Introduction

The genus Microbacterium is part of the family Microbacteriaceae, under the order Micrococcales. Originally, these bacteria were grouped with the genus Corynebacterium because of their similar morphological traits. However, in 1983, Collins et al. reclassified them into the genus Microbacterium following in-depth biochemical, genetic, and chemotaxonomic studies1. This genus has since expanded to include over 160 validated species, known for their broad ecological distribution and metabolic versatility2. Notably, our group has recently contributed to this growing diversity by characterizing four novel Microbacterium species from NASA’s Phoenix mission cleanroom facilities3 and one from the International Space Station (ISS)4.

Microbacterium are Gram-staining-positive, non-spore-forming rods, characterized by a high degree of physiological and biochemical diversity. Microbacterium species are primarily environmental organisms found in soils, aquatic environments, plant tissues, and decaying organic matter. They are widely distributed across terrestrial and marine ecosystems and are known to inhabit both moderate and extreme environments. Their ability to colonize such varied habitats is due to their versatile metabolism, allowing them to degrade a wide variety of organic compounds5. Some species also exhibit plant-associated behavior, contributing to plant growth promotion and acting as biocontrol agents6.

Microbacterium species are characterized by their high GC-content in DNA and the presence of L-lysine, ornithine and/or L-homoserine as distinct diamino acids in the cell wall4,7. Many species are facultative anaerobes, able to thrive in both oxygen-rich and oxygen-poor environments. Additionally, Microbacterium species can metabolize a wide range of carbon sources, which supports their survival in diverse ecological niches8. Their resilience extends to resistance against desiccation, radiation, and various chemical stressors, making them suitable for potential applications in biotechnology, such as bioremediation9.

While primarily known as environmental bacteria, some Microbacterium species are opportunistic pathogens in both humans and plants. Although infections caused by Microbacterium in humans are relatively rare, they have been reported in immunocompromised individuals, where the bacteria can cause bacteremia, endocarditis, and wound infections10. Notable species implicated in human infections include M. arborescens and M. paraoxydans, particularly in patients undergoing medical treatments such as dialysis or indwelling catheterization11. These cases, though infrequent, indicate their potential to act as opportunistic pathogens in clinical settings.

Certain Microbacterium species have also been identified as plant pathogens. For example, M. testaceum has been associated with galls and other disease symptoms in plants, particularly in agricultural settings. This genus can sometimes interfere with plant health, despite the more common association with plant growth promotion12.

ISS is a one-of-a-kind, isolated, built environment that orbits Earth in low-Earth orbit. As a closed system where air, water, and surfaces are recycled and reused, the ISS provides a unique opportunity to study the effects of long-term space habitation on both human health and the built environment13. Since 2014, ongoing microbial tracking studies have been conducted aboard the ISS, aimed at monitoring the microbial communities within this controlled ecosystem14. The Microbial Tracking series, particularly Microbial Tracking-2, has played a crucial role in collecting and analyzing environmental and human samples from the ISS15,16.

As part of the Microbial Tracking-2 study, four strains designated as F6_8S_P_2BT, F6_8S_P_3A, F6_8S_P_3B, and F6_8S_P_3C belonging to the family Microbacteriaceae were isolated from the ISS. The objectives of this study were to fully characterize these strains, generate whole genome sequences (WGS), and define the phenotypic and phylogenetic novelty of these ISS-derived Microbacterium strains using a polyphasic taxonomic approach. The WGS generated and annotated in this study were further analyzed to predict genetic determinants to explore functional traits. This research underscores the importance of understanding microbial adaptation and its implications for prolonged human missions in space.

Materials and methods

Sample collection and isolation

The surface samples (1 m2) were collected from the ISS during Microbial Tracking–2 flight experiments from 2017 to 2018. The methods for gathering samples, processing them, and isolating cultivable microorganisms have already been described and published14. The samples collected by the polyester wipes were transported to Earth for microbiological processing, then suspended in sterile phosphate-buffered saline (pH 7.4) solution and concentrated before plating onto R2A agar medium. The microbial cultures that were grown at 25 °C for 7 days were picked from the R2A plates, purified, and stored for further analysis. During this study, distinct colonies (n = 4) isolated from the outside wall of the crew sleeping quarters were characterized14. All four strains were isolated during Flight 6 (April 2018) at crew quarters in node 216. These ISS strains are designated as F6_8S_P_2BT, F6_8S_P_3A, F6_8S_P_3B and F6_8S_P_3C.

DNA extraction and whole-genome sequencing

Genomic DNA was extracted from 1 mg of biomass collected from each of the four Microbacterium strains grown on R2A agar at 25 °C for 3 days. DNA extraction was carried out using the ZymoBIOMICS 96 MagBead DNA kit (Zymo Research, USA), following the manufacturer’s protocol. WGS libraries were prepared using the Illumina Nextera DNA Flex kit (Illumina document #1,000000025416 v07) and sequenced on an Illumina NovaSeq 6000 S4 platform with a paired-end 2 × 150 bp configuration.

For all strains, raw reads were filtered using NGS QC Toolkit v2.317 to remove low-quality and adapter-contaminated sequences (minimum 80% read length and quality score cutoff of 20). For strain F6_8S_P_2BT, hybrid genome assembly was performed using both Illumina paired-end and Oxford Nanopore single-end reads. Nanopore data were used to polish the Illumina assembly using the Unicycler assembler18 under default parameters within the PATRIC BV-BRC genome assembly pipeline19, with trimming and normalization options enabled.

Assembly quality was evaluated using QUAST v5.0.220. Genome annotation was conducted using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) v4.11 and Rapid Annotation using Subsystems Technology (RAST)21.

Phylogenetic analysis

The phylogenetic analysis was performed using the 16S rRNA gene and seven housekeeping genes (to generate the concatenated phylogenetic tree), the genes being: ATP synthase F1 beta subunit (atpD), DNA strand exchange and recombination gene (recA), chaperone gene (dnaK), DNA-directed RNA polymerase subunit beta (rpoB), glutamine synthetase (glnA), polyphosphate kinase (PPK) and DNA gyrase subunit B (gyrB), for different species of Microbacterium. For the phylogenetic tree using the 16S rRNA gene, 48 Microbacterium species (in addition to the four ISS strains) were included and also six outgroups (for all phylogenetic trees): Pseudoclavibacter chungangensis, Agrococcus sediminis, Subtercola frigoramans, Clavibacter phaseoli, Leucobacter triazinivorans and Agromyces aureus. The outgroups are in the class Actinomycetes, order Micrococcales and family Microbacteriaceae. All gene sequences for the different Microbacterium species and outgroups were taken from the NCBI, except for the four ISS strains. For the phylogenetic trees of housekeeping genes and the concatenated phylogenetic tree, 32 species were analyzed (among the 48 that were analyzed by the 16S rRNA gene), in addition to the 4 strains taken from the ISS and the same 6 previous outgroups. To determine which species would be analyzed and compared with the four ISS strains, the sequences were aligned to the NCBI nucleotide database via blast search, through the 16S rRNA gene, and then the top 48 matches were chosen.

For the 16S rRNA gene tree and all other housekeeping genes, alignment was performed by R using the DECIPHER package22. Trimming and tree-building were performed in the IQ-TREE software23 using maximum likelihood and the substitution model GTR+I+G, which was selected through the search for the best model (MFP) and 1000 were included bootstrap replicates. A genome-based phylogenetic tree was constructed with all 194 species of the genus Microbacterium and the 4 strains. The assembly was performed by the BV-BRC platform (v3.42.3) using the Bacterial Genome Tree tool24. Phylogenetic trees were visualized using the Interactive Tree of Life (iTOL v5)25.

Average Nucleotide Identity (ANI), Average Amino Acid Identity (AAI), and digital DNA–DNA hybridization (dDDH) values were calculated to assess the phylogenomic relatedness between the four ISS isolates and the top 26 Microbacterium species with publicly available genome sequences. ANI values were computed using FastANI v1.3326, AAI values were estimated using EzAAI—High Throughput Prokaryotic AAI Calculator27, and dDDH values were determined using the Genome-to-Genome Distance Calculator (GGDC 2.1) v3.0 available through DSMZ28.

Pangenomic analysis

The pangenomic analysis was performed with nine Microbacterium genomes and four ISS strains: M. oxydans DSM 20578T, M. paraoxydans DSM 15019T, M. aurum DSM 8600T, M. thalassium DSM 12511T, M. profundi Shh49T, M. saperdae DSM 20169T, M. esteraromaticum DSM 8609T, M. keratanolyticum DSM 8606T, and Microbacterium sp. LTR1, using anvi’o v7.129. The annotation for the pangenome was generated with KEGG, Pfam, COGs, and KOfam database functions/pathways. Two pangenomes were generated, one with the 13 genomes and the second with only the four ISS strains and the Microbacterium species closest to the strains (M. paraoxydans DSM 15019T and Microbacterium sp. LTR1).

A phylogenetic tree using single-copy core genes of the pangenome was also generated. The core region with genes shared between all species and strains, with geometric homogeneity of 1 and a functional homogeneity below 0.9, were extracted and concatenated in a fasta file. After that, the gene sequences were aligned using DECIPHER package22 and the phylogenetic tree was made using IQ-TREE software23 using maximum likelihood and the substitution model LG+F+I+G4, which was selected through the search for the best model (MFP) with bootstrap of 1,000 and visualized using Interactive Tree of life (iTOL v5)25.

Functional gene characterization

The genomes of the four ISS strains, Microbacterium sp. LTR1, and M. paraoxydans DSM 15019T were annotated using the RAST platform (Aziz et al.21). The antiSMASH pipeline30 was used to determine the biosynthetic gene clusters (BGCs) and the ABRicate pipeline tool31 was also used to verify the presence of antimicrobial resistance or virulence genes. In addition, the dbCAN3 tool32 was used to determine the active carbohydrate enzymes and the CAZy Family (active carbohydrate enzymes, i.e. CAZymes) for each of the four ISS strains, M. paraoxydans LTR1, and M. paraoxydans DSM 15019T. Two chart models (Pie chart and Barchart) were generated using the tool COGclassifier v1.0.5 (https://github.com/moshi4/COGclassifier) to visualize the results.

Biochemical and phenotypic analyses

Cells were harvested from freshly prepared TSA plates (~ 24 h old culture) and transferred aseptically into the saline tube (aqueous 0.45–0.50% NaCl, pH 4.5 to 7.0) to prepare homogenous suspension having density of McFarland No. 0.5 (verified through VITEK® 2 DensiCHEK™ Plus). Vitek 2 GP ID (bioMérieux) card was used for the biochemical tests according to the manufacturer’s protocol. The tube containing bacterial suspension along with Vitek 2 GP ID card was placed in the cassette and incubated at 37 °C. Cassette loading, sample data entry, and retrieval of raw data (within 10 h of inoculation) were done according to the VITEK instrument user manual.

Phenotypic fingerprinting was carried out for freshly grown cells on TSA using GNIII MicroPlate according to BioLog’s protocol. Cells were transferred aseptically into the inoculum solution A (Cat # 72401; BioLog) to reach a McFarland No. 0.50. Wells of BioLog GNIII MicroPlate were loaded with the bacterial suspension (100 µl per well) and incubated at 37 °C for 24 h. OmniLog values (A590-A750) were recorded using MicroPlate reader (FLUO star Omega, BMG Labtech, Germany) within 24 h of incubation.

Catalase activity was determined by monitoring the bubble production by cells in the presence of 3% (v/v) H2O2. Oxidase activity was determined by using 1% (w/v) N, N, N, N,-tetramethyl-1, 4-phenylenediamine reagent (bioMérieux). Motility was assessed as per hanging drop method33.

Cells were grown on TSA for 3 days at 30 °C for carotenoid extraction. The wet biomass (100 mg) was added to 1 ml of methanol, mixed thoroughly, and incubated overnight at 4 °C under darkness. The mixture was centrifuged (12,400 g, 5 min, 4 °C), filtered through Millipore filter paper (PVDF, 13 mm, 0.22 μm), and subjected to a full-wavelength scan (250–700 nm) using a UV–visible spectrophotometer (U3010; Hitachi).

Physiological analyses

Temperature tolerance and optimum growth temperature were evaluated by culturing the cells in TSB at a range of temperatures: 4, 10, 20, 27, 35, 40, 50, 70, and 100 °C. Salt tolerance and the optimal sodium chloride concentration required for growth were determined by cultivating the cells in R2A broth supplemented with NaCl concentrations of 0, 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 12, 15, and 18% (w/v). To assess pH tolerance and identify the optimal pH, the strains were grown in media adjusted to pH values ranging from 4.0 to 10.0, at 1.0-unit intervals, using appropriate buffer systems. Specifically, 0.1 M citric acid and 0.1 M trisodium citrate were used for pH 4.0–5.0, 0.2 M Na₂HPO₄ and 0.2 M NaH₂PO₄ for pH 6.0–8.0, and 0.1 M NaHCO₃ with 0.1 M Na₂CO₃ for pH 9.0–10.0. All pH values were verified and adjusted following media sterilization to ensure consistency across experiments.

Chemotaxonomy

Cells were grown on TSA at 30 °C for 48 h until they reached the mid-exponential growth phase. Fatty acid methyl esters (FAME) were harvested from cell biomass using sequential saponification, methylation, and extraction34. Separation of FAME was done using a gas chromatographic instrument (Agilent 7890A) fitted with a flame ionization detector. Microbial Identification System (MIDI)35 with the Aerobe (RTSBA6) database (Sherlock version 6.0) was used for FAME identification while following the standard protocol36. Cells of F6_8S_P_2BT were grown on TSA for 3 days at 30 °C for the extraction and analysis of polar lipids, quinones, and peptidoglycans. The polar lipids and quinones were extracted and analyzed according to Minnikin et al.37. Polar lipid spots were identified by spraying the TLC plates with the following reagents: 10% (w/v) ethanolic phosphomolybdic acid for total lipids, 0.2% (w/v) ninhydrin in butanol for aminolipids, Dittmer and Lester’s Zinzadze reagent for phospholipids, and alpha-naphthol for glycolipids. Peptidoglycans were extracted and analyzed according to Staneck and Roberts38. DNA base composition was calculated through genome sequence data.

Results and discussion

Among 17 strains that were isolated during Flight 6 at crew quarters exterior walls, the WGS analyses revealed the presence of seven bacterial and 10 fungal cultures. Fungal species belonging to Aspergillus unguis (n = 2) and Penicillium palitrans (n = 3), as well as a lone yeast Rhodotorula mucilaginosa (n = 5) were identified and published39. However, all seven bacterial strains that comprised three species were either novel as reported before (Leifsonia virtsii [n = 1] and Leifsonia williamsii [n = 2]40 or as Microbacterium sp. (n = 4) described during this study. It is noteworthy to mention the dominance of the fungal population (105 CFU/m2) in the exterior walls of the crew quarters during Flight #6 when compared to the bacterial communities (104 CFU/m2). Also, it was documented that the crew quarters walls contained more fungal population among the eight locations sampled16. The co-occurrence of these novel actinobacterial species with fungi and forming any biofilms on the ISS surfaces should be carefully examined in the future.

Genome characteristics

Assembly statistics for the four strains are summarized in Table 1. The genome size of all four ISS strains is 3.46 Mb, GC content of 70.03%. The average coverage ranges from 575 for F6_8S_P_3A, 618 for F6_8S_P_3B, 706 for F6_8S_P_3C, and 823 for F6_8S_P_2BT. Furthermore, the N50 (bp) is 2,737,578 for the F6_8S_P_2BT strain, very similar to 2,737,578 for the F6_8S_P_3A strain and 2,817,653 for the F6_8S_P_3C strain, whereas for the F6_8S_P_3B strain it is markedly lower at 495,349.

Table 1 Genomic characteristics of the Microbacterium meiriae draft genomes, including genome size, GC content, number of coding sequences, and N50 values.
Full size table

ANI comparisons between ISS strains and several publicly available Microbacterium genomes revealed that Microbacterium sp. LTR1 displayed the highest similarity to the ISS isolates, with ANI values of 98.26% (Table 2). The closest recognized and validated species was M. paraoxydans DSM 15019T, showing ANI values of 93.14% across all four ISS strains. In digital dDDH analyses (formula 2), Microbacerium sp. LTR1 again exhibited high relatedness with the ISS strains (83.50%), whereas M. paraoxydans DSM 15019T showed dDDH values ranging from 49.70 to 49.80% (Table 2), supporting delineation of the ISS strains as members of a distinct species.

Table 2 Pairwise comparisons of Microbacterium meiriae strains with closely related Microbacterium species based on average nucleotide identity (ANI), average amino acid identity (AAI), and digital DNA–DNA hybridization (dDDH) values.
Full size table

When comparing the genomes of the Microbacterium F6_8S_P_2BT strain with all available WGS of Microbacterium (n = 1264 genomes), 15 isolates and three metagenome-assembled genomes (MAGs) showed > 95% ANI similarity (Supplemental Table S1). Among these 18 genomes, one isolate from Russia (LTR1 strain; 98.3%) and another from Bangladesh (BHS25 strain; 96.1%) were misidentified as M. paraoxydans. Additional Microbacterium isolates were recovered from diverse environments, including washroom surfaces at Red Deer Polytechnic, Alberta, Canada (n = 5), Florida, USA (n = 4), human feces in Madison, USA (n = 2), as well as from water in Algeria (n = 1) and the feces of a long-tailed goral in South Korea (n = 1). These findings confirm that the novel Microbacterium species identified on the ISS is widespread in nature, highlighting its ubiquity across diverse geographical regions and environments.

Phylogeny of the ISS Microbacterium strains

The 16S rRNA gene-based phylogenetic tree revealed that all four ISS strains clustered within a distinct monophyletic clade (> 99% similarity), positioned near M. paraoxydans DSM 15019T, albeit with low branch support values (20 for F6_8S_P_2BT and F6_8S_P_3B; 10 for F6_8S_P_3C; 59 for F6_8S_P_3A) (Fig. 1). In contrast, the gyrB gene phylogeny (Fig. 2) provided strong support for the exclusive clade formed by the ISS strains (branch support 95), which grouped closely with Microbacterium sp. LTR1 and M. paraoxydans DSM 15019T. WGS-based phylogenetic analysis (Fig. 3 top) and concatenated gene trees (Fig. 3 bottom) similarly grouped the four ISS strains in a well-supported clade (100%), adjacent to Microbacterium sp. LTR1 and M. paraoxydans DSM 15019T. Comparable clustering patterns were observed in the phylogenetic reconstructions based on individual housekeeping genes (recA, atpD, dnaK, glnA, ppk, and rpoB) (Supplemental Fig. S1–S6). Additionally, a phylogenetic tree derived from 741 single-copy core genes across the four ISS strains and nine Microbacterium species (Supplemental Fig. S7) corroborated the distinct clustering of the ISS strains and their close affiliation with Microbacterium sp. LTR1, further supporting their status as a novel species closely related to, but distinct from, M. paraoxydans.

Fig. 1
figure 1

Maximum-likelihood phylogenetic tree based on 16S rRNA gene sequences showing the position of Microbacterium meiriae strains in a distinct clade, separate from M. paraoxydans DSM 15019T. Bootstrap values (≥ 50%) are shown at branch nodes. Gordonia terrae was used as an outgroup.

Full size image
Fig. 2
figure 2

Maximum-likelihood phylogenetic tree based on gyrB gene sequences, placing the Microbacterium meiriae strains in a well-supported clade closely affiliated with Microbacterium sp. LTR1 and M. paraoxydans DSM 15019T. Bootstrap support values are indicated at branch points.

Full size image
Fig. 3
figure 3

Genome-based phylogenetic tree reconstructed from the alignment of 194 Microbacterium genomes including the four novel ISS strains (top). The M. meiriae strains cluster together and form a distinct lineage within the genus. The concatenated tree (bottom) of six housekeeping genes (recA, atpD, dnaK, glnA, ppk, and rpoB) similarly grouped the four ISS strains in a well-supported clade (100%), adjacent to Microbacterium sp. LTR1 and M. paraoxydans DSM 15019T. Bootstrap support values are indicated at branch points.

Full size image

Comparative genomic analysis

The pangenomic analysis also shows a high correlation between the four ISS strains and Microbacterium sp. LTR1 with M. paraoxydans DSM 15019T. Furthermore, for the analyzed genomes, no redundancy was found and the number of gene clusters is 3289 (between the four strains and the two strains of M. paraoxydans) and 3691 (for the pangenome analysis that included all thirteen genomes), and GC content of 70.0 for both pangenomes. In addition, the ANI values calculated in the pangenome indicated the high proximity of the four strains to each other and to the M. paraoxydans species, and in the dendrogram, the four strains were also grouped together in a clade and close to the M. paraoxydans species clade (Fig. 4 and Supplemental Fig. S8).

Fig. 4
figure 4

Pangenome analysis comparing the Microbacterium meiriae strains with Microbacterium sp. LTR1, M. paraoxydans DSM 15019T, and additional Microbacterium species. The distribution of core, accessory, and strain-specific gene clusters is shown.

Full size image

Pangenome analysis of the four ISS strains, Microbacterium sp. LTR1, and M. paraoxydans DSM 15019T delineated four major genomic regions: a conserved core region, an ISS strain-specific region, and unique gene clusters attributed individually to M. paraoxydans DSM 15019T and Microbacterium sp. LTR1. In the first comparison involving the four ISS strains, M. paraoxydans DSM 15019T, and LTR1, the core genome comprised 973 gene clusters, while 182 gene clusters were found exclusively in the ISS strains. In the expanded pangenome including seven additional Microbacterium species, 2861 gene clusters constituted the shared core genome. In this broader comparison, 225 gene clusters were unique to the ISS strains, 331 were exclusive to M. paraoxydans DSM 15019T, and 138 were specific to Microbacterium sp. LTR1. These findings reinforce the distinct genomic identity of the ISS isolates and support their delineation as representatives of a novel species within the genus Microbacterium.

Unique proteins

The majority of proteins unique to the ISS strains were assigned to the “unknown function” category (COG category “*”), comprising 16 proteins, indicating that a substantial proportion of their exclusive gene content remains functionally uncharacterized. Additional exclusive proteins were distributed across categories “K” (Transcription), “V” (Defense mechanisms), and “E” (Amino acid transport and metabolism), with 2–3 proteins identified in each. Comparison with Microbacterium sp. LTR1 revealed minimal differences, suggesting close genomic relatedness (Fig. 5A). However, in the comparative analysis with M. paraoxydans DSM 15019T, the ISS strains displayed 55 proteins in the “unknown” category, along with enrichment in category “G” (carbohydrate transport and metabolism; 14 proteins), which may reflect metabolic adaptations to low-nutrient spaceflight environments (Fig. 5B). Moderate representation was also observed in categories “K” and “L” (replication, recombination, and repair). A subset of these unique proteins, outlined in Supplemental Table S2, was further examined to distinguish genomic features exclusive to the ISS strains relative to Microbacterium sp. LTR1.

Fig. 5
figure 5

Functional classification of proteins unique to Microbacterium meiriae. (A) Comparison with Microbacterium sp. LTR1 based on COG categories. (B) Comparison with M. paraoxydans DSM 15019T highlighting enrichment in genes associated with carbohydrate transport and metabolism.

Full size image

Toxins and toxin-antitoxin (TA) systems

We identified four proteins found in ISS strains (and not found in Microbacterium sp. LTR1) that are part of the Toxins and Toxin-Antitoxin (TA) Systems.TA systems consist of compact genetic modules that include a toxin gene paired with its corresponding antitoxin. While toxins in TA systems are universally proteins, antitoxins can either be proteins or non-coding RNAs41. TA systems contribute to the virulence of pathogenic bacteria42 and are also involved in bacterial stress resistance by influencing processes such as biofilm development and the formation of persister cells43. Studies suggest that microbes on the ISS, exhibit an enrichment of functions associated with metal ion tolerance, dormancy, and antibiotic resistance 44. Furthermore, several studies report the presence of TA system genes in microorganisms collected from the ISS, including the RelBE genes44,45. The antitoxin component of the TA stability system, DNA-binding transcriptional repressor Phd, is part of a TA system regulating cellular stability. Phd serves two distinct roles: it acts as a transcriptional repressor by binding to its operator and also prevents host cell death by binding to and neutralizing the associated toxin46. Similarly, the predicted antitoxin YuzE, belonging to the DUF2283 family, functions as an antitoxin component within TA systems (UniProt:O32096). The VapBC TA systems, composed of the toxin VapC and the antitoxin VapB, represent the largest family of type II TA modules47. Although the precise physiological roles of most type II TA systems remain unclear48, studies have demonstrated that VapC toxins are differentially expressed under stress conditions commonly faced by M. tuberculosis during infection49.

Transcriptional regulators

The YhcF protein, a DNA-binding transcriptional regulator of the GntR Family and part of the YtrA subfamily, is implicated in metabolic regulation. GntR regulators are likely involved in controlling a specific subset of genes essential for adaptation to diverse environmental conditions50. Similarly, FrmR, a member of the FrmR family, functions as a formaldehyde-responsive transcriptional regulator, modulating gene expression in response to environmental changes51. Formaldehyde is a highly reactive and toxic chemical naturally produced by various organisms. Its generation occurs under several conditions, including: (i) as an intermediate in methylotrophic metabolism; (ii) during glycine degradation, either via methylglyoxal, a byproduct of glycolysis, or through Fenton reactions; (iii) as a result of heme breakdown, particularly in iron acquisition processes of some Gram-staining-positive bacteria; (iv) through lipid peroxidation or sugar metabolism; (v) during histone demethylation; (vi) as a byproduct of AlkB-mediated repair of methylated DNA; and (vii) via the activity of the enzyme N-methyltryptophan oxidase, encoded by the solA gene52,53,54. Because of these diverse pathways, organisms are exposed to both internal and external sources of formaldehyde. This exposure is harmful because formaldehyde interacts with and modifies vital cellular molecules, such as DNA and proteins, which can lead to significant cellular damage and impaired function. The HipB regulator, containing an XRE-family HTH domain, is associated with transcriptional control within TA systems. HipB antitoxin functions to neutralize the toxin55.

Carbohydrate active enzymes (CAZy) and other analyses

The CAZy analysis revealed that the enzymatic profiles are identical across the four ISS strains. Comparative analysis with M. paraoxydans DSM 15019T and Microbacterium sp. LTR1 (Supplemental Table S6) confirmed this pattern. A pie chart generated using the COGclassifier v1.0.5 template illustrated that the dominant CAZy classes in both ISS strains and M. paraoxydans species are glycoside hydrolases (GH) and glycosyltransferases (GT), highlighting conserved functional roles in carbohydrate metabolism (Fig. 6). GH and GT families represent around 43 and 36% of the CAZyme content, respectively, across M. meiriae and M. paraoxydans species. Within the glycosyltransferase (GT) families, GT2 and GT4 stand out due to their highest abundance in both ISS strains and M. paraoxydans DSM 15019T. These two enzyme families play essential roles in cellular membrane and cell wall processes56. Notably, the GT2 family is involved in key functions such as cellulose synthase, chitin synthase, dolichyl-phosphate synthase, β-mannosyltransferase, β-glucosyltransferase, β-N-acetylglucosaminyltransferase, β-N-acetylgalactosaminyltransferase, hyaluronan synthase, chitin oligosaccharide synthase, and β-1,3-glucan synthase.57 Meanwhile, the GT4 family is associated with functions such as sucrose synthase, sucrose-phosphate synthase, α-glucosyltransferase, α-N-acetylglucosaminyltransferase, α-mannosyltransferase, 1,2-diacylglycerol, α-3-glucosyltransferase, and trehalose phosphorylase.57 Another notable aspect of the results is the prominence of the GH13 family (Supplemental Table S6), particularly the subfamilies GH13_11, GH13_30, and GH13_9, which are characterized by enzymes such as isoamylase, α-glucosidase, and 1,4-α-glucan branching enzyme, all of which are involved in carbohydrate metabolism58.

Fig. 6
figure 6

Distribution of carbohydrate-active enzyme (CAZy) families in Microbacterium meiriae, Microbacterium sp. LTR1, and M. paraoxydans DSM 15019T. The pie chart reveals a predominance of glycoside hydrolases and glycosyltransferases across all strains.

Full size image

The analysis using the RAST platform revealed that only 24% of genes from the ISS strains and Microbacterium sp. LTR1 were associated with defined subsystems, indicating a high proportion of uncharacterized genes. Similarly, 23% of genes from M. paraoxydans DSM 15019T were assigned to subsystems. Functional annotations showed that both the ISS strains and M. paraoxydans species possessed genes involved in auxin biosynthesis, as part of their secondary metabolism pathways. Additionally, resistance genes related to cobalt-zinc-cadmium and fluoroquinolones were identified, suggesting adaptation to metal stress and antimicrobial compounds.

The antiSMASH analysis (Supplemental Table S4) identified five secondary metabolite biosynthetic gene clusters in the ISS strains: one RRE-containing cluster, one Type III polyketide synthase (T3PKS), two beta-lactone clusters (potential protease inhibitors), and one terpene cluster. The same profiles were observed in M. paraoxydans DSM 15019T and Microbacterium sp. LTR1. Notably, the RRE-containing and beta-lactone clusters remain uncharacterized. The T3PKS and terpene clusters showed low similarity—7 and 28%, respectively—to reference clusters from Micromonospora sp. HK160111 and Brevibacterium linens, suggesting potentially novel biosynthetic pathways.

Using the ABRicate pipeline, antimicrobial resistance gene Tet 42—conferring tetracycline resistance—was identified in M. meiriae and M. paraoxydans species with 82% sequence identity (Supplemental Table S8). This finding suggests a shared resistance trait among these closely related taxa, potentially contributing to their environmental resilience and highlighting the importance of monitoring such resistance determinants in cleanroom-associated microbiota.

Phenotypic characteristics of ISS strains

The phenotypic characters that discriminated F6_8S_P_2BT from other strains are listed in Supplemental Table S10. ISS strain F6_8S_P_2BT harbored 15:0 anteiso (40.8%), 17:0 anteiso (30.0%), 16:0 iso (14.8%), and 15:0 iso (7.6%) as a major fatty acid component. Similarly, ISS strains and M. paraoxydans CF36T also consistently displayed the predominance of these fatty acids. However, ISS strains can be distinguished from each other as well as from M. paraoxydans CF36T based on minor qualitative and quantitative variations in terms of their FAME profiles as shown in Supplemental Table S11.

F6_8S_P_2BT contained diphosphatidylglycerol (DPG), phosphatidylglycerol (PG), and an unidentified glycolipid (GL) as polar lipid component. ISS strains also displayed DPG, PG, and PL in predominant amounts. In addition, these strains revealed the heterogeneous presence of several unidentified lipids (Supplemental Table S10; Supplemental Fig. S9). The presence of DPG, PG, and GL was also reported in M. paraoxydans CF36T37 and several other established species of Microbacterium34,35,38. F6_8S_P_2BT produced predominant amounts of MK-12 (54%), followed by MK-11 (33%), MK-13 (8%), and MK-10 (5%) as isoprenoid quinones. Marginal quantitative variations were observed in terms of MK-12 (54–55%), MK-11 (33–34%), MK-13 (8–9%) and MK-10 (3–5%) in F6_8S_P_3A, F6_8S_P_3B and F6_8S_P_3C. The menaquinone profile of F6_8S_P_2BT found to be in line with that of M. paraoxydans CF36T37.

F6_8S_P_2BT and other tested strains of ISS consistently accommodated B2β type peptidoglycan with ornithine as a diagnostic cell wall diamino acid. The cell wall is also found to possess glycine, alanine, glutamate, hydroxyglutamate and homoserine as reported earlier in M. paraoxydans CF36T. Identification of B2β type peptidoglycan with ornithine as a diagnostic diamino are in accordance with the M. paraoxydans CF36T37 and several other species of Microbacterium34,36,37. Taken together, these chemotaxonomic parameters are in excellent agreement with the data reported earlier for closely related Microbacterium species.

Description of Microbacterium meiriae sp. nov

Microbacterium meiriae (meir’i.ae. N.L. gen. n. meiriae, named in honor of NASA astronaut (Jessica U. Meir) who is a marine biologist from USA.

Cells are strictly aerobic, Gram-staining-positive, non-spore-forming, chemoheterotrophic and mesophilic rods with pale yellow pigmentation. Catalase-positive and oxidase-negative. Cells are non-flagellated, slender rods with rounded ends. Cells are 1.5 µm in length and 1 µm in width. On TSA, after 1–2 days of incubation at 30 °C, colonies are circular with regular margins, convex, 0.5–1.0 mm in diameter. Cell growth occurs at 4‒40 °C (35 °C optimum), 0.5‒8% NaCl (4‒5% optimum) and at pH 6.0–9.0 (optimum, pH 7.5). Carotenoid pigments extracted in methanol show λmax at 434 nm and shoulder peaks at 414 and 530 nm.

Based on the available data, M. meiriae exhibits distinct biochemical characteristics in both Biolog GN III and VITEK GP2 profiles, setting it apart from other Microbacterium species. In the Biolog GN III system, M. meiriae consistently tests positive for the oxidation of substrates such as dextrin, d-maltose, d-trehalose, d-cellobiose, sucrose, d-turanose, d-galactose, l-rhamnose, d-sorbitol, glycyl-L-proline, l-histidine, l-serine, galacturonic acid, l-galactonic acid lactone, d-gluconic acid, mucic acid, quinic acid, tetrazolium violet, tetrazolium blue, p-hydroxy-phenylacetic acid, d-lactic acid methyl ester, l-lactic acid, citric acid, α-keto-glutaric acid, d-malic acid, d-malic acid, lithium chloride, α-hydroxy-butyric acid, α-keto-butyric acid, acetoacetic acid, propionic acid, acetic acid, and sodium butyrate. It is negative for the oxidation of raffinose, fusidic acid, d-aspartic acid, d-serine, niaproof 4, bromo-succinic acid, and sodium bromate, with variable reactions for several other substrates.

In the VITEK GP2 card, M. meiriae shows positive reactions for arginine dihydrolase 1, α-glucosidase, leucine arylamidase, alanine arylamidase, tyrosine arylamidase, utilization of d-maltose, d-mannose, saccharose/sucrose, and d-trehalose. It is negative for the utilization of d-amygdalin, d-xylose, cyclodextrin, d-sorbitol, d-ribose, lactose and N-acetyl-D-glucosamine; negative for phosphatidylinositol phospholipase C, Ala-Phe-Pro arylamidase, l-aspartate arylamidase, β-galactopyranosidase, α-mannosidase, phosphatase, β-glucuronidase, α-galactosidase, L-pyrrolydonyl-arylamidase, arginine dihydrolase 2 and urease reactions; negative for resistance to polymyxin B, bacitracin, O/129 and novobiocin; negative for growth at 6.5% NaCl, methyl-β-D-glucopyranoside, pullulan, d-raffinose, salicin. Strains exhibited variable reactions for β-galactosidase, L-proline arylamidase, d-galactose utilization, l-lactate alkalization, d-mannitol utilization and optochin resistance. These profiles suggest that M. meiriae possesses a unique metabolic fingerprint, distinguishing it from other members of the Microbacterium genus.

M. meriae produce 15:0 iso, 15:0 anteiso, 16:0 iso and 17:0 anteiso as major (> 5% of the total) fatty acids; MK-10, MK-11, MK-12 and MK-13 as respiratory quinones; D-ornithine, glycine, alanine, glutamate, hydroxyglutamate and homoserine as cell wall amino acids; B2β as peptidoglycan with ornithine as the diagnostic cell-wall diamino acid. All four strains contained diphosphatidylglycerol (DPG), phosphatidylglycerol (PG), and an unidentified glycolipid (GL) as polar lipid component.

Genomic DNA G+C content is 70.03 mol%. The type strain is (F6_8S_P_2BT = DSM 115935T = NRRL B-65668T), isolated from the crew quarters of ISS and the WGS accession number is CP133984. Other three M. meiriae strains are designated as F6_8S_P_3A, F6_8S_P_3B and F6_8S_P_3C and all four strains were isolated from the crew quarters of ISS.

Conclusion

This study presents the comprehensive taxonomic and functional characterization of Microbacterium meiriae sp. nov., a novel bacterial species isolated from the International Space Station. Phylogenomic and pangenomic analyses confirmed its distinct lineage within the genus Microbacterium, closely related to but genetically separate from M. paraoxydans and Microbacterium sp. LTR1. The ISS strains displayed unique metabolic, genomic, and chemotaxonomic features, including exclusive proteins, stress adaptation genes, and secondary metabolite clusters. The presence of toxin-antitoxin systems, transcriptional regulators, and tetracycline resistance genes underscores the ecological resilience and potential clinical relevance of this new species.