Introduction

The genus Mesorhizobium, with the type species Mesorhizobium loti, in the family Phyllobacteriaceae, phylum Proteobacteria, was first described by Jarvis et al. (1997). At the time of writing, the genus Mesorhizobium consists of 59 species with validly published names (https://lpsn.dsmz.de/genus/mesorhizobium). Members of the genus Mesorhizobium have been isolated from various habitats, such as different leguminous plants, deep-sea sediment (Yuan et al. 2016), soils and seawater (Wang et al. 1999; Fu et al. 2017). Their isolation source has been for the most part the root nodules of various leguminous plants. Members of the genus Mesorhizobium are Gram-stain-negative and aerobic. The typical characteristics of the cells include rod-shaped, non-motile and non-sporulating (Zhang et al. 2012). In addition, the genus Mesorhizobium was established for the growth rate which was slower than that of the genus Rhizobium. In this study, we describe a novel bacterial strain, designated lm94T, isolated from rhizosphere soil of Alhagi sparsifolia. The aim of the present investigation was to determine the taxonomic position of strain lm94T based on analysis of phenotypic, phylogenetic, genomic and chemotaxonomic characteristics.

Materials and methods

Isolation, maintenance and culture conditions

Strain lm94T was isolated from rhizosphere soil of Alhagi sparsifolia obtained from Alar city, located in Xinjiang province, China (80º 40ʹ 63ʺ E, 40º 25ʹ 23ʺ N). The sample was serially diluted in sterile water and plated on LB (yeast extract 5 g, tryptone 10 g, NaCl 10 g, water 1 L) agar and incubated at 30 °C for 10 days. One beige colony, designated lm94T, was picked and purified by repeated plate streaking. The strain was preserved at − 80 °C in sterile 1% (w/v) saline supplemented with 15% (v/v) glycerol. M. wenxiniae LMG 30254 T was obtained from previous study (Zhang et al. 2018). Mloti JCM 21464T (the type species of this genus) was purchased from the Japan Collection of Microorganisms (JCM). Both strains were used as reference strains.

Phylogenetic and genome sequence analysis

Genomic DNA extraction from strain lm94T was carried out using a bacterial genomic DNA kit (OMEGA) according to the manufacturer’s recommendations. The 16S rRNA gene was amplified by PCR with the primers 27F and 1492R (Mu et al. 2018). The amplicon was cloned into pMD18-T vector (Takara) and recombinant plasmids were reproduced in Escherichia coli DH5α cells. Sequencing was performed by TSINGKE Biotechnology (Qingdao, PR China). Then an almost full-length 16S rRNA gene was submitted to GenBank database (accession number is MN519466). Comparison of 16S rRNA with related strains was conducted by EzTaxon server (https://www.ezbiocloud.net/identify) (Yoon et al. 2017). Multiple alignments were performed via the MEGA version 7.0 and phylogenetic trees were reconstructed using the neighbour-joining, maximum-parsimony and minimum-likelihood methods in the computer program MEGA version 7.0 (Kumar et al. 2016). Bootstrap analysis was performed with 1000 replications. The draft genome sequence of strain lm94T was sequenced at Chinese National Human Genome Center (Shanghai, PR China) using Solexa paired-end sequencing technology. Genome data of M. wenxiniae LMG 30254T (NPKH00000000) and Mloti JCM 21464T (QGGH00000000) were obtained from the NCBI genome database. The predicted coding sequences of each genome were translated and annotated using the KEGG and COG. The average nucleotide identity (ANI) based on the whole genome sequence was calculated using the ANI calculator (www.ezbiocloud.net/tools/ani). The percentage of conserved protein (POCP) was calculated according to previous study (Qin et al. 2014). Digital DNA–DNA hybridization (dDDH) analysis was performed on the DSMZ Genome-to-Genome Distance Calculator platform (Richter et al. 2015). Furthermore, genome mining for the presence of secondary metabolites gene clusters was performed using antiSMASH program (version 5.0) (Blin et al. 2019).

Morphological, physiological and biochemical analyses

For phenotypic characteristics, strain lm94T was cultivated on LB at 30 °C for 3 days. Microscopic observations were performed using scanning electron microscopy (Hitachi SU-8100). Gram-staining was carried out as described by Cerny (1978). Gliding motility was examined according to the method described by Bowman (2000). The effects of different temperatures on growth were tested at 5, 10, 15, 20, 25, 30, 37, 40, 45, 50, 55, 60 °C for approximately 7 days in liquid medium. Then salt tolerance tests were tested at LB medium with a pH of 7.0, the concentrations of NaCl were ranging from 0 to 10% (w/v, at intervals of 1%). Then the initial pH range for growth was determined at pH 3.0–10.0 intervals of 0.5 units and at 30 °C in LB agar medium. The pH was adjusted by addition of citric acid–sodium citrate buffer (pH 3.0 and 3.5), NaH2PO4 buffer (4.0, 4.5 and 5.0), MES (pH 5.5 and 6.0), PIPES (pH 6.5 and 7.0), HEPES (pH 7.5 and 8.0), Tricine (pH 8.5) and CAPSO (pH 9.0, 9.5 and 10.0). The OD600 values of the cultures were measured after incubation for the growth tests mentioned above. Growth under anaerobic (10% H2, 10% CO2 and 80% N2) and microaerobic (5% O2, 10% CO2 and 85% N2) conditions were determined on LB medium with or without 0.1% (w/v) KNO3 after incubation for 14 days in an anaerobic jar. Oxidase activity was tested using the oxidase reagent kit (BioMérieux) according to manufacturer’s instructions. Catalase activity was detected by bubble production in 3% (v/v) H2O2. Susceptibility to antibiotics was examined on LB medium using the disc diffusion method as described previously, and according to procedures outlined by the Clinical and Laboratory Standards Institute (Institute 2007). Tests for other physiological and biochemical feature were determined using the API 20E and API 50CHB kits (BioMérieux). Enzyme activities were examined using the API ZYM test system (BioMérieux). The substrate-oxidation profiles were gained using Biolog GEN III MicroPlates. All the API and Biolog tests were performed in duplicate according to the manufacturer’s instructions, except that the suspension media were adjusted to 1% (w/v) NaCl.

Chemotaxonomic characterization

Strain lm94T and the reference strains were cultured at optimum growth conditions. Cells were harvested in the exponential growth phases and then lyophilized immediately for chemotaxonomic analyses. The fatty acids were analyzed according to the Sherlock Microbial Identification System (MIDI) Sherlock version 6.3. The fatty acid extracts obtained using the standard MIDI protocol were subjected to further analysis by GC–MS as previously described. Peaks were automatically integrated and fatty acid names and percentages were determined using MIS standard software with the TSBA40 database. Polar lipids were separated by two-dimensional silica gel TLC according to the methods of Xu et al. (2007). Polar lipids were then analyzed as described by Minnikin et al. (1984). For analyses of respiratory quinones, respiratory quinones were extracted from freeze-dried cells, and determined by HPLC (Kroppenstedt 1982).

Results and discussion

Phylogenetic and genome analysis

Pairwise comparison of the 16S rRNA gene sequence of strain lm94T (GenBank accession number MN519466) with the corresponding 16S rRNA gene sequences in the EzBioCloud database indicated that the strain shared 16S rRNA gene sequence similarities of 96.6% and 94.9% with Mesorhizobium wenxiniae LMG 30254T and Mesorhizobium loti JCM 21464T (the type species of this genus), respectively. The topological structure of the phylogenetic neighbor-joining tree (Fig. 1) clearly indicated that strain lm94T clustered with species of the genus Mesorhizobium and was distinctly separated from members of other genera.

Fig. 1
figure 1

Neighbour-joining tree based on 16S rRNA gene sequences, showing the phylogenetic position of strain lm94T among species of the genus Mesorhizobium. Bootstrap values (expressed as percentages of 1000 replications) > 50% are shown at branching points. Bradyrhizobium japonicum USDA 6 T (GenBank accession No. AP012206) was used as an outgroup. Filled circles indicate nodes also obtained in both maximum-likelihood and maximum-parsimony trees. Bar, 0.02 substitutions per nucleotide position

Draft genome sequencing of strain lm94T yielded a genome of 5 256 375 bp in length after assembly which produced 53 contigs, and the N50 value was 177 626 bp. The genome coverage was 561x. The genomic DNA G+C content of strain lm94T calculated from the draft genome sequence was 63.6 mol%, which was different from the related type strain Mesorhizobium wenxiniae LMG 30254T and Mesorhizobium loti JCM 21464T (Table 1). The whole genome data were used to calculate the average nucleotide identity (ANI; www.ezbiocloud.net/tools/ani) with M. wenxiniae LMG 30254T and Mloti JCM 21464T. The OrthoANI value of strain lm94T with M. wenxiniae LMG 30254T and Mloti JCM 21464T were 75.0 and 75.2%, respectively, which were lower than the species threshold of 95–96% (Liu et al. 2019). The POCP values between the genomes of strain lm94T and reference strains were 53.6 and 53.5%, respectively. The dDDH values (the recommended results from formula 2) between strain lm94T and M. wenxiniae LMG 30254T and M. loti JCM 21464 T were 20.0 and 20.1%, respectively. Both were significantly lower than their threshold values (dDDH, 70%) (Meng et al. 2020). Secondary metabolic genes were predicted by antiSMASH (https://antismash.secondarymetabolites.org/). The analysis of secondary metabolic gene clusters revealed that strain lm94T harboured seven gene clusters, which was different from reference strains (Table 1).

Table 1 Comparison between the genomes and gene clusters involved in biosynthesis of secondary metabolites of strain lm94T, M. wenxiniae LMG 30254T and Mloti JCM 21464T

Morphological, physiological and biochemical characteristics

The cells of strain lm94T grown on LB agar were circular, smooth, approximately 0.3–0.5 µm in width and 0.7–1.0 µm in length (Fig. S1) and cells without appendages were observed. Cells were Gram-stain-negative without gliding motility. The morphological characteristics between strain lm94T and reference strains were similar, which was consistent with the result of the phylogenetic analysis. However, the most important phenotypic feature, which can be significantly separated from reference strains, was the colour of cell mass. Anaerobic growth did not occur on LB agar. The catalase and oxidase test of strain lm94T was positive. Cells of the type strain lm94T were sensitive to kanamycin, chloramphenicol, streptomycin, gentamicin, neomycin, tetracycline, cefoperazone, spectinomycin, paromomycin, penicillin and erythromycin, while resistant to ampicillin, vancomycin, troleandomycin, rifamycin, minocycline and lincomycin. The detailed morphological, physiological, and biochemical characteristics of strain lm94T are summarized in the species description and in Table 2.

Table 2 Characteristics that differentiate strain lm94T from M. wenxiniae LMG 30254T and Mloti JCM 21464T

The major cellular fatty acids (> 10.0%) of strain lm94T were C19:0 cyclo ω8c and summed feature 8 (C18:1 ω6c and/or C18:1 ω7c), which were also the dominant fatty acids in the two reference strains. While the fatty acid composition was similar to that of two related genera, but there were differences in the proportions of some fatty acids. For example, C18:1 ω9c was found in reference strains, but was absent in strain lm94T. Moreover, iso-C15:0 3-OH was present in strain lm94T, but was absent in two related genera. Further comparative information between the proposed new species and two related genera are given in Table S1. The major polar lipids identified in strain lm94T were phosphatidylethanolamine (PE), phosphatidylglycerol (PG), unidentified phospholipid (PL), phosphatidylcholine (PC), diphosphatidylglycerol (DPG), unidentified aminolipid (AL), unknown glycolipid (GL), unidentified aminophospholipid (APL2) and unidentified polar lipid (L1 and L2). All three strains contained PE, PG, PC and DPG. However, strain lm94T contained GL that was absent in two reference strains. The polar lipid profiles of all the strains are represented in Fig. S2. Furthermore, the sole respiratory quinone of strain lm94T was Q-10, which was also the case with M. wenxiniae LMG 30254T and Mloti JCM 21464T.

Based on the results of the chemotaxonomic examination, phenotypic analysis and phylogenetic analysis from this study, strain lm94T represents a novel species of the genus Mesorhizobium, for which the name Mesorhizobium xinjiangense sp. nov. is proposed.

Description of Mesorhizobium xinjiangense sp. nov.

Mesorhizobium xinjiangense (xin.jiang.en’se. N.L. neut. adj. xinjiangense referring to Xinjiang province in China, the area where the type strain was isolated).

Cells are Gram-stain-negative, aerobic, non-flagellated, non-gliding, rod-shaped, 0.3–0.5 µm in width, 0.4–1.0 µm in length. Colonies on LB are circular, convex, smooth, opaque, beige-pigmented and approximately 1.0–1.5 mm in diameter after 3 days at 37 °C. Growth occurs at 20–45 °C (optimum 37 °C), pH 6.0–9.5 (optimum pH 7.0–7.5), and in the presence of 0–6% (w/v) NaCl (optimum 0–1%). Cells are catalase-positive and oxidase-positive. They are negative for H2S production, citrate utilization, indole production, gelatinase production and starch hydrolysing but positive for arabinose, tryptophan deaminase and lysine decarboxylase according to API 20E tests. Acid is produced from d-arabinose, l-arabinose, d-ribose, d-xylose, d-adonitol, d-glucose, d-fructose, l-rhamnose, mannose, ESC, sorbitol, d-lyxose, d-fucose, l-fucose, d-arbaitol and 5-keto-potassium gluconate (API 50 CHB). In API ZYM tests, positive for alkaline phosphatase, esterase (C4), leucine arylamidase, valine arylamidase, trypsin, acid phosphatase and naphthol-AS-BI-phosphohydrolase activities. In carbon source oxidation tests, cells are positive for d-sorbitol, d-mannitol, d-arabitol, glycerol, d-fructose-6-PO4, N-acetyl-d-glucosamine, l-alanine, l-arginine, l-glutamic acid, l-histidine, α-d-glucose, d-mannose, d-fructose, l-fucose, d-fucose, l-rhamnose, d-serine, d-glucuronic acid, glucuronamide, l-lactic acid, d-malic acid and l-malic acid. Q-10 is the sole respiratory quinone. The major polar lipids are phosphatidylethanolamine (PE), phosphatidylglycerol (PG), unidentified phospholipid (PL), phosphatidylcholine (PC), diphosphatidylglycerol (DPG), unidentified aminolipid (AL), unknown glycolipid (GL), unidentified aminophospholipid (APL2) and unidentified polar lipid (L1 and L2). The major cellular fatty acids (> 10.0%) are C19:0 cyclo ω8c and Summed Feature 8 (C18:1 ω6c and/or C18:1 ω7c). The genomic DNA G+C content of the type strain is 63.6 mol%.

The type strain is lm94T (=KCTC 72863T=CCTCC AB2019377T) and was isolated from rhizosphere soil of Alhagi sparsifolia, collected from Xinjiang province, China (80º 40′ 63″ E, 40º 25′ 23″ N).

The GenBank accession number for the 16S rRNA gene sequence of strain lm94T is MN519466. The whole genome shotgun project of strain lm94T has been deposited at DDBJ/ENA/GenBank under the accession WOAC00000000.