An operon consisting of a P-type ATPase gene and a transcriptional regulator gene responsible for cadmium resistances in Bacillus vietamensis 151–6 and Bacillus marisflavi 151–25

Background Cadmium (Cd) is a severely toxic heavy metal to most microorganisms. Many bacteria have developed Cd2+ resistance. Results In this study, we isolated two different Cd2+ resistance Bacillus sp. strains, Bacillus vietamensis 151–6 and Bacillus marisflavi 151–25, which could be grown in the presence of Cd2+ at concentration up to 0.3 mM and 0.8 mM, respectively. According to the genomic sequencing, transcriptome analysis under cadmium stress, and other related experiments, a gene cluster in plasmid p25 was found to be a major contributor to Cd2+ resistance in B. marisflavi 151–25. The cluster in p25 contained orf4802 and orf4803 which encodes an ATPase transporter and a transcriptional regulator protein, respectively. Although 151–6 has much lower Cd2+ resistance than 151–25, they contained similar gene cluster, but in different locations. A gene cluster on the chromosome containing orf4111, orf4112 and orf4113, which encodes an ATPase transporter, a cadmium efflux system accessory protein and a cadmium resistance protein, respectively, was found to play a major role on the Cd2+ resistance for B. vietamensis 151–6. Conclusions This work described cadmium resistance mechanisms in newly isolated Bacillus vietamensis 151–6 and Bacillus marisflavi 151–25. Based on homologies to the cad system (CadA-CadC) in Staphylococcus aureus and analysis of transcriptome under Cd2+ induction, we inferred that the mechanisms of cadmium resistance in B. marisflavi 151–25 was as same as the cad system in S. aureus. Although Bacillus vietamensis 151–6 also had the similar gene cluster to B. marisflavi 151–25 and S. aureus, its transcriptional regulatory mechanism of cadmium resistance was not same. This study explored the cadmium resistance mechanism for B. vietamensis 151–6 and B. marisflavi 151–25 and has expanded our understanding of the biological effects of cadmium.


Background
Soil contamination with heavy metals is becoming an increasingly urgent problem worldwide. Among all heavy metals, one of the most hazardous is cadmium (Cd) [1,2]. Cd is a non-essential heavy metal element that is toxic, teratogenic and carcinogenic to humans. When critical Cd levels in soil are reached, biodiversity, agricultural productivity, food safety and human heath can be threatened, as Cd can be accumulated in the food chain [3][4][5][6][7]. Consequently, solutions to remediate heavy metal-contaminated soil, in particular Cdcontaminated soil, are urgently needed. Many remediation techniques, such as chemical immobilization, electrokinetic extraction, phytoremediation, and bioremediation, have been proposed for soils contaminated with heavy metals [8]. Among these technologies, bioremediation is considered an innovative and promising remediation method based on its cost effectiveness and low environmental impact [9][10][11].
One of the most thoroughly characterized Cd 2+ resistance efflux systems is the czc (Cd 2+ , Zn 2+ , and Co 2+ resistance) system in the gram-negative bacterium Alcaligenes eutrophus [18,23]. The CzcA, CzcB and CzcC proteins comprise an active efflux pump complex driven by a cation-proton antiporter [18]. Briefly, CzcA, located in the inner membrane, is essential for cation transport and can remove heavy metals (Cd 2+ , Zn 2+ , and Co 2+ ) from the cytoplasm using a H + ion gradient [23,24]. CzcB belongs to a family of bacterial membrane fusion proteins. It may create a pathway for the removal of cations from CzcA [23,25]. CzcC relies on CzcB to function and may act as a substrate (Cd 2+ , Zn 2+ , and Co 2+ ) switch for the efflux pump [17].
Another well-characterized Cd 2+ resistance system is the cad system (cadA-cadC) in the gram-positive bacterium S. aureus [19,26]. The Cd-efflux ATPase is encoded by the cadA gene, which contains six predicted membrane-spanning regions. The fourth membrane span is thought to be involved in the cation translocation pathway, and includes a conserved Cys-Pro-Cys tripeptide [27]. CadC is a regulatory protein encoded immediately downstream of cadA, and is also required for Cd 2+ resistance in S. aureus [28]. CadC is a member of the ArsR/SmtB family [29,30], which can bind to the promoter-operator area of the cadA-cadC system and acts as a transcriptional repressor in vitro [31].
In present study, two different Cd 2+ resistance Bacillus sp. strains (B. vietamensis 151-6 and B. marisflavi 151-25) isolated from Cd-contaminated soil was grown in the presence of Cd 2+ at two distinct concentrations up to 0.3 mM and 0.8 mM, respectively. In order to compare the mechanism of Cd 2+ resistance for these two strains, their genome and transcriptome under Cd 2+ stress were analyzed. Moreover, a fosmid library from genomic DNA of B. vietamensis 151-6 was constructed to further delineate the Cd 2+ resistance-related genes. The functions of these genes were analyzed and verified through overexpression in E. coli and B. subtilis. The gene clusters of 4802-4803 in the plasmid p25 of B. marisflavi 151-25 and 4111-4112-4113 on the chromosome of B. vietamensis 151-6 played the major role on the Cd 2+ resistance.

Strain 151-6 and 151-25 exhibit different Cd 2+ resistances
To isolate Cd-resistant bacteria, a Cd-contaminated (59.561 ± 3.76 mg Cd kg − 1 ) soil sample was collected in Hunan Province. A total of 21 strains of bacteria were isolated using LB agar plates containing 0.5, 1 and 2 mM Cd 2+ . The MIC-Cd values of those isolates were determined and ranged from 0.4 to 1.0 mM (Additional file 3: Table S3). Among them, the strains 151-25 and 151-6, which showed the highest and lowest level of Cd 2+ resistance respectively, were selected for further study (Additional file 3: Table S3), and identified as Bacillus sp. through 16S rDNA analysis.
To compare the level of Cd 2+ resistance of two selected strains with other Bacillus sp. bacteria, we examined their MIC-Cd, as well as those of three other Bacillus sp. strains, B. subtilis WB600 (BS), B. amyloliquefaciens (BA) and B. licheniformis WX-02 (BL), in liquid medium supplemented with various Cd 2+ concentrations. 151-25 and 151-6 were grown in the presence of Cd 2+ at concentrations up to 0.8 mM and 0.3 mM, respectively ( Fig. 1), while the growth of the other three Bacillus sp. strains was inhibited in the presence of 0.3 mM Cd 2+ (Fig. 1). The growth curves at 0 mM Cd 2+ demonstrated that all four Bacillus sp. strains exhibited similar growth trends in liquid LB medium (Additional file 6: Figure S1a). However, under culture conditions with 0.1 mM or 0.3 mM Cd 2+ , 151-6 and 151-25 exhibited greater growth potential than the other three Bacillus sp. strains (Additional file 6: Figure S1b and c). And only 151-25 could grow at 0.5 mM Cd 2+ (Additional file 6: Figure S1d). These data indicated that compared to other Bacillus sp. strains, 151-6 and 151-25 exhibited stronger tolerance to Cd 2+ . And the Cd 2+ resistance of 151-25 was significantly higher than 151-6.

Genomic analysis of 151-6 and 151-25
To compare the mechanism of Cd 2+ resistance for 151-6 and 151-25, their genome were sequenced. The results showed that 151-6 contains a single chromosome (4,556, 861 bp and 4952 predicted genes) and one plasmid, designated p6 (40,946 bp and 66 predicted genes) (Additional file 7: Figure S2). And 151-25 also contains a single chromosome (4,411,234 bp and 4734 predicted genes) and one plasmid, designated p25 (138,020 bp and 156 predicted genes) (Additional file 8: Figure S3). Simultaneously, the results of non-redundant protein database (NR) annotation of the predicted coding genes presented that strain 151-6 can be identified as B. vietamensis (4093 of the 5018 genes are annotated to derive from B. vietamensis) and strain 151-25 canbe identified as B. marisflavi (4095 of the 4890 genes are annotated to derive from B. vietamensis).
Plasmid p25 in B. marisflavi 151-25 is important for Cd 2+ resistance, while p6 in B. vietamensis 151-6 is not relevant to the Cd 2+ resistance To determine whether plasmid p6 and p25 were responsible for Cd 2+ resistance in B. vietamensis 151-6 and B. marisflavi 151-25, we eliminated the plasmids and then examined their contribution to Cd 2+ resistance. The plasmid p6-eliminated strains 151-6△5 and 151-6△7 and the plasmid p25-deleted strains 151-25△3 and 151-25△29 were obtained (Additional file 9: Figure S4a and Additional file 10: Figure S5a, respectively) and further verified by PCR (Additional file 9: Figure S4b and Additional file 10: Figure S5b, 5c, respectively). In liquid media, the strains 151-6△5 and 151-6△7 exhibited the same MIC-Cd as the wild strain 151-6 ( Fig. 2a). And the growth curve of 151-6△5 and 151-6△7 did not show any differences with wild type 151-6 at the three different concentrations of Cd 2+ (Fig. 2b, c and d). These data indicated that the elimination of plasmid p6 does not influence the Cd 2+ resistance of B. vietamensis 151-6. However, eliminating plasmid p25 caused a significant decreased of Cd 2+ resistance in B. marisflavi 151-25. As shown in Fig. 3a, the MIC-Cd values of 151-25△3 and 151-25△29 were 0.2 mM, markedly lower than that of wild-type 151-25 (1.0 mM). Elimination of plasmid p25 did not affect the growth of the strain in liquid LB medium without added Cd 2+ (Fig. 3b). However, with addition of 0.1 or 0.5 mM Cd 2+ to the medium, the growth of 151-25△3 was inhibited, while the growth of B. marisflavi 151-25 was not affected ( Fig. 3c and d, respectively). Thus, we concluded that plasmid p25 is important for Cd 2+ resistance in B. marisflavi 151-25 while plasmid p6 is not relevant to the Cd 2+ resistance in B. vietamensis 151-6. To determine the genes related to the Cd 2+ resistance of B. marisflavi 151-25, its transcriptome under Cd 2+ induction was analyzed by RNA sequencing. Fragments per kilobase of exon model per million mapped reads values were used to measure transcript abundance. As shown in Additional file 4: Table S4 and Additional file 11: Figure S6, a total of 65 differentially expressed genes were identified in the presence of Cd 2+ (47 up-regulated genes and 18 down-regulated genes). Among the 47 up-regulated genes, 9 were located on plasmid p25, and the top three up-regulated genes in terms of fold change were also from plasmid p25 ( Table 1 and Additional file 4: Table S4). Briefly, transcripts of a transcriptional regulator gene (orf4803) and a copper-translocating P-type ATPase gene (orf4802) were increased 72.53-and 63.83-fold, respectively. These results were confirmed through qRT-PCR (Fig. 4b), and it further indicated that plasmid p25 is related to Cd 2+ resistance in B. marisflavi 151-25. Based on these results, we hypothesize that nine up-regulated genes located on plasmid p25, orf4774, 4775, 4776, 4777, 4779, 4781, 4782, 4802 and 4803 may be involved in Cd 2+ resistance of B. marisflavi 151-25. To confirm this hypothesis, we overexpressed these genes using their own promoter with five gene clusters ( Fig. 4a) in E. coli and B. subtilis, to examine their contributions to Cd 2+ resistance. As shown in Fig. 4c and d, the gene cluster containing orf4802 and orf4803 allowed recombinant E. coli and B. subtilis to exhibit higher Cd 2+ resistance than negative control strains. In particular, B. subtilis cells containing the recombinant plasmid 4802-4803-pUBC19 would grow in the presence of Cd 2+ at seven times the concentration tolerated by control cells. However, compared to control cells, the other four recombinant E. coli and B. subtilis strains did not show elevated Cd 2+ resistance. These results indicate that the gene cluster consisting of orf4802 and orf4803 confers high Cd 2+ resistance in B. marisflavi 151-25.
To find more genes responsible for high Cd 2+ resistance of B. marisflavi 151-25, we also compared gene expression between B. marisflavi 151-25 and 151-25△3 under Cd 2+ stress. Nineteen up-regulated genes located on the chromosome that exhibited a more than 4-fold change according to transcriptome data were analyzed using qRT-PCR. Compared to wild-type B. marisflavi 151-25, the transcript levels of seven genes showed  Figure S7a). These genes were orf666 (TetR family transcriptional regulator gene), orf667 (cysteine ABC transporter substrate-binding protein gene), orf668 (ABC transporter permease gene), orf1240 (ArsR family transcriptional regulator gene), orf1241 (copper-translocating P-type ATPase gene), orf3892 (hypothetical protein gene) and orf3894 (cation transporter gene). To confirm whether these genes were related toCd 2+ resistance for B. marisflavi 151-25, three gene fragments also with 4802-4803 were overexpressed in B. subtilis and the Cd-MIC values for those recombinant strains were determined. As shown in Additional file 12: Figure S7b, the operon containing orf4802 and orf4803 allowed recombinant B. subtilis to exhibit greater Cd 2+ resistance, while the other three fragments would not increase Cd 2+ resistance of the recombinant B. subtilis. Our results suggested that the operon of 4802-4803 plays a leading   Figure  S8 and Additional file 5: Table S5). Moreover, 585 upregulated genes in the presence in Cd 2+ were located on the chromosome, and among the 683 down-regulated genes, only 16 were located on plasmid p6. These data verified that plasmid p6 does not influence the Cd 2+ resistance of B. vietamensis 151-6. To further identify the key Cd 2+ resistance genes from those differentially expressed genes, a fosmid library of B. vietamensis 151-6 genomic DNA was constructed in E. coli. And the clones were screened by the Cd 2+ resistance to determine the genes that contribute to Cd 2+ resistance. Among the total of 1204 clones, there were 25 and 3 clones were able to grown on LB agar plates with 1.2 mM and 1.5 mM Cd 2+ , respectively. In liquid LB medium with 1.2 mM Cd 2+ , there were 27 clones could be grown. Only 2 clones, B2 and C2, could be grown in liquid LB medium with 1.5 mM Cd 2+ . The MIC-Cd of B2 and C2 were determined. As shown in Additional file 14: Figure S9, B2 and C2 exhibit higher Cd 2+ resistance than negative control strain EPI300-T1 R . The fosmid DNAs isolated from B2 and C2 were sequenced. Sequence analyses of B2 revealed that it contained 32, 263 bp insert fragment, which consisted 31 annotated genes ( Table 2) (And the fosmid DNA isolated from C2 was failed by sequencing many times, so the clone was not analyzed). Moreover, the analysis of the transcriptome showed that insert fragment of B2 contained 8 upregulated genes (orf4108, orf4109, orf4088, orf4087, orf4090, orf4093, orf4106, orf4107) and 3 down-regulated genes (orf4086, orf4101, orf4120). These results were confirmed through qRT-PCR and the transcription levels of other 16 genes in the insert fragment of B2 were also evaluated by qRT-PCR. The results presented that the transcripts of the orf4108, orf4109, orf4088, orf4087, orf4104, orf4106 and orf4107 were increased 156.64,   Fig. 5c and d, the gene cluster containing orf4111 (copper-translocating P-type ATPase gene), orf4112 (cadmium efflux system accessory protein gene, cadC) and orf4113 (cadmium resistance protein gene, cadD) allowed recombinant E. coli and B. subtilis to exhibit higher Cd 2+ resistance than the negative control strain. Especially, the B. subtilis cells containing the

Discussion
The present study showed that the gene cluster 4802-4803 located onA plasmid p25 and the gene cluster 4111-4112-4113 located on chromosome were involved in Cd 2+ resistance for strain B. marisflavi 151-25 and B. vietamensis 151-6, respectively. Many microorganisms have been reported to use heavy-metal-transporting ATPases, such as the proteins CadA and ZntA, to overcome Cd 2+ toxicity [19,33]. Specifically, CadA from S. aureus has a length of 727 amino acids [19], and its amino acid sequence is highly similar to 4111 (63.31%) and 4802 (65.21%) (Additional file 15: Figure S10). Alignment with CadA from S. aureus showed that the protein sequences of orf4111 from B. vietamensis 151-6 and orf4802 from B. marisflavi 151-25 all included a conserved (Cys-Pro-Cys) tripeptide for Cd 2+ , Pb 2+ or Zn 2+ binding [34]. CadC from S. aureus including 122 amino acids was related to the divalent cation ATPase [27]. Expression of CadA is regulated by CadC, which is a homodimeric repressor that dissociates from the cad operator/promoter upon Cd 2+ binding [19]. CadC is a member of the ArsR/SmtB family of metalloregulatory proteins [29,30]. Its crystal structure was resolved in 2005, and showed that two regulatory metal-binding sites for the inducer Cd 2+ are formed by Cys-7 and Cys-11 from the N terminus of one monomer and Cys-58 and Cys-60 of the other monomer [34]. Alignment with CadC from S. aureus showed that the protein sequence of orf4112 from B. vietamensis 151-6 and of orf4803 from B. marisflavi 151-25 contained these four Cys residues (Additional file 16: Figure S11). Moreover, the similarity between CadC and Orf4803 or Orf4112 are 86.89 and 82.64%, respectively. These analysis indicated that the orf4111, orf4112 from B. vietamensis 151-6 and orf4802, orf4803 from B. marisflavi 151-25 all encoded the proteins CadA and CadC, respectively. The protein CadA existed as the efflux transporter and was vital to the cadmium resistance. And the protein CadC was a transcription factor to regulate the expression of CadA. However, the expression mechanism of the two proteins in B. vietamensis 151-6 was different to the B. marisflavi 151-25. In B. marisflavi 151-25, the two genes (orf4802 and orf4803) were induced by the cadmium. While in B. vietamensis 151-6, the expression of CadC (Orf4112) was repressed by the cadmium, as the qRT-PCR Fold change (Cd/CK) was only 0.56 ( Table 2). As a result, few CadC was expressed in the cell and the expression of CadA (Orf4111) was activated by the cadmium and endowed the cadmium function of the organism, which qRT-PCR Fold change (Cd/CK) of orf4111 was 1.2 ( Table 2). Those analysis indicated that the cadmium resistance mechanism of B. marisflavi 151-25 was the same as cad system in S. aureus. The hypothetical expression model showed in Additional file 17: Figure S12. Additionally, we did not know the detailed mechanism about the expression of orf4111-41,112 in B. vietamensis 151-6. In the future study, we will clone the promoter and study the regulation of the cad system. In addition, the result of the transcriptome of B. marisflavi 151-25 under Cd 2+ induction showed that orf4608 (arsenical efflux pump membrane protein gene, arsB) and orf4779 (arsenic resistance protein gene) were all induced by Cd 2+ . However, overexpression of two genes in E. coli did not improve Cd 2+ resistance of recombinant E. coli. Maybe these genes can confer arsenic resistance for B. marisflavi 151-25. Presently, it has been reported that many Bacillus sp. strains (such as B. subtilis, B. safensis and B. cereus) have ability of resisting arsenic [35]. And the work of arsenic resistance for B. marisflavi 151-25 will be carried out in future study.

Conclusions
In our study, we isolated two different Cd 2+ resistance strains B. vietamensis 151-6 and B. marisflavi 151-25 isolated from Cd-contaminated soil, which could grow in the presence of Cd 2+ at two distinct concentrations up to 0.3 mM and 0.8 mM, respectively. Further study with whole genome sequencing, transcriptome sequencing and heterologous expression revealed a gene cluster that played a vital role in Cd 2+ resistance of the two strains. Orf4802 (ATPase transporter) and Orf4803 (transcriptional regulator) on plasmid p25, were found to be major contributors to Cd 2+ resistance in B. marisflavi 151-25. And Orf4111 (ATPase transporter), Orf4112 (cadmium efflux system accessory protein) and Orf4113 (cadmium resistance protein) on the chromosome, were proved to be critical for Cd 2+ resistance of B.vietamensis 151-6. In the end, we explored the cadmium resistance mechanism for B. vietamensis 151-6 and B. marisflavi 151-25. And we identified that the Cd 2+ resistance mechanism of B. marisflavi 151-25 was the same as cad system in S. aureus. Although Bacillus vietamensis 151-6 also had the similar gene cluster to B. marisflavi 151-25 and S. aureus, its transcriptional regulatory mechanism of cadmium resistance was not same. This study could expand our understanding of the biological effects of cadmium.

Isolation of cd-resistant bacteria
Cd-contaminated soil was collected from a former industrial site in Hunan Province (27°46′N, 112°52′E) and analyzed for Cd content through acid digestion followed by the use of a 7700 × inductively coupled plasma mass spectrometer (Agilent Technologies, Tokyo, Japan). To isolate Cd-resistant and bio-safe bacteria, aerobic Bacillus sp. were isolated from the soil by plating on LB agar plates containing progressively higher concentrations of cadmium chloride (0, 0.5, 1 and 2 mM). Then, in order to isolate aerobic Bacillus sp., the bacterial enrichment cultures were heat-shocked at 80°C for 20 min. Bacterial genomic DNA was isolated using the TIANamp Bacteria DNA kit (TIANGEN Biotech). The 16S rRNA gene was amplified from the extracted DNA using the universal primers 16S rRNA-F/16S rRNA-R (Additional file 2: Table S2) and the amplification products were cloned in the pGM-T (TIANGEN Biotech) vector using competent E. coli TOP10 cells (TIANGEN Biotech). Sequencing was carried out using T7 and SP6 primers and compared to the GenBank database using the NCBI BLAST program.

Evaluation of cadmium resistance and growth curve
To evaluate growth in a liquid medium of isolated bacteria, the minimum inhibitory concentration (MIC) of Cd 2+ (MIC-Cd) was determined. LB medium (800 μL) with different concentrations of Cd 2+ was dispensed into 96-well (12 × 8) microtiter plates (96 × 2-mL wells) with a multi-channel micropipette (rows A to H: 0 mM, 0.02 mM, 0.04 mM, 0.06 mM, 0.08 mM, 0.1 mM, 0.2 mM,0.3 mM, 0.4 mM, 0.5 mM, 0.6 mM, 0.7 mM, 0.8 mM, 0.9 mM, 1.0 mM and 1.1 mM). Single colonies of the test strains were inoculated into 3 mL of LB medium and cultured overnight. The test culture (15 μL) was then inoculated into each well of the prepared 96-well plate. After 24 h at 37°C and 750 rpm in an incubator (Heidolph, Viertrieb, Germany), 200 μL of the cell suspension was transferred to a 96-well plate and the turbidity at OD 600 was measured.
To determine the Cd 2+ tolerance of isolated bacteria, growth curves at different concentrations (0 mM, 0.1 mM, 0.3 mM and 0.5 mM) of Cd 2+ were analyzed. For the growth assay, single colonies of the test strains were cultured overnight and then diluted 1:100 into 100-well plates containing 200 μL of LB and various concentrations of Cd 2+ in quintuplicate. The growth curve was measured at 1-h intervals using a Bioscreen C automatic growth curve analyzer (Bioscreen, Helsinki, Finland).

Genome sequencing and analysis
Bacterial genomic DNA was extracted using the sodium dodecyl sulfate (SDS) method [36]. A total of 5 μg DNA was used to generate each library, and this DNA was sheared using Covaris g-Tubes to generate sheared fragments > 10 kb in length. The sheared DNA fragments were then prepared using the SMRT bell template preparation kit (Pacific Biosciences, Menlo Park, CA, USA) according to the manufacturer's instructions. Wholegenome sequencing was performed on the Pacbio RSII platform. All high-quality paired reads were assembled using SOAPdenovo (http://soap.genomics.org.cn/soapdenovo.html) onto a number of scaffolds [37]. Then, the filtered reads were transferred for the next step of gap closing. Transfer RNA (tRNA) genes were predicted with tRNAscan-SE [38]. Ribosomal RNA (rRNA) genes were analyzed using rRNAmmer [39]. Coding genes were identified with the GeneMarkS program [40]. The predicted coding genes were annotated based on the non-redundant protein database (NR) of the National Center for Biotechnology Information [41] and Gene Ontology (GO) [42].

RNA sequencing and transcriptome analysis
An overnight culture was diluted 1:100 in LB medium in the presence (0.1 mM) and absence of Cd 2+ , and these cultures were grown at 37°C and 200 rpm to the exponential phase. Total RNA from both groups (each group has three replicates) were extracted using a TRIzol kit (TIANGEN Biotech) according to the manufacturer's instructions. A total of 1 μg RNA per sample was used as input material for RNA sample preparation. Sequence libraries were generated using the Illumina® TruSeq® Stranded Total RNA Sample Preparation kit (NEB, Ipswich, MA, USA). TruSeq Stranded Total RNA was prepared with the Ribo-Zero™ Bacteria Kit (Epicenter, Madison, WI, USA). The clustering of index-coded samples was performed on a cBot Cluster Generation System using the TruSeq HiSeq 4000 PE 150 Cluster Kit (Illumina, San Diego, CA, USA) according to the manufacturer's instructions. After cluster generation, the prepared libraries were sequenced on the Illumina HiSeq 4000 platform by Allwegene Technology (Beijing, China) and 150-bp paired-end reads were generated. Bowtie2 was used to map mRNA reads to the genome, and HTSeq v 0.5.4 p3 was used to count the number of reads mapped to each gene. Then, the fragments per kilobase of exon model per million mapped reads (FPKM) of each gene was calculated based on the length of, and read count mapped to, the gene [43].

RT-PCR and qRT-PCR
Total RNA of B. vietamensis 151-6 and B. marisflavi 151-25 were isolated as described above under RNA sequencing. cDNA samples were prepared using Trans-Script One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China) with random primers according to the manufacturer's instructions. qRT-PCR was performed using the ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech, Beijing, China) according to the manufacturer's instructions. 16S rRNA, amplified with the primers 16S-F/16S-R, was used as an internal control. All reactions were performed in biological triplicate, and the normalized fold changes of the relative expression ratios were quantified using the 2 -△△CT method [45].
Construction and screening of an B. vietamensis 151-6 genomic fosmid library High-molecular-weight genomic DNA from B. vietamensis 151-6 was extracted using the bacterial genomic DNA extraction kit (BioTeke Biotech, Beijing, China) according to the manufacture's instructions. And a fosmid library was constructed with the CopyControl Fosmid Library Production Kit (vector pCC1FOS; Epicentre, Madison, WI, USA) according to the manufacture's instructions. A total of 1204 clones were used to construct the library in 13 96-well plates. To screen for clones with Cd 2+ resistance, overnight cultures of the 1204 clones were inoculated into LB liquid medium with 1.0 mM and 1.2 mM Cd 2+ , respectively. Meanwhile, these clones were spotted onto LB agar plates with 1.2 mM and 1.5 mM Cd 2+ . Fosmid DNA was isolated from the positive fosmid clones by Axyprep Plasmid Miniprep Kit (Axygen, USA) and sequenced by pCC1 sequencing primers pCC1-F/ pCC1-R (Additional file 2: Table S2). The results were aligned with the genome sequence of B. vietamensis 151-6 to confirm the sequence of the inserted fragments.

MIC-Cd and growth curve determination
MIC-Cd and the growth curve of plasmid-eliminated strains and wild strains of B. vietamensis 151-6 and B. marisflavi 151-25 were determined as described above in 2.3. MIC-Cd values for the recombinant strains of E. coli TOP10 and B. subtilis were also evaluated using 96well microtiter plates, as described in 2.3.