Identification and basic features of strain B5
The characterization of B5 strain was evaluated, and the colony was light yellow convex subtransparent with regular edges, sticky and 1-2 mm in diameter after incubation for 24 hours at 60 °C on LB plates (Fig. S1). The Congo red assay is a qualitative detection of reducing sugars, and commonly used to estimate cellulolytic activities. Here, clear zones were observed around the colonies on CMC-minimal media (Fig. 1A), showing remarkable cellulolytic activities. Under the microscope, it was shown that B5 was motile, Gram-positive, spore-forming, rod-shaped cell, and the oval spores were located terminally within a swollen sporangium (Fig. S1). The size of the cell was from 4 to 6 µm in length, 0.5 to 1µm in diameter with peritrichous flagella (Fig. 1B). 16S rRNA sequence of this strain indicated that B5 strain was closely related to members of Geobacillus. For Geobacillus, it has been demonstrated that recN gene is the most robust marker for assigning new bacterial strains at the species level [14], and homology search revealed that B5 strain was a member of genus Geobacillus, showing highest similarity (99.71%) to G. stearothermophilus DSM 458. Phylogenetic tree further indicated that B5 formed a distinct linkage with DSM 458 with 100% bootstrap support (Fig. S2). Combining the above results with physiological and biochemical characteristics (data not shown), B5 was identified as G. stearothermophilus B5 and the sequence data were stored in the NCBI SRA database (accession number CP034952).
The carbon source test results showed that B5 could utilize various carbon sources, including but not limited to sugars, amino acids, hexose acids, carboxylic acids, esters and fatty acids. The positive test included D-glucose, sucrose, D-mannose, D-cellobiose, D-mannitol, D-fructose, glycerol, D-turanose, inosine, L-pyroglutamic acid, pectin, glucuronamide, L-malic acid, L-lactic acid, and so on (Additional file 1 and Fig. S3). During composting process, microorganisms are responsible for transforming the organic matter into biomass, CO2, heat and humus-like end-products, and the broad carbon utilization illustrated that B5 could take advantage of kinds of carbon source during composting. B5 can grow at a broad temperature ranging from 40 °C to 73 °C, and the optimal temperature is between 55 °C and 65 °C (Fig. 1C). When the temperature raised above 55 °C, a mass of pathogens were killed, thus thermophilic composting is widely adopted for industrial application, and there is no doubt that exploring microorganisms with thermo-resistance is necessary and urgent. Here, the B5 strain just approached its optimal growth condition at 55 °C, and then it could degrade fats, celluloses, hemicelluloses and some lignin efficiently. Generally, microbial activity even for most thermophiles declined rapidly at temperatures above 63 °C, while B5 still maintained considerable activity even at 65 °C, indicating that B5 might play an important role during thermophilic composting process. As one of the critical parameters, the pH values ranging from 6.7 to 9.0 are convenient for most composting microbes, and the optimum pH are between 5.5 and 8.0 [15]. Interestingly, B5 can survive with the pH values ranging from 4.0 to 9.5, and the optimum growing pH value (Fig. 1D) is similar with the pH value for best composting efficiency. Furthermore, B5 could survive with the NaCl concentration from 0 to 3.5% (Fig. 1E). The growth curve of B5 under the optimal conditions was shown in Fig. 1F, and the logarithmic growth period fell within 6 to 12 h.
Determination of various enzyme activities
Different enzyme activities were detected, and the CMCase activity kept increasing and reached the peak (0.32±0.02 U ml-1) on the 6th day (Fig. 2A) and then decrease until the end. The xylanase activity increased sharply and the peak (0.14±0.01 U ml-1) was obtained at the 3rd day (Fig. 2B). The tendency of α-amylase was similar with that of CMCase, and the highest value was obtained on the 6th day (0.43±0.02 U ml-1) (Fig. 2C). The protease activity increased sharply like that of xylanase until reaching to 0.62±0.03 U ml-1 at the 3rd day, and then decreased gradually to the end (Fig. 2D). Factually, the biodegradation of lignocellulosic biomass in the composting process required the synergism of various enzymes including cellulase, hemicellulase, urease and protease [16]. It is known that filamentous fungi such as Aspergillus spp., Trichoderma spp. can secrete abundant extracellular hydrolytic enzymes with higher activities, while these mesophilic microbes and its enzyme activities tend to reduce or even completely lose at thermophilic phase [6, 17]. Interestingly, enzymes produced by thermophilic bacteria are usually more thermostable. Tai et al. [18] found that Geobacillus sp. could secrete CMCase retaining 90% activity after 1 h of incubation at 70 °C, and similar results were observed in another Geobacillus strain in which cellulase activity kept 100% stable after 24 h incubation at 60 °C [19]. As the member of Geobacillus spp., B5 also could secrete various enzymes with considerable thermostability. Furthermore, cell-bound effect always existed in some enzymes [20], if so the actual organic matters degrading efficiency would be even higher than that measured in cell-free supernatant. A bottleneck for B5 strain is the low yields of thermophilic cells related to their growth, and the highest OD600 value of B5 was only around 1.2 even in the optimal conditions (Fig. 1H). In fact, this is also the bottleneck for most of the thermophiles, and some measures have been taken to improve the cell yields thus increasing enzyme production, such as medium composition, process configuration and special equipment [21]. For example, the cellulase production of some Geobacillus sp. was 2-fold increased by optimizing the culture conditions with additions of ammonium sulfate and yeast extract [22]. Thus, more research about B5 is necessary so as to increase the cell yield and its enzyme productions.
Genomic analysis of G. stearothermophilus B5 and comparison of COG categories
The G. stearothermophilus B5 genome analysis results revealed a GC content of 52.46%, and a single contig of total 3.37 Mbp with 3371 CDS, 32 rRNA, and 90 tRNA (Fig. 3A). The gene length/genome ratio was 85% and the intergenetic region length/genome ratio was 15%. Genes were then annotated with different databases as follows: COG (2472), GO (2363), NR (3342), Swiss-Prot (2617), KEGG (1797), and CAZy (100). B5 strain was compared to the other four Geobacillus strains which were previously reported to own considerable capacity of cellulase or hemicellulase productions, and the genome features of the five strains are presented in Table 1. The number of orthologous genes between B5 strain and other four Geobacillus strains were 2615 (HTA426), 2414 (NG80-2), 2417 (NBRC 101842) and 2602 (Y412MC52), and the core genome of the five strains consists of 2202 orthologous genes and the pan-genome pool consists of 6175 genes, among which 465 special genes were unique for B5 strain (Fig. 3B). Conserved genes and gene pools were always used to evaluate the variation of the genus families. Zhang et al. [23] compared the genomic of five Bacillus strains (four Bacillus amyloliquefaciens and one Bacillus subtilis), and the results showed that 73.9% conserved genes and 5643 genes pools were observed in five Bacillus strains, which indicated low variation of Bacillus amyloliquefaciens. In this study, 65.3% conserved genes and 6175 genes pools were observed when comparing to these five Geobacillus strains, which indicated abundant variations for Geobacillus strains, and the variations of this specie might be due to the thermophilic environment.
Besides, 2472 genes in B5 strain were annotated to 1581 COGs, and all available CDSs from five Geobacillus strains were assigned into 20 COGs functional categories (Fig. 3C), and there were fewer differences in most of the cellular processes and signaling categories (D to V). These might be due to that those five strains belong to the same genus, and the major functional models were conserved. The major differences between these five strains were observed in metabolism categories (C to Q), especially in carbohydrate transport and metabolism (G). In order to deeply comprehend the potential roles of B5 strain during composting, specific COGs involved in carbon catabolic functions were also analyzed. Amino acid transport and metabolism, and carbohydrate transport and metabolism functions represented 10.76% and 7.24% of the COG categories, respectively (Additional file 2). With respect to amino acid transport and metabolism function, the top five abundant COGs were permeases of the major facilitator superfamily (COG0477), permeases of the drug/metabolite transporter superfamily (COG0697), aminotransferase (COG0436), deacetylase (COG0624), and lyase (COG0346); while the top five abundant COGs for carbohydrate transport and metabolism were permeases of the major facilitator superfamily (COG0477), permeases of the drug/metabolite transporter superfamily (COG0697), deacetylase (COG0726), glycosidase (COG0366), and phosphotransferase system IIC component (COG1263). COG0477 participated in encoding permeases of the major facilitator superfamily, which could catalyze the transport of kinds of substrates including carbohydrate, lipids, peptides, nucleotides, and some other molecules at the thermophilic condition [24]. COG0366 was versatile which could not only encode glycosidase responsible for the release of aromatic compounds [25], but also α-amylase which could destroy the alpha bonds between long-chain polysaccharides, like glycogen and starch [26]. Altogether, the deep investigation of COG categories indicated that B5 strain owned a good potential for degrading proteins and carbohydrates during composting. The extensive diversity of gene functions revealed considerable potential for G. stearothermophilus B5 in organic substance decomposition in the composting system.
CAZyme family analysis of B5 strain genome
CAZymes can cleave, build and rearrange oligo- and polysaccharides, which play important roles in bacteria and are vital for optimizing biomass degradation [27].The degradation capacity of B5 strain during the composting process was revealed through the gene annotation against the CAZy database. B5 strain encoded 100 CAZymes which unevenly distributed between glycoside hydrolases (GH, 29.0%), glycosyl transferases (GT, 36.0%), carbohydrate esterases (CE, 20.0%), auxiliary activities (AA, 4.0%) and carbohydrate-binding modules (CBM, 11.0%) (Fig. S4). The GH and GT family members take up the largest share and fulfil vita functions in the cleavage of polymeric substrates [28].
The GH family enzymes could hydrolyze the glycosidic bond between two carbohydrates or a carbohydrate and a non-carbohydrate moiety. Under thermophilic situation, GH members in B5 stain including cellulase (GH1, GH3, GH31), amylase (GH13), chitinase (GH18), together with some kinds of peptidoglycan hydrolase and oligosaccharide degrading enzymes were significantly up-regulated (Additional file 3). The genes encoding α-amylase (EC 3.2.1.1) were widely detected in B5 strain, which was considered as a crucial amylase. The biomass degradation pathways that GH families in B5 strain participated in included glycolysis (ko00010) and starch and sucrose metabolism (ko00500). The GH1 members encoded by B5 strain are 6-phospho-β-glycosidases, which were extremely thermostable and almost lost no activity after incubation at 60 °C for 7 days, and they could utilize cellobiose [29]. One GH4 member encoding 6-phospho-α-glucosidase and one GH5 member encoding endo-1, 4-β-glucanase also had obvious effect in cellulose degradation. Besides, B5 strain contained a significantly higher number of ten GH13 family genes, which participated in starch hydrolyzation mainly [30]. There were also some CEs detected in B5 strain, which exhibited potential in de-acetylating xylan and xylooligosaccharide. A CE3 discovered in T. reesei previously could encode acetyl xylan esterase, thus enhanced the solubilization of xylans [31]. Moreover, a CE7 from Thermoanaerobacterium sp. was also validated to degrade xylan [32]. The existence of these genes may be important for B5 strain during the biodegradation process of cellulose and hemicellulose. Valuably, six CE4 members associated with the destruction of plant polysaccharides were also detected in B5 strain. These CE4 were acetyl xylan esterases which could catalyze the de-acylation of galactoglucomannan and acetylated manno-compounds. By the way, the CE4 also possessed peptidoglycan N-deacetylates with the ability of chitin degradation [33]. Three AA4 family members including vanilly-alcohol oxidases (VAO) were detected, the function of which was to catalyze the conversion of multiple phenolic compounds bearing side chain at the para-position of aromatic rings [34]. As the vital components of CAZymes, glycosyl transferases could catalyze the transfer of sugar moieties from activated donor molecules to specific acceptor molecules, and thirty-one GTs were detected in B5 dominated by GT4 and GT2. Actually, the combination of GT4 and GT2 family members make up approximately 50% of all glycosyl transferases, and they might be the originals from which other GT families evolved [35, 36]. The GT4 and GT2 families could catalyze various reactions including some key steps in N-glycosylation pathways. Besides, GT35 family genes were also detected, which could catalyze the phosphorolysis of specific glycosidic bonds within maltodextrins through removing the non-reducing glucosyl residues of linear oligosaccharides [37]. The CAZymes identified here indicated that B5 strain possessed a considerable potential of metabolizing some recalcitrant and readily degradable biomass. Our results also provided genetic evidence of both strong hydrolytic and transglycosylatic capabilities for B5 strain, which harbored various kinds of CAZymes genes.
Global analysis of transcriptome and DEGs
The raw data were processed according to the tophat2-cufflinks workflow [38], and finally get the expression of all unigenes represented by FPKM values. The FPKM distribution and relationship of different treatments were shown in Fig 4. Multi-dimensional scaling (MDS) method was applied to identify the sources of variability within the data. An obvious divergence was observed among different temperatures, and the repeatability of the three biological replicates are credible enough for DEG analysis (Fig. 4B). The results were further verified by the Spearman coefficient of correlation among the 9 data sets, which shared a value of more than 0.93 within the same treatment (Fig. 4C).
DEGs were detected to identify the temperature responding genes between each treatment, and a diagram was constructed showing DEGs to further understand the interaction of these treatments for DEGs (Fig. 4E). There were 980, 1019 and 601 DEGs between different treatments. Interestingly, we found 171 DEGs in all treatments, and most of their expression level decreased when temperature decrease or increase to their limits (Additional file 4), which could be considered as the heat-response sensitive genes. Furthermore, the clustering analysis results showed that the treatments of T40 and T70 formed closer (Fig. 4F). It might due to that the two treatments were all extreme environments, so many genes tended to exhibit in similar patterns. Briefly, all DEGs were divided into two groups: in group I, most genes expressed at a very low level in T60, and there were also many differences between T40 and T70; in group II, almost all the genes in T60 expressed substantively, and DEGs in T40 and T70 showed small differences.
Metabolism characteristic of B5 strain at mesophilic situation
It is well known that the metabolism systems of microbes are appropriately regulated, especially in non-optimal environment. The translation and synthesis of proteins are energy consuming process, which should be limited at some level when suffering extreme conditions. Compared with T60 (Additional file 5), the isopropylmalate synthase (gene2488), isopropylmalate dehydrogenase (gene2487), and aminotransferase (gene2492) genes were significantly up-regulated in T40 treatment. These were key enzymes associated with the process of leucine biosynthesis, and played a critical role in “valine, leucine and isoleucine biosynthesis” and “valine, leucine and isoleucine degradation” pathways. Besides, the expression of branched-chain amino acid transporters (gene3091) belonging to ABC transport system was also up-regulated in T40, which mainly participated in transportation or assimilation of branched-chain amino acids (leucine et al.) into the cell. Meanwhile, the genes involved in the degradation of branched-chain amino acids were down-regulated, such as branched-chain alpha-keto acid dehydrogenase (gene0976, 2204), which illustrated that B5 strain would enhance the biosynthesis of leucine and inhibit its degradation at mesophilic phase of composting. It was reported that the D-amino acids (D-Leu, D-Met, and D-Phe) could regulate the synthesis of peptidoglycan [39]. Always, many bacteria cell walls contained a large amount of peptidoglycan, especially for the gram-positive bacteria. Thus, the down-regulation of UDP-N-acetylmuramoyl-L-alanyl-D-glutamate-2,6-diaminopimelate ligase (gene1033) involved in the cell peptides synthesis might result in the inhibition of the cell walls synthesis in B5 stain. Actually, the synthesis of D-amino acids might be a common strategy for bacteria to adapt to the non-optimal environment. Here, the uptake and synthesis of leucine might also contribute to compensate the down-regulation of genes related to peptidoglycan synthesis.
Aromatic amino acids including tryptophan, phenylalanine and tyrosine were necessary to all microorganisms during primary metabolism, and these aromatic compounds were mainly produced via shikimate pathway [40]. In T40 treatment, the expression of chorismate mutase genes (gene2048, 2622) showed no difference by comparing to T60, while the anthranilate phosphoribosyl transferase (gene2045) and the pyridoxal phosphate dependent enzyme (gene1050) were up-regulated. These findings illustrated that the synthesis of phenylalanine and tyrosine might not be affected at composting mesophilic phase, while the synthesis of tryptophan was strengthened. The results above were further supported by the up-regulated expression of tryptophan-tRNA ligase (gene0160), which might be a signal of increasing demand for tryptophan. Furthermore, the expression of phosphoglycerate dehydrogenase (gene2090) and phosphoserine aminotransferase (gene0536) was up-regulated in T40. These two enzymes were the key members during serine biosynthesis process, while serine could be taken as a precursor for the synthesis of tryptophan [41]. In addition, the up-regulated pyridoxal phosphate enzyme (gene1050) might also contribute to the conversion of serine to tryptophan. All above, the synthesis of tryptophan of B5 strain was up-regulated so as to get adapt to the low temperature and survive in the composting mesophilic phase. Similarly, the synthesis and uptake of tryptophan was also increased for Saccharomyces cerevisiae in cold stress, which enhanced the tolerance to low temperature [42].
The HSPs (heat shock proteins) and enrichment analyses of B5 strain at extreme thermophilic situation
B5 strain was isolated from the thermophilic composting, and it could survive at 70 °C or even higher temperature. Thus, it is certainly worth understanding its internal heat shock mechanism, which might be the prerequisite of B5 strain to secrete extracellular enzymes. Heat shock proteins performed chaperone functions by stabilizing proteins to ensure correct folding or helping refold denatured proteins so as to protect the cell from heat stress [43]. The expression changes of the heat shock proteins in T70 vs T60 treatment were displayed in Table 2, and their relative expression levels all increased at least 2 folds. Some HSPs were responsible for unfolding the insoluble protein aggregates, or serving as co-factor of Hsp70, such as clpB (gene0701), clpX (gene2482), clpP (gene1200) and dnaJ (gene2328). Hsp33 (gene0064) and dnaK (gene2329) belonged to class I heat shock proteins (chaperonin) with the functions of protein folding and unfolding, thus entrusted thermotolerance to cells exposed to extremely stressful conditions. Also, grpE (gene2330) was discovered in B5 strain, and it mainly serviced as co-factor of dnaK. It appeared that these genes played a pivotal role for B5 strain to survive in the thermal environment.
Except for the heat shock genes, the metabolism mechanism for B5 strain when confronting with the high temperature aroused more interests. To investigate how this strain responding to heat stress, the DEGs in T70 vs T60 treatment were taken to conduct the KEGG enrichment and protein-protein interaction analyses (Fig. 5 and Fig. 6). The most significant up-regulated pathway is ribosome (eco03010, q-value = 1.63×10-9) covering 33 genes. The participated genes could encode various ribosomal proteins (Additional file 6) including 30S ribosomal protein S3 (gene0114) and 50S ribosomal protein L13 (gene0141). Protein S3 could encircle the mRNA along with protein S4 (gene2613) when entering the ribosome, and they also play an important role in mRNA helicase processivity [44]; while protein L13 was very important during the early stage of 50S assembly [45]. These results could be further supported by GO enrichment analysis, whose top three up-regulated enrichment terms were all related to ribosome synthesis (GO:1990904, 0005840, 0003735). All the results referred above suggested that normal protein synthesis and vigorous growth strongly increased so as to intensify the high heat resistance capacity, which were different from the performance of thermolabile strain S. cerevisiae with down-regulated ribosome protein when facing heat stress [46]. This might be due to the specialty of B5 strain that the increasing number of expressed proteins, especially some key enzymes, were related to cell survival under heat stress. Furthermore, the pantothenate and CoA biosynthesis (eco00770), arginine biosynthesis (eco00220), and pyrimidine metabolism (eco00240) pathways were significantly up-regulated as well. Unlike the ribosome pathway, the genes participated in carbon metabolism (eco01200), fatty acid degradation (eco00071), fructose and mannose metabolism (eco00051), starch and sucrose metabolism (eco00500), biosynthesis of secondary metabolites (eco01110) and some biological process and cellular metabolic process related pathways were all down-regulated with various degrees. These results suggested that the catabolism of recalcitrant carbon sources could be synergistic with heat resistance of B5 strain because of the assistance from the high temperature which would destroy the lignocellulose thus providing more reducing sugar and polysaccharides.