Introduction

The human gut is a complex ecosystem harboring a diverse community of microbes, including bacteria, fungi, viruses, and archaea1. While bacteria are well described, and their role in gut health is well studied, the function and role of archaea in the human gastrointestinal tract (GIT) is still not well understood. Archaea found in the GIT are mainly represented by the order Methanobacteriales, with the most abundant species Methanobrevibacter smithii and Candidatus Methanobrevibacter intestini2,3,4.

These methanogenic archaea produce methane (CH4) as a metabolic byproduct while producing ATP via methanogenesis5,6,7,8,9. The yielded CH4 serves as an important indicator for the presence of methanogenic archaea10.

CH4 can be detected in humans’ breath, as it is partially absorbed into the blood and excreted through the lungs11. The impact of CH4 on the human GIT has already been addressed and it has been reported that CH4 leads to the reduced efficiency of intestinal smooth muscle activity, often associated with e.g., the constipation-predominant irritable bowel syndrome12,13,14. Further, variations of CH4 levels in the human breath were discussed as an indicator for enhanced oxidative stress and could therefore show a great potential in the context of in vivo diagnostics15,16,17. On the other hand, positive effects were also described, as CH4 could also act as antioxidant, protecting cells from reactive oxygen species (ROS) and demonstrating potential anti-inflammatory properties18.

Methanobrevibacter species show a prevalence of more than 90% in fecal samples with an average abundance of 0.56% for Methanobrevibacter smithii and 0.13% for Cand. Methanobrevibacter intestini3,19. In high-CH4 producing individuals (up to 40% of the middle European and Northern American population)20, this average abundance is raised to 2% for Methanobrevibacter genus in general21.

Notably, Methanobrevibacter species are highly correlated with certain bacterial representatives of the gut microbiome, e.g., Christensenellaceae, and have been shown to be syntrophically interactive in co-culture experiments with e.g., Christensenella minuta or Bacteroides thetaiotaomicron22,23.

Even though (methanogenic) archaea have been shown to be part of the human gut microbiome for more than 40 years24, knowledge about their function and role within the human gut microbiome is still sparse. The reason for this lies in several methodological issues as standard protocols are mainly bacteria-optimized, which are not optimally working for archaea due to their distinct cell structure, physiology, and metabolic activity. These distinct features and inappropriate protocols make it difficult to visualize, culture or analyze them with multi-omics approaches25.

Furthermore, the lack of comprehensive reference databases poses a challenge in adequately evaluating methanogenic archaea in the datasets. Additionally, the presence of a strong bacterial or host background further complicates the matter25. Therefore, it is of utmost importance to develop new protocols or modify existing ones to effectively study archaea in the complex microbiome environment.

Even though molecular biological techniques help to study the archaeome, isolating and culturing archaea is crucial for in-depth, mechanistic analysis of these microbes.

Based on the Global Catalog of Microorganisms (https://gcm.wdcm.org/) only few human host-associated archaeal isolates are available in culture (three Methanobrevibacter oralis strains from human oral cavity and subgingival plaque; and four Methanobrevibacter smithii isolated from human feces and large intestine), hindering detailed studies.

To tackle the lack of archaeal cultures for functional studies, we established herein a cultivation procedure to efficiently enrich and isolate methanogenic archaea from human stool samples in a targeted manner. By providing optimal growth conditions (e.g., by providing a suitable bacterial partner), followed by archaea-targeting Fluorescence-Activated Cell Sorting (FACS), dilution series and quality control through Nanopore sequencing, we were able to establish stable methanoarchaeal cultures from 14 out of 16 fecal samples obtained from both healthy and diseased subjects.

By this procedure, we obtained 9 novel methanoarchaeal strains, and were able to retrieve personalized representatives of Methanobrevibacter smithii and Cand. Methanobrevibacter intestini strains from stool samples of various participants with different health conditions, which were subsequently compared on genomic and functional levels.

As the role of archaea in health and disease is still unknown and it is still discussed whether they might have a potential beneficial or a pathogenic effect on human health, obtaining additional isolates from human samples will allow in-depth analysis, which might help to deepen our understanding of the role of methanogens in the human GIT.

Results

Set-up of a cultivation workflow and selection of bacterial syntrophic partners based on in silico predictions

For the cultivation of archaea from human fecal samples, we followed a combined strategy of: i) Providing in silico-customized media, partially including selected bacterial syntrophic partners to facilitate Methanobrevibacter growth, ii) pre-selection of sample donors for CH4 exhalation and such a higher methanogen load, iii) rigorous and standardized screening of enrichments for CH4 production and presence of (unique) methanoarchaeal key genes, and iv) FACS and antibiotic treatment for the mechanical and chemical enrichment of methanogens from the bacterial background of syntrophic partners or fecal samples. An overview of this strategy is provided in Fig. 1.

Fig. 1: General workflow of the cultivation strategy.
figure 1

Enrichment process encompassed strategies without (1) and with (2) bacterial partners. Created with BioRender.com.

Full size image

C. minuta and B. thetaiotaomicron were selected as potential bacterial syntrophic partners, based on previous reports of successfully established co-cultures with Methanobrevibacter species22,23, and the observed co-occurrence in metagenomic datasets21. Both species, C. minuta and B. thetaiotaomicron, were reported to provide hydrogen, acetate and possibly other metabolites to the archaeal partners, with C. minuta reported to be more efficient than B. thetaiotaomicron23. Nevertheless, we performed a more detailed prediction of the metabolic capacities of both bacteria to serve as a potential syntrophic partner for Methanobrevibacter species.

Metabolic models were generated for M. smithii, C. minuta and B. thetaiotamicron, based on their genomes. At this early stage, Cand. M. intestini was not included in the simulations. Single and co-cultures with MS and MpT1 media were simulated using flux balance analysis (FBA). Across all simulated co-cultures, the growth rate of M. smithii was increased compared to the single cultures on the same medium.

This growth boost is enabled by cross-feeding (Supplementary Fig. S5). Multiple metabolites could be identified as mediators of these interactions: The bacteria commonly provided H2, amino acids (alanine, lysine, glutamine, cysteine), nucleobases and nucleosides (uracil, inosine, adenosine, deoxyguanosine, xanthosine), vitamins and coenzymes (menaquinone 7, coenzyme A, NAD), as well as fructose, glycerol, formate, putrescine and ethanol (all output files are provided in our GitHub repository26). Based on the simulations, both B. thetaiotamicron and C. minuta were predicted to be fitting partners for M. smithii, without clear preference for either. As such, both bacteria and both media were selected for the enrichment cultures.

To optimize the enrichment cultures, we leveraged the methanoarchaeal autofluorescence of coenzyme F420 for FACS27, a method refined through the use of both stool samples and methanogenic reference strains (see “Materials and Methods”). This approach allowed us to establish effective sorting conditions using FACS, ensuring the accurate enrichment of target methanogens28.

Efficient enrichment of novel Methanobrevibacter from donors with detectable archaeal microbiome signatures

Fecal samples from all 16 study participants (Supplementary Data S1) were processed for enrichment of methanogenic archaea using suitable media with and without additionally added bacteria (see workflow overview Fig. 1). Overall, 8/8 samples of the healthy cohort and 6/8 of the diseased cohort yielded CH4-positive cultures (88% of all samples). Some samples yielded several enrichment cultures by different approaches (e.g., P15 and P86).

After initial pre-selection for unique methanoarchaea (according to their mcrA gene sequence) of all CH4-positive cultures, enrichment cultures of 14 participant samples were processed further (Table 1). FACS and visual monitoring for potential purity was performed, and after a minimum of 20 transfers they underwent Nanopore-based whole genome sequencing. An overview of the cultivation workflow and sequencing results (including taxonomic classification through kraken2/bracken) for each individual sample is provided on our GitHub repository26.

Table 1 Overview of participants, their health status, the derived cultures and the outcome of the metagenomic profiling of the original fecal samplea
Full size table

At this stage, 12 of 16 enrichment cultures were dominated by Methanobrevibacter representatives. In more detail, eight were predominated by M. smithii, and three by Cand. M. intestini. Further, two bacteria-dominated enrichment cultures (P51 and P46) contained a substantial amount of Cand. M. intestini (Fig. 2).

Fig. 2: Composition of enrichment cultures.
figure 2

Classified taxa and their relative abundance as assessed through Nanopore sequencing, followed by Kraken2/bracken with confidence score of 0.3. Small circles show composition after antibiotic treatment and dilutions series for selected cultures, which were ready at the time point of manuscript preparation. M. smithii is shown in green and Cand. M. intestini in yellow. Percentages of dominating methanogens are provided. All original bracken reports and culture purity updates are provided in our GitHub repository26. Source data are provided as a Source Data file.

Full size image

Notably, the dominance of either methanoarchaeon was evident in the original microbiome profiling. Samples were Cand. M. intestini was predominant in the profiling, exhibited similar dominance of this microbe in subsequent cultivation attempts. Remarkably, both Methanobrevibacter species were present in all samples initially. Yet during the enrichment process, one of these archaea emerged as the predominant species, consistent with their original relative abundance (Table 1). However, in the case of sample P86, we achieved successful enrichment of both Methanobrevibacter species separately (Fig. 2).

Noteworthy is the observation that samples P34 and P85 failed to yield archaeal cultures, despite positive methane breath test results of the donors (Table 1). While microbiome profiles were accessible for P85 (P34 had to be excluded due to technical reasons), it was evident that this individual lacked archaeal signatures, suggesting an alternative source of methane production, such as undetected methanogens in the oral cavity or small intestine. It shall be noted that these two individuals were the only ones with Crohn’s disease included in this study. It is indeed known that Crohn’s disease patients have a reduced amount of archaea in their GIT29.

To further purify the obtained enrichments and obtain monocultures of the methanogens, subsequent dilution series were and are being performed in combination with consequent antibiotics and lysozyme treatment (see “Materials and Methods”), leading to a substantial improvement of culture purity later on (e.g., P32: from 32.23% M. smithii to 95.05%; P 86 (MS + B.t.): from 80.48% Cand. M. intestini to 98.4%; Supplementary Data S3). In the meantime, we have improved the purity of more enrichments even further, for more details please visit our GitHub repository26.

We further visually examined the enrichments with Scanning Electron (SEM) and fluorescence microscopy. The SEM micrographs revealed the presence of mainly short rods (single or in chains) in each of the enrichment cultures, representing the known morphology of Methanobrevibacter species. Some rods appeared very bloated, as shown in Fig. 3 (arrows). All cultures were dense and revealed the presence of numerous, actively dividing cells. In agreement with the SEM examinations, we observed cells with Methanobrevibacter-typical morphology displaying the F420 autofluorescence in all cultures (Supplementary Fig. S6).

Fig. 3: Scanning electron micrographs visualizing the cell morphology of enriched archaeal cultures and co-enriched background bacteria.
figure 3

A enrichment culture P86, predominated by M. smithii. B enrichment culture P89, predominated by Cand. M. intestini. C enrichment culture P31, predominated by M. smithii. D enrichment culture P51, presence of Cand. M. intestini. E enrichment culture P86 (MS + B.t.), predominated by Cand. M. intestini, enlarged for better visibility of different cell phenotypes. Some rods appear bloated (arrows).

Full size image

Bacteroides thetaiotaomicron, but not Christensenella minuta, is a potent syntrophic partner for selective archaeal enrichment

Methane, being uniquely produced by methanogens, serves as an indicator for their presence in cultures. We noted the presence of CH4 in cultures grown in MS medium with and without supplementation with B. thetaiotaomicron after four days of incubation. Similarly, CH4 was detected in cultures grown in MpT1 medium, with or without supplemental B. thetaiotaomicron albeit only after 10 days of incubation.

The addition of B. thetaiotaomicron influenced our enrichment approach in a distinct manner. While its presence did not affect the overall success rate of methanogen enrichment - evidenced by 12 out of 16 positive enrichment experiments occurring both with and without B. thetaiotaomicron - it enabled the isolation of an additional strain (P66) that was not obtained in the abiotic medium (Fig. 2).

In contrast, cultures with additionally added C. minuta always remained CH4 negative (up to 8 weeks of incubation). Also, no autofluorescence of coenzyme F420 could be detected for these cultivation attempts. This finding was unexpected, as C. minuta had been described as an efficient syntrophic partner of M. smithii23 and we had originally expected positive effects in supporting the growth of these methanogens by providing C. minuta in the medium.

Co-enriched bacteria give indications on bacterial partners within the gastrointestinal microbiome

Methanobrevibacter species were enriched with background bacteria from the fecal samples. Interestingly, enrichment cultures with a predominance of M. smithii (Supplementary Fig. S7) showed only a few significant, weakly positive correlations with specific background bacteria, such as Phocaeicola dorei (q = 0.022052; rs = 0.3641) and UMGS75 sp900538885, a so-far uncultivated Bacilli (q = 1.389*10−9; rs = 0.4242; Spearman correlation and Benjamini Hochberg correction).

A number of negative associations of M. smithii were observed, including all Bacteroides (e.g., B. cellulosilyticus, q = 3.2267*10−6; rs = −0.4112) and E. coli species (e.g., E. coli_D, q = 7.8638*10−8; rs = −0.4926). These were, however, significantly positive with Cand. M. intestini (B. cellulosilyticus, q = 0.002058; rs = 0.4828; E. coli_D, q = 0.002042; rs = 0.2172; Supplementary Data S4; Supplementary Fig. S7)

Actually, in all found significant associations (Supplementary Fig. S8), the Methanobrevibacter species showed opposed preference for bacterial partners, indicating a somewhat different niche preference in the gastrointestinal setting. Also, M. smithii showed numerous negative correlations, whereas Cand. M. intestini, showed mostly positive correlations to co-cultivated bacteria (Supplementary Fig. S8).

Notably, B. thetaiotaomicron, although provided in the culture flasks to allow enrichment, was no longer predominant at later stages of the cultivation, indicating that the potential beneficial effect was rather on pre-priming the medium for methanogenic growth.

Although both methanoarchaeal species (M. smithii and Cand. M. intestini) co-exist in original gastrointestinal samples3, their growth in the enrichment cultures was significantly negatively correlated (Spearman correlation, rs = −0.6559, q = 1.967059*10−7), indicating a competitive behavior, at least under the cultivation conditions applied.

Cultured archaea represent nine novel, thus far uncultivated Methanobrevibacter strains

From all enrichment cultures, the respective archaeal genomes could be recovered by Nanopore sequencing and genome assembly (see “Material and Methods”). A summary of the genome classification according to the GTDB database, as well as the checkM2 output representing genome completeness and contamination, can be found in Supplementary Data S5. The obtained genomes were placed in the context of available genomes from human-associated Methanobrevibacter strains for further analysis2 (list of genomes provided in Supplementary Data S6; all genome sequences are available through our GitHub repository26).

In general, we preserved genomes from nine M. smithii and eight Cand. M. intestini strains. From the Cand. M. intestini clade, we obtained overall four different strain-clusters (cutoff: 99% similarity; Fig. 4; MIC-1 to MIC-4), while five clusters were obtained from the M. smithii clade (MS-1 to MS-5). All of these clades, except one (MIC-1) represent so far unknown genomes (below 99% identity compared to 1,167 genomes available; ANI values and matrix are given in Supplementary Data S7-S82;). All strains had no cultivated representative thus far, indicating that our enrichment strategy is well suited to expand the cultivable Methanobrevibacter diversity.

Fig. 4: Genome-based tree of all cultured Methanobrevibacter.
figure 4

Hierarchical clustering based on the ANI matrix, retrieved by the comparison of all cultured Methanobrevibacter (red letters) and already available genomes from isolates and MAGs (Supplementary Data S6). The M. smithii clade is indicated with a green background color; the Cand. M. intestini is shown with yellow background color. Cultures are accompanied by additional information on their source (healthy/ diseased donor, and culture conditions). Clusters representing one strain are indicated by black boxes (Supplementary Data S8). Source data are provided as a Source Data file.

Full size image

Differences of Methanobrevibacter smithii and Candidatus Methanobrevibacter intestini isolates revealed through sequence- and structure-based annotation

M. smithii and Cand. M. intestini show various differences in their genomic inventory2. Based on the sequence-based annotation obtained through eggNOG profiling, our findings here corroborate this observation. Including only high-quality genomes (> 90% completeness and < 10% contamination), we identified a significantly different functional profile (PERMANOVA, Ft = 2.2927, R2 = 0.12533, Pr(> F) = 0.002) among our enriched Methanobrevibacter representatives, clearly separating the two species (Supplementary Fig. S9). The analysis included reference genomes of M. smithii (Gut_genome132205; DSM2374) and Cand. M. intestini (Gut_genome143185; WWM1085; Supplementary Data S6-7). The most discriminative genes of all obtained genomes (present in ≥ 90% or ≤ 10% of the genomes) are shown in Table 2.

Table 2 Summary of the most discriminative genes found in genomes of M. smithii and Cand. M. intestini
Full size table

As typical for archaeal genomes, a high number of proteins cannot be assigned to a function (54% across 1.8 million proteins from 1167 genomes2), due to missing databases and mechanistic information. For that reason, we performed a structure-based protein function prediction via DeepFRI which is reliant on Alpha-fold30 for the most discriminative functions. By doing so, three previously unknown functional genes (2AVB5, 2C5WB and 2DDW5) from Cand M. intestini could be annotated (Table 2). Frequencies of the occurring GO terms in annotations are shown as word clouds in Supplementary Fig. S10.

Most discriminative functions were involved in substrate/ ion transport, such as for molybdate, as all M. smithii genomes analyzed in this study carried the modA and modB subunits for the ABC (molybdate) transporter. ModA binds molybdate and tungstate with high affinity, and together with ModB transport permease protein is widely distributed across methanogens. Molybdopterin is an important cofactor for the formate dehydrogenase, indicating some difference in the formate metabolism of both species31. Indeed, differences in the behavior of formate metabolism in Cand. M. intestini was found, as it, in contrast to M. smithii strains, accumulated formate in its medium, although the formate dehydrogenase and formate transporters are coded in its genome32. Further, unique nucleotide-sugar transmembrane transporters for each species were identified through DeepFri annotation.

Another finding was the potential presence of an oxalyl-CoA decarboxylase in the Cand. M. intestini; these enzymes are highly important to eliminate oxalates from human diet contained in e.g., coffee, tea and chocolate33, and would allow for the release of CO2 and formyl-CoA (https://www.ebi.ac.uk/QuickGO/).

Although Cand. M. intestini genomes were somewhat larger (1.8 Mbp) compared to M. smithii (1.7 Mbp; see also2), the overall gene count was similar (~1900 genes per genome). The percentage of core genes was also very similar (66.39% for M. smithii; 66.61% for Cand. M. intestini), but variations were observed in the percentage of non-core functions (27.79% and 29.47, respectively) (Supplementary Fig. S11). Notably, the representative Cand. M. intestini genome (Gut_genome132205) revealed double the amount (n = 36) of genes involved in defense mechanisms than the representative genome of M. smithii (Gut_genome132205, DSM2374, n = 18). The additional genes included mainly CRISPR-associated proteins or genes involved in Type I restriction modification (Supplementary Data S9), indicating a potentially more competitive environment for Cand. M. intestini (Supplementary Fig. S12)

Functional annotations indicate potential differences of Methanobrevibacter species according donor health status

We explored the functional profiles of the archaeal cultures obtained from both healthy and diseased participants. This analysis excluded all reference genomes due to the unknown health status of the source individuals and thus only initial insights can be retrieved. Please note further, that the following predictions for genome functions were only inferred from genomic data and not tested experimentally.

Overall, Methanobrevibacter genomes of healthy and diseased individuals did not differ significantly (Supplementary Fig. S13, eggNOG annotation), but the Cand. M. intestini population revealed variances according to the health status of the respective donor. Some potential unique functions for representatives stemming from the diseased (n = 4) and healthy cohort (n = 4) could also be identified (Supplementary Data S10). Unique functions found in Cand. M. intestini genomes from subjects with gastrointestinal issues were mainly CRISPR-associated proteins (arCOG05124: CRISPR-associated protein TM1802 (cas_TM1802)); COG3649: (CRISPR-associated protein Cas7)). Three unresolved annotations (arCOG07602, 2DP7E, 2DQUR) were further tackled through structure-based function prediction and identified functions involved, amongst others, hydrolase activities (Supplementary Fig. S14–15; Supplementary Data S10).

In general, the identified functions (DeepFri) for the unique COGs of Cand. M. intestini genomes derived from the diseased participants were mainly related to 5’−3’ nuclease activity, RNA polymerase activity and DNA polymerase complex, suggesting the alteration in gene expression, with the metabolic functions being mainly organonitrogen compound biosynthesis and macromolecule metabolism (HQ annotations only; Supplementary Fig. S16).

However, the Cand. M. intestini genomes obtained from the healthy cohort revealed a higher number of unique functions, in particular when DeepFRI-based annotation was performed. Annotations exclusive to healthy cohorts included a wider range of functions, e.g., peptidase regulation, lipid metabolism, glycosyltransferase, methyltransferase and nucleotidyltransferase, as well as regulation of signal transduction and cell communication. Other annotations were associated with functions associated with membrane, cytoskeleton and chromatin (Supplementary Fig. S16).

In summary, genes unique to the Cand. M. intestini genomes preliminarily identified in the healthy cohort in this study, encode a broad range of functions (cellular processing, metabolism, information storage and processing), whereas genes unique to the diseased cohort were mainly involved in CRISPR activities, DNA and RNA processing (Supplementary Fig. S16), warranting a closer look into this matter with additional cultures and genomes.

Discussion

Despite acknowledging archaea in the human gut microbiome for over 40 years14, our understanding of their functions and roles are still limited. The major bottleneck is the rare availability of cultivable representatives that would allow for deeper, mechanistic, comprehensive analyzes, including multi-omics and interaction studies34.

Development of an efficient cultivation procedure

In this study, we developed a targeted culturing procedure to efficiently enrich and isolate methanogenic archaea from human fecal samples. This approach enabled the stable cultivation of several M. smithii and Cand. M. intestini strains from participants with varying health conditions, facilitating genomic and functional comparisons. These cultures include nine previously uncultivated strains, with only one revealing overlap with one of more than 1000 recently recovered metagenome assembled genomes2.

From all donors, except Crohn’s disease patients, stable methanogen cultures could be recovered, which is a proof of the efficient enrichment pipeline. Such high efficiency could even allow personalized archaeome studies based on individual microbiomes.

Growth inhibition by C. minuta

(Human) methanogens often rely on syntrophy with bacterial partners that provide small carbon compounds and hydrogen for methanogenesis35. In particular Christensenella and Bacteroides species were considered potent partners, as they generally correlate positively with Methanobrevibacter species in natural gut microbiomes, in previous experiments23 and confirmation through in silico modeling22 (and this study). In our targeted approach, we utilized this positive interaction and added both bacterial species to the culture media in order to induce the low levels of methanogens to grow, as it has been proposed earlier34. Nevertheless, in contrast to cultures in which B. thetaiotaomicron was provided as syntrophic partner, all our enrichment attempts remained methane- and methanogen-negative when C. minuta was added, indicating a potential competition under the artificial culture-media conditions, despite their re-current co-existence in natural gastrointestinal microbiomes3,21. Pre-priming of the media with B. thetaiotaomicron, however, was slightly advantageous, as one Methanobrevibacter strain could only be cultured with its addition.

Co-enriched bacteria from human GIT microbiome

Besides those known interaction partners, we co-enriched background bacteria from feces, which might give indications on potential, natural bacterial interaction partners within the GIT microbiome. In contrast to M. smithii enrichments, Cand. M. intestini was revealing a strong co-occurrence with bacteria, such as E. coli and Bacteroides species. Both bacterial taxa produce a variety of substances from more or less complex sugar components, such as succinate, formate, acetate, lactate and ethanol36,37. This indicates a potential dependency of Cand. M. intestini on these or other components that are not well provided in the medium, or an inhibition of M. smithii by those or other components.

Retrieval of pure cultures

The provision of syntrophic partners, as well as the co-enrichment of bacteria from the original samples results in a minor, but stable background contamination in the enrichment cultures. To obtain final pure cultures, this background contamination has to be removed, by e.g., using optical tweezers, micromanipulators, encapsulation methods or high-throughput single cell isolation34. In our case, we used a combination of antibiotics/lysozyme treatment and dilution series, which substantially improved the archaeal predominance, and meanwhile led to several pure enrichments without bacterial contamination.

As all enrichment cultures are very stable and maintained for months and > 50 transfers in our laboratory now, we also might have to consider that some co-cultures that resist clonal purification, are reliant on the archaeal-bacterial interaction, i.e., by syntrophy or any other mechanisms. Co-cultures reflect natural conditions to a greater extent, and creative solutions are needed to still obtain pure cultures38,39, which are, for example, a prerequisite for the inclusion of these organisms in a culture collection. However, providing these strains to the scientific community is a declared goal, in order to extend the wealth of available Methanobrevibacter isolated from the human gut for future mechanistic studies.

Insights into the genomes and advantages of structure-based predictions

M. smithii and Cand. M. intestini stand out as distinct species, characterized by unique microbial networks, physiological roles, and genomic potentials28,32. Like many other archaea, much of their genetic landscape remains unexplored, as a significant portion of their genes lack annotation, and genetic models are yet to be established. To delve deeper into their characteristics, we employed structure-based predictions to complement sequence-based tools for annotation40. As previously observed41, also archaeal COG groups when further annotated using DeepFRI30, exhibit enhanced accuracy and coverage of annotated functions compared to orthodox orthology-based methods. As such, we were able to retrieve a higher resolution in the distinction of M. smithii and Cand. M. intestini genomes, and identified some novel functions (such as the putative oxalyl-CoA decarboxylase) that could be relevant for the gastrointestinal ecosystem.

The role of methanogenic archaea for health and disease are still a matter of ongoing studies (summarized in ref. 3, and ref. 18). In particular, the role of Cand. M. intestini remains to be explored, as earlier studies could not resolve both species separately, due to the high sequence similarity in the 16S rRNA gene32. In our study, the number of participants and the subsequent cultures and genomes was quite low, so the relevance of either species for disease/health cannot be concluded at this stage. Nevertheless, structure-based analyzes of the clusters of orthologous groups of the Cand. M. intestini genomes revealed some differences based on the health status of the donor, with a broader variety of functions associated with the healthy phenotype, and a number of functions associated with CRISPR, and DNA/RNA processing in the diseased phenotypes, also lacking a clear indication of their subcellular location, which might indicate a potential foreign origin (e.g., phage/virus). These results should be considered as preliminary evidence, showcasing the potential lying in the cultures and genomes, but also emphasizing the need for further investigation.

Conclusion and outlook

In this study, we have successfully extended the cultivable human archaeome through a targeted cultivation approach that relied on a strategic selection of samples and optimized culture conditions, including syntrophic bacterial strains. This approach facilitated the growth of Methanobrevibacter species, resulting in the stable cultivation of nine novel strains belonging to the two major archaeal species in the human gastrointestinal ecosystem, namely M. smithii and Cand. M. intestini. The derived genomes can be used to obtain first insights into the physiology and functions that might be relevant for human health or disease.

Furthermore, our method has proven highly efficient, as evidenced by ongoing work in our laboratory, as we were able to enrich methanogens even from individuals with methane content below 5 ppm in their breath. This highlights the robustness of our approach and its potential for widespread application. However, it should be noted that the success rate of obtaining enrichments was lower in low methane emitters (13% methanogen-positive enrichments) compared to a success rate of 88% in pre-screened donors.

Our work opens the way for comprehensive analyzes that promise to deepen our understanding of methanogen dynamics within the human gastrointestinal tract. Expanding the public isolate collection is essential for gaining further insights into the roles of methanogens in the human microbiome.

Methods

Ethics and participant recruitment

Sixteen participants with different health conditions (healthy or with any kind of gastrointestinal disorder) were recruited based on their baseline breath CH4 levels (> 5 ppm / high CH4 emitters). The recruitment was supported by medical doctors at the State Hospital Graz (Austria) and an independent physician also located in Graz, Austria. Participants were asked to undergo a breath test, provide a stool sample, and complete a questionnaire with health status information (Supplementary Data S1). Further, all participants signed an informed consent before participation. A list of exclusion and inclusion criteria can be found in Supplementary Data S2. The study was approved by the local ethics committee at the Medical University of Graz, Austria (EK-Nr: 31-452 ex 18/19). The sex of participants was recorded (Supplementary Data S1) but was not included in any analysis due to the low sample number processed and a different research focus.

Sample collection, processing and storage

For the pre-selection of participants, a breath test was performed21. For this purpose, participants performed the breath test either directly on a GastroCH4ECKTM Gastrolyzer (Bedfont Scientific Ltd., UK) or with a GastroCH4ECKTM breath bag (Bedfont Scientific Ltd., UK), which was used for a later measurement on the Gastrolyzer device. Measurements were performed according to the manufacturer’s instructions. Participants, who exhaled more than 5 ppm CH4, were considered as high CH4 emitters based on21 and were asked to provide a stool sample and personal data including their health status. For anaerobic cultivation the stool samples were collected with eSwabsTM (Copan, USA) and from the same source for metagenomics with a sterile stool sample container (Sarstedt, AUT). The stool sample container was pre-filled with 5 ml DNA/RNA free ethanol to stabilize the samples42.

The participants were asked to return the samples within 24 h after collection and store them at 4 °C in the meantime. After delivery, the samples for anaerobic cultivation were immediately transferred into sterile anoxic (N2) serum bottles (Glasgerätebau Ochs Laborfachhandel e.K, GER) with sterile syringes (InjektⓇ-F Solo, B.Braun SE, GER) and the same volume of glycerol containing ascorbic acid (50% glycerol (v/v), 50% 1x PBS (v/v), and 1 g/l ascorbic acid) was added. Afterwards, the samples were mixed by inverting, incubated for one hour at 4 °C and then stored at −80 °C until cultivation. Samples for DNA extraction were transferred in DNA-free 1.5 ml sterile reaction tubes (Eppendorf, GER) and frozen at −80 °C until further processing.

Microbiome profiling of the original samples

The DNA of the original stool samples was isolated using the PureLink™ Microbiome DNA Purification Kit according to manufacturer’s instructions (Invitrogen, Thermo-Fisher, USA).

Shotgun sequencing was conducted by the company Macrogen using the Illumina HiSeq device. Raw sequencing reads were quality filtered with ATLAS (version 2.16.2) which also included the removal of human reads43. For taxonomic annotation the following tools were used: Kraken2 (version 2.1.2; with a conf. 0.1) and Bracken (version 2.7; read length 100)44,45. To complete the hierarchical lineage information the script “kreport2mpa.py” was used that is publicly available at GitHub (https://github.com/jenniferlu717/KrakenTools/blob/master/kreport2mpa.py) before merging the feature tables of all samples.

General workflow of enrichment and cultivation

Selection of bacterial syntrophy partners and metabolic modeling

Metabolic modeling was used to test the suitability of bacterial syntrophy partners on MS46 and MpT147,48 media for enrichment of Methanobrevibacter smithii (full recipes are available in ref. 32).

Draft metabolic models were generated using gapseq (v1.2)49 on the genomes of the B. thetaiotamicron, C. minuta and M. smithii (GenBank IDs NC_004663.1, GCF_001652705.1 and ABYW00000000.1, respectively). The recipes for MS and MpT1 media were used to build a representation of media constituents in the ModelSEED50 namespace. The compositions of the undefined media ingredients yeast extract and casein hydrolysate were approximated using the definitions of yeast extract and tryptone by Marinos et al.51. The initial models were gap-filled with gapseq on each of the media separately, including the respective nutrient additives for C. minuta and B. thetaiotamicron.

For the simulations of the co-cultures, PyCoMo52 was used to build community metabolic models for each pair of organisms in each medium. The respective medium with bacterial additives were applied to the models. Growth of the communities was calculated at equal abundance of bacteria and M. smithii. Exchange metabolites and cross-feeding interactions were calculated from solutions of flux balance analysis with the same parameters.

Cross-feeding interaction profiles (metabolites produced by organism A and consumed by organism B in a given co-culture) were compared by calculating the fraction of the intersection of cross-fed metabolites divided by the union of cross-fed metabolites in two given profiles.

Cultivation

A general workflow of the enrichment process is shown in Fig. 1 and detailed flowcharts for each sample can be found on our GitHub repository (https://github.com/SDMUG/Methanobrevibacter_Enrichment)26. For the enrichment of methanogenic archaea, stool samples were gently thawed on ice and cultivated using the following two approaches: i) First, the samples were inoculated in MS and MpT1 media supplemented with antibiotics (0.1 mg/ml penicillin G potassium salt (Carl Roth GmbH + Co. KG, GER) and 0.1 mg/ml streptomycin sulfate (Carl Roth GmbH + Co. KG, GER) to reduce bacterial growth. ii) For the second approach, 0.2 ml of stool suspension was directly inoculated into MS and MpT1 media containing bacterial cultures that had already been grown for one day. For this co-cultivation, either Bacteroides thetaiotaomicron (DSM 2079) or Christensenella minuta (DSM 22607) (purchased from Leibniz Institute DSMZ - German Collection of Microorganisms and Cell Cultures GmBH, GER) were used, as these bacteria have been described to support the growth of the methanogenic archaea22,23, and we could also predict their supportive capacities through metabolic modeling (see below).

Inoculation for this approach followed the procedure described in ref. 23. For the cultivation, both media were prepared under anaerobic conditions. In case of cultures with B. thetaiotaomicron, the following supplements were additionally added before inoculation to MS and MpT1 media, following the instructions of the DSMZ for the growth of B. thetaiotaomicron (medium 78 + haemine + Vit K1, https://www.dsmz.de/collection/catalog/details/culture/DSM-2079): 0.2 ml hemin solution (50 mg hemin (Sigma-Aldrich, USA) in 1 ml 1 M NaOH (Carl Roth GmbH + Co. KG, GER), diluted to 100 ml with ddH2O); 0.2 ml vitamin K1 solution (0.1 ml vitamin K1 (Alfa Aesar GmbH & Co KG, GER) in 19 ml 95% (v/v) ethanol (Avantor Performance Material Deventer, NLD)) and 1 ml beef extract solution [100 μg/ml] (MP Biomedicals, USA). In case of C. minuta, the media were supplemented with 0.4 ml D-glucose solution [500 μg/ml] (Amresco Inc., USA). All supplements were also prepared under anaerobic conditions and pressurized with 1.0 bar N2.

The cultures were subjected to incubation at a temperature of 37 °C using an incubation shaker set at a shaking speed of 80 rpm. The growth of archaea was influenced by the medium and the presence of supporting bacteria. The incubation period ranged from 7 to 14 days.

The cultures from both approaches were screened for methane and F420-positive cells. In the case of approach two, screening occurred prior to FACS sorting, as the sorting process was based on F420 autofluorescence of the methanogens.

CH4 measurement and fluorescence microscopy

To detect the presence of methanogens, the CH4 sensor BCP-CH4 (BlueSens gas sensor GmbH, GER) was used to measure CH4 in the gas phase of the culture bottle according to the manufacturer’s manual. Any cultures with CH4 levels above 2% were classified as CH4 positive.

The presence of methanogens was further checked via the typical autofluorescence of coenzyme F420. For this purpose, fluorescence microscopy (Zeiss microscope Axio Imager A1, Carl Zeiss AG, GER) with fluorescence upgrade was used with Zeiss filter set 05: BP 395-440 (excitation), FT 460 (beam splitter), LP 470 (emission) and 100x magnification (objective Plan-NEOFLUAR).

Fluorescence activated cell sorting (FACS)

To physically enrich the methanogens within the samples, the fluorescence activated cell sorting (FACS) was used on CH4-positive cultures grown with B. thetaiotaomicron or C. minuta. Sorting was conducted via the Core Facility Flow Cytometry at the Center of Medical Research in Graz, Austria with FACSAriaTM IIu system (BD Biosciences, USA)53. Prior to sorting, samples were prepared in an anaerobic chamber (Whitley A95 Anaerobic Workstation, Don Whitley Scientific Limited, UK) as follows: 1 ml of culture was centrifuged at 4000 × g for 5 min in a 1.5 ml reaction tube (Eppendorf AG, Hamburg, GER), and the resulting cell pellet was washed twice with 500 μl 1x PBS containing 1 g/L L(+)-ascorbic acid (AppliChem GmbH, GER). The cell pellet was then resuspended in 1 ml of 1x PBS + ascorbic acid and further diluted with an additional 1 ml of 1x PBS + ascorbic acid and transferred to a Falcon® 5 ml Polystyrene Round-Bottom Tube with 35 µm Cell-Strainer Cap (Corning Science Mexico S.A. de C.V., MEX) before sorting to ensure a single cell suspension.

Cells were sorted based on the autofluorescence of coenzyme F420 (excitation wavelength at 420 nm) based on27. F420-negative and F420-positive cells were distinguished using a 405 nm laser for excitation and a 450/50 band pass filter for detecting fluorescence emission. The presence of autofluorescent cells was observed in CH4-positive cultures in contrast to its absence in CH4-negative cultures. In addition, original stool samples, enrichment cultures, and pure cultures of various methanogenic archaea (Methanobrevibacter smithii (DSM 2375), Methanosphaera stadtmanae (DSM 3091), and Methanomassiliicoccus luminyensis (DSM 25720)) were employed for identifying the sorting gate (Supplementary Figs. S1–S4). The sort time was chosen for each culture individually and ranged between 3 and 7 min, depending on the archaeal cell count.

Sorted cells (min. 800 events, details for each sample can be found in the flowcharts on our GitHub repository26) were transferred into a 1.5 ml microcentrifuge tube containing 250 μl of 1x PBS + ascorbic acid and inoculated into fresh medium. Afterwards the cultures were incubated at 37 °C with 80 rpm shaking speed under anaerobic conditions.

Antibiotics, Lysozyme, Dilution series

To reduce bacterial growth and enrich methanogens within the cultures without the addition of B. thetaiotaomicron or C. minuta and cultures after FACS, antibiotics (penicillin G potassium salt (0.1 mg/ml) and streptomycin sulfate (0.1 mg/ml) were added to the media directly before sample inoculation. For further reduction of antibiotic-resistant bacteria, lysozyme (0.5 mg/ml; all substances: Carl Roth GmbH + Co. KG, GER) was added after each culture transfer.

For further purification of the methanogens, dilution series from 10−1 to 10−13 were performed under anaerobic conditions. Growth was usually observed up to a dilution of 10−7 and 10−8. After one week of incubation (37 °C, 80 rpm shaking speed) the CH4 levels of positive growth cultures were measured and cultures were evaluated further via fluorescence microscopy.

Sequencing and sequence data analysis

Two sequencing approaches were followed to assess the composition of the cultures. Sanger sequencing was used to identify the (almost) pure archaeal and bacterial components of enrichment cultures, whereas Nanopore whole genome sequencing was used for a detailed assessment of metagenomic composition and genome reconstructions.

Sanger sequencing

For Sanger sequencing, the DNA was extracted from CH4-positive cultures using the MagNA Pure Compact system and MagNA Pure Compact Nucleic Acid Isolation Kit (Roche Diagnostics GmbH, GER). To prepare the samples, 1 ml of culture was centrifuged at 4000 × g for 10 min, then 800 μl of supernatant was discarded and 180 μl of MagNA Pure 96 External Lysis Buffer (Roche Diagnostics GmbH, GER) was added. The extraction was performed according to the manufacturer’s instructions.

For archaea, following primers were used: mcrA_forward (5’-GGH GGN GTH GGD TTY CAN CAR TA-3’) and mcrA_reverse (5’-CRT TCA TNG CRT ART TNG GRW AGT-3’) (Eurofins Genomics, Ebersberg (DE)), to amplify the mcrA gene of the methanogens (primers were modified from54 and adapted to the human archaea genome catalog2).

The amplification was carried out using Ex Taq® DNA polymerase, 10x ExTaq Buffer w/MgCl2, 2.5 mM dNTP Mix (Takara Bio Inc., JPN), 10 μM of each primer, 20-30 ng template, and PCR- grade water (Jena Bioscience GmbH, GER). The PCR conditions were as follows: initial denaturation for 5 min at 95 °C, 35 cycles of denaturation for 40 s at 95 °C, annealing for 2 min at 51.5 °C, extension for 1 min at 72 °C, and a final extension for 10 min at 72 °C.

To identify remaining bacteria in the CH4-positive cultures, the 16S rRNA gene was amplified using forward primer 9bf (5’-GRG TTT GAT CCT GGC TCA G-3’) and reverse primer 1406uR (5’- ACG GGC GGT GTG TRC AA-3’) (Eurofins Genomics, GER)55,56. The conditions for 16S rRNA gene PCR were as follows: initial denaturation for 2 min at 95 °C, 10 cycles of denaturation for 30 s at 96 °C, annealing for 30 s at 60 °C, extension for 1 min at 72 °C, 25 cycles of denaturation for 25 s at 94 °C, annealing for 30 s at 60 °C, extension for 1 min at 72 °C, and a final extension for 10 min at 72 °C. PCR success was determined by agarose gel electrophoresis.

Afterward, the PCR products were purified using the Monarch PCR & DNA Cleanup Kit (New England Biolabs GmbH, GER), according to the manufacturer’s manual and the DNA concentration of the purified samples was measured using a Nanodrop 2000c spectrophotometer (Thermo Fisher Scientific Inc., USA). To the purified products the respective primers with a concentration of 10 µM were added and the mixture was sent to Eurofins Genomics AT GmbH, Vienna (Austria) for sequencing.

The sequencing results were analyzed and identified using the standard nucleotide BLAST from NCBI. However, the mcrA gene sequence of Cand. Methanobrevibacter intestini was not yet available in NCBI, so the taxonomic affiliation was determined by constructing a phylogenetic tree of the mcrA gene sequences using an in-house database (see below).

Nanopore sequencing

To determine the complete genome of the enriched archaea and the remaining co-existing bacteria from the examined cultures, Nanopore sequencing on MinION Mk1C (Oxford Nanopore Technologies plc., UK) was performed (nanoporetech.com). Initially, DNA was extracted (InvitrogenTM PureLinkTM Microbiome DNA Purification Kit, Thermo Fisher Scientific Inc, USA), following the manufacturer’s protocol. The quality and concentration of extracted DNA were assessed using a Nanodrop 2000c spectrophotometer (Thermo Fisher Scientific Inc., USA) and an InvitrogenTM QubitTM 3 Fluorometer (Thermo Fisher Scientific Inc., USA), and DNA fragmentation was evaluated using agarose gel electrophoresis. The DNA was stored at −20 °C until further processing.

For library preparation, DNA was repaired using the NEBNext Companion Module (New England Biolabs GmbH, GER), and then prepared for the sequencing on a chemistry version 14 flow cell (R10.4.1, FLO-MIN114) following the Ligation sequencing gDNA–Native Barcoding Kit 24 V14 (SQK-NBD114.24) according to the manufacturer’s instructions (nanoporetech.com).

Sequence data analysis

McrA genes were aligned and integrated into an in-house database using ARB57 and MEGA1158. The in-house database is based on a collection of mcrA gene sequences from available genomes from human gut archaea2. Information on the mcrA gene of CH4-bearing enrichments was mainly used to identify unique enrichments and to weed-out isolate-duplicates at early stages. For that, distance matrices were calculated, integrating the generated sequences into the database. It should be mentioned that M. smithii and Cand. M. intestini cannot clearly be distinguished based on the 16S rRNA gene sequence, but the mcrA gene allows for a clear distinction of both species32.

The obtained data from Nanopore sequencing was analyzed using the following configurations: the MinION Mk1C device was set to run for 72 h, pore scan frequency of 1.5 h, minimum read length of 200 bp, high-accuracy base calling, and enabled active channel selection, reserved pores, read splitting, trimming barcodes and mid-read barcode filtering. The software versions used included MinKNOW v22.10.7, Bream v7.3.5, Configuration v5.3.8, Guppy v6.3.9, and MinKNOW Core v5.3.1 (nanoporetech.com). Obtained sequencing data from Nanopore was analyzed using the following specifications: Simplex base calling was carried out on a GPU node at the Life Science Compute Cluster (LiSC) UniversityofVienna (Austria),usingguppy-gpuv6.4.2andthe dna_r10.4.1_e8.2_260bps_sup.cfg model for superior base calling. Duplex base calling was achieved by following the recommended tutorial (https://github.com/nanoporetech/duplex-tools) using dorado v0.2.3 with the dna_r10.4.1_e8.2_260bps_fast@v4.1.0 fast model, followed by finding duplex pairs with duplextools v0.2.20-3.10 and dorado stereo/duplex base calling with the dna_r10.4.1_e8.2_260bps_sup@v4.1.0 superior model.

Further analysis of the merged simplex and duplex reads involved quality control using NanoPlot nanocomp v1.20.059, filtering with filtlong v0.2.1 (--min-length 200;--keep_percent 90, and --target_bases 20 million) (https://github.com/rrwick/Filtlong).

The filtered reads were then assembled using flye v2.9.1 in --nano_hq and --meta mode with a target genome size of 2 Mbp60. The reads were mapped using minimap2 v2.2461, and then read polishing was performed with racon v1.5.0 (https://github.com/isovic/racon), using the following settings: (-m -8 -x -6 -g -8 -w -500). A consensus contig was obtained using medaka v1.7.2 and the r1041_e82_260bps_sup_g632 model (https://github.com/nanoporetech/medaka). Basic genome statistics were generated using genometools calling gt seqstat62, and the genomes were classified using gtdbtk v2.2.6 with the --full_tree option enabled63. Genome completeness and contamination were estimated using checkm2 v1.0.1 with –allmodels64.

Taxonomic Classification

Reads were mapped against the Unified Human Gastrointestinal Genome (UHGG v.2.0.1) downloaded from MGgnify (https://www.ebi.ac.uk/metagenomics) consisting of more than 289.000 archaeal and bacterial RefSeq genomes using Kraken v.2.1.245. In order to increase the specificity and to compensate for the chance of returning the incorrect lowest common ancestor (LCA) of all genomes, a confidence threshold of 0.3 was chosen for Kraken2. To determine the relative abundance of bacterial and archaeal species, the Kraken2 output was subjected to analysis using Bracken v.2.7. (https://ccb.jhu.edu/software/bracken/) with default settings.

Scanning electron microscopy (SEM)

The cell morphology of the enriched archaea and their remaining co-existing bacteria was visualized using the Zeiss Sigma 500 VP scanning electron microscope (Carl Zeiss AG, GER). Culture samples of 2 ml (with the supernatant not removed) were centrifuged at 4000 × g for 10 min, and the resulting pellets were handed over to the Core Facility Ultrastructure Analysis from Medical University of Graz (Austria) for SEM preparation. Briefly, the cells were mounted on coverslips, fixed with 2% (wt/vol) paraformaldehyde in 0.1 M phosphate buffered saline, pH 7.4 and 2.5% glutaraldehyde in 0.1 M phosphate buffered saline, pH 7.4, and dehydrated stepwise in a graded ethanol series. Samples were post fixed with 1% osmium tetroxide (Electron microscopy Sciences) for 1 h at room temperature and subsequently dehydrated in graded ethanol series (30-96% and 100% (vol/vol) EtOH). Further, HMDS (Merck | Sigma - Aldrich) was applied. Coverslips were placed on stubs covered with a conductive double coated carbon tape65. SEM micrographs were taken with an electron voltage of 5 keV and a magnification of 3–21 kX, using the secondary electron detector (Zeiss Oberkochen).

Tree calculation, correlations and pangenome analysis

The ANI (average nucleotide identity) matrix of the genomes (all-vs-all) was calculated using the ANI Matrix calculator provided through (http://enve-omics.ce.gatech.edu/g-matrix/); the resulting phylogenetic tree was annotated using itol66.

Correlation heatmaps were generated in R. All scripts are provided in our GitHub repository (https://github.com/SDMUG/Methanobrevibacter_Enrichment)26. Statistical associations were based on spearman correlations, which are also included in the provided script.

Pangenome analyzes were performed using kbase67 based on RastK annotation, the Compute Pangenome and Pangenome Circle Plot (v1.2.0) function, including all available high-quality (> 90% completeness and < 10% contamination) M. smithii and Cand. M. intestini genomes (this study and2), with genomes of Gut_genome132205 and Gut_genome143185 serving as reference genomes.

Genomic, functional annotation and prediction using DeepFRI

Genome comparison was performed using drep v3.4.268, and genome annotation was carried out using eggnog-mapper v2.1.10 (default settings (e-value cut-off 0.001)69,70,71,72,73,74. Prodigal was used for gene identification.

To examine the functional differences between M. smithii and Cand. M. intestini as well as to explore the differences between functions related to a specific health condition, PCoA were performed using a Bray-Curtis distance matrix. The visualization was done using the R package ggplot2 (v. 3.4.4)75. PERMANOVA was performed with n = 999 permutations to identify statistical significance in the functional profiles.

For identifying the functions of protein-coding genes in archaeal genomes, protein-coding genes were first predicted using Prodigal (https://github.com/hyattpd/Prodigal). Then, gene sequences translated to amino acids were annotated using a metagenomic pipeline (https://github.com/bioinf-mcb/Metagenomic-DeepFRI/) with an embedded DeepFRI tool30. The pipeline generated a query contact map using results from mmseqs2 target database (https://github.com/soedinglab/MMseqs2) searching for similar protein sequences with known structures. Next, the contact map alignment was performed to use it as input to DeepFRI Graph Convolutional Network (GCN) and annotations without matches to known structures were processed with Convolutional Neural Network (CNN). DeepFRI version 1.1 was used in this study; in version 1.1 we retrained the original DeepFRI architecture using high-quality models from the AlphaFold Database, thus increasing the number of functions the method can predict by up to 4x76. For medium quality (MQ) and high quality (HQ) annotation, scores higher than or equal to 0.3 and 0.5 were used, respectively. For selected genes, a structure prediction was performed using AlphaFold database76,77, followed by DeepFRI annotation (https://github.com/SDMUG/Methanobrevibacter_Enrichment). Wordclouds were generated using the pip package (https://pypi.org/project/wordcloud/).