Identification of a virulence tal gene in the cotton pathogen, Xanthomonas citri pv. malvacearum strain Xss-V2–18

Background Bacterial blight of cotton (BBC), which is caused by the bacterium Xanthomonas citri pv. malvacearum (Xcm), is a destructive disease in cotton. Transcription activator-like effectors (TALEs), encoded by tal-genes, play critical roles in the pathogenesis of xanthomonads. Characterized strains of cotton pathogenic Xcm harbor 8–12 different tal genes and only one of them is functionally decoded. Further identification of novel tal genes in Xcm strains with virulence contributions are prerequisite to decipher the Xcm-cotton interactions. Results In this study, we identified six tal genes in Xss-V2–18, a highly-virulent strain of Xcm from China, and assessed their role in BBC. RFLP-based Southern hybridization assays indicated that Xss-V2–18 harbors the six tal genes on a plasmid. The plasmid-encoded tal genes were isolated by cloning BamHI fragments and screening clones by colony hybridization. The tal genes were sequenced by inserting a Tn5 transposon in the DNA encoding the central repeat region (CRR) of each tal gene. Xcm TALome evolutionary relationship based on TALEs CRR revealed relatedness of Xss-V2–18 to MSCT1 and MS14003 from the United States. However, Tal2 of Xss-V2–18 differs at two repeat variable diresidues (RVDs) from Tal6 and Tal26 in MSCT1 and MS14003, respectively, inferred functional dissimilarity. The suicide vector pKMS1 was then used to construct tal deletion mutants in Xcm Xss-V2–18. The mutants were evaluated for pathogenicity in cotton based on symptomology and growth in planta. Four mutants showed attenuated virulence and all contained mutations in tal2. One tal2 mutant designated M2 was further investigated in complementation assays. When tal2 was introduced into Xcm M2 and expressed in trans, the mutant was complemented for both symptoms and growth in planta, thus indicating that tal2 functions as a virulence factor in Xcm Xss-V2–18. Conclusions Overall, the results demonstrated that Tal2 is a major pathogenicity factor in Xcm strain Xss-V2–18 that contributes significantly in BBC. This study provides a foundation for future efforts aimed at identifying susceptibility genes in cotton that are targeted by Tal2.


Background
Cotton (Gossypium spp.) is an economically-important crop worldwide and is a significant source of fiber, feed, oil and biofuel [1]. The primary cotton production areas are located in the southern United States (USA), Central America, western Africa, and central and eastern Asia. According to the 2017/18 world ranking, China leads the world in cotton production followed by India, the USA and Pakistan [2]. Gossypium spp. contains over 50 species, including G. arboreum, G. herbaceum, G. hirsutum and G. barbadense. G. arboretum and G. herbaceum are diploid (2n = 26), whereas G. hirsutum and G. barbadense are tetraploid (4n = 52) [3,4]. G. hirsutum is the predominant species and produces with 90% of the world's cotton fiber production [5]. This species is impacted by a devastating bacterial disease known as bacterial blight of cotton (BBC), which is caused by Xanthomonas citri pv. malvacearum. The first detailed description of BBC was reported in the USA [6]. However, this disease currently occurs in all cotton production areas and causes significant yield losses (5-35%) either by injury to the plant or direct damage to the boll [7].
TALE proteins, functionally resemble eukaryotic transcription factors, are localized to the host plant nucleus where they bind to specific promoter sequences known as effector-binding elements (EBEs), thus regulating host gene expression [29][30][31]. TALEs belong to the avrBs3/pthA gene family [26], which is highly conserved among different Xanthomonas spp. TALEs contain an N-terminal T3S signal domain, a central repeat region (CRR), C-terminal nuclear localization signals (NLS), and an acid activation domain (AD) [30,31]. CRRs contain tandem repeats of 33-35 amino acids that differ only at residues 12 and 13; these are designated repeat variable di-residues (RVDs) and determine the specificity of DNA binding [30][31][32]. TALE-mediated activation of EBEs can induce host susceptibility (S) or resistance (R) genes [29,30]. For example, the TALEs PthXo1 and PthXo2 from X. oryzae pv. oryzae (Xoo) were shown to enhance the expression of rice genes OsSWEET11 and OsSWEET13, which are required for susceptibility to bacterial leaf blight [33,34]. However, rice cultivars were resistant to Xoo when they contained OsS-WEET11 and OsSWEET13 alleles lacking PthXo1 and PthXo2 EBEs [35][36][37]. Recently, a new rice S gene (OsERF#123) was shown to be targeted by TalB in African strains of Xoo [38]. Other examples of TALEs include AvrBs3 that targets the pepper resistance gene Bs3 and AvrXa10, AvrXa23 and AvrXa27 that interact with rice R genes Xa10, Xa23 and Xa27, respectively [29,[39][40][41]. Recently, Cai et al. [21] reported that Tal7 from Xoo binds and activates the expression of the rice gene Os09g29100, an interaction that suppresses avrXa7-Xa7-mediated resistance in rice. A number of truncated TALEs (truncTALEs) and interfering TALEs (iTALEs) have also been reported in Xoo that function as suppressors of Xa1-mediated defense in rice [42,43].
Resistance to Xcm has been identified primarily in G. hirsutum. The genetic nature of resistance to BBC was first revealed in 1939, and efforts to breed cotton plants for resistance ensued shortly thereafter [44]. About 20 major R genes or polygene complexes (B genes) participate in resistance to BBC in cotton [7,8]. Based on their virulence phenotype in differential cotton hosts, Xcm strains have been classified into 22 races that are named 1-22 [7]. Race 18 is the most common variant and was first isolated in 1973 [45,46]. In some cases, the outcome of interactions between Xcm strains and differential cotton varieties is dependent on the avrBs3/pthA gene family in Xcm, which indicates that Xcm-cotton interactions follow the gene-for-gene model for host plant resistance [7,10,47,48].
The aim of the current study is to identify a novel virulent tal-gene encoding TALE protein in a highly virulent cotton pathogen, Xcm strain Xss-V 2 -18 (from China).

TALEs of Xss-V 2 -18
Restriction fragment length polymorphism (RFLP) analysis was conducted to estimate the number and size of tal genes in Xcm Xss-V 2 -18. Since most tal genes retain two BamHI sites, Xcm Xss-V 2 -18 plasmid and genomic DNAs were digested with BamHI and analyzed by Southern blotting as described above. Six bands hybridized to the probe in BamHI-digested genomic and plasmid DNA, indicating that Xss-V 2 -18 contained six plasmid-encoded tal genes (Fig. 1a).
To better understand the features of Xss-V 2 -18 TALEs, we compared them with TALEs in Xcm strains MSCT1, H1005, N1003, MS14003 and AR81009 [27,46,59]. Phylogenetic tree of TALEs from Xcm strains were constructed by aligning TALE-CRR with DisTAL v1.1. All 53 TALEs (Xss-V 2 -18 = 6, MSCT1 = 8, H1005 = 12, N1003 = 9, MS14003 = 8 and AR81009 = 12) were classified into 6 major groups and 33 sub-groups. Tal2 of Xss-V 2 -18, TAL6 of MCST and Tal26 of MS14003 fall in same group (Fig. 2b). Fig. 1 Southern blotting, and Isolation and sequencing of Xss-V 2 -18 tal-genes. a Southern blot analysis of BamHI-digested genomic (gDNA) and plasmid DNA (pDNA) of Xcm strain Xss-V 2 -18. A 2.9-kb SphI fragment of pthXo1 (from Xoo) was labeled with digoxygenin (DIG) and used as a probe to detect tal genes in Xcm Xss-V 2 -18. b Plasmid DNA of Xss-V 2 -18 was digested with BamHI, and fragments were gel-purified and ligated into BamHI-digested and CIP-treated pBluescript II SK(−). Southern blot analysis was performed by the using internal SphI fragment of pthXo1 as a probe to confirm each clone (pB-tal1 -pB-tal6). c Schematic diagram of strategy used to sequence tal genes. After cloning into pBluescript II SK(−), the EZ-Tn5™ < KAN-2 > Tnp Transposome™ Kit was used to insert Tn5 into each tal gene. Clones with Tn5 insertions in the middle of the CRR were selected by SphI digestion and sequenced using primer pairs tal-F/RP and FP/tal-R Nearly identical RVD sequences were observed for the six TALEs in Xss-V 2 -18, MSCT1, H1005, MS14003 and AR81009 (Fig. 2a)  . Letters in red font indicate RVDs that differ between the two strains. The asterisk represents a missing amino acid residue b Construction of phylogenetic tree based on central repeat amino acid sequences of TALEs. A set of 53 TAL effector sequences from 6 different Xcm strains were used to construct tree with DisTAL program using default parameters. TALEs were classified into 6 major groups and 33 sub-groups showing the relationship of Xcm Xss-V 2 -18 to other Xcm strains published previously. Tal2 of Xss-V 2 -18, TAL6 of MCST and Tal26 (M26) of MS14003 fall in same group. Scale is shown below the tree mentioned in Fig. S1. RVDs in Xcm strains included NI, NG, NS, HD and NN; the latter RVD was absent in Tal1, Tal2, Tal3 and Tal4.

Xss-V 2 -18 tal deletion mutants
To assess the role of tal genes in the virulence of Xss-V 2 -18, we generated tal deletion mutants by homologous recombination using the suicide vector pKMS1 [60]. Fragments a (580 bp) and b (350 bp) were amplified on the left and right sides of DNA encoding the CRR, respectively, and cloned as a fused fragment in pKMSA1 (Fig. 3a, b). Construct pKMSA1 was introduced into Xcm Xss-V 2 -18; after homologous recombination, 41 putative mutants were selected for PCR amplification using primers pKMSA1-5F/pKMSA1-3R (Table S1). Four putative mutants designated M1, M2, M3 and M4 contained a 930-bp PCR product, which is consistent with the size of the insert in pKMSA1 (Table 1, Fig.  3c). Southern hybridization indicated that one or more tal genes were deleted in the four mutants (Fig. 3d). M1 and M2 were lacking tal3 and tal2, respectively, M3 was missing tal2 and tal4, and M4 lacked tal2, tal4, tal5, and tal6. These results indicated that four tal loci underwent homologous exchange via pKMSA1, and copies of the plasmid pKMSA1 functioned to delete multiple tal genes simultaneously in M3 and M4.
A second round of deletion mutagenesis was conducted with plasmid pKMSA2, which contains a fusion of fragments c (150 bp) and d (300 bp) on the left and right sides of the DNA encoding the CRR, respectively (Fig. 3a). Construct pKMSA2 was used to generate new deletions in the M4 mutant, and potential new mutants were analyzed by PCR with primer pairs pKMSA2-5F/ pKMSA2-3R (Table S1). Two mutants designated M5 and M6 contained a 450-bp PCR product that is consistent with the size of the insert in pKMSA2 (Fig. 3e). In addition to tal2, tal4, tal5, and tal6, Southern hybridization indicated that mutant M5 contained a deletion in tal3. M6 was lacking both tal1 and tal3 (Fig.  3f); thus, M6 lacked all six tal genes and can be considered a tal-free mutant of Xss-V 2 -18. (350 bp) were amplified on the left and right sides of the CRR, respectively, and cloned as a fused fragment in pKMSA1. Fragments c (580 bp) and d (150 bp) were amplified on the left and right sides of the CRR, respectively, and cloned as a fused fragment in pKMSA2. Constructs pKMSA1 and pKMSA2 were introduced into Xcm strain Xss-V 2 -18 by electroporation, and deletion of the CRR region was conducted as described in Methods. b Confirmation of 930-and 450-bp inserts in pKMSA1 and pKMSA2, respectively, by digestion with XbaI and SmaI. c PCR analysis of 41 putative mutants with primers pKMSA1-5F and pKMSA1-3R. A 930-bp fragment was amplified in M1, M2, M3, and M4, indicating that these four mutants underwent a homologous recombination and potential deletion of the CRR region. pKMSA1 was included as a control. d Southern hybridization analysis of Xss-V 2 -18 and mutant strains M1-M4. Plasmid DNA of WT Xss-V 2 -18 and mutants were isolated and digested with BamHI. The internal SphI fragment of pthXo1 (from Xoo) was used as a hybridization probe to detect tal genes. e PCR screening for putative mutants using primers pKMSA2-5F and pKMSA2-3R. pKMSA2 was included and used as a positive control. f Southern hybridization analysis of mutant M4 (used for second round of mutagenesis), M5 and M6. Plasmid DNA of M4, M5 and M6 were isolated and digested with BamHI, and the internal SphI fragment of pthXo1 was used a hybridization probe to detect tal genes

Virulence assays
Xss-V 2 -18 and mutants M1-M6 were inoculated into cotton leaves and phenotypes were observed 3-5 days post-inoculation (Fig. 4a). Xss-V 2 -18, M1, and M4 produced substantial water-soaked lesions in the inoculation sites; however, water-soaking was reduced in leaves inoculated with M2, M3, and M5 (Fig. 4a). In contrast, the region inoculated with the tal-free mutant M6 showed cell death and necrosis (Fig. 4a) signifying that the loss of tal genes affect the virulence of Xss-V 2 -18. On the second day post-inoculation, the populations of the M2 and M6 mutants were significantly lower than Xss-V 2 -18, M1, M3, M4 and M5 (Fig. 4b). On days 4 and 6 post-inoculation, the growth of Xss-V 2 -18 was significantly higher than mutants M1-M6 with no significant difference among the mutants. These results indicated that some of the tal genes are involved in Xss-V 2 -18 virulence, and the absence of selected tal genes impacted growth of the pathogen in planta.

Discussion
Until recently, BBC has been effectively controlled using classical R genes [63][64][65]; however, in 2011 the disease re-emerged with a vengeance [46]. A known virulence factors, transcription activator-like effectors (TALEs), in Xcm are important for BBC. In previous studies, 8-12   Wild-type (WT) and mutant strains were inoculated to the lower surface of cotton leaves (two-week-old plants) using a needleless syringe. Infiltration with simply 10 mM MgCl2 served as a mock. Phenotypes were observed 3-5 days post-inoculation. b In planta growth of WT Xss-V 2 -18 and mutants. Growth was measured at 0, 2, 4, and 6 days post-inoculation. Error bars represent means and standard deviations (means ± SD), and columns labeled with different letters represent significant differences (P < 0.05). c Western blot analysis of TALE production in Xcm M2. Plasmid pHZW-tal2 was transferred into Xcm M2 by electroporation. Production of TALE was analyzed by western blotting using an anti-FLAG primary antibody (see Methods). RNA polymerase subunit alpha (RNAP) from E. coli, was used as a loading control. d Symptoms in cotton leaves inoculated with Xss-V 2 -18, mutant M2, M2 containing empty vector and M2 containing tal2 in trans. Bacterial strains were inoculated to cotton leaves using a needleless syringe, and phenotypes were observed within 5-7 days post-inoculation. e In planta growth of the WT Xss-V 2 -18, mutant M2 and complemented strain. Growth was measured at 0, 2, 4, and 6 days post-inoculation. Error bars represent means and standard deviations (means ± SD), and columns labeled with different letters represent significant differences (P < 0.05) tal genes were reported in Xcm [26-28, 48, 59]. Some Xcm tal genes, notably avrB101, avrB102 and avrBln, are known to cause an hypersensitive response (HR) on cotton [28], whereas avrb6 elicits water-soaking [48]. In this study, RFLP-based Southern hybridizations indicated that the highly-virulent Xcm strain Xss-V 2 -18, which was originally isolated from Hainan, China, harbors six plasmid-borne tal genes (Fig. 1). In the genus Xanthomonas, the location and number of tal genes varies among species, pathovars and strains [55,66]. For example, strains of X. oryzae pv. oryzicola (Xoc) encode over 250 chromosomally-borne tal genes [55]; however, plasmid-encoded tal genes are common in other Xanthomonas spp. Examples include the tal genes in X. citri pv. citri, X. citri pv. aurantifolii and X. axonopodis pv. manihotis, which were identified on plasmids pXAC66, pXcB and pXam46, respectively [67][68][69]. Feyter and Gabriel [28] and Showmaker et al. [59] reported the existence of plasmid-borne tal genes in Xcm strains XcmH and MSCT1, respectively. A draft genome sequence of the Xanthomonas translucens pv. cerealis strain CFBP 2541 also indicate a plasmid borne tal-gene [70]. The presence of highly repetitive sequences in tal genes complicates efforts to obtain their nucleotide sequence; therefore, we used a Tn5 insertion method as a sequencing strategy. This sequencing strategy for talgenes was also used by others previously [21,71]. Normally the number of repeats in tal genes varies between 1.5 and 33.5, and each repeat encodes 33-34 amino acids that vary only at positions 12 and 13 (RVDs) [30]. In Xcm Xss-V 2 -18, we identified 27.5, 25.5, 21.5, 18.5, 15.5 and 13.5 tandemly arranged 102-bp direct repeats (encoding 34 amino acids) in tal1, tal2, tal3, tal4, tal5 and tal6, respectively. In order to understand how Xcm TALome differ from each other within and between strains, DisTAL and AnnoTALE were used to characterized [50,72]. Xcm encodes very diverse TAL effectors that were classified exclusively into 6 major groups and 33 sub-groups. TALE phylogenetic tree of Xcm strains showed that Tal2 of Xss-V 2 -18, TAL6 of MCST and Tal26 (M26) of MS14003 fall in same group. Furthermore, RVDs based analysis showed that the six TALEs in Xss-V 2 -18 were identical or nearly identical to plasmid-borne TALEs in Xcm MSCT1, MS14003, H1005 and AR81009 which suggests that these genes may have been horizontally transferred [67,73,74]. The number and location of tal genes varied in the six Xcm strains; MSCT1 possess eight (seven plasmid-borne) [59], XcmH1005 has 12 (six plasmid-and six chromosomallyencoded) [27], XcmN1003 has nine (four plasmidencoded) [27], MS14003 has 8 (7 plasmid-encoded) [46], AR81009 has 12 (six plasmid-encode) [46] and Xss-V 2 -18 has six plasmid-encoded tal genes (Figs. 1, 2). The variation in number, location and RVD sequence in Xcm TALEs could be important for maintaining virulence in cotton cultivars grown in different geographical regions.
To assess the role of tal genes in Xss-V 2 -18, we generated deletions in Xss-V 2 -18 by homologous recombination with pKMS1 [60], which was previously used to generate deletion mutants in the rice pathogen, Xoc [75][76][77]. This is the first report where pKMS1 was used to generate tal deletion mutants in Xcm, and the basic strategy was to replace the CRR (encoded by 102-bp repeat units) with up-and downstream fragments flanking the tal genes. Using construct pKMSA1, we obtained four mutants; M1 and M2 lacked tal3 and tal2, M3 had deletions in tal2 and tal4, and M4 lacked tal2, tal4, tal5 and tal6. We speculate that tal5 and tal6 might be located in the same gene cluster. The second knockout was obtained using pKMSA2 where up-and downstream flanking fragments (homology arms) were located closer to the CRR. Mutant M4 was used as a parental strain for the deletions generated with pKMSA2, and we recovered two new mutants designated M5 and M6. In addition to tal2, tal4, tal5 and tal6, mutant M5 also lacks tal3, whereas M6 contains deletions in all six tal genes (Fig.  3). The symptoms induced by M2, M3, M5 and M6 were significantly reduced relative to the wild-type, thus indicating that one or more tal genes contribute to symptom development in Xss-V 2 -18. Mutants M2, M3, M5 and M6 all lack the tal2 gene; thus the potential contribution of tal2 to symptom development was further investigated. Expression of tal2 in trans restored symptoms and growth in planta to the M2 mutant, thus confirming that Tal2 is a virulence factor (Fig. 4). Although the TALE repertoire of Xcm Xss-V 2 -18, MSCT1, MS14003, H1005 and AR81009 is somewhat identical, Tal2 of Xss-V 2 -18 differs at two repeat variable diresidues (RVDs) from Tal6 in MSCT1 and Tal26 in MS14003, inferred functional dissimilarity.
TALEs functionally resemble eukaryotic transcription factors that target and regulate the expression of host genes by binding to their promoter sequences. TALEtriggered susceptibility has been well-studied, and the contribution of TALEs to virulence has been evaluated in many Xanthomonas spp. [21-23, 27, 57, 78-81]. For example, the TALEs PthXo1 and PthXo2 from Xoo were shown to enhance the expression of rice genes OsS-WEET11 and OsSWEET13, which are required for susceptibility to bacterial leaf blight in rice [33,34]. However, rice cultivars were resistant to Xoo when they contained OsSWEET11 and OsSWEET13 alleles lacking PthXo1 and PthXo2 EBEs [35][36][37]. A recent study by Peng et al. [82] reported that Tal8 from Xtu target and induce the expression of host gene Ta-NCED-5BS, encode enzyme required for rate-limiting step in ABA biosynthesis, to promote disease susceptibility. In another new study, Wu et al. [83] shown that TAL-effector Brg11 from Ralstonia solanacearum enhance the expression of 5-truncated ADC (arginine decarboxylase) transcripts that subvert translational control and thereby inhibit competing pathogens. In Xcm, Avrb6 was the first TALE shown to be important for virulence [25]. Recently, the Xcm effector Avrb6 was shown to target and induce the expression of the cotton S gene, GhSWEET10, thus enhancing virulence and promoting disease [27]. The present study provides an important foundation for identifying potential S genes that interact with Tal2, which will ultimately help us develop better control strategies for BBC.

Conclusions
In this study, we identified genes encoding TALEs in the highly-virulent Xcm strain, Xss-V 2 -18 (from China), and assessed TALE roles in BBC. We found that Xss-V 2 -18 encodes six plasmid-borne tal genes. Knockout mutagenesis of Xss-V 2 -18 tal genes and complementation analysis demonstrated that Tal2 is required for full virulence of Xss-V 2 -18 on cotton. The identification of the Tal2 target in cotton will ultimately provide new avenues for developing BBC-resistant varieties.

Bacterial strains, growth conditions, and plasmids
The bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli strains were grown in Luria-Bertani (LB) medium (5 g yeast extract, 10 g NaCl, 10 g tryptone/L) or LB with agar at 37°C. Xcm strains were grown in nutrient broth (NB) (1 g yeast extract, 3 g beef extract, 5 g polypeptone and 10 g sucrose/L) or NB with agar at 28°C. Xcm transformants containing the first crossover event were grown on NAN (nutrient agar without sucrose) or NBN (NAN without agar) medium. For the second crossover event, transformants were plated on NAS agar (NAN with 10% sucrose) [60]. When appropriate, antibiotics were added at the following concentrations (μg/mL): ampicillin, 100; kanamycin, 20; spectinomycin, 25; and rifampicin, 50. The pH of both solid and liquid media was adjusted to 7.0-7.2.

DNA preparation
Total genomic DNA of Xss-V 2 -18 was isolated using the Bacterial Genomic DNA Extraction Kit (TaKaRa, China). The isolated gDNA pellet was re-suspended in double-distilled water. Bacterial plasmid DNA was isolated using the Plasmid Miniprep Kit (Omega, USA). The quality and quantity of genomic DNA and plasmid DNA were checked with NanoDrop spectrophotometer (Eppendorf). Routine plasmids isolation from E. coli was carried out by using the plasmid DNA Mini Kit (GBS Biotechnology, China).

Isolating, cloning and sequencing of Xss-V 2 -18 tal genes
The isolation and cloning of tal genes from Xcm strain Xcc-V 2 -18 followed a previously described procedure [21,71,79,84] with minor modifications. Plasmid DNA and genomic DNA (50 μg) were isolated from Xcm, digested with BamHI, and separated on 1.2% agarose gels. Specific tal DNA fragments were then gel-purified and ligated into pBluescriptII SK(−) that was digested with BamHI and treated with calf intestinal phosphatase (CIP). The ligated products were introduced into competent E. coli cells by the heat shock method according to the manufacturer's protocol (Bio-Rad, USA). The successful cloning of tal genes in pBluescript II was validated by restriction digestion, colony hybridization and sequence analysis.
The repeat units in tal genes complicate abilities to sequence the genes using conventional approaches. Thus, after cloning into pBluescript II SK(−), we used the EZ-Tn5™ < KAN-2 > Tnp Transposome™ Kit to insert Tn5 into each tal gene as recommended by the manufacturer (Epicentre, Madison, WI). Clones with Tn5 insertions in the middle of the repeat region were selected by SphI digestion and sequenced using primers pair tal-F/RP and FP/tal-R (Table S1).

TALEs phylogenetic tree construction and RVDs comparison
For TALEs phylogeny, available genome sequences of Xcm strains MSCT1, H1005, N1003, MS14003 and AR81009 were obtained from the NCBI. TALE genes were predicted and analyzed in each genome using AnnoTALE v1.4.1 [50]. DisTAL v1.1 were used to align and classify TALEs based on their central repeat region [72].
For the TALE RVDs analysis, we used AnnoTALE version 1.4.1. The TALEs are grouped into classes based on the RVDs that shows possible functional and evolutionary relationship [50,85].

Construction of Xss-V 2 -18 tal deletion mutants
The tal genes in Xcm Xss-V 2 -18 were deleted by homologous recombination using the suicide vector pKMS1 [60]. The 5′ and 3′ fragments that flank the CRR repeat in tal genes are conserved [66] and were used as sites for homologous recombination. The left-and right-flanking fragments of each tal gene were PCR-amplified using genomic DNA of Xcc strain Xcc049 (Table 1) as the template, and ligated into the MCS of pKMS1 [4], resulting in constructs pKMSA1 and pKMSA2, respectively. The new constructs were verified by restriction digestion and sequence analysis (TaKaRa, China). Constructs pKMSA1 and pKMSA2 were introduced into Xcm strain Xss-V 2 -18 by electroporation; cells were then plated on NAN medium supplemented with kanamycin and incubated at 28°C for 4 days. Single colonies were then cultured in NBN broth at 28°C to OD 600 ≤ 0.2 (~3 × 10 8 cells/mL), inoculated to NAS agar medium, and incubated for 2 days at 28°C. Single colonies that grew on NAS were then transferred to NA and NA containing kanamycin. Colonies that grew on NA, but not on NA Km , were selected as potential deletion mutants. The mutants were then analyzed by Southern blot hybridization and PCR with primer pairs pKMSA1-5F/pKMSA1-3R and pKMSA2-5F/ pKMSA2-3R (Table S1).

Southern hybridizations
Xcm plasmid and genomic DNA were extracted as described above. After BamHI digestion, DNA was separated on 1.2% agarose gels and then transferred onto Hybond N + nylon membranes (Roche, Germany). The 2898-bp internal SphI fragment of pthXo1 (GenBank accession no: AY495676) from Xoo [86] was labeled with digoxigenin (DIG) and used as a hybridization probe to detect the tal genes. Probe labeling and Southern blotting were performed using the DIG Probe Synthesis Kit as recommended by the manufacturer's instructions (Roche, Sweden).

Virulence assays
Cotton cultivar TM-1 (G. hirsutum) was used in this study. Plants were grown in a greenhouse at 23°C with a 12-h light/dark photoperiod and~80% RH. Two-weekold plants were used in virulence assays. Single colonies of Xcm were inoculated to 4 mL NB and cultured overnight at 28°C. Bacterial cells were harvested by centrifugation (5000 rpm, 3 min); pellets were washed twice in sterile 10 mM MgCl 2 and then re-suspended in 10 mM MgCl 2 buffer to OD 600 = 0.1 (~2 × 10 8 cells/mL). The suspensions were inoculated to the abaxial surface of leaves by infiltration with a sterile needleless syringe. Inoculation with simply 10 mM MgCl 2 buffer served as a mock. Leaf phenotypes were examined 4-5 days after inoculation. Three independent plants were used, and the experiments were repeated three times with similar results. For the quantification of bacterial growth in cotton, triplicate leaf samples (1cm 2 in diameter) were collected for each inoculated strain and washed with 70% ethanol and double-distilled water (ddw). Samples were macerated in 1 mL ddw and incubated for 30 min at room temperature. Serial dilutions were then plated on NB agar medium with appropriate antibiotics for colony counts. The experiment was repeated three times, and the significant differences were determined by using student's t-test.

Expression of tal2 gene in Xcm M2
The plasmid pZWavrXa7 (supplied by Dr. Bing Yang) was used to construct the plasmid for expression of tal2 in Xss-V 2 -18 strain. Plasmid pZWavrXa7 contains a FLAG-tag epitope immediately downstream of the second SphI site in the C-terminus of AvrXa7. The central SphI fragment of avrXa7 was replaced with the SphI fragment of Xss-V 2 -18 tal2 gene to generate pZW-tal2 ( Table 1). The recombinant plasmid was then fused with broad-host-range vector pHM1 at the HindIII site giving rise to pHZW-tal2. The constructs were transformed into Xcm M2 (Δtal2 strain, see below) by electroporation (2.5 kv, 4 ms).
The expression of tal2 in M2 was confirmed by western blotting with flag-labelled antisera. Briefly, the M2 strain containing pHZW-tal2 was cultured in NB to the logarithmic phase and harvested by centrifugation. The pellets were washed twice, and re-suspended in 1X PBS buffer to OD 600 = 1.0 (~3 × 10 9 cells/mL). SDS loading buffer (5X) was added to the bacterial suspensions and boiled in a water bath for 10 min. Proteins were separated on 8% SDS-PAGE gels and transferred to polyvinylidene difluoride membranes for immunoblotting using anti-FLAG (TransGene, Beijing, China) as the primary antibody. Primary antibodies were detected using goat anti-mouse IgG (H + L) (TransGen) and visualized with the EasySee Western Kit (TransGen). E. coli RNA polymerase subunit α (RNAP) was used as a loading control.
Additional file 1: Table S1. Primers used in this study. Figure S1.