Overexpression of MoSM1, encoding for an immunity-inducing protein from Magnaporthe oryzae, in rice confers broad-spectrum resistance against fungal and bacterial diseases

Potential of MoSM1, encoding for a cerato-platanin protein from Magnaporthe oryzae, in improvement of rice disease resistance was examined. Transient expression of MoSM1 in rice leaves initiated hypersensitive response and upregulated expression of defense genes. When transiently expressed in tobacco leaves, MoSM1 targeted to plasma membrane. The MoSM1-overexpressing (MoSM1-OE) transgenic rice lines showed an improved resistance, as revealed by the reduced disease severity and decreased in planta pathogen growth, against 2 strains belonging to two different races of M. oryzae, causing blast disease, and against 2 strains of Xanthomonas oryzae pv. oryzae, causing bacterial leaf blight disease. However, no alteration in resistance to sheath blight disease was observed in MoSM1-OE lines. The MoSM1-OE plants contained elevated levels of salicylic acid (SA) and jasmonic acid (JA) and constitutively activated the expression of SA and JA signaling-related regulatory and defense genes. Furthermore, the MoSM1-OE plants had no effect on drought and salt stress tolerance and on grain yield. We conclude that MoSM1 confers a broad-spectrum resistance against different pathogens through modulating SA- and JA-mediated signaling pathways without any penalty on abiotic stress tolerance and grain yield, providing a promising potential for application of MoSM1 in improvement of disease resistance in crops.


Results
Transient expression of MoSM1 in rice elicits HR and activates expression of defense-related genes. Before the investigation of function of MoSM1 in rice defense, we further examined the biochemical features of MoSM1 such as the role of signal peptide (SP) in activity and subcellular localization in planta. Because MoSM1 contains a 21 aa SP at its N-terminal 42 , we first examined whether this SP was necessary for its activity to induce HR. To do this, a SP-deletion vector, pINDEX3::MoSM1 ΔSP , in which the SP region was deleted, was constructed. At 24 hr after agroinfiltration, transient expression of MoSM1 in leaves of N. benthamiana plants induced a typical HR cell death while transient expression of MoSM1 ΔSP did not (Fig. 1a). We also examined the subcellular localization of MoSM1 in planta through generating an eGFP-MoSM1 fusion construct. When transiently expressed in leaves of N. benthamiana plants, significant accumulation of eGFP-MoSM1 fusion was detected by Western blotting, which was similar to the accumulation of eGFP alone, at 24 hr after agroinfiltration (Fig. 1b). Microscopic observation revealed that the eGFP-MoSM1 fusion targeted mainly to the plasma membrane, co-localized with a previously reported plasma membrane-targeted Arabidopsis aquaporin AtPIP1;4, whereas the eGFP alone distributed throughout the cells with distribution in nucleus, at 24 hr after agroinfiltration (Fig. 1c). These data suggest that SP in MoSM1 is required for its activity to induce HR and that MoSM1 targets to plasma membrane of cells in planta.
We next examined whether transient expression of MoSM1 in rice could induce defense responses such as HR and expression of defense-related genes. To this end, agrobacteria carrying pINDEX3::MoSM1 and empty pIN-DEX3::00 were infiltrated into leaves of 5-week-old rice plants and transient expression of MoSM1 was induced by exogenous application of 3 μ M DEX 42 . Transcript of MoSM1 in pINDEX3::MoSM1-infiltrated leaves was first detected at 12 hr after DEX treatment and the MoSM1 transcript levels gradually increased over a period of 12-48 hr (Fig. 1d). However, no MoSM1 transcript was detected in pINDEX3::00-infiltrated leaves (Fig. 1d). Water-soaked symptom initially occurred around the infiltration sites at 12 hr after DEX treatment in pINDEX-3::MoSM1-infiltrated leaves and further enlarged and expanded into large area, leading to dried and colorless tissues at 48 hr (Fig. 1e). No obvious water-soaked symptom was observed in pINDEX3::00-infiltrated leaves  Data presented in (f) are the means ± SD from three independent experiments and asterisks indicate statistically significant difference at p = 0.05 level between pINDEX3::MoSM1-and pINDEX3::00-infiltrated leaves. (Fig. 1e). Similarly, electrolyte leakage in pINDEX3::MoSM1-infiltrated leaves increased significantly by 13%, 26% and 39% at 12, 24 and 48 hr after DEX treatment compared to the level at 0 hr, while the electrolyte leakage in pINDEX3::00-infiltrated leaves had no significant change (Fig. 1f). These results indicate that transient expression of MoSM1 in rice leaves can induce a typical HR.
We further examined whether transient expression of MoSM1 activated defense response by comparing the expression patterns of some selected defense-related genes between pINDEX3::MoSM1-and pINDEX-3::00-infiltrated leaves after DEX treatment. As shown in Fig. 1g, induced expression of OsPR1b and OsPR10 was detected as early as 12 hr and the expression levels increased further at 24 and 48 hr after DEX treatment in pINDEX3::MoSM1-infiltrated leaves, although relatively lower expression levels of OsPR1b and OsPR10 were also detected in pINDEX3::00-infiltrated leaves. These data indicate that transient expression of MoSM1 can also activate the expression of defense genes in rice and thus transiently expressed MoSM1 may function as an elicitor of defense response in rice.
Generation and characterization of MoSM1-OE transgenic rice lines. Transgenic rice lines with overexpression of MoSM1 (MoSM1-OE) were generated to precisely evaluate the function of MoSM1 in rice immunity. For this purpose, the MoSM1 full coding sequence including SP was constructed into a binary vector under the control of the maize ubiquitin promoter (Fig. 2a) and transformed into rice cv. Xiushui11. A total 36 independent transgenic lines were obtained through the Agrobacterium-mediated transformation method and two independent lines (SM1-70 and SM1-73) with single copy of the MoSM1 gene, as revealed by Southern blotting (Fig. 2b), were chosen for further studies. qRT-PCR analysis indicated that the expression level of the MoSM1 gene in plants of the T3 generations of the MoSM1-OE lines SM1-70 and SM1-73 was estimated to be 9 and 22 folds of the level of a rice Actin gene, respectively (Fig. 2c). However, no transcript of MoSM1 was detected in nontransgenic wild type (WT) plants (Fig. 2c). During our study, we did not observe significant changes in morphological and developmental phenotypes in the MoSM1-OE lines SM1-70 and SM1-73 at seedling and adult stages as compared to the WT plants ( Fig. 2d and e). Furthermore, we also compared some important agronomic traits such as seed setting rate, panicle weight and grains per panicle of the MoSM1-OE plants with the nontransgenic WT plants grown under greenhouse condition. Overall, the panicles of the MoSM1-OE plants were comparable to those of the nontransgenic WT plants (Fig. 2f). No significant difference in some of key agronomic traits such as seed setting rate, panicle weight and grains per panicle was detected between the MoSM1-OE and nontransgenic WT plants ( Fig. 2g-i). Taken together, these data indicate that the MoSM1-OE plants do not have abnormal growth, development and grain production.

MoSM1-OE plants display an improved resistance against blast disease.
To explore the function of MoSM1 in rice disease resistance, we first evaluated the resistance level of the MoSM1-OE lines at seedling stage against blast disease, which is the most serious constraint on rice caused by M. oryzae. Two M. oryzae strains, 85-14 and 97-220, representing races ZB1 and ZE3, respectively, were used to inoculate the two-week-old transgenic and nontransgenic WT seedlings by foliar spraying of spore suspension and disease levels were evaluated by assessing disease scores and measuring in planta fungal growth. As shown in Fig. 3a,b and c, the overall blast disease phenotype on the MoSM1-OE lines SM1-70 and SM1-73 was less severe than those in the nontransgenic WT plants after inoculation with strains 85-14 and 97-220. Typical blast lesions were seen on leaves of nontransgenic WT plants, while almost no typical blast lesion was observed on leaves of the MoSM1-OE lines SM1-70 and SM1-73 (Fig. 3a,b and c). Similarly, significant cell death, as revealed by Trypan blue staining, in inoculated leaves of the nontransgenic WT plants was detected, while no remarkable cell death was seen in inoculated leaves of the MoSM1-OE lines SM1-70 and SM1-73 ( Fig. 3d). At 7 days post inoculation (dpi) with strains 85-14 and 97-220, the averages of the blast disease scores on the nontransgenic WT plants were 5.8 and 4.7 grades, respectively, while the averages of the blast disease scores on the MoSM1-OE lines SM1-70 and SM1-73 were 0.2-0.3 grades (Fig. 3e), resulting in 4-5 grades lower than those on the WT plants. Further measurement of in planta fungal growth, as revealed by analyzing genomic DNA level of the 28S rDNA gene of M. oryzae, indicated that the MoSM1-OE lines SM1-70 and SM1-73 supported less growth of M. oryzae in inoculated leaves as compared with those in the nontransgenic WT plants (Fig. 3f), leading to a reduction of > 110 folds in relative fungal growth. Taken together, these results and observations suggest that the MoSM1-OE lines showed an improved resistance to M. oryzae. MoSM1-OE plants exhibit an improved resistance against bacterial blight disease. We next evaluated the resistance level of the MoSM1-OE lines against bacterial leaf blight disease, which is the most devastating bacterial disease in rice caused by X. oryzae pv oryzae (Xoo). The MoSM1-OE and nontransgenic WT plants at booting stage were inoculated using the leaf-clipping method with 2 different Xoo strains PXO96 and Zhe817 and the disease levels were evaluated by measuring lesion length at 15 dpi and the dynamics of bacterial growth in inoculated leaves. As shown in Fig. 4a and b, the overall bacterial blight disease severity in the MoSM1-OE lines SM1-70 and SM1-73 was lower than that in the nontrangenic WT plants after inoculation with strains PXO96 and Zhe817. At 15 dpi, the average of the lesion length on leaves of nontransgenic WT plants was 10.5 cm, while the averages of the lesion length on leaves of the MoSM1-OE lines SM1-70 and SM1-73 were 3.7 cm and 3.1 cm, respectively (Fig. 4c), leading to 65-70% of reduction in lesion length compared to the nontransgenic WT plants. Similarly, measurement of in planta bacterial growth showed that, the growth rates of Xoo strains PXO96 in inoculated leaves of the MoSM1-OE lines SM1-70 and SM1-73 were significantly lower than that in the nontransgenic WT plants, resulting in reduction of 10-100 folds at 6, 10 and 14 dpi (Fig. 4d). These results indicate that the MoSM1-OE lines also displayed an improved resistance to Xoo. MoSM1-OE plants do not alter the resistance against sheath blight disease. We also evaluated the resistance level of the MoSM1-OE lines against sheath blight disease, another most important fungal disease on rice caused by R. solani, which is completely different in infection style from M. oryzae. Both in vitro detached leaf inoculation assay and in vivo sheath inoculation assay were used to evaluate the resistance level of the MoSM1-OE lines to R. solani. In the in vitro detached leaf inoculation assays, no significant difference in lesion formed on the R. solani-inoculated leaves between the MoSM1-OE and nontransgenic WT plants was detected (Fig. 5a). Measurement of in planta fungal growth, as revealed by analyzing genomic DNA level of the G-protein β -subunit gene of R. solani with a rice Actin gene, also indicated no significant difference in fungal growth between the MoSM1-OE and nontransgenic WT plants at 3 dpi (Fig. 5b). In the in vivo sheath inoculation assays, lesions on the R. solani-inoculated sheath of the MoSM1-OE lines SM1-70 and SM1-73 were comparable in size with those on the sheath of nontransgenic WT plants at 15 dpi ( Fig. 5c and d). Collectively, these results suggest that the MoSM1-OE lines did not alter the resistance to R. solani.
MoSM1-OE plants contain elevated SA and JA levels. SA and JA are most important plant hormones that play vital roles in regulating immune responses in rice 48 . To explore the molecular mechanism responsible for the improved resistance in MoSM1-OE lines, we examined whether the MoSM1-OE plants had altered levels of endogenous SA and JA under normal condition. The SA and JA levels in the MoSM1-OE lines SM1-70 and SM1-73 were significantly higher than those in the nontransgenic WT plants, leading to increases of 70-85% for SA and 90-141% for JA ( Fig. 6a and b), respectively. These data indicate that the MoSM1-OE plants grown under normal growth condition accumulated elevated SA and JA levels.  genes were selected to compare their expression in the MoSM1-OE and nontransgenic WT plants under normal conditions. As shown in Fig. 6c, the expression levels of OsICS1, OsEDS1, OsNPR1 and OsPR1a, which are involved in the SA biosynthesis, signaling pathway and defense response 49,50 , were markedly upregulated in MoSM1-OE plants, showing increases of 2.5, 30, 2.1 and 50 folds, respectively, as compared to the levels in nontransgenic WT plants. Similarly, the expression levels of OsAOS1, OsLOX1, OsACO7 and OsJAmyb, which are involved in JA biosynthesis and signaling pathway 51,52 , were also upregulated markedly in MoSM1-OE plants, leading to increases of 2.1, 2.8, 7.8 and 3.1 folds, respectively, as compared to the levels in nontransgenic WT plants (Fig. 6d). These results indicate that the MoSM1-OE plants activated constitutively the SA and JA signaling pathways.

MoSM1-OE plants do not have penalty on drought and salt stress tolerance.
To examine whether an improved disease resistance in MoSM1-OE lines leads to penalty on other biological processes such as abiotic stress tolerance, we compared the salt and drought tolerance of the MoSM1-OE and nontransgenic WT plants at vegetative stage. In salt stress tolerance assays, no significant difference in seed germination and seedling growth, as revealed by the shoot and root length, was observed between the MoSM1-OE and nontransgenic WT seedlings on 1/2 MS medium with or without supplement of 150 mM NaCl, although the growth of the MoSM1-OE and nontransgenic WT seedlings was markedly inhibited on 1/2 MS medium containing 150 mM NaCl (Fig. 7a,b and c). In drought stress assays, the MoSM1-OE and nontransgenic WT plants showed similar growth status before drought treatment (Fig. 7d). Comparable drought symptom such as leaf rolling and wilting was observed between the MoSM1-OE and nontransgenic WT plants after drought stress treatment by withholding water for 12 days (Fig. 7e) and, no significant difference in growth phenotype (Fig. 7f) and survival rates (Fig. 7g) were observed between the MoSM1-OE and nontransgenic WT plants after re-watering regularly for 12 days. Collectively, these data demonstrate that the MoSM1-OE plants did not have adverse effect on drought and salt stress tolerance.

MoSM1 levels in transiently expressed leaves and stably expressed transgenic rice.
To explain the phenomenon that transient expression of MoSM1 in rice leaves caused significant cell death but stable expression of MoSM1 in transgenic rice did not cause visible cell death, we detected the accumulation of MoSM1 in transiently expressed and stably expressed rice leaves. As shown in Fig. 8, a MoSM1 polyclonal antibody, prepared by immunizing rabbits with a peptide CNDLTNGQAGSLGRI, detected the accumulation of MoSM1 in both transiently expressed and stably expressed rice leaves whereas no signal was observed in transiently expressed eGFP leaves or in nontransgenic WT plants. Comparatively, the level of MoSM1 accumulation in transiently expressed rice leaves was relatively higher than those in the MoSM1-OE plants and this was much evident when the MoSM1 accumulation was normalized with the anti-β -tubulin antibody hybridized β -tubulin bands (Fig. 8). Thus, it is likely that the absence of visible cell death in leaves of the MoSM1-OE plants is largely due to a relatively low level of MoSM1, as compared with the amount of MoSM1 and the significant cell death in transiently expressed leaves.

Discussion
Many members that belong to the cerato-platanin family have been characterized from filamentous fungi including M. oryzae and B. cinerea and some of them have shown to trigger immunity in a number of plant species 22,23 . We and others have shown that, transient or stable expression of MoSM1 in Arabidopsis or application of  recombinant MoMSP1 in rice activated a series of defense responses, which conferred a broad-spectrum resistance against fungal and bacterial pathogens 32,42 . The present study further shows that expression of MoSM1 in transgenic rice confers an improved resistance against blast and bacterial blight diseases, probably through modulating SA/JA signaling pathways, demonstrating that MoSM1 is a promising elicitor that can be used in improvement of rice disease resistance.
No enzymatic activity has been reported to CPs including SM1 of T. virens 40 . Although not all reported CPs can induce necrosis in plants 33,37 , some CPs including a CP from C. platani, MoSM1/MoMSP1 from M. oryzae and BcSpl1 from B. cinerea have been shown to cause phytotoxic effects on different plants. For instance, when applied at 80 μ M 25 , purified C. platani CP induced significant necrosis in tobacco leaves and caused several structural and physiological changes, i.e. plasmolysis and accumulation of phenolic compounds 34,35 . The B. cinerea BcSpl1 was found to produce a faster and stronger necrosis in tobacco leaves than the C. platani CP at a lower concentration 31 . Similar to our previous observations that transiently or stably expressed MoSM1 caused typical necrosis in Arabidopsis, tobacco and tomato leaves 42 , we also observed that transiently expressed MoSM1 in rice leaves did cause necrosis symptom and increased leakage of electrolyte ( Fig. 1e and f). The difference in phytotoxic effects of the CPs may attribute to the intrinsic differences in CPs, the different plant species used in the assays and/or the methods used for application of the CPs to tested plants 31 . Surprisingly, stably expressed MoSM1 in transgenic rice did not produce any morphological and growth change as compared to the WT plants (Fig. 2d). The difference in appearance of HR-like necrosis in transient expression of MoSM1 in rice leaves and in stable expression of MoSM1 in transgenic rice may be caused by the amounts of MoSM1 accumulated in transiently or stably expressed rice plants. We previously found that, when induced with low concentrations of DEX, MoSM1 did not produce toxic effect in MoSM1-OE Arabidopsis plants 42 . Our Western blot assays confirmed that the level of MoSM1 accumulation in transiently expressed rice leaves was relatively higher than those in the MoSM1-OE plants (Fig. 8). This is probably due to the fact of only a single copy of the transgene MoSM1 in the MoSM1-OE plants vs a quite large number of copies of the MoSM1 gene in the transiently expressed rice leaves. Thus, unlike the necrosis produced by MoSM1 in transiently expressed rice leaves (Fig. 1e), the amount of MoSM1 in the tissues of the MoSM1-OE plants is not enough to cause observable necrosis in the MoSM1-OE plants. The lack of necrosis in the MoSM1-OE plants features an advantage for the use of these stable transgenic lines in development of novel rice varieties with improved disease resistance as significant necrosis may not be acceptable for practical use.
Most of CPs including MoSM1/MoMSP1 from M. oryzae have been shown to induce defense responses related to HR. Several HR characteristic features were observed for some CPs, e.g. the induction of ROS and autofluorescence, appearance of cytoplasm shrinkage and activation of defense gene expression 25,31,37,41,44 . The present and our previous studies 42 showed that MoSM1 can induce visible HR in both Arabidopsis and rice leaves, as verified by the appearance of rapid cell death and the induction of electrolyte leakage, and expression of defense genes (Fig. 1e,f and g). Similar results were also obtained with recombinant MoMSP1, which triggered cell death and induced H 2 O 2 production in both suspension-cultured cells and rice leaves 32 . On the other hand, it was shown that MoMSP1 is secreted into culture medium in vitro and into apoplast in planta during infection and that a high level of accumulation of MoMSP1 in rice apoplast is required for activation of defense response 30,32 . We observed that transient expression of MoSM1 ΔSP , which lacked the SP, in N. benthamiana leaves abolished HR (Fig. 1a), which is similar to the previous observation that deletion of SP led to loss of the activity of MoMSP1 to up-regulate expression of defense genes in rice leaves 32 . These data support the hypothesis that MoSM1/MoMSP1, along with other CPs studied so far, indeed function as typical PAMPs that can trigger PTI and that the secretion of MoSM1 into apoplast is critical for its PAMP activity in triggering immune response. Furthermore, our subcellular localization assays revealed that MoSM1 was targeted to the plasma membrane when transiently expressed in N. benthamiana leaves (Fig. 1c), which is similar to the observation that B. cinerea BcSpl1 associates with the plasma membrane of cells in tomato and tobacco 29 . It was found recently that Arabidopsis BAK1, a critical component of the PAMP signaling complex 53 , plays a role in the induction of necrosis by BcSpl1 31 . Similar to the common knowledge that some well-known PAMPs such as FLS2 for flg22 were found to be recognized by the extracellular regions of the plasma membrane surface-localized pattern recognition receptors 54 , a yet unknown receptor that recognizes MoSM1/MoMSP1 may be localized on the plasma membrane of plant cells. Therefore, it can be speculated that, during infection or when expressed in rice, MoSM1/MoSP1 is secreted through a SP-guided process into apoplast, and is then recognized by a plasma membrane-localized receptor(s), followed by the initiation of HR and PTI. This speculation explains, at least partially, the importance of MoSM1/MoSP1 secretion into apoplast in triggering the immunity 30,32 and the feature of the MoSM1/MoSP1 plasma membrane localization (Fig. 1c). Further identification of MoSM1/MoMSP1 receptor or receptor complex and its associated components will broaden our understanding of the molecular mechanism of MoSM1/MoMSP1, especially the clarification of the earliest events and downstream signaling pathway involved in MoSM1/MoMSP1-induced HR and immunity.
Several PAMPs such as flg22 and lipopolysaccharides have been shown to induce a systemic defense response with the characteristics of SAR. Indeed, a number of CPs from different filamentous fungi, e.g. SM1, SM2 and Epl1 from the biocontrol agents T. virens and T. atroviride 37,40,45,47 as well as BcSpl1 from B. cinerea and a CP from C. platani 29,44 , were shown to induce both local and systemic disease resistance in host and nonhost plants against different life style pathogens. In the present study, we found that overexpression of MoSM1 in transgenic rice plants confers an improved resistance against two strains representing difference races of M. oryzae and Xoo, as revealed by reduced disease severity and suppressed in planta pathogen growth (Figs 3 and 4). Similarly, application of recombinant MoMSP1 at a sub-lethal concentration was also found to potentiate resistance against M. oryzae, with no significant cell death on rice leaves 32 . However, the MoSM1-OE plants seem to be more effective to blast disease than to bacterial blight disease, as the blast disease severity and the in planta growth of M. oryzae were almost completely suppressed (Fig. 3) while the bacterial blight disease severity and the in planta Xoo growth were suppressed by ~70% (Fig. 4). In addition, MoSM1 was previously found to induce resistance in Arabidopsis against two necrotrophic fungi (i.e. B. cinerea and A. brassicicola) and a (hemi)biotrophic bacterial pathogen (i.e. P. syringae pv. tomato DC3000) 42 . These data indicate that MoSM1, like other CPs from both beneficial and pathogenic fungi 44,47 , confers a broad-spectrum resistance against pathogens with different life styles in host and nonhost plants, providing a great potential for application of MoSM1 in genetic improvement of disease resistance in other crop plants. The MoSM1-OE plants did not show a decreased resistance against sheath blight disease, although they did not gain an increased resistance (Fig. 5). This may be due to that the molecular mechanism responsible for defense response against R. solani differs from those for defense response against M. oryzae and Xoo. Similar results were also observed in OsWAK25-OE plants, which show resistance to both M. oryzae and Xoo but display increased susceptibility to R. solani 55 .
SA and JA signalings are effective against both biotrophic and necrotrophic pathogens in rice 48,56 . Given that MoSM1 functions as a PAMP, MoSM1 is hypothesized to be recognized by an unknown receptor, resulting in initiation of signal transduction pathway to activate defense response in rice. We found that the MoSM1-OE plants contained elevated levels of both SA and JA, two critical defense signaling hormones 48 , and upregulated expression of SA-and JA-related biosynthesis and signaling genes (Fig. 6), indicating that the MoSM1 plants constitutively activate the SA-and JA-mediated defense signaling pathways. This is similar to the function of MoSM1 in transgenic Arabidopsis plants 42 . Generally, rice plants contain high levels of endogenous SA that are weakly responsive to pathogen infection 57,58 ; however, SA can act in both of OsNPR1-and OsWRKY45-dependent pathways in rice defense response 56 . The improved resistance in MoSM1-OE plants may be due to the upregulated expression of SA signaling regulatory genes such as OsNPR1, whose overexpression in rice leads to constitutive activation of defense response 59 , rather than the elevated level of SA. The involvement of SA signaling in defense response of the MoSM1-OE plants is partially supported by the upregulated expression of OsPR1s and OsPR10 (Figs 1g and 6c), defense genes that are activated by the SA signaling pathway 60,61 , by transiently or stably expressed MoSM1 in rice plants. This is in agreement with the observations that a CP from C. platani in Arabidopsis and BcSpl1 from B. cinerea in tobacco were shown to function through SA signaling 44,46 but differs from the action of MoMSP1 in rice, in which SA suppressed MoMSP1-induced H 2 O 2 production and cell death in suspension cultured rice cells 32 . Whereas exogenous SA applied at a high concentration (e.g. 5 mM) suppressed MoMSP1-induced cell death and H 2 O 2 production in suspension cultured cells 32 , MoSM1 in transgenic rice plants promoted endogenous SA level (Fig. 6a). Alternatively, SA is differentially involved in MoSM1/MoMSP1-induced cell death and defense response in rice. On the other hand, manipulation of the JA biosynthesis by overexpression of OsAOS2, encoding allene oxide synthase that is a key enzyme involved in JA biosynthetic pathway, increased the endogenous JA level, which led to activation of defense gene expression and enhancement of resistance to M. oryzae 62 . MoSM1 activates JA signaling in the MoSM1-OE plants by upregulating the expression of JA biosynthesis and signaling genes ( Fig. 6b and d), similar to the observation that JA enhanced MoMSP1-induced cell death in rice 32 . It is therefore likely that JA signaling is involved in MoSM1/ MoMSP1-regulated defense response. This is inconsistent with a recent finding that function of the CP from C. platani triggers SA-and ethylene (ET)-signaling pathways, but not the JA-signaling pathway, as revealed by the expression pattern of marker genes 44 . Overall, the CP-triggered defense signaling pathways may be complicated and depend on specific CPs and/or the tested plant species.
The effect of CPs on plant growth and development as well as on the abiotic stress tolerance is not known yet. In the present study, we were unable to observe any morphological and developmental changes at seedling and adult stages and grain yield in the MoSM1-OE plants in comparison to the WT plants (Fig. 2), indicating that MoSM1 does not have effect on plant growth and development. On the other hand, we also did not observe distinguishable difference in salt and drought tolerance between the MoSM1-OE and WT plants (Fig. 7), indicating that overexpression of MoSM1 in transgenic rice did not disturb the abiotic stress tolerance.

Conclusion
We demonstrate in the present study that MoSM1, a member of cerato-platanin family from M. oryzae, can activate defense response in rice and that the MoSM1-OE plants gain an improved broad-spectrum resistance against fungal blast and bacterial blight, two most devastating diseases of rice worldwide. We also show that the enhanced resistance in the MoSM1-OE plants is regulated through a combination of SA-and JA-mediated signaling pathways. Furthermore, the MoSM1-OE plants do not have any deleterious effect on drought and salt tolerance as well as on important agronomic traits. Taken together, our data demonstrate that overexpression of MoSM1 in rice confers an improved broad-spectrum resistance against blast and bacterial blight diseases without any penalty on abiotic stress tolerance and grain yield, providing a promising potential for application of MoSM1 in improvement of disease resistance in other crop plants. Further identification of putative MoSM1 receptor(s), characterization of the MoSM1-triggered signaling pathway and genome-wide expression profiling of MoSM1-regulated genes will greatly help to elucidate the molecular mechanism by which MoSM1 regulates plant immunity. Transient expression assays. Infiltration suspension of agrobacteria carrying pINDEX3::MoSM1 or empty pINDEX3::00 construct was prepared as described previously 42 . To construct the SP-deletion vector, the fragment of MoSM1 without SP was amplified using primers MoSM1-Δ SP-F and MoSM1-Δ SP-R (Supplementary Table S1), digested with EcoRI/SalI, and then inserted into pINDEX3 vector 63 , yielding pIN-DEX3::MoSM1 ΔSP . Agrobacteria carrying pINDEX3::MoSM1, pINDEX3::MoSM1 ΔSP or empty pINDEX3::00 were infiltrated into leaves of 8-week-old N. benthamiana or 5-week-old rice plants using a 1-mL syringe without a needle. The agroinfiltrated plants were foliar sprayed with 1 μ M DEX solution 24 h later to induce the expression of MoSM1. Leaf samples were collected at indicated time points after DEX treatment and used for analyses of electrolyte leakage and defense-related gene expression. For measurement of electrolyte leakage, 200 mg leaf samples were put into 5 ml distilled water for 3 hr at room temperature and the initial conductivity was measured using a DDS-IIAT digital conductivity meter. The samples were then boiled for 15 min to disrupt plant tissue, and the total conductivity was measured using the same meter. Percentage of electrolyte leakage was calculated as (initial conductivity before boiling)/(total conductivity after boiling) × 100%. Experiments were repeated three times using at least 6 leaves from 6 individual plants per treatment in each experiment.
Western blotting assays. Samples from transiently expressed leaves of N. benthamiana and rice plants or from leaves of the stable transgenic rice plants were ground in 200 μ l extraction buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM DTT, 0.1% Triton X-100, 1× protease inhibitor cocktail, and 1 mM PMSF) and then added 100 μ l loading buffer. After boiling for 5 min, the extracts were centrifuged at 10000 g for 10 min at 4 °C and 20 μ l of the supernatant were separated on a 15% SDS-PAGE gel, followed by transferring onto PVDF membrane via wet electroblotting. Detection of GFP in transiently expressed N. benthamiana leaves was performed using a mouse monoclonal GFP antibody (Huaan Company, Hangzhou, China) whereas detection of MoSM1 in transiently expressed rice leaves or stably expressed in transgenic rice plants was carried out using a MoSM1 polyclonal antibody, prepared by immunizing rabbits with peptide CNDLTNGQAGSLGRI-KLH conjugate (GenScript Company, Nanjing, China). In detection of MoSM1 with MoSM1 polyclonal antibody, a β -tubulin monoclonal antibody anti-β -tubulin (Abmart, Shanghai, China), was used to detect endogenous tubulin as an internal equal-loading control.
Generation and characterization of MoSM1-OE lines. The full coding sequence of MoSM1 including SP was amplified using primers MoSM1-OE-F and MoSM1-OE-R (Supplementary Table S1), digested with BamHI/SmaI, and inserted into the binary vector pCoUm under the control of a maize ubiquitin promoter, yielding plasmid pCoUm-Ubi::MoSM1. The resulting construct was introduced into A. tumefaciens strain EHA105 and transformed into rice calli of cv. Xiushui11 through the Agrobacterium-mediated transformation protocol. The obtained MoSM1-overexpressing transgenic lines were screened by planting seeds on 1/2 MS medium containing 50 μ g/L hygromycin (Hgr) and lines showing 3:1 (Hgr-resistant: Hgr-susceptible) segregation were selected as candidates of single-copy transgenic lines. Individual lines of T3 generations showing 100% Hgr resistance on selective medium were selected as homozygous lines. To further characterize the single-copy transgenic lines, approximately 50 μ g of genomic DNA was digested completely with EcoRI, separated by agarose gel electrophoresis, and transferred onto a Hybond-N + nylon membrane (Amersham Biosciences, Little Chalfont, UK).
A 589 bp fragment of the HptII gene was amplified using a pair of primers HptII-Probe-F and HptII-Probe-R (Supplementary Table S1) and labelled with DIG by the random priming method using a DIG High Prime DNA Labeling and Detection kit I (Roche Diagnostics, Shanghai, China). Prehybridization, hybridization and detection were performed according to the manufacturer's recommendations. Homozygous lines with single-copy of transgene were used for all experiments.
Disease assays. M. oryzae strains 85-14 (race ZB1) and 97-220 (race E3) were grown on CM medium at 25 °C for 10 days and spores were collected to prepare inoculum. Two-week-old seedlings at three leaf stages were inoculated by spraying with 5 × 10 5 spores/mL spore suspension containing 0.02% Tween-20 65 . The inoculated plants were kept in the dark for 24 hr at 25 °C with 100% relative humidity and then moved to the growth room with same condition. Disease severity was evaluated using a standard international 0-9 scale (0 = resistant and 9 = susceptible) from at least 30 plants for each line at 7 dpi 65 . Fungal growth in inoculated leaves was estimated using qRT-PCR 66 by analyzing the genomic DNA level of a M. oryzae 28S rDNA and relative fungal growth was presented as ratios obtained by comparison of the genomic fungal 28 S rRNA gene levels with a rice OsEF1 genomic DNA levels. Trypan blue staining was carried out as previously described 67 . Briefly, leaf samples were submerged in trypan blue solution (2.5 mg/mL trypan blue, 25% lactic acid, 23% phenol and 25% glycerol in H 2 O), incubated for 10 min at 70 °C, heated for 2 min and then stained overnight. After de-staining in chloral hydrate solution (25 g in 10 mL of H 2 O) for 3 days, the samples were equilibrated with 70% glycerol before microscopy examination.
Xoo strains PXO96 and Zhe817 were grown in NA broth at 28 °C with shaking and the bacteria were collected by centrifugation and diluted in distilled water to OD 600 = 0.8. Rice plants at booting stage were inoculated using the leaf-clipping method 68 and the inoculated plants were moved to a greenhouse under environmental conditions at 30 °C in daytime/25 °C in night with natural sunlight. Disease phenotype and lesion length were recorded at 15 dpi. For measurement of bacterial growth in inoculated leaves, leaf samples were collected from three individual plants of each line, surface sterilized in 70% ethanol for 10 sec, homogenized in 200 μ L H 2 O and serially diluted to proper concentration. A hundred microliters of the final dilutions were plated on NA agar medium and incubated at 28 °C for 24 hr before counting the colony-forming units 69 .
Two methods were used to evaluate the resistance of the rice lines against sheath blight disease. For whole plant inoculation assays, R. solani strain GD-118 was grown in potato dextrose broth on a 28 °C shaker for 3 days and the mycelia were collected by centrifugation. Eight-week-old plants were inoculated by attaching mycelial balls (8 mm in diameter) inside the sheath with aluminium foil 70 and disease phenotype was investigated at 15 days after inoculation. For detached leaf inoculation assays, R. solani strain GD-118 was grown in potato dextrose agar medium and inoculation was performed by placing mycelial agar plugs onto detached leaves. Disease phenotype was examined at 3 days after inoculation. For measurement of fungal growth in inoculated leaves, the mycelial agar plugs inoculated on the detached leaves were removed and genomic DNA was extracted from the leaf tissues. The fungal amount was determined using qRT-PCR by analyzing the genomic DNA level of a R. solani G-protein β -subunit gene 71 and relative fungal growth was presented as ratios of the fungal G-protein β -subunit gene levels with a rice Actin gene level.
Quantification of SA and JA content. Quantification of SA and JA was carried out according to the method described previously 72 . Briefly, 300 mg leaf tissues were ground in liquid nitrogen with mortar and pestle, extracted with 2 mL 80% methanol and kept overnight at − 20 °C. After centrifugation at 4 °C for 10 min at 12,000 × g, the supernatant was collected, dried under nitrogen gas and then dissolved in 0.5 mL 2% ammonia solution. The collected extracts were loaded on preconditioned Oasis MAX SPE columns (Waters Corp., Milford, MA) and sequentially washed with ammonia solution (2%) and 2 mL methanol. SA and JA in the column were eluted with 4 mL methanol containing 1% formic acid. The eluent was dried under nitrogen gas and finally dissolved in 0.5 mL 80% methanol. Quantification of SA and JA was performed by a HPLC-Triple quadrupole liquid chromatography-mass spectrometry system (Model 1290/6460, Aglient Technologies, Santa Clara, CA) with stable isotope-labeled SA and JA as standards. The experiments were repeated three times and data were averaged.
qRT-PCR analyses of gene expression. Total RNA was extracted from frozen leaf samples using TRIzol reagent (Invitrogen, Shanghai, China) and treated with RNase-free DNase (TaKaRa, Dalian, China). First-strand cDNA was synthesized from 1 μ g of total RNA using AMV reverse transcriptase (TaKaRa, Dalian, China) according to the manufacturer's recommendations. The qRT-PCR was performed in a CFX96 real-time PCR system (BioRad, Hercules, CA, USA) using Fast Essential DNA Green Master kit (Roche Diagnostics, Shanghai, China). The rice Actin gene was used as an internal control to normalize the data and relative expression levels of genes of interest were calculated using the 2 −ΔΔCT method. Gene-specific primers used in qRT-PCR are listed in Supplementary Table S1. Three independent biological samples were performed.
Drought and salt tolerance assays. For drought tolerance assays, 4-week-old seedlings were grown in the same pot and were subjected to drought stress treatment by withholding watering for 12 days, followed by recovery with normal water supply for another 7 days. Plants with green and healthy young leaves at 12 days after re-watering were regarded as survivals and survival rate was calculated as percentage of the survived plants in total treated plants. At least 10 plants for each line were used in each experiment. For salt tolerance assays, 100 seeds were germinated on 1/2 MS medium supplemented with or without 150 mM NaCl and allowed the seedling growth on the same medium. At 6 days after germination under 28 °C/25 °C (day/night) with a 12 hr photoperiod, shoot height and fresh weight of seedlings of each line were measured. The experiments were repeated three times and data were averaged.
Evaluation of grain yield. Rice plants were grown in a restricted greenhouse with natural sunlight and agronomic traits were evaluated. Harvested grains were air-dried and store at room temperature for 1 month. Twenty random panicles from each line were chosen for measuring the panicle weight and total grains per panicle according to the conventional methods. Seed setting rate was calculated as percentage of fully filled seeds in total seeds of a single panicle. Experiments were repeated three times and data were averaged.