl-CoA into CO, which is then fully oxidized to CO2, and a methyl group which is transferred to tetrahydromethanopterin (H4MPT). The enzymes of the H4MPT methyl branch of the WL pathway then oxidize the methyl-H4MPT to CO2 19,20. Both Ca. Alkanophaga species encode and expressed several Acds and all enzymes of the WL pathway, except the methylene-H4MPT-deyhdrogenase (mtd) missing in the Ca. A. volatiphilum MAG (Figs. 4a and 5, Extended Data Fig. 7, Supplementary Table 4).
Notably, both Ca. Alkanophaga MAGs encode a canonical 5,10-methylene-H4MPT reductase (mer), unlike the other members of the class Syntrophoarchaeia, Ca. Syntrophoarchaeum and ANME-1. This enzyme catalyzes the oxidation of methyl-H4MPT (CH3-H4MPT) to methylene-H4MPT (CH2 = H4MPT) in the first step of the oxidative WL pathway45. The mer copies of Ca. Alkanophaga are highly similar to each other (91%). This close relatedness is supported by a phylogenetic mer analysis, which places the mer sequences of Ca. Alkanophaga next to each other and close to those of the hydrogenotrophic methanogens Methanocellales46 (Extended Data Fig. 8a). Because of this positioning, we suspect that the mer of Ca. Alkanophagales was inherited vertically from the methanogenic ancestor of Methanocellales. Another possibility is that Ca. Alkanophaga obtained the mer copies via lateral gene transfer from Methanocellales or a common ancestor. In Ca. Syntrophoarchaeum and ANME-1, the function of the missing mer seems to have been replaced with a methylene-tetrahydrofolate (H4F) reductase (metF) of the H4F methyl branch of the WL pathway20,47. Both Ca. Alkanophaga MAGs also encode copies of this enzyme, which share high identities (70–80%) with the ones of Ca. Syntrophoarchaeum, and cluster next to metF sequences of Hadesarchaeota from Jinze hot spring (China) and Yellowstone National Park (USA) (Extended Data Fig. 8b). While both mer and metF are transcribed, mer is especially expressed by Ca. A. liquidiphilum in the cultures oxidizing the longer alkanes C10, C12, and C14 (Fig. 5b,d, Extended Data Fig. 7e,f,k,l).
In general, the alkane oxidation genes of the Ca. Alkanophaga species that was dominant in the respective culture were strongly expressed, while the expression of the less abundant species was negligible. Interestingly, in the C6 culture, we observed an activity of both species, with a lower expression rate of the less abundant species Ca. Alkanophaga liquidiphilum (Fig. 5a).
Ca. Alkanophaga partner with a sulfate-reducing Thermodesulfobacterium
Ca. Alkanophaga lack the dissimilatory sulfate reduction (DSR) pathway and therefore have to transfer the electrons released during alkane oxidation to a sulfate-reducing partner organism. While several MAGs affiliated with the sulfate-reducing archaeon Archaeoglobus appeared in some cultures (Supplementary Table 3), their DSR genes were not transcribed. Another candidate for sulfate reduction was a species of the thermophilic bacterial genus Thermodesulfobacterium, which was enriched in all cultures. The genome of this organism, MAG 24, contains the three DSR genes ATP-sulfurylase (Sat), APS-reductase (Apr), and dissimilatory sulfite reductase (Dsr)48,49. These genes were actively transcribed, and especially enriched in the cultures growing on longer alkanes (Supplementary Table 4). Therefore, it is very likely that this Thermodesulfobacterium acts as syntrophic sulfate reducer in our cultures. We propose the name Candidatus Thermodesulfobacterium syntrophicum for this species. Ca. T. syntrophicum is closely related to the hyperthermophilic Thermodesulfobacterium geofontis, which was isolated from the Obsidian Pool in Yellowstone National Park50 (Extended Data Fig. 9).
Syntrophic microorganisms transfer reducing equivalents via molecular intermediates, such as hydrogen or formate51, or via direct interspecies electron transfer (DIET)52. Both Ca. Alkanophaga and Ca. T. syntrophicum encode membrane-bound [NiFe]-hydrogenases, including several hydrogenase maturation factors, which could facilitate electron transfer via molecular hydrogen. Some of these genes were substantially expressed (Fig. 5, Extended Data Fig. 7, Supplementary Table 4). Formate dehydrogenases, which are required to produce and consume formate as electron carrier, were also present in both partners, and partially enriched (Fig. 5, Extended Data Fig. 7, Supplementary Table 4). However, the addition of hydrogen or formate did not accelerate sulfide production in substrate-containing cultures (Extended Data Fig. 4c,d). Moreover, cultures in which sulfate reduction was inhibited by the addition of sodium molybdate produced only miniscule fractions (max. 2.1% for C6 and 0.8% for C14) of the hydrogen concentrations that would be necessary were hydrogen the sole electron carrier (Supplementary Table 6). Thus, Ca. Alkanophaga and Ca. T. syntrophicum use neither molecular hydrogen nor formate as primary agent for the exchange of reducing equivalents.
Alternatively, alkane oxidation and sulfate reduction might be coupled by DIET, as suggested for other alkane-oxidizing consortia19,20,53. DIET likely involves bacterial type IV pilin (PilA) and specific multi-heme c-type cytochromes (MHCs), which form filamentous nanowires that connect the partner organisms and allow a conductive exchange of reducing equivalents54–57. The precise role of both proteins in the wiring is still under study. While filaments formed by specific types of PilA and the highly similar archaeal protein flagellin B (FlaB) were shown to be electrically conductive without the presence of MHCs58,59, conductive nanowires in the bacterium Geobacter sulfurreducens consist exclusively of stacked hexa-heme cytochromes (OmcS)60. Both components are present and strongly expressed in previously established alkane-oxidizing consortia19,20,53.
Surprisingly, the Ca. Alkanophaga MAGs do not encode proteins with multiple CxxCH motifs, the typical heme-binding sequence61. MHCs were also absent in the previously published Ca. Alkanophaga MAG. This contrasts previous observations in the closest relatives of Ca. Alkanophaga, ANME-1 and Ca. Syntrophoarchaeum, which encode numerous MHCs20,40. The entire assembled metagenome contained 81 proteins with two or more CxxCH motifs, some of which might pertain to the missing 10% of the Ca. Alkanophaga MAGs. However, none of these MHCs was highly expressed. In contrast to Ca. Alkanophaga, Ca. T. syntrophicum encodes six MHCs. Only one of these cytochromes, containing four CxxCH motifs, is slightly enriched in all cultures (Supplementary Table 4). This suggests a minor role of MHCs in the interaction of both organisms.
Both Ca. Alkanophaga contain genes coding for PilA, and additionally encode several copies of FlaB62 for the formation of cell appendages for DIET. Remarkably, these cell appendage genes were among the most highly expressed genes of Ca. Alkanophaga. For instance, Ca. A. volatiphilum strongly expressed three genes annotated as PilA and/or FlaB in the C5, C6, and C7 culture samples. In Ca. A. liquidiphilum, five PilA/FlaB genes were strongly expressed, with one copy of PilA being the most expressed gene in the C14 culture. Ca. T. syntrophicum encodes several PilA genes as well, some of which were strongly enriched in the C10-C14 cultures (Supplementary Table 4). Because of the very high expression of cell appendage genes in Ca. Alkanophaga, and the unresponsiveness of the cultures to hydrogen and formate, we predict that electron transfer in our cultures is based on DIET. The lack of MHCs in Ca. Alkanophaga might be compensated by MHC production in the partner bacterium similar to observations in syntrophic AOM cultures, where only the bacterial partner expressed type IV pili genes53. DIET in these consortia might also be completely independent of MHCs. The previous observation that methanogenic Methanosarcinales archaea are capable of DIET without MHCs supports this possibility63.