Introduction

Leakage of crude oil and its derivatives to environment is one of the crucial contaminating factors of soil, air, and underground water (Bezza and Chirwa 2015; de Silva et al. 2014). Although some bacterial strains can degrade oil compounds, the low water solubility and hydrophobic characteristic of oil compounds cause low bio accessibility of these compounds for microbial digest (Liang et al. 2016).

Biosurfactant secretion is one of the main employed strategies in microorganisms for absorbing PAH aromatic hydro-carbons and hydrophobic compounds (Bezza and Chirwa 2015). Bacillus subtilis is an aerobic, rod-shaped, and GRAS (generally recognized as safe) bacterium (Sharma and Satyanarayana 2013). B. subtilis produces biosurfactant factors such as non-ribosome peptides (nrps) that is used for bioremediation of hydrocarbons (Bezza and Chirwa 2015) and improvement of enhanced oil recovery (Shibulal et al. 2014).

The produced peptide biosurfactant by B. subtilis has a range of activities from anti-microbial activities to eliminator agent in contaminated soils. Three main lipo-peptide compounds of surfactin, iturin, and fengycin families are produced by these bacterial strains (Ben Ayed et al. 2014). These molecules have various advantages compared to with chemical surfactants such as sustainability, lower toxicity, higher bio-degradation capability, ecological adaptability, higher foam ability, higher selectivity, and specific activity. Furthermore, these strains can work on harsh conditions of high temperature, salinity, and pH (Ben Ayed et al. 2014; Bezza and Chirwa 2015; Jha et al. 2016).

B. subtilis has been also recognized as a model organism (Kamada et al. 2015); whole genome sequencing provided valuable information relating to biological functions, gene conservation, variation among specious and involved metabolic pathways for producing biosurfactants, and also oil bioremediation through sequence annotation (Sharma and Satyanarayana 2013).

The high performance of new generation of sequencing technology, its reasonable costs, and its higher efficiency compared to the first-generation sequencing have elaborated the insights into the bacterial whole genome sequencing (Kamada et al. 2014). This technology will be an important sequencing tool in the microbial genome studies (Land et al. 2015).

Pac Bio sequencing technology, the third generation creates long reads with relative length of 8500–30,000 bp, which facilitate the manipulation of these reads in complex regions such as repetitive elements. Therefore, genomes can be assembled with higher accuracy and validity by using long reads of Pac Bio. (Hutchison et al. 2016; Koren et al. 2013).

After accurate genome assembly, the homology of sequences and the information from the other reference genome and close relatives can improve the annotation (Ali et al. 2013). As example, the biochemical characteristics of the biosurfactants can be revealed from functional genomics analysis of available Bacillus subtilis genomes (Shaligram et al. 2016).

In this study, we isolated as a new strain of B. subtilis from oil-contaminated soil in south of Iran and its full genome was sequenced by PacBio technology. Then, assembling, annotation, and genome comparison analysis carried out based on coding sequences in MJ01 strain.

Material and methods

Growth conditions and preparation of genomic DNA

Bacillus subtilis MJ01 was grown aerobically in Luria Bertuni (LB) medium for 24 h under 35 °C and 260 rpm. Genomic DNA of this bacterium was extracted from LB medium using genomic DNA purification kit MG™ (Macrogen; Seoul, Korea).

Genome sequencing and assembling

To this end, 8 μl of purified genomic DNA was used using segmented g-TUBE and AMPure Bp magnetic willows. SMRTbell Template Prep Kit 1.0 was used to prepare library. Then, sequencing carried out based on PacBio RS system in MACROGEN Company (Seoul, Korea). Sequenced reads were filtered, mapped, and assembled by HGAP3 protocol and SMRT analysis v.2.3.0.140936 software (Rhoads and Au 2015).

Submitting nucleotide sequences

The whole genome sequence of this bacterium was deposited in NCBI data base with accessibility Number CP-018173. The used version in this study is the first genome version.

Genome annotation, gene prediction, and coding zones

Genome interpretation, scanning, and gene prediction carried out by MicroScope platform (Vallenet et al. 2009; Vallenet et al. 2013). AntiSMASH v.3.05 (Weber et al. 2015) in MicroScope platform was used for identifying coding zones of secondary metabolites and non-ribosome peptides. In addition, BAGEL3 web-based database was used for predicting the coding sequences of Bacteriocin those assumed in B. subtilis MJ01 genome (Heel, Jong, Montalban-Lopez, Kok and Kuipers 2013). Resistance gene identifier (RGI) in CARD database was used for predicting coding sequences that are resistant to anti-biotic (McArthur and Wright 2015).

PHASTer was applied for recognition, interpretation, and indication of prophage sequences in MJ01 strains (Arndt et al. 2016). Moreover, genomic islands were detected by using on-line Web-based tool Island Viewer v.3 (Dhillon et al. 2015).

Comparative genomics analysis and protein classification

The comparative genomic sections of MicroScope platform such as pan and core genome and MAUVE alignment were used for comparing MJ01 genomes with reference genome strains of B. subtilis subsp. Subtilis str. 168 (NC-000964) and close relative B. subtilis subsp. spizizenii TU-B-10 (CP-002905). The classification of functional proteins carried out based on COG classification and total genome information by genomic tool of MicroScope platform. OrthoVenn platform was applied for comparing and interpreting orthologous gene cluster (Wang et al. 2015).

MJ01 genome was compared with whole genome of 45 strains of B. subtilis in NCBI database using GGDC v. 2.1 (Genome-Genome Distance calculator) online tools (Auch et al. 2010). Moreover, JSpecieWS was used to calculate ANI and four nucleotide correlation index (Richter et al. 2016).

Bacterial genome of MJ01 was entered to pubMLST online tool (http://pubmlst.org/) by using multi locus sequencing typing approach (MLST) for detecting taxonomic similarity in genetically loci of seven housekeeping genes (rpoD, tpiA, pycA, purH, glpF, pta, and ilvd) (Jolley and Maiden 2010).

Results

Genomic characteristics

MJ01 bacterial genome consists of a chromosome contig with sequence length of 4,108,293 bp with of 43.93% GC and 4269 coding sequences (Table 1). Table 1 shows total characteristics of MJ01 compared to spizizenii TU-B-10 and 168 strains. The length of MJ01 genome is 2% shorter than two other genomes of B. subtilis, while their GC percentages were relatively equal. Most of the strains of this group have 4 Mb length, and their GC percentages are between 43 and 44% (Fig. S1).

Table 1 Statistical comparison of information and characteristics of genomes of B. subtilis MJ01, spizizenii TU-B-10, and 168 strains
Full size table

This genome includes 10 operons with 3 genes for rRNA and totally 30 rRNA genes were predicted in genome. The comparative consideration of rRNA operon numbers showed that the number of rRNA were similar for all three strains. In addition, 86 tRNA coding genes were detected for 20 standard amino acids on MJ01 chromosome. In terms of tRNA number, MJ01 genome was similar to 168 and both of them had 86 tRNA coding genes. In contrast, 90 sequences that code tRNA were observed in spizizenii TU-B-10.

The analysis of 16SrRNA sequence showed that MJ01 bacterium was 100% similar to 16SrRNA sequence of B. subtilis strain AER314-2; on the other hand, B. subtilis subsp. Spizizenii TU-B-10, Bacillus sp. JS, and B. subtilis strain BS3902 were similar (Fig. 1).

Fig. 1
figure 1

Phylogenetic graph based on 16SrRNA sequence. The status of B. subtilis MJ01 is demonstrated versus the other relatives of B. subtilis

Full size image

COG (clusters of orthologus groups) analysis was classified 3972 proteins out of 4218 protein coding sequences, which were predicted in MJ01 genome (Fig. 2). COG class divided coding proteins of MJ01 to four main groups of signaling and cell processes, information processing and storage, metabolism and fewer known and 21 classes (Table S1). The most coding sequences (about 39%) were grouped in metabolism class. MJ01 bacterium is the producer genome of secondary metabolites similar to the other strains of B. subtilis. Therefore, the presence of sequences (107 coding sequences) that produce secondary metabolites in genome of this bacterium is significant (about 2.57%).

Fig. 2
figure 2

Classification of functional proteins of MJ01 genome based on clusters of orthologus groups (COG)

Full size image

Comparative genomics analysis of MJ01 strain with other B. subtilis strains

Whole-genome alignment by using MAUVE

Whole-genome alignment carried out for four genomes with the most similarity level, that carried out by BLAST search of MAUVE genome alignment software V. 20150226 and included spizizenii TU-B-10, W23, NRS 231, and T30 and MJ01 strains (Fig. 3). Three conserve blocks were detectable in full genome alignment.

Fig. 3
figure 3

Multi alignment of MJ01 genome compared to genomes of its close relatives B. subtilis, spizizenii TU-B-10, spizizenii W23, spizizenii NRS231 and T30. The same colored blocks mean conservation and homology between genomic zones

Full size image

Furthermore, MJ01 genome was aligned with the genome of 168 strain as reference genome of subtilis subgroup. The result of alignment of these two genomes showed that although there are conservative blocks between two genomes, non-homolog zones exist between two genomes (Fig. 4). Genome alignment shows that organization of MJ01 genome and its conservative zones are similar to strains in spizizenii group.

Fig. 4
figure 4

Dual alignment of B. subtilis MJ01 and reference strain of B. subtilis subsp. subtilis str. 168 using Mauve software. The same colored blocks mean conservation and homology between genomic zones. Some parts of MJ01 and 168 genomes had not homology

Full size image

Hybridization of DNA-DNA by using GGDC

Genome-to-genome distance calculate based amount of digital DNA-DNA hybridization (DDH) showed that MJ01 genome had the most DDH with the genomes of bacteria that are subgroup of spizizenii TU-B-10, W23, and NRS231 strains at the level of formula 1, 94.7, 93.6, and 93.6%, respectively (Table S2).

Average nucleotide identity

Analysis of four nucleotide correlations among all whole genomes and contings of MJ01 in JSpeciesws database (Richter et al. 2016) shows that the genome of MJ01 bacterium had correlation with 15 genomes with more than 0.999 in z-score domain (Table S3). The genome of B. subtilis subps. Spizizenii TU-B-10 with z-score = 0.9999 had the most correlation. Consequently, the contig of Jeotagalibacillus marinus DSM 1297 (separated from unknown sediment source) is located at the second rank and after that the contig of B. subtilis JRS7 (separated from dessert soil) with the most z-score (0.99959 and 0.9995, respectively). Among completed genomes, the most z-score was calculated for B. subtilis subsp. spizizenii W23 and NRS 231 (z-score = 0.99946 for both of them). The results of two methods of nucleotide similarity calculation (based on ANIb or BLAST and ANIm or Mummer) show that MJ01 had the most ANIb with the genome of bacteria that are subgroup of spizizenii such as TU-B-10, NRS 231, and W23, 99.13, 96.52, and 96.52%, respectively. In addition, ANIm amounts were 99.25, 96.74, and 96.73 for these strains. ANI analysis confirmed that the MJ01 genome can be located as a subgroup of spizizenii bacteria.

Multi locus sequence typing

Search of pubMLST data base for locus sequence of seven housekeeping genes rpoD, ilvD, pta, purH, pycA, glpF, and tpiA for B. subtilis in MJ01 genome shows that three genes of pta, ilvD, and rpoD had similar locus with same directions. Gene locus of tpiA, pycA, purH, and glpF did not follow same direction in the database. After considering locus sequencing typing of housekeeping genes on MJ01 genome by online tool in pubMLST web site, the result confirmed that the profile of MJ01 is exclusive to this bacterium and it has not been recorded in pubMLST web site so far. Therefore, the most similar profile belongs to B. subtilis bacterium subgroup of spizizenii and BGSC3A17 strain isolated No. 10 (Table 2).

Table 2 The results of searching profile of multi locus sequencing typing (MLST) in MJ01 genome of PubMLST data base
Full size table

Pan and core genome

Pan and core analysis of four genomes of natto BEST 195, W23, TU-B-10, and 168 strains along with MJ01 in MicroScope platform (Fig. 5) demonstrates that interpreted genes in 5 genomes have formed a gene pool with 21,704 genes. These genes have been classified into 6622 gene family based on 80% similarity of amino acid and 80% alignment coverage (Table 3).

Fig. 5
figure 5

Pan-genome analysis of B. subtilis close relatives with MJ01 genome in MicroScope platform. Distribution of gene families in the core and strain-specific genome was 70/30

Full size image
Table 3 The number of genes involved in pan-genome analysis of five relative bacterial strains. The results show that B. subtilis MJ01 genome has fewer strain-specific gene
Full size table

MJ01 strain-specific genes

Genome pan/core analysis showed that 279 coding sequences that are strain-specific have been detected in MJ01 genome. They include 2.66% of coding sequences in MJ01 genome. Fewer strain-specific genes in MJ01 compared to the other strains. The fewer amounts of strain-specific genes are not out of expectation according to shorter length of the genome and fewer predicted coding sequences. These strain-specific genes included phage, regulator, and unknown proteins (Fig. 6).

Fig. 6
figure 6

Classification of strain-specific genes for B. subtilis MJ01 strain. The result showed that the source of most of these genes is unknown and some of them are the results of horizontal transference of gene through prophages or transposon elements

Full size image

Non-ribosome proteins and anti-biotic coding sequences

Detection and prediction of secondary metabolites and non-ribosome lipo peptide

The results of MJ01 genome analysis by AntiSMASH on line tool V3.0.5, 11 gene zones for production of secondary metabolites, and non-ribosome lipo proteins are presented in Table 4. Four gene zones that were responsible for coding non-ribosome lipo proteins (NRPS) in MJ01 genome were detected including Bacillibactin, Surfactin, Bacilliacne, and Mycosubtilin. Moreover, other vulnerable sections that coding NRPS were detected such as Basilysin in position of 737,237 to 738,655 and a section that codes polyketide in positions of 3,273,643 to 3,274,740. Two involved gene zones in synthesis of lassopeptide and sactipeptide bacteriocins classes were predicted in the genome of MJ01 bacterium. In addition, there are two zones that produce terpenes in this genome.

Table 4 The result of analyzing B. subtilis MJ01 genome by using AntiSMASH online tool that predicts coding sequences of secondary metabolites
Full size table

Surfactin coding operon locates on the positions of 1,458,804 to 1,484,194 of MJ01 genome. This operon is similar to gene cluster of data base in terms of gene sequence (86%) (Figs. 7 and 8). Studying homolog gene cluster of this operon on relative bacteria showed that this operon has the most similarity (97%) with its homolog in B. subtilis subsp. spizizenii W23 bacteria. The homolog similarities of this operon with 168 and spizizenii TU-B-10 (the closest relative) are 91 and 78% in order (Fig. 7).

Fig. 7
figure 7

The cluster of gene that biosynthesizes surfactin lipo peptide and its predicted structure in MJ01 and its comparison with gene cluster of other bacterial specious existing in the data base

Full size image
Fig. 8
figure 8

The comparison of gene clusters for surfactin lipo peptide biosynthesis between B. subtilis MJ01 and other bacterial strains. The conservative synteny zones are observable comparing with area of interest and aimed sequences. Different genes are determined with different colors and homolog genes have same colors

Full size image

According to the obtained data from AntiSMASH and annotation of MJ01 genome, the position of SrfA operon and sfp gene were detected on the genome. This operon has 4 genes of srfA-ABCD in positions of 1,458,804–1,484,950. Sfp gene locates on 5017 bp at lower part of surfactin operon that biosynthesizes surfactin.

Mycosubtilin producer operon has been predicted on the positions of 3,020,638–3,063,549 with the length of 42,912 bp on MJ01 genome. The seeking on the sequence of this gene cluster in the data base shows that it has 100% similarity with mycosubtilin gene cluster. This gene cluster is 100% similar with its homolog in close relative bacterium (spizizenii TU-B-10), and it is also 94% similar with reference genome of 168. By using achieved data from AntiSMASH analysis and interpretation of MJ01 genome, the exact position of this operon was detected on the genome. This operon has four genes of mycA, mycB, mycC, and fenF on the positions of 3,027,547–3,064,769 and a negative strand.

Entrance of MJ01 genome to BAGLE3 online tool predicts two areas of interest for accepted bacteriocins. The first area locates in lasso-peptide class and position of 373,960–383,960 (Fig. 9). The BLAST search of this 10,000 nucleotide area in NCBI data base shows that there is not any nucleotide sequence in the data base that is significantly similar with it. The result of search is only a sequence including 28 nucleotides of Galdieria sulphuraria H+-translocating PPase (vocuolar) (Gasu-15,749), mRNA. Galderia is a unicellular red alga. It seems that the BLAST result of nucleotides in this predicted area indicates that this sequence specifies to the genome of MJ01 and a new bacteriocin.

Fig. 9
figure 9

The predicted areas that prone to produce bacteriocin in MJ01 genome by BAGEL3 on line tool. Coding sequences of bacteriocin (green), transporter (red), adjusted areas (yellow), and rectifier section (blue) have been shown

Full size image

The second area locates in sactipeptides class and positions of 697,075–707,075 on MJ01 genome (Fig. 10). The nucleotide BLAST search of this area in NCBI data base shows that this area has nucleotide similarity in the genomes of B. subtilis subsp. Spizizenii TU-B-10, B. subtilis T30, B. subtilis subsp. spizizenii NRS231 and W23; the similarities are 99, 97, 97, and 97% with 100% coverage.

Fig. 10
figure 10

The predicted areas that prone to produce subtilosin A bacteriocin n MJ01 genome by BAGEL3 on line tool. Coding sequences of bacteriocin (green), transporter (red), and rectifier section (blue) have been shown

Full size image

Resistance to anti-biotic

The search on MJ01 genome was performed to find involve genes that are resistant to anti-biotic. This search carried out by RGI (resistance gene identifier) through CARD on microscope platform that has led to detection of 22 coding sequence in the genome and 19 CDS that had homology with other existed genes in the data base and 3 CDS were identified as strain, which had single nucleotide polymorphism (SNP) in CARD database. Among these coding sequences, there are membrane pumps for resisting against peptide anti-biotics such as lmrB and mprF; in addition, ykkD and bmr genes were effective on resistance to Chloramphenicol anti-biotic and penR gene was effective on hydrolysis of beta Lactam chain of penicillin (Table S4).

Prophage sequences existed in B. subtilis MJ01 genome

Based on the searches that were carried out by PHASTer on line tool, three prophage areas are observable in MJ01 genome. One area includes a prophage section with the score of more than 70 and less than 90 and two areas including incomplete prophage genes (Table 5, Fig. 11).

Table 5 Prophage prone areas in B. subtilis MJ01 genome; two incomplete prophage areas and a relatively complete prophage area in B. subtilis MJ01 genome have been predicted by Phaster on line tool
Full size table
Fig. 11
figure 11

The view of detected prophage areas in MJ01 genome by phaster on line tool. There are two incomplete prophage areas (red) and a relatively complete prophage area (blue) in the structure of MJ01 genome

Full size image

Discussion

The long read sequences of Bacillus subtilis MJ01 genome that were prepared by PacBio RS SMRT provide high quality and accuracy of genome assembly and show success in a creation of a scaffolding circular chromosome. This high quality is the result of applying proper platform with preferable and long reading size more than 10,000 nucleotides and the absence of vague bases. Therefore, no gap has been created for contig assembly and genome scaffolding. Therefore, there is no need to use common methods such as PCR for filling the gaps, contrasting to the common studies that use short reads of other platforms (Koren et al. 2013).

Analysis in genome scale provides information about genome organization and its similarity with the other bacterial relative strains of B. subtilis group. Based on these analyses, it was determined that total organization of MJ01 genome is very similar to subgroups of spizizenii. Obtained information and results from studying whole alignment of genome that carried out by Mauve tool, DDH method, and ANI calculation have confirmed this point. Multi locus sequencing typing (MLST) results at the level of housekeeping genes and profile analysis showed that although there are close similarities between B. subtilis MJ01 genes and their homologs in B. subtilis spizizenii TU-B-10, the structure of MJ01 has been subjected to events such as point mutations and inversion. This issue not only caused differentiation of MJ01 bacterium from its close relatives such as spizizenii-TU-B-10 but also it confirmed that MJ01 is a unique strain and probably a subgroup of spizizenii.

Distribution of gene families in analysis with relative ratio of 70 to 30 among core genome and strain-specific genome was not in line with the findings of Yu et al. (2015). They used 13 genomes of B. subtilis and analyzed pan-genome to find that distribution of gene families follows a balanced ratio of 50/50 among core genome and strain-specific genome; they concluded that existed genes in core genome of B. subtilis were under higher purification selective pressure than strain-specific genes (Yu et al. 2015).

Ortholog protein clustering and gene classification based on pan-core have detected unique functional coding sequences in MJ01 genome. The searches of these unique sequences in non-repeated nucleotide sequence database showed that under studied sequence has not any similar sequence in the database. Therefore, the studies on new sequences can draw attentions toward interpretation of newfound genes of B. subtilis bacteria genetics in future, and also can provide better insight through genomic information about similar metabolic activities in relative bacteria that have similar genetic machine and coding sequences (Harvey et al. 2015).

In conclusion, using genomic comparative attitude has indicated that MJ01 genome has similar genomic structure with spizizenii-TU-B-10. However, MJ01 genome has distinctive differences with spizizenii-TU-B-10 strain such as new bacteriocin coding sequence.