Phylogenetic analysis
The assembled 16S rRNA gene amplicon of strain AK-R2A1-2T was 1457 base pairs (bp). Comparative sequence analysis showed that strain AK-R2A1-2T appeared to belong to the genus Subtercola, sharing the highest 16S rRNA gene sequence similarity with S. boreus DSM 13056T (97.3%), followed by 96.7% sequence similarity with Agreia bicolorata VKM Ac-1804T, Agreia pratensis VKM Ac-2510T, and S. lobariae KCTC 33586T. Sequence similarities with the other two members of the genus Subtercola were 96.2% and 96.1%. Strain AK-R2A1-2T was regarded as a novel species of the genus Subtercola based on the novel species recognition threshold value of 98.6%8. To identify the phylogenetic location of strain AK-R2A1-2T, phylogenetic trees based on 16S rRNA gene sequences were reconstructed using the MEGA program. The NJ phylogenetic tree showed that strain AK-R2A1-2T formed a cluster with S. boreus DSM 13056T, and all members of the genus Subtercola comprised a single group. The ML and MP phylogenetic trees supported this result (Fig. 1). Thus, strain AK-R2A1-2T was confirmed as a member of the genus Subtercola rather than a member of the genus Agreia. Based on the 16S rRNA gene sequence similarity and phylogenetic analysis, strains S. boreus DSM 13056T, S. vilae DSM 105013T, S. frigoramans KCTC 49696T, and S. lobariae KCTC 33586T were selected for further comparative study under the same conditions.
Phenotypic characteristics
Phenotypic characteristics including optimal growth medium, temperature, pH, and salt tolerance were investigated. Strain AK-R2A1-2T grew well on R2A, PD, MEA, YEP, ISP2, and GYM (optimum growth on PD and R2A), but did not grow on MA, LB, NA, or TSA. Cells of strain AK-R2A1-2T were short rod-shaped without flagella (0.2–0.3 µm in width and 0.3–1.6 µm in length, Fig. S1), and were Gram-stain-positive, non-motile, catalase-positive, and oxidase-negative. Growth of strain AK-R2A1-2T was observed on R2A broth medium at 4–25°C (optimal 20°C), at pH 5–8 (optimal 5), and with 0–1% NaCl (optimal 0%). Differential physiological characteristics of strain AK-R2A1-2T that distinguished it from closely related strains are shown in Table 1.
Chemotaxonomy features
The cellular fatty acid profiles (>1%) of strain AK-R2A1-2T and closely related strains are shown in Table 2. The major fatty acids (>10%) of strain AK-R2A1-2T were anteiso-C15:0 (12.3%) and summed feature 8 (76.7%). This was most similar to S. lobariae KCTC 33586T, which also contained anteiso-C15:0 (25.8%) and summed feature 8 (52.1%) as the major fatty acids. The major fatty acids of S. boreus DSM 13056T were anteiso-C15:0 (46.0%), iso-C15:0 (20.7%), 2-OH C14:0 (10.4%), and summed feature 8 (11.8%), while S. vilae DSM 105013T contained anteiso-C15:0 (64.1%), iso-C16:0 (12.6%), and 2-OH C14:0 (15.6%) as major fatty acids, and S. frigoramans KCTC 49696T contained summed feature 8 (40.4%) and summed feature 3 (30.2%) as major fatty acids. The differences in major and minor fatty acids can differentiate strain AK-R2A1-2T from other closely related members of the genus Subtercola (Table 2). The respiratory quinones of strain AK-R2A1-2T were menaquinone 9 (MK-9) and menaquinone 10 (MK-10), which is congruent with other closely related members of the genus Subtercola. The major polar lipids in strain AK-R2A1-2T were diphosphatidylglycerol (DPG) and three unknown aminolipids (AKL2, AKL3, AKL4). The polar lipids profile of strain AK-R2A1-2T was simple compared with those of type strains of other species of the genus Subtercola (Fig. S2).
Genome analyses
General and functional features of the genome of strain AK-R2A1-2T
After being assembled by Canu (version 1.7) de novo assembler, the whole genome of strain AK-R2A1-2T comprised a single circular chromosome of 4,318,731 bp, and the N50 value was 12,176 bases, with coverage of 95×. The G+C content calculated based on the respective whole-genome sequence was 65.8%. Comparison of two copies of 16S rRNA gene fragments with the whole-genome sequence indicated that DNA sequence contamination did not occur during genome assembly of strain AK-R2A1-2T. The whole-genome sequence of strain AK-R2A1-2T was deposited to NCBI under accession number CP087997.1. The genome sequence was annotated with RAST9, 10, and protein-coding sequences were determined using PGAP11. The NCBI PGAP annotation revealed that the genome of strain AK-R2A1-2T contains 3,874 protein-coding genes and 56 RNA genes, including two 5S rRNA genes, two 16S rRNA genes, two 23S rRNA genes, three non-coding RNA (ncRNA) genes, and 47 tRNA genes (Table S1). A total of 3,674 protein-coding sequences (CDs) were annotated by cluster orthologous group (Fig. 2). The largest group of CDs was classified as unknown (1,106 of total CDS, 30.1%), while the groups with the lowest numbers of CDs were classified as cytoskeleton and RNA processing and modification (1 of total CDS, 0.02%). Transcription (8.9%) and carbohydrate transport and metabolism (7.8%) were the groups with the next highest numbers of CDs annotated from the whole genome (Fig. 2 and Fig. S3).
A whole-genome-based phylogenetic tree was also reconstructed using the UBCG program (version 3.0). The whole genome-based phylogenetic tree indicated that strain AK-R2A1-2T forms a subgroup with S. lobariae 9583bT, and that all members of the genus Subtercola comprise one group (Fig. 3). Concurrently, members of the genus Agreia form a single cluster, consistent with the phylogenetic tree based on 16S rRNA gene sequences. These findings support that strain AK-R2A1-2T represented a novel species of the genus Subtercola.
Genomic features associated with cold-adaption of strain AK-R2A1-2T
The RAST server, PGAP, and BlastKOALA pipeline were used to perform genome annotation and determine the metabolic pathways in strain AK-R2A1-2T. Information in the following sections is predicted from the genome sequence analyses.
Stress response: Members of the genus Subtercola grow well in low temperatures, and although strain AK-R2A1-2T was isolated at 25°C, it was proven to grow at 4°C; the type strains of other members of this genus can even grow at –2°C5. The potential mechanism(s) by which bacteria invoke various stress responses to adapt to environmental changes have yet to be fully elucidated. Bacterial adaptations to low temperature to facilitate survival include cell membrane fluidity adaptation (increased unsaturated fatty acids), protein response (refolding cold-damaged proteins), and so on12. Cold-shock proteins (Csp) are produced by bacteria in response to a rapid decrease in temperature, with the Csps acting as RNA chaperones to help prevent mRNA misfolding during cold stress13. The PGAP analysis in the current study showed that the genome of strain AK-R2A1-2T contained two genes encoding a cold-shock protein (UFS59601.1; UFS60376.1) and a cold shock domain-containing protein (UFS59609.1). When whole-genome mining of strain AK-R2A1-2T was performed via RAST SEEDS and BlastKOALA, many well-studied, cold-inducible genes14-20 were annotated by the BlastKOALA pipeline, including aceE (encoding pyruvate dehydrogenase E1 component [EC:1.2.4.1]), aceF (pyruvate dehydrogenase E2 component, dihydrolipoamide acetyltransferase [EC:2.3.1.12]), cspA (cold shock protein), deaD (ATP-dependent RNA helicase DeaD [EC:3.6.4.13]), dnaA (chromosomal replication initiator protein), gyrA (DNA gyrase subunit A [EC:5.6.2.2]), hupB (DNA-binding protein HU-beta), infA (translation initiation factor IF-1), infB (translation initiation factor IF-2), infC (translation initiation factor IF-3), nusA (transcription termination/antitermination protein NusA), otsA (trehalose 6-phosphate synthase [EC:2.4.1.15 2.4.1.347]), otsB (trehalose 6-phosphate phosphatase [EC:3.1.3.12]), pnp (polyribonucleotide nucleotidyltransferase [EC:2.7.7.8]), rbfA (ribosome-binding factor A), recA (recombination protein RecA), and tig (trigger factor) (Table 3). Compatible solutes, a group of compounds that are important for osmotic stress and cold shock, can reduce membrane stress via various mechanisms. Glycine betaine, choline, and proline, which are recognized cold-protective solutes21, 22, were annotated from the whole genome of strain AK-R2A1-2T. In addition, genes encoding for the synthesis of the osmoprotectant proline from glutamate (proA, glutamate-5-semialdehyde dehydrogenase [EC:1.2.1.41]; proB, glutamate 5-kinase [EC:2.7.2.11]; proC (pyrroline-5-carboxylate reductase; [EC:1.5.1.2]), two choline dehydrogenases (betA, [EC:1.1.99.1]), and the glycine betaine transport system (opuC, opuBD, opuA) were identified in the genome of AK-R2A1-2T23-27 (Table 3). From the RAST SEED annotation of strain AK-R2A1-2T (Fig. S4), 22 genes were identified as related to the stress response. Among these genes, three of them were involved in osmotic stress (osmoregulation), 11 were related to oxidative stress [oxidative stress (6), glutathione: redox cycle (3), glutaredoxins (1), glutathionylspermidine and trypanothione (1)], and eight were related to detoxification. Some freezing-protective solutes, which can contribute to cold stress remission28, 29, such as amino sugar and nucleotide sugars (26), amino acids (96), starch, and sucrose (27) were also detected in the genome information of strain AK-R2A1-2T.
Motility: Two genes related to motility including one for flagellar assembly (rpoD) and one for bacterial chemotaxis (rbsB) were annotated from the whole genome of strain AK-R2A1-2T. The incomplete pathway of motility explains why strain AK-R2A1-2T was non-motile.
Fatty acid metabolism: The beta-oxidation-related biosynthesis pathway, including 3-hydroxyacyl-CoA dehydrogenase [EC:1.1.1.35], acyl-CoA oxidase [EC:1.3.3.6], acyl-CoA dehydrogenase [EC:1.3.8.7], acetyl-CoA acyltransferase [EC:2.3.1.16], and enoyl-CoA hydratase [EC:4.2.1.17], was detected in the whole genome of strain AK-R2A1-2T. Concurrently, long-chain acyl-CoA synthesis (EC:6.2.1.3), related to acyl-CoA synthesis, was reconstructed from the whole genome of strain AK-R2A1-2T by BlastKOALA pipeline.
Genome comparison and biosynthetic potential of the novel strain
The proposed threshold values of dDDH and ANI for delineating novel bacterial species were recommended as lower than 70% and lower than 95–96%, respectively30. In the current study, the ANI values between strain AK-R2A1-2T and type strains of members of the genus Subtercola were 76.8–80%, followed by 73.6–74% with members of the genus Agreia, and less than 74.7% with other genera in the family Microbacteriaceae. Consistently, strain AK-R2A1-2T exhibited a dDDH value of 22.5–24.5% with members of the genus Subtercola, 20.1–20.5% with members of the genus Agreia, and less than 20.7% with other genera in the family Microbacteriaceae (Table S2). Similar results were obtained for orthoANI values, which were less than 80.3% with most closely related strains in the family Microbacteriaceae (Table S2).
Bacteria can produce a large number of secondary metabolites that act as a reservoir of bioactive metabolites, with some exhibiting unique functions such as salt resistance31. In the genus Subtercola, there are limited studies on the production of specialized metabolites and the potential source of natural products in this genus remains untapped. In recent years, many computational methods have been developed to identify the biosynthetic gene clusters (BGCs) in genomic data. AntiSMASH is a platform that is widely used for genome mining of secondary metabolites. The secondary metabolite regions of the genome of strain AK-R2A1-2T were searched using antiSMASH (V. 6.0.1). Seven regions were identified from the whole genome, and these regions contained non-alpha poly-amino acids like e-Polylysine (NAPAA), beta-lactone (microansamycin), T3PKSS (alkylresorcinol), RRE-containing (kosinostatin), beta-lactone, terpene (carotenoid), and redox-cofactor (lipopolysaccharide) (Table S3 and Figs. S5–S11). Some of these compounds play vital roles in antioxidant, antimicrobial, and anti-inflammatory activities. For example, NAPAA is a cationic peptide proven to prevent microbe proliferation and is approved as a food-grade cationic antimicrobial metabolite32, which has natural antioxidant and antimicrobial activities33. Carotenoid can be used to protect against oxidative stress in model systems34-37, and regulates membrane fluidity at low temperature by cell membrane adaptation38, 39. Furthermore, increasing carotenoid levels might increase resistance against cold shock40. Alkylresorcinol can protect epithelial cells against oxidative damage, showed antioxidant activity41-43, and has antigenotoxic activity43 and potential antiglioma activity44. Kosinostatin has antitumor, antimicrobial, and antiproliferative functions45, 46, while beta-lactone is a natural product with significant clinical applications through its antimicrobial, anticancer, and anti-obesity properties47, 48. Microansamycin is a member of the macrolactams family, which are important bioactive compounds that have anti-tuberculous and antitumor activity49. Those secondary metabolites are highly matched with its host plant.
Effect of strain AK-R2A1-2T on rice seedling growth under low temperature
Rice is a cold-sensitive crop, and its exposure to low-temperature stress (<20 oC), during germination and early seedling growth, can negatively affect the initial stand establishment50, 51. To evaluate the strain AK-R2A1-2T on rice seedling growth under low temperature, rice seeds were co-cultivated with strain AK-R2A1-2T for 10 days at 20°C, and then total fresh weight, shoot weight, shoot length, root number, root length, and chlorophyll contents were examined in rice seedling (Fig. 4). Shoot weight and shoot length were increased 1.67-fold and1.36-fold, respectively compared to control (CK), while the root length and root number were increased 1.42-fold and 1.40-fold, respectively in rice seedling with strain AK-R2A1-2T (Fig.4). Strain AK-R2A1-2T significantly improved the rice seedling biomass and root morphological parameters under low temperatures.