Genomic and secretomic insight into lignocellulolytic system of an endophytic bacterium Pantoea ananatis Sd-1

Lignocellulose in vascular plant cell walls is the largest reservoir of organic polymers in terrestrial ecosystems and is a potential biofuel feedstock alternative to petroleum [1]. The main components of lignocellulose are polysaccharides and lignin. They are bound together in a matrix and the recalcitrance of structure gives rise to a major bottleneck in the efficient conversion of lignocellulosic biomass to biofuels [2]. Thus, exploring highly efficient lignocellulose decomposition microorganisms has attracted significant research attention [3]. During the past decade, only a limited number of microorganisms, such as various types of fungi, were found to breakdown the complex lignocellulose polymer structures [4]. Recently, bacterial systems have attracted an increasing amount of attention due to their unique mechanisms for lignocellulose decomposition and the high specificity of their lignocellulolases. Bacteria also have great advantages over fungi due to their rapid growth, wide range of environmental adaptability, and facilitative genetic manipulations [5, 6].

Bacterial candidates for degrading lignocellulosic polysaccharides have been identified by genomic and proteomic analysis. Caldicellulosiruptor bescii DSM 6725, Clostridium thermocellum ATCC 27405 and Streptomyces SDPB6 had been sequenced, and a lot of genes in these bacteria strains encoding putative carbohydrate-active enzymes (CAZy) were identified, and their presence were partially confirmed by proteomic analysis results [7–9]. Recently, qualitative and quantitative proteomic analysis methods were used to demonstrate that lignocellulolytic enzymes in Thermobifida fusca were highly up-regulated in the presence lignocellulosic biomass [10–12]. However, bacterium that can degrade both lignocellulosic polysaccharides and lignin was rarely found [13, 14], let alone a comprehensive research of the lignocellulolytic system, including cellulases, hemicellulases, and ligninolytic enzymes in the bacteria was reported.

Endophytes which spend the whole or part of its life cycle colonizing inside the healthy host plant tissues have been the subject of research attention worldwide, due to its wide distribution and the ability to produce a wide variety of secondary metabolites. They can secrete a large number of proteins, including lignocellulolytic enzymes, to assist their adaption and survival within higher plants [15]. A few endophytic fungi demonstrating lignocellulosic biomass degrading ability have already been discovered [16, 17]. However, endophytic bacteria as potential lignocellulose degradation resources have been largely overlooked.

We have isolated a rice endophytic bacterium Pantoea sp. Sd-1 (China General Microbiological Culture Collection Centre, no. 6698). It showed strong lignocellulosic biomass degradation capacity with 54.5 % of rice straw weight lost (cellulose, hemicellulose, and lignin content reduced 80.1, 59.6, and 33.1 %, respectively) after 6 days treatment and kraft lignin content reduced 69.1 % after 4 days incubation [18]. It was further identified as Pantoea ananatis Sd-1 by whole genome sequencing analysis. Generally the Pantoea genus is recognized for its plant pathogenesis [19]. However, it has also been reported as lignocellulose degradation bacterium. Pantoea sp. SL1_M5 isolated from invasive woodwasp Sirex noctilio and displayed degrading ability toward cellulose and its derivatives [20]. A few other P. ananatis strains were also reported to possess carbohydrate degradation enzymes [14, 21]. Pantoea sp. RCT2 isolated from pulp paper mill effluent and showed ligninolytic enzyme activity and strong pulp paper mill effluent decolorisation capacity [22]. The Pantoea sp NII-153 isolated from soil showed high strength lignin derivative degradation ability [23]. These indicated that Pantoea genus strains might be potential resources for lignocellulosic biomass biotransformation industry.

Pantoea ananatis Sd-1 possesses the degradation ability for both lignocellulosic polysaccharides and lignin. Uncovering corresponding genes and enzymes in its lignocellulolytic system will be essential for its efficient application in biofuel industry. In the present study, we sequenced the whole genome of P. ananatis Sd-1 and compared its characteristics to those of five typical cellulolytic and ligninolytic bacteria as well as eight other P. ananatis strains. We identified candidate genes involved in lignocellulose degradation through genomic analysis, and some of them were confirmed by analyzing their expression levels. Comparative proteome analysis was also conducted to differentiate the enzyme cocktails produced by P. ananatis Sd-1 when it grew on different carbon sources. In addition, this secretome was also evaluated for rice straw saccharification. Furthermore, several typical lignocellulolytic enzymes activity were detected and compared.

Lignocellulose degrading microorganisms and their degradative systems possess great potential for biofuels production [27]. To date few microorganisms with the exception of white-rot fungi have been reported to have the ability to degrade cellulose, hemicellulose, and lignin, simultaneously [28]. However, currently there is no commercial product for lignocellulose biodegradation, in part due to the practical challenges of fungal enzyme expression and fungal genetic manipulation [29]. Recent work indicates that the diverse bacterial lignocellulolytic systems may be more significant than previously thought, but their degradation catabolism has not been well characterized [30, 31]. Our previous study found that the endophytic bacterium P. ananatis Sd-1 had strong comprehensive ability to degrade cellulose, hemicellulose, and lignin [18]. Characterization of its degradation system will be very useful for utilization of plant lignocellulose for production of biofuels.

According to CAZy family classification, GHs, CEs, PLs, and CBMs reportedly play main roles in lignocellulosic polysaccharide degradation. Similarly, it has also been reported that AAs contribute largely to lignin breakdown [32]. Bioinformatic comparisons of CAZy in P. ananatis Sd-1 to other known cellulolytic and ligninolytic bacteria as well as P. ananatis strains confirmed its lignocellulolytic capabilities. Moreover, the biodegradation system in P. ananatis Sd-1 seems has its own characteristics.

In general, cellulose degradation is attributed to the synergistic action of three classes of glycosyl hydrolases: (1) endoglucanases, (2) exoglucanases, and (3) β-glucosidases [33]. CBM, a non-catalytic polysaccharide-recognizing module possessed by some GHs, influences enzymes activities by attaching to the substrate [34]. We found that only five GHs containing CBM domains in P. ananatis Sd-1. The value is far less than that reported in C. bescii DSM 6725 (14) and C. thermocellum ATCC 27405 (38). C. thermocellum ATCC 27405 is a typical cellulosome production strain [35], which might contain multiple CBM domains for cellulose substrates attachment. Its cellulosome subunits have been identified by the proteomic analysis [8]. Caldicellulosiruptor bescii genus are well-studied multifunctional enzymes production strains. Their multifunctional enzymes generally contain one or several CBMs [36]. Correspondingly, a group of glycosidases with CBM3 domains have been identified in the secretome of C. bescii when grown on crystalline cellulose. This implied that their enzymes might exist as free enzymes or are incorporated into cellulosomes [37]. In contrast, P. ananatis Sd-1 possesses typical cellulase genes encoding endoglucanase, β-glucosidase, and exoglucanase lacking CBM domains. Upon comparison of expression levels, we found that transcripts of all detected genes encoding these enzymes were significantly up-regulated in P. ananatis Sd-1 cultured in rice straw-containing medium (at least P < 0.05). Higher activities of these enzymes were also all detected in the supernatant of rice straw cultures relative to those in glucose cultures. Correspondingly, we found that endoglucanase (egl), exoglucanase (exgl), and two β-glucosidase enzymes (bgl1 and bgl2) all contain signal peptides. These four enzymes as well as another β-glucosidases (Bgl3) could only be identified in supernatant of rice straw cultures. These evidences indicate that the cellulase systems in P. ananatis Sd-1 might be composed of free single enzymes, which are soluble and diffuse away from the cell under culture conditions.

Several enzymes, including xylanases and esterases, are needed to completely degrade the hemicellulose polysaccharides [38]. No typical xylanase gene was found, but it was noticed that the CEs number (25) present in P. ananatis Sd-1 are strikingly higher than that of all the compared strains, including other eight P. ananatis strains. There are nine esterases belonging to CEs in P. ananatis Sd-1: Ets1 Y903_RS0120815 and est2 Y903_RS0117780 belonging to the CE10 family shared 85 and 75 % similarity with the endo-1,4-beta-xylanase (GenBank: CRH34589.1) and xynB (GenBank: WP_052270754), respectively. Both sequences of the two proteins contain signal peptides. Members of CE10 have been found to possess both carboxylesterase and xylanase activities [39]. In addition, two enzymes from CE10 have been identified in the secretome of Penicillium purpurogenum when it was cultured with acetylated xylan substrates [40]. Consistent with these reports, in P. ananatis Sd-1, the transcript levels of four detected esterase genes including est1 and est2 were significantly up-regulated (at least P < 0.01) in rice straw cultures, which corresponded with high xylanase-like activity detected in the same cultures. It should be mentioned that the maximum xylanase-like activity reached 25.29 ± 0.87 U/ml in rice straw cultures, which is higher than that was previously reported for cellulolytic fungus [26]. Furthermore, two esterases including est2 were only detected in rice straw cultures. These results are consistent with the previous reports, which indicated that P. ananatis Sd-1 might possess additional enzymes for efficient breakdown of hemicellulose and further characterization of these enzymes might broaden existing knowledge of hemicellulose degradation system. Interestingly, no CE10 was found in the eight other compared P. ananatis strains, except for typical xylanase genes. The difference between P. ananatis Sd-1 and other P. ananatis strains might be attributed to an evolutionary adaption of P. ananatis Sd-1 to the endophytic environment, which was different with that of other compared P. ananatis strains [21, 41].

Rice straw saccharification by the P. ananatis Sd-1 secretome from rice straw cultures confirmed the existence of its lignocellulosic polysaccharides degradation system. The highest yield of released reducing sugars was attained from the third day’s secretome, whereas the lowest yield was observed from the sixth day’s secretome. These were approximately consistent with results of enzymes activities detection. Analyses indicated that the secretome of P. ananatis Sd-1 release higher amounts of reducing sugars than that reported for Phoma exigua secretome which used pretreated paddy straw or wheat straw as substrates [42, 43]. These indicated that P. ananatis Sd-1 is deserved an in-depth exploration for its enzyme system, such as GHs and CEs.

Although bacterial enzymes associated with lignin degradation are not well characterized in comparison to their fungal counterparts, it has been reported that bacteria have their own unique enzymes for lignin degradation, e.g., Cα-dehydrogenase (LigD), glutathione-dependent β-etherase enzymes (LigE, F, G), demethylase enzyme (LigX), and DyP-type peroxidases [30, 44–46]. Hence it is not surprising that there is no classic fungi lignin degradation enzyme in genome of P. ananatis Sd-1. Recent reports of multicopper oxidases demonstrating laccase activities have increased [47, 48]. Secreted catalase/hydroperoxidase, which possess similar heme-containing domains to lignin peroxidases, is associated with strong peroxidase activity and was shown to play an important role in the lignin degradation process in E. lignolyticus SCF1 [25]. In P. ananatis Sd-1, significantly up-regulated transcript levels of 2 multicopper oxidase genes (P < 0.001) and catalase/hydroperoxidase gene (P < 0.05) were found to corresponded with higher laccase-like and lignin peroxidase-like activities in rice straw cultures relative to glucose cultures. These results suggest that these enzymes might fulfill similar roles to laccase and lignin peroxidase, respectively, in P. ananatis Sd-1. Notably, the maximum value (2.59 ± 0.11 U/ml) for lignin peroxidase-like activity produced by P. ananatis Sd-1 in rice straw culture exceeds previously reported values for any fungus or bacterium [49, 50]. Furthermore, the identification of multicopper oxidase (lac4) and catalase/hydroperoxidase (cat) proteins among secretome of rice straw cultures provides supporting evidence. These data provide a catalog of novel bacterial enzymes participating in the deconstruction of lignin. Indeed, we have cloned one multicopper oxidase gene lac4 (Y903_RS0119240) and the recombinant enzyme can efficiently degrade lignin and decolorize dyes in vitro [51].

During fungal degradation of lignin, GMC family oxidoreductases such as glucose oxidase, aryl-alcohol oxidase, methanol oxidase, and pyranose oxidase were reported to participate by providing the hydrogen peroxide required by ligninolytic peroxidases [52, 53]. According to our phylogenetic analysis, the pyranose oxidase enzyme might be the GMC family oxidoreductase equivalent in P. ananatis Sd-1. In addition, hydrogen peroxide production is a typical character in the Fenton chemistry pathway. Quinone oxidoreductases were found to drive Fenton chemistry systems participated in lignin degradation in the brown-rot fungus Gloeophyllum trabeum [54]. Interestingly, transcript levels of the GMC family oxidoreductase and quinone oxidoreductases in P. ananatis Sd-1 were all significantly up-regulated in rice straw cultures relative to glucose cultures (at least P < 0.01). These data provide clues that the Fenton chemistry pathway might also exist in bacteria, although this process was shown previously to only exist in fungi. Further research in support of this hypothesis is currently underway.

In our study, it was surprisingly to find that there are nine AAs in P. ananatis Sd-1, while they are absent from the other compared strains except for one. Considering their supposed role in lignin degradation [32], the enzyme system in P. ananatis Sd-1 might be a new resource for lignin decomposition. Although in this study, we only confirmed two AAs (AA2-Catalase/hydroperoxidase and AA3-GMC family oxidoreductase) were related to lignin degradation, we cannot exclude the possibility that other P. ananatis Sd-1 AAs are lignin degradation relevant enzymes. This speculation cannot be verified due to the limited reference information presently available.

Possessing unique enzymes for lignin degradation is an attractive characteristic in a bacterial degradative system. For example, GSTs (LigE, F, G) are able to cleave the β-aryl ether in Sphingomonas paucimobilis SYK-6 [44]. According to phylogenetic analysis, four GST genes in P. ananatis Sd-1 might possess β-etherase activity. Transcript levels of these four GST genes were significantly up-regulated when rice straw was used as substrate (at least P < 0.05). It was not surprising that GST remained undetected in the secretome of P. ananatis Sd-1, as none of the four GSTs possesses signal peptides. This is consistent with a report of GSTs in Sphingomonas paucimobilis SYK-6 [44]. Our results suggested that a distinct bacteria β-aryl ether degradation pathway might also exist for lignin degradation by P. ananatis Sd-1.

Our results confirmed that microbial lignocellulolytic enzymes production is dependent on the carbon source [11]. Growth comparisons of P. ananatis Sd-1 in different media indicted poorer growth in rice straw compared to glucose. However, the secreted protein contents from rice straw cultures were higher. These indicated that the discrepancy in enzymes activities was not related to cell density, but rather differences in gene expression in response to the carbon sources.

According to the work in this paper, further studies should focus on the individual enzyme function to aid in development of efficient biological treatment schemes. Elucidation of the mechanistic features of P. ananatis Sd-1 lignocellulolytic system is also indispensable for its further biotechnological application.

This study confirms the strong lignocellulosic biomass degradation capacity of rice endophytic bacterium P. ananatis Sd-1. Genomic analysis revealed that P. ananatis Sd-1 possesses abundant lignocellulolytic enzymes genes, especially for ligninolytic enzymes relevant genes compared to other cellulolytic and ligninolytic bacteria. A part of these genes transcript levels were found to be highly up-regulated when induced by lignocellulosic substrates, which was corroborated by the secretomic and enzyme activity profiles. In addition, the secretome of P. ananatis Sd-1 showed strong saccharification ability toward rice straw. Furthermore, we conclude that the cellulase system of P. ananatis Sd-1 might function as free single enzymes. The β-aryl ether degradation and Fenton chemistry pathways might exist in lignin degradation process of P. ananatis Sd-1. The discovery of lignocellulolytic system in P. ananatis Sd-1 is crucial for a deeper understanding of its degradation mechanisms and the regulation of their extracellular proteins. This work and further evaluation will therefore be a cornerstone for future applications of P. ananatis Sd-1 in biofuels production.