Transcription activator-like (TAL) effectors targeting OsSWEET genes enhance virulence on diverse rice (Oryza sativa) varieties when expressed individually in a TAL effector-deficient strain of Xanthomonas oryzae
Department of Bioagricultural Sciences and Pest Management and Program in Plant Molecular Biology, Colorado State University, Fort Collins, CO, 80523-1177, USA
Institut de Recherche pour le Développement, UMR Résistance des Plantes aux Bioagresseurs, IRD-CIRAD-UM2, Montpellier Cedex 5, France
Genomes of the rice (Oryza sativa) xylem and mesophyll pathogens Xanthomonas oryzae pv. oryzae (Xoo) and pv. oryzicola (Xoc) encode numerous secreted transcription factors called transcription activator-like (TAL) effectors. In a few studied rice varieties, some of these contribute to virulence by activating corresponding host susceptibility genes. Some activate disease resistance genes. The roles of X. oryzae TAL effectors in diverse rice backgrounds, however, are poorly understood.
Xoo TAL effectors that promote infection by activating SWEET sucrose transporter genes were expressed in TAL effector-deficient X. oryzae strain X11-5A, and assessed in 21 rice varieties. Some were also tested in Xoc on variety Nipponbare. Several Xoc TAL effectors were tested in X11-5A on four rice varieties.
Xoo TAL effectors enhanced X11-5A virulence on most varieties, but to varying extents depending on the effector and variety. SWEET genes were activated in all tested varieties, but increased virulence did not correlate with activation level. SWEET activators also enhanced Xoc virulence on Nipponbare. Xoc TAL effectors did not alter X11-5A virulence.
SWEET-targeting TAL effectors contribute broadly and non-tissue-specifically to virulence in rice, and their function is affected by host differences besides target sequences. Further, the utility of X11-5A for characterizing individual TAL effectors in rice was established.
Pathogenic Xanthomonas oryzae causes two important bacterial diseases of rice (Oryza sativa). Xanthomonas oryzae pv. oryzae (Xoo) enters through leaf hydathodes to invade xylem vessels and cause bacterial leaf blight, while Xanthomonas oryzae pv. oryzicola (Xoc) enters through stomata to colonize the leaf mesophyll and cause bacterial leaf streak. Xanthomonas oryzae is diverse, with distinct phylogenetic clades comprising Xoo from Asia, Xoo from Africa, and Xoc from Asia and Africa (Triplett et al., 2011; Hajri et al., 2012). Strains of weakly pathogenic X. oryzae, with no pathovar designation, were also isolated in the USA from rice showing mild symptoms of bacterial blight (Jones et al., 1989). These strains were recently placed in a clade distinct from all other subgroups, and it was hypothesized that they descended from a common ancestor of Xoc and Xoo (Triplett et al., 2011).
Asian and African Xoo and Xoc genomes contain multiple members of a gene family encoding transcription activator-like (TAL) effectors. TAL effectors enter host cells via the bacterial type III secretion system, and turn on specific host genes by binding to effector-specific sites in the host genome. Some TAL effectors contribute to virulence by activating host genes that enhance susceptibility, and some, not mutually exclusively, function as avirulence factors by activating resistance genes (Boch & Bonas, 2010; Bogdanove et al., 2010). TAL effector target specificity is determined by a central region of the effector that is composed of conserved, 33–34 amino acid repeats. Two variable amino acids at positions 12 and 13 in each repeat, called the repeat-variable diresidue (RVD), together specify a nucleotide in the target so that the number and composition of RVDs define the length and nucleotide sequence of the target (Boch et al., 2009; Moscou & Bogdanove, 2009).
Several Xoo TAL effectors that contribute to virulence have been identified and their relevant targets in rice characterized. PthXo6 is a TAL effector cloned from Philippine Xoo strain PXO99A that contributes moderately to virulence by activating the transcription factor TFX1 (Sugio et al., 2007). TAL effector PthXo7 from that strain contributes to virulence specifically in a variety that carries a mutation in the gene for the gamma subunit of general transcription factor TFIIA (Iyer & McCouch, 2004). This mutation confers resistance to many Xoo strains, hypothetically by reducing the ability of TAL effectors to recruit the host transcriptional machinery (Iyer & McCouch, 2004). PthXo7 induces expression of a TFIIAγ paralog that presumably restores full TAL effector activity (Sugio et al., 2007). Xoo TAL effectors with more substantial effects on virulence uniformly activate members of the OsSWEET family of sucrose transporter genes (Chen et al., 2010) to promote disease. PthXo1 from strain PXO99A activates OsSWEET11 (also called Os8N3) and TAL effectors PthXo3 from Japanese strain MAFF31101 and AvrXa7 from Philippine strain PXO86 activate OsSWEET14 (also called Os11N3) from overlapping binding sites (Yang et al., 2006; Antony et al., 2010; Romer et al., 2010). The TalC protein from the African Xoo strain BAI3 activates OsSWEET14 from a binding site distinct from those of AvrXa7 and PthXo3 (Yu et al., 2011).
In addition to the TFIIAγ mutation mentioned above (also known as the xa5 resistance gene), which confers resistance in a general way, plants have evolved highly specific mechanisms to defend against pathogens that deploy TAL effectors. Resistance genes Xa27 in rice and Bs3 in pepper (Capsicum annuum) are specifically transcriptionally activated by TAL effectors AvrXa27 and AvrBs3, respectively, and block disease progression by triggering a localized host cell death (Gu et al., 2005; Romer et al., 2007). Polymorphisms that provide resistance by destroying a TAL effector binding site in a major susceptibility gene promoter and preventing its activation also occur. For example, a large insertion in the promoter of OsSWEET11 prevents PthXo1-mediated susceptibility and constitutes the recessive xa13 gene for bacterial blight resistance (Yang et al., 2006). Xa7 confers a strong hypersensitive reaction (HR), a programmed cell death response, when inoculated with bacteria harboring the OsSWEET-activating TAL effector avrXa7 (Hopkins et al., 1992), but, to date, the resistance gene has not been cloned.
Despite these major advances in our understanding, the targets and roles of most TAL effectors from X. oryzae (and other Xanthomonas species) are still unknown. Sequenced X. oryzae genomes harbor between eight and 26 individual TAL effector genes (White et al., 2009; Bogdanove et al., 2011). The strain from which most TAL effectors have been characterized, PXO99A, has 11 additional yet uncharacterized TAL effectors. Furthermore, no studies of the virulence functions of TAL effectors from any Xoc strain have yet been reported. TAL effector functional studies are complicated by the possibility of functional redundancies and by the quantitative nature of their contributions to virulence (Yang et al., 1996; Bai et al., 2000). Also, as exemplified by PthXo7 discussed above, functions of some TAL effectors may only be apparent in certain host genotypes. Indeed, studies of X. oryzae TAL effector function to date have been conducted on only a relatively narrow range of rice varieties, including Azucena, Taipei, Nipponbare, and isogenic lines derived from IR24.
The weakly virulent US X. oryzae strains contain no TAL effector genes (Ryba-White & Leach, 1995; Triplett et al., 2011). Because these strains are otherwise highly genetically similar to African and Asian Xoo and Xoc strains (Triplett et al., 2011), they offer the opportunity to study the virulence (or avirulence) properties of TAL effectors individually. We used one of these, strain X11-5A, to ask whether the major virulence roles of Xoo TAL effectors that target OsSWEET11 and OsSWEET14 are conserved across a diverse collection of rice varieties, and whether any of these TAL effectors triggers defenses in one or more of these varieties that could reveal novel resistance loci. We also asked whether the virulence function of SWEET-targeting Xoo TAL effectors is specific to the xylem, or if they might enhance virulence of Xoc in the leaf mesophyll as well. Finally, we tested whether selected Xoc TAL effectors affect the behavior of X11-5A.
Materials and Methods
Strains, plasmids and rice varieties
The bacterial strains and plasmids used for this study are described Table 1. Escherichia coli cells were grown in Luria–Bertani (LB) medium at 37°C. Xanthomonas oryzae (Ishiyama 1922) Swings et al., 1990 was grown in peptone-sucrose agar (PSA) medium at 28°C (Karganilla et al., 1973). Antibiotics were used at the following concentrations: tetracycline, 2 μg l−1; gentamycin, 20 μg l−1; ampicillin, 50 μg l−1.
Table 1. Bacterial strains and plasmids used in this study
pKEB31 containing tal1c gene of Xoc BLS256 without repeat-containing SphI fragment; Tcr
pKEB31 containing tal1c gene of Xoc BLS256; Tcr
pKEB-tal1c with repeat-containing SphI fragment replaced by that of avrXa7 gene of Xoo PXO86; Tcr
pKEB-tal1c with repeat-containing SphI fragment replaced by that of pthXo1 gene of Xoo PXO99A; Tcr
pKEB-tal1c with repeat-containing BamHI fragment replaced by that of tal2a gene of Xoc BLS256; Tcr
pKEB-tal1c with repeat-containing BamHI fragment replaced by that of tal2g gene of Xoc BLS256; Tcr
pKEB-tal1c with repeat-containing BamHI fragment replaced by that of tal8 gene of Xoc BLS256; Tcr
pSKXI-2 derivative containing talC gene of Xoo BAI3; Gmr
B. Szurek, pers. comm.
Varieties of rice (Oryza sativa L.) used included a set of 20 diverse rice lines (the OryzaSNP set) that represent the genotypic and phenotypic diversity of cultivated rice (McNally et al., 2009). We also included Kitaake, an early-flowering japonica rice from northern Japan that is widely used in rice transformation studies (Manosalva et al., 2009). In addition, rice isogenic lines IRBB7 (with Xa7) and IR24 (the recurrent parent background) were used to assess delivery of AvrXa7 by X11-5A. Finally, UXO-3, a transgenic Kitaake line containing the Xa27 resistance gene driven by a custom promoter that is responsive to several TAL effectors from Xoo and Xoc, including PthXo1 and Tal2g (Hummel et al., 2012), was used to test translocation of those TAL effectors.
Cloning of TAL effector genes
Except for talC, the repeat region of each TAL effector gene was first cloned as a BamHI or SphI fragment into pCS466, a derivative of the Gateway entry vector pCR8-GW (Invitrogen) that contains a truncated form of the Xoc BLS256 tal1c gene, from which the SphI fragment that comprises the repeat region had been removed. SphI fragments were cloned into the single SphI site in the truncated tal1c gene. BamHI fragments replaced the BamHI fragment of the truncated gene. The reconstituted genes were then transferred to the broad host-range destination vector pKEB31 (Cermak et al., 2011; Addgene plasmid 31224; www.addgene.org) using Gateway LR Clonase (Invitrogen) according to the manufacturer's instructions, for expression in Xanthomonas (Table 1). In pKEB31, expression is driven by the lac promoter, which is constitutive in Xanthomonas.
Clones of avrXa7 and pthXo1 repeat regions were obtained from F. White (Kansas State University) and B. Yang (Iowa State University). Repeat region clones of tal1c, tal2a, tal2g, and tal8 were selected from a TAL effector gene repeat region library generated by digesting the genomic DNA of Xoc BLS256 with BamHI, gel-purifying all fragments ranging from 2.5 to 5 kb, ligating them to pBlueScript SK(-) (Agilent Technologies) linearized with BamHI, and then screening by PCR and sequencing. For talC, a complete gene clone in broad host range vector pSKXI was obtained from B. Szurek (pers. comm.).
Transformation of X. oryzae X11-5A
Plasmids were introduced into X. oryzae strain X11-5A by electroporation (Choi & Leach, 1994). For each transformation event, at least six single colonies were selected, purified twice on selective media, and confirmed by PCR. Templates for PCR amplification were derived by re-suspending cells from single colonies in sterile water and boiling for 6 min at 95°C. Each PCR reaction was prepared with 20.6 μl of water, 0.4 μl of each primer at 10 μmol l−1, 0.5 mmol l−1 of dNTPs, 2.5 μl buffer ×10 (Thermopol Buffer; New England BioLabs, Ipswich, MA, USA), and 3.5 μl of denatured cells. PCR amplification was performed in a GeneAmp® PCR System 2700 (Applied Biosystems, Carlsbad, CA, USA) using the following program: an initial denaturation step of 1 min at 95°C followed by 32 cycles of 30s at 95°C, 45s at 62.5°C, and 1 min at 72°C, and ending with a final elongation step of 5 min at 72°C. The primers used for amplification were P270 (5′-GCCAAGTCCTGCCCGCG-3′) and P271 (5′-CCTCCAGGGCGCGTGC-3′), which correspond to the N-terminus of talC. The 695-bp PCR products and an exACTGene 100-bp DNA ladder (ThermoFisher Scientific, Waltham, MA, USA) were separated in a 1% agarose gel and visualized by staining with ethidium bromide. Images were captured using the Gene Genius digital image capture system and the GeneSnap Software (SynGene, Cambridge, UK) version 7.12.
To confirm plasmid presence and stability, plasmid DNA was extracted from transformed X11-5A strains after several rounds of culture and after re-isolation from inoculated plant tissue using the PureYield™Plasmid Miniprep System (Promega, Madison, WI, USA). Then 200 ng of plasmid DNA was transformed into E. coli DH5α competent cells using a heat-shock protocol. Plasmids were extracted and digested with SphI, and the products were visualized as described in the previous paragraph.
Xanthomonas oryzae strains were grown on PSA medium supplemented with antibiotics for 24 h at 30°C, then re-suspended in sterile water at an optical density (OD600) of 0.2 (c. 108 cfu ml−1). Inoculations were conducted in growth chambers maintained at 28°C : 24°C, 12 h: 12 h, day : night with 85% relative humidity. For leaf clip assays, the two youngest, fully expanded leaves on each tiller of 4-wk-old rice plants were inoculated as described previously (Kauffman et al., 1973). Lesion length (centimeters) was measured 15 d post inoculation (dpi). Each strain was assessed on a set of 21 rice lines, including the OryzaSNP lines (McNally et al., 2009) and Kitaake. For each bacterial treatment, 10 plants per accession were evaluated, with at least two fully expanded leaves inoculated per plant. For syringe infiltration assays, strains were introduced with a needleless syringe into the intercellular spaces of rice leaves from 2-wk-old plants at concentrations of 108 cfu ml−1 as described previously (Reimers & Leach, 1991), HR was scored at 72 h after inoculation (hai), and lesions were measured at 10 d. Experiments were repeated three independent times.
Multiplication of Xoo X11-5A and X11-5A(pthXo1) was measured in planta at three time-points (0, 8 and 15 d after inoculation (dai)). Rice varieties Azucena and Nipponbare were inoculated by leaf clipping of 4-wk-old plants. The top 20 cm of each leaf was cut into four 5-cm sections, designated A, B, C, and D, starting from the point of inoculation ‘A’. The leaf pieces were ground in 1 ml of sterile water, and bacterial numbers were assessed in serial dilutions that were spread onto PSA agar plates supplemented with antibiotics. The plates were incubated at 28°C until single colonies could be counted. The experiment was repeated three times.
To visually differentiate lesions from resistance responses, the terminal 5 cm of each inoculated leaf was cleared by suspending overnight in a solution of 95% ethanol and 5% glycerol. A resistance response was visualized as a blackening of the inoculation site and rated as (+). Disease lesions that appear as light brown after clearing were rated as (−) while lesions with a light blackening and light browning were rated as (+/−) (Supporting Information Fig. S1).
Sequence, RT-PCR, and qPCR analysis of target promoters
Published primers (Romer et al., 2010) were used to amplify the sequences upstream of OsSWEET11 and OsSWEET14 from genomic DNA extracted from rice varieties Dular, Minghui, Nipponbare, Moroberekan, and FR13A (McNally et al., 2009). Fifty-microliter reactions containing 25 ng of genomic DNA, 1.5 mM MgCl2, 0.2 mM dNTPs, 0.5 μM for each primer, 1× PCR buffer (Life Technologies, Grand Island, NY, USA) and 2 units Taq DNA polymerase were cycled in the following conditions: 1 min at 94°C followed by 28 cycles of 30 s at 94°C, 30 s at 55°C, and 1 min at 72°C. PCR products were sequenced at the Colorado State University proteomics and metabolomics facility and deposited in GenBank (accession numbers JQ968614–JQ968623).
For reverse transcriptase (RT)-PCR and quantitative real-time RT-PCR (qPCR), RNA was extracted from 10 cm of infiltrated leaf tissue collected at 48 hai using the Spectrum™ Plant Total RNA Kit according to the manufacturer's recommendations (Sigma-Aldrich, St. Louis, MO, USA). The RNA was quantified by spectrophotometer (NanoDrop, ND-2000; ThermoFisher Scientific,Waltham, MA, USA).
Two micrograms of total RNA was treated with DNaseI (Invitrogen) and reverse-transcribed using SuperScript III (Invitrogen). For the semiquantitative RT-PCR analysis, 1 μl (5%) of the RT-PCR reaction was used as a template with Phire Hot Start II DNA polymerase (Thermo Scientific) together with gene-specifics oligos (OsSweet11F, CAAGCCCCACCAGGTCAAGGT; OsSweet11R, TTGTGCACGCCGAGGATCG; OsSweet14F, GCCTTCGCCTTTGGTCTCCT; OsSweet14R, ATCTGGATGACGCCGAAGGA; Actin AS1, CAAAATTCACGTCCGTACATCG; ActinS1, AAACTTTGTTACGTCGCGGC) as follows: 30 s at 98°C; 23 cycles of 10 s at 98°C, 5 s at 60°C, and 10 s at 72°C.
qPCR was performed using an iCycler Thermo Cycler (Bio-Rad) with 50 ng of total RNA as a template for cDNA synthesis and PCR amplification using the iScript One-Step RT-PCR kit with SYBR Green (Bio-Rad) according to the manufacturer's protocol. Gene-specific primers for OsSWEET11 (P941, ATGGCTAACCCGGCGGTCACCCT; P942, GCGTTGATGGTCAGCAGCGGCCT), OsSWEET14 (P943, GGCGACCGCCGCATCGTGGTT; P944, GCCCAGCACGTTGGGAAGAGCG), and rice Actin-6 for normalization (EU215044; P787, CCGGTGGATCTTCATGCTTACCTGG; P788, CGACGAGTCTTCTGGCGAAACTGC) were used for amplification. A minimum of three independent biological replicates, each with three qPCR technical replicates, was tested for each treatment. The method (Livak & Schmittgen, 2001) was used to quantify expression of OsSWEET11 or OsSWEET14 transcript for each treatment relative to abundance in mock-inoculated tissue.
X11-5A expresses and delivers heterologous TAL effector AvrXa7
The sequenced US X. oryzae strain X11-5A, which does not contain TAL effectors, was selected for heterologous TAL effector expression. To confirm the ability of this strain to express and deliver TAL effectors, we transformed it with a broad host range, avrXa7 expression construct, pKEB31-avrXa7. The resulting strain, X11-5A(avrXa7), was inoculated into rice variety IRBB7, which contains the cognate resistance gene Xa7, and into IR24, which does not. X11-5A(avrXa7) triggered an Xa7-dependent HR (Fig. 1). Thus, X11-5A functionally expresses and delivers TAL effectors from a plasmid vector.
OsSWEET-targeting TAL effectors increase X. oryzae X11-5A virulence to varying extents in different rice varieties
TAL effectors from Asian and African Xoo strains known to target members of the OsSWEET family (AvrXa7, PthXo1, and TalC) were selected for expression in X. oryzae X11-5A. Xanthomonas oryzae X11-5A elicits weak or delayed symptoms on susceptible rice, and we hypothesized that the absence of TAL effectors in the strain may partially account for the low virulence relative to Xoo and Xoc (Triplett et al., 2011). To determine whether the addition of TAL effectors could increase the virulence of X11-5A in diverse host backgrounds, the OryzaSNP set of 20 rice varieties (McNally et al., 2009) plus Kitaake were clip-inoculated with X11-5A and X11-5A containing avrXa7, pthXo1 and talC. X11-5A alone caused blight symptoms (lesion of ≥ 1 cm) on nine of the 21 varieties, six of which are in the O. sativa japonica group (Fig. 2, Table S1). Eight varieties showed increased susceptibility in response to all three TAL effectors, and five of these varieties are phylogenetically clustered in the japonica group (Fig. 3). Another seven varieties did not show a TAL effector-dependent increase in susceptibility (Fig. 3 and Table S1). Varieties Cypress, Aswina, and Pokkali were more susceptible in the presence of PthXo1 but not other TAL effectors. Dular susceptibility was affected by PthXo1 and AvrXa7, while Kitaake susceptibility was affected by PthXo1 and TalC. In all, 14 of 21 varieties showed increased susceptibility in response to at least one TAL effector, including five of the varieties that did not develop lesions (> 1 cm) in response to X11-5A alone.
We hypothesized that TAL effectors might be ineffective in some varieties if a strong resistance response were triggered by X. oryzae X11-5A. To discern resistance from short lesions, inoculated leaf tips were cleared with ethanol at 15 dpi and rated as dark brown (DB), suggestive of an HR, or beige, indicative of a susceptible response (Fig. S1). We also tested X11-5A(avrXa7) by using leaf infiltration assays, and observed an HR at 48 h post inoculation in those lines also exhibiting the DB response (Fig. S1). When TAL effectors did not contribute to virulence, the dark brown tips were observed in many cases; of the eight varieties in which TAL effectors made no virulence difference, five had a clearly blackened leaf tip in response to X11-5A (Table S1). On the other hand, seven of the eight varieties in which all TAL effectors increased virulence did not exhibit this resistance response. However, TAL effectors were ineffective in several varieties despite the lack of any obvious resistance response. In the varieties Minghui and Dular, the DB resistance response was observed in the absence, but not in the presence, of one or more TAL effectors (Table S1). AvrXa7 reduced lesion length and induced the DB response in the varieties Cypress and Sadu-cho, suggesting that these lines harbor resistance mediated by Xa7 activity. AvrXa7 also caused a reduction in lesion size in Pokkali plants without inducing leaf browning (Table S1). TalC also reduced lesion length slightly in Pokkali. Among those varieties responding to TAL effectors, the relative degree to which each TAL effector affected virulence varied (Table S1). Even though the TAL effectors each target one of two host genes in the same OsSWEET family, there was variation among the effects of different TAL effectors on the same varieties. For example, on some varieties, PthXo1 caused the largest increase in lesion size among the TAL effectors, while on others, TalC consistently caused a greater increase (Table S1). These experiments demonstrate that, while OsSWEET-targeting TAL effectors increase the virulence of X11-5A on many rice varieties, the degree of increase caused by each TAL effector varies among varieties.
Increased X11-5A virulence conferred by Xoo TAL effectors is associated with increased multiplication and movement
X11-5A(pthXo1) caused a substantial increase in lesion length on some varieties, such as Azucena, compared with the wild-type X11-5A, but no increase in lesion length on other varieties, such as Nipponbare (Fig. 2). To determine the effect of PthXo1 on bacterial population growth and movement, colony counts were determined from four 5-cm sections of clipped leaves (sections A, B, C, and D) at 0, 8, and 14 dpi. On Azucena, there was no difference in population growth between X11-5A and X11-5A(pthXo1) in the clipped section of leaf (Azucena leaf section A; Fig. 4). However, X11-5A(pthXo1) numbers were significantly greater than those of the wild type in leaf section B on days 8 and 14, and only X11-5A(pthXo1) spread to leaf sections C and D (Fig. 4). In Nipponbare, there was no significant difference between the population growths of X11-5A and X11-5A(pthXo1) in leaf section A, and neither strain was detected beyond this section (Fig. 4). Together, these results show that X11-5A bacterial spread and virulence are increased by pthXo1 in Azucena, but not in Nipponbare rice.
TAL effector binding sites in targeted OsSWEET promoter sequences are conserved and the target genes are activated in diverse rice varieties
The three TAL effectors that affect virulence activate transcription of two members of a family of sucrose transporters in rice. PthXo1 binds to the promoter of OsSWEET11, and AvrXa7 and TalC target the OsSWEET14 promoter. RT-PCR confirmed that both promoters are activated in leaves of Azucena infiltrated with X. oryzae X11-5A expressing the appropriate TAL effector (Fig. S2). TAL effectors similarly activated transcription in Nipponbare leaves, although there was no TAL effector-mediated increase in lesion development or bacterial multiplication in this variety.
We next tested whether the differential effectiveness of TAL effectors in some other varieties could arise from variations in the promoter TAL effector binding site or from differences in transcriptional activation levels. We sequenced roughly 500 bp upstream of the OsSWEET14 and OsSWEET11 coding sequences in five varieties with different patterns of TAL effector-mediated symptoms, including the indica variety Minghui, japonica varieties Nipponbare and Moroberekan, and aus varieties Dular and FR13A. Except for a single mutation in the Moroberekan AvrXa7 binding site, among these varieties, there were no other differences in the known binding sites for AvrXa7, PthXo1 or TalC (accession numbers JQ968614–JQ968623). The single difference in the Moroberekan AvrXa7 binding site is a G to A substitution at a position corresponding to an ‘NN’ RVD in AvrXa7; this RVD has dual specificity for G and A (Boch et al., 2009), so this substitution can be expected to be inconsequential. Thus, barring differences in epigenetic marks, there should be no difference in promoter binding efficiency among the varieties. Quantitative RT-PCR was performed to determine relative levels of TAL effector-dependent gene activation among four of these varieties (Minghui, Moroberekan, Dular and FR13A). Each TAL effector up-regulated its corresponding SWEET gene target strongly (> 50 fold vs mock inoculation) in each variety (Fig. 5). However, there was no observed relationship between lesion length increase and relative fold-change in target expression. For example, OsSWEET14 is up-regulated to a similar degree by AvrXa7 and TalC in Dular even though these effectors cause different increases in lesion length, and the effectors cause similar increases in lesion length in Minghui and Moroberekan despite differences in OsSWEET14 activation levels in that variety. Thus, the differences in TAL effector-mediated virulence levels are probably not attributable to differences in target activation levels.
OsSWEET-targeting Xoo TAL effectors enhance Xoc virulence
Although the US X. oryzae strains cause weak bacterial blight symptoms on rice, we previously found that US strains are phylogenetically distant from both Xoo and Xoc strains (Triplett et al., 2011). Therefore, it is not clear whether US strains are more like Xoo, and normally invade the vascular system, or more like Xoc, and colonize the spaces between leaf mesophyll cells. This raised the question of whether the contribution of the Xoo TAL effectors to X11-5A virulence might be tissue specific, selectively facilitating invasion of the xylem. To test this possibility, we assayed the effect of avrXa7 and pthXo1 on the virulence of Xoc BLS256, by leaf syringe infiltration typically used for quantitative measures of bacterial leaf streak susceptibility, and separately by leaf clip inoculation, used for bacterial blight. Both avrXa7 and pthXo1 significantly enhanced the lengths of lesions caused by Xoc BLS256 following syringe inoculation; however, they did not render BLS256 capable of causing bacterial blight symptoms following clip inoculation (Fig. 6).
Selected Xoc TAL effectors do not affect virulence of X11-5A in four rice varieties
We next asked whether X11-5A behavior could be modified by selected TAL effectors from Xoc. Preliminary mutagenic studies implicated Xoc effectors Tal2a, Tal2g, and Tal1c from strain BLS256 as potential virulence factors (R. A. Cernadas, L. Wang, and A. J. Bogdanove, unpublished). We transformed X11-5A with gene expression constructs corresponding to these effectors and syringe-infiltrated the resulting strains into leaves of rice varieties Nipponbare, Lemont, Kitaake and Azucena. Over 11 d, watersoaking localized to the inoculated spot was the only symptom that developed, and no differences were observed between the wild-type X11-5A strain and the transformants carrying the Xoc TAL effector genes, despite the fact that Xoc strain BLS256 developed lesions > 2 cm on all four varieties (Kitaake and Nipponbare results are shown in Fig. 7). Furthermore, no effects on X11-5A virulence were observed after clip inoculation, although Xoo strain PXO145 induced lesions of > 10 cm (Kitaake and Nipponbare results are shown in Fig. 7). The Xoc TAL effector constructs are identical to the Xoo TAL effector constructs except for the repeats, and would therefore be expected to be delivered by X11-5A. Nonetheless, to examine the possibility that the lack of effect of the Xoc TAL effectors on X11-5A virulence was a result of lack of delivery, we took advantage of transgenic Kitaake rice line UXO-3, which mounts an HR in response to several Xoo and Xoc TAL effectors, including PthXo1 and Tal2g (Hummel et al., 2012), to test whether X11-5A is indeed able to deliver the Xoc effector Tal2g. X11-5A expressing tal2g, pthXo1, a central repeat domain deletion control (tal1c ΔCRR), or no tal gene, as well as Xoc BLS256, was inoculated into UXO-3 plants by syringe infiltration (Fig. 7). Only X11-5A(tal2g), X11-5A(pthXo1), and Xoc BLS256 triggered the resistance gene-mediated HR. Thus, X11-5A expresses and delivers the Xoc effector Tal2g effectively.
Xanthomonas TAL effectors induce plant genes that promote disease, or, in some genotypes, genes that provide disease resistance. Understanding their roles in diverse plant genetic backgrounds will help to define key plant disease susceptibility factors, and may identify novel sources of resistance. In this study, we used a strain from the US clade of X. oryzae as a TAL effector-free platform for characterizing the effects of individual TAL effectors on diverse rice varieties. We found that three Xoo TAL effectors targeting the OsSWEET family of sucrose transporters conferred increased virulence to the weakly pathogenic US X. oryzae strain X11-5A on a majority of rice varieties tested. Importantly, the plant genetic background affected the level of virulence enhancement by these TAL effectors, despite sequence conservation of the targets. Variation did not correlate with quantitative differences in the fold-change of OsSWEET expression.
The increase in plant susceptibility to X11-5A that was conferred by the Xoo TAL effectors supports our previous hypothesis that the absence of TAL effectors is a major factor limiting the virulence of US X. oryzae strains (Triplett et al., 2011), and confirms that the OsSWEET family is a critical set of bacterial disease susceptibility genes in rice. Evidence suggests that SWEET family members are functionally equivalent; SWEET14-activating TAL effectors can rescue the SWEET11 activation deficiency of a pthXo1 mutant of Xoo strain PXO99A, and PthXo1 can rescue the SWEET14 activation deficiency of a talC mutant of Xoo strain BAI3 (Antony et al., 2010; Yu et al., 2011). Our data (Fig. 4) show that PthXo1 contributes in the same way to virulence as TalC (Yu et al., 2011), allowing invasive colonization of the xylem, further corroborating the functional equivalence of their respective targets.
The TAL effector binding sites in the OsSWEET promoters were identical among diverse rice varieties. Despite polymorphisms in other parts of the promoters (and possibly epigenetic differences), the fact that each gene was strongly activated by its corresponding TAL effector(s) in all varieties tested suggests that X. oryzae TAL effectors have evolved to target highly conserved, functional promoter elements. The quantitative differences in virulence contribution that we observed across varieties might be explained by differences in genes that modulate SWEET activity, in yet unidentified additional targets of the TAL effectors that affect the plant–bacterial interaction, or, except in the case of AvrXa7 and TalC, which target the same gene, in the coding sequences of the OsSWEET genes.
Some patterns of differential virulence contribution were associated with rice varietal group or phylogeny, suggestive of a genetic basis (Fig. 3, Table S1). For example, while most japonica varieties showed increased susceptibility to all three TAL effectors, Minghui is the only indica variety in which susceptibility was enhanced by all three. Single nucleotide polymorphism (SNP) analysis previously showed that Minghui has several large japonica introgressions not found in other indica varieties (McNally et al., 2009). These results show that characterization of plant responses to TAL effectors might be useful to identify susceptibility loci in breeding programs. One of the future goals of our group is to use advanced mapping lines to identify rice genetic regions associated with differential susceptibility to TAL effectors.
In addition to variation in susceptibility, this study identified several cultivars in which some TAL effectors either reduce virulence, trigger a resistance response, or have no effect; all of these are phenotypes that could be associated with resistance loci. The introgression of disease resistance (R) genes is currently the most economically feasible way to control bacterial diseases of plants and strain X11-5A can serve as a useful platform for screening for novel sources of major gene resistance to individual TAL effectors. As advances in high-throughput sequencing come to enable characterization of TAL effector gene inventories across a pathogen population, regionally conserved TAL effectors that would make the most logical candidates for such resistance screens might emerge (Bart et al., 2012). Although we did not identify any cases here, this approach might also identify sources of resistance such as xa13, in which the susceptibility gene promoter sequence for an important TAL effector is disrupted. While screening large germplasm collections for sources of resistance or susceptibility to individual virulence factors is an important step in selecting plant resistance genes for deployment, screening for the identification of individual resistance genes to Xoo strains has usually been conducted on near-isogenic rice lines in IR24 (indica) and Toyonoshiki (temperate japonica) cultivar backgrounds. X11-5A will be useful for characterizing interactions of multiple TAL effectors with diverse rice varieties to rapidly identify accessions for resistance gene discovery and deployment.
In addition to enhancing susceptibility to xylem invasion and bacterial blight development caused by X11-5A, each of two tested OsSWEET-targeting Xoo TAL effectors, avrXa7 (targeting OsSWEET14) and pthXo1 (targeting OsSWEET11), enhanced virulence of the mesophyll pathogen Xoc, resulting in increased length of bacterial leaf streak lesions. However, the effectors did not enable Xoc to cause bacterial blight. These observations indicate that the capacity of SWEET proteins to contribute to susceptibility is not tissue-specific and, at the same time, is not sufficient to allow the mesophyll pathogen Xoc to colonize the xylem. A role for SWEET proteins in phloem loading, as sugar transporters to export photosynthate from phloem parenchyma cells for uptake by companion cells, was recently demonstrated in Arabidopsis (Chen et al., 2012). Their activation in xylem parenchyma cells by Xoo TAL effectors was proposed to contribute to bacterial blight susceptibility by pumping sugar into the xylem for use by the bacterium (Chen et al., 2010). A similar mechanism could account for the contribution of those TAL effectors to bacterial leaf streak susceptibility, flooding the mesophyll apoplast with a ready carbon source for the pathogen. Curiously, though, no native OsSWEET-targeting TAL effectors have been identified in Xoc.
Our study also addressed whether selected Xoc TAL effectors could have a virulence role in X11-5A. These TAL effectors made no difference in susceptibility to X11-5A in either clip or syringe inoculations, on four tested rice varieties. Although the Xoc TAL effectors were not previously confirmed virulence factors like AvrXa7, PthXo1, and TalC, the effectiveness of Xoo but not Xoc TAL effectors in X11-5A suggests that this strain may be predisposed to act as a Xoo-like vascular pathogen. Because X11-5A branched off from the ancestor of both Xoo and Xoc (Triplett et al., 2011), we hypothesize that Xoo-like vascular pathogenicity is the ancestral state of the X. oryzae group, and that the Xoc group lost some factor that enables xylem invasion or gained one that precluded it to evolve towards efficient colonization of the leaf mesophyll. Furthermore, the absence of any TAL effector fragments or pseudogenes in the X11-5A genome suggests that TAL effectors were not present in the genome of the common ancestor. A single gene transfer event incorporating a TAL effector into the X. oryzae ancestral genome could have conferred a major selective advantage, increased by gene duplication and diversification. Whether the TAL effectors that emerged in Xoo and Xoc, and their respective plant targets, drove the divergence in plant tissue specificity between Xoo and Xoc, or were shaped by it, remains an open question.
This research was supported by a Marie Curie IOF Fellowship (EU Grant PIOF-GA-2009-235457 to V.V.), grants from the CSU Infectious Diseases Supercluster, the United States Agency for International Development Linkage Project, and the Colorado Agricultural Experiment Station (to J.E.L.), and by the National Science Foundation (award 0820831 to A.J.B.). The authors thank M. Peterson and P. Langlois for technical assistance, F. White, B. Yang, and B. Szurek for providing TAL effector clones, and Ralf Koebnik for valuable discussions.