Extracellular invertase is involved in the regulation of clubroot disease in Arabidopsis thaliana

Extracellular invertase is involved in the regulation of clubroot disease in Arabidopsis thalianampp_667 247..262 JOHANNES SIEMENS1,¶, MARIA-CRUZ GONZÁLEZ2,†¶, SEBASTIAN WOLF3,‡, CHRISTINA HOFMANN3, STEFFEN GREINER3, YEJIE DU3, THOMAS RAUSCH3, THOMAS ROITSCH2,§ AND JUTTA LUDWIG-MÜLLER1,* 1Department of Biology, Technische Universität Dresden, Zellescher Weg 20b, D-01062 Dresden, Germany 2Julius-von-Sachs-Institut, Julius-Maximilian-Universität Würzburg, Julius-von-Sachs-Platz 2, D-97082 Würzburg, Germany 3Heidelberg Institute for Plant Science, Ruprecht-Karls-Universität Heidelberg, Im Neuenheimer Feld 360, D-69120 Heidelberg, Germany SUMMARY Clubroot disease of Brassicaceae is caused by an obligate biotrophic protist, Plasmodiophora brassicae. During root gall development, a strong sink for assimilates is developed. Among other genes involved in sucrose and starch synthesis and degradation, the increased expression of invertases has been observed in a microarray experiment, and invertase and invertase inhibitor expression was confirmed using promoter::GUS lines of Arabidopsis thaliana. A functional approach demonstrates that invertases are important for gall development. Different transgenic lines expressing an invertase inhibitor under the control of two root-specific promoters, Pyk10 and CrypticT80, which results in the reduction of invertase activity, showed clearly reduced clubroot symptoms in root tissue with highest promoter expression, whereas hypocotyl galls developed normally. These results present the first evidence that invertases are important factors during gall development, most probably in supplying sugars to the pathogen. In addition, root-specific repression of invertase activity could be used as a tool to reduce clubroot symptoms. INTRODUCTION Clubroot disease within the family Brassicaceae is caused by the obligate biotrophic protist Plasmodiophora brassicae. During disease development, the root is transformed into a large gall, thus also influencing the aerial part of the plant (Siemens et al., 2002). Gall development is controlled, among other factors, by two plant hormones, i.e. auxin and cytokinin (Ludwig-Müller and Schuller, 2008). The pathogen is completely dependent on its host for nutritional purposes. Therefore, the root gall is transformed into a strong metabolic sink accumulating sugars and starch (Keen and Williams, 1969). Reallocation of assimilates from leaves to infected hypocotyls has been described in this phase of infection (Keen and Williams, 1969). Amyloplasts accumulate in infected host roots, but remain restricted to host cells (Mithen and Magrath, 1992), such that starch hydrolysis has to take place in the host cytoplasm and plastids. A change in partitioning of carbon, with a higher root to shoot ratio, is observed in infected plants (Evans and Scholes, 1995). In addition, the amount of starch in the leaves of infected plants is lower than in control plants, and both soluble and storage carbohydrates accumulate in infected tissues (Evans and Scholes, 1995). Brodmann et al. (2002) found that glucose and fructose, but not sucrose, increased in infected hypocotyls and roots, suggesting the induction of invertase-mediated active sink metabolism, because starch also increased simultaneously in clubroots at that time point. This is in accordance with observations that young vegetative secondary plasmodia of the pathogen are found near the vascular bundle (Kobelt et al., 2000). In a study using the Affymetrix ATH1 microarray, many genes associated with sugar transport and metabolism were differentially regulated (Siemens et al., 2006).Although the former indicated a strong change in trafficking of metabolites between different cells and compartments, the large number of genes involved in starch synthesis, which were upregulated, demonstrated the importance of assimilate accumulation in infected roots. In higher plants, the growth and metabolism of sink tissues are mainly sustained by sucrose synthesized in source leaves and transported through the phloem into sink tissues. The use of sucrose in sink tissues requires cleavage of the glycosidic bond, catalysed by both sucrose synthase and invertases. Sucrose synthase cleaves sucrose into uridine diphospho- (UDP-)glucose and *Correspondence: Email: [email protected] †Present address: Instituto de Bioquímica Vegetal y Fotosíntesis, Universidad de Sevilla and CSIC, Avda Américo Vespucio 49, E-41092 Sevilla, Spain. ‡Present address: Institut Jean-Pierre Bourgin, INRA, 78026 Versailles, France. §Present address: Institute for Plant Sciences, Karl-Franzens-Universität Graz, Schubertstr. 51, A-8010 Graz, Austria. ¶These two authors contributed equally to this work. MOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 DOI: 10.1111/J.1364-3703.2010.00667.X © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTD 247 UDP-fructose, whereas invertases hydrolyse sucrose into hexose monomers. Three types of invertase isoenzyme are distinguished on the basis of solubility, subcellular localization, pH optima and isoelectric point (Roitsch and González, 2004). Cell wall-bound invertases have been shown to play a crucial function in carbohydrate partitioning and supply of photoassimilates to sink tissues (Goetz et al., 2001; Hirsche et al., 2008;Tang et al., 1999; Weschke et al., 2003). Both vacuolar and cell wall invertases have been shown to play a key role in the active growth of young roots (Eschrich, 1980; Godt and Roitsch, 2006; Tang et al., 1999; Tymowska-Lalanne and Kreis, 1998), with a predominant role of vacuolar invertase in root elongation,especially in the lower part of the roots, where cell division and elongation occur (Seergeva et al., 2006). Vacuolar invertase could be a key regulator of cell expansion because of the doubled osmotic potential generated by sucrose cleavage in the vacuole. In addition, it could play a role in the stimulation of phloem unloading to parenchyma cells (Seergeva et al., 2006; Sturm and Tang, 1999). In addition, cell wall invertases could play a key function in providing an enhanced flow of utilizable carbohydrates to pathogens. Hexoses generated by the increased invertase activity, induced by pathogen attack, act, in turn, as signalling molecules, repressing photosynthetic genes (reviewed in Berger et al., 2007).The repression of cell wall invertase impaired and delayed defence reactions to Phytophthora nicotianae in leaves of tobacco (Essmann et al., 2008) and to Xanthomonas campestris in leaves of tomato (Kocal et al., 2008). In addition, in symbiotic interactions, such as in arbuscular mycorrhizal roots, increased expression and activity of cell wall invertases at stages with high carbohydrate demand support a key function of this enzyme in increasing sink strength by phloem unloading (Schaarschmidt et al., 2006). The observed reduction in mycorrhiza formation in tobacco roots with diminished apoplastic root invertase activity demonstrates the role of invertases in supplying hexoses to the mycorrhizal root (Schaarschmidt et al., 2007a, b). In tomato, the reduced mycorrhizal colonization in plants affected in the synthesis of jasmonic acid is related to a reduction in the expression and activity of cell wall invertase, suggesting that one of the mechanisms by which jasmonic acid modulates mycorrhiza formation may be through the influence of invertase activity in carbon partitioning in the plant (Tejeda-Sartorius et al., 2008). Invertases are part of a regulation network with cross-links to nearly all classes of phytohormones, which is important because of the increased demand of carbohydrates during hormone-promoted growth (Proels and Roitsch, 2009; Roitsch and González, 2004). Currently, different transgenic approaches based on altered sucrose metabolism by ectopic expression or antisense repression of an invertase have addressed the control of source–sink relationships and the effect of carbohydrate metabolism on disease development. Antisense expression of a vacuolar or cell wall invertase in tap roots of carrot resulted in an increased leaf to root ratio, with reduced tap root development in cell wall invertase antisense lines (Tang et al., 1999). The hexoses generated by the constitutive expression of a yeast invertase in the cell wall of tobacco and Arabidopsis resulted in enhanced resistance to tobacco mosaic virus (Herbers et al., 1996). The reduction of invertase activity by the expression of an antisense construct or the generation of single knock-out plants presents the limitation of repression of a specific invertase isoform, leading to a lack of clear effects, as has been described for resistance against Pseudomonas syringae or Alternaria brassicicola in Arabidopsis thaliana single knock-out plants for cell wall invertases (Berger et al., 2007). Proteinaceous invertase inhibitors constitute, in addition to transcriptional regulation, one of the main mechanisms regulating plant invertases (Rausch and Greiner, 2004). The use of an invertase inhibitor has proven to be a useful tool to circumvent the limitations of antisense inhibition for the prevention of cold-induced sweetening of potato tubers (Greiner et al., 1999) and to elucidate the role of invertases in cytokininmediated delay of senescence (Balibrea Lara et al., 2004). In this article, the role of invertases in the development of clubroot disease was analysed. In an attempt to modify photoassimilate allocation to the root, root invertase activity was reduced by tissue-specific expression of an invertase inhibitor. Reduced invertase activity in the root led to reduced clubroot symptoms, mainly in the root tissue, as a result of the rootspecific promoters used in this study, whereas hypocotyl galls developed normally. RESULTS Expression of invertases during clubroot disease The root galls caused by the obligate biotrophic pathogen P. brassicae are dependent on nutrients from the host plant (reviewed in Ludwig-Müller et al., 2009). Therefore, data obtained in previously conducted microarray experiments on galls of the compatible clubroot interaction in Arabidopsis thaliana ecotype Col-0 (Siemens et al., 2006) have been re-examined for differentially expressed genes associated with sugar metabolism, and presented in a MAPMAN graph (Figs S1 and S2, see Supporting Information). The two time points of the analysis reflect the establishment of the gall without visible symptoms and the pathogen mostly at the stage of young vegetative plasmodia (TP1), and the fully grown gall with mature plasmodia and resting spores of the pathogen (TP2), where nutrients from the host are needed for the growth of the pathogen. It was shown that, in general, sugar metabolism genes, including invertases, were upregulated during gall formation. As invertases are key enzymes for phloem unloading, there is reason to believe that they might be important for the nutrition 248 J. SIEMENS et al. © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTDMOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 of the obligate biotrophic pathogen; therefore promoter::GUS lines (GUS, b-glucuronidase) for four cell wall and two vacuolar invertases, as well as for three invertase inhibitors, were generated and analysed for expression during clubroot development (Fig. 1). In contrast with expression analyses, promoter::GUS lines would also provide information on the spatial expression of invertase genes in the galls. For the earlier time point, we chose the first visible GUS staining in the majority of young galls at 13 days after inoculation (dai). Quantitative real-time reverse transcription-polymerase chain reaction (RT-PCR) indicated expression in control roots of cell wall invertases AtcwINV5 and AtcwINV1 and vacuolar invertases Atßfruct4 and Atßfruct3, but not AtcwINV4 and AtcwINV2 (data not shown). GUS expression in galls could be shown for the cell wall invertase line pAtcwINV1::GUS (Fig. 1E,F), vacuolar invertases pAtßfruct3::GUS (Fig. 1A,B) and pAtßfruct4::GUS (Fig. 1C,D), and the invertase inhibitor pAtC/VIF1::GUS (Fig. 1G,H). In the other studied lines, pAtcwINV4::GUS,pAtcwINV5::GUS, pAtcwINV2::GUS,pAtC/VIF2::GUS and pAtC/VIF3::GUS, GUS expression in roots could not be visualized by staining under the growth conditions optimal for gall development, and these lines are therefore not shown [for more information on the promoter::GUS lines, see Table S1 (Supporting Information)]. The gene for extracellular invertase, AtcwINV1, was slightly upregulated in young galls (Fig. 1E,F), when the symptoms started to become visible, and strongly upregulated in developed galls. The same pattern was observed for the expression of the gene for vacuolar invertase, Atßfruct4 (Fig. 1C,D), although the intensity of GUS staining was lower for this invertase. In contrast, the line pAtßfruct3::GUS showed very weak expression level at 13 dai (Fig. 1A,B). Upregulation in galls could only be shown by GUS staining at TP2. The invertase inhibitor reporter line pAtC/VIF1::GUS revealed a weak expression level at 13 dai, but, at 23 dai, the expression was strong in the upper part of the root and hypocotyl of control plants and highly upregulated in infected roots (Fig. 1G,H). Functional approaches to reduce invertase activity in roots The fact that several invertases were upregulated according to promoter::GUS studies (Fig. 1) made the use of knockout lines for single genes not feasible. Therefore, in an attempt to modify the sink strength of the roots and to study the effect on the whole plant and disease development, we expressed different invertase inhibitors under the control of root-specific Fig. 1 Expression pattern analysis of selected invertase and invertase inhibitor genes in Plasmodiophora brassicae-infected (i) and control (c) roots using promoter::GUS fusions and X-Gluc staining of whole roots (GUS, b-glucuronidase; X-Gluc, 5-bromo-4-chloro-3-indoyl-b-D-glucuronide). The bar indicates 1 cm. The first symptoms became macroscopically visible at 12–14 days after inoculation (dai); therefore, in the left-hand panels, 13-dai plants are shown, which equals the first time point of the microarray experiment (for more details, see text). In the right-hand panels, 23-dai plants are shown. The expression pattern is shown for promoter::GUS lines of vacuolar invertases pAtßfruct3::GUS (A, B) and pAtßfruct4::GUS (C, D), the cell wall invertase pAtcwINV1::GUS (E, F) and the invertase inhibitor pAtC/VIF1::GUS (G, H). Invertase and clubroot disease 249 © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTD MOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 promoters. Two different root-specific promoters were used, the promoter C of the root-specific b-glucosidase PYK10 gene (Nitz et al., 2001) and a cryptic promoter described by Mollier et al. (2000). Both promoters have been shown to confer rootspecific expression by the use of GUS reporter plants. The fragment C of the Pyk10 promoter is active at different intensities throughout the whole plant in young developmental stages but, from 5 days after germination (dag) onwards, expression is restricted to the whole root, except for the elongation zone, with no detectable expression in the upper part of the plant at day 14 (Nitz et al., 2001). The 2.1-kb fragment of the cryptic promoter directed expression of the GUS reporter gene mainly to the root, with the exception of the elongation zone of the root tip, with maximal expression at 7 and 15 dag, more localized to the end of the roots and secondary roots from 26 dag onwards (Mollier et al., 2000). Two different Arabidopsis invertase inhibitors, with high homology to vacuolar (AtC/VIF1) and cell wall (AtC/VIF2) invertase inhibitors from tobacco, were used for the reduction of invertase activity in the root. Recent in vitro evidence has shown that AtC/VIF1 specifically inhibits vacuolar invertase, whereas AtC/VIF2 inhibits both acid invertase activities, although with a 10-fold higher activity for vacuolar than for cell wall invertase (Link et al., 2004). From the different independent transformants obtained for pPyk10::AtC/VIF2,pCrypticT80::AtC/VIF2 and pCrypticT80::AtC/ VIF1, several lines were chosen for further analysis of the effect of invertase inhibitor activity on plant phenotype. These lines were shown to have a statistically significant increased expression of the corresponding transgene compared with the wildtype by Northern blot analysis (Fig. 2). Lines 1 and 2 for pPyk10::AtC/VIF2 and pCrypticT80:AtC/VIF1 were homozygous lines obtained from the same heterozygous parental line; therefore, only one of the two lines is shown. The transcript analysis is presented for all three lines in the case of pCrypticT80::AtC/VIF2. The growth conditions affected the phenotype of the plants, and only the combination of long-day conditions and cultivation in Hoagland solution produced the phenotype referred to in Fig. 3 (data not shown). Expression of the transgene resulted in enhanced growth of transgenic plants during early development, although no apparent difference was observed in the time of flowering compared with wild-type control plants (Fig. 3). Transgenic plants did not show a consistent alteration in root fresh weight, but rather a tendency to an increased shoot weight because of greater branching (data not shown), and therefore an altered shoot to root ratio. In addition, the seed weight per plant showed a marked increase in comparison with wild-type plants. In soil-grown plants, greater branching was not observed and therefore the differences in shoot phenotypes were not so evident (data not shown); noninfected root phenotypes were similar to the wild-type (Table 1), which is important for the evaluation of infection experiments. Fig. 2 Analysis of transgene expression. RNA was isolated from roots of individual plants grown for 40 days. (A) Hybridization of RNA isolated from wild-type Col-0 pPyk10::AtC/VIF2 and pCrypticT80::AtC/VIF2 transgenic plants with a probe for AtC/VIF2. (B) Hybridization with an AtC/VIF1 probe of RNA isolated from wild-type Col-0 and pCrypticT80::AtC/VIF1 transgenic plants. Lines 1 and 2 for pPyk10::AtC/VIF2 and pCrypticT80:AtC/VIF1 were homozygous lines obtained from the same heterozygous parental line; therefore, only one of the two lines is shown here as an example. Data were similar for the second line. Transcript analysis is presented for three lines in the case of pCrypticT80::AtC/VIF2. At least three independent RNA samples per line were analysed and the mean values ⫾;standard error are presented as the relative transcript abundance (arbitrary units). Expression of the transgene was normalized on the expression of rRNA. 250 J. SIEMENS et al. © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTDMOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 Fig. 3 Characterization of transgenic lines expressing two invertase inhibitors under two root-specific promoters. The phenotype of the selected Arabidopsis transgenic plants was examined after 7 weeks of growth in long-day conditions. (A) Shoot and root fresh weights were quantified and the shoot to root ratio was determined for at least eight plants of each independent line. Values are referred to those of control plants (WT), considered as 100%. (B) Total seed weight per plant was determined in each individual plant after siliques were dried and the seeds collected. The normal distribution of seed weight was calculated for each plant and the results are plotted for all transgenic lines. The experiment was repeated three times. Table 1 Overexpression of an invertase inhibitor under the control of two different root-specific promoters (Pyk10,CrypticT80) leads to a reduction of the disease index after infection with Plasmodiophora brassicae in different independent lines. Line Club fresh weight (mg) Root fresh weight (mg) Root index Ri/Rni Disease index Col 47.0 ⫾11.6 10.7 ⫾2.1 4.3 ⫾0.5 100 Ws-0 43.7 ⫾5.3 10.2 ⫾1.7 4.1 ⫾0.4 97 pPyk10::invertase inhibitor AtC/VIF2 pPyk10::AtC/VIF2 line 1 30.3 ⫾5.9* 10.6 ⫾1.7 2.8 ⫾0.6* 82 pPyk10::AtC/VIF2 line 2 34.5 ⫾5.1* 10.8 ⫾1.9 3.5 ⫾0.5* 79 pPyk10::AtC/VIF2 line 3 36.6 ⫾4.7* 10.4 ⫾0.5 3.8 ⫾0.8* 81 pCrypticT80::invertase inhibitor AtC/VIF2 pCrypticT80::AtC/VIF2 line 1 29.2 ⫾5.1* 10.4 ⫾1.2 2.8 ⫾0.6* 88 pCrypticT80::AtC/VIF2 line 2 33.5 ⫾11.8* 9.8 ⫾1.9 3.4 ⫾0.6* 74 pCrypticT80::AtC/VIF2 line 3 44.8 ⫾6.2 10.5 ⫾0.2 4.4 ⫾0.6 89 pCrypticT80::invertase inhibitor AtC/VIF1 pCrypticT80::AtC/VIF1 line 1 33.7 ⫾9.3* 10.7 ⫾1.7 2.8 ⫾0.8* 89 pCrypticT80::AtC/VIF1 line 2 30.9 ⫾3.2* 9.5 ⫾1.3 3.3 ⫾0.5* 78 pCrypticT80::AtC/VIF1 line 3 30.0 ⫾7.2* 12.0 ⫾1.6 2.4 ⫾0.9* 83 The mean values and standard deviations of the club weight, root weight and root index (Ri/Rni), and disease indices, of the different host–pathogen combinations are given. Only experiments were presented with n>35. For the susceptible wild-type Col, with DI >95, all lines showed an infection rate of 100%. Inoculation (106spores/mL with single-spore isolate e3) was performed at 14 days after germination. *Significant difference compared with ecotype by Student’s t-test from at least eight experiments. Invertase and clubroot disease 251 © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTD MOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 In order to analyse the effect of ectopic invertase inhibitor expression in the root, in situ staining for invertase activity in transgenic plants was compared with that of wild-type control plants. This analysis prevents the described protective effect of sucrose on the inhibition of invertase activity in in vitro tests, usually performed at high concentrations of sucrose not reflecting the actual concentration of this sugar in tissue. The results showed a strongly reduced staining in the roots of transgenic plants (Fig. 4A–F), which expressed either one of the invertase inhibitors compared with control wild-type Col-0 and Wassilewskija plants (Fig. 4G,H). Staining in transgenic control plants, transformed with the empty binary plasmid, was similar to that of wild-type plants (data not shown), thus confirming that the absence of staining for invertase activity in the transgenic lines is caused by the activity of the inhibitor. As invertase enzyme activity, and not expression, was downregulated in this experiment, we did not carry out analysis on RNA or protein levels in infected roots. In addition, acid invertase activity, corresponding to extracellular and vacuolar invertases, was determined in the roots and shoots of the selected transgenic lines (pCrypticT80::AtC/VIF1-1 and pCrypticT80::AtC/VIF1-2) and compared with that in wildtype control plants (Fig. 5). No significant effect was observed on invertase activity in shoots. In roots, extracellular invertase activity was reduced in both transgenic lines. Subcellular localization of invertase inhibitors AtC/VIF1 and AtC/VIF2 In addition to measuring the activities of vacuolar and extracellular invertases in transgenic lines, we obtained further evidence that extracellular invertase is the target of the two invertase inhibitors used to downregulate invertase activity in this study (Fig. 5). Both invertase inhibitors, as well as a cell wall marker (invertase), were clearly localized in the cell wall in transient expression assays using Nicotiana benthamiana epidermal cells, as shown by confocal laser scanning microscopy. Together with invertase activity showing reduction only for the extracellular and not vacuolar form, this is good evidence that extracellular (cell wall) invertase is the target downregulated in our transgenic lines. Enhanced root-specific expression of invertase inhibitors causes a decrease in disease symptoms The clubroot disease symptoms of different transgenic lines overproducing invertase inhibitors AtC/VIF2 and AtC/VIF1 specifically in the root system under the control of two different root-specific promoters (Pyk10,CrypticT80) were compared with those of the susceptible ecotypes Col-0 and Ws-0 (Table 1). The CrypticT80 promoter caused root-specific expression and was also induced in clubroot galls, as shown in a promoter reporter line (Fig. S3, see Supporting Information). The expression pattern caused by the Pyk10 promoter revealed an elevated expression in the root system, but only weak expression in the hypocotyl (Fig. S4, see Supporting Information). Although expression was lower in Fig. 4 In situ staining of invertase activity in roots. Staining of invertase activity in roots of Col-0 (G) and Ws-0 (H) wild-type plants is shown in comparison with roots from line 2 (A), line 1 (B) and line 3 (C) of pCrypticT80::AtC/VIF1, line 1 (D) and line 2 (E) of pPyk10::AtC/VIF2, and line1ofpCrypticT80::AtC/VIF2 (F) transgenic plants. Roots shown correspond to 28-day-old grown plants. The results obtained were similar after 22 and 28 days of growth. The bar represents 1 cm. 252 J. SIEMENS et al. © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTDMOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 infected roots at a later time point, the expression level was estimated to be sufficiently high to drive root-specific invertase inhibitor expression in older galls. Attempts to determine invertase activity in situ in infected roots failed because of the high hydrogen peroxide levels present in root galls, which interfered with the enzyme assay (data not shown). As several invertases were upregulated during gall formation (Fig. 1), mutation of a single invertase gene did not result in any visible phenotype concerning gall development (data not shown). Overexpression of any invertase inhibitor gene/ protein should result in the repression of invertase activity, and thus in a lower sucrose import into infected cells. As we used 30 40 50 lc/min gFW) 3 4 5 (μg Glc/min gFW) HG 10 20 30 acuolar invertase (μg Gl 1 2 3 Extracellular invertase ( 0 WtCol0 At11-7 At11-9 Va 0 WtCol0 At11-7 At11-9 E Col-0 pCrypticT80:: pCrypticT80:: AtC/VIF1-1 AtC/VIF1-2 Col-0 pCrypticT80:: pCrypticT80:: AtC/VIF1-1 AtC/VIF1-2 AA BC FED Fig. 5 Subcellular localization of the invertase inhibitors used in this study. XFP fusions were transiently expressed in tobacco epidermal cells and analysed via confocal laser scanning microscopy. (A–C) AtC/VIF2 localization in tobacco epidermal cells; scale bar, 10 mm. (A) AtC/VIF2::YFP fluorescence. (B) Cell wall invertase::RFP fluorescence. (C) Overlap of (A) and (B). (D–F) AtC/VIF1 localization in tobacco epidermal cells; scale bar, 5 mm. (D) AtC/VIF1::YFP fluorescence. (E) Cell wall invertase::RFP fluorescence. (F) Overlap of (D) and (E). (G, H) Determination of extracellular (G) and vacuolar (H) invertase activity in wild-type Col and two transgenic lines (pCryptic::AtC/VIF1). White histograms, roots; grey histograms, shoots. Invertase and clubroot disease 253 © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTD MOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 invertase inhibitors localized to the cell wall, these will affect apoplastic invertases. Resistance to clubroot is defined by the reduction in diseased root weight, which is correlated with gall size and a reduction in resting spores, and a lower disease index (Siemens et al., 2002). Ecotypes Col-0 and Ws-0 showed the typical disease indices (DI >95) and root indices (Ri/Rni >4, roots of infected plants/roots of noninfected plants) of susceptible lines (Table 1). The plants showed clubs, which developed over the whole root and hypocotyl, and also extended in part into the rosette ground (Fig. 6). The overproduction of AtC/VIF2 and AtC/VIF1 caused an apparent tolerance to P. brassicae, indicated by reduced swellings of the root part, resulting in lower disease indices and lower root indices. All analysed overproducers of invertase inhibitor genes showed the different phenotypes presented in Fig. 6. The plants showed clear disease symptoms but, in comparison with ecotype Col-0, large clubs were induced less frequently and infected roots showed greater variation of disease symptoms. The transgenic plants Fig. 6 Phenotypes of transgenic plants after inoculation with Plasmodiophora brassciae and under control conditions. Shown are noninfected control (A) and infected roots of different lines at 28 days after inoculation (dai) (B–I). The bar indicates 1 cm. Clubroots of susceptible ecotypes Ws-0 (B) and Col-0 (C) are shown in comparison with clubroots of transgenic lines showing: hypocotyl gall but only slight infection of the root system [(D) pPyk10::AtC/VIF2 line 3 and (E) pPyk10::AtC/VIF2 line 1], weak hypocotyl and more pronounced root galls [(F) pCrypticT80::AtC/VIF2 line 2], hypocotyl and main root infected but the galls are slim [(G) pCrypticT80::AtC/VIF1 line 2 and (H) pCrypticT80::AtC/VIF2 line 1] and weak infection at lateral roots proving that root infection can also be developed [(I) pCrypticT80::AtC/VIF1 line 3] (h, hypocotyl; r, root). Photographs of these lines were selected because of their representative gall phenotypes. 254 J. SIEMENS et al. © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTDMOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 revealed reduced galls, either as hypocotyl galls with slight infection of the root system (Fig. 6D,E) or as slim galls in the hypocotyl and main root (Fig. 6G,H). In contrast with hypocotyl galls, strong gall development was rarely observed at lateral roots (Fig. 6F) but, in all lines, infected lateral roots could also grow into small galls (Fig. 6I). The disease indices and root indices indicate the susceptibility of the lines, and the mean root indices and disease indices in transgenic plants were significantly lower than those of Col-0 and Ws-0. In summary, the root indices of the investigated transgenic lines provided an indication of increased tolerance caused by the overexpression of invertase inhibitor genes. DISCUSSION Invertase gene expression is upregulated in root galls of Arabidopsis after P. brassicae infection Clubroot disease of Brassicaceae, caused by the obligate biotrophic pathogen P. brassicae, results in large root galls (clubs). For the nutrition of the pathogen during gall formation, the partitioning of assimilates into the roots has been proposed (Ludwig-Müller et al., 2009). Strong indications for the global involvement of sugar metabolism and transport originate from microarray analysis (Siemens et al., 2006). In this study, we re-examined the data for genes involved in sugar metabolism, generated promoter–reporter gene fusion lines of invertase and invertase inhibitor genes to verify the regulation of invertases and invertase inhibitors during clubroot disease (Figs 1 and 3) and produced transgenic lines with decreased invertase activity by the expression of two invertase inhibitor genes under the control of two root-specific promoters (Figs 2, 4 and 5) to investigate their role in gall formation (Fig. 6, Table 1). Invertases are key enzymes for sucrose metabolism and thus assimilate partitioning. They are encoded by small gene families in plants with different expression patterns and regulation, reflecting specific roles of the different isoenzymes in the regulation of plant growth and development. A number of studies have reflected the great impact of constitutively increasing (Büssis et al., 1997; von Schaewen et al., 1990; Sonnewald et al., 1991, 1997) or decreasing (Balibrea Lara et al., 2004; Goetz et al., 2001; Sturm and Tang, 1999; Tang et al., 1999) invertase activity in different cell compartments or specific developmental stages, resulting in marked alterations of plant phenotype not only because of the role of the enzymes in the control of assimilate partitioning, but also because of the control of sugar composition (hexose to sucrose ratio) and the crucial signalling role of sugar molecules (Koch, 1996; Rolland et al., 2006). Both soluble sugars (hexoses and sucrose) and starch accumulated in the galls of infected Arabidopsis (Brodmann et al., 2002; Evans and Scholes, 1995; Mithen and Magrath, 1992), consistent with an upregulation of expression of sucrose (sucrose synthase) and starch (starch synthase) biosynthetic genes (Siemens et al., 2006; Supporting Information). The visual display of MAPMAN-generated data in the current study organized the genes for starch/sucrose metabolism in contextual groups, which facilitates the discovery of general trends.This tool can be used to visualize the differences in sugar metabolism between an early time point (10 dai; TP1) and a later time point where galls are fully developed (23 dai;TP2) (Fig. S1A,B).At TP1, genes for starch synthesis, glycolysis and tricarboxylic acid (TCA) appeared to be consistently upregulated (also adenosine-5′- diphosphate glucose pyrophosphorylase and sucrose phosphatases), whereas sucrose synthase genes [sucrose-UDP glucosyltransferase (EC 2.4.1.13)] were mostly downregulated. At TP2, starch synthesis appeared to be further upregulated but, in part, degradation was also upregulated, reflecting the mixture of processes in the cells infected with different stages of the pathogen (Figs S1 and S2). In addition, several invertase genes were upregulated. The mechanism underlying an accumulation of soluble sugars and starch in clubroot galls remains to be elucidated, but may well be associated with increased cytokinin concentration in infected tissues (Devos et al., 2005; Siemens et al., 2006), as cytokinins induce both cell wall invertases and hexose transporters, enhance metabolic sinks and lead to increased starch synthesis (Balibrea Lara et al., 2004; Devos et al., 2006; Ehness and Roitsch, 1997; Siemens et al., 2006). The verification of microarray data was addressed by the generation of promoter–reporter gene fusion lines of invertases and invertase inhibitors and the monitoring of GUS activity during clubroot development (Fig. 1). The data using promoter:: GUS lines for several invertases partially confirmed the data obtained from microarray analysis. In particular, during later infection, when galls were clearly visible, invertase expression was upregulated (Fig. 1). In the current study, we therefore aimed to alter assimilate partitioning to the root in order to analyse the regulatory role of carbon availability on the development of clubroot disease in Arabidopsis (see below). Extracellular invertase activity is suppressed in transgenic plants With this aim, we decreased invertase activity specifically in the root by the overexpression of two different invertase inhibitor genes under the control of two different root-specific promoters, and analysed the effect on plant morphology and gall development. It should be noted that we did not imply, at this stage, a role for invertase inhibitors in clubroot formation; rather, we used these tools to downregulate invertase activity. Root specificity of the two promoters has been reported (Mollier et al., 2000; Nitz et al., 2001). The Pyk10 promoter was shown to be active in control roots and root galls (Fig. S4). As, for the gene Invertase and clubroot disease 255 © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTD MOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262 were extracted from a microarray (M-EXP-254) and analysed using MAPMAN software. (A) TP1: 10 days after inoculation (dai). (B) TP2: 23 dai. Fig. S3 Expression of b-glucuronidase (GUS) under the control of the crypticT80 promoter. (A) Control roots. (B) Roots infected with Plasmodiophora brassicae.ThecrypticT80 promoter is highly induced in root but not hypocotyl galls (r, root; h, hypocotyl). The photographs were taken by Cornelia Horn, Institute of Botany, Technische Universität Dresden, Germany. Fig. S4 Expression levels of PYK10 (At3g09260). (A) Root expression according to microarray data (E-MEXP-254). (B–D) Expression according to Arabidopsis eFP Browser (http:// www.bar.utoronto.ca; Winter et al., 2007). (B, C) Tissue-specific expression shows high levels only in the hypocotyls and roots of young seedlings and the roots of older seedlings, where the expression is mainly in the outer cell files. (D) Induction by hormones; no prominent induction by any of the tested hormones, at least at the seedling stage, was observed. dai, days after inoculation; dag, days after germination. Table S1 Description of the plant material used in this investigation. Table S2 List of primers used in this study. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. 262 J. SIEMENS et al. © 2010 THE AUTHORS MOLECULAR PLANT PATHOLOGY © 2010 BSPP AND BLACKWELL PUBLISHING LTDMOLECULAR PLANT PATHOLOGY (2011) 12(3), 247–262