By continuing to browse this site you agree to us using cookies as described in About Cookies
Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishing platforms. The new Wiley Online Library will be migrated over the weekend of February 24 & 25 and will be live on February 26, 2018. For more information, please visit our migration page: http://www.wileyactual.com/WOLMigration/
Author for correspondence: Thomas Lahaye Tel: +49 (0)89 2180 7470 Email: email@example.com
•Plant pathogenic bacteria of the genus Xanthomonas inject transcription activator-like effector (TALe) proteins that bind to and activate host promoters, thereby promoting disease or inducing plant defense. TALes bind to corresponding UPT (up-regulated by TALe) promoter boxes via tandemly arranged 34/35-amino acid repeats. Recent studies uncovered the TALe code in which two amino acid residues of each repeat define specific pairing to UPT boxes.
•Here we employed the TALe code to predict potential UPT boxes in TALe-induced host promoters and analyzed these via β-glucuronidase (GUS) reporter and electrophoretic mobility shift assays (EMSA).
•We demonstrate that the Xa13, OsTFX1 and Os11N3 promoters from rice are induced directly by the Xanthomonas oryzae pv. oryzae TALes PthXo1, PthXo6 and AvrXa7, respectively. We identified and functionally validated a UPT box in the corresponding rice target promoter for each TALe and show that box mutations suppress TALe-mediated promoter activation. Finally, EMSA demonstrate that code-predicted UPT boxes interact specifically with corresponding TALes.
•Our findings show that variations in the UPT boxes of different rice accessions correlate with susceptibility or resistance of these accessions to the bacterial blight pathogen.
Microbial plant pathogens deliver effector proteins into the host’s cytoplasm to promote their virulence or to suppress plant innate immunity (Göhre & Robatzek, 2008; Boller & He, 2009; Hogenhout et al., 2009). After delivery, microbial effectors are targeted to different subcellular compartments of the host cell. Recently it has become evident that the nucleus is targeted by effectors from various classes of plant microbial pathogens, including nematodes (Elling et al., 2007), oomycetes (Kanneganti et al., 2007), fungi (Kemen et al., 2005) and bacteria (Deslandes et al., 2003; Nissan et al., 2006; Bai et al., 2009). Transcription activator-like effectors (TALes) from the plant pathogenic bacterial genus Xanthomonas are among the most intensively studied class of nuclear-targeted microbial effectors (Kay & Bonas, 2009; White et al., 2009). The most characteristic structural feature of TALes is the central repeat domain that is composed of a variable number of tandemly arranged, imperfect copies of a 34/35-amino acid motif (Schornack et al., 2006). Differences between individual repeat units are found primarily at positions 12 and 13, the so-called repeat-variable diresidues (RVDs) (Moscou & Bogdanove, 2009). The repeat domain of the prototype TALe, AvrBs3, from Xanthomonas campestris pv. vesicatoria (Xcv), has been shown to interact with a corresponding promoter element, termed an UPA (up-regulated by AvrBs3) box, that is present in the promoter of the pepper transcription factor UPA20, a host susceptibility gene that appears to support bacterial spread (Kay et al., 2007). The presence of a UPA box in a promoter results in AvrBs3-mediated expression of the given host gene (Kay et al., 2009; Römer et al., 2009b). The promoter of the pepper resistance (R) gene Bs3 also contains a UPA box and thus is transcriptionally activated by AvrBs3 (Römer et al., 2007, 2009b). Expression of Bs3 triggers a cell death reaction, referred to as the hypersensitive response (HR), and results in resistance against Xcv. Thus, the R gene Bs3 represents a ‘promoter trap’ that coopts AvrBs3’s function in promoting virulence. Similarly, transcription of the rice R gene Xa27 is specifically induced by AvrXa27, a TALe from the bacterial blight pathogen, X. oryzae pv. oryzae (Xoo) (Gu et al., 2005). Recent studies uncovered that the Xoo TALe AvrXa27 binds to a matching promoter motif in the rice Xa27 promoter (Römer et al., 2009a). Thus the R genes Bs3 and Xa27 use identical mechanisms to detect their matching TALes. Promoter motifs that mediate TALe transcriptional activation have been collectively defined as UPT (up-regulated by TALes) boxes, with a subscript designation to define the specific TALe that targets the given UPT box (Römer et al., 2009a).
Although it was long known that TALe target specificity is defined by the number and order of repeat units that together form the repeat domain (Herbers et al., 1992), it was not clear how the repeat domain conferred target specificity at the molecular level. Recent studies demonstrated that RVDs specify the nucleotide target site of a given TALe with one RVD pairing to one specific UPT box nucleotide (Boch et al., 2009; Moscou & Bogdanove, 2009). This pairing code defined the interaction of TALes to colinear binding sites and was used to deduce functional UPT boxes for such TALes for which a corresponding host target promoter was not available (Boch et al., 2009).
Recently, a number of genes have been identified in rice that are targeted and transcriptionally activated by specific Xoo TALes to promote virulence of Xoo (Chu et al., 2006; Yang et al., 2006; Sugio et al., 2007; Antony et al., 2009; Yuan et al., 2009). The UPT boxes in the promoters of these rice genes have been predicted (Boch et al., 2009; Moscou & Bogdanove, 2009) by the use of the TAL code but not functionally validated. In the present study we analyzed if code predicted UPT boxes in the TALe-induced rice promoters of Xa13 (also known as Os8N3), OsTFX1 and Os11N3 are crucial to transcriptional activation by matching TALes. Furthermore we tested via electrophoretic mobility shift assay (EMSA) if TALes physically interact with the corresponding UPT boxes and how box mutations affect the TALe–DNA interaction. Our results show that resistance and susceptibility to Xoo in rice are influenced by UPT box sequences.
Materials and Methods
Generation of the promoter uidA fusion constructs
Promoter regions of OsTFX1, Os11N3, and Xa13 were PCR-amplified from genomic rice (Oryza sativa) DNA of cv IR24. The xa13 promoter region was amplified from genomic rice DNA of cv IRBB13. Amplification was carried out with Phusion high-fidelity DNA polymerase (New England Biolabs, Frankfurt, Germany) and primers provided in Supporting Information (Fig. S1). The PCR fragments were cloned into pENTR-D (Invitrogen GmbH, Karlsruhe, Germany), sequenced and transferred into the T-DNA vector pGWB3 (Nakagawa et al., 2007) by LR recombination (Invitrogen). pGWB3 derivatives were transformed into Agrobacterium tumefaciensGV3101 (Koncz & Schell, 1986) for in planta analysis.
Generation of the TALe constructs
For the generation of T-DNA vectors that contain avrXa7, pthXo1 or pthXo6, we used the vector pENTR-D-BamHI-avrXa27 (Römer et al., 2009a). The BamHI fragments of avrXa7, pthXo1 and pthXo6 were transferred into pENTR-D-BamHI-avrXa27, resulting in the pENTR-D-avrXa7, pENTR-D-pthXo1 and pENTR-D-pthXo6, respectively. The TALe genes were transferred via LR recombination in the binary vectors pGWB2 or pGWB5 (Nakagawa et al., 2007). pGWB2 and pGWB5 derivatives were transformed into A. tumefaciensGV3101 for in planta analysis. For EMSA we transferred pthXo1 and pthXo6 from pENTR-D-pthXo1 and pENTR-D-pthXo6 via LR recombination into pDEST17 (Invitrogen).
Insertion of UPT boxes in the Bs3 promoter
For the insertion of the predicted UPT boxes in the Bs3 promoter 5′ upstream of the UPTAvrBs3 box we used primers Xa13in30R-fwd-01-PR GATATNCATCTCCCCCT-ACTGTACACCACCAACTGGTTAAACAATGAACAC-GTTTGC, Xa13in30R-fwd-02-T-PR GATAGCATCT-CCCCCTACTGTACACCACCAACTGGTTAAACAATGAACACGTTTGC, OsTFX1in30R-fwd-01-PR ACCC-TATAAAAGGCCCTCACCAACCCATCGCCTGGTT-AAACAATGAACACGTTTGC, OsTFX1in30R-fwd-02-T-PR ACCCATAAAAGGCCCTCACCAACCCATCGC-CTGGTTAAACAATGAACACGTTTGC, Os11N3in-30R-fwd-03-PR GCACTATATAAACCCCCTCCAACC-AGGTGCTAAGCTCCTGGTTAAACAATGAACACG, Os11N3in30R-fwd-04-T-PR GCACATATAAACCCCC-TCCAACCAGGTGCTAAGCTCCTGGTTAAACAATGAACACG in combination with the primer 4in30R-rev-02-PR GGTGTGCAAATTGTGGTTTAACCC. All primers used are phosphorylated at their 5′ termini. Insertion was done using the Phusion site directed mutagenesis kit (New England Biolabs). As a template, we used pENTR-D containing 343 bp 5′ of the ATG start codon of the Bs3 gene. The promoter was amplified from genomic DNA of ECW-30R pepper plants using the Phusion high-fidelity DNA polymerase. After sequencing, the promoter constructs were transferred by LR recombination in the binary vector pGWB3 (Nakagawa et al., 2007). pGWB3 derivatives were transformed into A. tumefaciensGV3101 (Koncz & Schell, 1986) for in planta analysis.
Electrophoretic mobility shift assay (EMSA)
Electrophoretic mobility shift assays were carried out as described earlier (Römer et al., 2009a).
β-Glucuronidase (GUS) measurements
Leaves of three Nicotiana benthamiana plants were inoculated with a mixture of Agrobacterium delivering constructs for expression of TALes and the promoter-GUS reporter. Twenty-seven or 48 h postinoculation, two leaf discs (1 cm diameter) from separate infiltration spots of the same constructs on one plant were combined, ground in liquid nitrogen, and GUS assays were done as described previously (Kay et al., 2007). Samples were measured in a plate reader at 360 nm (excitation) and 465 nm (emission) with 4-methyl-umbelliferon (MU) (Carl Roth, Karlsruhe, Germany) dilutions as standard. Proteins were quantified using Bradford assays (Bio-Rad). Triplicate samples from three different plants were combined into one data point. In parallel, leaf discs from inoculated areas were sampled and incubated overnight in X-Gluc staining solution (Schornack et al., 2005). Leaf discs were cleared in 100% ethanol and dried using cellulose foil. Experiments were performed at least twice with similar results.
The promoters of the rice genes Xa13, OsTFX1 and Os11N3 are direct targets of the Xoo TALes PthXo1, PthXo6 and AvrXa7, respectively
Recent studies uncovered that the Xoo TALes PthXo1, PthXo6 and AvrXa7 transcriptionally activate the rice Xa13 (synonym: Os8N3), OsTFX1 and Os11N3 genes, respectively (Chu et al., 2006; Yang et al., 2006; Sugio et al., 2007; Antony et al., 2009; Yuan et al., 2009). To test if the rice OsTFX1, Os11N3 and Xa13 promoters are direct TALe targets, we amplified the corresponding promoter fragments from rice genomic DNA and cloned these in a T-DNA vector in front of an uidA reporter gene (Figs 1a, S1). The promoter::uidA fusion constructs were delivered into N. benthamiana leaves via transient A. tumefaciens-mediated T-DNA transformation (agroinfiltration) in combination with the cauliflower mosaic virus 35S (35S) promoter-driven TALe genes pthXo1, pthXo6, avrXa7 and avrBs3. GUS assays showed that the rice OsTFX1 and Os11N3 promoters are activated specifically by the matching Xoo TALes PthXo6 and AvrXa7, respectively, but not by the related Xcv TALe AvrBs3 (Fig. 1b,c). Furthermore, the GUS assays showed that the Xoo TALe PthXo1 transcriptionally activates only the rice Xa13 promoter from the rice cv IR24 but not the allelic xa13 promoter from the Xoo-resistant rice cv IRBB13 (Fig. 1d). Our GUS assays are in agreement with previous studies showing that Xoo delivering PthXo1 activates only expression of Xa13 but not xa13 alleles (Chu et al., 2006; Yuan et al., 2009). In our GUS assays, the pepper Bs3 promoter was not activated by any of the Xoo TALes but it was activated by the Xcv TALe AvrBs3 (Fig. 1b–d). These data demonstrate that the OsTFX1, Os11N3 and Xa13 promoters are direct targets of the Xoo TALes PthXo6, AvrXa7 and PthXo1, respectively.
TALes target the in silico predicted UPT boxes
We used the TALe code (Boch et al., 2009; Moscou & Bogdanove, 2009) to predict the UPTPthXo6, UPTAvrXa7 and UPTPthXo1 boxes of the rice OsTFX1, Os11N3 and Xa13 promoters, respectively (Figs S1, S2, Table S1). Regions potentially encompassing the distinct UPT boxes were introduced into the pepper Bs3 promoter and cloned in front of an uidA reporter gene. The Bs3 promoter-embedded UPT boxes were agroinfiltrated into N. benthamiana leaves in combination with the 35S promoter-driven TALe genes pthXo1, pthXo6, avrXa7 or avrBs3. GUS assays showed that a Bs3 promoter derivative containing a given UPT box is transcriptionally activated only by the matching Xoo TALe (Figs 2–4). For example, insertion of the UPTPthXo6 box from the rice OsTFX1 into the pepper Bs3 promoter (OsTFX1 in Bs3, Fig. 2b) made this promoter PthXo6- but not PthXo1-inducible. By contrast, the Bs3 wild-type promoter (Bs3) that lacks the UPTPthXo6 box was only AvrBs3- but not PthXo6-inducible. Similarly, insertion of the UPTAvrXa7 and UPTPthXo1 boxes into the Bs3 promoter resulted in promoter constructs that were AvrXa7- and PthXo1-inducible, respectively (Figs 3b, 4b). All Bs3 promoter derivatives contain the UPTAvrBs3 box and thus were AvrBs3-inducible, irrespective of whether or not a Xoo TALe box was present (Figs 2b, 3b, 4b). In summary, the TALe code enabled the identification of UPT boxes from rice promoters that are transcriptionally up-regulated by corresponding Xoo TALes.
Mutation of the conserved 5′ terminal T nucleotide of UPT boxes results in reduced TALe-mediated inducibility
All UPT boxes that have been predicted with the TALe code are preceded by a 5′ terminal T nucleotide (Boch et al., 2009; Moscou & Bogdanove, 2009). Mutations in the 5′ terminal T nucleotide of the UPTAvrBs3 or UPTHax3 box resulted in reduced inducibility by the matching TALe (Boch et al., 2009; Römer et al., 2009b). To study the functional importance of the 5′ terminal T nucleotide of the UPTPthXo6, UPTAvrXa7 and UPTPthXo1 boxes, we created T deletion mutants (ΔT) of the corresponding Bs3-promoter-embedded UPT boxes and cloned these in front of an uidA reporter gene. The UPT box ΔT mutants were delivered into N. benthamiana leaves via agroinfiltration in combination with the 35S promoter-driven TALe genes pthXo6, avrXa7, pthXo1 or avrBs3. Qualitative GUS assays showed that promoters containing the ΔT mutants of the UPTPthXo6, UPTAvrXa7 or UPTPthXo1 boxes were still induced by their matching TALes (OsTFX1ΔT in Bs3 (Fig. 2b), Os11N3ΔT in Bs3 (Fig. 3b) and Xa13ΔT in Bs3 (Fig. 4b)). However, quantitative GUS assays demonstrated that the three tested ΔT mutants in all cases produced a significantly reduced GUS activity in comparison to the wild-type UPT boxes (Figs 2c, 3c, 4c). Thus the 5′ terminal T nucleotide is important to the function of the UPTPthXo6, UPTAvrXa7 and UPTPthXo1 boxes.
Rice Xa13 and xa13 alleles differ in the predicted UPTPthXo1 box
Molecular analysis of a collection of rice xa13 and Xa13 rice genotypes uncovered that the pthXo1 expressing Xoo strain PXO99 transcriptionally activates only Xa13 but not xa13 genotypes (Chu et al., 2006; Yuan et al., 2009). We anticipated that Xa13 and xa13 genotypes are likely to differ in their UPTPthXo1 box region. Sequence analysis revealed that the PthXo1-inducible Xa13 alleles from rice cvs IR24, IR64, Nipponbare, Minghui and 93-11 were sequence identical within the UPTPthXo1 box (Figs S3, S4). By contrast, all studied xa13 alleles differed from the Xa13 alleles within the UPTPthXo1 box. In several xa13 alleles, the integrity of the UPTPthXo1 box was lost as a result of nucleotide insertions or deletions. For example, the xa13 alleles from rice cv AC 19-1-1 and Kalimekri 77-5 have lost 3′ terminal nucleotides of the UPTPthXo1 box as a result of a 34 bp deletion with respect to the IR24 Xa13 allele (Fig. S4). We also identified five xa13 genotypes (Tepa1, BJ1, Chinsurah 11484, Chinsurah 11760 and Chinsurah 50930) that showed only a G→T substitution in the second box nucleotide with respect to the UPTPthXo1 box from IR24 (Fig. S4a). According to the TALe code the second nucleotide of the UPTPthXo1 box is bound by the first PthXo1 repeat unit, which contains an NN-type RVD. Experimental studies with an in vitro constructed TALe consisting of NN-type RVDs only have shown that NN recognizes preferentially G (Boch et al., 2009). To clarify how polymorphisms in the second nucleotide of the UPTPthXo1 box influence PthXo1-mediated promoter activation, we replaced the G nucleotide of the Xa13 allele by A, C or T nucleotides and tested the activity of these boxes in the context of the Bs3 promoter (Fig. 4c). Quantitative GUS assays showed that G→A, G→C or G→T exchanges of the second box nucleotide resulted in significantly reduced PthXo1 inducibility in comparison to the nonmutated IR24 UPTPthXo1 box (Fig. 4c). Thus these experimental findings provide further support for the TALe code.
PthXo1 and PthXo6 bind in EMSA to matching UPT boxes
Previous studies have shown that the TALes AvrBs3, AvrBs3Δrep16 and AvrXa27 bind specifically to their matching UPT boxes (Römer et al., 2009a,b). Here we carried out EMSA to clarify if PthXo1 and PthXo6 would also bind specifically to their matching UPT boxes. EMSA showed that a His::PthXo1 fusion protein binds to a biotin-labeled Xa13 (cv IR24) promoter fragment containing the UPTPthXo1 box and, to a lesser extent, to the corresponding promoter region of the xa13 allele (cv IRBB13) (Fig. 5). Importantly, binding of His::PthXo1 to biotin-labeled Xa13 promoter fragments could be readily out-competed by nonlabeled Xa13 promoter fragments, whereas even a 100-fold excess of nonlabeled xa13 promoter fragments could not out-compete the binding (Fig. 5). Similarly, His::PthXo6 binds in EMSA to a biotin-labeled OsTFX1 promoter fragment containing the UPTPthXo6 box and, to a much lesser extent, to a mutated OsTFX1 promoter fragment (OsTFX1ΔT) that lacks the 5′ terminal T nucleotide of the UPTPthXo6 box (Fig. 6). Competition assays with biotin-labeled OsTFX1-derived promoter fragments and unlabeled OsTFX1 and OsTFX1ΔT promoter fragments further confirmed that His::PthXo6 has high affinity to the UPTPthXo6 box and only a very low affinity to a UPTPthXo6 box mutant variant that lacks the 5′ terminal T nucleotide (Fig. 6). Together these findings indicate that PthXo1 and PthXo6 bind specifically to their matching UPT boxes.
The TALe code and its limitations
We have demonstrated that the rice promoters Xa13, OsTFX1 and Os11N3 are activated by the Xoo TALes PthXo1, PthXo6 and AvrXa7, respectively (Fig. 1). Furthermore, we demonstrated that code-predicted UPT boxes are functional in the context of the pepper Bs3 promoter (Figs 2–4) and that TALes interact physically with code-predicted UPT boxes (Figs 5, 6). Given that functional UPT boxes could be reliably predicted for promoters that are known to be activated by given TALes, the question arises whether functional UPT boxes can also be identified from sequenced host genomes. One obvious limitation of the current version of the TALe code is that RVDs with low frequency of occurrence in sequenced TALes (e.g. HI, SS, NQ, NC and NV) have not yet been deciphered, although their specificity should be readily determined. The major limitation of the TALe code, then, is the uncertainty of the functional consequences of mismatches between UPT box nucleotides and individual RVDs. In this context it needs to be noted that our previous study on the TALe code (Boch et al., 2009) was focused on in vitro generated UPT boxes that show no or very few mismatches with respect to the given TALe. By contrast, all identified natural UPT boxes in plant promoters and their matching TALes contain many mismatches. For example, the TALe PthXo1 contains three NI-type RVDs that do not match the code-predicted A in the UPTPthXo1 box of the PthXo1-inducible Xa13 promoter (see PthXo1 repeat units 10 (NI→C), 19 (NI→C) and 21 (NI→C); Fig. S2). Whereas some mismatches have little effect on the magnitude of transcription activation, other mismatches have proved to be critical to TALe-mediated promoter activation. One striking example is the PthXo1-inducible Xa13 gene from the rice cv IR24 and the allelic, non PthXo1-inducible xa13 gene from the rice cv Tepa1. These Xa13/xa13 alleles differ only in a G→T substitution of the second nucleotide of the UPTPthXo1 box which pairs to a NN-type repeat (Figs S2, S4). Reverse transcription polymerase chain reaction (RT-PCR) analysis of rice leaf tissue that was infected with a pthXo1-expressing Xoo strain revealed transcriptional activation of the IR24 Xa13 but not the Tepa1 xa13 allele (Chu et al., 2006). Similarly, agroinfiltration assays revealed a significantly reduced PthXo1-mediated transcriptional activation of the Tepa1 xa13 allele as compared with induction of the IR24 Xa13 allele (Fig. 4c; the Tepa1 xa13 allele corresponds to ‘Xa13 mut T in Bs3’). This strong effect of a single mismatched NN-type repeat is somewhat unexpected considering that the UPTPthXo1 box of the IR24 Xa13 promoter, which mediates PthXo1-mediated promoter activation, contains seven mismatches compared with the code-predicted UPTPthXo1 box (Fig. S2). Thus it seems that correct pairing of the second (NN-type) repeat of PthXo1 is crucial in the context of the PthXo1–UPT box interaction than correct pairing of other RVDs.
We postulate that the sum of RVDs that pair to code-predicted nucleotides determines the overall affinity of a TALe to a given UPT box, with a minimum number of matching RVDs required to promote TALe-mediated transcriptional activation. This hypothesis is supported by the observation that longer TALes appear to tolerate more mismatches than shorter TALes. For example, AvrXa7 (26 repeat units) and PthXo1 (24 repeats units) transcriptionally activate the rice Os11N3 and Xa13 promoter despite the fact that there are eight and seven mismatches in the corresponding UPTAvrXa7 (Os11N3 promoter) and UPTPthXo1 (IR24 Xa13 promoter) boxes, respectively (Fig. S2). By contrast, the UPT boxes that are targeted by the shorter TALes AvrHah1 (14 repeats units; activates Bs3 promoter) (Schornack et al., 2008) and AvrBs3Δrep16 (14 repeats units; activates Bs3-E promoter) (Römer et al., 2007, 2009b; Boch et al., 2009; Moscou & Bogdanove, 2009) each contains a single mismatch as compared with the code-predicted UPT boxes.
Although longer TALes seem to target UPT boxes with multiple mismatches, it is conceivable that longer TALes also require a minimum number of RVDs that pair to matching nucleotides in order to promote transcriptional activation. Given that the UPTPthXo1 box from the IR24 Xa13 promoter contains seven mismatches as compared with the code-predicted UPTPthXo1 box (Fig. S2), one might speculate that any additional mismatch will result in reduced inducibility of the given box. Thus the reduced inducibility of the Tepa1 xa13 allele (G→T substitution of the second nucleotide of the UPTPthXo1 box) might be a consequence of the reduced overall affinity of PthXo1 to the Tepa1 xa13 allele and does not necessarily imply that correct pairing of this particular RVD is crucial to the TALe–UPT box interaction.
In summary, TALes target not only code-predicted UPT boxes but also closely related boxes. However, the functional consequences of mismatches between UPT box nucleotides and corresponding RVDs remain, to some extent, unpredictable. It remains to be clarified if all RVDs make an equal contribution to the TALe–DNA interaction or if certain RVDs are of particular importance. Obviously a crystal structure of a TALe and its matching UPT box will help to give further insights into the molecular basis of this interaction.
The 5′ terminal T of the UPT boxes is crucial to transcriptional activation by, and interaction with, its matching TALe
Previous studies uncovered that all functional UPT boxes contain a conserved, invariant 5′ terminal T nucleotide (Boch et al., 2009; Moscou & Bogdanove, 2009). Mutational studies of the conserved T in the UPT boxes of the TALes AvrBs3 and Hax3 resulted in reduced induction of the corresponding promoter mutant derivatives as compared with the promoters containing the conserved T nucleotide (Boch et al., 2009; Römer et al., 2009b). Analogously, our studies showed that a mutation in the conserved 5′ terminal T nucleotide of the PthXo1, PthXo6 and AvrXa7 UPT boxes also resulted in reduced inducibility of the corresponding rice Xa13, OsTFX1 and Os11N3 promoters (Figs 2c, 3c, 4c). Thus the functional relevance of the conserved 5′ terminal T nucleotide has by now been confirmed for five different TALes, suggesting that the invariant T is crucial to the function of most, or possibly all, UPT boxes.
Previous EMSAs on the TALe AvrBs3Δrep16 suggested that the 5′ terminal T nucleotide of the corresponding pepper Bs3-E promoter UPTAvrBs3Δrep16 box makes a significant contribution to the TALe–DNA interaction (Römer et al., 2007, 2009b). However, an EMSA-based comparison of identical DNA fragments that contain or lack the conserved 5′ terminal T nucleotide of a UPT box had not yet been carried out. We compared by EMSA the affinities of the wild-type UPTPthXo6 box from the rice OsTFX1 promoter and a corresponding mutant box lacking the conserved T nucleotide (OsTFX1ΔT), and found a drastically reduced interaction between PthXo6 and the mutant box as compared with the wild-type UPTPthXo6 box (Fig. 6). These findings demonstrate that the 5′ terminal T nucleotide of the UPTPthXo6 box is crucial to physical interaction between PthXo6 and the UPTPthXo6 box. Given that similar findings have been observed for the TALe AvrBs3Δrep16 (Römer et al., 2007, 2009b), it seems likely that, in general, the 5′ terminal T nucleotide of a UPT box is crucial to its physical interaction with a corresponding TALe. Future studies will have to clarify which TALe residues pair to the conserved T. Once this question is resolved, we may be able to modify TALes in such a way that pairing to nucleotides other than a 5′ terminal T is possible.
We thank Bing Yang and Frank White for providing Xoo TALe genes. We would like to acknowledge the technical support of Carola Kretschmer, Bianca Rosinsky, Marina Schulze and Heidi Scholze. Work in our laboratory has been supported by grants of the 2Blades foundation, the Exzellenznetzwerk Biowissenschaften (Ministry of Culture of Saxonia-Anhalt) and the Deutsche Forschungsgemeinschaft (SPP1212, SFB 648 and LA1338/2-2). We are grateful to Diana Horvath for helpful comments on the manuscript.