Comparative analysis of different host adaptation in two rice planthopper species

Background: Host adaptation is the major determinant of insect diversification. However, knowledge of different host ranges in very close-related species is scarce. The brown planthopper ( Nilaparvata lugens , BPH) and small brown planthopper ( Laodelphax striatellus , SBPH) are the most destructive insect pests belonging to Delphacidae. These two species differ in the host range, but the underlying mechanism remains unknown. Comparative analysis of planthoppers’ responses to different diets would improve our knowledge of host adaptation regarding herbivores insects. Results: SBPH could well colonize on rice and wheat plants, while BPH survived on rice plants only. We analyzed the gene expression change of SBPH that transferred from rice plants to wheat plants in a short term (rSBPH vs tSBPH) and colonized on wheat plants for a long term (rSBPH vs wSBPH). We found that the majority of differentially expressed genes in SBPH showed similar expression change between short-term transfer and long-term colonization. Based on the comparative analysis of BPH and SBPH after transferring, genes associated with sugar transporters and heat shock proteins varied similarly. However, most of genes were reciprocally regulated between two species. The detoxification-related genes were upregulated in SBPH after transferring, but downregulated in BPH under the same condition; ribosomal-related genes were down regulated in SBPH after transferring, but upregulated in BPH under the same condition. Conclusion: The results of this study provided evidence that host plants played dominant roles in shaping the gene expression, with alike responding of most genes between short-term transfer and long-term colonization in SBPH. In contrast, the majority of genes in BPH were reciprocally regulated in comparison to those of SBPH, and unfitness of BPH on wheat plants might be determined within 24 hours after transferring. This study deepens our understanding of different host ranges regarding two planthopper species, and may

3 provide a potential strategy for pest management.

Background
Herbivorous insects and their host plants have been engaged in a tight co-evolutionary arm race, leading to specialization of most insects restricted on a narrow host, whereas other insects can be adapted to a wider host range [1]. In general, successful host adaptation involves several essential traits of herbivorous insects, including sensing toward right plants, acquiring nutrients while avoiding intoxication, and overcoming plant defenses [2]. Changes in these essential features have been reported to cause failure of colonization or even new host specialization [3][4][5][6]. To date, studies of host adaptation have mainly focused on a specific insect that feeds on different diets [7]. The research findings showed that genes in association with detoxification, digestion, and transportation were significantly influenced after host transfer [7]. However, knowledge of the disparate host adaptation in different insect species was scarce, and most studies mainly focused on detoxification differences across species in response to plant secondary metabolites [8][9][10]. Comparative transcriptomic analysis of closely related taxa provided an ideal approach for unveiling the distinctions in host adaptation across herbivorous species.
Although they use rice plants as the primary food source, their host ranges differ. BPH is the monophagous insect pest restricted to rice plants, while SBPH is the oligophagous one that can feed on rice, wheat, and other gramineous plants [11]. Factors that are potentially implicated in different host range of planthoppers have been investigated for decades. Denno & Roderick (1990) presumed that it might be the chemical barrier instead of the physical one that prevented BPH feeding on non-adaptive plants, as the insects 4 could still settle and insert their stylets into plant tissues [12]. Liu et al. (2011) found that feeding on different plants greatly influenced the activity of detoxification enzymes in SBPH, but had little impact on BPH [13], which was in agreement with expanded gene families of cytochrome P450s and ABC transporters in the SBPH genome [14]. Additionally, the presence of endosymbionts, which provided essential amino acids, was also critical for planthoppers' colonization [15]. The sap compositions between rice and wheat plants were different, especially the amino acid concentration [16,17]. It was interesting to investigate the influence of nutrition supplements on host adaptation. So far, the genomic information of both planthoppers and their host plants (Oryza sativa and Triticum aestivu) became available [18], which provided an opportunity to study the mechanism of different host adaptation in these non-model insects.
The midgut is the main tissue where digestion and detoxification occurred, which exerts an important influence on insect feeding. Through "Omics" technologies, changes of gut physiology in response to different host plants have been investigated in several herbivorous species, especially the Lepidoptera insects [7]. The expression patterns of gut-associated transcripts, including digestive and detoxifying enzymes, transporters, immune and peritrophic genes were reported to be substantially altered when larvae were exposed to novel diets [19][20][21]. For generalist larvae, there would be profound transcriptional variations in guts to overcome detrimental effects of plant secondary metabolites; nevertheless, for specialist larvae, a specific detoxification system avoided activation of general stress responses and minimized the metabolic costs [19,22], which might restrain a specialist from adapting to a novel phytotoxin [23]. In our previous work, genes associated with wheat adaptation were demonstrated to be dramatically changed in the gut compared with other tissues, especially the CYP4DE1 [24]. It was intriguing to unveil the role of the midgut in planthopper adapting to different hosts.

5
In this study, we compared the performance of BPH and SBPH on different hosts, and profiled the global gene expression pattern of SBPH colonizing rice plants (rSBPH), colonizing wheat plants (wSBPH), and transferring from rice plants to wheat plants (tSBPH). Based on the gene expression changes, the differentially expressed genes (DEGs) of SBPH in response to different diets were grouped into four types, the majority of which were found to be changed in the same direction between short-term transfer and longterm colonization. In addition, the transcriptomes of BPH that colonizing rice plants (rBPH) and transferring to wheat plants (tBPH) were also investigated. By comparative analysis, the feasible mechanism of different host adaptation concerning BPH and SBPH was illustrated.

Performance of planthoppers on rice and wheat plants
Rice plants, other than wheat plants, can be well colonized by BPH, with only ricecolonized BPH strains (rBPH) generated in laboratory conditions. In contrast, SBPH could successfully colonize on both rice and wheat plants, and two SBPH strains (rSBPH and wSBPH) were maintained in our laboratory for more than 30 generations. According to survival analysis, more than 90% of rBPH survived on rice plants for 12 days, which was significantly higher than that on wheat plants (Fig. 1A). Apart from that, rBPH survived longer on wheat plants (LT 50 =6.1 day) than that provided with water only (LT 50 =3.3 day).
These results indicated that BPH could uptake wheat sap, and survived on wheat plants for a short time, but not for a long time. For SBPH, both rSBPH and wSBPH could survive on rice and wheat plants successfully (Fig. 1B) in accordance with previous reports [25].

Overview of RNA sequencing data
To explore the mechanism underlying different performances of two planthoppers, rice-6 colonized planthoppers were transferred to wheat plants for 24 hours (short-term transfer), or reared on wheat plants for over 30 generations (long-term colonization).
Then, guts of planthoppers that colonized on rice plants (rSBPH and rBPH) or wheat plants (wSBPH), or transfer from rice plants to wheat plants (tSBPH and tBPH) were isolated and underwent Illumina HiSeq2500 sequencing. A total of 15 libraries were generated, with clean reads exceeding 45 million in each library. The clean reads were mapped to their reference genomes [14,15], respectively. For BPH, there were 75% to 83% clean reads mapped to the reference genome; for SBPH, the percentage ranged from 60% to 66%.
Saturation analysis showed that the number of detected genes decreased as that of reads increased, and library capacity reached saturation when the number of sequence reads approached 20.0 million (Fig. S1, Supporting information). Furthermore, principal component analysis (PCA) demonstrated that the three replicates of each treatment were well clustered (Fig. 2). Compared with rSBPH, close-related expression patterns were found in wSBPH and tSBPH, indicating that host plants have a non-negligible influence on gene expression (Fig. 2).

Analysis of differentially expressed genes (DEGs)
Gene expression changes were analyzed via comparing rice-colony planthopper to transfer planthopper (tSBPH_vs_rSBPH and tBPH_vs_rBPH) and rice-colony planthopper to wheatcolony planthopper (wSBPH_vs_rSBPH) using a threshold of >2 fold change and an FDR adjusted p-value <0.05. For rice-colony planthoppers transferring to wheat plants, a total of 2,877 and 2,638 genes were differentially expressed in SBPH and BPH, respectively ( Fig. 3). There were 2,373 genes upregulated and 505 genes downregulated when rSBPH was transferred to wheat hosts (Table S1). Among them, genes participating in signal transduction were remarkably upregulated. CYP4DE1, which mediated wheat adaptation and ethiprole tolerance in SBPH [24] was also significantly induced after transferring (Fig.  4). On the contrary, 71 genes related to ribosomal proteins and 48 genes in relation to oxidative phosphorylation were significantly downregulated, indicating a decreased protein production and energy metabolism (Fig. 5). In BPH, the amount of genes downregulated (2,171 genes) exceeded that of genes upregulated (467 genes) (Table S2).
Contrary to SBPH, the majority of ribosomal proteins were upregulated in BPH (Fig. 5).
Cuticular proteins, which formed the insect cuticle and participated in insect molting, were dramatically upregulated after BPH transferred to wheat (Fig. 6, Table S2).
In the comparison of rSBPH and wSBPH, a total of 2,516 DEGs were identified (Fig. 2).
Strikingly, 90.9% of DEGs (2,287 genes) showed higher expression in wSBPH than those in rSBPH (Table S3), with peroxisomal biogenesis factor, nucleotide exchange factor, peptide transporter, and CYP6FK1 exhibiting most dramatic changes. As many as 37 genes participating in signal transduction were significantly enriched, similar to the patterns of rSBPH that transferred to wheat hosts. Among the 228 downregulated genes, there were the most dramatic changes in zinc metalloproteinase, UGT, and alpha-glucosidase. Other downregulated genes participating in chitin metabolism, carbohydrate derivative metabolism, starch and sucrose metabolism, oxidative phosphorylation were significantly enriched (Table S3).

Classification of SBPH genes associated with diet changes
Based on the gene expression changes in response to different diets, DEGs of SBPH were classified into four types: I) genes changed in the same direction between short-term transfer and long-term colonization, II) genes changed in the opposite direction between short-term transfer and long-term colonization, III) genes changed in short-term transfer, 8 but not long-term colonization, and IV) genes changed in long-term colonization, but not short-term transfer.
The number of type I responsive genes (1,558 genes) ranked the first (Table S4).
Enrichment analysis showed that genes participating in signal transductions and immune systems were significantly overrepresented. Nevertheless, only 22 genes belonged to type II response (Table S4). Four genes were downregulated after transferring, but dramatically upregulated during colonization, while other 18 genes showed reciprocal expression patterns. There were 1,297 genes belonging to type III response (Table S4). Enrichment analysis showed that ribosome pathway, oxidative phosphorylation pathway, and retrograde endocannabinoid signaling pathway were significantly overrepresented. It was noteworthy that the expression level of ribosome proteins was initially suppressed, but recovered when SBPH long-term colonized on wheat plants. A total of 936 genes belonged to type IV response (Table S4). Nevertheless, we failed to find GO terms or KEGG terms that were significantly enriched. Genes such as integrin alpha-PS4-like, integumentary mucin, and proliferation-associated protein showed a higher expression level in wSBPH.

Comparative genomics of genes in response to host transfer
To comprehend the different host adaptation across species, 6139 gene families with only one ortholog in BPH and SBPH were selected and compared. A total of 1,995 gene families showed altered expression in at least one planthopper species after transferring (Table   S5), among which 370 genes were responsive to host transfer in both planthoppers.
Interestingly, only 22 genes altered in the same direction, including heat shock protein, prophenoloxidase activating factor, MAP kinase-interacting serine/threonine-protein kinase, and small nuclear ribonucleoprotein. Nonetheless, other 348 genes showed reciprocal expression patterns between BPH and SBPH. Among them, 172 type I responsive genes were upregulated in SBPH after transferring, but downregulated in BPH; 25 9 ribosomal proteins (belong to type III response) were downregulated in SBPH after transferring, but upregulated in BPH.

qPCR validation
To confirm the validity of transcriptomic data, 10 genes from SBPH and BPH were selected for qPCR analysis. Nine SBPH genes (Fig. S2) and 10 BPH genes (Fig. S3) showed concordant direction of change between qPCR and transcriptomic results. Moreover, the Pearson correlation coefficient between qPCR and transcriptomic in each comparison ranged from 0.7847 to 0.8508 (Fig. 8), indicating the accuracy of DGE transcriptomic results. When rBPH and rSBPH transferred to wheat plants, the heat shock proteins were upregulated markedly. Other genes, such as chemosensory protein 12, nucleotide exchange factor SIL1, CYP4DE1, trehalose transporter Tret1, cuticle protein 16.5, and serine proteinase stubble showed altered expression in response to host transfer as well.

Discussion
In the present study, the adaptive differences of two planthoppers on wheat plants were investigated using comparative transcriptomic analysis. We identified a total of 2877, 2638 and 2287 DEGs when comparing tSBPH to rSBPH, tBPH to rBPH, and wSBPH to rSBPH, respectively. After transferring from rice plants to wheat plants, a large portion of DEGs in BPH and SBPH were found to be reciprocally changed, which might account for their adaptive differences. Our study presents comprehensive and accurate gene expression profiles for herbivores adapting to different hosts.
The wSBPH and rSBPH used in this study have been reared under laboratory conditions for more than 30 generations, while the tSBPH was collected by transferring rSBPH to wheat plants within 24 hours. Surprisingly, the expression pattern of tSBPH was bound up with wSBPH instead of rSBPH (Fig. 2), indicating the leading role of the host plant in shaping gene expression. Also, transferring of rSBPH to wheat plants gave rise to broad transcriptional readjustments, and the majority of DEGs were changed in the same direction between short-term transfer and long-term colonization. Because of a wheat-rice rotation cultivation system, SBPH shifted between rice and wheat plants every year [26].
Frequency in host transferring might help SBPH sophisticatedly respond to new hosts, generating successful wheat adaptation. Similar phenomena were also reported in Tetranychus urticae, which rapidly adapted to pre-exposed hosts [27]. In contrast, reciprocal expression patterns were revealed in 94.1% of homologous genes in BPH and SBPH when planthoppers transferred to wheat plants (Table S5). Thus, we speculated that the unfitness of BPH on wheat plants might be determined within 24 hours after transferring, with the majority of genes changed in the "wrong" direction.
Genes involved in detoxification were significantly influenced when planthoppers transferred to wheat plants (Fig. 4). In previous work, CYP4DE1 was found to be responsible for wheat adaptation, and knockdown of CYP4DE1 significantly lowered the intestinal cell viability of SBPH reared on wheat plants [24]. In this study, we found CYP4DE1, as well as other P450 genes was significantly upregulated after rSBPH transferred to wheat plants, indicating that induction of P450s might be critical. Similarly, 9 ABC transporters, which transport a large diversity of substrates across lipid membranes and out of cells, were also significantly upregulated in response to wheat plants.
Involvement of ABC transporters in xenobiotic transport and insecticide resistance have been well documented [28], and these proteins were also reported to participate in host adaptation in Manduca sexta [29], Aphis nerii [30], and Chrysomela populi [31]. The increased expression of ABC transporters signified that SBPH might accelerate secretion of metabolites when feeding on wheat plants. Comparative analysis showed that the majority of differentially expressed detoxifying genes were upregulated when SBPH transferred to wheat plants. However, these genes showed reciprocal expression in BPH (Fig. 4). As BPH also ingested sap from wheat plants, it was confused that these detoxifying genes were not induced by wheat secondary metabolites. We presumed that this might be associated with resource-based metabolic trade-off [32]. When coping with such adverse environment, it might be more critical for BPH to increase basic metabolism than the detoxification process. Further studies are needed to determine the detoxification variation between BPH and SBPH on wheat plants.
Large numbers of ribosomal proteins were differentially expressed when planthoppers transferred to wheat plants. However, the majority of these genes showed reciprocal expression change between SBPH and BPH (Fig. 5). Mostly associated with their standard roles in protein translation, ribosomal proteins are generally deemed to be stably expressed and used as housekeeping genes [33]. However, recent studies demonstrated that ribosomal proteins were differentially expressed in response to host transfers in Bemisia tabaci [34], Helicoverpa armigera [35], Polygonia c-album [36], and Cryptolaemus montrouzieri [37]. In an insect-plant model, regulations of ribosomal proteins are perceived to counteract the ribosome inactivating proteins (RIPs), which are produced by host plants and inhibit protein synthesis in insects [38]. In this study, the expression level of ribosomal proteins was significantly downregulated when rSBPH transferred to wheat, but upregulated in rBPH under the same condition. We presumed that the reciprocal expression changes of ribosomal proteins might be associated with different strategies of two planthoppers in dealing with RIPs in the food source. Additionally, ribosomal proteins played fundamental roles in the cellular process of translation [33]. The large-scale change of ribosomal proteins inevitably influenced the regulation of other genes, which might partially explain the distinct gene expression variation in two planthopper species.
Genes associated with amide/peptide biosynthetic process, amino sugar metabolic process, and aminoglycan metabolic process were significantly enriched in both SBPH and 12 BPH after host transferring. Planthoppers are piercing-sucking herbivores that depend on phloem sap containing high concentration of sugar, amino acid and inorganic ions as their food sources [15,39]. Previous work demonstrated that the sucrose concentration between rice phloem sap and wheat phloem sap was very similar, but the latter had higher concentration of total amino acids compared with the former [16,17]. In addition, the predominance of amides between two phloem sap was also very different [16,17]. In previous work, we found that the BPH genome lacks the ability to carry out de novo synthesis of some essential amino acids, which requires an additional supply from the yeast-like endosymbionts (YLS) or plant phloem sap [15]. Our presumption is that the alteration in phloem sap composition inevitably influenced the nutrition supplements, which might bring about the change of the basic metabolism (especially the amino acid metabolic process). This hypothesis has been partially verified by planthoppers reared on resistant rice, where the physiological adaptation of insects to novel rice variety was closely linked to a change in amino acid metabolism [40].
Masses of sugar transporters were significantly downregulated when planthoppers transferred to/colonized on wheat plants (Fig. 6). Mediating the movement of sugars into and out of cells in a diverse array of organisms, sugar transporters are vital for utilization of ingested sugars as a nutritional resource and maintenance of osmotic balance in insects [41]. We found that the majority of downregulated sugar transporters belonged to facilitated trehalose transporter. Trehalose is the major hemolymph sugar in most insects [42]. Besides, its concentration in hemolymph was determined by the balance between discharge from the fat body and uptake by other tissues via facilitated trehalose transporter [43]. In Anopheles gambiae, suppression of trehalose transporter significantly diminished the hemolymph trehalose concentration, making the insect more sensitive to stress environment [44]; similarly, the trehalose transporter from Polypedilum vanderplanki enhanced desiccation tolerance of cell lines [45]. Involvement of trehalose transporter in host adaptation was only reported in a few insects [19]. However, increased evidence demonstrated that trehalose, the regulatory product of trehalose transporter, played key roles in insects feeding on different diets. For example, trehalose concentration in haemolymph affected the food preference and carbohydrate intake in Heliothis zea and Manduca sexta [46,47]; trehalose homeostasis was critical for Drosophila adapting to unlike dietary conditions [48]. In this study, the expression levels of sugar transporters were suppressed in SBPH and BPH reared on wheat plants.
Accordingly, we inferred that the two planthoppers might use similar strategy to keep sugar homeostasis, but the underlying mechanism still remained elusive.
Heat shock proteins (HSPs) were prominently unregulated when two planthopper species transferred to/colonized on wheat plants (Fig. 6). HSPs are well-known stress proteins that respond to an array of stresses, including thermal hardening, oxidative hardening, chemical pesticides, and desiccation [49,50]. Induced expression of these proteins, as a significant molecular chaperone, prevented irreversible denaturation of substrate proteins, and promoted the coping capacity of insects in the face of stress environment [49]. Previously, the influence of diet quality on HSP expression has been unveiled in Drosophila melanogaster, and diet-induced HSPs increased heat and desiccation tolerance of flies [51]. Feeding on wheat was a biotic stress for BPH and SBPH [52], which might result in increased HSP expression. Interestingly, some SBPH overwintered in the wheat field after the rice-harvesting season [53]. Thus, further investigation concerning the influence of wheat-induced HSPs on SBPH cold tolerance is essential.

Conclusion
Overall, SBPH sophisticatedly coped with wheat hosts, with the majority of DEGs responded similarly between short-term transfer and long-term colonization. Compared with SBPH, BPH showed distinct gene expression changes after transferring to wheat plants. The reciprocal gene expression alteration between BPH and SBPH might account for their adaptive differences on wheat plants. Specifically, genes associated with sugar transporters and heat shock proteins showed similar expression trends between BPH and SBPH; other genes in association with detoxification, ribosomal proteins, and amino acid metabolism were reciprocally regulated between two planthoppers. Our work boosts our understanding of herbivore adapting to different food sources.

Insect strains
SBPHs and BPH populations used in this study were originally collected from a rice field at Huajiachi Campus, Zhejiang University, Hangzhou, China. The rSBPHs and wSBPHs were generated by rearing SBPHs on fresh rice (Xiushui 134) and wheat (Luyuan 502) plants for more than 30 generations. The rBPHs have been reared on rice (Xiushui) 134 plants for more than 40 generations. All insects were reared in large cage, and more than 300 individuals were used for propagation in each generation.

Survival analysis
Survival analysis of rSBPH on different hosts has been done previously [25]. However, the performance of wSBPH, as well as rBPH, on wheat plants was still unknown. For BPH, 3 rd instar nymphs (rBPH) were reared on 4-5-leaf stage rice plants or transferred to 4-5-leaf stage wheat plants, and the rBPH that provided with water only were used as a control.

Transcriptomic sequencing
To compare the planthoppers' responses to different dies, transcriptomic sequencing of rice-colonized BPH and SBPH to rice and wheat plants, as well as wheat-colonized SBPH to wheat plants were performed. For rSBPH, wSBPH, and rBPH populations, newly emerged 5 th instar nymphs were collected and continually reared on their original hosts for 24 h. After determination of RNA integrity and quantity, poly (A) + RNA was purified from 20 μg pooled total RNA using oligo (dT) magnetic beads. Fragmentation was conducted in the presence of divalent cations at 94 °C for 5 min and then reverse transcribed to doublestranded cDNA (dscDNA) by N6 random primers. After end-repairing and adaptor ligation, the products were PCR-amplified and purified using a QIAquick PCR purification kit (Qiagen, Hilden, Germany) to create a cDNA library. The library was sequenced on a BGISEQ-500 platform (Beijing Genomics Institute, Beijing, China) and the low quality reads were filtered using internal software. The clean reads from each cDNA library were aligned to reference genome sequences of Nilaparvata lugens (ftp://parrot.genomics.cn/gigadb/pub/10.5524/100001_101000/100139/) and Laodelphax striatellus (ftp://parrot.genomics.cn/gigadb/pub/10.5524/100001_101000/100361/) using Hierarchical Indexing for Spliced Alignment of Transcripts (HISAT). The clean data of rSBPH, rBPH, wSBPH, tSBPH, and tBPH have been submitted to the database of the NCBI Sequence Read Archive (http://trace.ncbi.nlm.nih.gov/Traces/sra).
For gene expression quantification, the mapped reads were calculated as FPKM using RESM software [54]. The differentially expressed genes were defined by the DEGseq [55,56] using the criteria of log2-ratio >1and adjusted p value< 0.05. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were based on the Gene Ontology Database (http://www.geneontology.org/) and KEGG pathway database (http://www.genome.jp/kegg/), respectively.

Gene family identification
Previous work demonstrated that genes associated with detoxification, digestion, and transportation played critical roles in host transfer [7]. To comprehensive analysis these genes, gene families of interest were initially identified based on the gene annotations. To avoid possibly incomplete annotations, genes from Drosophilamelanogaster, Acyrthosi phonpisum, and Bombyx mori were used as queries to search against genomic databases using BLAST (ftp://ftp.ncbi.nlm.nih.gov/blast) with a significance cut-off of E-value < 10 − 5 .
Gene annotations were further validated by aligning to the non-redundant (nr) National Center for Biotechnology Information (NCBI) protein database.

Comparative genomic analysis
The protein families of SBPH and BPH were analyzed using the ORTHOMCL software 2.0.3 (http://orthomcl.org/common/downloads/software/v2.0/). Briefly, the amino acid sequences derived from the coding sequences in genome were selected, and homologous pairs of sequences were found using the all-vs-all BLASTp algorithm with an E-value < 1e-5.
OrthoMCL then converted the BLASTp result into a normalized similarity matrix that was analyzed by a Markov chain clustering (MCL) for clustering according to Li et al. recommendation [57]. An inflation factor of 1.5 was used to regulate cluster tightness.

Real-time quantitative PCR (qPCR) analysis
To confirm the validity of transcriptomic data, total RNA was extracted as described above, and 1 μg RNA was reverse-transcribed in a 20-μl reaction system using HiScript II Q RT SuperMix (Vazyme, Nanjing, China). The primers used in qPCR were designed based on Primer Premier 6.0 (Table S6). The housekeeping genes for β-actin and GAPDH were used as internal control. Relative expression pattern was quantified by an ABI 7500 Real-Time PCR System (Applied Biosystems, Carlsbad, CA) using the SYBR Green Supermix Kit (Yeasen, Shanghai, China). The first-strand cDNA and a no-reverse-transcription control were used as templates under the following reaction program: denaturation for 5 min at 95°C, followed by 40 cycles at 95°C for 10 s and 60°C for 30 s. A relative quantitative method (2 -∆∆Ct ) was used to evaluate quantitative variation [58]. Three independent biological replicates were performed.  Classification of genes in SBPH based on the expression pattern. The expression pattern of identified differentially expressed genes in SBPH was classified into four types: Types I, genes changed in the same direction between short-term transfer and long-term colonization; Types II, genes changed in the opposite direction between short-term transfer and long-term colonization; Types III, genes changed in short-term transfer, but not long-term colonization; Types IV, genes changed in long-term colonization, but not short-term transfer.   Table S6.doc   Table S4.xlsx Figure S1.tif Table S2.xlsx   Table S5.xlsx   Table S3.xlsx   Table S1.xlsx