Introduction

The genus Brevibacterium was suggested by Breed (1953) based on Brevibacterium linens as the type species. Thereafter, many species of Brevibacterium were isolated and studied. The genus includes at least 49 recognized species according to (www.bacterio.net/brevibacterium.html) (Choi et al. 2018). Brevibacterium has been isolated from various environments such as food (e.g., fermented seafood [Choi et al. 2013] and cheese [Leclercq-Perlat et al. 2000]), humans (e.g., clinical specimens [Wauters et al. 2001] and the human body [Wauters et al. 2003]), and saline environments (e.g., saline soil [Tang et al. 2008], beach sediment [Lee 2006], marine environments [Lee 2008], salt-lake sediment [Guan et al. 2010]), and other environment such as soil of ginseng field (Jung et al. 2018).

Some Brevibacterium, as plant growth-promoting bacteria (PGPB), can improve plant growth under different environmental conditions. B. frigoritolerans strain Imbl 2.1 has potential for bioremediation of phorate both in liquid cultures and agricultural soils (Jariyal et al. 2015). Recent study indicated that Brevibacterium linens RS16 can confer salt tolerance to Oryza sativa (Chatterjee et al. 2018).

Maize is one of the most important foods, feed, and industrial crops worldwide, and is highly sensitive to drought and salt (Morari et al. 2015). To identify beneficial rhizobacteria that promote maize growth, we isolated a Brevibacterium strain, B. frigoritolerans ZB201705, from the maize rhizosphere in drought- and salt-stressed soil and analyzed its complete genome sequence. In addition, we assessed the activity of the strain under high-salt and simulated drought conditions. The information provided here will provide valuable insight into PGPB for agriculture under abiotic stress.

Materials and methods

Sample collection and bacteria isolation

Brevibacterium frigoritolerans ZB201705 was isolated from drought-stressed saline rhizosphere soil of maize in Bayan Nur, Inner Mongolia Autonomous Region, China (40° 13′–42° 28′, E 105° 12′–109° 53′). Soil samples were collected at a depth of 5–10 cm in the maize rhizosphere. Five points were chosen according to the “S” form five-spot sampling method (Lu et al. 2013). Then, 0.1 kg of soil was collected at each point and pooled into one composite soil sample.

To isolate bacteria, soil samples were placed in paper bags and stored at 4 °C. Next, 100 mL of sterile water and 0.01 kg of soil were transferred to a Waring blender. The sample was homogenized for 1 min and the supernatant was collected. The supernatant was centrifuged for 10 min at 5000×g and re-dissolved in 0.5 mL of sterile water. Samples diluted to 10−3, 10−4, and 10−5 were streaked onto soil extract agar plates containing 5 g/L of beef extract, 10 g/L of peptone, 5 g/L of NaCl, 100 g/L of mannitol, 15 g/L of agar, and 1 L of distilled water (pH 7.2). After 2 days of incubation at 30 °C, colonies of different morphologies and sizes were frozen in 15% glycerol and stored at −80 °C.

Identification of sequence similarity

To identify the colony strains, the bacteria were first identified by Gram staining. Then, a similarity search with the 16S rRNA gene nucleotide sequence (accession number MH490935) was performed using the online program EzBioCloud (Yoon et al. 2017) and two housekeeping genes (recA and atpD) were used to identify the strains. Finally, the average nucleotide identity (ANI) was calculated using JSpeciesWS (Richter et al. 2016) to identify the interspecific relationships among strains.

Genomic DNA preparation, genome sequencing, and assembly

For genomic DNA isolation, strains were inoculated into 50 mL of liquid medium and cultivated overnight at 30 °C in a shaker at 150 rpm. Genomic DNA was then extracted using a Rapid Bacterial Genomic DNA Isolation Kit (Sangon Biotech Shanghai Co., Ltd., Shanghai, China).

Whole-genome sequencing was performed using a combination of the PacBio RS and Illumina sequencing platforms. The Illumina PE library and PacBio Library (8–10 kb) were each constructed. First, 52 Mb of continuous long read (CLR) PacBio sequences were sequenced on the PacBio RS platform using an SMRT cell with a sequence N50 length of 12,035 bp, of which 47.2 Mb of sequences had lengths ≥ 5 kb. Then, a 400 bp Illumina sequencing library was constructed, and 2.9 Gb of paired-end sequences were sequenced on the Illumina HiSeq 2000 platform. The Illumina data were used to evaluate the complexity of the genome and were assembled using Velvet ver. 1.2.10 genome assembler with a k-mer length of 99 (Zerbino and Birney 2008).

The complete genome sequence was assembled using both the PacBio and Illumina reads. The assembly was initially produced using an in-house assembly solution, in which a de Bruijn-based assembly algorithm and CLR correction algorithm were integrated. The final circularization step was completed manually using Circos ver. 0.64 software (http://circos.ca/) (Krzywinski et al. 2009).

Genome annotation

Prodigal software (Hyatt et al. 2010) was used to predict bacterial genes. Then, the protein sequences of the predicted genes were compared with the Clusters of Orthologous Groups (COG) protein database (http://www.ncbi.nlm.nih.gov/COG/) (Tatusov et al. 2003), and the corresponding COG annotation results were obtained. BLAST (blastx/blastp 2.2.24+) was used to compare the predicted genes with the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.genome.jp/kegg/genes.html). Specific pathways involving the corresponding genes were obtained according to the KEGG Orthology number in the comparison results. Finally, the BLAST results were analyzed using Gene Ontology annotations (http://www.geneontology.org/) with Blast2go.

Identification of B. frigoritolerans ZB201705 abiotic stress responses

For the anti-abiotic stress susceptibility tests (e.g., salt and drought), different NaCl concentrations (1, 3, and 5 M) and d-sorbitol concentrations (0, 1, 2, and 3 M) were tested on LB agar plates. For the salt and drought stress challenges, cultures were multiplied at 30 °C to a density of about 5 × 107 cells/mL. The cell concentrations were determined by plating appropriate dilutions onto agar containing 1, 3, and 5 M NaCl or 1, 2, or 3 M d-sorbitol for at least 12 h at 30 °C.

$$ \mathrm{Survival}\ \mathrm{rate}\ \left(\%\right)=\frac{\mathrm{Colonies}\ \mathrm{of}\ \mathrm{stress}\ \mathrm{group}}{\mathrm{Colonies}\ \mathrm{of}\ \mathrm{CK}\ \mathrm{group}}\left(\mathrm{CK},0\kern0.1em \mathrm{mol}/\mathrm{L}\ \mathrm{NaCl}\ \mathrm{or}\ \mathrm{D}-\mathrm{sorbitol}\right). $$

Statistical analyses

The obtained data were subjected to an analysis of variance using the general linear model software Agres and Agdata, and means were compared using the least significant difference test at a probability level ≤ 0.05.

Results

Organism information and classification

According to the Gram stain results, the cells were Gram-positive bacteria (Fig. S1). Based on the similarity of the 16S rRNA gene sequence determined using EzBioCloud, twelve strains with the greatest similarity were selected. The results (Table 1) showed 100% identity with B. frigoritolerans FJAT-2396, 98% with Bacillus psychrosaccharolyticus ATCC 23296, and 97% with Bacillus manliponensis JCM 15802. Nevertheless, analyses based on 16S rRNA gene sequencing alone are insufficient for species classification within this group. Therefore, the identities of two housekeeping genes (recA and atpD) in B. frigoritolerans ZB201705 and 12 other type strains were determined, which revealed two identities > 90% (Table 1). To classify the strains further, the twelve most similar strains were selected and their whole genome sequences were analyzed with ANI using JSpeciesWS (Richter et al. 2016). The highest ANI value was 97.1% for B. frigoritolerans FJAT-2396, which exceeded the species cutoff threshold (95%) (Goris et al. 2007). The whole genome sequences of the two strains were compared using Circos ver. 0.22 (Krzywinski et al. 2009), which revealed numerous differences between the two strains (Fig. 1). A comparison with the available type strain genomes indicated that our strain should be considered a subspecies of B. frigoritolerans. Based on these organism information and classification, we named the strain B. frigoritolerans ZB201705.

Table 1 Sequence similarity (%) with Brevibacterium frigoritolerans ZB201705
Full size table
Fig. 1
figure 1

Comparison of whole genome sequence from the two strains was performed using the software Circos V0.22

Full size image

Genome sequencing results

To obtain a genome sequence with no gaps, we used a combination of the PacBio RS and Illumina sequencing platforms. The assembled results contained 350 contigs with lengths ≥ 200 bp, with an N50 length of 53,736 bp, and indicated that the B. frigoritolerans ZB201705 genome consists of one circular chromosome with a size of 5,475,560 bp (Fig. 2) and one plasmid sequence (Fig. S2), with no gaps. The guanine–cytosine content of the genome was 40.5%. In total, 5,266 genes were predicted, accounting for 79.73% of the genome. Moreover, 39 ribosomal RNAs and 81 transfer RNAs were detected. Of the total number of predicted genes (5,266), 4,391 (83.38%) were putative protein-coding genes and 72.52% were assigned a putative function (Table 2).

Fig. 2
figure 2

The circular chromosome of Brevibacterium frigoritolerans ZB201705

Full size image
Table 2 Genome features of Brevibacterium frigoritolerans ZB201705
Full size table

Based on the COG analysis (Tatusov et al. 2003), the identified proteins were classified into 25 functional categories (Fig. 3). Among these categories, many proteins were involved in salt and drought stress responses. For example, the coenzyme transport and metabolism category (H) contained a sodium/proline symporter (Jung et al. 2012), the signal transduction mechanism category (T) contained the signal transduction histidine-protein kinase/phosphatase DegS (Kunst and Rapoport 1995), and the transcription category (K) contained the transcriptional regulatory protein DegU (Kunst and Rapoport 1995) and cold-shock protein CspB, which improves maize grain yield under water-limited conditions (Castiglioni et al. 2008). A previous study indicated that stress-induced reactive oxygen species accumulation is counteracted by enzymatic antioxidant systems comprised of a variety of scavengers, including superoxide dismutase (SOD), catalase (CAT), and glutathione peroxidase (GPX) (Wang et al. 2017). According to the COG functional categories of the genome, there were several related genes. For example, the inorganic ion transport transmembrane category (P) contained three SOD- and three CAT-related genes, and the secondary metabolite biosynthesis, transport, and catabolism category (O) contained one GPX-related gene. These results suggest that B. frigoritolerans ZB201705 has an important role in environmental pathogen defense and abiotic stress responses.

Fig. 3
figure 3

COG functional classifications of Brevibacterium frigoritolerans ZB201705 coding sequences

Full size image

Activity of the strain under high salt and simulated drought conditions

We identified several environmental abiotic stress-related genes based on our COG analysis. With the aim of detecting its resistance to high salt and drought conditions, we cultured B. frigoritolerans ZB201705 on LB media with different concentrations of NaCl and d-sorbitol. The strain can tolerate up to 3 M d-sorbitol (~ 55%, w/v) and 5 M NaCl (~ 29%, w/v) with survival rates of 3.5% and 12.6%, respectively (Fig. 4).

Fig. 4
figure 4

Identification of the activity of Brevibacterium frigoritolerans ZB201705 under salt and drought stress

Full size image

Nucleotide sequence accession numbers

The complete genomic sequence and 16S rRNA gene sequence of B. frigoritolerans ZB201705 have been deposited in the GenBank database under Accessions CP030063 and MH490935.

Discussion

Brevibacterium has been isolated from numerous environments, including saline environments such as saline soil, beach sediment, marine environments, and salt-lake sediment (Lee 2006; Lee 2008; Tang et al. 2008; Guan et al. 2010). However, few studies have referred specifically to B. frigoritolerans. One strain of B. frigoritolerans, Imbl 2.1, has shown potential for the bioremediation of phorate (Jariyal et al. 2015). Another strain, SMA23, may improve growth and nutrient uptake in wheat (Meena and Saharan 2017). As for plant growth-promoting bacteria (PGPB), free-living plant growth-promoting bacteria form symbiotic relationships with specific plants or bacterial endophytes that colonize some or part of the interior tissues of plants (Basu et al. 2017). Therefore, PGPB are generally used as inoculants for the biostimulation, biocontrol, and biofertilization of plants, and may improve plant growth under various environmental stress conditions (Numan et al. 2018).

The present study shows that B. frigoritolerans ZB201705 produces a large number of proteins involved in salt and drought stress responses, including DegS, DegU, CspB, and SOD. Anti-abiotic stress susceptibility tests revealed that this strain can tolerate up to 3 M d-sorbitol (~ 55%, w/v) and 5 M NaCl (~ 29%, w/v). These results suggest that B. frigoritolerans ZB201705 is a vital bacterium with potential environmental abiotic stress resistance applications, including the promotion of salt and drought tolerance in maize. The information provided here will provide valuable insight into PGPB for agriculture under abiotic stress.