Introduction

The genus Paenibacillus in the family Paenibacillaceae was first described by Ash et al. based on comparative 16S rRNA sequence analysis of group 3 bacilli (Ash et al. 1991), and its description was later amended by Shida et al. (1997) and Behrendt et al (2010). Currently, over 200 species have been identified and published as members of the genus Paenibacillus (https://www.bacterio.net/genus/paenibacillus). Some Paenibacillus species serve as effective plant growth-promoting rhizobacteria (PGPR), which can act as biofertilizers and as antagonists of recognized root pathogens, such as bacteria, fungi, and nematodes (Padda et al. 2017).

Iron is an essential element for the maintenance of metabolic processes and serves as a selective determining of the microbial population in the rhizosphere. PGPR exerts their plant growth-promoting activity by depriving native microflora and bacterial pathogens of iron (Kloepper et al. 1980). Despite being abundant in various environments, iron is oxidized to its ferric form, which is inaccessible for microbial acquisition. To overcome this limitation, microorganisms produce and secrete small, high-affinity, iron-chelating compounds, known as siderophores, which help solubilise iron from organic compounds and improve its bioavailability (Sujatha and Ammani 2013). Siderophores have applications in nuclear fuel processing, biotechnology (De Serrano 2017), and medicine (Ali and Vidhale 2011). Moreover, siderophore production in Paenibacillus species has been reported before and reference some species (Wen et al. 2011; Chen et al. 2016).

In the present study, a polyphasic taxonomic approach was undertaken to characterize a novel strain isolated from marine coastal sediment collected in Busan, Republic of Korea, designated CAU 1523T. The study findings support a novel siderophore-producing species within the genus Paenibacillus, for which the name P. arenosi sp. nov. is proposed.

Materials and methods

Bacterial isolation and strains

Strain CAU 1523T was isolated from coastal sediment samples collected from Haeunda in Busan, Republic of Korea (35° 09′ 19.4″ N 129° 09′ 16.9″ E) on 3rd October, 2018. Sand samples were serially diluted with a sterilised 0.9% NaCl solution, and aliquots (100 mL) were spread onto marine agar 2216 (MA) plates (BD Difco, Franklin Lakes, NJ, USA). The plates were incubated at 30 °C for 2 weeks, and single colonies were re-streaked thrice on MA plates for purification, resulting in the isolation of CAU 1523T. The strain was stored in a 25% (v/v) glycerol suspension at − 80 °C in marine broth (MB).

Five type strains closely related to CAU 1523T, based on 16S rRNA gene sequence analysis, were selected and assessed for their biochemical and chemotaxonomic characteristics. The references strains, P. assamensis JCM 13186T, P. taiwanensis DSM 18679T, P. profundus NRIC 0885T, P. aquistagni LMG 29561T, and P. marinisediminis KACC 16317T were obtained from the Japan Collection of Microorganisms (JCM, Tsukuba, Japan), the German Collection of Microorganisms and Cell Cultures (DSMZ, Brunswick, Germany), the NODAI Culture Collection Center (NRIC, Okada, Japan), the Belgian Coordinated Collections of Microorganisms/Bacteria Collection (BCCM/LMG, Ghent, Belgium), and the Korean Agricultural Culture Collection (KACC, Jeonju, Republic of Korea), respectively.

Morphological, phenotypic and biochemical characteristics

Cells grown to exponential phase, on MA plates at 30 °C for 2 days, were observed using a DM 1000 light microscope (Leica, Wetzlar, Germany), and their morphology (cell size, shape, appearance, and texture) was identified using a JEM-1200EX transmission electron microscope (JEOL, Tokyo, Japan). Gram staining of cells was performed using a Gram staining kit (bioMérieux, Craponne, France). Motility was determined in MB cultures using the hanging-drop method (Bowman 2000). Endospore formation was observed using malachite green staining after cell growth for 4 days. The growth of CAU 1523T was evaluated on several media, including MA, brain–heart infusion agar (BHI; Difco), nutrient agar (NA; Difco), and tryptic soy agar (TSA; Difco). Growth was examined on MA for 10 days at various temperatures (4, 10, 20, 25, 30, 37, and 45 °C) and pH values (4.0–12.0, at intervals of 1.0 pH unit) using sodium acetate buffer for pH 4–5, phosphate buffer for pH 6–8, and carbonate buffer for pH 9–10. Oxidase activity was determined using an oxidase reagent kit (bioMérieux) with 0.1% (w/v) tetramethyl-p-phenylenediamine. Catalase activity was assessed by observing bubble production in a 3% (v/v) H2O2 solution (Cappuccino and Sherman 2010). Casein and gelatin hydrolysis, nitrate reduction, and indole and urease production were examined using API 20NE test strips (bioMérieux). Assimilation of carbon sources and enzyme activities were evaluated using API 50CH, and API ZYM systems (bioMérieux) at 30 °C for 2 days, according to the manufacturer’s instructions.

Siderophore production

The isolate culture was grown on MA plates at 30 °C until cells reached stationary growth, and cells were subsequently harvested via centrifugation at 5000×g for 15 min. The supernatant was carefully removed, passed through a 0.45 µm pore size filter, and stored at − 20 °C for siderophore determination. Siderophore production was determined using the CAS assay (Louden et al. 2011). Briefly, the blue CAS dye was prepared by dissolving 60.5 mg CAS in 50 mL distilled water (Solution 1), and 27 mg of FeCl3–6H2O was dissolved in 10 mL of 10 mM HCl (Solution 2). Solution 1 and 2 were then mixed, and 73 mg hexadecyl trimethyl ammonium bromide dissolved in 40 mL distilled water was added to the resulting solution. The prepared blue CAS dye solution was autoclaved and stored at 20 °C. CAS agar was prepared by mixing 100 mL of M9 Minimal Salt solution with 750 mL distilled water, to which 32.24 g piperazine bis(2-ethanesulphonic acid) PIPES and 15 g Bacto agar were added. This mixture was autoclaved and cooled to 50 °C; then 30 mL sterile casamino acid was added, followed by 100 mL of the prepared blue CAS dye solution. The presence of yellow/orange zones around cell growth indicated siderophore production (John and Thangavel 2016).

Phylogenetic analysis

Genomic DNA of CAU 1523T was extracted and purified using a genomic DNA extraction kit (iNtRON Biotechnology, Seongnam, Republic of Korea). The 16S rRNA gene sequence of CAU 1523T was amplified using the primers 27F and 1525R (Lane 1991), sequenced using the BigDye Terminator 66 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA), and assembled using an automated 3730 DNA Sequencer (Applied Biosystems). Multiple sequence alignment of CAU 1523T was performed via ClustalX 2.1 software (Larkin et al. 2007), and neighbour-joining analysis was performed using MEGA version 7.0 software (Kumar et al. 2016). The 16S rRNA gene sequence of CAU 1523T was aligned with that of other Paenibacillus species obtained from the NCBI and EzBioCloud database. The phylogenetic trees of CAU 1523T and closely related strains were constructed using neighbour-joining (Saitou and Nei 1987), maximum-likelihood (Felsenstein 1981), and maximum-parsimony (Fitch 1971) methods based on the algorithms of the Jukes–Cantor model (Jukes and Cantor 1969). Confidence values of branches on the phylogenetic trees were assessed by performing bootstrap resampling with 1000 replicates (Felsenstein 1985).

Genome sequencing and functional analysis

The genome of strain CAU 1523T was sequenced by Macrogen (Seoul, Republic of Korea) using the Illumina Hiseq 2500 platform (Illumina Inc., San Diego, CA, USA). Library preparation was conducted using the Illumina TruSeq DNA Kit, and de novo assembly of the sequence data was performed using SPAdes version 3.13.0 software (http://cab.spbu.ru/software/spades). Comparative genomic analysis of these strains (except P. marinisediminis KACC 16317T) was performed using genome information provided from the NCBI database.

The genome of CAU 1523T was annotated through the NCBI Prokaryotic genome annotation pipeline (Tatusova et al. 2016). The genomic G + C content was determined from the whole genome analysis of CAU 1523T. The average nucleotide identity (ANI) value was calculated using the orthoANI algorithm (http://www.ezbiocloud.net/tools/ani). The in silico digital DNA–DNA hybridisation (dDDH) value was determined using the recommended formula 2 value via the Genome–Genome Distance Calculator software (http://ggdc.dsmz.de/distcalc2.php). The average amino acid identity (AAI) was determined using the AAI calculator developed at the Kostas Lab (Rodriguez-R and Konstantinidis 2014). Phylogenomic analysis was performed using the up-to-date bacterial core gene set (UBCG) pipeline (Na et al. 2018). OrthoVenn 2 software was used to identify and compare orthologous genes in the draft genome sequence of the isolate with those in the genomes of closely related strains (Xu et al. 2019). The secondary metabolic gene clusters were predicted using antiSMASH version 5.1.0 software (Blin et al. 2019).

Chemotaxonomic characterization

For cellular fatty acid profile analysis, CAU 1523T and closely related strains were grown on TSA for 3 days at 30 °C; fatty acids were separated on a gas chromatograph fitted with an autosampler (Agilent Technologies, Santa Clara, CA, USA), according to the standard protocol of the Sherlock Microbial Identification System version 6.1 (MIDI Inc., Newark, DE, USA). Peaks were analysed using the MIDI software package based on the TSBA6 database (MIDI Inc.). Isoprenoid quinone was isolated according to a previously described method (Minnikin et al. 1984) and separated via reversed-phase high-performance liquid chromatography (HPLC). Polar lipids were extracted from the cells of CAU 1523T grown on MA and analysed via two-dimensional thin-layer chromatography (Minnikin et al. 1984) using 10% ethanolic molybdatophosphoric acid (for total lipids), molybdenum blue (for phospholipids), ninhydrin (for aminolipids), and α-naphthol/sulphuric acid reagent (for glycolipids) as previously described (Kim et al. 2015). The peptidoglycan cell wall was analysed as previously described (Schumann 2011).

Results and discussion

Phenotypic properties and screening for siderophore production

Strain CAU 1523T was pale pink colour when routinely cultured on MA agar. Cells were Gram-positive, rod-shaped, and endospore-forming at the terminal position in swollen sporangia. Cell sizes were 0.3–0.5 µm × 1.0–1.5 µm in diameter (Fig. S1). Cells could grow on TSA, R2A, NA, and MA media at 30 °C. The detailed morphological, phenotypic and biochemical characteristics between strain CAU 1523T and the related type strains are listed in Table 1, whereas the all the negative properties of strain CAU 1523T in commercial kits are given in Table S1. CAU 1523T differed from the most closely related species, P. assamensis JCM 13186T, based on its inability to use d-mannose, potassium gluconate, and dulcitol as carbon sources. Furthermore, CAU 1523T was able to produce siderophores in the CAS agar assay than reference strain P. taiwanensis DSM 18679T as shown in Fig. S2, demonstrating orange-coloured colonies due to siderophore-mediated removal of Fe from the blue dye.

Table 1 Differential characteristics of strain CAU 1523T and closely related type strains of the genus Paenibacillus
Full size table

16S rRNA gene phylogenetic analysis

The nearly complete 16S rRNA gene sequence of CAU 1523T was 1515 bp long (GenBank accession no. MN097941). This 16S rRNA gene sequence indicated that CAU 1523T belonged to the genus Paenibacillus, sharing high similarity with P. assamensis JCM 13186T (98.0%), P. taiwanensis DSM 18679T (95.3%), P. profundus NRIC 0885T (95.0%), P. aquistagni LMG 29561T (94.7%), and P. marinisediminis KACC 16317T (94.6%). The phylogenetic position, based on 16S rRNA gene sequences analysis, indicated that CAU 1523T clustered with P. assamensis JCM 13186T (Fig. 1), but was clearly separated from other members of the genus Paenibacillus. This result was further supported by the genomic features described below.

Fig. 1
figure 1

Neighbour-joining (NJ) phylogenetic tree based on the 16S rRNA gene sequence from Paenibacillus arenosi CAU 1523T. Dots show that the corresponding nodes were recovered in phylogenetic trees generated with the maximum-likelihood (ML) and maximum-parsimony (MP) algorithms. The numbers at the nodes indicate levels of bootstrap values as percentages obtained from the NJ/ML/MP analyses of 1000 resampled datasets; only values > 70% are shown. Bacillus amyloliquefaciens DSM7T was used as an outgroup organism. Scale bar represents 0.01 substitutions per nucleotide position

Full size image

Genomic features

The draft genome sequence of CAU 1523T had a total length of 5,307,553 bp (GenBank accession no. JACYTN000000000), containing 57 contigs, 4585 protein-coding genes, 14 rRNAs (six 5S rRNAs, five 16S rRNAs, and three 23S rRNAs), 69 tRNAs, and an N50 length of 172,680 bp. The sequence contained 57 scaffolds and a G + C content of 43.2 mol%. The general genomic characteristics of CAU 1523T in this study and closely related strains obtained from the NCBI database are listed in Table 2.

Table 2 Genomic features of strain CAU 1523T and closely related type strains of the genus Paenibacillus
Full size table

The phylogenomic tree based on the 92 concatenated core genes revealed that CAU 1523T formed a distinct clade with P. assamensis JCM 13186T (Fig. 2). The ANI and dDDH values between CAU 1523T and the most closely related species, P. assamensis JCM 13186T, were 93.2 and 51.1%, respectively. These values were below the 94–96% ANI and 70% dDDH cut-off values for delineation of a bacterial species (Meier-Kolthoff et al. 2013). The average AAI value between CAU 1523T and P. assamensis JCM 13186T was 94.3%, which was also below the threshold value of 96% (Konstantinidis and Tiedje 2005), as shown in Table S2. Based on the core genomes, the four related strains shared 7602 gene clusters comprising 7598 orthologous gene clusters (at least containing two species) and four single-copy gene clusters (Fig. 3A). The genomes of all strains shared 1560 gene clusters, while the genome of CAU 1523T contained 3788 gene clusters, of which 412 were unique to it. The heatmap matrix for pairwise genomic comparisons in Fig. 3B illustrates the orthologous gene clusters between CAU 1523T and closely related strains. These genomic comparisons provided strong evidence that CAU 1523T represented a novel species in the genus Paenibacillus.

Fig. 2
figure 2

Phylogenomic tree based on 92 core gene sequences from Paenibacillus arenosi CAU 1523T. The core genome sequence of Bacillus amyloliquefaciens DSM7T was used as an outgroup. Gene support index (GSI) values are shown at branching points. Bar, 0.05 substitutions per position

Full size image
Fig. 3
figure 3

Comparison of orthologous genes among Paenibacillus strains. A Venn diagram showing the distribution of shared orthologous gene clusters. B Heatmap showing the orthologous clusters between any pair of genomes

Full size image

Biosynthetic gene clusters for secondary metabolites

The antiSMASH software predicts biosynthetic gene clusters (BGCs) involved in the production of secondary metabolites based on whole-genome sequences. In this study, the genomes of CAU 1523T and closely related type strains were analyzed for BGCs to compare their biosynthetic potential, as shown in Fig. 4. The genomes of all strains contained high numbers of clusters associated with secondary metabolites, including non-ribosomal polyketide synthetases and trans-acyltransferase polyketide synthase fragments. Interestingly, the genome of CAU 1523T contained a siderophore gene cluster that was predicted to encode vibrioferrin, which was not identified in the other reference strains. Moreover, other gene clusters identified in the genome of CAU 1523T were associated with the synthesis of paenilamicin (100% gene similarity) and bacitracin (22% gene similarity), which have antibacterial properties (Garcia-Gonzalez et al. 2014). In addition, the genome of CAU 1523T contained a cluster associated with bacillibactin (53% gene similarity), a catechol-based siderophore secreted by the genus Bacillus. The BGC results suggested that CAU 1523T had a greater potential to synthesise secondary metabolites, notably siderophores, than the other closely related type strains.

Fig. 4
figure 4

Biosynthetic gene clusters (BGCs) identified in the genomes of CAU 1523T and closely related type strains using antiSMASH version 6.0.1 software

Full size image

Chemotaxonomic characterization

The major polar lipids of CAU 1523T were diphosphatidylglycerol and phosphatidylethanolamine, which were also the predominant polar lipids found in the other closely related species (Lee et al. 2013; Romanenko et al. 2013). In addition, the polar lipid profile of CAU 1523T included one unidentified phospholipid (PL1), one unidentified lipid (L1), and two unidentified aminophospholipids (APL1–3) (Fig. S3). The predominant isoprenoid quinone of CAU 1523T was menaquinone-7 (MK-7), which is one of the common features of the genus Paenibacillus (Ash et al. 1993). The major fatty acids (> 10% of the total) of CAU 1523T were anteiso-C15:0 (46.4%) and C16:1 ω11c (10.5%) (Table S3). The fatty acid profile of CAU 1523T was distinguished from that of closely related strains by its higher proportion of C16:1 ω11c. The diagnostic diamino acid present in the cell wall of CAU 1523T was meso-diaminopimelic acid, which is a common feature among members of the genus Paenibacillus (Ash et al. 1993).

Conclusion

The physiology, phenotypic characteristics, and genomic features of CAU 1523T support this strain as a novel species in the genus Paenibacillus, for which the name Paenibacillus arenosi sp. nov. is proposed.

Description of Paenibacillus arenosi sp. nov.

Paenibacillus arenosi (a.re.no’si. L. gen. n. arenosi, of a sandy place).

Cells are Gram-strain-positive, strictly aerobic, motile, siderophore-producing, endospore-forming (terminal), and rod shape approximately 0.3–0.5 µm in diameter and 1.0–1.5 µm in length. After incubation on MA, colonies are pale pink in colour, circular, round and convex. Optimum growth occurs in the presence of 0–3.0% (w/v) NaCl (optimum, 1%), at temperature range of 20–45 °C (optimum, 30 °C), pH range of 6.5–9.0 (optimum, 7.5). Hydrolyses starch and aesculin, but not gelatin, urea and casein. Nitrate and nitrite are not reduced. In the API 20NE tests, 4-nitrophenyl-βd-galactopyranoside and d-glucose are positive. In the ZYM tests, alkaline phosphate, esterase lipase (C8), acid phosphatase and naphthol-AS-BI-phosphohydrolase are positive. In the 50CH tests, glycerol, d-ribose, Methyl-α d-glucopyranoside, N-acethylglucosamine, amygdalin, d-maltose, sucrose, d-trehalose, glycogen, gentiobiose and d-turanose are assimilate. The major fatty acids are anteiso-C15:0 and C16:1 ω11c. The polar lipids are diphosphatidylglycerol, phosphatidylethanolamine, one unidentified phospholipid, one unidentified lipid, and two unidentified aminophospholipids. The diagnostic diamino acid of peptidoglycan is meso-diaminopimelic acid. The DNA G + C content is 43.2 mol%.

The type strain CAU 1523T (= KCTC 43108T = MCCC 1K04063T) was isolated from sea sand, Republic of Korea.