Arthrobacter pokkalii sp nov, a Novel Plant Associated Actinobacterium with Plant Beneficial Properties, Isolated from Saline Tolerant Pokkali Rice, Kerala, India

A novel yellow colony-forming bacterium, strain P3B162T was isolated from the pokkali rice rhizosphere from Kerala, India, as part of a project study aimed at isolating plant growth beneficial rhizobacteria from saline tolerant pokkali rice and functionally evaluate their abilities to promote plant growth under saline conditions. The novel strain P3B162T possesses plant growth beneficial traits such as positive growth on 1-aminocyclopropane-1-carboxylic acid (ACC), production of indole acetic acid (IAA) and siderophore. In addition, it also showed important phenotypic characters such as ability to form biofilm and utilization of various components of plant root exudates (sugars, amino acids and organic acids), clearly indicating its lifestyle as a plant rhizosphere associated bacterium. Taxonomically, the novel strain P3B162T was affiliated to the genus Arthrobacter based on the collective results of phenotypic, genotypic and chemotaxonomic analyses. Moreover, molecular analysis using 16S rRNA gene showed Arthrobacter globiformis NBRC 12137T, Arthrobacter pascens DSM 20545T and Arthrobacter liuii DSXY973T as the closely related phylogenetic neighbours, showing more than 98% 16S rRNA similarity values, whereas the recA gene analysis displayed Arthrobacter liuii JCM 19864T as the nearest neighbour with 94.7% sequence similarity and only 91.7% to Arthrobacter globiformis LMG 3813T and 88.7% to Arthrobacter pascens LMG 16255T. However, the DNA-DNA hybridization values between strain P3B162T, Arthrobacter globiformis LMG 3813T, Arthrobacter pascens LMG 16255T and Arthrobacter liuii JCM 19864T was below 50%. In addition, the novel strain P3B162T can be distinguished from its closely related type strains by several phenotypic characters such as colony pigment, tolerance to NaCl, motility, reduction of nitrate, hydrolysis of DNA, acid from sucrose, cell wall sugars and cell wall peptidoglycan structure. In conclusion, the combined results of this study support the classification of strain P3B162T as a novel Arthrobacter species and we propose Arthrobacter pokkalii sp.nov.as its name. The type strain is P3B162T (= KCTC 29498T = MTCC 12358T).

The genus Arthrobacter represents one of the most divergent heterotrophic bacterial groups of actinobacteria, because of their metabolic versatility they are reported to exist in a diverse range of environments like soils, plants, freshwater, clinical specimens and marine habitats [3,5,6]. Numerous studies have revealed the association of diverse strains of Arthrobacter with different plants by both culture-dependent and -independent methods [7][8][9][10][11]. More importantly it was found in higher portions in plants which are grown in saline, drought, polluted and low nutrient agricultural soils, where they were found to be beneficial to the plants by protecting them from abiotic stress and improve plant nutrition, health and yield [12][13][14][15]. Because of these plant beneficial properties, they were considered to be an important member among the plant growth promoting rhizobacteria that are present in the rhizosphere microflora [9,13,16]. Despite their importance towards plant health under stressful conditions, the knowledge about the association of Arthrobacter with different crop plants that are growing in stressful conditions and its functional significance is far from adequate. The discovery of a new Arthrobacter species from these crop plants as in this case pokkali rice, a highly saline tolerant rice variety and dealing with its plant probiotic properties is therefore still significant.
Till date, there are more than 70 described species with validly published names in the genus Arthrobacter that were isolated from various sources [17]. Notably recovered from plant rhizosphere are Arthrobacter cupressi [18], A. siccitolerans [19], A. bambusae [6], A. gyeryongensis [20], A. oryzae [21] and A. humicola [21]. However, none of these above studies provided any evidence that they possess plant growth beneficial properties even though they are isolated from different plant rhizosphere.
During an investigation on the taxonomy and functional characterization of rhizosphere bacteria of a saline tolerant pokkali rice, a strain designated as P3B162 T was isolated. The current study describes the isolation, plant beneficial functions and taxonomic position of this novel Arthrobacter strain P3B162 T using a polyphasic approach as recommended by Schumann et al. [22]. Based upon the collective results generated in this study, a new species of the genus Arthrobacter, Arthrobacter pokkalii sp. nov., is proposed.

Ethics statement
There is no requirement for any authority permit to collect the plant rhizosphere samples. The rhizosphere samples were collected from various rice fields with the knowledge and prior permission from the respective owners of the field. This study doesn't involve any endangered or protected plant species.
Isolation of bacterial strain P3B162 T Strain P3B162 T was isolated from the rhizosphere soil of a hybrid pokkali rice variety VTL-6 collected from Alappuzha in Kerala, India. A standard serial dilution procedure was followed in isolating the bacterial strain P3B162 T . Briefly, two grams of soil attaching to the root portion was collected carefully and mixed with 20ml of sterile 0.85% (w/v) NaCl solution. The resulting rhizosphere soil suspension was then vortexed vigorously for 10-30min and was allowed to stand still for 5min. A fivefold serial dilutions were done with the above rhizosphere soil suspension and 0.1ml aliquots of each dilution was then spread plated onto 1/1000 dilutions of Luria Bertani (LB) agar medium (Himedia, India), pH 5.5 (maintained with 0.1N HCl). These plates were then incubated for 1-2 weeks. Strain P3B162 T was regularly subcultured on full strength LB agar medium at 28°C for 3-4 days. For storage purposes, the strain was kept at 4°C as active plates for 2-3 weeks as short-term storage and as 10% glycerol suspensions stored in deep freezer at −80°C as long-term storage. The following reference strains were included in this study; Arthrobacter globiformis LMG 3813 T , A. pascens LMG 16255 T , A. liuii JCM 19864 T , A. humicola DSM 25587 T , A. oryzae DSM 25586 T and A. cupressi DSM 24664 T .
Phylogenetic analysis using 16S rRNA and recA genes The genomic DNA was extracted from overnight grown bacterial cells using the DNA extraction kit (QIAamp DNA Mini Kit; Qiagen) as per the protocol mentioned in the manufacturer's instruction manual. The 16S rRNA gene amplification with primers 27F and 1492R [23] were performed as reported before [24]. The recA gene was amplified using the primers GPRA-UF2 (5 0 -GGSAAGGGSKCNGTNATGCG-3 0 ) and GPRA-UR2 (5 0 -CCTTSCCCTGSCCNARYT-3 0 ) [25] using PCR program of 5 min at 95°C, 33 cycles of 1 min at 95 C, 1 min at 56°C and 1 min at 72°C, and 10 min at 72°C as final extension step. The PCR product size of 1500bp (16S rRNA) and 800bp (recA) was gel eluted using Gel Extraction kit (Qiagen) following the manufacturer's protocol and bi-directionally sequenced using the ABI Prism BIGDye Terminator v3.1 cycle sequencing ready reaction kit (Applied Biosystems). For 16S rRNA gene sequencing, the 16S rRNA internal primers as mentioned by Rameshkumar et al. [24] were used and for recA gene sequencing GPRA-UF2 and GPRA-UR2 primers were used. The Applied Biosystems 3500 DNA Genetic Analyzer was used for sequencing purposes. The quality gene sequences thus obtained were matched against the reference 16S rRNA and recA gene sequences that are available in the National Center for Biotechnology Information (NCBI) database using BLASTn search tool. Using Eztaxon [26], nearest phylogenetic neighbours and its 16S rRNA gene sequence similarity values were determined. The sequences were aligned using CLUSTAL X1.8 [27] and corrected by manual inspection. Phylogenetic trees were reconstructed using three different tree making algorithms; neighbour-joining, maximum parsimony and maximum-likelihood methods with Kimura two-state parameter model analyses as described before [28] except that the program MEGA version 5 was used [29]. Using NCBI ORF finder, the recA nucleotide sequences were translated to amino acid sequences based on the putative open reading frames and NCBI BLASTp search tool was used to find identical protein sequences that are present in the protein database.

Genomic fingerprinting
Repetitive extragenic palindromic PCR (rep-PCR). Whole DNA fingerprinting was performed by using four (rep-PCR) primers; REP1R-I & REP2-I, enterobacterial repetitive intergenic consensus (ERIC1R & ERIC2), BOXA1R, and (GTG) 5 as described [30,31,24] using PCR program of 5 min at 95°C, 35 cycles of 1 min at 95°C, 1 min at 40°C and 8 min at 65°C, and then a final extension step of 10 min at 72°C. The experiment was repeated three independent times to make sure the fingerprints generated are reproducible.

GC content and DNA-DNA hybridization
The procedure of Marmur [32] was followed to isolate high molecular weight genomic DNA for GC content analysis. By determining the mid-point (T m ) of the thermal denaturation curve of the given DNA [33] and by using the equation of Owen & Hill [34], the G+C base composition was calculated. DNA-DNA hybridization (DDH) was conducted using fluorometric direct binding method [35] and the hybridization experiments were performed in micro dilution wells. For labelling the DNA as probes, photoprobe biotin from Vector Laboratories, Inc., was used. The temperature used for DNA-DNA hybridization was 53°C.

Chemotaxonomy
For analysis of cellular fatty acids, cells were grown on Tryptic Soy Agar (TSA-Himedia) for 2 days at 30°C and cells were harvested at similar physiological age. The fatty acids were extracted and determined as mentioned before [28,36] except MIDI Sherlock TSBA6 (version of the database) was used for identification of the fatty acid methyl esters. The peptidoglycan structure and whole cell wall sugars analyses were determined according to the protocol described by Schumann [37] at DSMZ using service facility provided for bacterial identification. For polar lipids and quinone analyses stationary phase grown cells were used and 3.3xPYE broth which contains 1.0% peptone from casein, and 1.0% yeast extract, (pH 7.2) was used to raise the biomass. The extraction and analyses of quinones and polar lipids were done as reported recently [38,39] and for quinone analysis the HPLC equipment reported by Stolz et al. [40] was applied.

Plant growth functions
The utilisation of 1-aminocyclopropane carboxylic acid (ACC) as sole nitrogen source was determined using the method as reported by Penrose & Glick [42] with slight modification. The isolate was grown in a minimal agar medium consisting of 5g/L each of glucose and sucrose, K 2 HPO 4 -0.8g/L, KH 2 PO 4 -0.2g/L, MgSO 4 .7H 2 O-0.2g/L, CaCl 2. 2H 2 O-0.2g/L, NaCl-5g/ L, agar-18g/L and 3mM ACC as sole nitrogen source. After a week incubation, visible growth on the medium considered as positive for ACC utilisation [43]. To check Indole Acetic acid (IAA) production, the isolate was grown in a minimal broth medium consisting of 5g/L of sucrose, K 2 HPO 4 -0.8g/L, KH 2 PO 4 -0.2g/L, MgSO 4 .7H 2 O-0.2g/L, CaCl 2. 2H 2 O-0.2g/L, NaCl-5g/ L and supplemented with 1g/L tryptophan. After an incubation period of 3 days at 30°C, the quantification of IAA was determined using the Salkowski assay [44]. Ability to form biofilm was determined using the method reported by Christensen et al. [45] with the modifications as mentioned by Sarkar and Chakraborty, [46] the tests were performed in McCarthy tubes (HiMedia). Siderophore production was determined using the method as mentioned before [47]. Ability to grow in semi-solid nitrogen free medium was tested by growing the isolate in minimal medium devoid of any nitrogen source: 5g/L of sucrose, K 2 HPO 4 -0.8g/L, KH 2 PO 4 -0.2g/L, MgSO 4 .7H 2 O-0.2g/L, CaCl 2. 2H 2 O-0.2g/L, NaCl-5g/L and agar-2g/L, incubated for 5 days at 30°C. A visible pellicle formation on the medium was considered as positive. Presence of nitrogen fixing gene was checked by PCR amplification of nifH gene using primers 19F (5'-GCIWTYTAYGGIAARGGIGG-3') and 407R (5'-AAICCRCCRCAIACIACRTC-3') as reported by Rameshkumar et al. [24]. Ability of the isolate to grow under drought stress was analysed as cited by Marasco et al. [48] except that the isolate was grown in 10% and 20% Polyethylene Glycol (PEG) prepared in Luria Bertani broth medium.

Nucleotide sequence accession number
The

Results and Discussion
The 16S rRNA phylogenetic analysis, comprising 1424bp revealed that strain P3B162 T belongs to the genus Arthrobacter (Fig 1), displaying higher pairwise 16S rRNA gene sequence similarities to the type strains of two Arthrobacter species which are representatives of the genus Arthrobacter sensu stricto [3]; Arthrobacter globiformis NBRC 12137 T (98.17%), and A. pascens DSM 20545 T (98.10%). Followed by A. liuii DSXY973 T with 98.02% sequence similarity, very recently described species so far has not been assigned to any Arthrobacter group. No 16S rRNA similarity values higher than 98% were found with other type strains of established Arthrobacter species. It is worth to mention here that other Arthrobacter species namely A. phenthrenivorans Sphe3 T , A. defluvii 4C1-a T , A. niigatensis LC4 T , representatives of Arthrobacter oxydans-group and A. tecti LMG 22282 T , a representative of Arthrobacter agilis-group shared relatively high 16S rRNA similarity values between 97.5 to 97.8% with strain P3B162 T . However, all these 16S rRNA similarity values are lesser than 98.5%, the mean value considered to be the threshold for the existence of a novel species within a genus [49]. Altogether these informations clearly provided the evidence for assignment of strain P3B162 T as a novel species in the genus Arthrobacter. However such observations should be carefully validated by a polyphasic taxonomy approach which involves characterization based on phenotypic and genotypic methods [22,50]. In 16S rRNA maximum-likelihood phylogenetic tree the novel strain P3B162 T occupied a distinct phylogenetic position moving away from A. globiformis NBRC 12137 T and A. pascens DSM 20545 T but clustering stably with A. liuii DSXY973 T (Fig 1). When other treeing algorithms were applied (maximum parsimony, neighbour-joining) again A. liuii DSXY973 T was phylogenetically identified as the next relative of strain P3B162 T (S1 and S2 Figs). However, the phylogenetic analyses do not provide reliable evidence that P3B162 T is related to any other species of the genus Arthrobacter. Based on the high 16S rRNA similarity and considering the phylogenetic positioning of P3B162 T the following type strains were used for further taxonomic comparison: Arthrobacter globiformis LMG 3813 T , A. pascens LMG 16255 T , A. liuii JCM 19864 T , A. humicola DSM 25587 T , A. oryzae DSM 25586 T and A. cupressi DSM 24664 T .
The results of 16S rRNA gene sequence comparisons and its corresponding phylogeny did not provide information on relationship to any Arthrobacter group as defined by Busse et al. [3]. Therefore, we analysed the housekeeping gene encoding recombinase A (recA) of strain P3B162 T and its related reference Arthrobacter species. This gene is considered to be of greater taxonomic resolution than the 16S rRNA gene, when describing and differentiating closely related strains up to species level [22,[50][51][52][53]. A partial (770) bp of recA sequence of strain P3B162 T was obtained after PCR amplification and sequencing. The resulting sequence was compared with recA sequences available from the databases and determined in the course of this study. The result also identified A. liuii JCM 19864 T as the next relative of P3B162 T with highest recA sequence similarity (94.7%). Other Arthrobacter species showed recA sequence similarities of 88.7-91.7% with strain P3B162 T (  Table). Likewise, very high amino acid sequence similarities were also observed among several type strains of Arthrobacter species notably 100% amino acid sequence similarity between four species of Arthrobacter oxydans-group namely A. oxydans ATCC 14358 T , A. phenanthrenivorans Sphe3 T , A. polychromogenes ATCC 15216 T and A. siccitolerans 4J27 T , between A. pascens ATCC 13346 T and A. oryzae DSM 25586 T and between A. globiformis ATCC 8010 T and A. chlorophenolicus A6 T (S1 Table), even though these type strains of Arthrobacter species share very low similarity at the nucleotide level among each other (Table 1). However, when longer stretches (212-250 amino acids) of the same gene of selected species were compared, the amino acid similarity values between aforementioned species decreased significantly. For instance, in this comparison A. phenanthrenivorans Sphe3 T shared only 97.2% similarity with A. pascens ATCC 13346 T and A. chlorophenolicus A6 T shared 97.2% similarity with A. globiformis LMG 3813 T instead of 98.7 and 100% similarity observed in the shorter amino acid sequences, respectively well reflecting the relationships indicated from 16S rRNA phylogeny and its assignment to respective Arthrobacter groups. From these observations it is obvious that at least some relationships indicated  by high RecA similarities in the short amino acid sequences are misleading and no taxonomic conclusions should be drawn without good support from other data. It is also clear that the trimmed corresponding recA nucleotide sequences provide better taxonomic resolution to closely related Arthrobacter strains than its amino acid sequences. This observation is most likely due to the degeneracy of the genetic code, where in many cases a mutation at the third position of a nucleotide triplet does not affect the translated amino acid. Hence, in pairwise sequences comparisons, the amino acid sequence similarity value was always higher than that of its corresponding DNA sequences (Table 1 and S1 Table). Similar findings were obtained by other researchers when using other protein coding genes for similar phylogenetic studies; HSP60 gene [54,55], gyrB gene [56]. Since only rather short recA sequences (450 nucleotides) were accessible we did not carry out phylogenetic calculations. However, both recA and RecA data identified A. liuii JCM 19864 T as the next relative of P3B162 T . The results from recA gene sequence comparisons also supported the status of P3B162 T as a novel species because the similarity in this gene sequence was lower (94.7%) than that found between the very closely related Arthrobacter species; A. polychromogenes ATCC 12516 T and A. oxydans ATCC 14358 T (97.10%).
In addition to the phylogenetic analysis using 16S rRNA and recA genes, we used rapid DNA typing methods based on whole genome to check whether the strain P3B162 T is genetically different from its closest type strains. For this purpose, repetitive extragenic palindromic genomic fingerprinting methods using different set of primers was adapted and the results clearly proved that the DNA fingerprints of strain P3B162 T was evidently different from the DNA fingerprinting patterns of closely related Arthrobacter species [A. globiformis LMG 3813 T , A. pascens LMG16255 T , A. liuii JCM 19864 T , A. humicola DSM 25587 T , A. oryzae DSM 25586 T and A. cupressi DSM 24664 T (S3 Fig)]. Though different genomic fingerprint do not necessarily identify different species but the results support the conclusion derived from the sequence analyses using two genes encoding 16S rRNA and recA, respectively that strain P3B162 T may likely to represent a novel species in the genus Arthrobacter.
In order to validate distinct species status for strain P3B162 T , the DNA-DNA hybridization between the strain P3B162 T and its closely related phylogenetic neighbours sharing more than 98% 16S rRNA similarity; A. globiformis LMG 3813 T , A. pascens LMG 16255 T and A. liuii JCM 19864 T were chosen and performed. The results showed that, at the DNA-DNA level, the relatedness between P3B162 T and its closest phylogenetic neighbours were found to be 40.8% with A. liuii JCM 19864 T , 31.9% with A. globiformis LMG 3813 T , and 41.5% with A. pascens LMG 16255 T . This result was further well supported by the recA sequence analysis, where the novel strain P3B162 T shared low level of gene sequence similarities (<95%) with its closely related species of the genus Arthrobacter (Table 1), clearly indicating the suitability of using recA gene as a substitute in describing new Arthrobacter species before considering DNA-DNA hybridization experiments. However, the results of DNA relatedness undoubtedly pointed towards the fact that strain P3B162 T does not belong to any of its closest Arthrobacter species when considering the cut-off value of 70% DNA-DNA relatedness for the definition of species is considered [57]. These data clearly indicate that strain P3B162 T should represent a novel species status in the genus Arthrobacter.
Nevertheless, description of novel species involves merely not based on the genomic differences, but also should be validated through a polyphasic approach which includes both phenotypic and chemotaxonomic data [53].
Results of phenotypic data validated the phylogenetic analysis results by placing the strain P3B162 T in the genus Arthrobacter. This assignment of novel strain P3B162 T to the genus is based on the phenotypic characters which are shared by almost all species of the genus Arthrobacter such as non-motility, strict requirement of oxygen for growth, positive in Gram staining, displaying a rod to coccus cell morphology in its growth cycle and no production of acid from glucose [3]. Colonies of the novel isolate grown on LB agar medium are yellow pigmented (S4 Fig) and other characters including biochemical and physiological features of strain P3B162 T were detailed and presented in its species description. The identified phenotypic characters which distinguish strain P3B162 T from it's phylogenetically closely related Arthrobacter species are mentioned in Table 2.
Next, we evaluated the strain P3B162 T for its plant growth beneficial traits in plate assays. The results revealed that strain P3B162 T able to produce plant growth phytohormone, Indole acetic acid (IAA) (106.7±6.48 μg/ml) from L-tryptophan, which is a well-known plant growth regulator, showed positive growth on 1-aminocyclopropane-1-carboxylic acid (ACC) (S5A Fig), indicating its ability to utilize ACC, a precursor of ethylene biosynthesis thereby protects the plants from the harmful effects of stress ethylene produced during abiotic stress where many studies have shown ACC utilizing rhizobacteria known to help plants under abiotic stress conditions [59,60], positive for siderophore production (S5B Fig) and can form biofilm Colony pigment a Yellow ------ Hydrolysis: Utilisation of Larabinose

HICarbohydrate kit
Utilisation of: Acid Production:  [61,62]. The ability of the isolate to grow in 10% and 20% PEG solutions clearly indicated its possibility to perform its function under drought conditions as well [48,63]. However the novel strain P3B162 T found negative for growth in semisolid nitrogen free medium and for nitrogen fixation gene (nifH) amplification. These results altogether indicate the possible role of strain P3B162 T towards plant growth promotion as previous reports state that rhizosphere bacteria which possess these plant growth beneficial activities have positive influence on the plant growth [64][65][66].
We also identified a number of phenotypic characters which might contribute to the successful adaptability of this novel strain P3B162 T towards its existence in plant rhizosphere. It grows in pH 5.5, tolerates up to 8% NaCl, clearly depicting the acid saline nature of the site from where this novel isolate P3B162 T was isolated. It can grow both in oligotrophic as well as in high nutrient conditions clearly indicating its metabolic versatility of the novel isolate P3B162 T towards surviving under the fluctuating nutrient levels in the rhizosphere environments [67]. It utilises humic acid which is a major component of the soil [68], utilises certain plant derived components which are secreted as root exudates such as sucrose [69], citric acid, DL-malic acid [70], Larabinose, D-cellobiose, citric acid, D-dextrose, D-fructose, D-galactose, D-dextrose, maltose, xylose [71], tryptophan and putrescine [72][73][74]. Additionally, it utilises components which are majorly present in root exudates when the plant is under stress conditions; proline and mannitol [75,76]. It produces enzymes such as pectinase (S5D Fig), proteases, lipase and amylase which might be useful in hydrolysing the plant polymers [77]. These results altogether gives a positive indication that this novel strain P3B162 T possesses typical characters which can be selective advantage when it is residing in plant rhizosphere, were similar findings identified in other plant growth promoting rhizobacteria [78,79]. Therefore, it will be interesting to explore further to know more in detail about the beneficial interactions between the novel strain and its host plant, the saline tolerant pokkali rice plant and how it protects the plant from salinity stress and is being explored through ongoing plant inoculation studies.
Chemotaxonomically strain P3B162 T contained fatty acids majorly dominated by anteiso-C 15:0 (53.98%), anteiso-C 17:0 (14.62%), iso-C 16:0 (8.64%), iso-C 15:0 (8.06%), C 16:0 (7.14%), a fatty acids profile which is similar to that of representatives of the genus Arthrobacter. The fatty acids profile of strain P3B162 T and its phylogenetically closely related Arthrobacter species are mentioned in Table 3. The quinone system of P3B162 T contained predominantly menaquinones, 88.1% MK-9(H 2 ), 9.1% MK-8(H 2 ) and 2.8% MK-10(H 2 ) which is in line with the quinone system of numerous Arthrobacter species. The polar lipid profile (Fig 2) showed the major lipids as diphosphatidylglycerol, phosphatidylglycerol, dimannosylglyceride, digalactosyldiacylglycerol and trimannosyldiacylglycerol. In addition, minor to moderate amounts of phosphatidylinositol, an unidentified glycolipid (GL4) and three lipids (L1, L2 and L7) only detectable after total lipid staining. This polar lipid profile is similar to the profile of Arthrobacter globiformis, Arthrobacter pascens, Arthrobacter humicola, (Arthrobacter sensu stricto), Arthrobacter polychromogenes ('Arthrobacter oxydans-group), Arthrobacter histidinolovorans (Arthrobacter aurescens-group), A. liuii and A. cupressi and clearly distinguishes from profiles of other Arthrobacter species (Busse, unpublished results; [58]). In detail, strain P3B162 T could be distinguished from A. globiformis NBRC 12137 T based on the presence of glycolipid GL4, lipids L1, L2, L7 and higher amount of trimannosyldiacylglycerol (identified in the image of the polar lipid profile of A. globiformis NBRC 12137 T based on its chromatographic motility) in its polar lipid profile as given in [58]. Furthermore, the presence of phospholipid PL1 in the profile of A. globiformis NBRC 12137 T was not detectable in P3B162 T . Polar lipid profiles of strain P3B162 T and A. liuii DSXY973 T were more similar to each other especially based on the presence of GL4 designated GL3 in A. liuii DSXY973 T [58]. However, also these two strains could be distinguished from each other based on the presence of lipids L1, L2 and L7 in the polar lipid profile strain P3B162 T . The cell wall peptidoglycan analysis showed that the cell wall of strain P3B162 T contains A3α peptidoglycan type [4] with L-Lys-L-Ser-L-Thr-L-Ala as interpeptide bridge, which is similar to the cell wall peptidoglycan structure of A. liuii JCM 19864 T . However, this peptidoglycan structure clearly distinguishes P3B162 T from Arthrobacter sensu stricto and places it in the vicinity of 'Arthrobacter oxydans-group' as defined by Busse et al. [3]. The cell wall peptidoglycan structure plays a major place in the Arthrobacter taxonomy as many Arthrobacter species have been grouped based on their cell wall peptidoglycan structure Busse et al. [3]. Considering the peptidoglycan structure which is in line to that of A. oxydans group, the phenotypic characters that are differentiating P3B162 T from the Arthrobacter oxydans group were given in S2 Table. The cell-wall sugars of strain P3162 T was composed of galactose, rhamnose and mannose an important character that differentiates strain P3B162 T from its closest neighbours (Table 2 and S2 Table).

Taxonomic conclusion
In summary, the phenotypic, chemotaxonomic and phylogenetic data obtained in this study showed that strain P3B162 T belongs to the genus Arthrobacter. The results of phylogenetic analysis using 16S rRNA and recA, genomic fingerprinting using different DNA typing methods, DNA-DNA hybridization, and phenotypic data distinguished the strain P3B162 T from its closest neighbours A. globiformis LMG 3813 T , A. pascens LMG16255 T , A. liuii JCM 19864 T , A. humicola DSM 25587 T , A. oryzae DSM 25586 T and A. cupressi DSM 24664 T . Therefore, strain P3B162 T is classified as a novel species in the genus Arthrobacter, for which the name Arthrobacter pokkalii sp. nov. is proposed.
The type strain P3B162 T (= KCTC 29498 T = MTCC 12358 T ) was isolated from the rhizosphere of saline tolerant pokkali rice of Alappuzha, Kerala, India. The DNA G+C content of strain P3B162 T is 64.0 mol%.  Table. recA amino acid sequence similarity values between strain P3B162 T and its related Arthrobacter type strains. (DOCX) S2 Table. Phenotypic and genotypic characters differentiating strain P3B162 T from Arthrobacter strains having similar peptidoglycan interpeptide bridge of L-Lys-L-Ser-L-Thr-L-Ala. (DOCX)