Gene expression and plant hormone levels in two contrasting rice genotypes responding to brown planthopper infestation

The brown planthopper (BPH; Nilaparvata lugens Stål) is a destructive piercing-sucking insect pest of rice. The plant hormones salicylic acid (SA) and jasmonic acid (JA) play important roles in plant–pest interactions. Many isolated rice genes that modulate BPH resistance are involved in the metabolism or signaling pathways of SA, JA and ethylene. ‘Rathu Heenati’ (RH) is a rice cultivar with a high-level, broad-spectrum resistance to all BPH biotypes. Here, RH was used as the research material, while a BPH-susceptible rice cultivar ‘Taichung Native 1’ (TN1) was the control. A cDNA microarray analysis illuminated the resistance response at the genome level of RH under BPH infestation. The levels of SA and JA in RH and TN1 seedlings after BPH infestation were also determined. The expression pattern clustering indicated that 1467 differential probe sets may be associated with constitutive resistance and 67 with the BPH infestation-responsive resistance of RH. A Venn diagram analysis revealed 192 RH-specific and BPH-inducible probe sets. Finally, 23 BPH resistance-related gene candidates were selected based on the expression pattern clustering and Venn diagram analysis. In RH, the SA content significantly increased and the JA content significantly decreased after BPH infestation, with the former occurring prior to the latter. In RH, the differential genes in the SA pathway were synthesis-related and were up-regulated after BPH infestation. The differential genes in the JA pathway were also up-regulated. They were jasmonate ZIM-domain transcription factors, which are important negative regulators of the JA pathway. Comparatively, genes involved in the ET pathway were less affected by a BPH infestation in RH. DNA sequence analysis revealed that most BPH infestation-inducible genes may be regulated by the genetic background in a trans-acting manner, instead of by their promoters. We profiled the analysis of the global gene expression in RH and TN1 under BPH infestation, together with changes in the SA and JA levels. SA plays a leading role in the resistance response of rice to BPH. Our results will aid in understanding the molecular basis of RH’s BPH resistance and facilitate the identification of new resistance-related genes for breeding BPH-resistant rice varieties.


Background
As a staple food for more than half of the world's population, rice (Oryza sativa) is the most important crop in Asia [1], and insect pests are a main factor affecting rice production. Rice brown planthopper (BPH; Nilaparvata lugens Stål), which sucks sap from the plant's phloem, is highly destructive [2]. In addition to direct damage, BPH, as the intermediate vector, can infect rice with pathogens, such as grass stunt virus and ragged stunt virus, leading to further yield losses [3,4].
Since the 1960s, rice breeders have endeavored to identify BPH-resistant germplasms and develop BPH-resistant rice varieties. At least 30 BPH-resistance quantitative trait loci (QTLs) have been identified in rice. Among these, the chromosomal locations of 21 have been determined, and 12 have been fine-mapped. Most of these BPH-resistance genes were identified from wild rice, few from indica rice and none from japonica rice [5][6][7].
Four major BPH-resistance QTLs, Bph14, Bph3/ Bph17, BPH26/BPH2 and Bph29, were cloned using map-based techniques [8][9][10][11]. Bph14 and BPH26 encode coiled-coil, nucleotide-binding and leucine-rich repeat proteins that resemble the R genes of the nucleotidebinding-leucine-rich repeat family, which mediate plant resistance to pathogens [8,10]. Bph14 and BPH26 are presumed to induce effector-triggered immunity, like the R genes, because they share similar conserved protein domains. The expression of Bph14 is induced by BPH infestation, and it then activates the salicylic acid (SA) signaling pathway, induces calluses deposition in phloem cells and enhances trypsin inhibitor production [8]. BPH26, like Bph14, also mediates sucking inhibition in the phloem sieve element. The Bph3 locus is a cluster of three genes encoding plasma membrane-localized lectin receptor kinases (OsLecRK1-OsLecRK3). Bph3 was proposed to play a critical role in priming the patterntriggered immunity response to BPH infestation by perceiving herbivore-associated or damage-associated molecular patterns due to the functions of recently discovered lectin receptor kinases [9]. Bph29 is a recessive resistance gene that encodes a B3 DNA-binding domain that contains a protein that has a DNA mutation in the B3 domain. Bph29 was proposed to have lost the function of the dominant allele, which is required for the settling of insects, and thus confers an antixenosis resistance in conjunction with an anther recessive locus [11]. The cloning of major BPH-resistance QTLs facilitated the understanding of the molecular mechanisms of resistant rice to BPH, although the biological mechanisms are not well understood.
In addition to the major BPH-resistance QTLs, many other rice genes that modulate BPH resistance (resistance-related genes), such as OsHI-LOX [12], Bphi008a [13], OsERF3 [14], OsPLD α4 and α5 [15], OsHPL3 [16], OsACS2 [17], OsAOC [18], Osr9-LOX1 [19], and OsJMT1 [20], have recently been isolated. Most of them were identified using a reverse genetics strategy, and all of them are involved in the metabolism or signaling pathways of the plant hormones SA, jasmonic acid (JA) and ethylene (ET), indicating that these hormones play important roles in rice defense responses to BPH. SA is generally considered to positively regulate plant defense responses against pathogens and piercing-sucking insects, which cause minimal levels of physical injury to the host plants. The major BPH-resistance QTLs Bph14 and BPH29 enhance the expression of the SA synthesisrelated genes and the homolog of Arabidopsis nonexpressor of pathogenesis-related genes 1, a key regulator of SA-dependent systemic acquired resistance after BPH infestation. They also suppress the expression levels of the JA synthesis-related genes and the ethylene signaling pathway receptor gene ethylene insensitive 2 [8,11].
JA is associated with wounding responses and positively regulates the plant defense response against chewing insects that cause extensive damage to the host. The interactions between SA and JA are commonly antagonistic, with SA having a suppression effect on JA accumulation and signaling. ET is considered to finetune the JA-induced responses, and JA and ET are synergistic. Many rice genes involved in JA metabolism, including OsHI-LOX [12], OsPLDa4 and -a5 [13], OsHPL3 [16], AOC [18], Osr9-LOX1 [19] and OsJMT1 [20], can modulate BPH resistance. JA and its metabolites have diverse functions in BPH resistance. The ET pathway is generally considered to negatively modulate the BPH resistance of rice [21]. The ET-responsive gene OsERF3 encodes a nucleus-localized protein, which is rapidly up-regulated in response to infestations of the rice striped stem borer (SSB; Chilo suppressalis Walker). Transgenic rice overexpressing OsERF3 exhibit improved SSB resistance but are more susceptible to BPH. OsACS2 encodes a rice 1-aminocyclopropane-1-carboxylic acid synthase gene, which is inducible by both SSB and BPH infestations. Suppression of OsACS2 expression in rice reduced the elicited ethylene emissions and SSB resistance, but enhanced BPH resistance [17].
'Rathu Heenati' (RH), an indica rice cultivar from Sri Lanka, has a high, durable resistance to all of the BPH biotypes. RH is regularly used as a positive control when identifying BPH resistant rice varieties and is an important BPH-resistance donor for rice breeding. In this study, a cDNA microarray analysis was performed to profile the whole genome expression of RH, and a BPHsusceptible rice cultivar 'Taichung Native 1' (TN1) was used as the control. All of the rice plants were grown under three treatment conditions, natural growth (untreated), needle puncturing (mocking a simple mechanical wound) and BPH infestation, and two sampling time points, 6 and 24 h after treatment, were used. The changes in the levels of SA and JA in RH and TN1 seedlings were also determined at 6, 24 and 48 h after BPH infestation. The ET content is difficult to measure because of its volatility and was, therefore, not determined in this study. This study revealed the resistance responses of RH to BPH at the genomic transcription level and through changes in the JA and SA contents. The results of this study aid in understanding the biological foundation of the BPH resistance of RH and in identifying new BPH resistance-related genes for rice breeding.

Results
Genome-wide comparison of differentially-expressed probe sets between RH and TN1 under BPH infestation The experimental design included two rice varieties (RH and TN1), three treatments (untreated control, needle puncturing and BPH infestation) and two sampling time points (6 and 24 h after BPH infestation). There were 12 samples in total, with 3 biological replicates per sample. The abbreviations R6C, R24C, R6N, R24N, R6P, R24P, T6C, T24C, T6N, T24N, T6P and T24P were used to represent these 12 samples, where the first letter, R or T, indicates the rice varieties RH or TN1, respectively; the Arabic number in the middle indicates the 6 or 24 h post-treatment collection time points; and the last letter C, N or P, indicates one of the three different treatments, untreated control, needle puncturing or BPH infestation, respectively. The Affymetrix GeneChip Rice Genome Array, containing 57,194 probe sets, was used for the microarray analysis, and the results of all 36 chips were submitted to the NCBI website (http:// www.ncbi.nlm.nih.gov/geo/; accession number: GSE74 106; Additional file 1).
There were 8019 differentially-expressed probe sets [fold change (FC) ≥ 2] detected in this study. These included 3682 differential probe sets between different treatments, 3339 between different varieties (RH and TN1), and 5757 between different sampling times, with the two non-differentiating variables being the same (Additional file 2). To compare the gene expressional differences between RH and TN1, all 8019 differential probe sets were clustered into different groups based on their expression patterns using Multi-Experiment Viewer ( Fig. 1). Significantly, 785 differential probe sets in Group k had low expression levels in TN1 and high expression levels in RH under all of the conditions; in contrast, 682 differential probe sets in Group p had low  levels in RH and high expression levels in TN1 under all of the conditions. Therefore, Groups k and p roughly represent the genetic background differences between RH and TN1. Genes related to the constitutive resistance of RH should be included in these two groups. Groups a and m represent probe sets with BPH infestation-specific down-(6 differential probe sets) and up-regulated (61 differential probe sets) expression levels in RH, respectively. Genes in these two probe sets may be related to the BPH infestation-responsive resistance of RH. Differential probe sets in Groups l, q, r and s showed regular fluctuations associated with sampling time, indicating that their expression is mainly controlled by sampling time (circadian rhythm or other environmental factors). These four types of expression patterns included a total of 3793 differential probe sets, accounting for almost a half of the 8019 differential probe sets.

Best reference genes for BPH resistance research
Reference genes are key factors that ensure the accuracy of a gene expressional analysis under specific conditions. It was very important to select a stably expressed gene under BPH infestation as the reference for quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis [22][23][24]. Figure 2 shows eight frequentlyused reference genes, actin-1 (Os03g50890), LSD1 (Os12 g41700), GAPDH (Os02g38920), SDHA (Os07g04240), TPB (Os02g45410), RPS27α (Os01g22490), HSP (Os03g 31300) and Ubiquitin (Os03g03920), and their respective signal values extracted from our microarray data under different treatments. Actin-1 and RPS27α were downregulated in both TN1 and RH within 24 h after BPH infestation; GAPDH showed an obvious expressional difference between RH and TN1; and the expression level of LSD1 was stable but relatively low. Among the remaining four references, TBP and Ubiquitin showed better overall consistency compared with HSP and SDHA. Therefore, TBP and Ubiquitin were determined to be suitable reference genes for BPH resistance research. Ubiquitin was used as the reference gene for all of the qRT-PCR analysis in this study.
To find additional stable reference genes for the gene expression analysis under BPH-infestation conditions, 373 probe sets with total variance < 5% in the microarray data were screened. Two probe sets with low expressional variances and appropriate expressional levels were selected, Os.1322.1.S1_at (Os05g23860) and Os.145.1.S1_ a_at (Os02g56000) (Fig. 2), and the genes corresponding to these two probe sets were annotated to be rab GDP dissociation inhibitor alpha and 26S protease regulatory subunit 6A, respectively, in the MSU Rice Genome Annotation Project Release 7 (http://rice.plantbiology.msu.edu/index.shtml). They were also stably expressed during the entire life cycles of the indica rice varieties 'Minghui 63' and 'Zhenshan 97' [23].

Different defense responses, at the genome level, between RH and TN1 to BPH infestation
At 6 h after needle puncturing, there were 92 differential probe sets in RH compared with the samples under natural growth, while there were 211 in TN1, indicating TN1's more sensitive reaction to simple mechanical damage than RH. However, at 24 h after needle puncturing, fewer differential probe sets were detected in both RH (27) and TN1 (38) ( Table 1), indicating that the Fig. 2 Signal values of actin-1, LSD1, GAPDH, SDHA, TPB, RPS27α, HSP, Ubiquitin, Os.1322.1.S1_at and Os.145.1.S1_a_at in the microarray data. Among the eight frequently-used reference genes, TBP and Ubiquitin were more stably expressed than the others under BPH infestation. Os.1322.1.S1_at (Os05g23860) and Os.145.1.S1_a_at (Os02g56000) are two newly identified references with stable expressions under BPH infestation. Error bars indicate standard deviations of three biological replicates reactions to simple mechanical damage had almost recovered by 24 h.
At 6 h after BPH infestation, there were more differential probe sets in RH (209) than in TN1 (173). The number of differential probe sets with high FCs (>5) was 35 in RH and 12 in TN1 (Table 1). Although TN1 was relatively more sensitive to simple mechanical damage, more differential probe sets were detected in RH at 6 h (early stage) after BPH infestation, implying that the defense responses to BPH infestation in RH was more intensive than in TN1. At 24 h after BPH infestation, more differential probe sets (614) were detected in RH compared with at 6 h, and the number of probe sets with FCs > 5 also increased from 35 to 73, indicating that the signals of BPH infestation were progressively transmitted to the related downstream genes. However, 24 h after BPH infestation, the number of differential probe sets in TN1 dramatically increased to 3356, and the number with FCs > 5 also increased dramatically from 12 to 349 (Table 1). Thus, more differential probe sets were detected in RH at 6 h after BPH infestation; however, many more differential probe sets were detected in TN1 at 24 h after BPH infestation. Many of the differential probe sets detected in TN1 at 24 h after BPH infestation may have represented the responses to wounding and various physiological stresses resulting from the serious damage caused by the BPH, instead of a resistance response. In RH, the majority of differential probe sets were up-regulated at either 6 or 24 h after BPH infestation, indicating that the inducible expression of resistance-related genes may be the main molecular basis of the defense responses in RH to BPH.
The Venn diagram analysis of differential probe sets is shown as Fig. 3. Among the differential probe sets of RH at 6 h after BPH infestation, 53 were also present at 6 h after needle puncturing. Among the differential probe sets of TN1 at 6 h after BPH infestation, 60 were also present at 6 h after needle puncturing. At 24 h after the treatments, most of the differential probe sets under needle puncturing, in either RH or TN1, were also present under BPH infestation. By excluding the differential probe sets of RH and TN1 under needle puncturing and those of TN1 under BPH infestation, a total of 192 RH-specific and BPH-inducible probe sets were determined. These 192 probe sets may represent the basis of the inducible defense in RH to BPH (Additional file 3).

Gene Ontology (GO) analysis
The differential probe sets of RH and TN1 under needle puncturing (the differential probe set groups of R6N/ R6C and T6N/T6C) or BPH infestation (groups of R6P/ R6C, R24P/R24C, T6P/T6C and T24P/T24C) were subjected to GO analysis using GOEAST software, which includes two principal GO categories: biological process (BP) and molecular function (MF) (Additional file 4). The GO analysis of the groups R24N/R24C and T24N/ T24C was not conducted because very few differential probe sets were identified at 24 h after needle puncturing. The first level of BP categories and MF categories with significant differences and the percentages of related differential probe sets with annotations are shown in Fig. 4. As expected, in the needle-puncturing groups (R6N/R6C and T6N/T6C), 'response to wounding' was the GO term found having significant differences. Of the BP categories, the differential probe sets of 'response to biotic stimulus' , 'response to abiotic stimulus' , 'response to wounding' , 'defense response to fungus' , 'cell wall macromolecule catabolic' and 'chitin catabolic process' were mainly up-regulated, while those of 'lipid transport' were mainly down-regulated (Fig. 4a). Of the MF categories, the differential probe sets of resistance-related projects, such as 'serine-type endopeptidase inhibitor activity' , 'chitinase activity' , 'beta-glucosidase activity' and 'tetrapyrrole binding' were mainly up-regulated, while those of 'RNA methyltransferase activity' were mainly downregulated (Fig. 4b).
Seeking BPH resistance-related gene candidates in RH Because too many differential probe sets were identified to test them individually in this study (Fig. 1), strategies were employed to reduce the number of BPH resistancerelated gene candidates. One strategy was to combine the microarray analysis with QTL mapping. Four BPH-resistance QTLs were previously reported [25][26][27] (Additional file 5). In this study, we identified 106 differential genes between RH and TN1 that were located within the four BPH-resistance QTL regions (Additional file 6), and 14 of them are likely to be associated with the BPH resistance according to their functional annotations (Table 2). Among them, five genes were located in the Qbph3 region, five in Bph17, one in Bph3 and three in the Qbph10 region. A major BPH-resistance QTL, BPH17/BPH3, which was mapped in the Bph17 region, was cloned and contains three tandem genes encoding lectin receptor kinases (OsLecRK1-OsLecRK3) [9]. However, the expression levels of these three genes were too low to be detected in this study.
The second strategy was to determine BPH resistancerelated gene candidates by examining gene expression patterns. Clustering the gene expression patterns divided the differential probe sets into 20 groups (Fig. 2). Group m represented those differential probe sets that were BPH-inducible only in RH but not in TN1. A Venn diagram analysis also helped to remove those differential probes sets induced by both mechanical damage and BPH infestation (Fig. 3). Finally, 23 gene candidates were selected based on the expression pattern clustering and Venn diagram analysis. These selected gene candidates shared a similar expression pattern (Fig. 5). They were inducible by BPH infestation only in RH, but not by mechanical damage or in TN1. Among the 23 BPH resistance-related gene candidates, Os03g12660, which encodes a cytochrome P450 protein, was the only gene also located in a BPH resistance QTL region (Qbph3).
Dynamic changes of SA and JA levels and differential genes associated with SA, JA and ET pathways under BPH infestation The SA and JA levels (free status) in RH and TN1 were determined at 6, 24 and 48 h after BPH infestation. Overall, the SA level in both RH and TN1 increased significantly after BPH infestation (Fig. 6a); while the JA level in both RH and TN1 decreased significantly after BPH infestation (Fig. 6b). Moreover, the change in the SA content in RH occurred more rapidly than that in TN1. The SA content in RH had increased at 6 h after BPH infestation, while that in TN1 did not increase until 24 h (Fig. 6a).
All of the differential genes under BPH-infestation treatment involved in the SA, JA and ET pathways are listed in Table 3. The information concerning the genes involved in the SA, JA and ET pathways refer to Kyoto Encyclopedia of Genes and Genomes annotations. Six differential genes involved in SA metabolism were detected, and they were up-regulated, except for one gene (Os02g19970) that was involved in both ET and SA metabolism.
Among 42 JA pathway-related genes, 17 were upregulated under BPH infestation in either RH or TN1. However, the expression level patterns of differential genes involved in the JA pathway were diverse between RH and TN1 ( Table 3). All six differential genes in RH were up-regulated and were jasmonate ZIM-domain (JAZ) transcription factors, which are important negative regulators in the JA signal transduction pathway, implying that the JA pathway was repressed in RH under BPH infestation (Table 3). All 11 differentially expressed genes only presented in TN1 under BPH infestation were JA metabolism-related genes, including several JA synthesis-related lipoxygenase (LOX) genes. Additionally, 16 differential genes related to ET metabolism were detected. Most of them were present in TN1 and were down-regulated at 24 h after BPH infestation (Table 3). Comparatively, the ET pathway in RH was less affected after BPH infestation.

Isolating BPH-inducible promoters
Microarray analysis facilitates the isolation of BPHinducible promoters. Six representative BPH-inducible genes were selected from 23 BPH resistance-related gene candidates (Fig. 5). The expression levels of these six genes were only induced by BPH infestation in RH, and they were only slightly, or not at all, induced by needle puncturing or in TN1. A qRT-PCR analysis showed that these six genes were also up-regulated by a SA or JA treatment (Fig. 7).
The genomic sequences, including the promoter regions and coding sequences of these six selected BPHinducible genes, were isolated by PCR from both RH and TN1, and then subjected to sequencing. No significant sequence differences in the promoter regions of these six genes were found between RH and TN1. Furthermore, a DNA sequence comparison of the coding regions showed no differences between RH and TN1. It was hypothesized that the inducibility of these genes in RH was not determined by the promoter sequences, but by the genetic background.
Additional rice varieties, including four BPH-resistant ('PTB33' , 'R644' , 'B5' and 'IR54751') and four BPHsusceptible varieties ('Bai56' , 'Nipponbare' , 'Zhenshan97B' and '9311'), were used for further examination. Similar to in RH and TN1, the expression levels of the six selected genes were relatively high and significantly induced by BPH infestation in all of the resistant varieties, but was relatively low and barely induced by BPH infestation in all of the susceptible varieties (Additional file 7). The genomic sequences of these six resistance-related genes were also isolated from all eight additional rice varieties and then compared. Neither the promoters nor the coding sequences had any significant sequence differences between the resistant and susceptible varieties.  [28,29]. However, the study of Wang et al. did not incorporate an untreated control for each sampling time point (0, 8 and 24 h after BPH infestation), and they simply used the samples collected at 0 h as the control for the microarray analysis. This design is problematic because the expression levels of many genes are regulated by a circadian rhythm or other environmental factors. The expression pattern clustering in our study clearly showed that 3793 differential probe sets, such as groups l, q, r and s in Fig. 2, accounting for almost a half of all detected differential probe sets, were obviously affected by a circadian rhythm or other environmental factors. These were possibly mistaken for having an association with BPH infestation because of the lack of a control for each sampling time. Second, our study included the The immune responses of rice to BPH involve molecular interactions between rice and BPH, and the resistance reactions of rice are triggered by specific effectors in BPH saliva, which are delivered into rice cells by feeding and are recognized by specific receptors in rice [30]. The needle-puncturing treatment can be used to exclude those differential genes that respond to simple mechanical wounding.
There are several other studies that also profiled gene expression in rice varieties responding to BPH infestation using RNA-sequencing (RNA seq) or microarray analyses [31,32]. In the study of Lv et al., RNA-seq of BPH susceptible WT '9311' and resistant '9311' introgression lines containing BPH15 was conducted to help identify candidate genes of BPH15, which are located in a recombination cold spot [31]. The study of Wang et al. compared gene expression profiles between Bt and non-Bt rice in response to BPH infestation [32]. In these two studies, the genetic backgrounds of rice materials used for the gene expressional analyses were both BPH susceptible '9311' [31] or 'Xiushui 11' [32]; therefore, their results could not present the responses of a rice variety resistant to BPH infestation .
The microarray analysis showed significant response differences between RH and TN1 to BPH infestation. The response of RH was more active at the early stage of BPH infestation compared with TN1 because more differential probe sets (209) were detected in RH at 6 h after BPH infestation than in TN1 (173). In contrast, many more differential probe sets (3356) were detected in TN1 at 24 h after BPH infestation than in RH (614). Many of the differential probe sets presented at 24 h reflected the physiological and metabolic changes in TN1 under BPH attack caused by its lack of resistance. The GO analysis also confirmed that many of the differential probe sets in TN1 at 24 h after BPH infestation were related to abiotic and biotic stresses. Most differential probe sets detected in RH at either 6 or 24 h after BPH infestation were up-regulated, implying that the up-regulation of the defense-related genes was possibly the molecular basis of BPH resistance in RH. The change in the SA level in RH occurred more rapidly than in TN1, and the increase in the SA level occurred prior to the decrease in the JA level, indicating that the SA pathway plays a leading role in triggering the BPH-resistance response. The decrease in the JA content was probably caused by the antagonist effect of SA, which increased after BPH infestation. The upregulation of SA synthesis-related genes was the main change trend in the SA pathway, which was consistent with the change in the SA content after BPH infestation. However, differential genes in the JA pathway showed significant diverse change patterns between RH and TN1. In RH, all six differential genes were up-regulated and were JAZ transcription factors, which are important negative regulators of the JA pathway. While in TN1, 11 differential genes involved in JA synthesis were upregulated and only two were JAZ transcription factors. JA is also associated with wounding responses. As a susceptible variety, TN1 lacks BPH resistance-related genes and undergoes much more serious damage under BPH infestation. Therefore, the up-regulation of JA synthesis-related genes probably represented a response to wounding or another stress caused by BPH, instead of a resistance response.
Although the four major BPH resistance QTLs were cloned by map-based techniques, many BPH resistancerelated genes, including OsHI-LOX [12], Bphi008a [13], OsERF3 [14], OsPLD α4 and α5 [15], OsACS2 [17], Osr9-LOX1 [19], and OsJMT1 [20], were also isolated using a reverse genetics strategy. Almost all of these BPH resistance-related genes are inducible to BPH or SSB infestation, and are involved in SA, JA, or ET pathways. A  BPH-responsive promoters have potential applications in developing BPH-resistant genetically modified rice varieties. Therefore, the approximate 2-kb upstream promoter regions of some representative RH-specific and BPH-inducible genes were isolated and fused with the GUS reporter gene. However, when tested in BPHsusceptible rice varieties, these putative BPH-inducible promoters were either not, or were poorly, induced under BPH infestation. A DNA sequencing analysis showed that these RH-specific and BPH-inducible genes showed no significant sequence differences in both the promoter and coding regions between BPH-resistant and -susceptible rice varieties. Thus, the promoters of most of these BPH-inducible genes were probably not BPH-inducible, and they may be regulated by their upstream trans-regulators in the BPH-resistance network. In other words, their expression pattern was determined by the genetic background instead of the promoter sequence.

Conclusions
In this study, a cDNA microarray analysis was conducted to reveal the genome-wide response differences between RH and TN1 under BPH infestation. Expression pattern clustering of differential probe sets demonstrated that 1467 differential probe sets (Groups p and k) may be associated with the constitutive resistance of RH and 67 (Groups a and m) with the BPH infestationresponsive resistance. The Venn diagram analysis determined 192 RH-specific and BPH-inducible probe sets. Finally, 23 genes were selected as BPH resistance-related gene candidates based on the expression pattern clustering and Venn diagram analysis. The overall response of RH to BPH infestation was more prompt than that of TN1, at the global gene expression and the SA levels. In RH, the significant increase in the SA (6 h after BPH infestation) occurred prior to the significant decrease in Fig. 7 Expression pattern testing of six selected BPH resistance-related genes in RH and TN1 after independent SA and JA treatments. All six genes were up-regulated after independent SA and JA treatments. The gene expression levels in TN1 without a treatment were normalized as the calibrators. Error bars indicate standard deviations of three biological replicates the JA (24 h after BPH infestation), implying a leading role of SA in mediating BPH-resistance responses. In RH, differential genes in the SA pathway were synthesisrelated and up-regulated after BPH infestation. The differential genes in the JA pathway were up-regulated and were JAZ transcription factors, which are important negative regulators of the JA pathway. These results were consistent with the changes in the SA and JA levels in RH seedlings infested with BPH. Comparatively, genes involved in the ET pathway were less affected by a BPH infestation in RH. The results of this study aid in understanding the molecular foundation of BPH resistance in the RH and facilitate the identification of new resistancerelated genes for breeding BPH-resistant rice varieties.

Plant materials
Rice varieties RH and TN1 were kindly provided by Prof. Hongxia Hua. RH is an indica rice cultivar from Sri Lanka, with a high-level, broad-spectrum resistance to all BPH biotypes. The indica rice cultivar TN1 is highly susceptible to all BPH biotypes. Four BPH-resistant rice varieties 'PTB33' , 'R644' , 'B5' and 'IR54751' were identified and provided by Prof. Yuqing He. Four BPH-susceptible rice varieties 'Bai56' , 'Nipponbare' , 'Zhenshan97B' and '9311' were previously collected and maintained in our laboratory. The BPH resistance of all involved rice materials were validated in our laboratory prior to the current research.

Plant sample preparation for the microarray analysis
Two rice varieties, BPH-resistant RH and BPHsusceptible TN1, were used. The BPH population for this experiment was collected from a local rice field in Wuhan, China, and raised on TN1 in cages. This BPH population is a mix of BPH biotypes 1 and 2, which represent the main natural BPH populations present locally.
Seeds of RH and TN1 were sown in small pots and grown under natural conditions. Approximately 15 seeds were sown per pot, and after 2 weeks only 10 wellgrown rice seedlings were retained in each pot. Three treatments, untreated (naturally grown), needle puncturing and BPH infestation, were implemented on the 2week-old RH and TN1 seedlings. For needle puncturing, each plant was pricked with a needle 15 times at the bottom of the seedling, and then grown normally. For BPH infestation, approximately 100 third-instar BPH nymphs were introduced per pot (an average of 10 nymphs per seedling). Each pot was placed in a single netted cage to prevent nymphs from escaping. Sampling time points of 6 and 24 h after BPH infestation were used for all of the treatments. There were three replications for each treatment and sampling time point. All 10 seedlings from each pot were collected as a replication. The seedling shoots (the aerial part) were collected after the treatment, immediately placed in liquid nitrogen and then stored at -80°C. Finally, all of the samples were shipped to CapitalBio Corporation (Beijing, China) for the microarray determinations.
RNA extraction, microarray hybridization and data analysis RNA extraction, purification, microarray hybridization and gene annotation were conducted by CapitalBio. After RNA extraction, the total RNA was purified using a Nucleospin RNA Clean-up kit (Macherey-Nagel, Germany) following the manufacturer's instructions. The RNA quality was determined by electrophoresis on 1% agarose gel with formaldehyde. Biotin-labeled cRNA was synthesized using a MessageAmp™ II-Biotin kit (Ambion, TX, USA). Under the guidance of the Affymetrix Gene-Chip Expression Analysis technical manual, the biotinlabeled cRNA was linearized into fragments of 35-300 bp in length. The linear cRNA was hybridized with an Affymetrix Rice Genome array in an Affymetrix GeneChip Hybridization Oven 320 at 45°C for 16 h. After hybridization, the chip was washed and dyed in an Affymetrix Fluidics Station 400. Finally, the chip was scanned using a GeneChip Scanner 3000. The Single Array Analysis, Comparison Analysis and Molecule Annotation System were performed by CapitalBio.

Expression profiling following plant hormone treatments
The RH and TN1 plants were grown under greenhouse conditions at 25-30°C and a photoperiod of 14 h light/ 10 h dark. At the trefoil stage, the RH and TN1 plants were transferred into hydroponic nutrient solutions containing either 0.1 mM SA or JA. At each time point of 0, 0.5, 2, 6, 12 and 24 h after treatment, three to five whole plants were collected for RNA isolation and extraction.