General features of the genome of Klebsiella sp. WL1316 and functional gene prediction
In this study, genome sequencing and characterisation were conducted on Klebsiella sp. WL1316 to gain a comprehensive understanding of the functional genes of this strain, especially those response to glucose and xylose metabolism and hydrogen production. The genome characteristics are listed in Table 1. The genome of Klebsiella sp. WL1316 is a chromosomal DNA with a total length of 5.2 Mb and a GC content of 57.6%. A total of 4915 coding genes were predicted, up to 86.7% of the whole genome, and 76.7% of the predicted coding genes encoded products possessed definite functions, while the rest were hypothesised proteins or unknown proteins. The functional protein-coding genes were annotated by aligning the sequences to the COG, GO, and KEGG databases. The COG and GO annotated protein-coding gene numbers reached 3768 and 3622, respectively, representing 76.7% and 73.7% of the total gene numbers, respectively. The COG and GO annotated functional categories are illustrated in Supplementary Fig. 1 and Supplementary Fig. 2. In addition, the KEGG annotated protein-coding genes reached 3408 genes, which were involved in 152 metabolic pathways. The matching genes were mined and discovered to locate them in the variable KEGG metabolic pathways, which covered the basic metabolic pathways, such as carbohydrate metabolism, lipid metabolism, amino acid metabolism, energy metabolism, and some special metabolic pathways (Supplementary Fig. 3).
Table 1
Genome characteristics of Klebsiella sp. WL1316
General features
|
|
Genome size (Mb)
|
5.20
|
Gene number (#)
|
4,915
|
GC content (%)
|
57.6
|
N50 Length (bp)
|
136,619
|
N90 Length (bp)
|
51,942
|
Number of scaffolds
|
76
|
Properties of gene annotation
|
|
Number of rRNA genes
|
5
|
Number of tRNA genes
|
76
|
Number of sRNA genes
|
41
|
COG annotation
|
3768
|
GO annotation
|
3622
|
KEGG annotation
|
3408
|
Annotation of the metabolic pathways for glucose and xylose utilisation and bioconversion
In a previous study, we reported that Klebsiella sp. WL1316 intrinsically possesses the ability to utilise glucose and xylose produced by lignocellulosic hydrolysate (Li et al. 2018). Therefore, we firstly focused on the sugar metabolic pathway, and found that glucose could be metabolised and subsequently converted in the fermentation process via the EMP and PPP pathways, as illustrated in Supplementary Fig. 4. However, there was no reference pathway for xylose metabolism that could be blasted from the KEGG database. Therefore, several functional proteins related to xylose transportation and metabolism were further mined based on the gene annotation results. As a result, we discovered that the Klebsiella sp. WL1316 was equipped with the ribose/xylose/arabinose/galactoside ABC-type transport system, as well as the D-xylose transport system ATP-binding component, D-xylose ABC transporter, and periplasmic D-xylose-binding protein, which are conducive to the transmembrane transport of xylose in bacteria. Furthermore, in terms of xylose metabolism, this strain was found to possess xylose isomerase and xylulokinase, which isomerises xylose to xylulose and further phosphorylates xylulose-5-phosphate to enter the pentose phosphate pathway for metabolism. The schematic pathways are illustrated in Fig. 1.
Identification of the genes coding for key functional proteins for hydrogen production
According to the literature, Klebsiella is a facultative anaerobic bacterium that is able to generate hydrogen gas via mixed acid fermentation pathway, in which the key enzyme for hydrogen production is formate-hydrogen lyase (Jawed et al. 2016). In addition, hydrogenase is also an important enzyme that catalyses the hydrogen production process. As such, we analysed the hydrogenase and formate-hydrogen lyase-related functional proteins of Klebsiella sp. WL1316 based on genome annotation results. As shown in Fig. 2, this hydrogen-producing bacterium possesses diverse hydrogenase and formate-hydrogen lyase related functional proteins. The hydrogenase functional proteins included the protein subunits of hydrogenase, the components of the Fe-S cluster and nickel-iron ion binding protein in the hydrogenase active center, and the assembly protein in the metal ion binding center of hydrogenase. The hydrogenase-related gene cluster is illustrated in Fig. 2a. The formate-hydrogen lyase functional proteins included multiple protein subunits, formate-hydrogen lyase regulatory proteins and transcriptional activators, and formate-hydrogen lyase Fe-S protein (Fig. 2b). The distributions of gene clusters of hydrogenase and formate-hydrogen lyase were all located in the genome (Fig. 2c). The annotated functional proteins and their Gene_ids are listed in Supplementary Table 2. The diversity of hydrogenase and hydrogen formate lyase indicated that the strain had a strong hydrogen production potential at the nucleic acid and protein levels. Hydrogenase is the most important functional enzyme that catalyses the hydrogen production process in hydrogen-producing bacteria. Therefore, we further compared the difference in the category and number of hydrogenase functional proteins between Klebsiella sp. WL1316 and the five reference bacteria based on the bacterial genome information (Fig. 3). Figure 3a reveals that Klebsiella sp. WL1316 in this study obtained abundant hydrogenase functional proteins, while the hydrogenase3 large subunit was dominant in Klebsiella pneumoniae subsp. NTUH-K2044, Klebsiella oxytoca KONIH1, Enterobacter sp. R4-368. Clustering analysis indicated the category diversity of hydrogenase functional proteins in Klebsiella sp. WL1316 was more similar to the two reported hydrogen-producing bacteria Enterobacter ludwigii EcWSU1 and Enterobacter sp. R4-368. The statistical results of gene numbers of hydrogenase functional proteins indicated that Klebsiella sp. WL1316 contained the highest gene number of hydrogenase functional proteins, which is close to the hydrogen-producing bacterium Enterobacter sp. R4-368 (Fig. 3b). This result indicated that Klebsiella sp. WL1316 could be a hydrogen-producing bacterium with a high hydrogen production potential, with more hydrogenase functional proteins than the five reference bacteria.
Construction of the genome scale metabolic model for hydrogen production
Based on the above analysis of glucose metabolic pathways, xylose metabolism-related enzymes and functional proteins, as well as enzymes and functional proteins related to biological hydrogen synthesis, the metabolic network for hydrogen production by Klebsiella sp. WL1316 from the fermentation of cotton stalk hydrolysate was reconstructed. This metabolic network covers not only the glucose and xylose metabolic pathways, but also the biohydrogen synthesis pathway and some competitive metabolic branches, as illustrated in Fig. 4. The biohydrogen synthesis pathway included the mixed acid fermentation pathway and hydrogenase catalytic pathway, while the competitive metabolic branches included the succinic, lactic, acetic acids, and ethanol generation pathways, which acted as the branches of the mixed acid fermentation pathway.
Metabolic analysis for hydrogen production of Klebsiella sp. WL1316 from the fermentation of cotton stalk hydrolysate
Pyruvic acid is an important intermediate metabolite of Klebsiella sp. WL1316 for the metabolisis of glucose and xylose in cotton stalk hydrolysate. This strain enters the mixed acid pathway for fermentative hydrogen production, which was initiated by the decarboxylation of pyruvate. Therefore, pyruvate decarboxylase (PDC) is the key enzyme involved in this metabolic pathway. Supplementary Fig. 5 shows that PDC activity was relatively high after 24 h of fermentation, indicating that this enzyme promoted pyruvate decarboxylation, and then entered the fermentation pathway. Thereafter, PDC activity decreased gradually to 96 h, and the PDC activity declined to a relatively low level, almost 0, after 120 h of fermentation. The lactate dehydrogenase (LDH) activity was low at 24–48 h and increased rapidly to the highest level (0.027 U/104 cells) at 72 h, indicating that high LDH activity in 72-h fermentation may lead to a large accumulation of lactic acid and compete for the distribution of metabolic carbon skeleton in the biohydrogen synthesis pathway. Such a competitive effect for biohydrogen synthesis controlled by LDH was consistent with that reported by Jung et al. (2012). The aldehyde dehydrogenase (ALDH) activity reached the highest level after fermenting 48‒96 h, indicating that the metabolic pathway of acetic acid and ethanol synthesis mainly competed for the metabolic carbon skeleton of the biohydrogen synthesis in the mid-to-late fermentation stage.
The metabolic analysis was first developed toward the key metabolites in the conversion of sugars, such as glucose, xylose, and xylulose (Fig. 5). Glucose and xylose consumption rates decreased with fermentation time during the fermentation stage, and the results showed that glucose was metabolised effectively, which promoted the synthesis and fermentation of cell material. According to the biohydrogen synthesis metabolic network of Klebsiella sp. WL1316, xylose was metabolised by converting xylose to xylulose through the PPP pathway, therefore the generation rate of xylulose was further monitored. Although the generation rates of xylulose were lower than consumption rates of xylose during the whole fermentation stage, its generation rates also gradually decreased, indicating that xylulose was also rapidly metabolised via the PPP pathway, which was provided as a carbon skeleton for the fermentation process.
A survey of the literature showed that key node metabolites, such as succinic, lactic, acetic, formic acids, and ethanol, are typically present in the mixed acid fermentation pathway (Jenni et al. 2013; Jung et al. 2012; Yoshida et al. 2005; Lu et al. 2009; Zhao et al. 2009). Therefore, the time-dependent generation profiles of these key node metabolites were analysed (Fig. 6). Pyruvic acid is the most important metabolic intermediate, which initiates the mixed acid fermentation pathway, especially for hydrogen gas generation. Pyruvate is also a key node metabolite for mixed acid fermentation. In this study, the pyruvate node obtained a high generation rate at 24 h of fermentation, and then decreased rapidly, indicating that pyruvate underwent rapid decarboxylation transformation, with more carbon skeletons flowing to the fermentation process. The splitting of formic acid is the most important hydrogen-producing pathway for facultative anaerobes (Zhao et al. 2009; Lu et al. 2009; Yoshida et al. 2005). The results revealed that the generation rate of formic acid was relatively low at 24 h of fermentation, but reached a peak value at 48 h, indicating that a large amount of formic acid accumulated between 24 and 48 h, and that the metabolic carbon skeleton mostly flowed to the mixed acid fermentation pathway at this fermentation stage. Subsequently, the generation rate of formic acid decreased sharply, indicating that formic acid may be used for biohydrogen synthesis in large quantities during the mid-to-late fermentation stage. The shift in citric acid concentration revealed that there was a high generation rate after 24–48-h fermentation at the citric acid metabolic node, which enabled many of the metabolic carbon skeletons to originate from the flow of pyruvate nodes to the TCA cycle. The generation rates of succinic acid were relatively low during the whole fermentation stage, indicating that succinic acid accumulated less under anaerobic fermentation conditions, indicating that it may compete for the metabolic carbon skeleton of biohydrogen synthesis with relatively weak competitiveness. The generation rate of the lactic acid node plays an important role in the whole fermentation stage, which was enhanced sharply after fermenting for 48 h, and reached a peak value at 72 h, indicating that the lactic acid production pathway may be the main competitive branch for biohydrogen synthesis. In addition, the generation rates of the ethanol and acetic acid in the same metabolic branch also significantly affected biohydrogen synthesis. The generation rate of the ethanol remained at a high level during the whole fermentation stage and decreased gradually with the fermentation time but increased to a certain extent at 120 h of fermentation. This indicates that ethanol may accumulate more in the late fermentation period at this time, and it may be an important competitive branch for biohydrogen synthesis. The metabolic flux of acetic acid was higher during the fermentation period of 48‒72 h, indicating that the acetic acid production was higher in this fermentation stage, which may be strongly competitive for the metabolic carbon skeleton. In general, acetic acid was produced in a large quantity after 48 h of fermentation, and ethanol was accumulated in large quantities in this metabolic branch, indicating that this metabolic branch is an important branch in competing for the carbon skeleton of biohydrogen synthesis after 48 h of fermentation.