Introduction

Members of the genus Arthrobacter are aerobic, Gram-stain positive, catalase-positive actinobacteria with a high DNA G + C content (Keddie et al. 1986; Stackebrandt et al. 1983), contain MK-9(H2) or MK-8/MK-9 as the major menaquinones. Two major cell-wall peptidoglycan structural types in Arthrobacter species are A3α and A4α (Busse 2016). The genus Arthrobacter strains currently comprises 69 species with validly names, frequently isolated from various environments including soils (Gupta et al. 2004; Ganzert et al. 2011), marine environments (Pindi et al. 2010; Cobet et al. 1970), lake water (Dubinina and Zhdanov 1975), marsh (Zhang et al. 2018) and sewage (Kim et al. 2008). It has also been found in the filtration substrate (Ding et al. 2009), the surface of smear-ripened cheeses (Irlinger et al. 2005) and mural paintings (Heyrmanl et al. 2005). Psychrophilic members of the genus Arthrobacter including A. psychrochitiniphilus (Wang et al. 2009), A. alpinus (Zhang et al. 2010), A. cryoconiti (Margesin et al. 2012) and Arthrobacter ruber (Liu et al. 2018) were obtained from Antarctica and other cold environments. The wide distribution of Arthrobacter members is probably due to their nutritional versatility and their resistance to environmental stress factors. Recently, cold-active enzymes such as β-D-galactosidase and α-Amylase have been isolated from Arthrobacter species (Kim et al. 2017; Rutkiewicz et al. 2019), which indicated that psychrophilic species of the genus Arthrobacter have important potential sources of cold-active enzymes.

During an exploration of Antarctic microbial resources, a Gram-staining-positive, aerobic bacterium, designated strain Z1-20 T was isolated from a soil sample collected in Zhongshan station, Antarctic. In this study, strain Z1-20 T was characterized taxonomically using a polyphasic approach, and a novel species of the genus Arthrobacter is proposed.

Materials and methods

Isolation of microorganism

The strain Z1-20 T was isolated from a soil sample collected from the Zhongshan station, Antarctic. The soil sample was serially diluted in 1% (w/v) NaCl solution and inoculated onto R2A agar plates (0.5 g yeast extract, 0.5 g peptone, 0.5 g casamino acids, 0.5 g glucose, 0.5 g soluble starch, 0.3 g K2HPO4, 0.024 g MgSO4, 0.3 g sodium pyruvate, pH7.0, 15 g agar in 1L) and then incubated at 15 °C for 7–10 days. The isolated strain was routinely cultured on R2A agar plates at 20 °C and maintained at 4 °C. It was also stored as glycerol suspensions (20%, v/v) at − 80 °C.

Phenotypic characterization

Strain Z1-20 T was grown for 4 days in R2A broth at 20 °C and 180 rpm. An aliquot of culture was stained with crystal violet and Gram’s iodine. The morphological characteristics of strain Z1-20 T were observed using transmission electron microscope (model JEM-1200EX) after 4 days of growth on R2A agar plates. Growth at different temperatures (4, 10, 15, 20, 25, 28 and 37 °C) and different pH (pH 4.0–12.0 at intervals of 1.0 pH unit) were tested on R2A agar plates for 7–10 days. Media with different pH values were prepared using the buffer system described by Xu et al. (2005). The tolerance to different NaCl concentrations (0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 15 and 20%, w/v) was tested on R2A agar plates as the basal plates by incubating the cultures for 7–10 days at 20 °C. Oxidase production was tested as described by McCarthy and Cross (1984). Catalase activity was detected by the production of bubbles following addition of a drop of 3% (v/v) H2O2. Hydrolysis of casein, gelatin, starch and tween 80 and nitrate reduction were determined as described by Cowan and Steel (1965). Carbon source utilization was determined using the API 50CH tests (BioMérieux, Marcy-l’Étoile, France) according to the manufacturer’s instructions.

Phylogenetic and genome analyses

PCR amplification of the 16S rRNA gene fragment of the novel strain was performed. The 16S rRNA gene sequence of strain Z1-20 T was compared against a database of cultured species via the EzBio-Cloud Database (http://www.ezbiocloud.net/) (Yoon et al. 2017a, b) and multiple sequence alignment was programed via CLUSTAL W (MEGA 7.0) (Larkin et al. 2007). Neighbor-joining phylogenetic and maximum-likelihood phylogenetic trees were constructed using MEGA 7.0 (Kumar 2016; Felsenstein 1985; Saitou and Nei 1987), and PHYML phylogenetic tree was generated using the web (http://www.atgc-montpellier.fr/phyml/) after multiple alignment of data by CLUSTAL W (Larkin et al. 2007). Distances (distance options according to the Kimura two-parameter model) and clustering were determined using the neighbor-joining method. Bootstrap analysis was used to evaluate the tree topology of the neighbor-joining data by performing 1000 resamplings (Fitch 1970).

Genomic DNA for the genome sequencing was extracted via the Genomic DNA Rapid Isolation Kit for Bacterial Cell (TIANGEN). The genomic DNA G + C content was determined from whole genome sequence data of the strain Z1-20 T. Digital DNA-DNA hybridization (dDDH) analysis to establish the precise taxonomic position of strain Z1-20 T was performed using the GGDC web (http://ggdc.dsmz.de) (Meier-Kolthoff et al. 2013). Average nucleotide identity (ANI) was calculated using the web services available at EzBio-cloud (https://www.ezbiocloud.net/tools/ani) (Yoon et al. 2017a, b). A maximum-likelihood phylogenetic tree was constructed based on the alignment of 900 single-copy core genes with 26 Arthrobacter species, constructed by RAxML-NG (https://github.com/amkozlov/raxml-ng/) with an inferred model by ModelTest (https://github.com/ ddarriba/modeltest/), with 100 bootstrap replicates (Croucher et al. 2015).

Online tool antiSMAH (https://antismash.secondarymetabolites.org/#!/start) was used to predict the gene clusters of secondary metabolites in strains Z1-20 T, A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T. For comparative analyses, Z1-20 T, A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T were selected based on neighbor-joining phylogenetic tree and integrity and similarity of their genomes. GeneMarkS software (Besemer et al. 2001) was used to predict the protein-coding genes of bacterial genome. Comparing the protein sequence of the predicted gene with NR, eggNOG and KEGG databases were performed by diamond blastp. For this analysis, the comparison result with the highest score was selected for annotation and the following cutoff values were applied (e-value < 1e-6 and amino acid sequence identity of at least 40%).

Chemotaxonomic characterization

Biomass for chemotaxonomic analysis was obtained by cultivation in shake flasks (with shaking at about 180 rpm) using R2A liquid medium (1% (w/v) NaCl, pH 7.0) at 20 °C for 4 days. The analysis of peptidoglycan structure was carried out as described by Schumann (Schumann 2011) in DSMZ. Polar lipids were extracted, examined by two-dimensional TLC and identified using published procedures (Minnikin et al. 1984a, b; Minnikin et al. 1984a, b). Menaquinones were isolated using the methods of Minnikin et al. (Minnikin et al. 1984a, b; Minnikin et al. 1984a, b) and separated by HPLC (Kroppenstedt 2004). The cellular fatty acid compositions were determined according to the instructions of the Sherlock Microbial Identification System (MIDI Sherlock version 4.5, MIDI database TSBA40 4.10) (Kroppenstedt 2004; Sasser 1990).

Results and discussion

Morphological and phenotypic characteristic

Bacterial cells of strain Z1-20 T are ellipsoidal, non-spore-forming, approximately 0.43–0.79 μm wide and 0.71–1.4 μm long (Fig. S1). When tested on R2A agar plates, growth of strain Z1-20 T was observed at temperatures between 10 and 28 °C, with an optimum growth temperature of 20 °C and within the pH range pH 6.0–9.0, with an optimum of pH 8.0. The strain grew on R2A agar with NaCl concentrations range from 0 to 7% (w/v), with an optimum of 1%. Positive for oxidase, catalase, nitrate reduction, but negative for hydrolysis of casein, gelatin, starch, Tween80. The following substrates in the API 50CH tests were utilized: mannitol, D-mannose, D-glucose, D-fructose, aesculin, iron citrate, potassium gluconate and 2 potassium ketogluconate. Z1-20 T and the reference strains are shown in Table 1 and Table S1.

Table 1 Characteristic s useful for differentiating strain Z1-20 T from the most closely related species of the genus Arthrobacter and Pseudarthrobacter
Full size table

Phylogenetic characteristics

The almost-complete 16S rRNA gene sequence (1497 bp) of strain Z1-20 T was determined and compared against the EzBio-cloud Database (Kim et al. 2012) to retrieve most similar sequences of recognized bacteria. Strain Z1-20 T showed high sequence similarities to Arthrobacter species. The most closed 16S rRNA gene sequence similarity was found to A. glacialis HLT2-12-2 T (97.05%) (Liu et al. 2019), followed by A. psychrochitiniphilus GP3T (96.92%) (Wang et al. 2009), Pseudarthrobacter siccitolerans 4J27T and P. phenanthrenivorans Sphe3T. Neighbor-joining and maximum-likelihood phylogenetic trees showed that the novel strain Z1-20 T formed a separate branch from the two closely related Arthrobacter type strains and Pseudarthrobacter clusters (formerly ‘Arthrobacter oxydans’ group) (Fig. 1 and Fig. S3). The high consistency of phylogenetic position of strain Z1-20 T was also supported by the phylogenomic tree (Fig. 2) and other algorithms tested (Fig.S4). Based on the 16S rRNA gene sequence similarity and phylogenetic clustering, the closely related type strains A. glacialis, A. psychrochitiniphilus and the relative P. siccitolerans and P. phenanthrenivorans, as well as the type strain P. polychromogenes of the genus Pseudarthrobacter were chosen as reference strains for further comparative study to determine the taxonomic position of the novel strain Z1-20 T.

Fig. 1
figure 1

Neighbor-joining phylogenetic tree based on 16S rRNA gene sequences (1408 bp) showing the phylogenetic position of strain Z1-20 T within the family Micrococcaceae. Dots (•) indicates the clades that were conserved in the maximum-likelihood phylogenetic tree and PHYML phylogenetic tree. Bootstrap values are shown on nodes in percentages of 1,000 replicates, when greater than 50%. Bar, 5 substitutions per 1000 nt

Full size image
Fig. 2
figure 2

A maximum-likelihood phylogenetic tree was generated from the alignment of 900 single-copy core genes, with bootstrap support values from 100 replicates. Bold font indicates the position of strain Z1-20 T

Full size image

Genomic characteristics

The genome size of strain Z1-20 T was about 4.43 Mb with 4351 genes and 69 tRNA genes and a single 16S rRNA gene. The DNA G + C content of the strain Z1-20 T was calculated to be 63.0% from genome data. Digital DNA-DNA hybridizations (dDDH) values between strain Z1-20 T against the reference strains were 13.8–20.8%, which were clearly below the 70% threshold generally accepted for species delineation (Wayne et al. 1987). Average nucleotide identity (ANI) score between strain Z1-20 T against the reference strains ranged from 72.1 to 72.9%. The ANI values lower than 95% (Kim et al. 2014) found between the new strain and its closest related species could also support strain Z1-20 T representing a novel species. In addition, maximum-likelihood phylogenomic tree based on the alignment of 900 single-copy core genes showed high consistency of the phylogenetic position of strain Z1-20 T, which support the designation of strain Z1-20 T as a novel species (Fig. 2). The above genomic analysis results confirm that the strain Z1-20 T represents a new taxon.

AntiSMASH analysis of the genome of Z1-20 T found six type biosynthesis gene clusters: bacteriocin, butyrolactone, terpene, betalactone, T3PKS and NRPS, showing a slight difference with the two reference strains (Table S2). Genes associated with cold adaptation in the genome of strain Z1-20 T have been revealed (Table 2 and Supplementary file 1) which include cold shock response, osmoprotection protection, oxidative stress risistance and membrane adaptations. Z1-20 T has an uptake system for organic osmoprotectants, as suggested by the identification of several genes coding for ABC-type glycine betaine/proline transporters, as well as a putative betAB pathways for the formation of glycine betaine/proline. It is well known that glycine betaine often acts as a cryoprotectant to protect bacteria from cold environments by preventing cold-induced aggregation of cellular proteins and maintaining membrane fluidity at low temperature (Ko et al. 1994; Chattopadhyay 2002). Besides accumulating osmoprotectants from the environment, strain Z1-20 T is also capable of synthesizing compatible solutes. Two putative pathways are present for the formation of trehalose. Genes predicted to code for trehalose-6-phosphate synthase and trehalose-6-phosphate phosphatase, comprising the OtsAB pathway for trehalose synthesis from UDP-D-glucose and α-D-glucose-6-phosphate, as well as the genes for the alternative TreYZ pathway (Table 2 and Supplementary file 1). Interestingly, there are a higher number of tRNA and copies of gene encoding cold shock protein CapA compared with strain A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T. CapA can be induced following the cold shock and remain overexpressed several hours after the temperature downshift which could be the key factors contributing to the survival ability in cold Antarctic environments (Phadtare et al. 1999). Sufficient number of tRNA in the genome also explains its translational efficiency and survival in the unfavorable conditions in Antarctic.

Table 2 Genes linked to environmental stress response in strain Z1-20 T and the reference strains
Full size table

Chemotaxonomic characterization

The purified amino acids in total hydrolysates (4 N HCl, 100 °C, 16 h) of peptidoglycan were alanine, glutamic, threonine and lysine in a molar ratio of 2.7:1.0:0.4:0.3. The peptidoglycan type was A3α (L-Lys-Thr-Ala2, A11.27) and lysine was the diagnostic diamino acid. The menaquinones of strain Z1-20 T were MK-9(H2) (55.5%), MK-8(H2) (37.7%) and MK-7(H2) (6.8%), while the polar lipids were phosphatidylglycerol (PG), diphosphatidylglycerol (DPG), phosphatidylinositol (PI) and one unidentified glycolipid (GL) (Fig. S2). There is a distinct difference in the peptidoglycan structures between Z1-20 T and Pseudarthrobacter strains. The peptidoglycan interpeptide chain of Z1-20 T detected in DSMZ is Lys-Thr-Ala2, corresponding to peptidoglycan structure A11.27, while Pseudarthrobacter sp. strains have A3α peptidoglycan type with Lys-Ser-Thr-Ala in the interpeptide corresponding to A11.23 which distinguished distinctly from Arthrobacter sp. Strain ZS1-20 T was composed of the major lipids PG, DPG, PI and one GL, which was also found in the most closed A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T. While the relative Pseudarthrobacter strains contained the major compounds DPG, PI, PG and three GL. The almost identical polar lipids profiles and peptidoglycan variations supported that strain Z1-20 T belongs to the genus Arthrobacter. The cellular fatty acids (> 10% of the total fatty acids) were anteiso-C15:0 (45.4%), anteiso-C17:0 (20.7%) and iso-C15:0 (16.6%) (Table 3). The composition profile with anteiso-C15:0 as the main fatty acid is typical of species of Arthrobacter genus (Lee et al. 2003) that confirmed strain Z1-20 T belongs to the genus Arthrobacter. The above chemotaxonomic characteristics of strain Z1-20 T, such as peptidoglycan type, major fatty acids, major menaquinone and phospholipids were consistent with its assignment to the genus Arthrobacter.

Table 3 Cellular fatty acid compositions (%) of strain Z1-20 T and the two most closely related type strains of the genus Arthrobacter
Full size table

Strain Z1-20 T differed greatly from the closely related Arthrobacter type strains in terms of some of its phylogenetic, biochemical, chemotaxonomic and physiological data (Tables 1, 3 and Table S1). The NaCl tolerance range of Z1-20 T was slightly higher than A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T. The growth under 10 °C was nonvisible. The ability to oxidase production was different from A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T. Meanwhile, the content of iso-C15:0 of the fatty acids of strain Z1-20Tdiffered distinctly from the two reference strains, although a relative high proportion of iso-C15:0 is common on Arthrobacter members such as A. koreensis and A. globiformis (Lee et al. 2003; Hahne et al. 2019). Furthermore, digital DNA-DNA hybridizations values between strain Z1-20 T against the type strains of A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T were 20.3% and 13.8%, respectively and average nucleotide identity score between strain Z1-20 T against A. glacialis HLT2-12-2 T and A. psychrochitiniphilus GP3T were 72.5% and 72.1%, respectively. Based on the phenotypic and genotypic results obtained in this study, it is concluded that strain Z1-20 T represents a novel species of the genus Arthrobacter, for which the name Arthrobacter terrae sp. nov. is proposed.

Description of Arthrobacter terrae sp. nov

Arthrobacter terrae (ter’rae L. gen. fem. n. terrae, of the earth), Z1-20 T (= CCTCC AA 2019079 T = KCTC49361T), was isolated from a soil sample collected at Zhongshan station, Antarctic. Cells are Gram-staining-positive, non-spore-forming, non-flagellated, ellipsoidal, approximately 0.43–0.79 μm wide and 0.71–1.4 μm long (Fig. S1). Colonies of the novel strain Z1-20 T are yellow, convex, dry and opaque, with smooth and entire margins and the diameter is 1.0–1.5 mm after incubation on R2A agar for 72 h at 20 °C. Growth occurs at 10–28 °C (optimum 20 °C), at pH 6.0–9.0 (optimum pH 8.0) and in the presence of 0–7% (w/v) NaCl (optimum, 1%). Major polar lipids are phosphatidylglycerol (PG), diphosphatidylglycerol (DPG), phosphatidylinositol (PI) and one unidentified glycolipid (GL). The major fatty acids are anteiso-C15:0, anteiso-C17:0 and iso-C15:0 and the predominant menaquinone is MK-9(H2). The peptidoglycan type is A3α. The DNA G + C content of the genomic DNA of the type strain is 63.0%.