•ToxA, a host-selective toxin of wheat, can be detected within ToxA-sensitive mesophyll cells, where it localizes to chloroplasts and induces necrosis. Interaction of ToxA with the chloroplast-localized protein ToxABP1 has been implicated in this process. Therefore, we hypothesized that silencing of ToxABP1 in wheat would lead to a necrotic phenotype. Also, because ToxABP1 is highly conserved in plants, internal expression of ToxA in plants that do not normally internalize ToxA should result in cell death.
•Reduction of ToxABP1 expression was achieved using Barley stripe mosaic virus (BSMV)-mediated, viral-induced gene silencing. The BSMV system was modified for use as an internal expression vector for ToxA in monocots. Agrobacterium-mediated expression of ToxA in a dicot (tobacco-Nicotiana benthamiana) was also performed.
•Viral-induced gene silencing of ToxABP1 partially recapitulates the phenotype of ToxA treatment and wheat plants with reduced ToxABP1 also have reduced sensitivity to ToxA. When ToxA is expressed in ToxA-insensitive wheat, barley (Hordeum vulgare) and tobacco, cell death ensues.
•ToxA accumulation in any chloroplast-containing cell is likely to result in cell death. Our data indicate that the ToxA–ToxABP1 interaction alters ToxABP1 function. This interaction is a critical, although not exclusive, component of the ToxA-induced cell death cascade.
Ptr ToxA (ToxA) (syn. Ptr toxin, Ptr necrosis toxin, and ToxA (Ciuffetti et al., 1998)) is a proteinaceous toxin produced by the necrotrophic, fungal wheat (Triticum aestivum) pathogen Pyrenophora tritici-repentis (Ballance et al., 1989; Tomas et al., 1990; Tuori et al., 1995). A gene that is highly similar to the gene that encodes ToxA has recently been described as an important pathogenicity factor for another significant wheat pathogen, Stagonospora nodorum (Friesen et al., 2006). Pyrenophora tritici-repentis isolates that secrete ToxA into the apoplastic space of infected wheat leaves induce the formation of necrotic lesions on cultivars that are sensitive to the toxin (Lamari et al., 1995; Ciuffetti et al., 1997). If the cultivar is not sensitive to ToxA, the typical ‘tan spot’ lesion will not form. Isolation, purification and infiltration of ToxA results in necrosis in sensitive leaves, but no symptoms are evident in insensitive leaves. Therefore, ToxA is characterized as ‘host-selective’ because it affects only specific wheat cultivars. Sensitivity to ToxA is conferred by the Tsn1 locus on the 5BL chromosome (Faris et al., 1996; Stock et al., 1996; Gamba et al., 1998; Anderson et al., 1999). Although the function of the product of the Tsn1 locus is unknown, we have shown that ToxA can be detected inside mesophyll cells of ToxA-sensitive cultivars, but not ToxA-insensitive cultivars, indicating that internalization is required for ToxA activity (Manning & Ciuffetti, 2005). Interestingly, intracellular expression of ToxA in ToxA-sensitive and ToxA-insensitive cultivars via biolistic introduction of a ToxA expression plasmid shows that cell death occurs in both ToxA-sensitive and ToxA-insensitive cells, and suggests that the internal site-of-action is conserved. Once inside sensitive cells, ToxA localizes to the chloroplast, implicating the chloroplast as the site-of-action. This inferred site-of-action is further supported by the finding that ToxA interacts with a chloroplast-localized protein that we call ToxABinding Protein 1 (ToxABP1) (Manning et al., 2007).
The ToxA gene encodes a protein that contains a signal peptide for secretion (Ballance et al., 1996; Ciuffetti et al., 1997), a 4.3 kDa N-domain that has been shown to be important for proper folding of the protein (Tuori et al., 2000) and a 13.2 kDa C-domain that encodes the mature peptide (Tuori et al., 1995). The three-dimensional structure of the mature peptide has been resolved (Sarma et al., 2005) and a conspicuous feature of the structure is a solvent-exposed, RGD-containing loop that is required for internalization into sensitive cells (Manning et al., 2008). In addition to the ToxA structural elements required for internalization, ToxA-sensitive cells have a high-affinity binding site for ToxA, which is also likely required for toxin internalization. Once internalized, ToxA induces cell death in a light-dependent manner likely by altering photosystem homeostasis, leading to an accumulation of reactive oxygen species (ROS) (Manning et al., 2009). Recently, we have shown that ToxA vs H2O treatment of sensitive leaves results in the downregulation of a number of photosystem genes and the regulation of genes involved in oxidative stress (Pandelova et al., 2009). The contribution of these transcriptional changes on photosystem homeostasis is currently under investigation.
ToxABP1, also known as Thf1 in Arabidopsis, has been characterized in Arabidopsis as a chloroplast-localized protein implicated in thylakoid formation and/or photosystem II (PSII) biogenesis (Wang et al., 2004; Keren et al., 2005; Manning et al., 2007). Knockout mutants of the ToxABP1 homologs in Arabidopsis and Synechocystis (Psb29) have a reduced growth rate and reduced photosystem II (PSII) activity in high light conditions and Arabidopsis knockouts have variegated leaves with variegations lacking chlorophyll and organized thylakoids (Wang et al., 2004; Keren et al., 2005). The similarity of the phenotype displayed by mutants in plants and cyanobacteria suggests that ToxABP1 function is highly conserved. Owing to the light-dependent nature of the mutant phenotypes it has been suggested that the actual role for ToxABP1 homologs in these organisms is in PSII biogenesis, and because biogenesis is affected, ROS can accumulate in the chloroplasts during exposure to light, resulting in disorganized thylakoids (Keren et al., 2005). The compelling similarities between how ToxA induces cell death and how knockouts of ToxABP1 homologs behave suggest that ToxA could be interfering with ToxABP1 function, resulting in the accumulation of ROS and eventual cell death.
Barley stripe mosaic virus (BSMV) is a tripartite, positive-sense RNA virus with a broad monocot host range that has been developed for use as a vector for virus-induced gene silencing (VIGS) in grasses (Holzberg et al., 2002; Scofield et al., 2005; Scofield & Nelson, 2009). Infectious transcripts can be produced in vitro from plasmids representing the α, β and γ viral RNAs (Petty et al., 1989). Insertion of sequences downstream of the γb gene has proven effective for silencing genes that are associated with the barley (Hordeum vulgare) response to powdery mildew and stem rust (Hein et al., 2005; Meng et al., 2009; Zhang et al., 2009a), as well as leaf rust in wheat (Scofield et al., 2005). BSMV has also been used to express green fluorescent protein (GFP) by fusion of the GFP open reading frame downstream of the γb open reading frame (Lawrence & Jackson, 2001). These fusions were successfully used to study BSMV genes required for cell-to-cell spread in monocots and dicots and symplastic phloem unloading in barley (Haupt et al., 2001).
We use the BSMV vector system for expression of ToxA in wheat cultivars that are either sensitive or insensitive to external application of ToxA and confirm that internal expression of ToxA in wheat, regardless of a cultivar’s sensitivity to externally delivered ToxA, induces cell death. In addition, we show that ToxA expression in barley as well as Agrobacterium-mediated expression of ToxA in tobacco (Nicotiana benthamiana) also leads to cell death. Because we had previously shown that ToxA interacts with ToxABP1 in wheat, we use yeast two-hybrid analysis to confirm that ToxA can interact with tobacco ToxABP1 and can therefore be a putative interacting partner regardless of plant species. These data suggest that the internal site-of-action of ToxA is conserved in monocots and dicots. We also use the BSMV vector system for silencing ToxABP1 and find that silencing of ToxABP1 partially recapitulates the phenotype of internal expression of ToxA. In addition, we find that ToxA-sensitive wheat leaves in which ToxABP1 has been silenced are less responsive to ToxA treatment. These data suggest that complete ToxA activity relies on interaction with ToxABP1, and that this interaction compromises ToxABP1 function.
Materials and Methods
Plant growth conditions and toxin purification
Wheat and barley plants were grown in 16 h light at 22°C : 8 h dark at 19°C at approx. 150 μmol m−2 s−1 light intensity. Nicotiana benthamiana plants were grown in a 12-h photoperiod at 22°C at approx. 50 μmol m−2 s−1 light intensity. Toxin was isolated as previously described (Tuori et al., 1995). To document symptom development, leaves were harvested and scanned on an Epson 1600 scanner (Epson, Long Beach, CA, USA).
Constructs for BSMV-mediated viral-induced gene silencing were described in Scofield et al., 2005. For construction of BSMV-mediated expression of ToxA, the stop codon of the γb gene, present on the plasmid pSL038-1, was converted to an EcoRI site by site-directed mutagenesis (QuickChange; Stratagene, LaJolla, CA, USA) using the primers Gb:TGAtoRI-1 (5′-TCCGTTGCTAGCGAATT-CGCCGCCCCGGGTTAA-3′) and Gb:TGAtoRI-2 (5′-TTAACCCGGGGCGGCGAATTCGCTAGCAACGGA-3′). His-tagged ToxA (Tuori et al., 2000) was PCR amplified using primers Gb-HisToxA5′ (5′-TCCTTTGAATT-CATGATGCACCAT-3′) containing an EcoRI site and Gb-HisToxA3′ (5′-CGCGCCCGGGCTAATTTTGTAG-3′) containing a SmaI site downstream of the ToxA stop codon and cloned as a fusion into the γb open reading frame. Fusions were confirmed by sequencing at the Central Services Laboratory, Center for Genome Research and Biocomputing, Oregon State University, Corvallis, USA.
For construction of ToxABP1 silencing vectors, 5′, center and 3′ fragments of the coding region of ToxABP1 were PCR amplified with the following primer pairs, respectively: Vigs-1F (5′-CCATATCGTCGCTTCCTTTC-3′) and Vigs-1R (5′-CTTGAAGATGGCATCCCTGT-3′); Vigs-2F (5′-CAGGGATGCCATCTTCAAGT-3′) and Vigs-2R (5′-AGTTGCGGTAAACGTCAAGG-3′); Vigs-3F (5′-TCAGCTACAGCCGGTTCTTT-3′) and Vigs-3R (5′-AAGATGCCCCCGATAGTTCT-3′). Fragments were cloned into pGEM-T Easy (Promega, Madison, WI, USA) and sequences verified. Fragments were excised with NotI and cloned in the sense and antisense orientation into the NotI site of pSL038-1.
Viral inoculation and Agrobacterium infiltration
For agroinfiltration, plasmids containing the β-glucuronidase and the N- and C-domains of ToxA were transformed into AGL-1 cells and GV2260 cells. The AGL-1 cells were not effective at transforming the tobacco cells so the GV2260 cells were used. Single colonies were grown overnight in 5 ml low salt Luria broth (Research Products International Corp, Prospect, IL, USA) at room temperature (RT) with constant shaking. Fifty microlitre of the culture was used to inoculate 25 ml of YEP (5 g peptone, 2.5 g NaCl, 5 g yeast extract in 500 ml dH2O) and cultures grown overnight at RT with constant shaking. Cells were pelleted at 5000 g and resuspended in 25 ml induction media consisting of M9 Media, pH 5.2 (6 g Na2HPO4, 3 g KH2PO4, 0.5 g NaCl, 1.0 g NH4Cl, 1 ml 1 M MgSO4, 0.1 ml 1 M CaCl2 and 10.0 ml of 20% glucose in 1 l) plus 0.1 mM acetosyringone and 10 mM 2-(N-morpholino)-ethanesulfonic acid (MES), pH 5.2. Cultures were left on the benchtop overnight, spun down and resuspended at an OD600 of approx. 0.5 in infiltration media (10 mM MgCl2, 10 mM MES, pH 5.2 and 1.5 mM acetosyringone). Nine to ten leaves of 4- to 5-wk-old N. benthamiana plants were infiltrated using a 1 ml syringe. Infiltrations were repeated four times with similar results each time.
Viral inoculations were performed on primary leaves of wheat and barley plants as described in Scofield et al. (2005) when second leaves had emerged approx. 2 cm (6–7 d post seeding). For a single experiment, each set of transcripts was inoculated on at least four plants. Each inoculation was performed a minimum of two times.
For RNA preparation from barley and wheat, two tertiary leaves were pooled, and for tobacco an approx. 5 cm square of leaf section was cut, frozen in liquid nitrogen and stored at −80°C. RNA was prepared using a Qiagen RNeasy Plant Mini Kit, DNased with the DNA-free kit (Ambion) and cDNA synthesized with iScript (Bio-Rad). Primer pairs used for PCR were: 18S, 18SF (5′-GTGACGGGTGACGGAGA-ATT-3′) & 18SR (5′-GACACTAATGCGCCCGGTAT-3′); ToxA, V22 (5′-GCGCGGATTCGGAAGCTGCATG-3′) and p14B (5′-GCTTGGATCCTAATTTTCTAGCTGC-ATTCT-3′); γb, gamma1-F (5′-CCTGAATGGAATCGTT-GGAC-3′) and gamma1-R (5′-GCTCCTGCATCTTCTT-CTGG-3′); GUS, gusA1F (5′-CTGATAGCGCGTGACA-AAAA) and gusA1R (5′-GGCACAGCACATCAAAGAGA).
For qRT-PCR, RNA quality was confirmed by Bioanalyzer at the Central Services Laboratory, Center for Genome Research and Biocomputing, Oregon State University. The efficacy of DNase treatment was verified by PCR of the RNA template. The 18S primers described above were used as an internal control for relative quantification. ToxABP1 transcript was detected with primers ToxABP1-1F (5′-CTGGTGATATCCCACCTACTG-3′) and ToxABP1-1R (5′-GTTGCTGCACCAAAAGTTCT-3′). Reactions consisted of 2 × iQ SybrGreen (Bio-Rad), primers and 1 μl and 5 μl of cDNA used per 18S or ToxABP1 reaction, respectively. The PCR program for real-time was performed on a MyiQ iCycler: 95°C-1.5 min and 95°C-30 s, 55.5°C-30 s, 72°C-30 s (40×).
Protein extraction and western blot analysis
Protein extractions were performed on 1 cm2 or a 3-cm length of tobacco or wheat leaves, respectively, based on a protocol by Curtis & Wolpert (2002). Western blot analysis was performed on 80 μg of total tobacco protein (Pierce BCA assay kit (Rockford, IL, USA) with BSA as the standard) or the entire sample of total wheat, as previously described (Manning et al., 2008).
Cloning of tobacco ToxABP1 and yeast two-hybrid analysis
RNA from untreated N. benthamiana leaves was prepared, DNased and cDNA synthesized as above. Primers V106 (5′-GGGAATTCATGGCGGCAGTTAC-3′) and V107 (5′-ATGGATCCCTTACCTCCCAGCATA-3′) for PCR amplification of ToxABP1 from N. benthamiana cDNA were designed based on EST TA15813_4097 retrieved via a blastn search (Altschul et al., 1990) of the TIGR plant transcript assembly database (http://plantta.jcvi.org/). The restriction sites for EcoRI and BamHI (given in bold type) were included in the primers to facilitate cloning into the yeast two-hybrid vector, pGADT7. The ToxABP1 open reading frame was PCR amplified, cloned into pGEM-T Easy and the sequence verified. A fragment containing the correct sequence was excised from pGEM-T Easy with EcoRI/BamHI and cloned into pGADT7 cut with the same enzyme. This created a fusion construct (pCVM121-3) of ToxABP1 and the activation domain of GAL4 (ToxABP1:AD).
The yeast cell line Y2HGold (Clontech, Mountain View, CA, USA) was transformed with pCVM121-3, a control plasmid (pGBT9) that contains no insert, and constructs pCVM40 (GAL4 DNA-binding domain::ToxA), pCVM42 (GAL4 DNA-binding domain::ToxA(T137A)) and pCVM35 (wheat ToxABP1:AD), previously described in Manning et al. (2007), either singly or in combination, with the Pichia EasyComp kit (Invitrogen). Transformants were selected on synthetic dropout (SD) media lacking the appropriate amino acid for nutritional selection (Clontech): minus tryptophan – pGBT9, pCVM40 and pCVM42 – and minus leucine and tryptophan – pGBT9/pCVM35, pGBT9/pCVM121-3, pCVM40/pCVM35, pCVM40/pCVM121-3 and pCVM42/pCVM121-3. Three colonies from each transformation were tested for plasmid insert using ToxA V22b (5′-CGCGGAATTCCAGGGAAG-CTGCATG-3′ & p14b, wheat ToxABP1 V25c (5′-CAGTGAATTCATGGCGGCCATATCG-3′) & V26c (5′-CGCGGATCCGGCCGCTCTAGAGTCGAC-3′) and tobacco ToxABP1 specific primers (V106 & V107). One colony representing a pCVM40/pCVM121-3 cotransformant showed an unusual ToxABP1 PCR pattern and was not used further. All colonies were tested for the ability to induce transcription by the ADE1 and HIS3 promoters by plating on SD media minus histidine, adenine and the amino acids used for nutritional selection. The X-α-Gal assay was performed following manufacturer’s instructions (Clontech).
Five agroinfiltrated leaves were harvested at day 12 and leaf sections of infiltrated areas were removed using a #6 borer. Each section was added to 1 ml of 95% ethanol and chlorophyll extracted overnight at RT in the dark. Total chlorophyll was calculated as previously described (Manning et al., 2004).
The internal site-of-action of ToxA is conserved between monocots and dicots
We previously showed that infiltration of ToxA into the apoplastic space of ToxA-sensitive and ToxA-insenstive wheat leaves results in detection of ToxA within ToxA-sensitive but not ToxA-insensitive mesophyll cells. This internalization correlates with the cell death that occurs only in ToxA-sensitive cultivars. Furthermore, internal expression of ToxA by ToxA-sensitive and ToxA-insensitive cells via introduction of a ToxA expression plasmid using a biolistics approach, which bypasses the requirement for ToxA internalization, results in cell death, regardless of whether cells are sensitive or insensitive to external application of the toxin (Manning & Ciuffetti, 2005). This suggests that the internal site-of-action of ToxA is conserved in ToxA-sensitive and ToxA-insensitive wheat. To determine if this site-of-action is conserved in other monocots, we constructed a ToxA expression system for monocots based on the ability of BSMV to infect both wheat and barley. We adopted an approach similar to that of Lawrence & Jackson (2001) for the expression of green fluorescent protein by BSMV, where we converted the stop codon of the γb protein of BSMV to an EcoRI site and fused, in-frame, the N- and C-domains of the ToxA coding region (Fig. 1a). The signal sequence was excluded to ensure that ToxA remains internalized and the N-domain was included to ensure proper folding of the protein (Tuori et al., 2000). All BSMV inoculations included transcripts prepared from the same α and β templates, but different γ templates. For example, vector alone inoculations (BSMV:00) consisted of α, β and unmodified γ transcripts, phytoene desaturase (PDS)-silencing inoculations (BSMV:PDSas) consisted of α and β transcripts and a modified γb transcript with an antisense fragment of PDS cloned downstream of the γb gene (Scofield et al., 2005), and ToxA-expression inoculations (BSMVγb:ToxA) consisted of α and β transcripts and a γ transcript encoding the γb:ToxA fusion. To test whether BSMV-mediated expression of ToxA results in cell death in wheat, we inoculated the first leaves of the ToxA-sensitive wheat cultivar Glenlea with BSMV:00 as a vector alone control, BSMV:PDSas to confirm viral spread in this wheat cultivar and BSMVγb:ToxA as the test construct. Second and third leaves were harvested 12 d post inoculation (dpi). Fig. 1(b) shows representative leaves from these inoculations. Three leaves are included to show the range of symptoms. Plants inoculated with BSMV:00 showed chlorosis on the second leaves, but third leaves were relatively free of any symptoms (Fig. 1b, top panels). As previously described by Scofield et al. (2005), PDS silencing via inoculation of BSMV:PDSas resulted in photobleaching of third leaves (Fig. 1b, middle panels). The second leaves of plants inoculated with BSMVγb:ToxA were highly affected and showed cell death throughout the leaf and third leaves exhibited patchy necrosis (Fig. 1b, bottom panels). Therefore, BSMV-mediated expression of ToxA in the ToxA-sensitive cv Glenlea results in cell death. To confirm the expression of ToxA, RT-PCR was performed on RNA extracted from third leaves harvested 9 dpi and 12 dpi that had been inoculated with BSMV:00 (Fig. 1c, lanes 1, 3) or BSMVγb:ToxA (Fig. 1c, lanes 2, 4). Primers were designed to amplify the RNAs of wheat 18S, to assess the quality of the RNA extraction, the viral γb gene to confirm viral spread, and ToxA. ToxA transcript was detected at 9 dpi and 12 dpi in BSMVγb:ToxA inoculated leaves only (Fig. 1c, middle panel) whereas 18S and γb transcripts were detected in both BSMV:00 and BSMVγb:ToxA inoculated leaves (Fig. 1c, top and bottom panel). To confirm that the γb:ToxA fusion protein was produced, we performed western blot analysis with ToxA antisera on total protein extracted from third leaves harvested 9 dpi and 12 dpi inoculated with BSMV:00 (Fig. 1d, lanes 1, 3) or BSMVγb:ToxA (Fig. 1d, lanes 2, 4). As expected, the approx. 37 kDa fusion protein was detected in BSMVγb:ToxA inoculated leaves only. Together, these data suggest that production of the γb:ToxA fusion protein is responsible for the cell death that occurs due to inoculation with BSMVγb:ToxA.
Because the spread of BSMV is cultivar dependent (Hein et al., 2005) and to confirm that internal expression of ToxA can induce cell death in a ToxA-insensitive cultivar, the effect of BSMV-mediated ToxA expression was further tested on the ToxA-sensitive cv Katepwa and a ToxA-insensitive cv 6B365 (Fig. 1e). These cultivars were also inoculated with BSMV:PDSas, and, as expected, third leaves exhibited photobleaching (data not shown). As seen with the ToxA-sensitive cv Glenlea (Fig. 1b), the second leaves of these cultivars exhibited extensive necrosis caused by ToxA expression. ToxA-expression in the third leaves of these cultivars leads to a range of symptoms. Sometimes, the leaves do not emerge (data not shown) and leaves that do emerge are often stunted, do not expand or are severely necrotic and collapsed (Fig. 1e). If leaves do not collapse, there can be patchy necrosis (Fig. 1e). Therefore, BSMV-mediated ToxA-expression leads to cell death in all wheat cultivars tested, confirming biolistic transformation experiments, which showed that internal expression of ToxA leads to cell death. Therefore the internal site-of-action of ToxA is conserved in ToxA-sensitive and ToxA-insensitive wheat cultivars (Manning & Ciuffetti, 2005). These data also suggest that BSMV-mediated expression of ToxA can be used to test whether ToxA-expression in other BSMV-susceptible monocots leads to cell death.
The cv Black Hulless was used to determine the outcome of BSMV-mediated expression of ToxA in barley. Primary leaves of plants were inoculated with either BSMV:00, BSMV:PDSas or BSMVγb:ToxA and leaves harvested 12 dpi. Second leaves of BSMVγb:ToxA-inoculated plants exhibited severe necrosis throughout the leaf (Fig. 2a, bottom left) compared with second leaves of BSMV:00- or BSMV:PDSas-inoculated plants that showed only slight chlorosis in some leaves (Fig. 2a, top, middle left). The third leaves of BSMV:00-inoculated plants, although slightly discolored, remained broad (Fig. 2a top right), whereas third leaves of BSMV:PDSas-inoculated plants were less broad and exhibited photobleaching, as expected (Fig. 2a, middle right). The third leaves of BSMVγb:ToxA-inoculated plants were also less broad and showed streaks of necrosis (Fig. 2a, bottom right). RNA was extracted from second leaves of inoculated plants 9 dpi to test whether ToxA expression correlated with necrosis. Second leaves 12 dpi were already so compromised that it would have been difficult to isolate the RNA of ToxA-expressing cells, therefore, RNA was extracted from third leaves at this time-point. ToxA transcript was detected in RNA extracted from second leaves 9 dpi and in third leaves 12 dpi only in plants inoculated with BSMVγb:ToxA (Fig. 2b, lanes 2, 4). The 18S and γb transcripts were readily detected in both PDS-silenced and ToxA-expressing leaves.
Agrobacterium-mediated transformation of N. benthamiana was used to determine if internal expression of ToxA in tobacco leads to cell death. Tobacco leaves were infiltrated with Agrobacterium containing a construct expressing the gene for β-glucuronidase (GUS) or the N- and C-domains of the ToxA gene (Fig. 3a) and the status of the leaf monitored for 12 d (Fig. 3b). Leaves infiltrated with the Agrobacterium strain carrying the ToxA-expression construct began to show yellowing by day 5 (Fig. 3b, bottom left). These symptoms increased over time such that by day 12 cell death had spread throughout the infiltration zone (Fig. 3b, bottom right). Previously, we used the reduction of total chlorophyll as a measure of ToxA activity (Manning et al., 2004). Employing this method for GUS- and ToxA-expressing tobacco leaves, we found that ToxA-expressing leaves had less total chlorophyll than GUS-expressing leaves (Fig. 3c). To confirm that the visualized cell death and the reduction of chlorophyll resulted from the expression of ToxA, RT-PCR was used to confirm expression in the infiltration zone (Fig. 3d). The GUS transcript was detected at day 5 and day 7 only in leaves infiltrated with the Agrobacterium strain containing the control GUS-expression construct (Fig. 3d, top panel) and ToxA transcript was detected only in leaves infiltrated with the Agrobacterium strain containing the ToxA expression construct (Fig. 3d, middle panel). The 18S transcript was readily detected in both treatments (Fig. 3d, bottom panel). Expression of ToxA was not tested at day 12 due to extensive leaf damage. Western blot analysis confirmed that ToxA was produced in the leaves infiltrated with the Agrobacterium strain containing the ToxA-expression plasmid and not in the control infiltrated plants (Fig. 3e). These data suggest that internal expression of ToxA in tobacco leads to cell death and that the site-of-action of ToxA is conserved in monocots and dicots. We have reported that ToxA interacts with the chloroplast protein ToxABP1 (Thf1) and that interaction with this protein is likely to contribute to cell death (Manning et al., 2007). Comparison of the protein sequence of ToxABP1 from wheat with that from barley and tobacco shows that the barley protein is 98% identical and 99% similar to that of wheat when the comparison includes or excludes the sequence that corresponds to the chloroplast transit peptide (cTP), respectively. Tobacco ToxABP1 is 67% identical and 85% similar with the cTP and 78% identical and 94% similar without the cTP. This high degree of similarity suggests that ToxA could interact with both barley and tobacco ToxABP1.
Yeast two-hybrid analysis was used to test whether ToxA can interact with tobacco ToxABP1 (NbToxABP1). We had previously shown that a GAL4 DNA-binding domain ToxA fusion (BD:ToxA) could interact with a wheat ToxABP1:GAL4 activation domain fusion (TaToxABP1:AD) and induce transcription in this system, whereas a mutant form of ToxA, BD:ToxA(T137A) could not (Manning et al., 2007). To test the NbToxABP1–ToxA interaction, the ToxABP1 open reading frame was PCR-amplified from reverse transcribed RNA isolated from N. benthamiana leaf tissue, and cloned as a fusion with the GAL4 activation domain (NbToxABP1:AD). We used a newly developed yeast cell line, Y2HGold, to test the interaction because this cell line has four GAL4-dependent reporters expressed from three unique GAL4-dependent promoters (M1, G1 and G2) for testing protein–protein interactions: ADE1, nutritional selection on Adenine minus media (G2 promoter); HIS3, nutritional selection on Histidine minus media (G1 promoter); MEL1, colorimetric conversion of clear X-α-Gal substrate to blue; and AUR1-C, growth in media supplemented with the antibiotic Aureobasidin A (AbA) (both driven by the M1 promoter). Y2HGold cells were transformed with: plasmids pGBT9, BD:ToxA and BD:ToxA(T137A) to confirm that the product of these plasmids, while binding DNA, do not induce transcription; pGBT9/TaToxABP1:AD and pGBT9/NbToxABP1:AD to confirm that neither wheat nor tobacco ToxABP1 induce transcription in the absence of an interacting partner that contains a DNA-BD; BD:ToxA/TaToxABP1:AD as a positive interaction control that should induce transcription; BD:ToxA/NbToxABP1:AD, the test interaction; and BD:ToxA(T137A)/NbToxABP1:AD to confirm the specificity of the test interaction. Transformants were selected on the appropriate media (Table 1) and the presence of the correct expression plasmids confirmed via PCR (data not shown). At least three transformants each were tested for the ability to induce expression from GAL4-dependent reporter constructs (Table 1). Only cotransformants that contained BD:ToxA/TaToxABP1:AD & BD:ToxA/NbToxABP1:AD were able to grow on media deficient in histidine and adenine; therefore, these combinations were capable of activating the G1 and G2 promoters. These cotransformants were also able to cleave X-α-Gal, resulting in dark blue colonies 1 d after plating, suggesting a strong activation of the M1 promoter. These data confirm our previous results that ToxA interacts with wheat ToxABP1 (Manning et al., 2007) and suggest that ToxA can interact with tobacco ToxABP1. BD:ToxA(T137A):NbToxABP1:AD does not follow this pattern; therefore, it appears that like wheat ToxABP1, tobacco ToxABP1 cannot interact with the ToxA(T137A) mutant. Transformants that contain pGBT9, BD:ToxA and BD:ToxA(T137A) had a light blue coloration on media containing X-α-Gal several days after plating (Table 1), as did untransformed Y2HGold cells (data not shown). This was surprising as untransformed cells and control transformants that contain the plasmid that encodes the DNA binding domain alone (pGBT9) should not convert the X-α-Gal substrate. This indicates that there is leaky expression from the M1 promoter in Y2HGold cells. Consistent with this finding was an unpredictable growth pattern of transformants in AbA (data not shown). Therefore, the G1 and G2 promoters present in the Y2HGold yeast line were more reliable in the detection of protein–protein interactions than was the M1 promoter.
Table 1. Yeast two-hybrid confirmation of the interaction between tobacco ToxABP1 and ToxA
Construct for transformationa,b
Nutritional requirements for transformant selectionc
Promoter for confirming protein-protein interaction
G1 and G2d
Nutritional selection on −Ade/−His mediac
Conversion of substrate (X-α-Gal) on transformant selection mediac,e
aDNA-BD refers to the GAL 4 DNA-Binding domain, which binds to the M1, G1 and G2 promoters of the yeast two-hybrid reporter construct genes. AD refers to the GAL4 activation domain required for inducing transcription from the above promoters.
cAll nutritional media was made from synthetic dropout (SD) minimal media minus (−) the respective amino acid(s).
dThe G1 promoter drives expression of the ADE2 gene, G2 drives HIS3 and M1 drives MEL1.
e(−/+) indicates that a slight blue coloration appears after several days of incubation with substrate, (−) no blue appears and (+) blue appears after overnight incubation with substrate.
Control plasmid (pGBT9)
TaToxABP1:AD + pGBT9
NbToxABP1:AD + pGBT9
DNA-BD:ToxA + TaToxABP1:AD
DNA-BD:ToxA + NbToxABP1:AD
DNA-BD:ToxA (T137A) + NbToxABP1:AD
ToxABP1 silencing in wheat partially recapitulates the phenotype of internal expression of ToxA
If ToxA interacts with ToxABP1 and alters ToxABP1 function, we would expect that silencing ToxABP1 would lead to a similar phenotype as internal expression of ToxA. To test this, we designed constructs in BSMV that would lead to silencing of ToxABP1. In this case, we inserted 3′, middle, and 5′ regions of the ToxABP1 coding sequence in both the sense and antisense orientation downstream of the stop codon of the γb gene (Fig. 4a). Each of these constructs was inoculated onto the first leaves of the ToxA-sensitive cv Glenlea and symptom development monitored for 12 d. On second leaves, the most common phenotype with all constructs was discontinuous necrosis (Fig. 4b) seen on 60–86% of the leaves and this was dependent on the construct (Table 2). BSMV:00 and BSMV:PDSas inoculations led to occasional chlorosis on second leaves (Figs 1b, 4b). Third leaves of ToxABP1-silenced plants showed a variety of phenotypes including: necrosis/chlorosis throughout leaf, interspersed with green areas; necrotic streaks; tip senescence with necrotic flecks below the tip, not extending throughout the leaf; and flecks of necrosis/chlorosis (Fig. 4c, bottom panel). The construct that consistently showed the most severe phenotype (phenotype 1, Fig. 4c) was BSMV:ToxABP1(AS2), where the γ transcript contained an antisense fragment from the middle of the coding region of ToxABP1 (Table 2, Fig. 4a). The extent of silencing in third leaves caused by inoculation of BSMV:ToxABP1(AS2) compared with inoculation of BSMV:00 (vector alone) was determined by comparative qRT-PCR using 18S as the internal control, and fold change calculated by the ΔΔCT method (Livak & Schmittgen, 2001) (Table 3). Three biological replicates were performed for each time-point. Reduction of ToxABP1 transcript at day 9 ranged from 37% to 78% and at day 12 the reduction was approx. 90%. The phenotypes of ToxA expression (Fig. 1) and ToxABP1 silencing (Fig. 4) were similar, although ToxA expression induced more severe necrosis. We also inoculated BSMV:ToxABP1(AS2) onto the first leaves of the wheat cv 6B365 to determine the effect of ToxABP1 silencing on a ToxA-insensitive cultivar (Fig. 4d). Second leaves of silenced plants showed slight necrosis and some third leaves did not emerge, and a subset of those that did emerge were severely stunted or collapsed. These effects were similar, though less severe than the leaves of cv 6B365 that express ToxA (Compare Figs 1e and 4d). Therefore, ToxABP1 silencing in the ToxA-insensitive cultivar is similar to ToxA expression, but less severe, similar to what was seen with the ToxA-sensitive cv Glenlea.
Table 2. ToxABP1-silenced wheat leaves display a range of phenotypes
% Leaves displaying phenotype
% of 3rd leaves showing each phenotype
aRefers to the fragment of the ToxABP1 open reading frame (Genbank # AY37791) present in the silencing vector. S = sense and AS = antisense orientation. 1 = bp 8–393, 2 = bp 375–715 and 3 = bp 569–953 of ORF.
bLeaves become necrotic.
cLeaves display any of phenotype 1–4(d–g).
dNecrotic/chlorotic areas throughout leaf interspersed with green areas.
eStreaks of necrosis/chlorosis.
fTip senescence with necrotic areas below the tip, not extending throughout leaf.
The wheat cultivar used was the ToxA-sensitive cv Glenlea. Leaves were inoculated 6–7 d after planting and harvested 12 d post inoculation. Three biological replicates were performed for each construct. Each replicate represents at least four inoculated plants.
Table 3. Fold change of ToxABP1 transcript in the third leaf of ToxABP1-silenced wheat as determined by quantitative reverse-transcription polymerase chain reaction analysis
The wheat cultivar used was the ToxA-sensitive cv Glenlea.
Infiltration of ToxA into the apoplastic space of ToxA-sensitive leaves produces necrosis. There is no visible symptom in water-treated ToxA-sensitive leaves or in ToxA-treated, ToxA-insensitive leaves (Ballance et al., 1989; Tomas et al., 1990; Tuori et al., 1995). If ToxA interaction with ToxABP1 is required for ToxA activity then we would expect that ToxABP1-silenced, ToxA-sensitive wheat leaves would show a decreased response to infiltrated toxin. To test this, the ToxA-sensitive cv Glenlea was inoculated with BSMV:00 and BSMV:ToxABP1(AS2) and third leaves infiltrated with ToxA 10 dpi. Leaves were harvested 2 d after toxin treatment. BSMV:00 inoculated leaves showed severe necrosis indicating that viral inoculation does not affect ToxA-induced necrosis in ToxA-sensitive leaves. However, ToxABP1-silenced leaves infiltrated with ToxA showed less ToxA-induced necrosis than control leaves (Fig. 5).
Intracellular ToxA induces cell death in monocots and dicots
Our data indicate that the intracellular site-of-action of the host-selective toxin, ToxA, is likely conserved in both monocots and dicots. To demonstrate this, we developed a means by which we could reproducibly express ToxA internally and easily visualize the results of that expression. For monocot expression, we chose to use BSMV as it has a broad monocot host range and can be modified to express proteins as fusions to the viral γb protein. BSMV-induced internal expression of ToxA leads to severe necrosis in second leaves by day 12 in wheat and barley, regardless of the sensitivity of the plant to external application of ToxA (Fig. 1). Studies in barley with the BSMV.gfp fusion virus on which the ToxA-expression vector is based have shown that BSMV travels through the vascular tissue and is unloaded in discrete areas along veins, producing discrete, discontinuous flecks or streaks of gfp-expressing virus (Haupt et al., 2001). In the wheat cv Glenlea and in barley, third leaves show discontinuous flecks of necrosis and streaks of necrosis, respectively, consistent with a pattern of viral spread. There are several possible explanations as to why the symptoms in the third leaves are less severe compared with the second leaves. First, it is likely that the recombinant region of the virus is undergoing mutagenesis, which would alter the function of the toxin. It was observed that the pattern of fluorescence in leaves systemically infected with the BSMV.gfp fusion was erratic, suggesting that mutations in the gfp gene produces some viral particles that vary in fluorescence as the progeny spread throughout the leaves (Lawrence & Jackson, 2001). Also, it has been shown that in plants inoculated with BSMV:PDS recombinant viruses, the PDS insert mutates over time and is sometimes completely deleted (Bruun-Rasmussen et al., 2007). Second, the ability to express the ToxA fusion may be transient, as has been shown for BSMV-mediated silencing (Holzberg et al., 2002; Hein et al., 2005; Bruun-Rasmussen et al., 2007). Third, the very nature of expressing ToxA in cells and inducing cell death would reduce the number of viable cells capable of producing viral progeny. In the other two wheat cultivars tested – Katepwa (ToxA-sensitive) and 6B365 (ToxA-insensitive) – third leaves often did not emerge and when they did, were often collapsed and necrotic. In these cases, leaf tips appear to be more affected than the rest of the blade. This is consistent with the inability of the leaf to grow because of alteration of chloroplast function, a known result of ToxA internalization in ToxA-sensitive wheat (Manning et al., 2004, 2009). This is also consistent with ToxA interaction with ToxABP1 altering ToxABP1 function resulting in a reduced growth phenotype Keren et al. (2005). The difference in the phenotypes of the different cultivars could be explained by the difference in the host plants’ ability to tolerate virus accumulation. It has been shown that different barley cultivars show a range in their ability to develop a BSMV-mediated VIGS response (Hein et al., 2005).
Tai & Bragg (2007) also used a modified BSMV vector for expression of ToxA in wheat and barley. Their approach was to insert the sequence corresponding to the mature ToxA peptide onto the N-terminus of the γb sequence, deleting the majority of the γb protein, but preserving the C-terminus. Cell death was apparent in wheat and barley when this construct was used; however, symptom development required 21 d and was not as extensive as seen with our construct. Possible reasons for this difference in results include: the activity of ToxA produced in the absence of the N-domain is approx. 10–20% of the activity of ToxA produced in the presence of the N-domain, most likely because of the requirement of the N-domain for proper folding of the mature protein (Tuori et al., 2000); ToxA C-terminal fusions are far less active than N-terminal fusions (VA Manning & LM Ciuffetti, unpublished); γb deletions can have an effect on systemic movement of the virus (Petty et al., 1990) most likely because of the function of γb as a suppressor of silencing (Yelina et al., 2002); and BSMV RNA accumulation and viral pathogenesis is diminished when γb is deleted (C Cakir & SR Scofield, unpublished). In addition, in the Tai & Bragg study mentioned, ToxA transcript could not be detected in the necrotic tissue and therefore necrosis could not be correlated with ToxA expression. We therefore felt that it was important to repeat these studies with a construct that expressed ToxA in the presence of the N-domain to facilitate greater activity of the toxin and to correlate necrosis with ToxA expression. As a result, we were able to see necrosis in leaves inoculated with the ToxA-expression construct within 7 dpi (data not shown). In addition, we harvested leaves before complete necrosis to test for ToxA transcript and protein and found that ToxA accumulation correlated with necrosis (Fig. 1). Agrobacterium-mediated expression of ToxA in tobacco with coding regions that did not contain the N-domain has also been previously attempted (Tai et al., 2007), but no cell death occurred. When the N-domain is included in the construct, cell death is induced and can be correlated with the accumulation of ToxA (Fig. 3). Tobacco cell death requires more time than is necessary for monocot hosts and this could be because of weaker protein–protein interactions at the site-of-action. Yeast two-hybrid analysis of the interaction of ToxA with the chloroplast protein ToxABP1 from both wheat (Manning et al., 2007) and tobacco suggest that ToxA interacts with ToxABP1 in both monocots and dicots (Table 1); however, whether this interaction is significant in ToxA-induced cell death in tobacco is not known. Studies in our laboratory are underway to determine the proteins required for ToxA-induced necrosis in tobacco. These studies should help us to determine the significance of the interaction of ToxA with ToxABP1 in dicots.
ToxABP1 silencing in wheat supports the significance of the ToxA-ToxABP1 interaction in ToxA-induced cell death
Knockouts of the ToxABP1 homolog in Arabidopsis (Thf1) show variegations where chloroplasts lack thylakoid structure (Wang et al., 2004; Keren et al., 2005). However, other areas of the leaf are green and show normal chloroplast structure, suggesting that plants can overcome the deficiency. In wheat, the ToxABP1 silenced phenotype is somewhat different with extensive chlorosis and necrosis (Fig. 4). These differences in phenotype could be the result of the method used for reducing gene expression. In BSMV-mediated silencing of ToxABP1 there is likely to be an enhanced affect of viral replication and silencing occurring in the same cells. These additional stresses on the cells could lead to cell death. Thf1 knockouts in Arabidopsis also show a reduced growth rate as did ToxABP1 silencing in the ToxA-insensitive wheat cultivar, 6B365 (Fig. 4d). However, it is important to note that the reduction of the expression of ToxABP1 in ToxA-sensitive wheat also reduces the sensitivity of these wheat leaves to external application of ToxA (Fig. 5). This strongly suggests that ToxABP1 interaction with ToxA is necessary for ToxA activity. Furthermore, the similarity in the phenotypes seen with BSMV-mediated expression of ToxA and BSMV-mediated silencing of ToxABP1 suggests that inactivation of ToxABP1 function from the system is one of the features of ToxA-induced necrosis. Interestingly, although there is a high degree of similarity in phenotypes, internal expression of ToxA has a stronger necrotic phenotype than ToxABP1 silencing in both ToxA-sensitive and ToxA-insensitive cultivars (Fig. 4). The partial recapitulation of the ToxA-expression phenotype by ToxABP1 silencing and the partial reduction of ToxA activity in ToxABP1 silenced wheat leaves suggest that ToxA interacts with additional plant proteins for a complete response. The discovery of these proteins is an area of active investigation.
Exactly how the ToxA–ToxABP1 interaction results in cell death is unknown and is currently under investigation in our laboratory. However, recent studies by several groups suggest how the Arabidopsis homolog of ToxABP1, Thf1, might function and therefore, why disruption of function might lead to the altered homeostasis of photosystems and the accumulation of reactive oxygen species observed following ToxA treatment of wheat (Manning et al., 2009). Thf1 has been shown to interact at the plasma membrane with the Gα-subunit of the G-protein heterotrimer complex where it is thought to play a role in G-protein-coupled d-glucose sensing (Huang et al., 2006). Zhang et al. (2009b) expanded this observation to show that Thf1 was involved in G-protein regulation of chloroplast development. In their study, it was shown that thf1-null mutants in Arabidopsis have reduced FtsH protease activity, which is required for regulation of photosystem II D1 protein turnover. FtsH protease is required to remove photodamaged D1 protein and the authors speculate that accumulation of photodamaged D1 would lead to oxidative stress and inhibition of chloroplast development and/or photosynthesis (Zhang et al., 2009b). An investigation into the status of FtsH protease activity and D1 protein turnover following ToxA treatment of ToxA-sensitive wheat would be valuable in determining whether ToxA is effectively preventing ToxABP1 from functioning in the chloroplast, and if this is the means by which photosystem homeostasis is being affected by ToxA treatment. As more information about Thf1 (ToxABP1) function emerges and as we gain a more comprehensive understanding of ToxA-induced changes in wheat, a more complete picture of how the ToxA–ToxABP1 interaction contributes to ToxA-induced cell death should evolve. However, results presented here indicate that the interaction of ToxA with ToxABP1 plays a significant role in the mode of action of this important pathogenicity factor.
The authors thank Dr Melania Figueroa Betts and Dr Jennifer Lorang for useful conversations and Dr Tom Wolpert for manuscript review. This project is supported by a grant to L. Ciuffetti from the National Research Initiative (NRI) Microbial Biology: Microbial Associations with Plants Program of the USDA Cooperative State Research, Education and Extension Service (CSREES, grant number 2006-55600-16619).