The gaseous pollutant SO2 readily reacts with water to form sulfite that impacts deleteriously on animal and plant health. By modulating the level of sulfite oxidase (SO) that catalyzes the transformation of sulfites to the non-toxic sulfate, we show that Arabidopsis and tomato plants can be rendered resistant or susceptible to SO2/sulfite. Plants in which sulfite oxidase expression was abrogated by RNA interference (RNAi) accumulated relatively less sulfate after SO2 application and showed enhanced induction of senescence and wounding-associated transcripts, leaf necrosis and chlorophyll bleaching. In contrast, SO overexpression lines accumulated relatively more sulfate and showed little or no necrosis after SO2 application. The transcript of sulfite reductase, a chloroplast-localized enzyme that reduces sulfites to sulfides, was shown to be rapidly induced by SO2 in a sulfite oxidase-dependent manner. Transcripts of other sulfite-requiring enzymatic activities such as mercaptopyruvate sulfur transferases and UDP-sulfoquinovose synthase 1 were induced later and to a lesser extent, whereas SO was constitutively expressed and was not significantly induced by SO2. The results imply that plants can utilize sulfite oxidase in a sulfite oxidative pathway to cope with sulfite overflow.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Sulfur dioxide (SO2) is a gaseous pollutant emitted by natural sources, such as microbial and volcanic activities, and by anthropogenic combustion of sulfur-containing fossil fuels. The problem in developing industrial countries is exacerbated by combustion of ubiquitous sulfur-containing coal, with China leading the world as an SO2 emitter (Marshall, 2002). In water, SO2 readily hydrates to form the sulfite ions, HSO31 − and SO32 −, strong nucleophiles that can deleteriously react with a wide variety of cellular components, affecting human and plant health (Murray, 1997). SO2 enters plants via their stomates, and damage is correlated with the degree of stomatal opening (van der Kooij et al., 1997; Rennenberg and Herschbach, 1996). At below toxic levels, plants are able to utilize SO2. Indeed, sulfur assimilation and biomass production are correlated with SO2 in the air (Rennenberg, 1984). However, above a certain threshold, which differs between plant species, SO2 toxicity leads to visible effects that include chlorosis (chlorophyll destruction), necrosis (plant tissue death) and long-term yield reduction (van der Kooij et al., 1997; Legge and Krupa, 2002; Noji et al., 2001). Debilitation of plants by SO2 can also facilitate pathogen ingress. For example, in the UK, a retrospective study showed that proliferation of the necrotrophic pathogen Phaeosphaeria nodorum, which is responsible for destroying millions of tons of grain worldwide, correlated with sulfur dioxide pollution (Bearchell et al., 2005).
Sulfite can be oxidized to sulfate by the molybdenum co-factor-containing enzyme, sulfite oxidase (SO; EC 18.104.22.168). The enzyme catalyzes a two-electron transfer reaction in which the electrons from sulfite reduce the molybdenum co-factor redox center. The electrons are subsequently transferred to molecular oxygen with simultaneous formation of hydrogen peroxide in addition to sulfate as demonstrated recently using recombinant sulfur oxidase from Arabidopsis thaliana (AtSO;Eilers et al., 2001; Hansch et al., 2006). Mutations in the molybdenum co-factor biosynthetic loci cnxA–cnxF in Nicotiana plumbaginifolia and the nar2a mutant in barley simultaneously abrogate the activities of SO and the other known plant molybdenum co-factor-containing enzymes, nitrate reductase (NR; EC 22.214.171.124), xanthine dehydrogenase (XDH; EC 1.1.204) and aldehyde oxidase (AO; EC 126.96.36.199; Eilers et al., 2001; Gabard et al., 1988; Müller and Mendel, 1989; Walker-Simmons et al., 1989). Specific mutations in structural genes have also been described for NR that diminish nitrate assimilation (Wilkinson and Crawford, 1993), for XDH that abrogate superoxide production and probably purine catabolism (Yesbergenova et al., 2005), and for AO that diminish the biosynthesis of the phytohormone ABA (Seo et al., 2000). However, specific SO mutations have yet to be reported in plants.
The vertebrate sulfite oxidase is a mitochondrial enzyme containing a heme domain, with cytochrome c serving as the physiological electron acceptor. Human sulfite oxidase deficiency leads to severe neurological abnormalities that often result in death in infancy (Garrett et al., 1998). Among the eukaryotes, plant SO is the smallest molybdenum co-factor-containing enzyme known so far. The enzyme lacks contiguous redox-active centers such as FAD, heme or Fe-S (Eilers et al., 2001). In Arabidopsis, AtSO is localized in peroxisomes (Eilers et al., 2001; Nowak et al., 2004), and is thus distinct from the multi-enzyme sulfur assimilatory pathway localized to the chloroplast. It has been speculated that SO is required for removing excess sulfite that accumulates upon decomposition of sulfur-containing amino acids or sulfated metabolites (Hansch and Mendel, 2005; Heber and Huve, 1998). In contrast to this pathway, in sulfur assimilation, plants reduce the ubiquitous sulfate ion through a series of steps that includes activation by ATP sulfurylase and subsequent reduction to the sulfite form by APS reductase (for recent reviews, see Leustek et al., 2000; Saito, 2004). The sulfite is then reduced by sulfite reductase (SiR; EC 188.8.131.52) by a process that transfers six electrons from ferredoxin to produce the fully reduced sulfide form for incorporation into sulfur-containing amino acids (Garsed and Read, 1977; van der Kooij et al., 1997; Leustek et al., 2000Saito, 1999).
In addition to SO and SiR, additional enzymatic activities in plants are capable of catalyzing sulfite conversion. The UDP-sulfoquinovose synthase 1 (SQD1) protein is localized in the chloroplast and can participate in detoxifying SO2/sulfite as it catalyzes the transfer of sulfite to UDP-Glc, giving rise to UDP-sulfoquinovose (Sanda et al., 2001), an intermediate product for the biosynthesis of sulfolipids required for proper function of the photosynthetic membranes (Saito, 2004; Yu et al., 2002). Additionally, the mitochondrion- and cytosol-localized mercaptopyruvate sulfurtransferases (MST1 and MST2, respectively), also known as rhodaneses, have been shown to catalyze synthesis of the less toxic compound thiosulfate in the presence of 3-mercaptopyruvate and sulfite (Papenbrock and Schmidt, 2000a,Papenbrock and Schmidt, 2000b; Tsakraklides et al., 2002).
Thus, sulfite can be processed in plants by multiple pathways, and the exact physiological role of SO activity has yet to be established. Here we show that SO is constitutively expressed and is not significantly induced by SO2, whereas the SiR transcript involved in sulfur assimilation is highly induced by SO2 in an SO-dependent manner. By modulating SO levels through RNA interference (RNAi) and overexpression (OE) lines in Arabidopsis and tomato (Solanum lycopersicum), we provide physiological and molecular evidence that SO can protect plants from sulfite overflow and toxic doses of SO2 gas.
SO is present as a single gene in Arabidopsis. Digital Northern and response activities of a 2507-slide size microarray collection by GENEVESTIGATOR (http://www.genevestigator.ethz.ch, Zimmermann et al., 2004) showed that SO transcript is present in all plant organs, and its level did not change significantly under any of the 75 diverse experimental conditions. Western blot analysis of SO in leaves of Arabidopsis wild-type (Col) plants sampled either immediately after SO2 exposure (2 ppm for 2 h) or 24 h later showed only slight enhancement in SO protein in response to SO2 treatment (Figure 1a). To obtain a higher-resolution image of SO expression, a 1562 bp promoter fragment was used to direct the expression of the β-glucuronidase (GUS) reporter gene in Arabidopsis. Out of 20 transformed lines, five had significant levels of expression. Histochemical analysis showed GUS expression in all tissues, with a higher degree of staining in stem, hypocotyls, root vasculatures (Figure 1b), root tip and young inflorescences (Figure 1c,d). Immuno-detection confirmed constitutive expression of a 45 kDa polypeptide in all plant organs (Figure 1e) except root tissue, in which SO migrated in two distinct forms. The different polypeptides may represent the products of different post-translational modifications. The results show an overlap between expression potential as determined by microarray analysis, promoter fusion expression and protein accumulation.
SO activity in wild-type and SO-modified plants
In order to evaluate the role of SO, its expression levels were modulated in Arabidopsis (AtSO) and tomato (LeSO) plants. A full-length tomato SO cDNA was isolated (LeSO, GenBank accession number DQ853413, see Experimental procedures) that exhibited 77% identity with AtSO. The sequence includes the C-terminus tripeptide ANL, which is in agreement with the consensus peroxisomal targeting signal type 1 (A/C/G/S/T–H/K/L/N/R–I/L/M/Y; Mullen et al., 1997a,b). In order to assess SO function, AtSO and LeSO RNA interference (RNAi) lines lacking immuno-detectable SO polypeptide (13 and three independent lines in Arabidopsis and tomato, respectively) were generated (Ri lines, Figures 2a and 3a). In addition, constitutive overexpression (OE) lines exhibiting up to eightfold higher protein levels (five and three independent lines in Arabidopsis and tomato, respectively) were generated (OE lines, Figures 2a and 3a). Extracts of wild-type, RNAi and OE lines were examined for SO activity by employing an assay that measures the reduction of ferricyanide (Eilers et al., 2001). The results showed that the activity in Arabidopsis and tomato RNAi and OE extracts was either >68% lower or >231% higher, respectively, compared with that in wild-type plants (Figure 2b, upper left, and Figure 3b, upper left). The correlation between activity and the measured SO protein levels was not exact and may be due to SO non-specific activity. Such activity has been reported in the non-leaky molybdenum co-factor mutants of Nicotiana plumbaginifolia that should have no SO activity (Eilers et al., 2001).
To further examine SO-dependent activity, an independent assay was developed based on the ability of SO to generate H2O2 during sulfate formation. In this case, the measured H2O2 product was 56 or 73% higher in wild-type than in RNAi extracts of Arabidopsis and tomato, respectively, and the levels in OE modified plants were 2.3- and 3.3-fold higher than in Arabidopsis and tomato wild-types, respectively (Figure 2b, upper right, and Figure 3b, upper right). If the residual activity measured in RNAi lines was considered as non-SO dependent activity, then the activities in OE lines were 4.6- and 6.4-fold higher than the wild-types in Arabidopsis and tomato, respectively, which is consistent with the range of protein fold increase seen in the Western blot analysis (Figures 2a and 3a).
In an attempt to directly correlate the H2O2 generation with SO, we used a chromogenic horseradish peroxidase (HRP) in-gel assay in which accumulating H2O2 serves as a proton-accepting substrate while o-dianisidine serves as a proton donor. This method was a modified version of those described by Yesbergenova et al. (2005) and Manchenko (1994). Extracts were fractionated under native PAGE conditions, and broad orange-coloured bands of H2O2-generating activity could be detected in Arabidopsis and tomato (Figure 2b, lower left, and Figure 3b, lower left). The band was less intense in wild-types Col and RR, and was absent in Ri lines, but was significantly greater in intensity in OE leaves (four and sixfold). The area of the activity bands was excised and re-fractionated by denaturating SDS–PAGE, and immunoblotted using SO-specific antisera. In this case, activity bands from OE and RR but not from RNAi plants yielded cross-reacting polypeptides with SO antisera of correct molecular weight. The intensities of immunoblot activity were three- and fivefold higher in OE plants than in Arabidopsis and tomato wild-types, consistent with the activity gel measurement (Figure 2b, lower right, and Figure 3b, lower right). The results indicate that SO polypeptide levels correlate with SO-dependent H2O2 production.
We wished to consider in planta changes in the levels of the immediate substrate and product of SO activity in vivo under dynamic conditions of sulfur dioxide application. Sulfite is difficult to measure as it is maintained at low levels in plant tissues and rapidly oxidizes in extracts (Tsakraklides et al., 2002). We therefore chose to monitor the sulfate concentration. Leaves of Arabidopsis and tomato plants were treated with 2 ppm SO2 for 2 or 4 h, respectively. The basal level of sulfate in OE modified plants was lower than that in wild-type plants (15% for both Arabidopsis and tomato; P < 0.05). In order to accurately assess the involvement of SO in channeling sulfite overflow, we compared the sulfate content of plants before and after their fumigation by SO2. Increases of 15 and 39% for Arabidopsis and 24.7% and 43.5% for tomato were detected in wild-type and OE plants. In contrast, the effect of SO2 treatment on sulfate levels in RNAi plants resulted in smaller increases of 2.7% and 6.7% in Arabidopsis and tomato, respectively (Figures 2c and 3c). This result shows that the amount of sulfate that accumulates during SO2 application is influenced by the SO level in plants.
Sulfur dioxide toxicity in wild-type and SO-modified plants
The results above indicate that transgenic plants harboring increased or reduced amounts of SO show differential activity in processing sulfite. It was therefore of interest to examine the response of SO-modified plants to sulfite toxicity. Leaf discs of wild-type and transgenic plants were treated with 7 mm Na2SO3, and chlorophyll content, a sensitive indicator of leaf health, was monitored. Leaf discs of RNAi lines showed significantly higher chlorosis and damage than wild-type and OE lines (Figure 4a and Figure S1). After 24 h, reductions in the level of chlorophyll of 30% and 50% were detected in RNAi lines, compared with reductions of 10% and 0% in wild-type and OE lines, respectively (Figure 4b). Tomato RNAi and OE lines exhibited a similar response to Na2SO3 as Arabidopsis (Figure S1). Hence, sensitivity to sulfite is correlated with loss or gain of SO activity.
SO2 is a highly cell-permeable toxic gas that can reach levels of 2 ppm in heavily polluted geographical regions (Legge and Krupa, 2002). When Arabidopsis wild-type and transgenic lines were exposed to 1 ppm of SO2 for 2 or 4 h, no significant damage was observed in the leaves within 4 days (data not shown). However, when plants were exposed to 2 ppm SO2 for 2 or 4 h and examined 4 days later, increasing levels of leaf damage were seen, correlating with the duration of exposure (Figure 5a,b and Figure S2). Quantitative analysis of the leaves of RNAi lines showed damage index values that were 5–10 times greater than the damage levels in wild-type plants after 2 h of exposure to 2 ppm (Figure 5b, top left). After 4 h of exposure, wild-type plants were also damaged, while OE lines showed nearly threefold less damage (Figure 5b, top right). Interestingly, after 4 h of exposure, leaf growth was severely arrested in wild-type and RNAi lines (50% of non-treated controls), but <10% growth retardation occurred in OE lines (Figure 5b, middle right). The index of residual chlorophyll showed a similar differential change (Figure 5b, lower panels). Importantly, RNAi tomato leaves were more damaged than RR (wild-type) and OE plants when exposed for 4 h to 2 ppm SO2 (Figure S3a). Although the calculated relative leaf area after treatment was similar to that of RR, RNAi leaves were significantly more damaged, while OE plants showed a minimal reduction in size and leaf damage (Figure S3b). The results indicate a distinct threshold for SO2 toxicity in wild-type plants, and imply that the physiological capacity for the plant to detoxify SO2 is enhanced by increasing SO activity.
Senescence and wound-associated gene expression in SO-modified plants
The symptoms of SO2 poisoning are reminiscent of leaf senescence and wounding stress. We therefore monitored the activity of genes that are known to be associated with these processes. WRKY6 (senescence-related transcription factor), ERD/SAG15 (senescence-associated gene) and ACX1 (acyl CoA oxidase 1) are triggered during early senescence and plant defense responses (Castillo et al., 2004; He et al., 2002; Laloi et al., 2004; Robatzek and Somssich, 2001). Treatment of wild-type and mutant SO expression plants with 2 ppm SO2 showed a rapid 4–10-fold accumulation of these transcripts in all lines (Figure 6, left). However, the induction level remained elevated 24 h later in Arabidopsis RNAi lines but not in wild-type or OE lines. A different set of marker genes that emphasize late processes in senescence and stress were also monitored. These include ER5/LEA (ethylene-responsive 5/late embryogenesis-like protein), XERO1/TAS14 (dehydrin) and SRG1 (senescence-related gene 1). The genes were shown to be activated in later stages of leaf senescence, drought and wounding (Alsheikh et al., 2005; Callard et al., 1996; Reymond et al., 2000). In this case, wild-type and OE lines showed little changes; however, RNAi lines displayed high levels of induction after 24 h (Figure 6, right). Taken together, as the levels of these transcripts reflect the status of cellular stress responses, the results indicate that AtSO plays an important role in preventing stress induced by toxic levels of SO2.
Expression of genes that utilize sulfite in SO-modified plants
SO, SiR, SQD1 and MST are enzymes that use sulfite as substrate and could play pivotal roles in SO2 metabolism. We therefore monitored the expression level of these genes. When Arabidopsis plants were exposed to 2 ppm SO2 for 2 h and examined immediately, the levels of AtSO transcripts were reduced by at least twofold in wild-type lines. After 24 h the levels returned to normal (Figure 7, upper left). Similar results were obtained for LeSO (Figure 7, lower left). In addition, immuno-detection of SO revealed only a slight increase in the amount of polypeptide in wild-type plants (Figure 1a for Arabidopsis; data for tomato not shown). Thus, AtSO and LeSO transcript and protein levels are not highly sensitive to application of SO2. The relative levels of AtSO and LeSO transcripts in RNAi plants changed to a greater extent but their absolute levels are inherently very low (approximately 104 and 108 times lower than wild-type and OE transcripts, respectively), and any change in these levels may be a reflection of changes in general RNAi-specific processes. In contrast to SO, the AtSiR transcript was enhanced more than 30-fold immediately after treatment in wild-type plants and between seven- and 10-fold in RNAi plants (Figure 7, upper right; note different scales). However, the level of AtSiR in OE lines was induced to a lower extent (1–3-fold). When measured 24 h later, AtSiR expression was elevated in RNAi lines by twofold, but not in wild-type and OE lines where it returned to less than normal levels. In this respect, tomato plants showed a different response. The LeSiR transcript was not induced in response to SO2 treatment in wild-type RR plants, although a significant increase was obtained in RNAi plants but not in OE plants (Figure 7, lower right). These results indicate that the AtSiR transcript level, and to lesser extent LeSiR, are regulated by SO2, particularly under conditions of limitation in SO activity.
SQD1, MST1 and MST2 represent genes that catalyze the diversion of sulfite to other assimilatory pathways (Papenbrock and Schmidt, 2000a,Papenbrock and Schmidt, 2000b; Tsakraklides et al., 2002). It was therefore of interest to examine whether their transcripts were also regulated by fumigation with SO2. As shown in Figure 8, no significant differences in transcript of these three genes were detected immediately after exposure (0 h). However, after 24 h, in both Arabidopsis and tomato, RNAi plants but not wild-type or OE plants displayed elevated levels of transcript of the three genes after exposure to SO2/sulfite treatment. These results indicate that the late-responsive (24 h) SO-dependent upregulation of SQD1, MST1 and MST2 transcripts is distinct from that of the early-responsive (0 h) SiR transcripts.
SO expression and activity
SO is thought to play a role in cellular sulfur turnover (Hansch et al., 2006; Heber and Huve, 1998). It is shown here to be constitutively expressed in all Arabidopsis plant organs (Figure 1e). Although the proportion of plants exhibiting significant GUS expression was low (only 20% of the plants), all plants showed an identical pattern of expression. The relative enrichment of SO in the region of tissue vasculature and the plant-wide distribution of SO activity are consistent with a ‘housekeeping’ function in basal sulfur catabolism (Heber and Huve, 1998; Johnson and Rajagopalan, 1979). However, the exact role played by SO may differ depending on its tissue localization. For example, it was also enriched in root tips, an organ that is unlikely to participate in sulfur turnover. In this case, it may function to detoxify incoming soil sulfites. The juxtaposition with root sulfate transporters such as SULTR4;1 and SULTR4;2 would mean that the resultant sulfate product of SO activity would be immediately available for tissue transport (Kataoka et al., 2004).
SO is present as a single gene in Arabidopsis. However, downregulation of its activity in Arabidopsis and tomato lines to levels that cannot be detected by immunological methods revealed no obvious phenotype other than hypersensitivity to exogenous sulfite. This indicates that, under the prevailing non-stress conditions that exist in growth room conditions, SO activity is dispensable. Alternatively, redundant non-SO activities may exist that can contribute to the detoxification of sulfite. For example, detoxification of SO2 by apoplastic peroxidases has been described in barley leaves (Pfanz et al., 1990), and a chloroplast-based sulfite oxidizing activity has been described in wheat and spinach leaves (Jolivet et al., 1995a,b). The ferricyanide method revealed discrepancies between expected rates of SO activity and SO protein levels. While this may be attributable to non-SO-related activities, the alternative SO activity detected in extracts may be due to limitations in the specificity of the ferricyanide method. We have demonstrated direct release of H2O2 from isolated SO using the in-gel technique, a result that is consistent with the activity detected in recombinant SO (Hansch et al., 2006). By taking advantage of this aspect of SO activity, an assay was devised that is based on quantitative measurement of H2O2 production in extracts (Figures 2b and 3b). This yielded a greatly improved correlation between protein levels and SO enzymatic activity. The enhancement in sulfate accumulation during fumigation for 2 h was 2.1 versus 4.6 μmol g−1 FW for Arabidopsis and 7.3 versus 10.9 μmol g−1 FW for tomato, in wild-type and OE plants, respectively (Figures 2c and 3c). This represents 2.2- and 1.5-fold increases in sulfate accumulation in Arabidopsis and tomato OE plants, respectively, which are lower than the differences in SO protein levels and enzymatic activities between OE and wild-type extracts (Figures 2a,b and 3a,b). As SO protein level was not significantly affected during the fumigation (for Col SO, see Figure 1a; data not shown for RR and OE plants), the discrepancy may be attributed to competing consumption of sulfate via the reductive pathway (Leustek et al., 2000; Saito, 2004).
Ectopic expression of SO enhances tolerance to sulfur dioxide
We show that plant lines lacking SO are more susceptible to the application of sulfites and sulfur dioxide, and that ectopic overexpression of SO enhances tolerance. The protection afforded by overexpression of AtSO and LeSO does not interfere with normal sulfur reductive metabolism, as these opposing metabolic demands probably occur in different subcellular compartments. Other metabolic engineering approaches have been shown to have an impact on sulfite metabolism. For example, overexpression of cysteine synthase or yeast MET25, which encodes O-acetylhomoserine sulfhydrylase activity (Matityahu et al., 2006; Noji et al., 2001), conveyed a degree of protection to exogenous sulfite. This protection may be due to metabolic detoxification of sulfite by its enhanced fixation into sulfur-containing products, e.g. cysteine, or increasing the pool of glutathione, a known reactive oxygen species scavenger (Dittrich et al., 1992; Heber and Huve, 1998; Saito et al., 1994). The protective effect was marginal, and, by their very nature, the use of multi-step pathways to dissipate toxicity can lead to a wide range of effects on native metabolic profiles. Thus, plants with an excess of sulfur assimilatory flux show deleterious phenotypes (Hacham et al., 2002; Sirko et al., 2004; Tsakraklides et al., 2002). Furthermore, in those cases, the long-term physiological ramifications of fumigation tests with SO2 gas were not reported, so that it is difficult to gauge the potential protection afforded by the reductive detoxification pathway. SO2 is highly permeable, and its application as a gas better mimics the natural sulfur pollutant–plant interaction. Its greater permeability compared to sulfite is explained by the fact that the latter negative oxosulfur species have greater difficulty negotiating membrane barriers compared to the neutral SO2.
Regulatory steps in sulfite metabolism
The observation that some tolerance can be afforded by enhancing the biosynthesis of sulfur-containing amino acids implies that flux through sulfate reduction would not be rate-limiting (Hawkesford and De Kok, 2006). However, increasing sulfite production by overexpression of the bacterial enzyme adenosine phosphosulphate (APS) reductase (EC 184.108.40.206) enhanced the incorporation of reduced sulfur into a variety of compounds (Tsakraklides et al., 2002). Furthermore, plant APS reductase and, to a lesser extent, ATP sulphurylase (EC 220.127.116.11) exert control over the sulfur assimilatory pathway (Kopriva and Koprivova, 2004). In addition, sulfate uptake and APS reductase are regulated by thiols, e.g. reduced gluthione (GSH) (Vauclare et al., 2002). Thus, multiple nodes in the regulation of sulfate metabolism exist. This work exposes an additional regulatory step in the sulfite assimilatory pathway in Arabidopsis by showing that AtSiR is highly regulated by the presence of SO2/sulfite (30-fold induction of Arabidopsis transcript level, Figure 7). Indeed, in AtSO OE lines, the AtSiR transcript increased much less after exposure to SO2 than in wild-type or RNAi lines, indicating that the local cellular SO2/SO32 − concentrations were lower due to their rapid conversion to sulfate (Figure 2c). This is also evident in the direct measurement of sulfate levels, which were enhanced in OE plants after SO2 exposure (Figure 2c). Additional support for this regulatory step was obtained in tomato, but, in this case, enhanced levels of LeSiR were achieved only in LeSO-compromised plants (Figure 7). More SO2/sulfite was converted to sulfate in wild-type and LeSO OE tomato compared to LeSO RNAi plants (Figure 3c). The increased transcript level does not necessarily indicate increased SiR protein or activity levels, but may indicate increased turnover of SiR to maintain normal rates of flux.
Our results indicate that, under conditions of excess sulfite flux, SiR regulation is connected to SO activity levels, and that transcription of AtSiR and LeSiR is rapidly induced by the residual levels of sulfite in the cell. The SO-dependent upregulation of SQD1, MST1 and MST2 is distinct from that of AtSiR and LeSiR and is of a late-responsive type (24 h). Elevated SiR transcripts were detected immediately after SO2 application, indicating that this promoter may either be directly sensitive to SO2 or that it was induced by the early damage that was incurred. Such damage is evidenced by the early appearance of WRKY6 and other wound-responsive transcripts. However, direct wounding has not been shown to modify the behavior of AtSiR using GENEVESTIGATOR (Zimmermann et al., 2004).
Taken together, the results imply that (i) SO is a major player in SO2/sulfite conversion, (ii) the absence of SO uncovers possible roles for SiR in rapid SO2/sulfite detoxification, and (iii) SQD1, MST1 and MST2 may play a further role in late conversion of SO2 and sulfite derivates in plants. The modulation of sulfite levels by SO expression demonstrates the physiological cross-relationship between the oxidative and reductive pathways in sulfite metabolism in which AtSO serves as a physiological safety valve. For short-term exposure (2 h), assimilatory sulfur pathways may be able to cope and prevent toxic effects (compare damage levels in wild-type after 2 h exposure in Figure 5b and transcript levels in Figure 7). However, above a certain level of exposure (4 h), there is a need for further detoxification that only AtSO can provide, which is enhanced in overexpression lines (see damage level in OE lines compared to wild-type lines, Figure 5a,b). One may have expected AtSiR levels to be highest in RNAi plants. While this is true for LeSiR in tomato, it is not the case in Arabidopsis. In that case, either direct damage or secondary stress effects, as evidenced by the enhanced stress-related transcript levels, serve to prevent the full AtSiR increase (Figure 7).
SO levels moderate SO2 toxicity
The symptoms of SO2 poisoning are reminiscent of leaf senescence and wounding stress. Premature senescence leading to the induction of other senescence-related genes has been shown to be induced by environmental stresses such as drought, UV-B irradiation and ozone, and may indicate that a common molecular nexus is affected (Buchanan-Wollaston et al., 2003). We show here that a high constitutive level of AtSO can prevent the induction of early-responsive and late-responsive senescence transcripts by SO2, whereas, in the absence of AtSO, these transcripts are highly induced by SO2 poisoning (Figure 6). The upregulation of early-responsive genes that lasted 24 h after SO2 poisoning in RNAi but not in OE plants may lead to additional regulation of wound, senescence and pathogen defense-associated expression. Indeed, the WRKY6 transcription factor has been shown to control senescence and pathogen defense-associated genes (Robatzek and Somssich, 2001, 2002). The protease regulator, ERD1 that encodes caseinolytic protease regulatory subunit, was shown to be involved in the degradation of chloroplast proteins subjected to artificial senescence related to ABA and ethylene (Weaver et al., 1998). Upregulation of the ACX1 protein that is the first specific β-oxidation pathway enzyme involved in JA biosynthesis is an indication of stimulated stress pathways related to wounding (Castillo et al., 2004). Transcription of the late-responsive genes XERO1, LEA and SRG1 was significantly upregulated only in RNAi plants 24 h after SO2 poisoning, but not in wild-type and OE plants (Figure 6). Late-responsive senescence transcripts were detectable under stress conditions (Alsheikh et al., 2005) and under conditions where tissue damage was irreversible (Gepstein et al., 2003). SRG1, which is active at late leaf senescence stages (Callard et al., 1996), was upregulated 500-fold in RNAi plants 24 h after SO2 poisoning. Interestingly, inspection of digital Northern blots using GENEVESTIGATOR (Zimmermann et al., 2004) disclosed that the late-responsive gene SRG1 and the early-responsive gene ERD1 were highly induced when Arabidopsis cell cultures underwent programmed spontaneous or heat-induced cell death (Swidzinski et al., 2002). Significantly, the presence of constitutive levels of AtSO protected wild-type and OE plants from the onset of this programmed cell death-like process (Figure 6).
SO2 phytotoxicity strongly depends on the level and duration of exposure to SO2, combined with inter- and intra-specific variation in plant susceptibility (van der Kooij et al., 1997). We show here that the constitutive level of SO and the degree of SiR inducibility differ slightly between Arabidopsis and tomato and may be one source of this variation. Thus, the modulation of cellular SO levels revealed the potential for sulfur-based metabolic cross-talk between different organelle compartments, and has implications for achieving enhanced tolerance to SO2 toxicity.
Plant materials and growth conditions
A. thaliana plants (ecotype Columbia) were grown in 50% Murashige and Skoog (MS) agar plates or trays containing low-nutrient soil. Tomato plants (Lycopersicon esculentum/Solanum lycopersicum Mill. cv. Rheinlands Ruhm) were grown in pots filled with a peat and vermiculite (4:1 v/v) mixture containing slow-release high-N Multicote 4 with micro-elements (0.3% w/w; Haifa Chemicals Ltd; http://www.haifachem.com/). A. thaliana and tomato plants were grown in a growth room under 16 h light/8 h darkness, 22°C, 75–85% relative humidity, and 100 μ mol m−2 s−1 as described previously (Yesbergenova et al., 2005).
Preparation of RNA
For quantitative RT-PCR and AtSO gene silencing, total RNA was prepared using an Aurum™ total RNA mini kit (Bio-Rad; http://www.bio-rad.com/). For cloning of tomato SO (LeSO) cDNA, total RNA was extracted using an RNeasy plant mini kit (Qiagen; http://www.qiagen.com/). Total RNA was extracted according to the manufacturer’s instructions.
cDNA synthesis, overexpression and gene silencing constructs
For gene silencing, a 254 bp PCR product of AtSO was isolated from Arabidopsis using an iScript™ cDNA synthesis kit (Bio-Rad). The fragment was introduced in the sense or antisense oritnations in pRNA69 plasmid containing a CaMV 35S promoter upstream of the sense and antisense multi-cloning site separated by a 631 bp intron (gift from Dr Yuval Eshed, Weizmann Institute of Science, Rehovot, Israel). The forward primer was CGGGATCCCTCGAGGCTCGTTCGGTCAAAT, containing BamHI and XhoI restriction sites (underlined), and the reverse primer was CCATCGATGAATTCCTTTCTATCCCGCGTCCA, containing ClaI and EcoRI sites (underlined). The sense fragment was ligated to pRNA69 plasmid via the restriction sites XhoI and EcoRI, and then the antisense fragment was ligated to the plasmid through the restriction sites BamHI and ClaI. The resulting construct was digested, and the fragment containing the 35S promoter and the inserted AtSO fragments was inserted via NotI flanking sites into the binary vector pML-BART (gift from Dr Yuval Eshed). The construct was introduced into Agrobacterium tumefaciens strain GV3101 by electroporation, and transformed into Arabidopsis ecotype Columbia plants using the floral dip method (Clough and Bent, 1998). For verification of AtSO interference lines, the antisense-specific fragment was amplified using the primer GGGCTTTGACATCTTTGAAGAAAAC that spans the intron region of the pRNA69 plasmid, with TCAATTGGGATAATATCAACTGGTCCTC as the reverse primer. The sense-specific fragment was amplified using the reverse primer AAAACTTACATTCTTGGCAGCAGTG that spans the intron region of the pRNA69 plasmid, with TCAATTGGGATAATATCAACTGGTCCTC as the forward primer. Transformed lines were selected by resistance to Basta® (glufosinate ammonium; Aventis CropScience; http://www.aventis.com). For transgene verification, genomic DNA was examined for the presence of the 255 and 202 bp PCR products flanking the prokaryotic intron and the antisense and sense cDNA inserts, respectively, and these were separated on a 2% agarose gel, excised from the gel and sequenced.
For, LeSO gene silencing, tomato cDNA was synthesized from total RNA (1.5 μg), subjected to first-strand synthesis using SuperScript II reverse transcriptase (Gibco BRL; http://www.gibcobrl.com), according to the manufacturer’s procedure, and the GeneRacer™ (Invitrogen; http://www.introgen.com) oligo(dT) primer GCTGTCAACGATACGCTACGTAACGGCATGACAGTG(T)24. PCR amplification of the full-length LeSO cDNA was conducted on one-tenth of the reaction using forward primer CAAGTCACACAGCACCGTTT and reverse primer GCTGTCAACGATACGCTACGTAACG, resulting in a 1581 bp PCR product. The resultant full-length cDNA of LeSO was directly ligated into pGEM-T Easy (Promega; http://www.promega.com/) and sequenced (GenBank accession number DQ853413). For LeSO gene silencing, a 262 bp PCR product was introduced into pRNA69 plasmid as described for AtSO gene silencing. Forward and reverse primers were CGGGATCCCTCGAGAGACTTGTTTATGAAG and CCATCGATGAATTCCTTACACTTGTCAATGCT. The resulting construct was digested and ligated into the NotI site in the binary vector pART27 (a gift from Dr Yuval Eshed), introduced into Agrobacterium tumefaciens strain GV3101 and used to transform tomato plants (Lycopersicon esculentum/Solanum lycopersicum Mill. cv. Rheinlands Ruhm) (McCormick et al., 1986). For LeSO interference line verification, the antisense- and sense-specific fragments were amplified using the primers that span the intron region of the pRNA69 plasmid, as described for Arabidopsis, and GGATATTGCTGCTTTAGGAAATGCTGT and GGATATTGCTGCTTTAGGAAATGCTGT as reverse and forward primers, respectively. Genomic DNA extracted from resulting transgenic plants resistant to kanamycin was employed as template. The 240 and 187 bp PCR products were separated and sequenced for verification as described above.
For LeSO overexpression, the full-length LeSO cDNA was introduced into the pART7 plasmid using forward primer ACACTCGAGATGCCTGGGATTAAAGGGCC containing an XhoI restriction site (underlined) and reverse primer TACGAATTCCTAAAGATTTGCTTGACCAAC containing an EcoRI site (underlined). The resulting construct was digested, and the fragment containing the 35S promoter upstream to LeSO was inserted via NotI digestion into the binary vector pART27, introduced into Agrobacterium tumefaciens strain GV3101 and transformed into tomato as described above. For verification, the 185 bp PCR product was amplified using genomic DNA as template, forward primer ATCATTGCGATAAAGGAAAGGCTATCA that spans the multi-cloning site region of the pART7 plasmid and reverse primer GAATAATCGGAAGGCCCTTTAATCC for the cDNA insert. The PCR product was separated and sequenced as described above.
To generate the AtSO overexpression lines, full-length AtSO cDNA was obtained using the forward primer AGTCTCGAGATGCCTGGAATTAGAGGTCCTTCGG containing an XhoI restriction site (underlined) and the reverse primer TACGAATTCCTACAAGTTAGAGTGGCCAAGCCGG containing an EcoRI site (underlined). The resulting 1182 bp PCR product spanning the AtSO coding region was ligated to pART7 and thereafter inserted into pML-BART and transformed into Arabidopsis (ecotype Columbia) plants as described above. Transformation verification was performed, as described for tomato, by identification of a 185 bp PCR product, using GAGTATTCCGAAGGACCTCTAATTCCA as the specific reverse primer for the cDNA insert.
Genetic characterization of the lines was based on resistance of the self-pollinated T1 progenies to Basta® or kanamycin and verified by PCR analysis. Homozygous Arabidopsis and tomato SO-modified lines that contained single-site transgene insertions were used for further experiments.
Plants harboring the pAtSO:GUS construct and histochemical GUS staining
For construction of the AtSO promoter–GUS fusion, a forward primer AGACTCGAGTATGACCTTGGGATATGGTCCTGTC containing an XhoI restriction site (underlined) and a reverse primer TCCAAGCTTTCTTCTTTCGAGGAGGAGATACCGAG containing a HindIII site (underlined) were used to amplify the Arabidopsis BAC F1C9 template (accession number AC011664; obtained from the Arabidopsis Biological Research Center, Colombus, OH). The resulting 1562 bp PCR product containing the AtSO promoter was ligated to pRITA plasmid (a gift from Dr Yuval Eshed), via the XhoI and HindIII restriction sites upstream of the β-glucuronidase (GUS) reporter gene, and introduced by NotI ligation into pML-BART. The resulting construct was transformed into Arabidopsis (ecotype Columbia) plants as described above. For verification of plant modification, the primer AGGAAACAGCTATGACCATGATTACGA that spans the multi-cloning site region of the pRITA plasmid and the reverse primer TTTGTGGTAGACGGAGGTATACGAGTG for the promoter region were used to amplify a 189 bp PCR product that was confirmed by sequencing. T2 plants harboring the constructs were stained using 5-bromo-4-chloro-3-indolyl-β-d-GlcUA.
Treatment of plants and leaf discs
Exposure to SO2 was carried out in a 40-liter growth chamber continuously supplied with a calibration gas cylinder containing 250 ppm SO2 in air (Scientific & Technical Gases; http://www.stgas.eu/) with an SO2 control system (WGA-50-MAS, Emproco; http://www.emproco.com/). The control system was designed to maintain a stable SO2 concentration in the growth chamber. The tomato and Arabidopsis wild-type and SO-modified plants were exposed to 1 and 2 ppm SO2 for 2 or 4 h, respectively, under light (40 μmol m−2 s−1) at 25°C with a relative humidity of 85–95%, and sampled immediately or after 24 h for expression analysis. For chlorophyll content and leaf damage analysis, tomato and Arabidopsis plants were analyzed after 1 or 4 days’ recovery in the growth room. Plants under identical conditions without exposure to SO2 served as controls. For leaf disc treatment, discs removed from 3- to 4-week-old wild-type and transgenic Arabidopsis plants and 5-week-old wild-type and transgenic tomato plants, 7 and 9 mm in diameter, respectively, were placed in 90 mm diameter plates on a filter paper moistened with 2 ml of 50% MS salt solution (Duchefa Biochemie BV; http://www.duchefa.com) with or without 7 mM Na2SO3, for 24 h under constant illumination (100 μmol m−2 s−1) and then were photographed and analyzed for chlorophyll content.
Determination of sulfate, chlorophyll, leaf damage level and relative leaf area
For sulfate determination, leaves of Arabidopsis and tomato wild-type and SO-modified plants were sampled immediately after exposure to 2 ppm SO2, extracted in double-distilled water (1:3 w/v), heated for 5 min at 95°C (Hansch et al., 2006) and the sulfate levesl determined by an ion exchange chromatography system (DX 600; Dionex; http://www.dionex.com) using an IonPac® column (AS 4A-SC; Dionex) for separation and an electrochemical conductivity detector (ED 50; Dionex) combined with an upstream-inserted micromembrane suppressor (ASRS-Ultra II 4 min; Dionex). The retention times of 3.57 and 3.90 distinguished sulfite from sulfate, respectively. Plant sulfite levels were below the system’s detection limit (<1 ppm).
For statistical analysis, each treatment was compared to its own control using a two-tailed t test. Total chlorophyll content was measured in extracts of the fully expanded leaves as described previously (Graan and Ort, 1984). The severity scale for leaf damage was as follows: 1, no damage; 2, <30%; 3, 30–50%; 4, >50% of the leaf area damaged. The average leaf damage was then multiplied by the total No. of damaged leaves to determine the damage level. Relative leaf area was calculated as the ratio of leaf area of the treated plants (leaf length multiplied by leaf width) divided by the leaf area of untreated plants and multiplied by 100. Means ± SEM for each treatment are presented.
Protein extraction, immunoblot and in-gel SO activity
For protein extraction, leaves of tomato and Arabidopsis plants were ground using a pestle and mortar in extraction buffer (4 ml g−1 fresh weight) containing 0.25 m sucrose, 50 mm Tris–HCl (pH 8.5), 3 mm EDTA, 1 mm sodium molybdate and a cocktail of protease inhibitors including aprotinin (10 μg ml−1), leupeptin (10 μg ml−1) and pepstatin (10 μg ml−1). The homogenate was centrifuged at 4000 g for 5 min. The resulting supernatant was subjected to centrifugation at 18 000 g for 20 min, and the pellet was dissolved in the above extraction buffer supplemented with Triton X-100 to a final concentration of 0.025%. The total soluble protein content was determined according to the Bradford (1976). For direct measurement of SO protein, proteins were separated by SDS–PAGE carried out in 10% polyacrylamide gels and blotted to polyvinylidene difluoride membranes (Immun-Blot membranes, Bio-Rad). Blotted proteins were subjected to immuno-detection with antibodies raised against an SO synthetic polypeptide, RHPSLKINAKEPFNAE. Primary antibodies were diluted 500-fold in TBS, and secondary antibodies (anti-guinea pig IgG, Sigma; http://www.sigmaaldrich.com/) were diluted 1000-fold in TBS. Phosphatase activity was visualized by staining with 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium (NBT). In-gel assay of SO activity, following H2O2 production, was examined after protein refractionation with by native PAGE. A modified chromogenic horseradish peroxidase (HRP) assay was employed in which H2O2 serves as a proton-accepting substrate while o-dianisidine serves as a proton donor (Manchenko, 1994; Yesbergenova et al., 2005). The modified reaction mixture contained 2.5 mM o-dianisidine, 4.5 U ml−1 HRP and 0.4 mm sodium sulfite. The reaction was stopped by immersion of the gels in double-distilled water. For verification, the detected activity bands were excised and subsequently refractionated by denaturating SDS–PAGE and immunoblotted with SO-specific antisera. The bands detected after Western blot and in-gel assay were scanned using an Arcus 1200 scanner (Agfa; http://www.agfa.com) and quantified by NIH Image software (version 1.6).
Kinetic assays of SO and ROS-generating activities
H2O2-generating activities in leaf extracts of wild-type and SO-modified plants were detected in reaction mixtures containing 10 μg soluble protein, 0.85 mm 4-aminoantipyrine, 3.4 mm 3,5 dichloro-2-hydroxobenzene sulfonate and 4.5 U ml−1 HRP in 1 ml of 50 mm phosphate buffer (pH 7.5). The colorimetric assay is based on 3,5 dichloro-2-hydroxobenzene sulfonate, which couples oxidatively to 4-aminoantipyrine in the presence of H2O2 and HRP to yield a red quinonemine dye (Fossati et al., 1980; Yesbergenova et al., 2005). The H2O2-generating activity was assayed spectrophotometrically at 515 nm after the addition of 0.4 mm sodium sulfite. SO activity was determined by the reduction of ferricyanide at 420 nm, in reaction mixtures containing 10 μg soluble protein, 0.395 mm ferricyanide, 0.4 mm sodium sulfite in 1 ml of 20 mm Tris–HCl buffer (pH 8). One unit of SO activity was defined as the conversion of 1 μmol of sulfite min−1 (Eilers et al., 2001). For both assays, reaction mixtures without sodium sulfite served as controls.
Quantitative real-time RT-PCR
For each RT reaction, 0.5 μg of Arabidopsis or tomato total RNA was reverse-transcribed with an iScript™ cDNA Synthesis Kit (Bio-Rad; http://www.bio-rad.com) using modified MMLV-derived reverse transcriptase (Bio-Rad), and a blend of oligo(dT) and random hexamer primers according to the manufacturer’s instructions. Quantitative RT-PCR reactions contained 1:13 v/v first-strand cDNA as the template, specific primers (see Table S1 for Arabidopsis and Table S2 for tomato) and iQ™ SYBR® Green Master Super Mix (Bio-Rad) in a final volume of 15 μl. Amplification was performed for 40 cycles, consisting of initial pre-heating at 95°C for 3 min, 20 sec at 95°C, 20 sec at 65°C and 30 sec at 72°C. Fluorescence increments of each reaction were simultaneously monitored using the iCycler iQ Multicolor real-time PCR detection system (Bio-Rad). The PCR products were separated on a 2% agarose gel, excised from the gel and sequenced for verification. Reactions normalized with ACTIN 2 (At3g18780) and elongation factor 1-α (At5g60390) for Arabidopsis or ACTIN Tom41 (U60480) and elongation factor 1-α (SGN-U196120) for tomato as housekeeping genes revealed similar results. The expression in each treated line was compared to that in the untreated line after normalization to ACTIN 2 or ACTIN Tom41 and presented as relative expression (means ± SEM, n = 3).
Sequence analysis was performed using the ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit on an ABI Prism 310 cycle sequencer (PE Applied Biosystems; http://www.appliedbiosystems.com).
M.S. and R.F. gratefully acknowledge a grant from the Israel Science Foundation (grant 417/03) in partial coverage of the costs. Our gratitude to Dr Yvonne Ventura and Ms Evelyn Farfan for assistance with SO activity, sulfate and chlorophyll assays is acknowledged.
Accession number: The GenBank accession number for tomato sulfite oxidase is DQ853413.