Diagnostic kit for rice blight resistance

Blight-resistant rice lines are the most effective solution for bacterial blight, caused by Xanthomonas oryzae pv. oryzae (Xoo). Key resistance mechanisms involve SWEET genes as susceptibility factors. Bacterial transcription activator-like (TAL) effectors bind to effector-binding elements (EBEs) in SWEET gene promoters and induce SWEET genes. EBE variants that cannot be recognized by TAL effectors abrogate induction, causing resistance. Here we describe a diagnostic kit to enable analysis of bacterial blight in the field and identification of suitable resistant lines. Specifically, we include a SWEET promoter database, RT–PCR primers for detecting SWEET induction, engineered reporter rice lines to visualize SWEET protein accumulation and knock-out rice lines to identify virulence mechanisms in bacterial isolates. We also developed CRISPR–Cas9 genome-edited Kitaake rice to evaluate the efficacy of EBE mutations in resistance, software to predict the optimal resistance gene set for a specific geographic region, and two resistant ‘mega’ rice lines that will empower farmers to plant lines that are most likely to resist rice blight.

R ice is the most important food crop in the world, but biotic and abiotic stresses can reduce yields. Xoo causes the rice disease bacterial blight, which results in substantial yield losses across Asia 1 . Bacterial blight has been classed as the most serious bacterial disease of rice 2 , with the highest social impact. Epidemics severely affect smallholders: about 70% of farms in India are ~ 0.39 ha (ref. 3 ), roughly similar to 80% of farms in sub-Saharan Africa ( < 2 ha). Bacterial blight is a major problem in India that has increased in severity year on year since 2000 (ref. 4 ). Increased severity has been attributed in part to climate change (increased rainfall and higher cyclone frequency) 4 . Modeling indicates that climate change might result in an increased effect of Xoo-mediated disease in Africa, where rice production is rising, and researchers have suggested that losses due to rice blight would eventually exceed those caused by rice blast 5 . Moreover, the United States is concerned about the potential for introduction of Asian or African bacterial blight strains; currently, only low-virulence strains that lack TAL effector genes are present in the United States 6 . Xoo is on the US Select Agent list as a potential bioterrorism agent 7 .
Genetic resistance to disease reduces the need for pesticides 8 . The best known plant resistance (R) genes encode proteins that interact with specific pathogen effectors and endow plants with dominant resistance to pathogens. More than 40 R genes for bacterial blight have been identified; a few of these have been cloned, and several have been associated with modular TAL effectors 8 . Complete genome sequences of Xoo enable identification of TAL effectors, information that can then be used together with the TAL effector recognition code 9,10 to identify the respective TAL EBEs throughout the rice genome. In some cases, however, resistance is recessive and caused by effectors that target susceptibility factors, as in the case of host sucrose transporter (SWEET) genes. Xoo produces TAL effectors that ectopically induce the SWEET genes that make rice susceptible to infection and disease 1,12 .
DNA polymorphisms in EBEs can prevent TAL effectors from binding to target promoters, and the respective rice lines with altered EBEs in a SWEET promoter are resistant to bacterial blight 11,12 . The first identified SWEET resistance variant was xa13, a naturally occurring promoter variant in SWEET11 (refs. 13,14 ). xa13 resistance is recessive and is used in rice breeding programs, as the xa13 promoter variants do not negatively affect yield 4,15 . The TAL effector PthXo1, which is present in several Xoo strains (PXO99 and PXO71), binds an EBE in the SWEET11 promoter 11 . The resistant rice line IRBB13 (xa13) carries a 38-bp deletion and a 252-bp insertion in the SWEET11 promoter, which abrogate binding of the EBE by PthXo1. Other resistant varieties carrying xa13 include Chinsurah Boro2, Tepa1, Aus274, AC-19-1-1, Long Grain (35023), Kalimekri77-5, Long Grain (64950) and BJ1 (ref. 13 ). xa13 resistance can be overcome by Xoo strains that produce alternative TAL effectors that bind to the promoter of a different SWEET paralog. For example, SWEET14 promoter elements can be bound by PthXo3 or AvrXa7 (ref. 16 ; Fig. 1a). Naturally occurring recessive resistance has also been identified for SWEET13 (xa25) 17,18 and SWEET14 (xa41) 19 . Altogether, six EBEs in three SWEET genes that are targeted by naturally occurring TAL effectors (PthXo1 targeting SWEET11; PthXo2 and variants targeting SWEET13; PthXo3, AvrXa7, TalC and TalF targeting SWEET14) have been characterized. The TAL effectors target three of the five clade III SWEET genes; the other two SWEET genes can function as susceptibility (S) genes when artificially induced, but no Xoo strains targeting these genes have been identified 20 . SWEET genes in other clades do not function as S genes 20 .
The pool of naturally occurring EBE variants is small, meaning that extensive breeding efforts are needed to identify and introduce such resistance variants into elite rice varieties. TALEN-or CRISPRbased genome editing has been used to engineer EBE variants of SWEET11 and SWEET14 in the japonica variety Kitaake 18,19,21 . Additionally, five EBEs have been targeted, alone or in combination, in Kitaake, IR64 and Ciherang-Sub1 (refs. 18,[21][22][23][24], generating resistance to a collection of 94 Xoo strains of Asian origin 24 . These R-gene-containing lines can now be deployed, individually or in combination, for broad resistance. However, Xoo will likely evolve TAL effector variants that can bind to other promoter sequences to enable plant infection.
To enable the strategic deployment of R genes to block novel virulent strains of Xoo when they emerge, we developed the SWEET R kit 1.0, a multicomponent tool set that integrates pathogen diagnostics, genetic resistance and customizable resistance line deployment to overcome bacterial blight.

Results
Content of the SWEET R kit 1.0. The SWEET R kit 1.0 (Supplementary Table 1) contains a SWEET promoter database (SWEETpDB), SWEET up primers for detecting mRNA accumulation for each SWEET gene, three SWEET acc -rice tester lines that report spatial SWEET protein accumulation during infection using translational reporter fusions 26 , single-and combined-knockout SWEET ko mutants to identify which SWEET genes are required for susceptibility to individual Xoo strains, SWEETp R tester lines in the R-gene-free Kitaake background to evaluate whether particular EBE variants (or combinations thereof) are sufficient for resistance against an Xoo isolate, SWEET PathoTracer, a decision tool 26   diagnostic surveys and population information to develop the most effective and most durable resistance for a specific region, and a total of 32 transgene-free EBE-edited lines of two mega varieties: IR64 and Ciherang-Sub1 (ref. 24 ).
SWEET promoter database. To predict resistance against, or susceptibility to, any Xoo strain with TAL effector(s) that target SWEET promoters, the EBE sequences in SWEET promoters must be characterized. We analyzed EBE variations for SWEET11, SWEET13 and SWEET14 in 4,726 accessions (in the first 400 bp of the SWEET promoters) 27 Fig. 4) and in infected leaves in a strain-specific manner (Fig. 1b). Quantitative RT-PCR (qRT-PCR) showed that SWEET13 mRNA levels were highest among the five clade III SWEET genes in uninfected leaves, with lower levels of SWEET14. SWEET13 may have a role in phloem loading, as shown for its maize homologs 30 . SWEET11 mRNA levels were very low in leaves, consistent with its roles in seed filling 26 . Validated primer pairs are available in the kit for testing new Xoo isolates.
SWEET acc -rice tester lines for SWEET protein accumulation. First, we engineered transcriptional reporter lines to visualize SWEET RNA accumulation by histochemical β-glucuronidase (GUS) staining. However, reporter lines for all three promoters had nonspecific reporter activity ( Supplementary Fig. 5). Previously, similar observations were reported for AtSWEET11 and AtSWEET12 in Arabidopsis; only translational SWEET reporters had cell-specific reporter activity 12 . To monitor SWEET protein accumulation, translational promoter reporter lines were engineered. The constructs included a 2-kb fragment from the SWEET promoter, the coding region of the SWEET gene including introns and a GUSPlus reporter (Supplementary Figs. [6][7][8]. Consistent with the absolute mRNA levels measured by qRT-PCR, SWEET13 and SWEET14 translational promoter reporter activities were detected in uninfected leaves, whereas SWEET11 translational promoter fusion lines showed no detectable GUS activity ( Fig. 2a-d). SWEET13 and SWEET14 translational reporter lines showed vein-specific protein accumulation, consistent with the roles of SWEET13 and SWEET14 in phloem loading ( Fig. 2b-d). Three reporter lines for SWEET11, SWEET13 and SWEET14, named SWEET acc -rice tester lines, were each infected with five Xoo strains (PXO61, PXO71, PXO86, PXO99 and PXO112), which are known to induce specific SWEET genes. The Xoo strain ME2, lacking TAL effectors for SWEET induction 14 , was used as a control and did not trigger SWEET acc reporter activity. box limits indicate the 25th and 75th percentiles as determined by R software; and whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles. c, One-thousand-grain weight of greenhouse-grown Kitaake, sweet13-1 and sweet13-2 (mean ± s.e.m.; n = 4 biological replicates). The experiment was repeated at least three independent times with similar results. Center lines show medians; box limits indicate the 25th and 75th percentiles as determined by R software; and whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles. No significant differences (P = 0.051 for sweet13-1 and P = 0.758 for sweet13-2) were identified by Student's t-test. d, Phenotypes of wild type and the sweet13;sweet14 double knockout grown in the greenhouse. No significant differences were identified. e, Length of lesions at 14 days after inoculation (DAI) caused by ME2 (negative control), PXO99 (positive control) and AXO1947 on single-, double-and triple-knockout (sweet11, sweet13 and sweet14) mutants relative to Kitaake wild type (mean ± s.e.m.; n = 10 inoculated leaves). The experiment was independently repeated twice with similar results. The difference observed for AXO1947 virulence between sweet14 and sweet11;14 in a single experiment was not significant when compared over a larger number of experiments ( Supplementary Fig. 14).
SWEET ko knock-out lines to detect Xoo SWEET targets. Tools that could identify the SWEET genes targeted by a specific Xoo strain, without previous knowledge of the targeted EBE, would inform efforts to engineer resistance. This is especially important because variant Xoo strains can target either different promoter elements or other clade III SWEET paralogs. Moreover, it is conceivable that promoter-edited lines or variants could impair yield, if the promoter variations affect normal gene function in uninfected plants.
Knowledge of the specific defects is thus important to judge possible negative effects of such mutations on plant performance. SWEETmutant lines could also provide insight into the role(s) of SWEET genes in resistance and yield. To diagnose which SWEET was targeted by any Xoo strain and to predict the yield effects of alterations in SWEET genes, knockout mutants were created for four of the five described clade III SWEET genes (SWEET11, SWEET13, SWEET14 and SWEET15) using CRISPR-Cas9 (Supplementary Fig. 10 and ref. 26 ). In all cases, lines containing frameshift mutations in the sequence corresponding to transmembrane domain I (TM I) were identified (Supplementary Table 6). Frameshifts that result in early termination should create nonfunctional transporters. For example, premature termination in the last transmembrane domain, TM VII, of OsSWEET11 results in defective transporters 12 . sweet13 mutant lines had reduced SWEET13 mRNA levels, which often occur concurrently with early termination (Fig. 3b). Knockout mutants for SWEET11 and SWEET15 were reported to have defects in seed filling 26 . Although SWEET11 and SWEET15 have important roles in seed filling in rice, widely used promoter variants, such as the resistance-conferring xa13 variant (SWEET11), do not seem to affect yield 4,15 .
Because maize ZmSWEET13 paralogs have roles in phloem loading 30 , the role of rice SWEET13 was investigated. SWEET13 is the most highly expressed SWEET gene in rice leaves, and the encoded protein localizes to the plasma membrane and, in common with SWEET14, transports sucrose (Supplementary Figs. 4 and 11) 12,18 . Analysis of GUS reporter fusions showed that SWEET13 accumulates in the phloem (Fig. 2b,d). SWEET14 also accumulated in the phloem but had substantially lower mRNA levels in leaves than SWEET13 ( Fig. 2c; transcript levels not shown). Nevertheless, CRISPR-Cas9 sweet13 and sweet14 knockout mutant lines did not show detectable growth or yield defects under greenhouse conditions (Fig. 3a,c and Supplementary Fig. 12), nor were obvious differences observed in a single-season field experiment (based on visual inspection during the growth period and after harvest). To identify potential compensatory activity from other SWEET genes in the knockout mutants, the expression levels of other sucrose-transporting SWEET genes were analyzed. Only the weakly expressed SWEET14 and SWEET15, and none of the other clade III SWEET genes, showed substantial increases in mRNA accumulation in the leaf blade of sweet13 knockout lines ( Supplementary Fig. 13). To test whether upregulation of SWEET14 could compensate for the loss of SWEET13 and thereby restore apoplasmic phloem loading, sweet14 single-knockout and sweet13;sweet14 double-knockout lines were generated. Double mutants did not show obvious growth differences relative to controls in the greenhouse (Fig. 3d). Because mutant lines had no clear defects in plant growth or yield, EBE-edited lines in which the normal promoter function of SWEET13 and SWEET14 is affected are not hypothesized to have yield penalties. Further, our data indicate that apoplasmic phloem loading in rice, in contrast to maize and Arabidopsis, either is not crucial to plant performance or does not entirely depend on SWEET function 12,30 .
Knockout lines can serve as diagnostic tools for testing whether Xoo strains require specific SWEET genes. The knockout mutants proved to be valuable tools for characterizing the virulence of a collection of 95 different Xoo strains with diverse geographic origins 24 . In this analysis, we observed that an African strain, AXO1947, which contains the effector TalC and can induce SWEET14, but does not induce SWEET13, was still able to infect a Kitaake mutant line edited in the TalC EBE present in the SWEET14 promoter 24 . Although technically weakly virulent on lines carrying TalC EBE variants, AXO1947 showed substantially reduced virulence in the quintuplemutant promoter lines, which are likely sufficiently resistant in field conditions 24 . The quintuple-mutant line was moderately resistant, with lesion lengths of 5-7 cm upon infection by AXO1947, as compared to 18-25 cm in controls. A systematic screen for resistance using sweet13 and sweet14 single-knockout and sweet13;sweet14 double-knockout mutants showed that AXO1947 lost some virulence in sweet14-knockout lines but was unable to infect sweet13;sweet14 double mutants ( Fig. 3e and Supplementary Fig. 14). These data demonstrate the utility of knockout lines for testing resistance (Fig. 3e, Table 1 and Supplementary  Tables 8 and 9).

SWEET PathoTracer visualization. Geographic Information
System (GIS)-based platforms that incorporate pathogen monitoring and resistance profiles of rice varieties are useful for management of local disease outbreaks 26 . We integrated the near-isogenic IR64 and Ciherang-Sub1 lines into the PathoTracer platform (http://webapps. irri.org/pathotracer/index.html). PathoTracer displays the predicted involvement of SWEET11, SWEET13 and SWEET14 on the basis of the Xoo population that is present in geographic regions and suggests effective edited variants for planting in the next season. For example, a dataset containing analyses about the ability of Xoo strains to infect irri.org/pathotracer/index.html) is an online repository that integrates genotypic and phenotypic pathogen data with the resistance profiles of rice accessions to support strategic deployment of varieties in the region. Tester-strain-based prediction of SWEET targets is provided here for SWEET14, using Xoo populations collected from 1972 to 2012 in Laguna, a disease-endemic area in the Philippines (n = 1,294 isolates). A screenshot of the full interface with the same map is shown in Supplementary Fig. 15.
the R-gene near-isogenic IRBB lines 32 was compared to the proportion of endemic strains from an area of the Philippines that might activate SWEET14 (Fig. 4 and Supplementary Fig. 15). On the basis of this information, 47% of the strains in the Xoo population are predicted to be controlled by one or more of the SWEET14 EBE variants. By planting the recommended varieties, farmers can minimize the risk of infection and the resulting yield losses in the next season.
Genome-edited Xoo-resistant mega variety lines. Mega rice varieties are defined as varieties planted on more than 1 million hectares. Although genome-edited bacterial-blight-resistant SWEETp R Kitaake lines can be used by breeders, SWEET R mega variety lines would reduce the need for further breeding efforts. This is of particular relevance because breeding efforts are more extensive in this context, owing to the recessive nature of SWEET-based resistance. CRISPR-Cas9 was used to edit five of the six EBE sites in the three SWEET promoters in widely used indica rice mega variety IR64 and in Ciherang-Sub1, a new flooding-tolerant elite line 24,33,34 . We generated edited lines with alterations in single or multiple EBEs. Together, 32 Cas9-free lines were produced, encompassing 35 single variations in the three SWEET promoters. Agronomic assessment and pathogenicity trials validated resistance against single or multiple Xoo strains ( Fig. 5 and Supplementary

Discussion
Genetically narrow germplasm and extensive mono-cropping are two hallmarks of modern agriculture that favor disease and its spread 35 . Genome editing could provide efficient tools for rapid engineering of pathogen resistance in cropping lines. However, detection of pathogen strains, their virulence factors and cropping line susceptibilities will be crucial if we are to effectively reduce the impact of plant diseases. Here we present a diagnostic kit to enable breeders, crop management teams and farmers to reduce the effect of bacterial blight on rice yields worldwide. Genome editing can create large pools of R-gene variants, providing new ways to implement and manage long-term genetic resistance. Diseases that use TAL effectors are excellent candidates for reducing yield loss by engineering resistant plant lines, because TAL effector binding depends on highly conserved and short cognate EBE sequences present in host S-gene promoters. One crucial finding in the fight against bacterial blight is that Xoo strains use different TAL effectors (eight effectors have been characterized to date) to directly target five EBEs in three SWEET gene promoters 24 . Pathogen-mediated induction of one of three SWEET genes is sufficient to cause disease. Extant Xoo strains harbor just eight SWEET-targeting TAL effectors, indicating that effectors with novel EBE recognition motifs may not evolve quickly. Therefore, genome editing could provide durable as well as broad bacterial blight resistance.
The success of genome-edited rice lines resistant to bacterial blight will be threatened by the emergence of pathogenic strains that have adapted to recessive resistance. The resistance endowed by SWEET promoter variants that prevent TAL effector binding can be overcome by the emergence of pathogenic Xoo strains. For example, resistance based on the natural promoter variant xa13 can be overcome by Xoo strains that produce alternative TAL effectors, such as PthXo3 or AvrXa7, to induce either the same or different SWEET genes 16 . SWEET promoter variants will likely need to be deployed in combination with other R genes, and Xoo strain emergence may, therefore, depend on multiple genetic alterations.
The underlying key to durable resistance may be preventing bacteria from multiplying in monocultures 36 . Genome editing enabled us to establish a portfolio of recessive resistance variants with different promoter sequences that are available as R-gene variants to use when a novel pathogen emerges. The portfolio allows us to swap between R-gene variants. Tracking pathogens, using diagnostics and GIS, will likely also help. Because the edited lines are isogenic, they can be deployed as mixtures to reduce pathogen spread 37,38 , using mosaic crop plotting in which different R genes are deployed in adjacent fields or R genes are rotated over different seasons 38 . Here we provide a diagnostic tool set to accompany a suite of resistant rice lines 24 . The SWEET R kit comprises a SWEETpDB and SWEET up primers to detect which SWEET is targeted by an Xoo isolate. If suitable expertise is not available locally, leaf samples can be preserved and analyzed in central test labs 39 . We also include four SWEET acc -rice tester lines containing full-gene translational reporters that detect a particular isolate that induces a SWEET gene 20,25 ; this set includes a line for analysis of SWEET15 induction, although so far no naturally occurring Xoo strains induce this SWEET. The kit further includes single-and combined-knockout SWEET ko mutants to identify which SWEET genes are required for susceptibility and SWEETp R tester lines in Kitaake. We added a component to a web-based decision tool 26 named PathoTracer, which uses population information, to aid development of the most effective and most durable R-gene combination for a specific region through breeding. We also include 31 transgene-free EBE-edited lines in the two mega varieties IR64 and Ciherang-Sub1 (Supplementary Table 10) 24 . Figure 6 provides a roadmap for using the kit for customized deployment of the SWEET R lines.
Recently, a new Gibson cloning technology for TAL effectors was developed that enables determination of the full TAL effector complement of a new isolate followed by bioinformatic target prediction, a tool that will be integrated into the kit 40 .
In the future, we will use our diagnostic kit to track newly emerging Xoo strains, use them to challenge R-gene variants and engineer promoter variants if TAL effectors evolve. Although only three of five clade III SWEET genes have been targeted 24 , we have also Reactions of IR64 SWEET-promoter-edited lines to three representative Xoo strains (data from ref. 24 ). Lesion lengths were measured at 14 DAI with strains PXO99A, PXO339 and PXO86. Infections were carried out at the maximum tillering stage by inoculating 3-6 leaf samples using a leaf clipping. Four replicate experiments with two plants each were performed per strain (four replicates per strain, two plants per replicate (n = 8)) and scored for 3-6 inoculated leaf samples per plant. The experiment was repeated three times independently.
developed sweet15 knockout mutants that have no obvious growth defects in greenhouse and field conditions 24 . We plan to improve our kit by adding sweet knockout lines to cover all clade III SWEET genes. There is one report of an Xoo strain that does not induce SWEET genes 41 , but it is challenging to prove a lack of induction, as exemplified here for AXO1947.
SWEET-based resistance also occurs for bacterial blights of cotton and cassava, so our approach of providing a disease diagnosis and management kit together with a suite of genome-edited lines may prove useful for other pathogens that decimate crops such as cassava and cotton 42,43 .

Online content
Any methods, additional references, Nature Research reporting summaries, source data, statements of code and data availability and associated accession codes are available at https://doi.org/10.1038/ s41587-019-0268-y.  Fig. 6 | Customized deployment of SWEET R lines with the help of the SWEET R kit 1.0. Farmers with Xoo-infected rice fields will send samples to local breeders or pathologists, who will isolate respective Xoo strains. Pathologists will then identify induced and critically important SWEET genes using SWEET up primers for mRNA accumulation and SWEET knock-out (SWEET KO ) mutants. After validation with SWEET EBE-edited Kitaake lines (SWEETp R ), pathologists will identify the optimal SWEET R line, which is then provided to local breeders. In parallel, labs will isolate Xoo DNA from infected leaves, identify TAL effectors (the TALeome) and predict SWEET targets using SWEETpDB. PathoTracer provides additional region-specific recommendations for deployment of SWEET R variants.

RNA isolation and transcript analyses.
Total RNA was isolated using Spectrum Plant Total RNA kits (Sigma) or TRIzol (Invitrogen), and first-strand cDNA was synthesized using a QuantiTect Reverse Transcription kit (Qiagen). qRT-PCR was performed using a LightCycler 480 (Roche), with the 2 −ΔCt method for relative quantification 44 . Primers for SWEET11, SWEET13, SWEET14 and UBI1 are listed in Supplementary Table 5.

DNA constructs and plant transformation. Generation of GUS reporter constructs.
The method for constructing pSWEET11:gSWEET11-GUSplus was described 26 . In short, the promoter and coding region were fused in frame to the GUSplus coding sequence with the NOS terminator, and the resulting constructs were PCR amplified and inserted into pC1300intC (GenBank AF294978.1) 45 . For tissue specificity analysis, a 4,354-bp genomic clone of SWEET13 containing 1,919 bp of the 5′ region upstream of the translational start codon (ATG) and 2,435 bp of the entire coding region without a stop codon and a 4,365-bp genomic clone of SWEET14 containing 2,176 bp of the 5′ upstream region and 2,189 bp of the entire coding region without a stop codon were amplified by PCR using Kitaake genomic DNA as a template (primers are listed in Supplementary Table 5). The PCR amplicons were subcloned into pJET2.1/blunt (Thermo Fisher), and resulting inserts were confirmed by DNA sequencing. Sequences are shown in Supplementary Table 11. The cloned fragments digested with XbaI and KpnI for SWEET13 or AvrII and XmaI for SWEET14 were subsequently inserted in front of the GUSplus coding sequence of a promoterless GUSplus coding vector 26 restricted with XbaI and KpnI for SWEET13 and XbaI and XmaI for SWEET14. The resulting pSWEET13:gSWEET13-GUSplus and pSWEET14:gSWEET14-GUSplus constructs were used to transform Oryza sativa L. ssp. japonica Kitaake. Nine independent events were obtained for pSWEET13:gSWEET13-GUSplus and pSWEET14:gSWEET14-GUSplus. Whereas GUS activity levels were different in the independent lines, the GUS patterns were similar.
Plant materials and growth conditions. Kitaake wild-type and mutant plants were grown either in field conditions (20 individual plants per genotype, paddy field in summer 2016, Carnegie) or in greenhouses under long-day conditions of 14-h day/10-h night at 28-30 °C and 50% relative humidity, with a light intensity of 500-1,000 µmol m -2 s -1 . Independent transformants for the GUS fusions were characterized (transcriptional GUS fusions: SWEET11, n = 8; SWEET13, n = 8; SWEET14, n = 15; translational GUS fusions: SWEET11, n = 18; SWEET13, n = 9; SWEET14, n = 29). All independent transformants showed similar tissue specificity by histochemical staining. Analyses were performed for a minimum of six plants per infection of each Xoo strain. Three independent experiments were performed for specific induction in a growth chamber (12-h day at 28 °C/12-h night at 25 °C, 80% relative humidity). IR64 lines were grown in a small-scale field environment in a screenhouse (28 ± 7 °C during the day/23 ± 4 °C at night; 80-85% relative humidity) at IRRI.
Xoo strains and infection protocols. The Xoo strains collected from different geographic regions were reported 24 . Plasmid-containing Xoo strains were obtained through electroporation of competent cells 46 with respective pHM1derived plasmids (e.g., pHM1/ZWpthXo1 for the pthXo1 gene) 47 . For infection experiments, bacterial inocula were prepared by growing bacterial cells on tryptone sucrose plates with appropriate antibiotics. Cells were scraped from the plates and resuspended in sterile distilled water at OD 600 ~ 0.5. (i) Leaf clipping 48 : the two youngest fully expanded leaves of 4-to 5-week-old rice plants were clipped about 1-2 cm from the tip with scissor blades that were immersed in bacteria immediately before clipping. Five plants were used for inoculation of each strain (ten leaves in total). Lesion length (distance from the cut to the leading edge of (gray) symptoms) was measured for each inoculated leaf at 12-14 DAI. The mean lesion length of ten leaves was used for each treatment. The Tukey test for analysis after ANOVA was used for statistical analyses. Leaf tissues were mounted in laminating film and photographed under white light. (ii) Syringe infiltration: bacterial suspensions were infiltrated into leaves from the bottom by pressing the opening of a needleless syringe to the leaf. Leaf fragments with inoculated spots were cut off 48 h after inoculation for RNA extraction and GUS staining analysis.
Statistical analysis. Data were plotted using BoxPlotR (http://shiny.chemgrid. org/boxplotr/) or are presented as mean ± s.e.m as specified in respective tables or figures. One-sided ANOVA was conducted on measurements. The Tukey honestly significant difference test was used after ANOVA pairwise tests for significance, which was set at P t-test was used. Exact P values, the statistical test used and sample number n can be found in figure legends or graphs.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated Our web collection on statistics for biologists contains articles on many of the points above.

Software and code
Policy information about availability of computer code Data collection

Data analysis
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A list of figures that have associated raw data -A description of any restrictions on data availability Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.