The Brassicaceae-Specific EWR1 Gene Provides Resistance to Vascular Wilt Pathogens

Soil-borne vascular wilt diseases caused by Verticillium spp. are among the most destructive diseases worldwide in a wide range of plant species. The most effective means of controlling Verticillium wilt diseases is the use of genetic resistance. We have previously reported the identification of four activation-tagged Arabidopsis mutants which showed enhanced resistance to Verticillium wilt. Among these, one mutant also showed enhanced resistance to Ralstonia solanacearum, a bacterial vascular wilt pathogen. Cloning of the activation tag revealed an insertion upstream of gene At3g13437, which we designated as EWR1 (for Enhancer of vascular Wilt Resistance 1) that encodes a putatively secreted protein of unknown function. The search for homologs of Arabidopsis EWR1 (AtEWR1) in public databases only identified homologs within the Brassicaceae family. We subsequently cloned the EWR1 homolog from Brassica oleracea (BoEWR1) and show that over-expression in Arabidopsis results in V. dahliae resistance. Moreover, over-expression of AtEWR1 and BoEWR1 in N. benthamiana, a member of the Solanaceae family, results in V. dahliae resistance, suggesting that EWR1 homologs can be used to engineer Verticillium wilt resistance in non-Brassicaceae crops as well.


Introduction
Verticillium species belong to the phylum Ascomycota that comprises the largest group of fungal pathogens and contains several plant pathogenic species such as V. dahliae and V. longisporum [1], [2], [3]. While V. dahliae has an extremely broad host range that contains hundreds of mainly dicotyledonous plant hosts, V. longisporum is pathogenic on Brassicaceae only. Verticillium spp. are soil-borne and cause vascular wilt diseases [1], [2], [4]. Controlling Verticillium wilt diseases is difficult for several reasons: Verticillium spp. produce resting structures that can survive in the soil for many years [5], and soil fumigation has largely been banned due to environmental concerns. A commonly used alternative control method, crop rotation, is ineffective due to the wide host range of the pathogen. Finally, once into the xylem, the fungus is not affected by fungicides. Consequently, the preferred method to control Verticillium wilt disease is the use of genetic resistance.
Genetic resistance against Verticillium wilt diseases has been reported for several crop species [1], [6]. The first Verticillium wilt resistance locus that has been cloned and functionally characterized is the tomato Ve locus that contains the Ve1 gene that provides resistance in tomato against race 1 isolates of V. dahliae [7], [8].
Recently, it was shown that transgenic expression of tomato Ve1 in Arabidopsis provides resistance against Verticillium race 1 isolates [9]. Over the years, Arabidopsis has increasingly been used as a model host to study Verticillium-host interactions [1], [10], [11], [12], [13], [14]. In addition to screening germplasm for resistance [12], [15], mutagenesis followed by screening for enhanced resistance with a pathogen of interest is a means to identify novel resistance traits. Several molecular techniques have been used to generate random mutants in Arabidopsis, such as EMS-and radiation-induced mutation, and transposon and activation tagging. Activation tagging involves the random integration of promoter or enhancer sequences in a genome, using either a T-DNA or a transposon, generally leading to enhanced expression of genes near the integration site and generating gain-of-function mutations [16], [17], [18]. Insertion of enhancer sequences in the genome may positively regulate gene expression, even when inserted at a considerable distance to the target gene [19]. Some of the advantages of activation tagging over knock-out strategies include that activation tagging generates dominant instead of recessive mutations, it generates viable mutants for those genes where knock-outs would lead to lethal phenotypes and it is also applicable to dissect phenotypes of redundant genes [18].
Transposon-based activation tagging has been successfully employed in various plant species to identify novel genes involved in various physiological processes [17], including pathogen defence [20], [21], [22]. In an attempt to identify sources of Verticillium wilt resistance using Arabidopsis, we have screened an activation-tagged Arabidopsis mutant collection with V. dahliae. Previously, we have reported the identification of four mutants with enhanced resistance to Verticillium wilt disease [23]. Here, we pursued functional characterization of one of the mutants and demonstrate that enhanced activation of the At3g13437 gene, encoding a protein of unknown function, is responsible for the enhanced Verticillium wilt resistance phenotype. This gene is designated as Enhancer of vascular Wilt Resistance 1 (EWR1).

Identification of the Enhancer of vascular Wilt Resistance 1
Previously, we have reported the identification of four activation-tagged Arabidopsis mutants, A1 to A4, that displayed enhanced resistance to Verticillium wilt disease [23]. Of these, mutant A2 not only displayed resistance to V. dahliae (Figure 1), but also to the bacterial vascular wilt pathogen Ralstonia solanacearum [23]. Here, we investigated this mutant further and determined the position of the activation tag insertion site using thermal asymmetric interlaced PCR (TAIL-PCR) [24]. The tag was found to be inserted on chromosome 3 at a position 376 bp upstream of the translational start codon of gene At3g13435. To identify the gene responsible for the enhanced Verticillium resistance of the A2 mutant, we analysed the expression of genes flanking the T-DNA insertion site to detect transcriptional changes. The analysis of expression of 11 genes spanning a region of ,14 kb upstream to ,17 kb downstream of the activation tag insertion site showed that four of these genes, namely At3g13432, At3g13435, At3g13437 and At3g13445, were induced in the A2 mutant when compared to wild-type plants ( Table 1, Figure S1 in File S1). Simultaneously, we analysed homozygous knock-out alleles of all genes flanking the activation tag insertion site for susceptibility towards V. dahliae. Interestingly, the knock-out allele of At3g13437 showed clearly enhanced susceptibility to V. dahliae (Figure 2A), whereas none of the other knock-out alleles showed enhanced susceptibility to Verticillium wilt when compared to the Col-0 wild-type (Table 1). Therefore, we tentatively named At3g13437 EWR1, for Enhancer of vascular Wilt Resistance 1. To validate the enhanced susceptibility of the knock-out allele of At3g13437, ewr1, we quantified the ratio of rosette leaves showing Verticillium wilt symptoms at 14 and 20 days post inoculation (dpi), showing that the percentage of diseased rosette leaves of the ewr1 mutant is significantly higher when compared with wild-type plants ( Figure 2B). We further validated the enhanced susceptibility of ewr1 by quantifying V. dahliae colonization of the rosette leaves using real-time PCR. As expected, more fungal DNA was detected in ewr1 plants when compared with wild-type plants ( Figure 2C). Overall, the gene expression data, combined with the Verticillium wilt phenotyping, strongly suggests that enhanced expression of EWR1 causes enhanced Verticillium wilt resistance in the activation-tagged mutant A2.
We have previously shown that the induction of AHL19, which encodes an AT-hook DNA binding protein, is causal to the enhanced Verticillium wilt resistance in the A1 mutant [23]. To investigate whether EWR1 over-expression can explain the enhanced Verticillium wilt resistance of the A3 and A4 mutants, we assessed EWR1 expression in these mutants in absence of pathogen challenge. This analysis showed that EWR1 is not overexpressed in these mutants ( Figure S2 in File S1), showing that constitutive activation of EWR1 cannot explain the enhanced Verticillium wilt resistance in the A3 and A4 mutants.

EWR1 over-expression provides resistance to Verticillium wilt
To corroborate whether the enhanced expression of EWR1 is causal to the enhanced Verticillium wilt resistance of mutant A2, we generated EWR1 over-expressing lines in Arabidopsis ecotypes Col-0 and WS. Similar to the activation-tagged mutant A2, which displays compact and rounded rosette leaves with short petioles [23], also EWR1 over-expressing plants displayed altered plant morphology ( Figure 3A, Figure S3A in File S1). EWR1 overexpressing plants show compact, dark green and slightly thicker leaves than wild-type plants with short petioles, and shorter and fewer inflorescences ( Figure S4 in File S1).
In order to learn if EWR1 over-expression contributes to Verticillium wilt resistance, three independent EWR1 overexpressing lines of the Col-0 ecotype (EWR1-1, EWR1-2, and EWR1-3) were challenged with V. dahliae along with ewr1 and wildtype plants. While Col-0 and ewr1 plants showed clear wilting, chlorosis, and stunting symptoms at 14 dpi that were significantly increased by 21 dpi, EWR1-expressing plants showed only mild symptoms at these time points ( Figure 3A; Figure S3A in File S1). Thus, when assessing the impact of V. dahliae inoculation by comparing mock-inoculated and V. dahliae-inoculated plants for each of the genotypes it is evident that EWR1 over-expressing lines show relatively little impact of the pathogen. The degree of stunting induced by V. dahliae on wild-type plants (when comparing mock-inoculated and V. dahliae-inoculated Col-0 plants) and on ewr1 plants (when comparing mock-inoculated and V. dahliaeinoculated ewr1 plants) exceeds the degree of stunting induced by V. dahliae on EWR1 over-expressing lines (when comparing mockinoculated and V. dahliae-inoculated EWR1 plants) by far. Moreover, quantification of fungal colonization in planta using real-time PCR showed only little V. dahliae biomass in EWR1transgenic lines ( Figure 3C). Similar phenotypes were observed on EWR1 over-expressing lines of the WS ecotype ( Figure S3 in File S1). These data further confirm that the constitutive activation of EWR1 expression is causal to the enhanced Verticillium resistance of the A2 mutant.

Transcriptional regulation of EWR1
We showed that constitutive over-expression of EWR1 enhances Arabidopsis resistance to Verticillium wilt disease ( Figure 1A, 3A, Figure S3A in File S1). To understand how EWR1 is regulated at transcriptional level during the course of the Verticillium-Arabidopsis interaction, we performed a time course experiment where we challenged the wild-type Col-0 and WS plants with V. dahliae. Subsequently, we assessed transcription of EWR1 using real-time PCR. This analysis showed that EWR1 expression is transiently induced upon V. dahliae inoculation in both Col-0 and WS ecotypes ( Figure S5 in File S1). Subsequently, the expression level of EWR1 was assessed in roots and shoots of non-inoculated WS, A2 mutant and EWR1-5 mutant plants. Except for EWR1-4, which showed slight induction, EWR1 expression was hardly detected in roots of WS and mutant A2 whereas in shoots, EWR1 was strongly expressed in mutant A2 and EWR1-4 when compared with wild-type plants ( Figure S6 in File S1).

EWR1 provides resistance to other vascular wilt pathogens
To investigate whether the enhanced Ralstonia resistance in the A2 mutant can similarly be attributed to EWR1 over-expression, we challenged two EWR1 over-expressing lines (EWR1-1 and  EWR1-2), along with the wild-type Col-0 and the ewr1 mutant with R. solanacearum strain GMI1000. While Col-0 plants showed mild disease symptoms at 3 dpi which aggravated by 6 dpi (Figure 4 A, B), resulting in death of the inoculated plants by 10 dpi, most rosette leaves of GMI1000-inoculated ewr1 plants showed clear wilting at 3 dpi and completely collapsed by 6 dpi, indicating that ewr1 plants show enhanced susceptibility to R. solanacearum. Conversely, EWR1 over-expressing plants were completely resistant to R. solanacearum and did not show any disease symptoms throughout the assay up to 10 dpi ( Figure 4A, B). With real-time PCR it was confirmed that the amount of disease symptoms observed on the various genotypes correlates with the degree of R. solanacearum colonization ( Figure 4C). Hardly any bacterial DNA was detected in DNA extracts of EWR1-1 and EWR1-2 rosette leaves, indicating that EWR1 over-expression provides a high level of R. solanacearum resistance.   We have previously shown that the activation-tagged mutant A2 displayed wild-type levels of susceptibility to Fusarium oxysporum f. sp. raphani [23]. In this study, we have also challenged the EWR1 over-expressing plants with F. oxysporum. While wild-type plants that were inoculated with F. oxysporum showed clear wilting of rosette leaves and overall stunting of the plants at 10 dpi which led to a complete collapse of the plants by 14 dpi, EWR1 overexpressing plants showed enhanced resistance to this pathogen ( Figure 4A, D). Inoculated EWR1-1 and EWR1-2 plants hardly showed any symptoms of disease throughout the assay. To explain the discrepancy in disease phenotypes between the A2 mutant and the EWR1 over-expressing lines, we compared the EWR1 expression levels in these plants, showing that EWR1 expression is significantly higher in the over-expression lines when compared to the A2 mutant ( Figure S2 in File S1).

EWR1 over-expression enhances drought tolerance
In addition to enhanced Verticillium wilt resistance, we investigated whether over-expression of EWR1 plays a role in drought stress tolerance. To this end, we tested the drought stress resistance of EWR1-2, the Col-0 wild-type and the ewr1 mutant. After 3 weeks of growth with regular watering, we stopped watering the plants and evaluated the response of the lines to drought stress. The assay showed that Col-0 plants and ewr1 mutants started to show wilting symptoms 10 days after the last watering, while the rosette leaves were collapsed after 14 days. In contrast, EWR1 expressing plants did not show any drought symptoms up to 14 days after the last watering ( Figure 5). We similarly evaluated the drought stress resistance of two EWR1 over-expressing lines in WS background along with the WS wildtype and the activation-tagged mutant A2 and found similar results. EWR1 over-expressing plants and A2 mutant showed enhanced drought tolerance when compared to the wild-type WS. This indicates that EWR1 over-expression enhances drought stress tolerance in Arabidopsis.

EWR1 encodes a protein of unknown function
The full-length genomic DNA sequence of EWR1 consists of 599 bp containing two exons of 97 and 116 bp separated by an intron of 87 bp ( Figure 6A). The ORF encodes a protein of 70 amino acids with a predicted molecular mass of 7.93 kDa which is annotated as unknown. With SignalP 3.0 (http://www.cbs.dtu.dk/ services/SignalP/) [25] it is strongly predicted that EWR1 contains a signal peptide of 21 amino acids. Searching for recognizable protein domains or signatures of conserved motifs in the EWR1 protein sequence in publicly available databases did not result in any significant hits.
Based on the sequence conservation, we designed primers to amplify the EWR1 homologue from yet another Brassicaceous species, Brassica oleracea, using genomic DNA. Indeed a homolog, designated BoEWR1, was amplified which reinforced the suggestion that EWR1 is a Brassicaceae-specific gene. We next aimed to test whether BoEWR1 expression in Arabidopsis also provides Verticillium wilt resistance. To this end, BoEWR1 was amplified from B. oleracea cDNA and constitutively expressed in the Arabidopsis Col-0 and WS ecotypes. As expected, BoEWR1 expressing plants displayed similar leaf morphology as AtEWR1 expressing plants ( Figure 7A). Subsequently, BoEWR1 expressing plants were challenged with V. dahliae, revealing enhanced resistance to Verticillium wilt when compared with wild-type plants ( Figure 7A, B). Real-time PCR analysis confirmed reduced Verticillium colonization on BoEWR1 expressing plants when compared with wild-type plants ( Figure 7C). These data show that the EWR1 homologs of Arabidopsis thaliana (AtEWR1) and Brassica oleracea (BoEWR1) are functional homologs with respect to their role in Verticillium wilt resistance.

AtEWR1 and BoEWR1 over-expression in N. benthamiana confers Verticillium wilt resistance
To investigate whether AtEWR1 over-expression results in Verticillium wilt resistance in non-Brassicaceae plant species, we over-expressed AtEWR1 and BoEWR1 in the Australian tobacco species Nicotiana benthamiana. This Solanaceous plant species has been used as a model system to study interactions with various plant pathogens [26]. Unlike in Arabidopsis, AtEWR1 or BoEWR1 over-expression did not cause any obvious changes in the morphology of N. benthamiana plants when compared to nontransgenic control plants, except from perhaps slight stunting ( Figure 8A, B). Based on the expression level of EWR1 ( Figure S7 in File S1), three independent AtEWR1 (AtEWR1-a, AtEWR1-b, and AtEWR1-c) and BoEWR1 (BoEWR1-a, BoEWR1-b, and BoEWR1-c) expressing T2 lines were selected and challenged with V. dahliae. Interestingly, AtEWR1 and BoEWR1 over-expressing N. benthamiana plants showed reduced Verticillium wilt symptoms when compared with inoculated wild-type plants ( Figure 8A; B). The wild-type plants showed severe wilting, stunting, and chlorosis symptoms at 7 dpi, and leaves are completely collapsed by 10 dpi, whereas AtEWR1 as well as BoEWR1 over-expressing plants showed only mild wilting symptoms only on the older, lower leaves at 10 dpi ( Figure 8A; B). When we counted the number of plants (n = 10 plants per assay) that showed any symptoms of Verticillium wilt disease at 10 dpi, irrespective of the severity of the symptoms, 100% of the wild-type plants showed signs of infection, whereas only 40 to 60% of the AtEWR1 and BoEWR1 over-expressing plants displayed symptoms ( Figure 8C). This indicates that EWR1 homologs from Brassicaceae species can be used to establish Verticillium wilt resistance in non-Brassicaceae plant species in the absence of significant developmental phenotypes.

Discussion
Arabidopsis has increasingly been used as a model host for the identification of Verticillium resistance sources and studying the molecular mechanisms of Verticillium-host interactions [9], [10], [11], [12, [13], [14], [27]. In a phenotypic screening for gain-offunction mutants, we have previously reported the identification of four activation-tagged Arabidopsis mutants (A1-A4) which showed enhanced Verticillium wilt resistance [23]. We have also reported that the specific activation of the gene encoding the AThook DNA binding protein AHL19 causes enhanced Verticillium resistance in the A1 mutant [23]. Here, we cloned the activationtag insertion site in the A2 mutant that showed enhanced resistance not only to Verticillium spp., but also to R. solanacearum. Among the 11 genes found within a 31 Kb window spanning the activation tag insertion site, four genes showed induced expression in the A2 mutant, of which At3g13437 (EWR1) showed the strongest over-expression when compared to wild-type plants. Intriguingly, only the KO allele of EWR1 showed increased Verticillium susceptibility when compared to wild-type, whereas over-expression of EWR1 in wild-type Arabidopsis provided resistance to V. dahliae, F. oxysporum and R. solanacearum. Interestingly, in addition to enhanced resistance towards the V. dahliae, V. albo-atrum and V. longisporum, mutant A2 shows wild-type susceptibility to the necrotrophic foliar pathogens B. cinerea, and P. cucumerina and to the bacterial foliar pathogen P. syringae [23]. This suggests that EWR1 over-expression does not lead to overall enhanced plant defence against a wide range of pathogens.
Soil-borne vascular wilt pathogens share several important features with respect to their biology and infection style. The pathogens enter their hosts via the roots, invade xylem vessels and spread rapidly to the aerial part of the plants [1], [2], [4], [28]. Thus, any plant resistance mechanism that prevents root penetration, xylem colonization or systemic spread could potentially contribute to resistance towards vascular wilt pathogens. We previously showed that the A2 mutant has similar root morphology as wild-type plants [23], suggesting that EWR1-mediated pathogen resistance is unlikely to be caused due to altered root morphology. Moreover, the stronger EWR1 induction in shoots than in roots of the A2 mutant suggests that EWR1-mediated resistance may occur in shoots rather than in roots. Previous studies in Arabidopsis as well as in tomato have also shown that Verticillium wilt resistance is established once the fungus has entered and colonized the xylem vessels [8], [9], [23], [28], [29]. Despite absence of a significant difference in root morphology between the A2 mutant and the wild type plants, a significant morphological alteration in the rosette of the A2 mutant, which was even stronger in the rosette of EWR1 over-expressing lines, was observed. The severity of the developmental phenotypes correlates with the level of AtEWR1 expression that is significantly higher in the AtEWR1 over-expression lines than in the A2 mutant ( Figure S2 in File S1). However, irrespective of the extent of phenotypical deviations, all EWR1-expressing plants were found to display resistance that is specific to vascular wilt pathogens and does not concern other (foliar) pathogens. Furthermore, the developmental phenotypes that are observed upon EWR1 overexpression are not observed upon over-expression in N. benthamiana, although vascular wilt resistance was maintained in these plants. Finally, all assays were performed in well-watered plants that did not experience drought stress, and no macroscopically visible signs of stress, such as anthocyanin accumulation, were observed. These findings suggest that the induction of developmental aberrancies and Verticillium wilt resistance can be uncoupled.
EWR1 encodes a mature protein of 49 amino acids with unknown function. Homologs are only found in Brassicaceae species, showing high sequence conservation at the C-termini and more diversity at the N-termini. The C-termini of the B. rapa, B. oleracea and S. irio homologs, but not of the A. lyrata homolog, contain two adjacent cysteine residues. Although cysteine residues are often implicated in disulphide bond formation to enhance protein stability, the adjacent localization of these residues in EWR1 makes intramolecular disulphide bond formation unlikely. However, possibly the cysteine residues might be involved in EWR1 homodimerization.
The search for recognizable protein domains in EWR1 did not result in any significant hits in publicly available databases. The absence of a functional annotation and any known motif or domain in EWR1 complicates the prediction of EWR1 function. The presence of an N-terminal signal peptide, an overall net positive charged (+2), and a relatively high number of hydrophobic amino acids (41%) are typical features that are shared with many antimicrobial peptides (AMPs) [30], [31], [32]. AMPs are found in all living organisms [32], [33]. In plants, six different AMPs families have been described, comprising thionins, defensins, lipid transfer proteins, knottins, heveins, and snakins, of which defensins are the largest group and best characterised [30], [31], [32], [33]. In Arabidopsis, 825 small cysteine-rich proteins with typical features of antimicrobial peptides have been predicted [34]. Several lines of evidence indicate that AMPs play role in plant defence against viral, bacterial and fungal pathogens [30], [31], [32], [33], [35]. AMPs are expressed in plants both constitutively and in response to pathogen attack [30], [36]. It has been shown that constitutive over-expression of AMPs increases plant defence against bacterial and fungal pathogens. For instance, the constitutive over-expression of the alfalfa defensin (alfAFP) in potato provides resistance against V. dahliae [37]. Similarly, constitutive expression of the radish defensin in tobaco and tomato, provides resistance against Alternaria longipes and Alternaria solani, respectively [30]. An in vitro EWR1 antimicrobial activity assay should answer whether EWR1 function as an AMP. So far, attempts for heterologous EWR1 protein production using Escherichia coli or the yeast Pichia pastoris system have not been successful.
Vascular wilt symptoms such as wilting, stunting, chlorosis and leaf defoliation are similar to those symptoms caused by drought stress. Indeed, the physical presence of vascular wilt pathogens in the xylem vessels, enzymes secreted by the fungus or plant defence responses may interfere with water transport in the xylem [28], [38]. In potato, it has been shown that Verticillium resistant potato cultivars exhibit drought stress tolerance [39]. We observed that EWR1 over-expressing Arabidopsis plants similarly show drought stress tolerance. Leaf morphology such as size, thickness and shape has direct implication on water loss through transpiration [40], [41]. EWR1 over-expressing plants have a smaller leaf size; have thicker and curly leaves than wild-type plants, which all can contribute to the amount of water loss through transpiration. Determining the effect of EWR1 over-expression on the number of open stomata and the amount of water loss through transpiration in EWR1 over-expressing plants when compared to the wild-type may provide insight in how EWR1 regulates drought stress resistance.
For years, it has been a major focus for plant breeders to identify effective and durable genetic resistance to a wide range of pathogens. However, most of the resistance genes identified so far are either race-or species-specific and thus provide resistance to a limited number of pathogens. Thus, to obtain effective and durable resistance, it requires the transfer of multiple resistant genes into a cultivar. Since most AMPs have both antibacterial and antifungal activities and can be used across eukaryotic kingdoms [30], [37], [42], [43], [44], [45], they can potentially be used for developing effective resistance in plants against a broad spectrum of pathogens. Here we identified an Arabidopsis gene, EWR1, possibly encoding an AMP, which is effective at least against three vascular wilt pathogens. Moreover, heterologous expression of AtEWR1 and BoEWR1 in N. benthamiana, a plant species that belongs to the Solanaceae, confers Verticillium wilt resistance, making EWR1 a potential gene to control vascular wilt pathogens in Brassicaceae and non-Brassicaceae plant species.

Determination of the activation-tag insertion site
The activation-tag insertion site in mutant A2 was determined using thermal asymmetric interlaced PCR (TAIL-PCR) [46]. The PCR was performed with a combination of nested primers [47] and 10-mer random primers [48]. The secondary and tertiary TAIL-PCRs were separated on 1.2% agarose gel, stained with ethidium bromide, and visualized using the ChemiDoc XRS system (Bio-Rad). Specific product, judged based on the size differences generated by the nested primers, was excised, cleaned using the QIAquick Gel Extraction Kit (QIAGEN), cloned into the pGEM-T Easy Vector (Invitrogen), and sequenced. Blastn search of the TAIR database using the PCR sequences was performed to identify the genomic insertion site. Based on the putative insertion site, the primer pair MPR15F and MPR15R were designed and used to amplify the flanking genomic region. By sequencing this region in the wild-type and the mutant A2, the exact insertion site was determined.

EWR1 over-expression
The EWR1 CDS was amplified with the primer pair dMRP15-F1 and dMRP15-R1 that contain BamHI and AscI restriction sites, respectively, using Pfu DNA polymerase (Promega). The amplicon was cloned into the BamHI-and AscI-pre-digested binary vector pmk40, a variant of the vector pmog800 [8], [49]. The resulting P35S:EWR1 vector construct was transformed into A. tumefaciens strain GV3101 and eventually in to WS and Col-0 Arabidopsis ecotypes using the floral dip technique [50].

Cloning of EWR1 homologs
Primer pair EVR1H-BrF0 and EVR1H-BrR1 was used to amplify BoEWR1 from genomic DNA (gDNA) of Brassica oleracea (Brussels sprout). The PCR product was excised from the gel, cleaned (GE Healthcare) and cloned into the pGMET-easy vector (Promega) and sequenced. Based on the sequence alignment of the PCR sequence and the B. rapa sequence in the database, primer EVR1H-BrR3 was designed and used in combination with EVR1H-BrF0 to amplify the predicted full length CDS of BoEWR1 from B. oleracea cDNA. As a control, the same primer combination was used to amplify BoEWR1 from gDNA. The PCR fragments were sequenced to confirm the full length CDS. To generate an BoEWR1 over-expression construct, the full length CDS of BoEWR1 was amplified from cDNA using primer pair EVR1H-BaF1 and EVR1H-AsR1 containing BamHI and AscI custom restriction sites, respectively, and cloned into BamHI and AscI pre-digested binary vector pB7K40 [23]. Subsequently, the binary vector construct was transformed into A. tumefaciens (strain GV3101) and eventually into Arabidopsis ecotypes WS and Col-0.

Expression of EWR1 homologs in N. benthamiana
In order to test whether expression of AtEWR1 and BoEWR1 results in Verticillium wilt resistance in non-Brassicaceae plants as well, the binary vectors containing AtEWR1 or BoEWR1 (described above) were transformed into N. benthamiana, a Solanaceae family member, following a standard N. benthamiana transformation protocol [51]. AtEWR1 and BoEWR1 transformed calli were selected on Kanamycin (50 mg/ml) and ammonium glufosinate (Basta = 25 mg/ml) plates, respectively. After root generation, about 20 independent transformants per constructs were transferred to a soil for seed production. Subsequently, transformants were tested with PCR for transgene using Kanamycin and Basta specific primers, respectively. T2 seeds were harvested and three PCR-positive lines were selected and used in the preliminary Verticillium assay. Before inoculation, both the wild-type, AtEWR1, and BoEWR1 over-expressing N. benthamiana plants were grown for four weeks in a greenhouse. Subsequently, plants were carefully uprooted, the roots were rinsed in water, and eventually inoculation was performed by root-dipping method as described for tomato and Arabidopsis [9], [23].

Pathogen quantification in planta
Real-time PCR was used for quantification of pathogen colonization in planta using an ABI7300 PCR machine (Applied Biosystems) in combination with the qPCR Core kit for SYBR Green I (Eurogentec, Maastricht, The Netherlands) and analyzed using the 7300 System SDS software (Applied Biosystems). Unless described otherwise, the primer pair AtRub-F4 and AtRub-R4 targeting the gene encoding the large subunit of RuBisCo was used as endogenous control. Verticillium and R. solanacearum colonization was assessed as previously described [13], [23].

Expression analysis
Both reverse transcription PCR and real-time PCR were used to analyze gene expression. Unless described otherwise, the primer pair Act2-F2 and Act2-R2 targeting the Arabidopsis Actin 2 gene was used as endogenous control. A list of primers used in this study and their targets is presented in Table S1 in File S1. The real-time PCR conditions consisted of 2 min incubation at 50uC and 10 min at 95uC followed by 40 cycles of 95uC for 15 sec. and 60uC for 1 min.

Supporting Information
File S1 Figure S1. Expression analysis of genes flanking the insertion site in the A2 mutant when compared to wild-type (WS) plants in absence of pathogen inoculation. The gene encoding EWR1 is boxed (At3g13437). Reactions to amplify the Actin 2 gene and a non-template control (NTC) were included as controls. Figure S2. Relative quantification (RQ) of EWR1 transcription. (A) Relative quantification (RQ) of EWR1 transcription in wild-type (WS) and activation-tagged mutants (A1-A4). Bars represent averages with standard deviation of three biological replicates. (B) Relative quantification (RQ) of EWR1 transcription in wild-type (Col-0) plants, two independent EWR1 over-expressing plants (EWR1-1 and EWR1-2), the A2 mutant, and of the EWR1 knockout line (ewr1). Bars represent averages with standard deviation of three biological replicates. Figure S3. EWR1 over-expressing Arabidopsis plants are resistant to V. dahliae. (A) Typical symptoms of V. dahliae on the wild-type (WS) and three independent EWR1 over-expressing lines in WS background (AtEWR1-4, AtEWR1-5, and AtEWR1-6) at 21 days post inoculation (dpi). Representative of three experimental replicates is shown. (B) Disease severity score for the wild-type (WS) and three independent EWR1 overexpressing lines in WS background (AtEWR1-4, AtEWR1-5, and AtEWR1-6) at 14 (white bar) and 21 (grey bar) dpi. The total number of rosette leaves and number of rosette leaves that showed Verticillium symptoms was counted at least from eight plants and percentage of the disease leaves were calculated as an indication of disease severity. The bars represent averages of three independent experiments with standard deviation and asterisks indicate significance differences (Dunnett t-test at P = 0.05). Figure S4. AtEWR1 over-expression alters Arabidopsis leaf morphology when compared to the wild-type (Col-0). Figure S5. Transcriptional regulation of EWR1 gene during Verticillium infection. Relative quantification of EWR1 transcription levels in the wild-type WS (white bar) and Col-0 (grey bar) plants at 0 (before inoculation), 4, 8, 12, and 17 days post Verticillium inoculation. The bars represent average and standard deviation of three technical replicates. Representative of three independent experimental replicates is shown. Figure S6. Relative quantification of AtEWR1 transcription in the root and shoot of non-inoculated wild-type (WS) (white bar), the activation-tagged mutant A2 (light grey bar) and AtEWR1 over-expressing line (EWR1-4) (dark grey bar). The EWR1 transcript level in the shoot of WS is set at one and used for calibration. A representative of two independent biological replications is shown and bar indicates average of three technical replicates and standard deviation. Figure S7. Relative quantification of EWR1 transcript levels in AtEWR1 (A) and BoEWR1 (B) over-expressing N. benthamiana plants. The real-time PCR assays were normalized to the Arabidopsis actin transcript level as an internal control. The experiment was repeated at least three times with similar result and the bar indicates average of three technical replicates and standard deviation. Table S1: Primers used in this study. (DOC)