Effects of Endobacterium (Stenotrophomonas maltophilia) on Pathogenesis-Related Gene Expression of Pine Wood Nematode (Bursaphelenchus xylophilus) and Pine Wilt Disease

Pine wilt disease (PWD) caused by the pine wood nematode (PWN), Bursaphelenchus xylophilus, is responsible for devastating epidemics in pine trees in Asia and Europe. Recent studies showed that bacteria carried by the PWN might be involved in PWD. However, the molecular mechanism of the interaction between bacteria and the PWN remained unclear. Now that the whole genome of B. xylophilus (Bursaphelenchus xylophilus) is published, transcriptome analysis is a unique method to study the role played by bacteria in PWN. In this study, the transcriptome of aseptic B. xylophilus, B. xylophilus treated with endobacterium (Stenotrophomonas maltophilia NSPmBx03) and fungus B. xylophilus were sequenced. We found that 61 genes were up-regulated and 830 were down-regulated in B. xylophilus after treatment with the endobacterium; 178 genes were up-regulated and 1122 were down-regulated in fungus B. xylophilus compared with aseptic B. xylophilus. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses were used to study the significantly changed biological functions and pathways for these differentially expressed genes. Many pathogenesis-related genes, including glutathinone S-transferase, pectate lyase, ATP-binding cassette transporter and cytochrome P450, were up-regulated after B. xylophilus were treated with the endobacterium. In addition, we found that bacteria enhanced the virulence of PWN. These findings indicate that endobacteria might play an important role in the development and virulence of PWN and will improve our understanding of the regulatory mechanisms involved in the interaction between bacteria and the PWN.


Introduction
Pine wilt disease (PWD) is a worldwide forest disease affecting several species of pine trees (Pinus spp.), and it has been epidemic in Asia (Japan, China, and Korea) and in Portugal [1][2][3][4]. In these countries, PWD has killed a large number of pine trees and caused serious economic losses and ecological damages [5]. The pine wood nematode (PWN), Bursaphelenchus xylophilus [6], is considered the causal pathogenic agent of PWD and is designated an important quarantine species [7]. The PWN invades healthy pines via sawyer beetles (Monochamus spp.), particularly Monochamus alternatus during feeding or oviposition [8,9]. After PWNs invade the trees, they move inside the host rapidly via mapped to the reference genome and PWN genes, respectively. Over 70% of our reads were uniquely mapped reads.

Differentially Expressed Genes in Transcriptome Sequence of PWNs Treated with Endobacterium and Fungus PWNs
Differentially expressed genes (DEGs) between aseptic PWNs and PWNs treated with S. maltophilia NSPmBx03 were examined. Compared to aseptic PWNs, a total of 891 significant DEGs were identified with at least a two-fold change at the expression level and false discovery rate (FDR) <0.001. Of these genes, 61 were up-regulated and 830 down-regulated. A total of 1300 genes were identified as DEGs between aseptic PWNs and fungus PWNs. Of these genes, 178 were up-regulated and 1122 down-regulated. The quantity of DEGs in PWNs treated with one endobacterial strain was less than that in fungus PWNs. The number of DEGs between PWNs treated with endobacterium and fungus PWNs was 625, which was less than that between aseptic PWNs and fungus PWNs ( Figure S1). These results implied that bacteria affected expression of some genes of the PWN.

Functional Annotation and Classification for DEGs of the PWN
Gene Ontology (GO) analysis was used to search for significantly enriched GO terms of DEGs. The main GO categories included molecular function, cellular component, and biological process. Compared to aseptic PWNs, the "catalytic activity" (GO:0003824), "binding" (GO:0005488), "hydrolase activity" (GO:0016817), "structural molecule activity" (GO:0005198), and "transporter activity" (GO:0005215) categories were among the top molecular function terms in both PWNs treated with NSPmBx03 and fungus PWNs (Figure 1). In the cell component category, most DEGs were involved in the "cell" (GO:0005623), "cell part" (GO:0044464), "organelle" (GO:0043266), "membrane" (GO:0016020), and "membrane part" (GO:0044425). In addition, "single-organism process" (GO:0044699), "developmental process" (GO:0003006), "multicellular organismal process" (GO:0032501), and "metabolic process" (GO:0008152) were the top categories enriched in the biological process. In general, the aforementioned GO terms accounted for the majority of the DEGs. Furthermore, some GO terms play an important role in PWN development, such as embryo development, juvenile development, regulation of growth rate and hydrolase activity, in response to external stimuli. were mapped to the reference genome and PWN genes, respectively. Over 70% of our reads were uniquely mapped reads.

Differentially Expressed Genes in Transcriptome Sequence of PWNs Treated with Endobacterium and Fungus PWNs
Differentially expressed genes (DEGs) between aseptic PWNs and PWNs treated with S. maltophilia NSPmBx03 were examined. Compared to aseptic PWNs, a total of 891 significant DEGs were identified with at least a two-fold change at the expression level and false discovery rate (FDR) <0.001. Of these genes, 61 were up-regulated and 830 down-regulated. A total of 1300 genes were identified as DEGs between aseptic PWNs and fungus PWNs. Of these genes, 178 were up-regulated and 1122 down-regulated. The quantity of DEGs in PWNs treated with one endobacterial strain was less than that in fungus PWNs. The number of DEGs between PWNs treated with endobacterium and fungus PWNs was 625, which was less than that between aseptic PWNs and fungus PWNs ( Figure S1). These results implied that bacteria affected expression of some genes of the PWN.

Functional Annotation and Classification for DEGs of the PWN
Gene Ontology (GO) analysis was used to search for significantly enriched GO terms of DEGs. The main GO categories included molecular function, cellular component, and biological process. Compared to aseptic PWNs, the "catalytic activity" (GO:0003824), "binding" (GO:0005488), "hydrolase activity" (GO:0016817), "structural molecule activity" (GO:0005198), and "transporter activity" (GO:0005215) categories were among the top molecular function terms in both PWNs treated with NSPmBx03 and fungus PWNs (Figure 1). In the cell component category, most DEGs were involved in the "cell" (GO:0005623), "cell part" (GO:0044464), "organelle" (GO:0043266), "membrane" (GO:0016020), and "membrane part" (GO:0044425). In addition, "single-organism process" (GO:0044699), "developmental process" (GO:0003006), "multicellular organismal process" (GO:0032501), and "metabolic process" (GO:0008152) were the top categories enriched in the biological process. In general, the aforementioned GO terms accounted for the majority of the DEGs. Furthermore, some GO terms play an important role in PWN development, such as embryo development, juvenile development, regulation of growth rate and hydrolase activity, in response to external stimuli.

Analysis of Cell Wall Degradation-Related Genes
Enzymes of cell wall degradation are intrinsically related to nematode pathogenicity in host pines [25]. Compared with aseptic PWNs, the expression levels of six cell wall degradation-related genes changed significantly (over two-fold) in PWNs treated with S. maltophilia NSPmBx03. Of these DEGs, three pectate lyase genes (BUX.s01259.20, BUX.s01661.75, and BUX.s01530.1) and one cellulase gene (BUX.s01259.61) were up-regulated over two-fold ( Figure 2). In fungus PWNs, eight cell wall degradation-related genes were DEGs (over two-fold). The expression of BUX.s01530.1 was up-regulated 5.5-fold. However, expression levels of cellulase genes were significantly down-regulated ( Figure 2). These results indicated that endobacteria could affect the expression levels of cell wall degradation-related genes of PWNs and these genes changed much more evidently in fungus PWNs.

Analysis of Cell Wall Degradation-Related Genes
Enzymes of cell wall degradation are intrinsically related to nematode pathogenicity in host pines [25]. Compared with aseptic PWNs, the expression levels of six cell wall degradation-related genes changed significantly (over two-fold) in PWNs treated with S. maltophilia NSPmBx03. Of these DEGs, three pectate lyase genes (BUX.s01259.20, BUX.s01661.75, and BUX.s01530.1) and one cellulase gene (BUX.s01259.61) were up-regulated over two-fold ( Figure 2). In fungus PWNs, eight cell wall degradation-related genes were DEGs (over two-fold). The expression of BUX.s01530.1 was up-regulated 5.5-fold. However, expression levels of cellulase genes were significantly down-regulated ( Figure 2). These results indicated that endobacteria could affect the expression levels of cell wall degradation-related genes of PWNs and these genes changed much more evidently in fungus PWNs.

Detoxification-Related Gene Analysis
The capacity for detoxification is a key factor in PWN colonization of pines [27]. Compared with aseptic PWNs, the expression level of several detoxification-related genes changed significantly in PWNs treated with S. maltophilia NSPmBx03 and fungus PWNs, and their functions included glutathinone S-transferase (GST) activity, ATP-binding cassette (ABC) transporter, venom allergen-like protein VAP1, rethinol dehydrogenase, cytochrome P450 (CYP), and carboxylesterase ( Figure 3). GST and CYP genes were primary DEGs related to detoxification. These results suggested that the endobacterium and other bacteria genera have an important role in regulation of detoxification-related gene expression, especially CYP and GST. The expression level of BUX.s01147.131 (ABC transporter) and BUX.s00647.111 (GST) were up-regulated over three-fold in PWNs treated with NSPmBx03. The expression of BUX.s00974.3 (ABC transporter) and BUX.s01092.237 (CYP) were up-regulated 8-fold and 5.5-fold in wild-type PWNs, respectively. In addition, the expression levels of BUX.s01147.131 and BUX.s00647.111 were more greatly up-regulated in PWN treated with NSPmBx03 than in fungus PWNs. These results indicated that the bacteria had an important function in the expression of these detoxification-related genes of PWN.

Detoxification-Related Gene Analysis
The capacity for detoxification is a key factor in PWN colonization of pines [27]. Compared with aseptic PWNs, the expression level of several detoxification-related genes changed significantly in PWNs treated with S. maltophilia NSPmBx03 and fungus PWNs, and their functions included glutathinone S-transferase (GST) activity, ATP-binding cassette (ABC) transporter, venom allergen-like protein VAP1, rethinol dehydrogenase, cytochrome P450 (CYP), and carboxylesterase ( Figure 3). GST and CYP genes were primary DEGs related to detoxification. These results suggested that the endobacterium and other bacteria genera have an important role in regulation of detoxification-related gene expression, especially CYP and GST. The expression level of BUX.s01147.131 (ABC transporter) and BUX.s00647.111 (GST) were up-regulated over three-fold in PWNs treated with NSPmBx03. The expression of BUX.s00974.3 (ABC transporter) and BUX.s01092.237 (CYP) were up-regulated 8-fold and 5.5-fold in wild-type PWNs, respectively. In addition, the expression levels of BUX.s01147.131 and BUX.s00647.111 were more greatly up-regulated in PWN treated with NSPmBx03 than in fungus PWNs. These results indicated that the bacteria had an important function in the expression of these detoxification-related genes of PWN.

Reproduction-Related Gene Analysis
A number of reproduction-related genes were differentially expressed in PWNs treated with S. maltophilia NSPmBx03 and fungus PWNs compared with aseptic PWNs (Figure 4). The functions of reproduction-related genes were involved with juvenile development, embryo development, anatomical structure morphogenesis, endothelin-converting enzyme, growth differentiation factor, developmental process involved in reproduction, regulation of growth rate, and multicellular organismal aging. Most juvenile and embryo development-related genes were down-regulated. The expression level of BUX.s01513.215 (developmental process involved in reproduction) and BUX.s00116.877 (juvenile development) were down-regulated over six-fold in PWNs treated with NSPmBx03, and the expression of BUX.s00532.5 (juvenile development) was down-regulated over six-fold in fungus PWNs. These results suggested that PWN-associated bacteria could slow the growth of PWN.

Reproduction-Related Gene Analysis
A number of reproduction-related genes were differentially expressed in PWNs treated with S. maltophilia NSPmBx03 and fungus PWNs compared with aseptic PWNs (Figure 4). The functions of reproduction-related genes were involved with juvenile development, embryo development, anatomical structure morphogenesis, endothelin-converting enzyme, growth differentiation factor, developmental process involved in reproduction, regulation of growth rate, and multicellular organismal aging. Most juvenile and embryo development-related genes were down-regulated. The expression level of BUX.s01513.215 (developmental process involved in reproduction) and BUX.s00116.877 (juvenile development) were down-regulated over six-fold in PWNs treated with NSPmBx03, and the expression of BUX.s00532.5 (juvenile development) was down-regulated over six-fold in fungus PWNs. These results suggested that PWN-associated bacteria could slow the growth of PWN.

Reproduction-Related Gene Analysis
A number of reproduction-related genes were differentially expressed in PWNs treated with S. maltophilia NSPmBx03 and fungus PWNs compared with aseptic PWNs (Figure 4). The functions of reproduction-related genes were involved with juvenile development, embryo development, anatomical structure morphogenesis, endothelin-converting enzyme, growth differentiation factor, developmental process involved in reproduction, regulation of growth rate, and multicellular organismal aging. Most juvenile and embryo development-related genes were down-regulated. The expression level of BUX.s01513.215 (developmental process involved in reproduction) and BUX.s00116.877 (juvenile development) were down-regulated over six-fold in PWNs treated with NSPmBx03, and the expression of BUX.s00532.5 (juvenile development) was down-regulated over six-fold in fungus PWNs. These results suggested that PWN-associated bacteria could slow the growth of PWN.

Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Enrichment Analysis
The pathway enrichment for DEGs was performed using KEGG. From the transcriptome of PWNs treated with S. maltophilia NSPmBx03, 75 DEGs were involved in detoxification pathways containing drug metabolism cytochrome P450, metabolism of xenobiotics by cytochrome P450, lysosome, and glutathione metabolism (p < 0.05) ( Figure 5). Twenty-eight genes were enriched in turpentine-related metabolism including ether lipid metabolism, arachidonic acid metabolism, polyketide sugar unit biosynthesis, and retinol metabolism. These metabolic pathways may be important for the PWN to parasitize pines. In fungus PWNs, 238 genes (28.4%) were enriched in detoxification pathways containing lysosome, metabolism of xenobiotics by cytochrome P450, drug metabolism cytochrome P450, peroxisome, glutathione metabolism, drug metabolism-other enzymes, ascorbate and aldarate metabolism. Fifty-nine genes were involved in turpentine-related metabolism containing arachidonic acid metabolism, rethinol metabolism, ubiquinone and other terpenoid-quinone biosynthesis, and riboflavin metabolism. Amino acid metabolism and fat metabolism were the prominent pathways in both PWNs treated with NSPmBx03 and fungus PWNs. The results indicated that the endobacterium and other bacteria could affect detoxification-related metabolism. The bacteria of fungus PWNs affected gene expression to a greater extent than S. maltophilia NSPmBx03. In addition, endobacteria and other bacteria may help nutrition metabolism of PWNs.

Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway Enrichment Analysis
The pathway enrichment for DEGs was performed using KEGG. From the transcriptome of PWNs treated with S. maltophilia NSPmBx03, 75 DEGs were involved in detoxification pathways containing drug metabolism cytochrome P450, metabolism of xenobiotics by cytochrome P450, lysosome, and glutathione metabolism (p < 0.05) ( Figure 5). Twenty-eight genes were enriched in turpentine-related metabolism including ether lipid metabolism, arachidonic acid metabolism, polyketide sugar unit biosynthesis, and retinol metabolism. These metabolic pathways may be important for the PWN to parasitize pines. In fungus PWNs, 238 genes (28.4%) were enriched in detoxification pathways containing lysosome, metabolism of xenobiotics by cytochrome P450, drug metabolism cytochrome P450, peroxisome, glutathione metabolism, drug metabolism-other enzymes, ascorbate and aldarate metabolism. Fifty-nine genes were involved in turpentine-related metabolism containing arachidonic acid metabolism, rethinol metabolism, ubiquinone and other terpenoid-quinone biosynthesis, and riboflavin metabolism. Amino acid metabolism and fat metabolism were the prominent pathways in both PWNs treated with NSPmBx03 and fungus PWNs. The results indicated that the endobacterium and other bacteria could affect detoxification-related metabolism. The bacteria of fungus PWNs affected gene expression to a greater extent than S. maltophilia NSPmBx03. In addition, endobacteria and other bacteria may help nutrition metabolism of PWNs.

Symptoms of Pines Inoculated with PWNs
Pines inoculated with different PWNs showed differences in symptoms. The time of initial symptoms of fungus PWNs and PWNs treated with bacteria appeared 13 days after inoculation and 17 days for aseptic PWNs. After 21 days of inoculation, pines inoculated with fungus PWNs had an infection rate of 100% and a disease severity index (DSI) of 83.3, while S. maltophilia NSPmBx03 treatment infection rate was 66.7% and DSI was 58.3; aseptic PWN treatment infection rate was 33.3% and DSI was 33.3 (Table 1). After 26 days of inoculation, infection rates of all pines with different treated PWNs were 100%, but aseptic PWNs had the lowest DSI. After 30 days of inoculation, only one pine inoculated with aseptic PWNs was not completely wilted. After 35 days of inoculation, all pine needles were brown and wilting. Pines inoculated with S. maltophilia NSPmBx03 and sterile water remained healthy (Figure 7). These results indicated that the S. maltophilia NSPmBx03 enhanced the virulence of PWNs.

Symptoms of Pines Inoculated with PWNs
Pines inoculated with different PWNs showed differences in symptoms. The time of initial symptoms of fungus PWNs and PWNs treated with bacteria appeared 13 days after inoculation and 17 days for aseptic PWNs. After 21 days of inoculation, pines inoculated with fungus PWNs had an infection rate of 100% and a disease severity index (DSI) of 83.3, while S. maltophilia NSPmBx03 treatment infection rate was 66.7% and DSI was 58.3; aseptic PWN treatment infection rate was 33.3% and DSI was 33.3 (Table 1). After 26 days of inoculation, infection rates of all pines with different treated PWNs were 100%, but aseptic PWNs had the lowest DSI. After 30 days of inoculation, only one pine inoculated with aseptic PWNs was not completely wilted. After 35 days of inoculation, all pine needles were brown and wilting. Pines inoculated with S. maltophilia NSPmBx03 and sterile water remained healthy (Figure 7). These results indicated that the S. maltophilia NSPmBx03 enhanced the virulence of PWNs.

Discussion
Symbiotic relationships between nematodes and bacteria are common in nature. Recent studies showed that bacteria carried by the PWN were likely to play a role in PWD [18,21,28]. Zhang et al. [29] reported that Stenotrophomonas was the most dominant group in the PWN. Wu et al. [13] isolated several endobacterial strains from PWNs, and S. maltophilia was a dominant species. The specific and functional diversities of endobacteria from various isolates of B. xylophilus have been studied [13]. Tian et al. [21] reported that endobacteria S. maltophilia NSBx.14 could decrease the reproduction of ZL1 in fungus-conditions. However, the molecular mechanisms between endobacteria and the PWN are poorly understood. In this study, we sequenced the transcriptome of B. xylophilus treated with S. maltophilia NSPmBx03 for the first time and found that NSPmBx03 affected the gene expression level of B. xylophilus (Table S3).
Pectinases are the primary cell-degrading enzymes of the PWN secretome, intrinsically related to nematode pathogenicity in host pines [30,31]. Pectin is the most complex component of the pine cell wall polysaccharides, and plays a critical role in resisting invasion by PWN [32]. To invade the host successfully, PWN needs to break down this barrier. Therefore, pectinases are essential to allow PWN to infect its plant host. Qiu et al. found that pectin lyase genes of PWNs were highly expressed when PWNs infected P. thunbergii [33]. From the secretion of the PWN esophageal gland, DeBoer et al. [34] cloned Bxpel1 and Bxpel2, which enhanced the ability of PWNs to feed and migrate in the host pine. Our results showed that the endobacterium S. maltophilia NSPmBx03 affected expression of genes of carbohydrate active enzymes in the PWN. The expression of the pectate lyase gene BUX.s01530.1 was up-regulated 5.5-fold in fungus PWNs and 5-fold in PWN treated with endobacteria compared to control treatment ( Figure 2). After P. massoniana was inoculated with aseptic PWNs, PWNs treated with endobacterium, and fungus PWNs, respectively, we found that the fastest group lead to pine trees wilting was fungus PWNs, followed by PWNs treated with endobacterium, and the last was the pines inoculated with aseptic PWNs. As previously reported, the bacteria species associated with B. xylophilus AmA3 were Stenotrophomonas, Rhizobiaceae (unclassified), Chitinophaga, Oxalobacteraceae (unclassified), and Ochrobactrum [23]. These results indicated that S. maltophilia NSPmBx03 and some other bacteria associated with the PWN could affect pectate lyase gene expression, which may play an important role in parasitizing pines.
Phytotoxins such as benzoic acid, 8-hydroxycarbotanacetone, and 10-hydroxyverbenone were identified and characterized from PWN-infested pines [35]. These toxic substances could defend against invasion and decrease the reproduction and migration rate of the PWN in host pines [36].

Discussion
Symbiotic relationships between nematodes and bacteria are common in nature. Recent studies showed that bacteria carried by the PWN were likely to play a role in PWD [18,21,28]. Zhang et al. [29] reported that Stenotrophomonas was the most dominant group in the PWN. Wu et al. [13] isolated several endobacterial strains from PWNs, and S. maltophilia was a dominant species. The specific and functional diversities of endobacteria from various isolates of B. xylophilus have been studied [13]. Tian et al. [21] reported that endobacteria S. maltophilia NSBx.14 could decrease the reproduction of ZL1 in fungus-conditions. However, the molecular mechanisms between endobacteria and the PWN are poorly understood. In this study, we sequenced the transcriptome of B. xylophilus treated with S. maltophilia NSPmBx03 for the first time and found that NSPmBx03 affected the gene expression level of B. xylophilus (Table S3).
Pectinases are the primary cell-degrading enzymes of the PWN secretome, intrinsically related to nematode pathogenicity in host pines [30,31]. Pectin is the most complex component of the pine cell wall polysaccharides, and plays a critical role in resisting invasion by PWN [32]. To invade the host successfully, PWN needs to break down this barrier. Therefore, pectinases are essential to allow PWN to infect its plant host. Qiu et al. found that pectin lyase genes of PWNs were highly expressed when PWNs infected P. thunbergii [33]. From the secretion of the PWN esophageal gland, DeBoer et al. [34] cloned Bxpel1 and Bxpel2, which enhanced the ability of PWNs to feed and migrate in the host pine. Our results showed that the endobacterium S. maltophilia NSPmBx03 affected expression of genes of carbohydrate active enzymes in the PWN. The expression of the pectate lyase gene BUX.s01530.1 was up-regulated 5.5-fold in fungus PWNs and 5-fold in PWN treated with endobacteria compared to control treatment (Figure 2). After P. massoniana was inoculated with aseptic PWNs, PWNs treated with endobacterium, and fungus PWNs, respectively, we found that the fastest group lead to pine trees wilting was fungus PWNs, followed by PWNs treated with endobacterium, and the last was the pines inoculated with aseptic PWNs. As previously reported, the bacteria species associated with B. xylophilus AmA3 were Stenotrophomonas, Rhizobiaceae (unclassified), Chitinophaga, Oxalobacteraceae (unclassified), and Ochrobactrum [23]. These results indicated that S. maltophilia NSPmBx03 and some other bacteria associated with the PWN could affect pectate lyase gene expression, which may play an important role in parasitizing pines.
Phytotoxins such as benzoic acid, 8-hydroxycarbotanacetone, and 10-hydroxyverbenone were identified and characterized from PWN-infested pines [35]. These toxic substances could defend against invasion and decrease the reproduction and migration rate of the PWN in host pines [36]. Futai et al. [37] found that tannic acid significantly accumulated in parenchymal cells of P. thunbergii after PWN invasion. Additionally, PWN inhabits the resin canals of host pines [38] and the resin is a complex mixture of compounds, including terpenoids and cyclic aromatic compounds [39]. These compounds are likely to have nematocidal activity. PWNs must detoxify these toxic compounds in order to invade the pines successfully. Our results showed that abundant derived enzymes were assigned to xenobiotic metabolism pathways in the transcriptome of PWNs treated with NSPmBx03 and fungus PWN. The major pathways were drug metabolism CYP, metabolism of xenobiotics by CYP and glutathione metabolism ( Figure 5). CYP and GST were the primary DEGs of detoxification ( Figure 3). These results were consistent with those of Yan et al., who reported that CYP and GST pathways were the primary detoxification metabolic pathways in PWN [40]. The detoxification process of PWN has been divided into three successive phases [27]. First, CYPs are the most important group of the first phase proteins, making molecules more suitable substrates for downstream. Second, the actual detoxification reaction occurs with GSTs and UDP-glucuronosyl transferases (UGTs) as essential enzymes. In the third phase, an ABC transporter is responsible for the efflux of detoxification molecules. Vicente et al. [41] reported that some PWN-associated bacteria could increase nematode survival under strong oxidative stress. Xu et al. [42] identified the function of CYP genes (BxCYP33C9, BxCYP33C4, and BxCYP3D3) in B. xylophilus, and revealed that these genes may affect their viability, proliferation, pesticide metabolism and pathogenicity. In our study, a number of detoxification-related genes, including GST, ABC transporter, venom allergen-like protein VAP1, retinol dehydrogenase, and CYP, were greatly up-regulated in the PWNs treated with S. maltophilia NSPmBx03 (Figure 3). These results indicated that NSPmBx03 played an important role in the xenobiotic detoxification of PWN, which may help PWN cope with a difficult environment.
A recent study showed that Pseudomonas spp. could promote the reproduction of B. xylophilus [43]. However, some other bacteria, such as Pantoea sp., Serratia marcescens, Buttiauxella agrestis and Enterobacter amnigenus had an inhibitory effect [43]. Tian et al. [21] reported that endobacteria isolated from highly virulent PWN could partially promote the development of nematodes, while endobacteria isolated from lowly virulent PWN may retard the growth of nematodes. Additionally, S. maltophilia NSBx.14, obtained from the highly virulent ZL1 strain of B. xylophilus, can prolong the development time of nematode embryos [22]. We found that the expression of juvenile and embryo development-related genes in PWNs treated with S. maltophilia NSPmBx03 and fungus PWNs were significantly down-regulated ( Figure 4). As in previous results, we found that aseptic PWN had a higher reproduction rate than PWN treated with bacteria on B. cinerea, but completely opposite results were obtained for P. massoniana [44]. Cheng et al. [45] reported that Stenotrophomonas was the most common genera of symbiotic bacteria with PWN, and may play an important role in regulating ecological function. Qiu et al. [33] found that the expression levels of the pectate lyase, CYP, UGT, and ABC transporter genes, and reproduction-related genes of PWN associated with bacteria were up-regulated dramatically by growth on P. thunbergii compared with growth on B. cinerea. Therefore, we speculated that the effect of bacteria on PWN differed in fungus-conditions and parasitic conditions. In fungus-conditions, bacteria might be detrimental to nematodes to some extent. In parasitic conditions, the high-level expression of cell wall degradation-related and detoxification-related genes could help the PWN successfully parasitize pines and grow better in the host. These results will help us to perceive the underlying nature of the relationship between pine wood nematode and bacteria. However, the role of endobacteria in PWD requires further experimental verification.

The PWN and Its Endobacteria
PWNs (AmA3) were isolated from infected P. massoniana in Maanshan, Anhui, China [46]. An endobacterium S. maltophilia NSPmBx03 was initially isolated from inside the body of PWNs extracted from infected P. massoniana in Nanjing, Jiangsu, China. This bacterium was frequently isolated from PWN strains in our previous study [13]. The strains of the PWN and the endobacterium were provided by the Jiangsu Key Laboratory for Prevention and Management of Invasive Species.

Preparation of Bacteria-Free PWNs and Treatment of Aseptic PWNs with Endobacterium
A suspension of 6000 PWNs was used to inoculate six-year-old P. massoniana in the field (Nanjing Forestry University, Nanjing, China) [33]. The PWNs were isolated from dead P. massoniana using a Baermann funnel under aseptic conditions. Then, the PWNs were pipetted into a potato dextrose agar plate (6-cm diameter) containing B. cinerea. After growth at 25˝C for 7 days, the nematodes were collected in 10-mL centrifuge tubes and treated as fungus PWN (Bx_fungus).
The above-mentioned suspension (1 mL) of PWNs was poured onto a sterilized cover slip in a Petri dish and incubated at 25˝C for depositing eggs during 4-6 h, and the nematodes discarded after egg deposition [47]. The nematode eggs were collected and soaked in 15% H 2 O 2 for 60 min at 25˝C. The eggs were rinsed three times with sterile water and placed on mycelia of B. cinerea grown on PDA medium at 25˝C in darkness. The hatched nematodes were collected in 10-mL centrifuge tubes under aseptic conditions and treated as aseptic PWNs (Bx_a). Aseptic PWNs was the control treatment.
The endobacterial strain of S. maltophilia NSPmBx03 maintained on slant nutrient agar medium in tubes was transferred to 50 mL of beef extract-peptone liquid medium and cultured on a shaker for 24 h. The concentration of S. maltophilia NSPmBx03 was adjusted to 8.0ˆ10 8 CFU (Colony-Forming Units) and 200 µL of suspension was sprayed onto PDA plates containing B. cinerea [21]. Each plate was inoculated with 1000 bacteria-free nematodes (AmA3) and incubated for 7 days at 25˝C. Then PWNs were collected into 10-mL centrifuge tubes with a Baermann funnel under aseptic conditions. The nematodes (Bx_b) were washed three times with sterile water.

RNA Extraction, cDNA Preparation and Illumina Sequencing
Nine samples (each treatment had three replicates) were immediately frozen in liquid nitrogen and stored at´75˝C until use. Total RNA was respectively extracted from frozen nematode samples (Bx_a, Bx_b, and Bx_fungus), using a RNAprep Kit (Tiangen, Beijing, China) and purified using a RNAclean Kit (Tiangen) according to the manufacturer's instructions. RNA integrity and quantity were determined with an Agilent 2100 Bioanalyzer (Agilent Technologies, Redwood City, CA, USA) before cDNA synthesis. One of the three RNA samples from each treatment PWN was used for RNA-seq. The mRNA was isolated from total RNA by Oligo (dT) magnetic beads. Isolated mRNA strands were fragmented and used as templates to synthesize cDNA using a cDNA Synthesis SystemKit (Roche Applied Science, Mannheim, Germany) [48]. The short fragments that were purified and resolved with EB buffer were subjected to end repair and addition of single nucleotide A (adenine). After that, the short fragments were connected with adapters. The qualified fragments of three cDNA samples were respectively selected for PCR amplification as templates. The quality of cDNA libraries was determined with an Agilent 2100 Bioanalyzer and ABI StepOnePlus Real-Time PCR System (Life Technologies, Grand Island, NY, USA). Each cDNA library was sequenced in a single lane of the Illumina HiSeq TM 2000 system using paired-end protocols according to the manufacturer's instructions at BeiJing Genomics Institute (Shenzhen, China). The RNA-Seq experiments have three technical replicates. The raw data were deposited in the NCBI Short Read Archive (SAR) and are accessible through SAR accession numbers: SRR2336946, SRR2342509, and SRR2342510.
To obtain clean reads, adapter, reads in which unknown bases were more than 10%, low quality reads were removed from raw reads using the SeqPrep program (https://github.com/jstjohn/SeqPrep).

Read Mapping and DEG Analysis
The clean reads of three PWN samples with different treatment methods were mapped to the genome of B. xylophilus (http://www.ncbi.nlm.nih.gov/genome) using SOAP aligner/SOAP2 [49], allowing for maximum mismatch of 5 bp. Gene expression levels of three PWN samples were calculated by reads per kilobase transcriptome per million mapped reads (RPKM) method [50]. DEGs were defined as those with changes of at least two-fold between samples (Bx_a vs. Bx_b, and Bx_a vs. Bx_fungus) and at FDR ď 0.001 [51,52], excluding genes where the RPKM was 0 in either sample. Cluster analysis of gene expression patterns was conducted with cluster [53] and Java Treeview [54] software.
Finally, the DEGs were analyzed for enrichment of GO and KEGG pathway terms. All DEGs were mapped to GO terms in the database (http://www.geneontology.org). Then, gene numbers in every term were calculated, and hypergeometric test was used to detect the notably enriched GO terms [55]. The calculated p-value was subject to Bonferroni correction [56]. Set the corrected p-value ď0.05 as a threshold. GO terms meeting this threshold were selected as notably enriched in DEGs. KEGG (www.genome.jp/kegg/) was used to perform metabolic pathways or signal transduction pathways enrichment analysis of DEGs. Significantly enriched pathways were identified in DEGs by comparing with the whole genome background. The method of calculation was used as that in GO analysis.

qRT-PCR Validation
To verify the RNA-seq results, nine extracted RNA samples of different treatment PWNs were used for qRT-PCR analysis. The first-strand cDNA was synthesized with Prime Script 1st strand cDNA synthesis Kit (TaKaRa, Shiga, Japan), and the cDNA samples were diluted to 20 ng/µL. Sixteen gene-specific primers (Table S2) were designed using Primer Premier 5.0 software. The Actin gene of B.xylophilus was used as internal control: Actin forward primer 5 1 -GCAACACGGAGTTCGTTGTA-3 1 and reverse primer 5 1 -GTATCGTCACCAACTGGGAT-3 1 . The qRT-PCR was performed using the ABI 7500 system (Applied Biosystems, Waltham, MA, USA), with a 20-µL final reaction volume, which contained 2 µL of template, 10 µL of SYBR Premix Ex Tap, 0.4 µL of ROX Reference Dye II, 0.4 µL of forward primer, 0.4 µL of reverse primer, and 6.8 µL of ddH 2 O. The PCR amplification was performed as follows: 40 cycles of 95˝C for 30 s, and 60˝C for 34 s. The relative expressed levels in selected genes of three PWN samples were evaluated using the relative quantification method (2´∆ ∆Ct ) [57]. These experiments were biological (n = 3) and technical (n = 3) repeated.

Observation of P. massoniana Symptoms for Different Treatments of PWNs
The experiment was conducted in a greenhouse. All P. massoniana (2 years old) were disinfected by spraying with 75% ethyl alcohol. After that, 0.5 mL of suspension (3000 nematodes) of different treatments of PWNs (Bx_a, Bx_b, and Bx_fungus) were respectively pipetted into cutting wounds (0.5 cm in length) in P. massoniana seedlings at about 20 cm above the soil level. Sterile water was used as control. The wounds on P. massoniana were sealed by Parafilm. Each treatment contained three replicates. The inoculated seedlings were placed in the greenhouse. PWD symptoms were evaluated and categorized as 0-4 [58]. The categories were as follows: 0 = all needles were green; 1 = 0%-25% needles discolored and turned yellow; 2 = 25%-50% needles turned yellow; 3 = 50%-75% needles turned yellow; and 4 = 75%-100% needles turned yellow. The infection rates and DSI were calculated with the formulae as follow: Number of infected plants Total number of plantsˆ1 00% DSI " ř Number of disease plantsˆsymptom stage Total number of plantˆhighest symptom stageˆ1 00

Conclusions
The results demonstrated that the endobacterium S. maltophilia NSPmBx03 could affect expression levels of genes related to cell wall degradation, detoxification, and reproduction genes in PWNs and enhance the virulence of nematodes. This suggests that S. maltophilia NSPmBx03 played an important role in the gene expression of PWN. These results provide insights into the regulatory mechanisms involved in the interaction between nematodes and bacteria, which will facilitate a better understanding of the role of bacteria in PWD.