Characterization of Mycobacterium antarcticum sp. nov., a psychrotolerant member of the genus Mycobacterium isolated from Antarctic soil

Abstract

Most species of the genus Mycobacterium are widely distributed in soil and aquatic environments; however, there is limited understanding of their distribution and adaptation to permanently cold environments. Psychrophilic/psychrotolerant bacteria isolated from cold environments represent an important resource that can be used to understand these aspects. Here, we report representative psychrotolerant mycobacteria strains, designated TUM20985^T, TUM20983, and TUM20984, from Antarctic soils. Phylogenetic analysis using the 16S rRNA gene and average nucleotide identity confirmed that these three strains belong to the same species, and are most closely related to Mycobacterium hodleri JCM 12141^T. These strains are psychrotolerant and grow at 4–25 °C, but not above 30 °C. All three strains exhibited a higher proportion of unsaturated fatty acids compared with closely related species, with C_18:1ω9c, C_17:1ω8c, C_17:0, and C_16:0 identified as the major fatty acids. The predominant respiratory quinone was MK-9(H2). Comparative genomic analysis showed potential functional divergence in lipid metabolism genes, and revealed the presence of nitrate reduction and potential methane or ammonia oxidation genes that were absent in closely related species. Based on these findings, we propose that our strains represent a novel species, designated Mycobacterium antarcticum sp. nov., with TUM20985^T (RIMD 4000203^T = NCTC 15058^T) as the type strain.

Introduction

Characterization of microorganisms isolated from different environments is important for understanding phenotypic diversity and environmental adaptation. As of March 2025, the genus Mycobacterium comprised more than 200 validly published species¹ some of which have been extensively studied as pathogens. Animal hosts are the primary ecological niche for some of these pathogenic species, such as M. tuberculosis and M. leprae; however, the majority of species are commonly found in soil and water. Notably, some of these environmental species are also capable of causing human infections. These species, which are known as nontuberculous mycobacteria, are a growing concern because of their increasing prevalence worldwide². Therefore, it is important to elucidate the factors that determine the environmental niches of this genus. Comparing bacterial strains with different environmental preferences can contribute to this understanding.

A prior culture-independent study³ revealed a diverse and high relative abundance of this genus in a low-temperature environment, suggesting its distribution in such habitats. In Antarctica, studies based on both culture-independent and culture-based analyses showed the presence of mycobacteria in soils^4,5,6,7. Moreover, a few previously described species of the genus Mycobacterium such as M. fluoranthenivorans, M. frederiksbergense⁵ and M. hodleri⁶ have been isolated from Antarctica. Through our attempt to culture psychrotolerant bacteria of Antarctica, we also isolated three psychrotolerant mycobacteria strains from soils of ice-free areas in East Antarctica⁸. We previously reported the whole genome sequences of these strains, and they were considered candidate novel species owing to their low genomic similarity to other described species.

However, it is not clearly understood how these psychrotolerant species adapt to low-temperature conditions. Cold-adapted bacteria generally exhibit features that differ from their relatives such as regulation of cell membrane fluidity, synthesis of cold adaptation proteins, regulators and metabolic changes, energy supply, and reactive oxygen species⁹. Meanwhile, it remains unclear which of these or other mechanisms are specifically involved in cold adaptation by species of the genus Mycobacterium. Understanding these adaptation mechanisms could enhance our knowledge of bacterial evolution of this genus in response to environmental challenges.

Characterizing cold-adapted bacteria and comparing them with mesophilic relatives is an effective approach to understanding cold adaptation mechanisms. Such comparative analyses can reveal specific genomic features and potential metabolic pathways that have evolved for cold adaptation, while also providing insights into the ecological plasticity of mycobacteria. In this study, we characterized strains TUM20985^T, TUM20983, and TUM20984 using a polyphasic approach that included comparative genome analyses with their closest mesophilic relatives. Note that we previously reported these strains as novel members of Mycolicibacterium⁸ which was one of the proposed five genera (Mycolicibacterium, Mycolicibacter, Mycolicibacillus, Mycobacteroides, and Mycobacterium) that had split from the genus Mycobacterium at the time of the report based on sequence-based classifications¹⁰. However, considering the ongoing debate regarding reconstituting the genus Mycobacterium^11,12 we followed the proposal by Meehan et al.¹¹ and included the split genera as sub-genera within Mycobacterium. Here, we propose that our strains represent a novel species, Mycobacterium antarcticum sp. nov. To the best of our knowledge, this is the first new species of Antarctic-derived Mycobacterium.

Materials and methods

Morphological characterization

The Antarctic strains TUM20985^T, TUM20983 and TUM20984 were first isolated from soil supernatant collected in East Antarctica: TUM20983 from Langhovde (69.2404 S, 39.7655 E), and TUM20984, and TUM20985^T from Skallen (69.6044 S, 39.7458 E). Sampling was conducted within the framework of the Japanese Antarctic Research Expedition (JARE), following receipt of approval from the Ministry of the Environment, Government of Japan. Isolation was performed through cultivation on Reasoner’s 2 A (R2A) agar plates (Becton, Dickinson and Company, Franklin Lakes, NJ, USA)⁸. However, the strains were confirmed to grow similarly on Trypticase Soy Agar (TSA, Becton, Dickinson and Company); therefore, TSA was used for subsequent tests. Antarctic strains were grown on TSA for 7 days at 25 °C, at which time the cell morphology was examined by light microscopy (Stemi 2000-C, Carl Zeiss. Oberkochen, Germany) after Gram staining using a Favor G kit (Nissui Pharmaceutical Co., Ltd, Tokyo, Japan). Strains were also evaluated by transmission electron microscope after preparation as described below. Each pellet of evaluated strains, cultured on TSA and obtained by centrifugation at 4,000 x g for 10 min, was fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer (PB) for 2 h, washed with 0.1 M PB, then post-fixed in 1% osmium tetroxide buffered with 0.1 M PB for 2 h. After washing again with PB, cells were resuspended in 2% gelatin and pelleted again (4,000 x g, 10 min). The microcentrifuge tubes were then submersed in ice-cold water to quickly solidify the gelatin with the cells. Next, the tips of the tubes were cut open and the cell pellets were cut into 1–3 mm blocks. The blocks were subsequently dehydrated in a graded series of ethanol and then embedded in Epon 812. Finally, ultrathin section (70 nm) were collected on copper grids, double-stained with uranyl acetate and lead citrate, and then examined by transmission electron microscopy (JEM-1400Flash, JEOL, Tokyo, Japan).

Physiological and biochemical characterization

The growth range of temperatures (4 °C, 15 °C, 25 °C, 30 °C, and 37 °C) and NaCl concentrations (0–8% [wt/vol]) were assessed on TSA and pH (4 to 10) was assessed in Trypticase Soy Broth (TSB, Becton, Dickinson and Company). The pH was adjusted by adding either 6 M HCl or 5 M KOH prior to sterilization. Growth curves at 25 °C were monitored under aerobic conditions in test tubes containing 5 mL of TSB supplemented with 0.05% Tween 80 (FUJIFILM Wako Chemicals, Osaka, Japan). The initial inoculum was 50 µL of bacterial suspension adjusted to an optical density at 600 nm (OD₆₀₀) of 0.400 to 0.500. The OD₆₀₀ was measured at 1-hour intervals using an OD-Monitor C&T (TAITEC Corporation, Saitama, Japan) for up to 240 h. Doubling times were estimated by linear regression of log-transformed OD₆₀₀ values during the exponential growth phase. The slope of the fitted regression line was used to calculate doubling time as log (2) divided by the slope of the regression line. All physiological and biochemical characteristics were tested at 25 °C. Additionally, a range of biochemical characteristics were measured using API 20E (bioMérieux, Marcy l’Etoile, France) according to the manufacturer’s instructions, with the exception that the incubation period was extended to 10 days. Enzymatic activity was further evaluated using API ZYM (bioMérieux) following the manufacturer’s protocol. Tween 80 hydrolysis was evaluated as previously described¹³. Catalase activity was determined using 3.0% H₂O₂, with gas production taken to indicate a positive reaction. Nitrate reduction was evaluated using a tube-based method¹⁴ at 25 °C, using sulfanilic acid and α-naphthylamine reagents, followed by zinc if no color developed.

Determination of chemotaxonomic properties

The chemotaxonomic properties of Antarctic strains were characterized based on their genomic GC content, major respiratory quinones, cellular fatty acid composition, and mycolic acids (MAs) content. Quinone and fatty acid methyl esters samples were prepared and analyzed by TechnoSuruga Laboratory Co. Ltd. (Shizuoka, Japan). Quinone extraction and analysis were conducted as previously described¹⁵. Briefly, cells were grown on TSA at 25 °C for 5 days, then harvested and freeze-dried. Total lipids were subsequently extracted from the cells using a modified method¹⁶ after which quinones in the crude extract were purified using Sep-Pak plus silica (Waters Corp., Milford, MA, USA). Quinones were then analyzed by ultraperformance liquid chromatography using a liquid UPLC H-class system (Waters) connected to a PDA eλ detector (Waters). Samples were separated using an Eclipse plus C18 column (2.1 × 150 mm, 1.8 μm; Agilent Technologies), after which data were analyzed using the Masslynx v4.2 software (Waters). Quinone species were determined by analyzing the linear relationship between log retention times (UPLC) and isoprene unit numbers¹⁷. For measurement of the fatty acid composition, fatty acid methyl esters were prepared and analyzed using the manufacturer’s protocol for the Sherlock Microbial Identification system v6.0 (MIDI, Inc., Newark, DE, USA), after which fatty acid profiles were identified by comparison with the TSBA6 6.20. The genomic GC content was calculated from the whole genome sequences of the strains⁸. The MA subclass patterns were determined by silica gel thin-layer chromatography (TLC, Uniplate, 20 × 20 cm, 250 μm; Analtech) developed three times with n-hexane/diethyl ether (90:15, v/v) and benzene. The composition ratio of MA subclasses was evaluated using the ImageJ software (https://imagej.net/ij//index.html). Each MA subclass was purified from TLC, and the molecular species of each MA subclass was analyzed by Matrix-assisted laser desorption/ionization time-of-flight mass-spectrometry (MALDI-TOF/MS. ultraflexXtreme, Bruker Daltonics Tokyo, Japan) as previously described^18,19.

Phylogenetic and comparative genomic analysis

Extraction of genomic DNA and whole-genome sequencing by MiSeq (Illumina Inc., San Diego, CA, USA) and MinION (Oxford Nanopore Technologies, Oxford, UK) were conducted as previously described⁸. A phylogenetic tree was then constructed from 16S rRNA gene sequences of reference species downloaded from the NCBI GenBank database, with multiple alignments produced using MAFFT v7.450²⁰. Subsequently, a maximum-likelihood (ML) phylogenetic tree was constructed under estimation by the GTRGAMMA model using the RAxML v8.2.12 software with 100 bootstrap replicates²¹. The average nucleotide identity (ANI) value was calculated with FastANI v1.34²² against genomes of the genus Mycobacterium downloaded from the NCBI RefSeq database (downloaded in February 2024). A phylogenetic tree was visualized using Figtree v1.4 (http://tree.bio.ed.ac.uk/software/figtree/).

Comparative genomic analyses were conducted to identify genetic features with the potential to contribute to the differences in physiological and biochemical characteristics found between closely related species. To accomplish this, the genomes of TUM20983, TUM20984, TUM20985^T, M. hodleri JCM 12141^T, and M. madagascariense JCM 13574^T were reannotated using the DDBJ Fast Annotation and Submission Tool (DFAST) v1.2.18 with the default parameters²³. To group orthologous protein sequences, pan-genome analysis was conducted using the Build Pangenome with OrthoMCL (v2.0)^24,25 which is available on the KBase server (https://www.kbase.us/)²⁶. Upset plots to visualize core, accessory, and singleton genes were created using the ComplexUpset package v1.3.3 (https://CRAN.R-project.org/package=ComplexUpset) in R v3.6.3. In addition, functional pathway reconstruction was performed using KEGG Mapper (https://www.genome.jp/kegg/mapper.html). Comparative metabolic profiling of the genomes was conducted using the Distilled and Refined Annotation of Metabolism (DRAM)²⁷ tool, implemented on the KBase server²⁸ and run with default parameters. This enabled the identification of strain-specific and shared metabolic features. Gene arrangement and synteny were visualized using the genoPlotR package v0.8.11 in R²⁹.

Results and discussion

Genome analysis overview and sequence-based species identification

The basic genomic features of the Antarctic strains determined from their genome sequences are summarized in Table S1. The strains possessed identical 16S rRNA gene sequences; however, a BLASTn search against the NCBI nr/nt database (last accessed in May 2025) showed that the closest match was Mycobacterium sp. strain AT1 (accession number: KY649294.1), which was isolated from Colne estuary water and sediment in UK, with a sequence identity of 98.22%.

A maximum likelihood tree based on the 16S rRNA gene showed that M. hodleri JCM 12141^T was the closest type strain supported by a bootstrap value of 93% (Fig. 1a). However, the highest ANI value among the relatives of Antarctic strains was found in M. hodleri JCM 12141^T, with an ANI value of only 83.5%, indicating a significant genetic distance. This was followed by M. madagacariense JCM 13574^T with an ANI value of 82.2% (Fig. 1b). Additionally, the ANI values between Antarctic strains were greater than 99%, suggesting they belong to the same species. Based on these results, we further evaluated the morphological, physiological, and chemotaxonomic characteristics of Antarctic strains by comparing them with these two reference strains.

Fig. 1

(a) Maximum-likelihood tree based on 16S rRNA gene sequences of strain TUM20983, TUM20984, and TUM20985^T and members of the genus Mycobacterium. Bootstrap values (expressed as percentages of 100 replications). Values above 50% are shown at the branch points. Nocardia abscessus DSM 44432^T was used as an outgroup. The scale bar indicates the branch length value. (b) The ANI value matrix is shown as a heat map. ANI values of reference strains, which showed particularly high similarity with the three strains, are shown in bold.

Physiological and morphological characteristics

TUM20983, TUM20984, and TUM20985^T grew at temperatures between 4 °C and 25 °C, but failed to grow at above 30 °C. Additionally, strains grew well on TSA and R2A medium while showing slightly weaker growth on 7H10 medium supplemented with OADC. TUM20983, TUM20984, and TUM20985^T formed smooth, orange colonies on TSA medium. Three strains tolerated up to 2% NaCl and pH values between 6.0 and 8.0. The growth curves at 25 °C were assessed in TSB medium based on OD₆₀₀ measurements (Figure S1). The doubling times of TUM20983, TUM20984, and TUM20985^T were estimated to be 38.8, 34.7, and 42.5 h, respectively, whereas that of the reference strain M. madagascariense JCM 13574^T was considerably shorter, at 13.2 h. For M. hodleri JCM 12141^T, the doubling time could not be determined due to sedimentation interfering with OD₆₀₀ measurement.

In terms of physiological traits, all three Antarctic strains were positive for catalase activity, but negative for oxidase and Tween 80 hydrolysis. All reactions in the API 20E (bioMérieux), including gelatinase and nitrate reduction, were negative. However, nitrate reduction was reevaluated using a tube-based method¹⁴ and yielded a positive result, suggesting that the activity may not have been detectable under the API 20E test conditions. Enzymatic profiling using API ZYM (bioMérieux) revealed multiple enzyme activities in the Antarctic strains, including alkaline phosphatase, esterase (C4), esterase lipase (C8), leucine arylamidase, valyl arylamidase, cystinyl aminopeptidase, acid phosphatase, naphthol-AS-BI-phosphohydrolase, α-glucosidase, and β-glucosidase. The physiological and biochemical test results are summarized in Table 1 and Table S2.

Table 1 Physiological and biochemical characteristics of strains TUM20983, TUM20984, TUM20985^T and their closest relatives.

Morphological analyses showed that the Antarctic strains were gram-stain-positive, rod shaped, non-spore forming bacteria. TEM images of the strains grown at 25 °C on TSA medium showed that they were approximately 1.6 μm in length and 0.3 μm in diameter. The strains differed from their relatives in that they retained numerous internal vacuole-like structures (Fig. 2). Similar structures have been observed in several other Mycobacterium species and identified as lipid droplets³⁰. Such droplets, which are excess lipids stored in the cytoplasm, have been observed in some actinomycetes, including Mycobacterium sp. such as M. tuberculosis and M. smegmatis. In a previous study of M. tuberculosis, lipid droplets were suggested to be biomarkers of slow growth, low energy state and persistence in restricted environments³¹. In general, lipid droplets contribute to cell growth, cell division and stress response, while also serving as stored energy for survival³². Whether the structures observed in the Antarctic isolates are lipid droplets has yet to be confirmed, and further research is needed to understand the relationship of these structures to adaptation to Antarctic environmental conditions.

Fig. 2

Representative TEM images of the cells of Antarctic strains (TUM20983, TUM20984, and TUM20985^T) with their close relatives, M. hodleri JCM 12141^T and M. madagascariense JCM 13574^T. Arrows indicate vacuole-like structures found in Antarctic strains, which are not observed in the reference strains. Each scale bar indicates 200 nm.

Chemotaxonomic characteristics

The G+C content of the genomes of strains TUM20983, TUM20984, and TUM20985^T were 67.5%, 67.3%, and 67.3%, respectively (Table S1). The major respiratory quinone of TUM20985^T was MK-9(H2) (100%), a common feature among Mycobacterium spp.³³. The MA subclasses of TUM20983, TUM20984 and TUM20985^T were composed of α-MA, keto-MA, and wax (dicarboxy)-MA, with patterns identical to those observed with M. hodleri JCM 12141^T and M. madagascariense JCM 13574^T (Fig. 3, Figure S2, Table S3). The major molecular species of each MA subclass in Antarctic strains were α-MA; C_74:2, C_75:2, keto-MA; C_77:0, C_78:0, ω-dicarboxy-MA; C_56:1, C_58:1, and C_59:1 (Figure S3, Table S4). As shown in Table 2, the dominant cellular fatty acids of the three strains were distinguishable from those of their relatives. The three strains had dominant compositions of C_18:1ω9c and C_17:1ω8c and tended to have less saturated fatty acids than their relatives, with C_17:1ω8c being a distinctive fatty acid that was not observed in the reference strains (Table 2). The species with the highest similarity in the MIDI’s TSBA library database was Dietzia maris, but the similarity index was low (< 0.268). Regulation of cell membrane fluidity is generally one of the main strategies employed by cold-adapted microorganisms. The composition of membrane lipids affects antifreeze ability, with branched-chain and unsaturated fatty acids playing crucial roles in maintenance of membrane fluidity and enabling adaptation to low temperature³⁴. The distinct characteristics of the fatty acid composition found in Antarctic strains are consistent with their psychrotolerant growth profiles, supporting their adaptation to cold environments.

Fig. 3

TLC profiles of mycolic acids methyl esters of Mycobacterium sp. TUM20983, TUM20984, and TUM20985^T compared with reference strains. Mycolic acid fractions of each strain were methylated and separated on TLC plates developed after (a) three runs in hexane: diethyl ether (90:15 by volume) and (b) three runs in benzene. The evaluated strains were as follows: M. tuberculosis ATCC 35801 (lane 1), M. avium ATCC 25291^T (lane 2), M. smegmatis mc²155 (lane 3), M. hodleri JCM 12141^T (lane 4), M. madagascariense JCM 13574^T (lane 5), TUM20983 (lane 6), TUM20984 (lane 7), and TUM20985^T (lane 8).

Table 2 Whole cell fatty acid composition of strains TUM20983, TUM20984, TUM20985^T and their closest relatives.

Comparative genomic analysis

We conducted comparative genomic analyses to search for genetic features that may contribute to the differences in physiological and biochemical characteristics found between closely related species. Basic genomic features are summarized in Table S1. Based on comparison to the Clusters of Orthologous Group (COG) functional annotation, a total of 3,179, 3,220, and 3,303 protein-coding genes were assigned to COG functional categories for TUM20983, TUM20984 and TUM20985^T, respectively (Table S5). These Antarctic strains showed a high proportion of lipid transport and metabolism (I), transcription (K), amino acid transport and metabolism (E), carbohydrate transport and metabolism (G), and energy production and conversion (C), with the same trends observed in the reference strains. However, the Antarctic strains had a lower proportion of transcription (K) and higher proportion of inorganic ion transport and metabolism (P) and signal transduction mechanism (T) than the reference strains (Table S5). Pan-genome analysis of TUM20983, TUM20984, TUM20985^T, M. hodleri JCM 12141^T, and M. madagascariense JCM 13574^T revealed 3,476 clusters assigned to core genes and 788 clusters assigned to accessory genes shared among all three Antarctic strains. Among the Antarctic strains, 46 clusters were shared between TUM20983 and TUM20984, 244 clusters between TUM20984 and TUM20985^T, and 68 clusters between TUM20983 and TUM20985^T. Additionally, there were 235, 192, and 167 singletons observed for TUM20983, TUM20984, and TUM20985^T, respectively (Fig. 4a).

Fig. 4

(a) An UpSet plot representing the results of a pangenome analysis. Horizontal bars on the left represent the total number of genes annotated in the genome of each strain. The vertical bars represent core, accessory, and singleton genes, as indicated by the connected circles. (b) Heatmap showing the distribution of COG categories among gene clusters shared by subsets of Antarctic strains. The horizontal axis indicates COG functional categories, and the vertical axis denotes the strain combinations sharing each gene cluster. The color gradient represents the number of gene clusters annotated to each category. Only gene clusters annotated to COG categories are included. (c) The graph represents the number of gene clusters annotated to the same COG ID, showing COG IDs with two or more gene clusters assigned among accessory genes of Antarctic strains.

The core genome included a variety of genes associated with environmental stress adaptation. These included cold shock-related genes such as cspA (TUM20985_08920, TUM20985_32900), cspB (TUM20985_19550), dnaK (TUM20985_06990), glpK (TUM20985_06330), groES (TUM20985_46020) and groEL (TUM20985_46010, TUM20985_50950), as well as many genes involved in host interaction such as mammalian cell entry family proteins and components of the ESX-1 type VII secretion system. The presence of these genes in the core genome indicates conserved host interaction mechanisms among their close relatives. However, the natural hosts of the Antarctic strains remain unknown, and further studies are needed to clarify their behavior within hosts under low temperature conditions.

The 788 accessory gene clusters shared only among Antarctic strains included 303 clusters annotated to COG categories. A large proportion of these genes belonged to categories related to lipid transport and metabolism (I), secondary metabolite biosynthesis, transport, and catabolism (Q), and general function prediction only (R). Notably, functional variation was observed even within the Antarctic strains, indicating intra-species metabolic diversity (Fig. 4b). Among these, genes annotated as NAD(P)-dependent dehydrogenases (COG1028) were the most abundant, comprising 24 clusters, which was followed by DNA-binding regulators (COG1309), with 12 clusters (Fig. 4c).

Genome-based metabolic profiling

Based on the pangenome results, each metabolic pathway of the Antarctic strains was further explored through genome-based comparisons with the reference strains.

Genes involved in MA biosynthesis were highly conserved and were identified as part of the core genome of the compared strains. This included meromycolic acid 3-oxoacyl-acyl-carrier protein reductase, FabG1 (also known as MabA; KEGG Orthology [KO]: K00079), a key enzyme in the FAS II system responsible for generating precursors for mycolic acid synthesis (Table S6). In addition to FabG1, multiple genes were annotated as FabG family members (KO: K00059), which was part of the genes annotated as COG1028. This family exhibited varying strain-level conservation, with some shared among all strains and, others specific to the Antarctic strains, and one gene unique to TUM20985^T (Table S7). This result suggests potential functional diversification both between and within species. Several FabG homologs (FabG1–FabG5) have also been identified in Mycobacterium tuberculosis³⁷ however, their physiological roles remain unclear in this genus. Studies in M. tuberculosis indicate that FabG4 can become essential under stress conditions, such as nutrient limitation³⁸ or antibiotic exposure³⁹ suggesting condition-dependent regulation among FabG homologs. The characteristic of possessing distinct fatty acid metabolism-related genes is similar to the situation we previously reported in Legionella antarctica, which possesses distinct fab family genes compared with its mesophilic related species⁴⁰ suggesting that this may be an important features observed in Antarctic cold-tolerant bacteria. Further evaluation of the expression of these genes in Antarctic strains may provide further insight into their fatty acid metabolism and mechanisms of cold adaptation.

Other enzymes involved in metabolic pathway such as carbon utilization, energy metabolism, transport, and nitrogen metabolism were annotated using DRAM²⁷. The major differences between the Antarctic strains and reference strains were the presence of genes involved in nitrogen metabolism and methane or ammonia oxidation, which were conserved among the Antarctic strains but absent in the reference strains (Fig. 5a). These gene clusters were found to be organized as operons.

In the case of the nitrate reductase operon (narKGHIJ), multiple transposase genes were located adjacent to the cluster, suggesting that this operon may have been acquired through horizontal gene transfer. Although the insertion sequences exhibited structural variation among the three Antarctic strains, the nitrate reductase gene region itself was conserved (Fig. 5b). Nitrate reductase activity has been detected in some Mycobacterium spp. and plays a key role in survival under hypoxic or nitrogen-limited conditions, such as latent infection or in nutrient-poor environments⁴¹. Among the Antarctic strains, conservation of the nitrate reductase operon may also suggest functional importance. Indeed, all three strains tested positive in nitrate reduction assays, supporting this enzymatic activity. The ability to reduce nitrate may offer a selective advantage in oligotrophic Antarctic soils where nitrogen is limited.

Fig. 5

(a) Heatmap showing completeness (left panel) or presence/absence (right panel) of metabolic pathways annotated by DRAM across five strains. Only pathways that were present in at least one genome are included. (b) Genomic structure of nitrogen metabolism-related gene clusters uniquely present in the Antarctic strains but absent from closely related species. The gene cluster includes narK (nitrate/nitrite transporter), narG, narH, and narI (subunits of nitrate reductase/nitrite oxidoreductase), and narJ (nitrate reductase molybdenum cofactor assembly chaperone), with adjacent tnp (transposase) and int (integrase) elements suggesting possible mobility. Arrows represent individual genes, and gray blocks indicate conserved gene arrangements among the strains. (c) Gene clusters associated with methane or ammonia oxidation conserved across the Antarctic strains. This region includes amo/pmoB, amo/pmoA, and amo/pmoC (ammonia/methane monooxygenase subunits), SIP (siderophore-interacting protein), adhD (aldehyde dehydrogenase D), adh (alcohol dehydrogenase), and hp (hypothetical protein). As in (b), arrows represent genes, and gray shading denotes conserved synteny among the three strains.

Similarly, a set of genes putatively involved in methane or ammonia oxidation was conserved across all three strains (Fig. 5c). In particular, the ammonia/methane monooxygenase-like genes (amo/pmoCAB) were located in a conserved syntenic arrangement with adjacent genes encoding a putative alcohol dehydrogenase and a siderophore-interacting protein. No mobile gene elements were detected near this cluster, however, its absence in related species suggests that these genes may also have been acquired through horizontal gene transfer as part of adaptive evolution. For the monooxygenase-like gene cluster, enzymatic activity has not yet been demonstrated, but its conserved presence suggests a potential role in the oxidation of small molecules such as methane or ammonia, which may also be advantageous for survival in nutrient-poor environments. Notably, a methane-oxidizing Mycobacterium species (Ca. Mycobacterium methanotrophicum) was recently identified from a volcanic cave⁴² expanding the known metabolic capabilities of actinobacteria in extreme environments. The presence of monooxygenase-like genes in the Antarctic strains may reflect ecological adaptation to nutritionally challenging environments. Nevertheless, further detailed bioinformatic and experimental analyses are required to confirm their function.

To assess the consistency between genome-based annotations and results obtained from chemical and physiological tests, we evaluated the correspondence between predicted enzyme-coding genes and observed phenotypic activities (Table S2). Several enzymatic activities, such as α- and β-glucosidases, catalase, and nitrate reductase, were consistent with genome-based predictions, whereas other results were incongruent. Some enzymes detected in these assays could not be linked to annotated genes with corresponding EC numbers, and conversely, certain metabolic genes present in the genome did not show detectable activity under the test conditions. In some cases, this discrepancy may be attributable to limitations in the sensitivity of the test systems or to the conditional expression of genes that require specific environmental or physiological cues for activation. Further investigation, including expression analyses under varied culture conditions, will be necessary to clarify the functional relevance of these genome-encoded enzymes.

Proposal of a novel species for strain TUM20985^T, TUM20983 and TUM20984

Our Antarctic strains (TUM20983, TUM20984, and TUM20985^T) isolated from Antarctic soil showed low 16S rRNA gene sequence identities (< 98.2%) and ANI values (< 85%) when compared with the closely related type strains belonging to the genus Mycobacterium. Phenotypic differences were also observed when compared with the closest strains, M. hodleri JCM 12141^T and M. madagascariense JCM 13574^T. Antarctic strains were found to have different growth temperature preferences (4 °C to 25 °C) and distinctive fatty acid profiles dominated by unsaturated fatty acids (C_18:1ω9c, C_17:1ω8c). The common features of Antarctic strains and other members of the genus Mycobacterium were a similar GC content (67%) and use of the same respiratory quinone (MK-9(H2)). Antarctic strains shared identical 16S rRNA gene sequences and showed high ANI values (> 98.9%) among themselves, along with similar growth temperature preferences and fatty acid compositions, indicating that they represent the same species. The growth temperature and habitat preferences suggest that these strains are well-adapted to the Antarctic environment, although further studies are needed to fully understand their cold-adaptation mechanisms. Based on these distinct phenotypic features and the phylogenetic placement, we propose the name Mycobacterium antarcticum sp. nov. for strains TUM20985^T, TUM20983 and TUM20984.

Description of Mycobacterium antarcticum sp. nov.

Mycobacterium antarcticum (ant.arc’ti.cum. L. neut. adj. antarcticum, southern, Antarctic, where the type strain was isolated).

Cells are gram-positive, rod-shaped bacilli. Grows to maturity in 7–10 days at 25 °C, pH 7.0. Colonies on 7H10 medium, R2A medium and TSA medium are orange, smooth and scotochromogenic. Growth occurs in the presence of up to 0.5% NaCl, pH 6–8 and 4–25 °C (optimum 25 °C). Strains were positive for catalase and nitrate reductase, but negative for urease, gelatin hydrolysis, oxidase and Tween 80 hydrolysis. Strains possess MA (major molecular species; C_74:2, C_75:2), keto-MA (C_77:0, C_78:0), and wax (dicarboxy)-MA (ω-dicarboxy-MA; C_56:1, C_58:1, C_59:1) subclasses, with the major fatty acids (> 15%) being C_18:1ω9c and C_17:1ω8c. The major respiratory quinones are MK-9(H2). Genome based phylogenetic analyses indicate that the species are related to Mycobacterium hodleri JCM 12141^T.

The type strain is TUM20985^T (= RIMD 4000203^T = NCTC 15058^T) isolated from the soil of the Skallen ice-free area, East Antarctica.