In this study, activated sludge, freshwater sediment, and landfill cover soil were used as sources for isolation of new Methylococcus strains. Activated sludge samples were collected in bottles (500ml each) from two municipal wastewater treatment plants in Moscow, Russia. Freshwater sediment samples were collected in 50ml falcon tubes from beneath shallow water in an unnamed lake in Krasnodar region, Russia. Surface (depth of 0–5cm) soil samples were collected from a landfill site in Khanty-Mansiysk region. Aliquots of collected samples were used as inoculum to obtain enrichment cultures of methanotrophic bacteria. The latter were obtained using modified AMS medium (mAMS), containing (in grams per liter of distilled water) NH₄Cl, 0.1; MgSO₄×7H₂O, 0.2; CaCl₂×2H₂O, 0.02; 100mM phosphate buffer, pH 5.8, 1% (vol/vol); and trace element solution 0.1% (vol/vol), containing the following (g/L): EDTA, (in grams per liter) EDTA, 5; FeSO₄×7H₂O, 2; ZnSO₄×7H₂O, 0.1; MnCl₂×4H₂O, 0.03; CoCl₂×6H₂O, 0.2; CuSO₄×5H₂O, 0.1; NiCl₂×6H₂O, 0.02; and Na₂MoO₄, 0.03. After inoculation, 500ml bottles were sealed with silicone rubber septa, and methane was added aseptically using a syringe equipped with a disposable filter (0.22μm) to achieve a 10–20% mixing ratio in the headspace. Bottles were incubated on a rotary shaker (150 r.p.m.) at 42°C. After 1week of incubation, the cultures enriched with methanotrophic bacteria were subjected to serial dilutions. After several serial dilution steps, cell suspensions were plated on agar-solidified mAMS medium. The plates were incubated at 42°C in desiccators containing approximately 30% methane in air. The colonies appearing on the plates were picked and restreaked on the same agar medium. The set of finally selected colonies was subjected to several additional serial dilution steps in a liquid mAMS medium at 42°C until isolates of methanotrophic bacteria were obtained. Culture purity was verified by examination using phase-contrast microscopy and by plating on 10-fold diluted Luria–Bertani agar (1.0% tryptone, 0.5% yeast extract, 1.0% NaCl).

Prior to cell size measurements and growth experiments, the isolates were maintained in liquid mAMS medium with 20% (v/v) CH₄ at 42°C for 10days, with regular transfers each 2–3days. Methylococcus capulatus Bath was used as a reference organism in all experiments. Morphological observations and cell-size measurements were made with a Zeiss Axioplan 2 microscope and Axiovision 4.2 software (Zeiss). Cell sizes were measured for 20 randomly selected cells of all strains. Comparative analysis of growth characteristics of the isolates was performed by monitoring their growth dynamics in liquid mineral medium mAMS with 20% methane in the headspace within the temperature range of 25–55°C. Variations in the pH were achieved by mixing 0.1 M solutions of H₃PO₄, KH₂PO₄, and K₂HPO₄. All incubations were performed in triplicate. Growth dynamics was determined by measuring OD₆₀₀ of the cultures on an Eppendorf Biophotometer AG spectrophotometer. The ability to grow on methanol was tested in mAMS medium containing 0.01–3% (v/v) CH₃OH. Nitrogen sources were tested by replacing NH₄Cl in mAMS with 0.01% (w/v) NaNO₃, NaNO₂, urea, methylamine, glutamine, glycine, alanine, peptone, and yeast extract. Growth was examined after 3days of incubation. Salt tolerance was examined by adding NaCl to mAMS medium in concentrations of 0–2% (w/v).

Experiments on continuous cultivation were conducted in a 1.5L bioreactor filled with 1L of mineral medium. The inlet gases were methane with a flow rate of 100mlmin⁻¹ and air pumped in with an air compressor with a flow rate of 500mlmin⁻¹. The pH level of 5.6 was controlled by titration with 0.8% NH₄OH solution. Agitation was kept constant at 1,000 r.p.m. All gases were filtered with a 0.22μm sterile membrane. Each of the growth experiments started as a batch cultivation. When the culture reached exponential phase, bioreactor was switched to a continuous mode at a dilution rate of 0.05h⁻¹. The later was increased with an increment of 0.05h⁻¹ every 1–2days until a culture washout was observed. After that, the dilution rate was decreased by 0.05h⁻¹ and bioreactor was operated in a continuous mode for the next 5days.

Cells of Methylococcus strains were pre-grown to the exponential phase in 120-ml serum bottles containing 20ml of mAMS medium. Hydrogen utilization was tested with and without the addition of methane under fully aerobic conditions. Headspace of experimental bottles was filled with either 2% CO₂, 2% H₂ and 20% CH₄ or only 2% H₂ and 2% CO₂. Strains were cultivated for 24h using the same cultivation conditions as described above. Control treatments contained 2% CO₂ or 20% CH₄ and 2% CO₂ in the headspace. Methane and hydrogen were measured by gas chromatography at the beginning of the experiment and after 24h of cultivation. The optical density was measured at 600nm wave length (OD₆₀₀) using Eppendorf Biophotometer UV/Vis spectrophotometer (Eppendorf, Germany).

Cultures of new isolates were grown in the liquid mAMS as described above. The cells were harvested after incubation at 42°C on a rotary shaker at 150rpm for 2days. Genomic DNA extraction was done using the standard CTAB and phenol-chloroform protocol (Wilson, 2001).

Genomic libraries suitable for MiSeq sequencing were prepared with a NEBNext ultra II DNA Library kit (New England Biolabs). On average, a total of 1.68 million paired-end reads (2×300; 250nt) were obtained for each genome. Nanopore sequencing library was prepared using the 1D ligation sequencing kit (SQK-LSK108, Oxford Nanopore, United Kingdom). Sequencing was performed on an R9.4 flow cell (FLO-MIN106) using MinION device. Hybrid assembly of short and long reads was performed using Unicycler v.0.4.8 (Wick et al., 2017). Assemblies were evaluated with Quast 5.0 (Gurevich et al., 2013) and Busco 5.1.2 (Simão et al., 2015).

Annotation was performed using the RAST server (Aziz et al., 2008) and Prokka (Seemann, 2014). In addition, the presence of genes encoding enzymes of the primary and central metabolism as well as some secondary metabolic pathways discussed below was verified manually by web version of NCBI blastp using 35% identity and 50% coverage of amino acid sequence as a cut-off. The same criteria were used for searching genes of interest in the genomes of other methylotrophs. Assembled genomes have been deposited in NCBI GenBank under the accession numbers CP079095–CP079098.

A genome-based tree for members of the Methylococcus group was reconstructed using the Genome Taxonomy Database and GTDB-toolkit,¹ release 04-RS89. The maximum likelihood phylogenetic tree was constructed using MegaX software (Kumar et al., 2018). The Pyani program was used to estimate average nucleotide identities across Methylococcus genomes (Pritchard et al., 2016). The resulting distance matrices were further visualized as a heatmap in R (R Core Team, 2020).

The pan-genome was reconstructed using microbial pangenomics workflow in Anvi’o (Eren et al., 2015). The genomes were annotated in Prokka, after which the genes were organized using MCL algorithm into core, shell, and singleton clusters (Distance: Euclidean; Linkage: Ward). The core, shell, and singleton genes were separately annotated by BLASTp against the NCBI COG database using eggNOG-mapper (Cantalapiedra et al., 2021). Heatmap based on the annotated COG functions of the core and singleton gene clusters was then plotted in R. The Tettelin best-fit curves of the core- and pan-genomes were constructed using OMCL v1.4 implemented in GET_HOMOLOGUES pipeline (Contreras-Moreira and Vinuesa, 2013).

Whole-genome synteny was computed using Sibelia (Minkin et al., 2013). Synteny blocks were visualized using Circos (Krzywinski et al., 2009).

Insertion sequences (IS) were identified and classified into IS families using ISsaga pipeline (Varani et al., 2011) with IS finder database (Siguier et al., 2006) and ISEScaner pipeline (Xie and Tang, 2017). ICEfinder and ICEBerg v2.0 were used for IME detection (Liu et al., 2018b). Integron Finder (Cury et al., 2016) was used to identify regions containing integrons. Potential prophage regions were searched with PHASTER server² using the settings described in Arndt et al. (2016). Gene functions were determined by homology with known viral proteins in the NCBI GenBank database and the VirFam package³ using the settings described in Lopes et al. (2014).

In total, four novel isolates of Methylococcus-like methanotrophs were obtained and characterized in this study. Strains IO1 and KN2 were isolated from activated sludge samples of the Moscow municipal sewage plant (Table 1). These isolates displayed nearly identical 16S rRNA gene sequences and were closely related to M. capsulatus Bath (99.93–99.97% 16S rRNA gene sequence similarity). Strain BH was obtained from a sediment of an unnamed freshwater lake in Krasnodar region, South Russia and was also closely related to M. capsulatus Bath (99.02% 16S rRNA gene sequence similarity). Finally, strain Mc7 was isolated from a landfill cover soil in Khanty-Mansiysk region, West Siberia, Russia. In contrast to other three isolates, strain Mc7 displayed highest 16S rRNA gene sequence similarity (98.56%) to M. geothermalis IM1^T.

Four novel isolates were represented by non-motile, Gram-negative cocci, with a characteristic diplococcoid arrangement (Figure 1). Cell sizes of strains IO1, KN2, and BH were the same as those reported for M. capsulatus, approximately 1μm in diameter (Table 1; Figure 1A). Cell sizes of strain Mc7 (1.6μm in diameter; Figure 1B) were larger than those of other three strains. The colonies formed by these isolates on agar mineral medium after 2weeks of incubation with 20% (v/v) methane were small (1–2mm in diameter), round, milk-white (for strain Mc7) or cream colored (for strains IO1, KN2, and BH).

Methane and methanol were used as growth substrates, although the isolates differed in their ability to grow on methanol. Similar to M. capsulatus Bath, strains IO1 and Mc7 displayed only a trace growth (up to OD₆₀₀ 0.10–0.12) in a relatively narrow range of CH₃OH concentrations (Table 1). By contrast, strains KN2 and BH were capable of a relatively good growth (up to OD₆₀₀ 0.3) on methanol in a wide range of concentrations. The isolates were quite uniform with regard to their pH preferences (Table 1). Most strains grew in the pH range of 5.5–8.0. Strain KN2 demonstrated slightly better adaptation to moderately acidic conditions growing between pH 4.6 and 8.5. All strains used nitrate and ammonium as nitrogen sources. In addition, strains BH and KN2 were able to utilize glutamine, while strain IO1 showed weak growth on peptone as a nitrogen source. Strain Mc7 displayed highest tolerance to NaCl, up to 1% (w/v), while other isolates tolerated up to 0.75% (w/v) NaCl. For comparison, NaCl tolerance determined in our experiments for M. capsulatus Bath was 0.5% (w/v).

The ability to utilize hydrogen was tested using incubations with either H₂ and CO₂ or H₂, CO₂ and CH₄ in the headspace (see Materials and Methods). No growth as well as no H₂ consumption were observed in the absence of CH₄, with H₂ as the only energy source. With the only exception of strain KN2, the presence of H₂ in addition to CH₄ increased the growth yield of all strains examined in this study by 2–20% (Supplementary Figure S1A) as well as the corresponding CH₄ consumption (Supplementary Figure S1B). H₂ utilization in the presence of CH₄ was observed for all studied strains, including strain KN2 (Supplementary Figure S1C). These results agree well with the previous report of the absence of autotrophic growth of M. capsulatus Bath in liquid medium with H₂ and CO₂ (Baxter et al., 2002; Henard et al., 2021). Apparently, Methylococcus strains are capable of using H₂ as an alternative energy source during their growth on methane. Recently, the ability to grow mixotrophically on H₂ and CH₄ was also reported for verrucomicrobial (Methylacidiphilum sp. RTK17.1) and proteobacterial (Methylocystis sp. strain SC2) methanotrophs (Carere et al., 2017; Mohammadi et al., 2019; Hakobyan et al., 2020).

The special attention in our study was given to assessing growth characteristics of novel Methylococcus isolates. The results of measuring specific growth rates of novel isolates and M. capsulatus Bath within the temperature range of 25–55°C are shown in Figure 2. The two strains obtained from sludge samples, IO1 and KN2, were clearly more thermotolerant than M. capsulatus Bath and displayed growth optima at 48–52°C. The lowest temperature growth optimum of 40°C was revealed for strain Mc7, which is reasonable given that it was isolated from a landfill soil in West Siberia. Highest specific growth rates, 0.22–0.23h⁻¹, which were comparable to that of M. capsulatus Bath, were recorded for strains KN2 and Mc7. We also performed additional experiments in order to examine specific growth rates of novel isolates during continuous cultivation on methane in 1.5L bioreactor. The highest specific growth rate in a fermenter culture (0.3h⁻¹) was recorded for strain KN2. Other strains also demonstrated stable growth under conditions of continuous cultivation. Their growth rates, however, were below that demonstrated by strain KN2 (Table 1).

The genomes of four novel isolates were sequenced using a hybrid approach. Oxford Nanopore sequencing yielded 192,565–278,437 reads with a total length of 1.05–1.59 Gb (Supplementary Table S1). Sequencing on Illumina MiSeq platform generated a total of 415,244–3,477,384 paired-end reads, with a mean read length of 250bp. Both short and long reads were combined to perform a hybrid assembly using Unicycler, resulting in circular genomes. Genome characteristics of new isolates are summarized in Table 2.

Genome sizes varied from 3.2Mb in strain BH to 4.0Mb in strain Mc7. The DNA G+C content was highly similar in all examined genomes and constituted 63.44–63.52%. Each genome contained two copies of rRNA operon, two copies of pMMO, and one copy of sMMO. The number of protein-coding genes varied between 2,912 and 3,752. No plasmids were detected. Lowest number (17) of insertion (IS) elements was detected in strain KN2, while substantially higher number of IS elements (78) was observed in strain Mc7.

The genome-based phylogeny of four novel Methylococcus-affiliated isolates was determined using the comparative sequence analysis of 120 ubiquitous single-copy proteins (Figure 3). Three isolates (strains IO1, KN2, BH) were clustered together with two strains of M. capsulatus, Texas^T, and Bath, whereas strain Mc7 and M. geothermalis IM1^T constituted two separate branches within the genus Methylococcus. The complete genome of strain EFPC2, which is annotated in the GenBank as belonging to Methylococcus species, was also included in the analysis. However, as seen from Figure 3, strain EFPC2 clustered together with Methyloterricola oryzae 73a^T and did not affiliate with Methylococcus clade. The genome of strain EFPC2, therefore, was excluded from further comparative analysis of finished genomes available for Methylococcus species.

The average nucleotide identity (ANI) values were estimated for each pair of genomes (Supplementary Figure S2). ANI values calculated for isolates IO, KN2, BH, and M. capsulatus Bath were within a range of 98.75–99.73% which indicates that these organisms belong to the same species as intra-species level is defined at ≥95% ANI (Konstantinidis and Tiedje, 2005; Goris et al., 2007). ANI value determined for separately clustered strain Mc7 and M. geothermalis IM1^T was 88.56%. Thus, strain Mc7 may potentially represent a novel species of the genus Methylococcus.

The genomes of novel isolates, M. capsulatus Bath and M. geothermalis IM1^T, were included in the whole-genome synteny analysis in order to explore the evolution of genome structure. Overall, 216 shared synteny regions were identified. Among those, 86 regions were common for all six genomes. Figure 4 depicts only the synteny blocks affiliated with M. capsulatus Bath.

The number of synteny blocks revealed between M. capsulatus Bath and each of the isolates IO1, KN2, and BH was 102–103. Strain Mc7 and M. geothermalis IM1^T exhibited slightly lower levels of synteny to M. capsulatus Bath by displaying 97 and 93 synteny blocks, respectively. One of M. capsulatus-related isolates, strain KN2, and Methylococcus sp. Mc7 were chosen for more detailed analysis of whole-genome synteny patterns (Figure 5). As expected, the calculations made for two genomes resulted in lower number of synteny blocks, which were larger in size. Synteny regions shared between strain KN2 and M. capsulatus Bath covered 88.9 and 94.9% genomes of these bacteria, respectively. The largest synteny block in the genome of strain KN2 spanned about 1.1Mb and contained 978 genes. Two relatively large synteny break regions in the genome of strain KN2 spanned 224 and 76kb. Synteny breaks are caused by rearrangements, the insertion of novel genes, or the presence of genes that are too diverged to establish an orthologous relationship or have undergone expansion or loss (Liu et al., 2018a). Of 224 genes in the first synteny break region, 187 genes encoded hypothetical proteins. This region also included several genes coding for transposases, formate dehydrogenase displaying 95.5% amino acid sequence identity to NAD-dependent formate dehydrogenase from Methylocaldum marinum, flagellar transcriptional regulator FlhC, lysozyme RrrD, several helicases, and endonucleases. Of 84 genes in the second synteny break region in the genome of strain KN2, 76 genes encoded hypothetical proteins, while others coded for transport proteins. Comparison of the genomes of strain Mc7 and M. capsulatus Bath revealed 81 synteny blocks, which corresponded to nearly 50% of their genome length (Figure 5). The largest synteny break regions spanned 98.9, 98.5, 79.7, and 74.4kb. Approximately half of the genes within these regions encoded hypothetical proteins. Interestingly, 98.9kb region contained genes encoding small and large subunits of ribulose bisphosphate carboxylase.

The Anvi’o pangenomics workflow was used to cluster protein-coding sequences into core, shell, and singleton genomes. Of 4,485 identified gene clusters, the Methylococcus pan-genome core comprised 2,331 genes (on average 51.9% of each genome), with the shell genome containing 846 gene clusters (18.9% of total gene clusters) and the cloud containing 1,308 gene clusters (29.2% of total gene clusters; Figure 6).

Pan-genome of the genus Methylococcus is open. Supplementary Figure S3 displays the number of core genes (A) and the pan-genome size (B) as a function of the number of included genomes. Fitting the curve in Supplementary Figure S3B to a power law (3,146.3n^0.171) allows extrapolating calculations to n genomes. According to approximation, sequencing of one additional genome would provide 100 new genes to pan-genome. If the number of available genomes constituted 50 or 100 genomes, sequencing of one additional genome would contribute 21 and 12 genes, respectively. The core, shell, and cloud gene clusters were further annotated into COG classes (Supplementary Figure S4). The core genome was mostly conserved in the following: energy production and conversion (11.5% of total core gene clusters); translation, ribosomal structure, and biogenesis (8.6%); cell wall/membrane/envelope biogenesis (8.3%); coenzyme transport and metabolism (7.3%); inorganic ion transport and metabolism (6.3%); amino acid transport and metabolism (5.3%); posttranslational modification; protein turnover, chaperones (4.9%). However, the most abundant category was the one with no functional prediction (19.9%). The most abundantly represented functional categories in shell genome were replication, recombination, and repair, coenzyme transport and metabolism, signal transduction mechanisms, transcription, inorganic ion transport and metabolism, cell wall/membrane/envelope biogenesis, energy production and conversion. The majority of cloud genes were concentrated in the genomes of strain Mc7, M. geothermalis IM1^T, and M. capsulatus KN2 and represented replication, transcription, cell wall/membrane/envelope biogenesis, and signal transduction mechanisms. These functional categories were also affiliated with few cloud genes found in M. capsulatus Bath and strain BH. A number of studies demonstrated that core genes systematically represented the smallest fraction of pan-genome (Bazinet, 2017; Livingstone et al., 2018; Kumar et al., 2020; Oshkin et al., 2020). The large proportion of core genes in Methylococcus pan-genome may be explained by both the small genome size of individual Methylococcus strains and the low number of available genomes that could be taken for analysis. However, approximation for core- and pan-genome (Supplementary Figure S3) suggests that even if 30 genomes were included into pan-genome analysis, the fraction of core genes would still be high (41.1% of the total gene clusters).

Two complete copies of the pmoCAB gene cluster encoding pMMO and one additional copy of the pmoC (pmoC3) gene were revealed in the genomes of all novel Methylococcus strains. The corresponding pmoA sequences were nearly identical to each other and to the pmoA sequences from M. capsulatus Bath and M. geothermalis IM1^T. The sMMO-encoding gene cluster mmoXYBZDC was also present in all examined genomes, with MmoX sequences exhibiting high level of homology (100% identity between MmoX from M. capsulatus Bath and strains KN2, IO1, BH, and 97% identity with that from strain Mc7). This chromosomal locus in M. capsulatus Bath and strain IO1 contained also one gene encoding a hypothetic protein between mmoB and mmoZ genes. The gene cluster mmoGQSR coding for the large subunit of bacterial chaperonin GroEL (mmoG), the two-component sensory/regulatory system (mmoQ and mmoS), and the transcription activator (mmoR; Csáki et al., 2003) was found in all examined genomes.

Two PQQ-dependent MDHs catalyze the second stage of C₁-oxidation. The structural components of heterodimeric Ca²⁺-dependent MDH (MxaFI type) and the proteins required for its catalytic activity are encoded by the mxaFJGIRACKLD cluster with gene organization completely identical in all Methylococcus strains. Namely, the mxaG gene encodes the specific electron acceptor cytochrome c_L, the mxaACKL gene cluster encodes the proteins required for Ca²⁺ incorporation into the active site, while the mxaRS gene cluster encodes two proteins of unknown functions (Chistoserdova, 2011; Vuilleumier et al., 2012). The four strains harbor gene cluster xoxFJ, which encodes an alternative MDH (XoxF) containing a rare earth element in the active site (Hibi et al., 2011), and a periplasmic solute-binding protein (XoxJ). No other genes relevant to methanol oxidation were present in this locus. The xoxG gene coding for a putative electron acceptor from XoxF and being only distantly related to the mxaG gene (Keltjens et al., 2014; Yu et al., 2017; Zheng et al., 2018) was found in all genomes and was located separately from xoxFJ.

Oxidative transformations of formaldehyde are mediated by the enzymes of the tetrahydrofolate (THF)-based pathway whose amino acid sequences in the novel isolates are almost identical to those in M. capsulatus Bath. The complete sets of genes for the tetrahydromethanopterin (THMP)-dependent pathway of formaldehyde oxidation are present in all Methylococcus genomes. Each strain possesses the three-subunit formate dehydrogenase, which catalyzes the last stage of methane oxidation, and a five-gene operon encoding the cytoplasmic NAD⁺-dependent formate dehydrogenase composed of (abc)₂ subunits that catalyze the reversible reaction of formate oxidation to CO₂ (Hartmann and Leimkühler, 2013). All strains of M. capsulatus (Bath, KN2, IO1, and BH) possess also the two-subunit formate dehydrogenase, whereas this enzyme is missing in strain Mc7 and M. geotermalis IM1^T. The genomes of three strains, Mc7, KN2, and M. geothermalis IM1^T, encode the one-subunit formate dehydrogenase, which is not present in M. capsulatus Bath, IO1 and BH. Thus, strain KN2 possesses four formate dehydrogenases in comparison with other Methylococcus strains possessing three orthologs each. Since no other evident differences in the arrays of C1 oxidizing enzymes of these strains were found, one may hypothesize that the extended array of formate dehydrogenases helps increasing removal of C1 metabolites via their oxidation to CO₂. However, the exact role of different formate dehydrogenases in metabolism of methanotrophic bacteria remains to be elucidated.

Integrons are gene-capturing platforms that play a significant role in generating phenotypic diversity and shaping adaptive responses in microorganisms, including the spread of antibiotic resistance genes (Mazel, 2006; Gillings, 2014). The integron structure includes variable gene cassette array (recombination site attC with corresponding genes) and a stable platform containing integrase (IntI), recombination site (attI) and promoter (Boucher et al., 2007). Of all novel Methylococcus isolates, integron-like elements were detected only in the genome of strain Mc7. This strain contains one complete integron with one attC site, one separate attC site, large attC-cassette (nine sites), and stand-alone integrase. The gene encoding 6-phosphogluconate phosphatase was found in a large attC-cassette; the functions of other genes near attC sites could not be predicted. A complete integron containing three attC sites was also revealed in the genome of M. geothermalis IM1^T.

Integrative and conjugative elements (ICEs) are widespread mobile DNA that transmit both vertically, in a host-integrated state, and horizontally, through excision and transfer to new recipients. Recent findings indicated that the main actors of conjugative transfer are not the well- known conjugative or mobilizable plasmids but the integrated mobilizable elements (IMEs; Delavat et al., 2017; Guédon et al., 2017). IMEs encode their own excision and integration and use the conjugation machinery of unrelated co-resident conjugative element for their own transfer (Guédon et al., 2017). One putative IME was found in the genome of strain Mc7; unfortunately, the ends of the element could not be identified. A putative IСE region of 123,456bp was revealed in the genome of strain KN2. This region contained the genes of the type IV secretion system, which plays a key role in conjugation (Lawley et al., 2003). Interestingly, some fragments of a prophage were also found in this ICE region (Figure 8).

One to six prophage regions were found in the genomes of the examined Methylococcus strains (Table 3). Potentially complete prophages were detected in the genomes of strains IO1 and KN2. In the prophage region of strain IO1, the attL and attR sites were displaced, and the lysozyme genes were also absent (Figure 7). This prophage is structurally highly similar to the prophage 2 from M. capsulatus Bath. The prophage region in the genome of strain KN2 was, most likely, divided into large fragments as a result of integrase activity. This suggests that the conversion of these prophages to the lytic cycle is unlikely.

Insertion sequences are transposable DNA segments ranging in length from 0.7 to 3.5kb, generally including a transposase gene encoding the enzyme that catalyzes IS movement. The number of complete IS in the genomes of studied strains varied from 17 in KN2 to 78 in Mc7 (Figure 9). Similar to M. capsulatus Bath, the genomes of strains IO1 and BH contained IS elements belonging to IS256, IS3, IS5, and ISAs1 families (Siguier et al., 2014). The total number of complete IS elements in these three strains was also similar. By contrast, the genomes of strains IO1 and IM1^T did not contain IS elements of the ISAs1 family but contained one IS of the IS110 and IS630 families. Strain IM1^T also contained IS of IS91, ISNCY, IS200_IS605 families, which were not revealed in other studied genomes. Among the studied strains, the Mc7 genome was noticeably distinguished by both the total number and the variety of IS elements. IS massive expansion is accompanied by gene inactivation and decay, genome rearrangement, and genome reduction (Siguier et al., 2014). Some IS elements, for example, IS30 (Dalrymple and Arber, 1985) are capable of reactivating the expression of nearby genes. Therefore, IS elements can play an important role in the adaptation of the host cell to new lifestyle, such as continuous cultivation for the needs of industry (Gaffé et al., 2011). High similarity of the IS family composition between strains Bath, IO1, and BH suggests low influence of a lateral gene transfer on the evolution of their genomes. The change in the IS composition in other strains may be associated with the activity of other mobile elements, i.e., integrons in the genomes of strains Mc7 and M. geothermalis IM1^T, and ICE in the genome of strain KN2.

In summary, this study expanded the pool of phenotypically characterized Methylococcus strains with good-quality genome sequences by contributing four novel isolates of these bacteria from various habitats. Three novel isolates, strains KN2, IO1, and BH, were assigned to the species M. capsulatus and displayed surprisingly high similarity in gene content to that in M. capsulatus Bath. As evidenced by the original study of Foster and Davis (1966) and our work, methanotrophs of this species tend to inhabit nutrient-rich freshwater habitats, such as wastewater, sewage or sediments. They possess relatively small (3.2–3.6Mb) genomes and display high growth rates on methane in the temperature range between 42 and 52°C. The ability to grow on methanol, however, may vary in different members of this species. Thus, M. capsulatus Bath displayed only a trace growth on methanol. The same was true for one of the isolates obtained in our study, strain IO1, which showed poor growth only at low methanol concentrations (0.05–0.1%, v/v). By contrast, two other strains, KN2 and BH, were capable of growth in a wide range of methanol concentrations, up to 3.0% CH₃OH in case of strain KN2. Notably, high resistance of strain KN2 to methanol correlates with the occurrence of four different isozymes of formate dehydrogenase in this methanotroph. This array of formate dehydrogenases may potentially be responsible for the efficient oxidation of С1 metabolites to CO₂, thus determining high resistance to methanol. However, the exact role of different formate dehydrogenases in metabolism of methanotrophic bacteria remains to be clarified. Strain KN2 was also distinct with regard to its temperature optimum (48–52°C), which was higher than that in other isolates and strain Bath. Finally, strain KN2 showed the best performance during its continuous cultivation in a bioreactor. Taken together, M. capsulatus KN2 has the highest potential for the biotechnologies implying high growth rates on methane and/or abilities to grow on methanol.

The fourth isolate obtained in our study from a landfill cover soil, strain Mc7, may potentially represent a novel species of the genus Methylococcus. This is suggested by the results of phylogenomic analysis, comparative genome analysis as well as by the phenotypic difference of strain Mc7 and two previously described Methylococcus species. Thus, cells of strain Mc7 (~1.6μm in diameter) were larger that cells of M. capsulatus (1.0–1.1μm) or M. geothermalis (0.7–1.0μm; Awala et al., 2020). The genome size of strain Mc7 (4.0Mb) also exceeded that in other described Methylococcus species. Notably, the genome of this methanotroph from a landfill cover soil contained high number and variety of mobile elements, which may have played an important role in the adaptation of strain Mc7 to highly variable conditions of its natural habitat. Thus, strain Mc7 possessed a number of genome-encoded features, which were absent in strains of M. capsulatus, such as sucrose biosynthesis and the ability to scavenge phosphorus and sulfur from the environment. These metabolic capabilities may represent important components of the genome-determined environmental adaptations of this methanotroph. Further examination of phenotypic and chemotaxonomic features is needed to establish the taxonomic position of strain Mc7.

Thus, our study contributes to the current knowledge of Methylococcus-like methanotrophs, the bacteria of high environmental and biotechnological relevance. As suggested by the pangenome analysis, the metabolic diversity within this genus remains underestimated and calls for further cultivation- and genome-based studies.