Suppression of a phospholipase D gene, OsPLDbeta1, activates defense responses and increases disease resistance in rice.

Phospholipase D (PLD) plays an important role in plants, including responses to abiotic as well as biotic stresses. A survey of the rice (Oryza sativa) genome database indicated the presence of 17 PLD genes in the genome, among which OsPLDalpha1, OsPLDalpha5, and OsPLDbeta1 were highly expressed in most tissues studied. To examine the physiological function of PLD in rice, we made knockdown plants for each PLD isoform by introducing gene-specific RNA interference constructs. One of them, OsPLDbeta1-knockdown plants, showed the accumulation of reactive oxygen species in the absence of pathogen infection. Reverse transcription-polymerase chain reaction and DNA microarray analyses revealed that the knockdown of OsPLDbeta1 resulted in the up-/down-regulation of more than 1,400 genes, including the induction of defense-related genes such as pathogenesis-related protein genes and WRKY/ERF family transcription factor genes. Hypersensitive response-like cell death and phytoalexin production were also observed at a later phase of growth in the OsPLDbeta1-knockdown plants. These results indicated that the OsPLDbeta1-knockdown plants spontaneously activated the defense responses in the absence of pathogen infection. Furthermore, the OsPLDbeta1-knockdown plants exhibited increased resistance to the infection of major pathogens of rice, Pyricularia grisea and Xanthomonas oryzae pv oryzae. These results suggested that OsPLDbeta1 functions as a negative regulator of defense responses and disease resistance in rice.

Phospholipase D (PLD) is one of the representative phospholipases in plants and hydrolyzes phospholipids to generate phosphatidic acid (PA) and a freehead group. In recent years, PLD has been suggested to be involved in many plant cellular processes, such as membrane degradation (Ryu and Wang, 1995), membrane tethering (Dhonukshe et al., 2003;Gardiner et al., 2003), and signaling for stress and hormone responses (Testerink and Munnik, 2005;Wang, 2005;Wang et al., 2006). PA, which can be directly generated by the action of PLD, has been recognized as an important lipid second messenger. PA mediates the generation of reactive oxygen species (ROS; Sang et al., 2001;Yamaguchi et al., 2003) and activates wound-activated mitogenactivated protein kinase (Lee et al., 2001). Several PA targets have been functionally characterized in plant cells (Testerink and Munnik, 2005).
Plant PLDs are a family of heterologous enzymes (Wang, 2005). In Arabidopsis (Arabidopsis thaliana), 12 PLD genes were reported, seven of which were directly cloned and the others were predicted by searching the BLAST database (National Center for Biotechnology Information; http://www.ncbi.nml. nih.gov/). They can be classified into six types, PLDa(3), b(2), g(3), d, «, and z(2), based on their gene architectures, sequence similarities, domain structures, and biochemical properties (Wang, 2005). The function of each plant PLD has been studied using knockout, knockdown, and overexpression of corresponding genes, resulting in the identification of the physiological functions of several PLDs. PLDa has been reported to be involved in the signaling for abscisic acid, ethylene, wounding, freezing, drought, and hyperosmotic stress (Testerink and Munnik, 2005;Wang, 2005;Hong et al., 2008). PLDa was also shown to positively regulate ROS generation (Sang et al., 2001) and to be involved in the loss of seed viability by aging (Devaiah et al., 2007). PLDb was suggested to be involved in the regulation of seed germination in rice (Oryza sativa; . Involvement of PLDd in the regulation of various responses such as drought, hydrogen peroxide, freezing, and UV irradiation has been reported (Testerink and Munnik, 2005;Wang, 2005). PLDd has also been shown to play an important role in the reorganization of microtubules as well as their association with plasma membrane (Dhonukshe et al., 2003;Gardiner et al., 2003). PLDz has been suggested to regulate root hair growth and patterning (Ohashi et al., 2003), phosphate recycling during phosphate starvation (Li et al., 2006), vesicle trafficking, and auxin response .
For defense responses in plants, activation of PLD was observed by the challenge of a pathogen in rice leaves (Young et al., 1996) and the addition of defense elicitors in tomato (Solanum lycopersicum) cells ( Van der Luit et al., 2000) and rice cells (Yamaguchi et al., 2003). In addition, the pathogen-and elicitor-induced PLD activation involved the recruitment of PLD to the membrane (Young et al., 1996;Yamaguchi et al., 2005). Concerning the expression of PLD genes in defense responses, induction of PLDb genes was reported in the elicitor treatment of cultured tomato cells (Laxalt et al., 2001) and tobacco (Nicotiana tabacum) cells (Suzuki et al., 2007). De Torres Zabela et al. (2002) reported the induction of Arabidopsis PLDa, -b, and -g genes by the challenge of a pathogen in the leaves. McGee et al. (2003) also showed the induction of rice PLDa genes by the challenge of a pathogen in the leaves. In addition, the pathogen-and elicitor-induced PLD activation involved the recruitment of PLD to the membrane (Young et al., 1996;Yamaguchi et al., 2005).
Concerning the function of the reaction product of PLD, PA, in defense responses, it has been reported that PA can induce ROS generation in Arabidopsis (Sang et al., 2001) and rice (Yamaguchi et al., 2003) and the expression of defense-related genes (Yamaguchi et al., 2005). On the other hand, 1-butanol, which inhibits the generation of PA by PLD, suppressed the elicitor-induced ROS generation and phytoalexin production in rice cells (Yamaguchi et al., 2005). Based on these findings, it has been expected that PLD posi-tively regulates defense responses in plants. On the contrary,  reported that xylanase elicitor-induced ROS generation was increased in PLDb-suppressed tomato cells. Although these findings indicate that plant PLDs play important roles in defense signaling, most evidence remains circumstantial and the potential role of each PLD isoform is still obscure. Moreover, there is no evidence for the involvement of PLDs in the disease resistance against pathogens.
In this study, we generated a series of knockdown plants for each PLD isoform in rice and analyzed their phenotypes for defense responses. We report here that the knockdown of a single gene, OsPLDb1, resulted in ROS generation and expression of defense genes, followed by spontaneous lesion formation and phytoalexin production in the absence of pathogen infection. Furthermore, the knockdown transgenic plant showed a marked increase in the disease resistance to the blast fungus, Pyricularia grisea, and the bacterial blight, Xanthomonas oryzae pv oryzae (Xoo). These results clearly indicate that OsPLDb1 functions as a negative regulator of defense responses and is involved in the disease resistance in rice.

Analysis of the Rice PLD Gene Family
Multisequence alignment analysis was carried out to examine the phylogenetic relationship of the 17 PLD genes from rice and 12 genes from Arabidopsis. The phylogeny is shown as a rooted UPGMA tree in Figure  1. In Figure 1, we basically followed the terminology for PLD family members proposed by Wang (2005) for Arabidopsis PLDs to facilitate the comparison of corresponding genes in these two model plants. Rice PLDs can be grouped into eight types, PLDa,), based on their predicted sequence homology with Arabidopsis PLDs (Fig. 1). The 16 PLD genes except OsPLDh2 have two HKD Figure 1. A phylogenetic tree of PLDs in rice and Arabidopsis. The phylogenetic tree was constructed from the matrix of amino acid sequence similarities calculated with the UPGMA program of the Wisconsin University Genetics Computer Group. Rice PLDs were classified based on their similarity to Arabidopsis PLDs. The numbers following rice PLD genes indicate The Institute for Genomic Research locus identifiers and GenBank accession numbers of the corresponding genes. [See online article for color version of this figure. ] motifs, which constitute the highly conserved domain in the PLD family and are used to define the PLD superfamily (Qin and Wang, 2002). Fourteen PLD genes except OsPLDz1, OsPLDz2, and OsPLDi contain a C2 domain that is a Ca 2+ -and phospholipid-binding domain. OsPLDz1 and OsPLDz2 contain the PX and PH domains, which are present in mammalian PLDs as well (Wang, 2005). The remaining one, OsPLDi, contains two HKD motifs but does not contain C2 or PH-PX domains. Homologues of OsPLDi are present in the genomes of Phytophthora species (Meijer and Govers, 2006), Vitis vinifera, and Medicago truncatula. The phylogenetic relationship of rice PLDs in Figure  1 was partially different from the result recently reported by . In our analysis, OsPLDh1 and OsPLDh2 or OsPLD« and AtPLD« were grouped into the same class, which corresponds to the results reported by Elias et al. (2002).
To characterize the expression patterns of PLD genes in rice, expression of the 16 rice PLD genes except OsPLDi was analyzed using cultured cells, roots, leaf sheaths, leaf blades, and immature seeds under normal growth conditions. The accumulation of their mRNA in each tissue was evaluated using quantitative reverse transcription (RT)-PCR. The expression of 16 PLD genes was detected in most tissues analyzed (Supplemental Table S1). OsPLDb1 was highly expressed in all tissues analyzed (Fig. 2). OsPLDa1 and OsPLDa5 were also highly expressed in most tissues, although a higher expression of OsPLDa1 mRNA in the cultured cells was observed. Higher expression of OsPLDa4 and OsPLDd1 was detected in roots and leaf sheaths, respectively. These results demonstrated that the PLD isoforms have overlapping but distinct patterns of expression in the different tissues/organs under normal growth conditions. On the other hand, the expression levels of OsPLD«, OsPLDd2, OsPLDd3, OsPLDh1, and OsPLDu were very low in all tissues studied (Supplemental Table S1).
Similar studies in Arabidopsis and tomato showed that AtPLDa1 and AtPLDg1 in Arabidopsis and LePLDa2 and LePLDa3 in tomato were all highly expressed in most tissues. In contrast to the rather ubiquitous expression of OsPLDb1 in rice, the expression of both AtPLDb1 and LePLDb1 was very low under normal growth conditions (Fan et al., 1999;Laxalt et al., 2001). This may relate to the fact that rice lacks g-type PLD in the genome, which was extensively expressed in Arabidopsis and tomato. For rice PLDs,  reported similar results with those reported here, although the expression profile was significantly different for some PLD isoforms, probably reflecting the differences in the plant materials such as cultivar and growth conditions.

Accumulation of ROS in the OsPLDb1-Knockdown Transgenic Rice
As an initial step to characterize the function of each PLD isoform in rice, we made knockdown plants for eight PLD genes in rice, OsPLDa1, -a3, -a4, -a5, -«, -b1, -b2, and -d2, by introducing gene-specific RNA interference (RNAi) constructs. More than 10 lines were established for each gene, and at least five independent lines were used for the analysis of the phenotypes of the T1 transgenic plants grown under normal conditions in a greenhouse. Among them, OsPLDb1knockdown plants showed a marked enhancement in the generation of ROS. The amount of hydrogen peroxide (H 2 O 2 ) in the leaves of OsPLDb1-knockdown plants (20 or 30 d after seeding) was approximately 2-fold higher than that in the vector control plants (Fig.  3A). Quantitative analysis of OsPLDb1 mRNA in the leaves of the plant confirmed a clear suppression of OsPLDb1 expression (Fig. 3B). Five independent knockdown lines gave similar results (Supplemental Fig. S1). No difference in the appearance of the leaves was observed between the OsPLDb1-knockdown plants and the vector control plants at this stage (data not shown).
The analysis of PLD activity in the leaf extracts showed that the PLD activity of the OsPLDb1-knockdown plants was decreased to half of the vector control level. However, this decrease of PLD activity was observed only for phosphatidylcholine-specific (PC-) PLD activity but not for phosphatidylethanolaminespecific (PE-) PLD activity (Fig. 3C). These results indicate that the knockdown of OsPLDb1 really reflects the change of in planta PLD activity and that the in vivo substrate of OsPLDb1 is PC.
To examine whether the higher accumulation of H 2 O 2 in the OsPLDb1-knockdown plants is caused by the down-regulation of scavenging activity of ROS, we analyzed the catalase-and peroxidase-like activities in the leaf extracts of the OsPLDb1-knockdown and vector control plants. No difference in the enzyme activities was observed between these plants (data not shown), indicating that the knockdown of OsPLDb1 probably affected ROS generating rather than scavenging activity.

Up-Regulation of Defense-Related Genes in the OsPLDb1-Knockdown Plants
We also analyzed whether the expression of defenserelated genes was up-regulated in these OsPLDb1knockdown plants. Quantitative RT-PCR analysis of three typical defense-related genes, probenazoleinducible protein1, chitinase, and thaumatin-like protein, in the leaves of 30-d-old plants (i.e. 30 d after seeding) showed that the expression of these genes was dramatically increased in the OsPLDb1-knockdown plants, concomitant with the decrease of OsPLDb1 expression (Fig. 4A). Global changes in the gene expression induced by OsPLDb1 knockdown were further analyzed for the leaves from 30-d-old plants using a rice 44K DNA microarray. Figure 4B summarizes the up-and down-regulation of those genes in each category, whose expression was changed more than three times compared with that of the vector control plants. Approximately 1,400 genes changed their expression significantly in the OsPLDb1-knockdown plants under such conditions (Supplemental Table S2). More than 600 genes with the annotated functions, notably those associated with defense responses, signal transduction, and a number of transcription factors, were upregulated in the OsPLDb1-knockdown plants. These defense-related genes included those for PR-1, b-glucanase, chitinase, PR-4, thaumatin-like protein, and probenazol-inducible proteins. In addition, a number of transcription factor genes belonging to the WRKY and ERF families were also up-regulated. These genes are well known to be up-regulated during the defense responses (Van Loon and Van Strien, 1999;Maleck et al., 2001). On the other hand, many cell cycle-related genes, such as aand b-expansin genes, were down-regulated ( Fig. 4B; Supplemental Table  S2). These results are similar to the previous observations on the elicitor-induced gene responses in rice cells (Akimoto-Tomiyama et al., 2003;Desaki et al., 2006). These data, therefore, indicate that the knockdown of OsPLDb1 resulted in the induction of defense-related genes in the absence of pathogen infections, a novel observation that, to our knowledge, has not been reported for any plant PLD. Recently,  reported that OsPLDb1 plays a positive role in the abscisic acid response and that the suppression of OsPLDb1 resulted in the up-regulation of GAmyb (X98355) and a-amylase (AK101744) genes during seed germination. However, the up-regulation of the GAmyb gene as well as the a-amylase gene was not observed in either the seedlings or leaves of OsPLDb1-knockdown plants by DNA microarray analysis (data not shown), probably reflecting the differences in the experimental systems.
It was also observed that the expression of OsPLDb2, a closely related PLD isoform in rice, was significantly up-regulated in the OsPLDb1-knockdown plants (Supplemental Fig. S2). The exact biological significance of this phenomenon is not clear, but it might reflect the cellular response to compensate for the knockdown of OsPLDb1. In spite of the up-regulation of OsPLDb2 expression, PC-PLD activity of the knockdown plant was clearly suppressed (Fig. 3C), indicating that OsPLDb1 is the major PLD isoform contributing to this activity.

Induction of Hypersensitive Response-Like Reactions in the OsPLDb1-Knockdown Plants
When the knockdown plants for eight PLD genes were grown further in a greenhouse, small reddish brown lesions became visible over the surface of the leaves only in the OsPLDb1-knockdown plants approximately 40 d after seeding, although such a phenotype was not visible for the 30-d-old plants, where the up-regulation of various defense genes and the H 2 O 2 accumulation were already started (Fig. 5A). Similar results were observed for 51 of 55 independent OsPLDb1-knockdown lines, although the density of the lesions was different among them. Quantitative analysis of OsPLDb1 mRNA in the leaves of OsPLDb1knockdown lines, all of which developed the lesions, showed a clear suppression of OsPLDb1 expression, as shown in Figure 5B

Suppression of OsPLDb1 Induces Phytoalexin Production
To examine whether phytoalexin production is up-regulated in the OsPLDb1-knockdown plants, we analyzed the amount of momilactone A, a major phytoalexin of rice (Cartwright et al., 1977), in the leaves of 50-d-old OsPLDb1-knockdown plants and vector control plants. As shown in Figure 6 (RNAi b), momilactone A concentration in the OsPLDb1-knockdown plants with visible lesions was approximately 100 times higher compared with the vector control plants.
On the other hand, the amount of momilactone A in the leaves of the OsPLDb1-knockdown plants without the lesions was only slightly increased compared with the vector control plants (RNAi a). These results indicate that the phytoalexin production is upregulated in the OsPLDb1-knockdown plants but starts mostly after the development of the lesions.

Suppression of OsPLDb1 Enhances Disease Resistance against Rice Bacterial Blight as Well as Rice Blast
Susceptibility of the OsPLDb1-knockdown rice plants to a virulent blast fungus was evaluated to see whether the activation of defense responses induced by the OsPLDb1 knockdown resulted in an increase of disease resistance. The OsPLDb1-knockdown and vector control plants (at the three-to six-leaf stage) were inoculated with P. grisea (teleomorph Magnaporthe grisea) race 007, a major fungal pathogen and compatible to the rice variety used in this experiment, and the disease symptoms were evaluated 7 d later. Contrary to the development of extensive lesions in the leaves of the vector control plants, the OsPLDb1-knockdown plants displayed remarkably decreased lesion formation, irrespective of the age of the tested plants (20 or 30 d old; Fig. 7, A and C). Quantification of the lesion area indicated that the development of susceptibletype lesions in the OsPLDb1-knockdown plants was decreased more than 7-fold compared with the vector control plants (Fig. 7, B and D), indicating an evident increase of disease resistance against rice blast infection. Similar results were obtained with other OsPLDb1-knockdown lines, where the degree of disease resistance seemed to correlate with the degree of OsPLDb1 knockdown (Supplemental Fig. S3; correlation coefficient = 0.904). To examine whether the OsPLDb1-knockdown plants also show disease resistance against other pathogens, we tested their response to rice bacterial blight infection, Xoo, using the uppermost fully extended leaves of 50-d-old plants (Fig. 7, E and F). Similar to the case of blast infection, lesion formation caused by the bacterial blight was  dramatically decreased (4.4-fold decrease) in the OsPLDb1-knockdown plants (Fig. 7E). In addition, the population of Xoo in the OsPLDb1-knockdown plants was decreased more than 10-fold compared with the vector control plants at 15 d after Xoo inoculation (Supplemental Fig. S4). These results clearly show that the suppression of OsPLDb1 enhanced disease resistance against a broad range of pathogens, including bacteria and oomycetes. It is noteworthy that these experiments were performed with the uppermost fully extended leaves, where the hypersensitive response (HR)-like cell death and the phytoalexin accumulation were not yet visible.
The expression of OsPLDb1 itself was induced by the inoculation of P. grisea to the wild-type rice leaves (Fig. 7G). The expression of OsPLDb1 reached a maximum (3.6-fold increase) at 1 d after inoculation, preceding the susceptible lesion formation.

Knockdown of OsPLDb1 Activates Defense Responses and Increases Disease Resistance in Rice
In this paper, we show that the knockdown of a rice PLD gene, OsPLDb1, resulted in the activation of defense responses and increased disease resistance in rice. We establish the knockdown plants for eight PLD genes in rice, OsPLDa1, -a3, -a4, -a5, -«, -b1, -b2, and -d2, which were listed in the database when we started this work, using RNAi. Among them, only OsPLDb1knockdown plants showed the activation of typical defense responses such as ROS generation and defense gene expression in the absence of pathogen infection, followed by the development of HR-like lesions and phytoalexin production in a later growth stage. The fact that the induction of ROS generation as well as defense gene expression in the OsPLDb1-knockdown plants preceded the HR-like lesion formation indicates that the knockdown of OsPLDb1 directly affected the upstream part of these defense responses, rather than through HR cell death. Concerning the induction of HR-like lesion formation, various mutants or transgenic plants that exhibit a lesion-mimic phenotype have been reported, providing a good model system to study the molecular mechanisms of HR in plants (Lorrain et al., 2003). However, it has not been reported that the knockdown or knockout of the genes involved in phospholipid signaling leads to the lesion-mimic phenotype associated with defense responses.
The OsPLDb1-knockdown plants exhibited increased disease resistance to major pathogens of rice, P. grisea and Xoo, before the formation of visible spontaneous lesions (Fig. 7). In this respect, the OsPLDb1-knockdown plants are similar to the barley (Hordeum vulgare) mlo (Wolter et al., 1993) and rice cdr3 (Takahashi et al., 1999) mutants, which exhibit resistance to fungal pathogens before the formation of visible lesions. Interestingly, the up-regulated genes in the OsPLDb1-knockdown plants included those genes for transcription factors that have been reported to be induced by the pathogen infection/elicitor treatment and play an important role for defense signaling, such as WRKY45 and WRKY71. Shimono et al. (2007) recently showed that the overexpression of WRKY45 enhanced blast resistance, and Liu et al. (2007) showed that the overexpression of WRKY71 enhanced resistance to virulent bacterial pathogens. The expression levels of OsWRKY71 and OsWRKY45 genes in the OsPLDb1-knockdown plants increased 8.7-and 3.8fold compared with that in the vector control plant, respectively (Supplemental Table S2), indicating that the up-regulation of these genes may contribute to the induction of defense-related genes and the enhancement of blast resistance.
No significant difference was observed between the growth of the vector control and OsPLDb1-knockdown plants, such as plant height, number of leaves, and heading date after seeding, although the seed weight of the OsPLDb1-knockdown plants seemed to be slightly decreased compared with that of the vector control plants (Supplemental Table S3). These results suggested that the suppression of OsPLDb1 increased disease resistance of rice without a significant effect on growth and development.
Concerning the up-regulation of OsPLDb1 by the infection of a compatible pathogen (Fig. 7G), it is worth remembering that some negative regulator genes for defense responses, such as rice spl11 and Arabidopsis WRKY48, were up-regulated in response to pathogen infection (Zeng et al., 2004;Xing et al., 2008). As the induction of defense reactions such as HR potentially has an adverse effect on normal growth, these responses must be induced when plants are infected with pathogens and also be down-regulated properly. Thus, the induction of negative regulators such as OsPLDb1 by pathogen infection may constitute part of such a negative feedback loop.

OsPLDb1, a Negative Regulator of Defense Responses?
It has been suggested that the activation of PLD and the resulting PA generation during pathogen infection/elicitor treatment play a positive role in the mobilization of defense responses. We previously reported that the elicitor treatment of rice cells activated PC-PLD and induced PA generation (Yamaguchi et al., 2003(Yamaguchi et al., , 2005. The addition of PA itself directly induced ROS generation and expression of defenserelated genes in the absence of the elicitor treatment. Sang et al. (2001) also reported that the suppression of AtPLDa1 decreased the level of ROS generation and that the addition of PA recovered the ROS generation in Arabidopsis leaves. Our results described here, however, indicate that OsPLDb1 negatively regulates defense responses, probably at the upstream part of the defense signaling cascade.  also reported that the knockdown of a tomato PLDb gene, LePLDb1, resulted in the up-regulation of ROS generation, although in the presence of xylanase elicitor. Although the contribution of other isoforms of rice PLDs, such as OsPLDb2, in the up-regulation of defense responses in the OsPLDb1-knockdown rice cannot be excluded, these results seem to indicate the involvement of PLDb1 in the negative regulation of defense responses in plants. Our results also indicate that OsPLDb1 seems to suppress defense responses in rice even under normal growth conditions; however, it is not clear whether this is the case for tomato as well. As already described, the expression level of OsPLDb1 was very high in most tissues studied (Fig. 2), whereas that of LePLDb1 was very low (Laxalt et al., 2001). These differences in the expression levels of PLDb1 genes might cause the differences in their regulation of defense responses in the corresponding plants under normal growth conditions.
These results indicate that plant PLDs are involved in the positive and negative regulation of defense responses, probably depending on the corresponding isoforms. Similar observations have been reported in the responses against freezing or oxidative stress in Arabidopsis. It was reported that the knockdown of PLDa1 in Arabidopsis enhanced freezing tolerance, whereas the knockout of PLDd weakened freezing tolerance (Wang, 2005). Wang (2005) also indicated that AtPLDa1 is involved in ROS generation, whereas AtPLDd is involved in the suppression of ROSpromoted programmed cell death. Which isoforms of PLDs are involved in the opposite regulation of these defense responses should be clarified further in future studies.

Possible Mechanisms for the Different Functions of PLD Isoforms
If plant PLDs play opposite roles in plant defense signaling, as discussed above, what can be the explanation for such mechanisms? One possibility is the different localization of each PLD isoform, which enables the regulation of different target proteins downstream of each PLD. Several papers have indicated that the different PLD isoforms localize in different organelles/membranes in plant cells. In rice, OsPLDa1 was detected in the cell wall, membranes, and chloroplasts, whereas OsPLDa4 and OsPLDa5 were predominantly detected in the chloroplasts of the leaves (McGee et al., 2003). AtPLDa1 in Arabidopsis was detected in the plasma membrane, intracellular membranes, mitochondria, and clathrin-coated vesicles (Fan et al., 1999). AtPLDb1 was reported to bind actin in vitro, which is a major component of a microfilament in the cells (Kusner et al., 2003).  showed that the fusion protein of tomato LePLDb1 and GFP was detected in the cytosol of cultured cells, whereas treatment with xylanase elicitor induced the relocalization to punctate structures within the cytosol. They proposed a possibility that plant PLDb1 may form a tether between vesicular membranes and the actin cytoskeleton . These results suggest that the cellular localization is different between PLDa and PLDb, indicating the different physiological functions of these PLDs.
In addition to the different localization of PLD isoforms, different modes of regulation of each PLD may be a factor controlling the function of each PLD. For example, it is well known that the Ca 2+ dependence of activation is very different among PLD isoforms (Qin and Wang, 2002). Similarly, PLDb1, PLDg1, and PLDz1 require phosphatidylinositol 4,5-bisP for the activation, whereas other isoforms do not (Qin and Wang, 2002). These results indicate that different PLDs are involved in the different physiological responses downstream of these second messengers.
Different molecular species of PA may also be generated by different PLD isoforms, as differences in the substrate preference among PLD isoforms have been reported (Qin and Wang, 2002). More recently, Hong et al. (2008) reported that the molecular species of PA generated by the action of AtPLDa3 are clearly different from those by AtPLDa1 in Arabidopsis. This observation suggests that the differences in the molecular species of PA generated by these PLDs might reflect differences in the functions of the PLD isoforms. Furthermore, different phospholipid compositions of organelle membranes might also contribute to the generation of different PAs by PLD isoforms, resulting in the differences in PA target proteins.
Although the results obtained in this study with knockdown transgenic plants indicate that OsPLDb1 negatively regulates defense responses in rice, it is not clear whether OsPLDb1 is actually involved in the defense regulation under physiological conditions, and this should be clarified in future studies. The fact that the knockdown of OsPLDb1 resulted in the activation of various defense responses, including the up-/down-regulation of numerous genes, ROS generation, followed by HR cell death and phytoalexin production, indicates that OsPLDb1 regulates the defense responses at the upstream part of the signal transduction pathways. The analysis of the mechanism of negative regulation of defense responses by OsPLDb1, including the survey of downstream components in OsPLDb1 signaling, should be the subject of future studies as well.

Plant and Fungal Materials
The rice (Oryza sativa japonica 'Nipponbare', blast resistance gene: Pik-s) was used in this study. The sterilized transgenic or wild-type rice seeds were germinated for 2 weeks on Murashige and Skoog (MS) agar medium with or without 50 mg L 21 hygromycin, respectively. The resulting plants were transferred to soil and grown in a greenhouse (25°C/27°C, solar radiation).

DNA Database Search and Computer Analysis
To identify the members of the PLD gene family in rice, we searched the following databases: DNA Data Bank of Japan (http://www.ddbj.nig.ac.jp), National Center for Biotechnology Information (http://www.ncbi.nml.nih. gov/blast/Genome), and National Institute of Agrobiological Sciences (http://cdna01.dna.affrc.go.jp/cDNA/). Multisequence alignment analysis for the deduced amino acid sequences was carried out using Genetyx (Software Development).

RNA Extraction and Real-Time PCR
Total RNAs for first-strand cDNA synthesis were isolated from rice tissues by the SDS-phenol method. The isolated RNA (5 mg) was reverse transcribed (SuperScript II; Invitrogen) using an oligo(dT) 13 primer. An aliquot of the firststrand cDNA mixture corresponding to 10 ng of the total RNA was used as a template for real-time PCR analysis. The reaction was carried out on the Smart Cycler System (Cepheid) with the SYBER premix ExTaq (Takara Bio) according to the manufacturer's instructions. The gene-specific primers designed for each gene are listed in Supplemental Table S4. The amplified bands were cloned directly into pGEM-T vector (Promega) and sequenced to confirm that they were indeed the fragments of the intended genes. A calibration curve for each gene was obtained by performing real-time PCR with several dilutions of the cloned cDNA fragment. The specificity of the individual PCR amplifica-tion was checked using a heat dissociation protocol from 65°C to 95°C following the final cycle of PCR. The results obtained for the different cDNAs were standardized to the expression level of a rice polyubiquitin gene (RUBIQ1; Wang et al., 2000).

DNA Constructs and Rice Transformation
The OsPLDb1-RNAi vector was constructed by generating an inverted hairpin loop into the pZH2Bik binary vector, which was modified to contain the rice ubiquitin promoter (OsmUbiP), the intron of a rice aspartic protease gene (RAP intron; Asakura et al., 1995), and the Nos terminator (T-Nos), in the sequence from pPZP202 (Hajdukiewicz et al., 1994). A 343-bp fragment corresponding to nucleotides 2,323 to 2,665 of OsPLDb1, which does not overlap with the amplified OsPLDb1-specific fragment (1,973-2,314) in the real-time PCR experiments described in Supplemental Table S4, was amplified using the following oligonucleotides: 5#-GAATGCATGAGACGAGTTCGC-3# (forward) and 5#-CCTTTTCTTGTCATGTCCTACT-3# (reverse). The PCR fragment was ligated in the reverse orientation into pZH2bik vector between OsmUbiP and RAP intron and between RAP intron and T-Nos. The resulting vector contains the RNAi intermediate sequence downstream of the rice ubiquitin promoter. The empty vector, pZH2Bik, was also used to generate vector control plants. The other rice PLD gene-specific RNAi vectors were also constructed according to the same methods described above. To confirm the gene specificity, part of the cDNA including the 3# noncoding region, the sequence of which is different for each PLD gene, was amplified using each primer set (Supplemental Table S5) and inserted into the vector.
Agrobacterium tumefaciens EHA105 carrying the above construct was used to transform Nipponbare rice. The Agrobacterium-mediated transformation was performed according to the method of Toki (1997) using vigorously growing callus derived from mature embryos. Transformed calli were selected on MS agar medium containing 50 mg L 21 hygromycin. Regenerated plants were grown in a greenhouse, and T0 transgenic seeds were harvested 40 d after flowering. T1 transgenic plants carrying the transgene were selected by germinating seeds on MS agar medium containing 50 mg L 21 hygromycin for 2 weeks, and then selected plants were transferred to soil and grown in a greenhouse (25°C-30°C, solar radiation).

Determination of H 2 O 2 Accumulation and Scavenging Activity
A 50-mg sample of leaf tissue harvested from 30-to 50-d-old plants was cut into small pieces (5 3 5 mm) and immediately homogenized with 5% (w/v) aqueous trichloroacetic acid containing EGTA (10 mM) on ice. The extract was centrifuged at 10,000g at 4°C for 10 min, and the amount of H 2 O 2 in the supernatant was analyzed using a quantitative H 2 O 2 assay kit (Bioxytech H 2 O 2 -560; Oxis International).
ROS-scavenging activity was determined using 10 mM H 2 O 2 as a substrate. A 20-mg sample of leaf tissue harvested from 30-to 50-d-old plants was cut into small pieces (5 3 5 mm) and immediately homogenized with 0.5 mL of 25 mM Tris-HCl buffer (pH 7.5) containing 10 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 0.3 M Suc for 1 min on ice. The homogenate was centrifuged at 10,000g at 4°C for 10 min, and the supernatant was stored as the cellular extract fraction at 220°C until analysis. Protein content in each fraction was determined according to Bradford (1976). The standard reaction mixture contained 1.0 to 1.5 mg of cellular protein and 10 mM H 2 O 2 in a final volume of 50 mL. Reactions were performed at 37°C for 10 min, and the amount of H 2 O 2 in the supernatant was analyzed using a quantitative H 2 O 2 assay kit.

Measurement of in Vitro PLD Activity
PLD activity in the cellular extract fraction described above was determined using fluorescent lipids (Avanti Polar Lipids) as a substrate as described previously (Yamaguchi et al., 2004) with a slight modification. 7-Nitro-2-1,3-benzoxadiazol-4-yl amino (NBD)-PC (14:0 12:0) and NBD-PE (14:0 12:0) were stored at 220°C in chloroform. Prior to use, they were dried and emulsified by sonication in 5 mM sodium cholate. The standard reaction mixture contained 25 mM HEPES (pH 7.0), 10 to 15 mg of cellular protein, 0.1 mM NBD-PC or NBD-PE, and 1-butanol (final concentration, 0.1% [v/v]) in a final volume of 50 mL. Reactions were performed at 37°C for 30 min and then stopped by the addition of 0.25 mL of CHCl 3 :methanol:HCl (100:100:0.6, v/v). After 0.1 mL of 1 N HCl containing 5 mM EGTA was added, the mixture was vortexed and centrifuged at 10,000g for 1 min. The solvent layer was then recovered and dried under vacuum. The samples were applied to a thin-layer chromatography plate (20 3 10 cm) and developed with the organic upper phase of ethylacetate:iso-octane:acetic acid:water (12:2:3:10, v/v). Fluorescently labeled phosphatidylbutanol (PtdBut) was visualized using a UV transilluminator, and the regions corresponding to PtdBut were marked. The marked spots were scraped from the plates and placed in 0.5 mL of CHCl 3 : methanol:water (5:5:1, v/v), vortexed, and centrifuged for 5 min at 10,000g. The supernatant was dried under vacuum and emulsified by sonication in 5 mM sodium cholate. The fluorescence (excitation, 460 nm; emission, 534 nm) from the eluted PtdBut was measured with a fluorescence spectrophotometer (Versa Fluor; Bio-Rad). The spots of PtdBut were identified by comparison with the standards (Avanti Polar Lipids) visualized by iodine vapor. A standard curve was constructed using a range of NBD-PC concentrations.

Microarray Analysis
Microarray analyses were performed using a 60-mer rice oligomicroarray containing 44K features (Agilent Technologies). The total RNA used for the analysis was prepared from leaves of 30-d-old plants by the SDS-phenol method. The RNA extracts were fluorescently labeled and hybridized to the microarray slides according to the manufacturer's protocol. The slides were then scanned by an Agilent scanner. To assess reproducibility, two independent experiments were conducted using two independent plants. Data indicating a less than 3-fold increase in fluorescence between Cy5 and Cy3 in each experiment were excluded from further analysis. Genes whose expression in the knockdown plants was increased or decreased more than three times compared with that in the vector control plants are listed in Supplemental Table S2.

Determination of Phytoalexin
A 50-mg sample of leaf tissue harvested from 50-d-old plants was cut into small pieces (5 3 5 mm) and extracted with 0.5 mL of 70% (v/v) methanol at room temperature for 24 h. After centrifugation at 10,000g for 10 min, the supernatant was analyzed with a liquid chromatography-mass spectrometry system. An HP 1100 HPLC apparatus (Hewlett-Packard) equipped with an Inertsil ODS 3 column (4.6 3 250 mm; GL Science) was used. Elution with 80% (v/v) aqueous acetonitrile (containing 0.1% [v/v] formic acid) was carried out at a flow rate of 1.0 mL min 21 . Mass spectrometry was performed using a Mariner electrospray ionization-time of flight mass spectrometer (Applied Biosystems) in the positive-ion mode.

Inoculation of P. grisea and Xoo on Rice Plants
Inoculation of blast fungus on rice plants was performed according to the methods of Ashizawa et al. (2005). The isolate of P. grisea (Ina 86-137) was transferred to oatmeal agar plates and incubated at 25°C for 14 d in the dark. The aerial hyphae on these plates were removed by gently brushing the agar surface, and then the plates were incubated for 72 h at 25°C under white fluorescent light. To prepare the conidial suspension, the conidia formed on the surface were collected by adding distilled water containing 0.02% (w/v) Tween 20. The OsPLDb1-knockdown and vector control plants grown for 20 or 30 d after seeding (the three-to six-leaf stage) under normal growth conditions were sprayed with 10 mL of a conidial suspension per plant (10 4 conidia mL 21 in distilled water containing 0.02% Tween 20) using an atomizer with a fine nozzle. The sprayed plants were kept in a moist chamber at 100% relative humidity in the dark at 25°C for 20 h and then moved to a greenhouse (25°C/ 27°C, solar radiation). The disease symptoms were observed for 1 to 2 weeks after inoculation, and the area of diseased leaf lesions was measured.
Inoculation of bacterial blight was performed according to the methods of Kauffman et al. (1973). The isolate of Xoo (T-7133) was grown on PPGA medium (Nishiyama and Ezuka, 1977). The uppermost fully extended leaves of OsPLDb1-knockdown and vector control plants grown for 50 d after seeding under normal growth conditions were inoculated with a bacterial suspension (10 8 cells mL 21 ) by the leaf-clipping method. The disease symptoms were observed for 2 weeks after the inoculation, and the area of the diseased leaf lesions was measured. To determine the bacterial populations, the inoculated leaves were harvested at 0, 5, and 15 d after Xoo inoculation and surface sterilized with 70% (v/v) ethanol and 3% (v/v) sodium hypochlorite solution. The leaves were cut into small pieces and resuspended in 10 mL of water to harvest bacteria separately. Diluted extract (0.05 mL) was incubated on PSA medium (5 g L 21 Bacto peptone, 15 g L 21 Suc, 5 g L 21 soluble starch, 0.5 g L 21 calcium nitrate, and 15 g L 21 agar, pH 7.0) at 28°C for 4 to 6 d, and the number of colonies formed on the agar plate was counted.
Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers listed in Supplemental Tables S4 and S5.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Accumulation of H 2 O 2 and OsPLDb1 mRNA in the leaves of 30-d-old OsPLDb1-knockdown plants.
Supplemental Figure S2. Accumulation of OsPLDb1 and OsPLDb2 mRNA in the OsPLDb1-knockdown plants.
Supplemental Figure S3. Comparison between blast resistance and mRNA levels of OsPLDb1 in the OsPLDb1-knockdown plants.
Supplemental Figure S4. Growth curve of Xoo in the OsPLDb1-knockdown and vector control plants.
Supplemental Table S1. The expression analysis of rice PLD genes.
Supplemental Table S3. Growth and development parameters of the vector control and OsPLDb1-knockdown plants.
Supplemental Table S4. Gene-specific PCR primers used for quantitative RT-PCR amplification.
Supplemental Table S5. Gene-specific PCR primers used for the construction of RNAi vectors.