Characterization of B. megaterium FDU301
A strain of B. megaterium tolerant to arid condition (15% PEG200 (w/w), aw 0.985) was isolated from plaque area on the surface of a leaflet in an old book, and was named as FDU301. The sequence of its 16S rDNA gene was identical to that of B. megaterium NBRC15308 and B. megaterium QMB1551 (data not shown). The whole genome of B. megaterium FDU301 was sequenced on a combination of Illumina HiSeq and PacBio RSII platforms (31). The assembled genome was 6,872,701bp in length, comprising one chromosome and nine plasmids.
Comparing with B. megaterium NBRC15308, FDU301 had a larger genome, and more predicted genes. Under the conditions of amino acid sequence identity being greater than 40% over at least 80% of the full sequence length (32), FDU301 and NBRC15308 had 5335 homologous genes. Among 1561 genes unique to FDU301, 305 genes were annotated with KEGG database to be related to signaling and cellular processes, environmental information processing, genetic information processing, carbohydrate metabolism, etc. The full genome data of B. megaterium FDU301 can be found in NCBI GenBank (accession numbers CP045267-CP045276).
As shown in Fig.1a, FDU301 showed a typical "S" type growth curve in normal LB medium, with a short incubation period of 2 h, and reached a plateau around 10 h. In the presence of 5% PEG200, the FDU301 grew faster and reached higher cell density than that in normal LB medium. As the concentration of PEG200 increased, the growth of bacteria slowed down and reached much lower cell density than that in normal LB medium. The bacteria hardly grew in the medium with 20% PEG200, indicating the limit of the strain to tolerate. Compared to B. megaterium NBRC15308, B. megaterium FDU301 grew much better in the arid medium (15% PEG200 (w/w)) (Fig.1b).
Global overview of the RNA-Seq data
The transcriptome of B. megaterium FDU301 in the growth phase (4 h) under control (LB medium, L) and simulated arid condition (LB medium with 15% PEG200, P) was analyzed with RNA-seq. The RNA-seq data have been submitted to NCBI SRA (accession numbers PRJNA649685). After filtration, a total of 51,893,124 and 61,892,804 reads were obtained from L and P samples, respectively. For both samples, more than 95% of the reads were mapped to the B. megaterium FDU301 reference genome (Table 1). In order to verify the transcriptomic results, ten differentially expressed genes (DEGs) were randomly selected and their transcriptional level were determined with quantitative reverse transcription PCR (RT-qPCR). The results of RNA-seq and RT-qPCR were generally consistent with each other, indicating that the transcriptomic results reflected the differences in gene expression under the arid and normal conditions (Additional file 1: Fig. S1).
As shown in Fig.2a, the correlation between the three biological replicates of each sample (L and P) was high, indicating that the sequencing data was highly reproducible. Meanwhile the difference between treatment groups was obvious. Two groups were also well separated from the other in the principal component analysis (Fig.2b). These showed that arid stress had a significant effect on the gene expression of FDU301.
The volcano map of DEGs is shown in Fig.2c. Compared with the control group, the expression levels of 2941 genes were significantly different under the simulated arid conditions (FDR<0.05 & |log2FC|≥1), of which 1422 genes were upregulated and 1519 genes were downregulated (Additional file 2: Table S1).
Annotation analysis of DEGs
The 2941 DEGs (FDR<0.05 & |log2FC|≥1) were annotated with COG and KEGG databases. According to COG annotation, DEGs were seen in most of the COG categories, which meant that the response of FDU301 to the arid stress was a complicated process (Additional file 3: Table S2). As shown in Fig.3, the category with the highest proportion of upregulated genes was inorganic ion transport and metabolism (P, 41.56%). In terms of downregulated genes, categories with more than 30% genes downregulated included carbohydrate transport and metabolism (G); translation, ribosomal structure and biogenesis (J); energy production and conversion (C); lipid transport and metabolism (I); and amino acid transport and metabolism (E). These results suggested FDU301 had general suppressions in metabolism and protein production, and enhancement in the transport for inorganic ion, such as Fe, Zn, Ni, in face of the arid stress.
KEGG annotation showed that the arid stress significantly upregulated the transcription of genes associated with ABC transporters (Additional file 4: Table S3), and significantly downregulated that of oxidative phosphorylation and glycolysis pathways (Additional file 5: Table S4).
Major changes in gene expression under arid stress
Selected genes with significant changes in transcription under simulated arid (15% PEG200 (w/w)) and normal conditions were further analyzed with RT-qPCR.
Oxidative stress-responsive genes. PerR is a key regulatory protein for oxidative stress response in Bacillus spp. (33). As shown in Fig.4a, perR was upregulated under arid condition. Several genes known to be regulated by perR were also upregulated significantly, including fur, dps, and katE (Fig.4a). Fur encodes a major suppressor for the expression of many ferrous uptake operons, whereas dps and katE are related to avoiding DNA damage and removing ROS, respectively. These results were consistent with the level ROS in FDU301 cells grown in the medium with different concentration of PEG200 (Additional file 6: Fig. S2), suggesting that oxidative stress was one of the main challenges for the bacteria in the simulated arid condition.
Fe2+ transportation genes. Under oxidative stress, Fe2+ will react with H2O2 through fenton reaction to form hydroperoxide, which was highly active in destroying DNA. Upregulations of many Fe2+ transportation-related genes were seen in B. megaterium FDU301 under arid condition. In fact, Fur is a suppressor for Fe2+ uptake, and Dps functions by binding Fe2+ and reduces the level of free Fe2+ in the cell. Meanwhile, the Fe2+ uptake gene feoB (34), was also found to be greatly upregulated in arid condition (Fig.4a), suggesting a delicate balance of Fe2+.
Ectoine biosynthesis genes. The ectB and ectA were significantly upregulated by about 23.10-fold and 8.40-fold, respectively in arid condition (Fig.4b). The two genes were involved in the biosynthesis of compatible solute ectoine (35). Meanwhile, genes related to the transportation and biosynthesis of another commonly used compatible solute, glycine-betaine, were not significantly changed, or even downregulated (Table 2, Additional file 2: Table S1).
Sporulation genes. Under the simulated arid condition, genes related to sporulation stage II (spoIIB, spoIIE, spoIIGA) were upregulated by about 8.57 to 29.24-fold. SspD, the gene encoding small acid-soluble spore proteins (SASP), which is a major protective component of Bacillus spores, were also highly expressed (Fig.4c).
TipA gene. TipA encodes a transcriptional regulator activated by cyclic thiopeptide antibiotics, such as thiostrepton, and promothiocin in Streptomyces (36). Under the simulated arid condition, tipA was one of the most dramatically upregulated genes, in terms of the fold of change in the transcriptional level (Fig.4d). TipA had not been noticed in previous studies on bacterial arid tolerance.
Respiratory and glycolysis genes. As shown in Fig.4e, under the simulated arid condition, genes related to oxidative phosphorylation (atpB, atpE, atpF, atpH, atpA, atpG, atpD, atpC) were downregulated to various degrees. Meanwhile, some glycolysis-related genes (pgk, tpiA, frmA) were also downregulated (Fig.4f). This might reflect the slow growing status of bacteria in the simulated arid condition.
Changes in gene expression under 5% PEG200
As shown in Fig. 1a, FDU301 grows slightly faster in medium with 5% PEG200 than in LB medium without PEG200. We compared the expression of selected genes with 0%, 5%, and 15% PEG200. Although genes upregulated in 15% PEG200 were also upregulated in 5% PEG200, the levels of upregulation for most of them were significantly lower than that in 15% PEG200 (Fig.5a,b). In contrast, genes downregulated in 15% PEG200, including those in oxidative phosphorylation (atpB, atpE, atpF, atpH, atpA, atpG, atpD, atpC) and glycolysis (pgk, tpiA, frmA) pathways, were instead slightly upregulated (Fig.5c,d). It seems that under 5% PEG200, the bacteria sense the change and increase their metabolisms to be prepared for worsen environments.