Bacterial strains and plasmids used in this study are listed in Supplementary Table S1, and primers used are provided in Supplementary Table S2. Escherichia coli strains were grown in LB medium at 37°C, and X. translucens strains were cultured in GYE (20 g/l glucose, 10 g/l yeast extract) at 28°C. Plasmids were introduced into E. coli and X. translucens by heat shock and electroporation, respectively. Antibiotics used for selection were as follows: ampicillin at 100 μg/ml, spectinomycin at 25 μg/ml, kanamycin at 25 μg/ml for E. coli and 50 μg/ml for X. translucens and tetracycline at 10 μg/ml for E. coli and 2 μg/ml for X. translucens.

For SMRT sequencing, total genomic DNA of strain ICMP11055 was extracted from 30 ml of bacterial culture grown in GYE for 48 h at 28°C on a rotary shaker at 250 rpm as follows (modified from Ausubel et al., 1994). Bacterial suspension was harvested at 3,220 × g for 10 min at 4°C. The pellet was washed twice in 20 ml NE buffer (0.15 M NaCl/ 50 mM EDTA) and resuspended in 2.5 ml of 50 mM Tris pH 8.0, 50 mM EDTA. The suspension was frozen at -20°C and thawed at room temperature. Then, a solution of 0.5 ml 25 mM Tris pH 8.0, 10 μl ReadyLyse, and 50 μl RNase A (10 mg/ml) was added to the suspension and mixed thoroughly. Following incubation for 45 min on ice, one ml of STEP (0.5% SDS, 50 mM Tris pH 7.5, 40 mM EDTA, 2 mg/ml protease K) was added to the sample, and the sample was shaken well and incubated at 37°C for 1 h, mixing every 10–15 min. Then, 1.8 ml 7.5 M ammonium acetate was added and mixed gently. An equal volume phenol/chloroform (1:1 V/V) was added to each tube and mixed well by repeated inversions for 20 min. Following centrifugation at 7,200 × g for 10 min, the aqueous phase was transferred to a new tube and chloroform/isoamyl alcohol extraction was repeated twice more. The final aqueous phase was transferred to a sterile tube, and DNA was precipitated by adding two volumes of cold 95% ethanol and centrifugation at 1,800 × g for 5 min. The pellet was washed with 70% ethanol, dried at room temperature, and then resuspended in TE (10 mM Tris/HCl, 1 mM EDTA, pH 8.0).

To check for the existence of small plasmids, DNA was extracted and examined by agarose gel electrophoresis as described (Chakrabarty et al., 2010). X. axonopodis pv. vesicatoria strain 85-10 (Thieme et al., 2005) and X. oryzae pv. oryzicola BLS256 (Bogdanove et al., 2011) were used as positive and negative controls, respectively.

For PCR analysis of Xtu ICMP11055 mutant strains, genomic DNA was extracted using the GenElute Bacterial Genomic Kit (Sigma–Aldrich, St. Louis, MO, United States). PCR was conducted in 20-μl volume using Phire Hot Start II DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, United States).

Routine plasmid isolation from E. coli was carried out using the E.Z.N.A.^® Plasmid DNA Mini Kit (Omega Bio-Tek, Norcross, GA, United States).

The ICMP11055 genome was sequenced to 120× coverage using two SMRT cells and the P5-C3 chemistry. De novo assembly was done using the HGAP (Chin et al., 2013) v3.0 software package. The assembly was verified with an independent assembly of the TALE gene regions with the PBX toolkit as described (Booher et al., 2015) and annotated by NCBI with the NCBI Prokaryotic Genome Annotation Pipeline (PGAP). The tal genes were named according to the scheme described by Salzberg et al. (2008). The ICMP11055 genome sequence is available in GenBank under accession PRJNA264445.

Whole-genome discovery of single nucleotide polymorphisms (SNPs) and phylogenetic tree construction was carried out using the XT4699 genome sequence as the reference as described by Peng et al. (2016).

Type III effector genes were identified as described by Peng et al. (2016), using blastp against the protein sequences derived from the NCBI annotation of the ICMP11055 genome and tblastn against the whole genome sequence.

Clones of tal1, tal4a, and tal4b were obtained using a shotgun cloning approach, as described by Cernadas et al. (2014) with minor modifications. SphI sites flank the CRR of most of the ICMP110055 tal genes, at short distances from the start and end of the CRR. Ten micrograms of genomic DNA of ICMP11055 were digested with SphI. DNA fragments from 1.5 to 3 kb were then gel purified using 1% agarose and ligated into SphI-linearized and alkaline phosphatase (Calf Intestinal; New England Biolabs, Ipswitch, MA, United States)-treated pTAL1 (Cermak et al., 2011). pTAL1 carries the tal1c gene of X. oryzae pv. oryzicola strain BLS256 missing the SphI fragment that encompasses its CRR. Ligation products were introduced into E. coli TOP10 cells by electroporation, and clones carrying the central SphI fragment of an ICMP110055 tal gene in the correct orientation were identified by colony PCR using oligonucleotide primers 1571 and 1593, which target sequences immediately 5′ of the CRR and within the SphI fragment that are conserved across the ICMP11055 tal genes. The PCR conditions were as follows: initial denaturation at 98°C for 30 s, followed by 30 cycles of 98°C for 5 s, 62°C annealing for 5 s, 72°C extension for 1 min/kb and a final extension of 1 min at 72°C. The cloned tal gene fragments were identified based on their size and 5′ and 3′ sequencing with primers B235 and B236, respectively. Because the SphI site downstream of the CRR is at a different location in the ICMP110055 tal genes from that in tal1c of BLS256, the ligation resulted in loss of a small stretch of coding sequence. This was corrected in each clone by replacing an AatII to Bpu10I fragment spanning the ligation junction with the corresponding fragment from pTAL1. Each clone was confirmed by Sanger sequencing.

Clones equivalent to tal2, tal3a, tal3b, and tal5 were assembled into pTAL1 using the Golden Gate kit of Cermak et al. (2011). RVDs NN and NG, already represented among the repeat modules of the kit, were used in place of HN and KG, because they show the same respective nucleotide specificities (Yang et al., 2014). For RVDs Y^∗ and YK, new repeat modules were created using the Q5^® Site-Directed Mutagenesis Kit (New England Biolabs) and NN repeat modules as templates. Similarly, repeat module QD5 was created using HD5 as template. A last (truncated) repeat module for RVD QD (LR-QD) was synthesized and cloned in plasmid pCR8 as described (Cermak et al., 2011).

For expression in Xanthomonas, all reconstituted tal genes in pTAL1 were introduced into pKEB31 (Cermak et al., 2011) using Gateway LR Clonase (Life Technologies, Carlsbad, CA, United States) (see Supplementary Table S1).

The suicide plasmid pSM7 (Makino, 2005) was used to make a library of tal gene knockout strains of ICMP11055 by targeted integration. pSM7 carries a 4.5-kb PstI fragment with all but the first 80 bp of tal gene aB4.5 (Bai et al., 2000) disrupted in the CRR by an insertion of the EZ-Tn5, NotI/KAN-3 transposon (Epicenter). Insertion endpoints in kanamycin resistant transformants were investigated by sequencing the distal ends of PCR amplicons extending in either direction from the transposon to the regions flanking the CRR. The primers used for amplifying the 5′ flanking DNA were primer 1527 (forward), matching a conserved 5′ sequence in Xtu tal genes, and primer 395 (reverse), corresponding to the 3′ end of the transposon. For the 3′ fragment, forward primer 397, corresponding to the 5′ end of the transposon, and reverse primer 398, originally designed based on a conserved 3′ sequence of X. oryzae tal genes, were used. The amplicon ends farthest from the CRR were sequenced using tal gene primers 1527 and 398, respectively. This process yielded a knockout mutant for tal1, tal2, and tal3b. For each, the 5′ and 3′ PCR product sequences indicated integration by a single crossover at the 5′ end of the tal gene.

For tal3a and for tal4a and tal4b, plasmid pK18mob::sacB (Schäfer et al., 1994) was used to generate a marker-free deletion mutant of the tal3 cluster (containing tal3a and tal3b) and separately a mutant with the tal4 cluster (containing tal4a and tal4b) deleted, following the approach of Kvitko and Collmer (2011). For the tal3 cluster, a unique 1000-bp upstream flanking sequence and a unique 624-bp downstream flanking sequence were amplified by PCR using the primer sets 1623/1624 and 1627/1628, respectively (Supplementary Figure S1 and Table S2). Following verification of the upstream and downstream fragments through sequencing by the use of the primers 1625 and 1629 (Supplementary Figure S1 and Table S2), the fragments were digested with SmaI and then ligated together using T4-DNA ligase. The ligation product was separated on 1% agarose by electrophoresis, and the gel-purified fragment was digested with EcoRI-HF and XbaI and cloned into the suicide plasmid pK18mob::sacB digested with the same enzymes. The recombinant plasmid was electrotransferred into ICMP11055. Transformants were plated on GYE containing kanamycin, and two colonies were cultured overnight without selection and then plated on GYE agar containing 10% sucrose. Resulting colonies were screened by colony-PCR using the 1626/1630 primer pair (Supplementary Figures S1, S2A). PCR products were verified via sequencing using primers 1625 and 1629 and the corresponding mutant tested to confirm kanamycin sensitivity. One of the confirmed mutants was chosen for further study and designated as Δtal3. A tal4 cluster deletion mutant, designated as Δtal4, was obtained in the same way, using primer sets 1631/1632 and 1635/1636 to amplify flanking sequence, primers 1633 and 1637 to sequence-confirm those fragments, primer set 1634/1638 to screen sucrose-resistant colonies, and primers 1633 and 1637 to sequence-confirm the PCR products resulting from the screening (Supplementary Figures S1, S2B).

A tal5 deletion mutant was generated by marker exchange as follows. The chloramphenicol resistance gene was amplified from pdcas9 (Bikard et al., 2013) using primer pair 1916/1917, tal5 flanking sequences were amplified using primer sets 1914/1915 and 1918/1919, and the marker was cloned between the flanking sequences using NEBuilder^® HiFi DNA Assembly Cloning Kit (New England Biolabs) (Supplementary Figure S1). The construct was cloned into pK18mob::sacB using the NEBuilder High-Fidelity DNA Assembly Cloning kit (New England Biolabs). Following electrotransfer into ICMP11055, chloramphenicol- and sucrose-resistant but kanamycin-sensitive colonies were identified and the targeted mutagenesis confirmed by PCR using the primers 1877/1879 (Supplementary Figure S2C). One of the confirmed mutants was chosen for further study and designated as Δtal5.

Wheat cv. Chinese Spring was grown in a growth chamber under a cycle of 16 h light at 28°C and 8 h dark at 24°C with 75% relative humidity (seed kindly provided by M. Sorrells, Cornell University). Inoculum was prepared by suspending log-phase bacterial cells in sterile 10 mM MgCl₂ to an optical density of 0.2 at 600 nm. Suspensions of wild-type cells and mutant cells (or, for complementation assays, mutants transformed with plasmid) were inoculated side by side across the midrib into the youngest two to three leaves of 3-week old plants by infiltration using a disposable syringe with a 30- or 31-gauge hypodermic needle inserted at a sharp angle into the leaf mesophyll tissue, from the adaxial side, and the edges of the infiltrated areas were marked using a permanent marker. For each mutant or transformed mutant, the expansion of lesions beyond the inoculated area was measured 9 days post-inoculation and results were expressed as the ratio of that length to the length resulting from the wild type inoculated opposite on the same leaf. Significant differences were determined using the paired Student’s t-test. The experiment was repeated twice.

The ICMP11055 genome was sequenced using SMRT sequencing. It consists of a single circular chromosome of 4,561,583 bp (Figure 1 and Supplementary Figure S3). The genome has 67.8% average GC content and contains 3,953 annotated protein-encoding genes, 54 tRNA genes, and two rRNA operons, similar to other Xanthomonas genomes.

Genome-wide SNP comparison of ICMP11055 with the XT4699 genome and several other available X. translucens genome sequences (Figure 2) showed that, with three exceptions, strains accessioned as Xtc, Xtt, and Xtu group in distinct clades, and that within the Xtu clade, ICMP11055 resides on a branch separate from the other Xtu strains. The exceptions are three strains accessioned as Xtt that group with the Xtu strains; these are likely misnamed, Xtu strains. A fourth clade contains strains accessioned under pathovars graminis (Xtg), poae (Xtp), and arrhenatheri (Xta). With the exception of ICMP11055 and one Australian Xtu strain, DAR61454, all the strains in the Xtu clade are from North or Central America, and DAR61454 groups tightly with the American strains.

A comparison of general features and selected gene content of the ICMP11055 and XT4699 genomes is presented in Table 1. A schematic alignment is shown in Figure 3. The ICMP11055 genome is 200 kb larger than that of XT4699. Nine extra genomic regions (>10 kb) are present in ICMP11055 (Figure 3). These include regions containing prophage elements (Regions D, F, G, H), a CRISPR gene cluster (Region C), VGR (valine-glycine repeats)-related genes of the Type VI secretion system (Regions A, F, I), genes encoding cyclolysin secretion ATP-binding protein, hemolysin secretion protein D and alkaline phosphatase (Region E), and a gene cluster for Type II and Type IV secretion systems (Region B). Content in each of these regions is unique to ICMP11055 except the VGR-related genes (Regions A, F, I): VGR-related genes are present in other regions shared by ICMP11055 and XT4699. Only one genomic region (>10 kb) is unique to XT4699, being a gene cluster predicted to encode a type III restriction and modification system (Region J). IS elements are frequently found at the ends of the extra genomic regions in ICMP11055, implicating possible involvement of transposons in the acquisition or loss of these regions. Similarly, IS elements are found at the ends of two rearranged regions in the alignment between the two genomes (Figure 3), indicating that the rearrangements might have been mediated by transposon-related sequences. ICMP11055 and XT4699 contain seven and eight tal genes, respectively. Based on RVD sequence, four of these are perfectly or near perfectly conserved between the two strains, and a fifth is conserved except in the last five RVDs of its CRR (Figure 3).

Xanthan, an exopolysaccharide characteristic of Xanthomonas spp., has been implicated in Xanthomonas pathogenicity or virulence in several studies (Büttner and Bonas, 2010). Xanthan biosynthesis is directed by several genetic loci including the highly conserved gum gene cluster, gumB to gumM. (Katzen et al., 1998; Vojnov et al., 1998). This cluster is present in ICMP11055, however, gumG is missing. The gumG gene is missing from the cluster in all other available X. translucens genomes as well. The missing gene was reported to be gumF in Xtc strain ART-Xtg29 Xtg29 (Wichmann et al., 2013), however, based on sequence alignment, gumF is present and gumG absent in that assembly (not shown). The gumO and gumP genes, which are present in group II Xanthomonas that infect eudicots but absent from those that infect monocots (Lu et al., 2008; Jacobs et al., 2015), are also missing from ICMP11055 and each of the other X. translucens strains examined.

The type III secretion system (T3SS) is key to the pathogenicity of Xanthomonas spp. with the notable exceptions of X. albilineans (Pieretti et al., 2009) and X. cannabis (Jacobs et al., 2015). A strain of X. sacchari associated with banana plants was found that lacks the T3SS but its pathogenicity was not tested (Studholme et al., 2011). The T3SS is encoded by a chromosomal cluster of more than 20 hrp (HR and pathogenicity) genes in several operons (Büttner and Bonas, 2010). The cluster is, as expected, present in ICMP11055. However, in contrast to many studied Xanthomonas strains, in ICMP11055, the hrpG and hrpX genes, which encode regulators of the T3SS, are located inside the hrp gene cluster (Supplementary Figure S4). This characteristic of ICMP11055 is shared by XT4699 (Peng et al., 2016) and by Xtg ART-Xtg29 (Wichmann et al., 2013).

The T3SS delivers a suite of proteins called effectors into plant cells. These type III effectors (T3E) act as virulence factors by targeting host metabolic, defense, or other components to promote disease or, depending on the host genotype, as avirulence factors by triggering plant defense. A typical Xanthomonas genome may encode 2–30 (non-TALE) T3E (Studholme et al., 2010). Candidate T3E of ICMP11055 were identified and compared with those of XT4699 and a third Xtu strain, XT-Rocky, as well as one strain each of Xtc, Xtt, Xtg, Xtp, and Xta (CFBP 2541, DSM 18974, ART-Xtg29, B99, and UPB455, respectively; Supplementary Table S3). Thirty-two putative T3E genes were predicted from the ICMP11055 genome sequence, Among these are ten core T3E genes, including avrBs2, xopF, xopK, xopL, xopN, xopP, xopQ, xopR, xopX, and xopZ, that were previously reported present in the genomes of all examined Xanthomonas¹ except X. albilineans, X. cannabis, X. sacchari, and X. campestris pv. armoraciae, the last of which carries only xopP and xopR. Each is present also in the other X. translucens genomes examined, except xopK, which is disrupted by a frameshift mutation in the Xtg, Xtp, and Xta strains, and xopR, which is frameshifted in the Xtc strain. All the strains have two copies of xopF, though one copy is disrupted by a frameshift in the Xtp and Xtu strains. Each also has multiple copies of xopX. The avrBs2 gene is present in two copies in each strain except the Xtg and Xtp strains, in which it is single-copy. The xopL and xopP genes are dual-copy in each of the Xtu strains, and their copy number ranges from one to four across the representative strains of the other pathovars. This copy number variation contrasts with group II Xanthomonas genomes, in which most of the core T3Es are present in single copy¹. Seven other T3E genes found in ICMP11055, hpaH (hopP1), xopB, xopC2, xopG, xopV, xopY, and xopAM, are each also present in all of the other X. translucens strains examined, all single-copy. Four more are present in each of the three Xtu strains and all but one of the other strains: xopAA is present in all but the Xtg strain, at single copy; xopAD is present in each strain at single copy but is frameshifted in the Xtp strain; xopAF is single copy in ICMP11055 and one or two copies in the other strains except the Xta strain, from which it is absent; and xopAP is present at single copy in all strains but is frameshifted in the Xtt strain. ICMP11055, the two other Xtu strains, and the Xtt strain each also carries xopAH, which is missing from the other strains, and xopAK, which is also present in the Xtg strain but not in the others. The Xtu strains and the Xtc strain each carry two members of the xopE family, xopE1 and xopE5, each at single copy. Other members of the family, xopE2, xopE3, and xopE4, are distributed unevenly across the strains representing other pathovars, at single copy where they occur. Finally, four T3E not found in ICMP11055, avrBs1, xopI, xopJ1, and xopJ1, are present in one or more of the other strains examined. Overall, there is substantial variation in T3E content across the pathovar representatives examined, however, with the exception of the presence of xopJ1 in XTU-Rocky, T3E content is identical across ICMP11055 and the other Xtu strains.

Many members of the genus Xanthomonas harbor one or more tal genes, and individual strains in some species may harbor several tal genes (up to 29 tal genes have been observed in individual X. oryzae pv. oryzicola strains) (Bonas et al., 1989; Yang and White, 2004; Boch and Bonas, 2010; Wilkins et al., 2015). In the comparison of T3E content described above, presence of tal genes, or in the draft genomes evidence thereof, was also examined (Supplementary Table S3). Similar to ICMP11055 and XT4699, the third Xtu strain, XTU-Rocky, harbors at least seven tal genes. The Xtc strain (CFBP 2541) carries two. Partial sequences were detected in the draft genome assemblies for the Xtt, Xtp, and Xta strains. None was found in the Xtg sequence. The seven ICMP11055 tal genes encode proteins with 15 to 18 RVDs. The TALEs of ICMP11055 are partially distinct from those of XT4699 in their DNA binding specificities (RVD sequences) (Figure 4). Four are perfectly or nearly perfectly conserved (1 or zero differences in RVDs) between the two strains. One ICMP11055 TALE, Tal2, shares the first 13 RVDs with Tal2 of XT4699, but the two TALEs diverge in the remaining five RVDs. There is no obvious similarity among the remaining two TALEs of ICMP11055 and three TALEs of XT4699.

The TALEs of ICMP11055, like those of XT4699 (Peng et al., 2016), show unusual RVD composition. Across all TALEs, HD, NG, NI, and NN are the most common RVDs; several others occur less frequently (Moscou and Bogdanove, 2009; Boch and Bonas, 2010; Rinaldi et al., 2017). In addition to the common RVDs, ICMP11055 and XT4699 TALEs harbor some RVDs so far not observed in TALEs in other Xanthomonas species, including NF, KG, QD, Y^∗, YD, and YK (Figure 4). Also distinct from TALEs in other Xanthomonas species, HN is relatively frequent in ICMP11055 and XT4699 TALEs, as it is in Ralstonia RipTALs (Lange et al., 2013). On the other hand, the RVD NK, which occurs at higher frequencies in Ralstonia than in Xanthomonas, is not found in the ICMP11055 or XT4699 TALEs. The TALEs encoded by the few genes individually cloned from other Xtu strains (Peng et al., 2016) align well with ICMP11055 and XT4699 TALEs, while the two available Xtc CFBP 2541 TALE sequences (Pesce et al., 2015) and six of the eight identified Xtt DSM 18974 TALE sequences (Jaenicke et al., 2016) are unique (Supplementary Figure S5). One of the remaining Xtt DSM 18974 TALEs aligns well with ICMP11055 Tal2 up to RVD 14 but diverges thereafter, distinctly from XT4699 Tal2. The other aligns well with ICMP11055 Tal3b (and XT4699 Tal7). The Xtc CFBP 2541 TALEs contain the RVDs KI and GI, not found in any of the other TALEs. The CFBP 2541 TALEs and one of the DSM 18974 TALEs also contain an NK. Of all the TALEs, only one, from CFBP 2541, contains an NS, an RVD that is generally common in Xanthomonas TALEs but not yet found in RipTALs. Thus RVD composition appears to set X. translucens TALEs apart both from TALEs of group II Xanthomonas species and from RipTALs.

Transcription activator-like effector repeats are typically 34 amino acids in length, or 33 when the RVD is missing a residue. An exception is Hax2 of X. campestris pv. armoraciae strain 5, which consists solely of 35 aa repeats; Hax3 and Hax4 from that strain, however, contain only the more typical 34 aa repeat type (Kay et al., 2005). AvrHah1 from X. gardneri comprises both 34 and 35 aa repeat types (Schornack et al., 2008). Ralstonia RipTAL repeats are 35 amino acids long (Bogdanove et al., 2010). With the exception of ICMP11055 Tal5 and its ortholog in XT4699, which contain exclusively repeats of the 35 aa type, the ICMP11055 and XT4699 TALEs (and the other X. translucens TALEs for which sequence is available) each comprise both 34 and 35 aa repeat types. Interestingly, although Tal4a in ICMP11055 and Tal6 in XT4699 have identical RVD sequences, they differ at two repeats in the repeat length, 34 vs. 35 aa. As in the RipTALs, the extra amino acid in the 35 aa repeats of the X. translucens TALEs is at position 32 or 33, within the inter-repeat loop (Mak et al., 2012). And, whereas the penultimate two residues of the repeat consensus for TALEs of X. oryzae, X. citri, X. euvesicatoria, and X. campestris are HG, they are PY in the repeat consensus for ICMP11055, as in the repeat consensus of a typical RipTAL (Figure 5). However, like other Xanthomonas TALEs and in contrast to RipTALs, the non-RVD, backbone residues of the ICMP11055 TALE repeats are relatively invariable. Thus, the overall X. translucens TALE repeat structure appears to be somewhat intermediate between that of RipTALs and that of TALEs of group II Xanthomonas species. The N- and C- terminal domains outside the repeats, however, are typical of TALEs in group II Xanthomonas species (Supplementary Figure S6).

The tal5 gene of ICMP11055, in addition to being the only ICMP11055 tal gene comprising exclusively repeats of the 35 aa type, has a promoter sequence distinct from the other ICMP11055 tal genes (not shown), and the encoded N- and C-terminal regions also differ (compare Tal5 and the representative Tal3a in Supplementary Figure S6). These differences suggest that tal5 was acquired through horizontal gene transfer sometime after the other tal genes were established in the lineage that gave rise to ICMP11055, and that those other tal genes had arisen through duplication of a distinct, ancestral tal gene.

To determine whether any of the ICMP11055 TALEs contribute to virulence, tal gene knockout mutant derivatives of ICMP11055 were generated. The mutant strains include single-crossover, knockout mutants M2 (tal1), M10 (tal2), and M55 (tal3a) and deletion mutants Δtal3 (tal3a and tal3b), Δtal4 (tal4a and tal4b), and Δtal5 (tal5) (see Materials and Methods). The mutants were inoculated to wheat cv. Chinese Spring leaves, and the lengths of resulting lesions were compared to those caused by the wild type strain at 9 days after inoculation. M10 and Δtal4 caused shorter lesions than the wild type strain, and the lengths of the lesions produced by the other mutants were not significantly different from those of the wild type strain (Figure 6A and Supplementary Figure S7). A dTALE equivalent of tal2 (see Materials and Methods) introduced on a plasmid into M10 restored full virulence (Figure 6B). Cloned tal4b, but not tal4a, likewise restored full virulence to Δtal4 (Figure 6B). These results indicate that tal2 and tal4b encode major TALEs that contribute non-redundantly to the virulence of ICMP11055.

To detect virulence contributions of any ICMP11055 TALEs that might function redundantly, we next attempted a gain-of-function approach by expressing each ICMP11055 tal gene individually in the Xtc strain CFBP 2541, which is weakly virulent on Chinese Spring. In Chinese Spring leaves, watersoaking caused by CFBP 2541 on Chinese Spring leaves remains limited to the infiltrated area, in contrast to the spreading watersoaking caused by ICMP11055. None of the ICMP11055 tal genes, however, changed the extent of watersoaking caused by CFBP 2541, including tal2 and tal4b (data not shown).