Glucosinolates are plant secondary metabolites involved in responses to biotic stress. The final step of their synthesis is the transfer of a sulfo group from 3′-phosphoadenosine 5′-phosphosulfate (PAPS) onto a desulfo precursor. Thus, glucosinolate synthesis is linked to sulfate assimilation. The sulfate donor for this reaction is synthesized from sulfate in two steps catalyzed by ATP sulfurylase (ATPS) and adenosine 5′-phosphosulfate kinase (APK). Here we demonstrate that R2R3-MYB transcription factors, which are known to regulate both aliphatic and indolic glucosinolate biosynthesis in Arabidopsis thaliana, also control genes of primary sulfate metabolism. Using trans-activation assays we found that two isoforms of APK, APK1, and APK2, are regulated by both classes of glucosinolate MYB transcription factors; whereas two ATPS genes, ATPS1 and ATPS3, are differentially regulated by these two groups of MYB factors. In addition, we show that the adenosine 5′-phosphosulfate reductases APR1, APR2, and APR3, which participate in primary sulfate reduction, are also activated by the MYB factors. These observations were confirmed by analysis of transgenic lines with modulated expression levels of the glucosinolate MYB factors. The changes in transcript levels also affected enzyme activities, the thiol content and the sulfate reduction rate in some of the transgenic plants. Altogether the data revealed that the MYB transcription factors regulate genes of primary sulfate metabolism and that the genes involved in the synthesis of activated sulfate are part of the glucosinolate biosynthesis network.
Plants produce a great variety of natural compounds with various functions. One of the best-studied groups of secondary metabolites is the glucosinolates (GSs). Glucosinolates are a large group of sulfur-rich amino acid-derived metabolites, found in the Brassicaceae (Fahey et al., 2001; Halkier and Gershenzon, 2006). Glucosinolates play an important role in plant defense against herbivores and insects (Halkier and Gershenzon, 2006). Upon tissue damage, the vacuole-stored GSs come into contact with the enzyme myrosinase, which catalyzes their degradation into the active deterrents, volatile isothiocyanates or nitriles (Matile, 1980; Rask et al., 2000). The reaction with myrosinase, and thus the biological activity of GSs, depends on the sulfate group (Ratzka et al., 2002). Glucosinolates are also important nutritionally; they are responsible for the smell and taste of cruciferous vegetables (Fenwick et al., 1983). In addition, their breakdown products have been shown to have anticarcinogenic activity (Verhoeven et al., 1997; Mithen et al., 2003). They can, however, have a negative impact when present in large amounts in animal feed derived from rapeseed (Schöne et al., 1997).
The biosynthesis of GS is closely connected to primary sulfur metabolism. The final step of GS core synthesis is the sulfation of the desulfo-GS precursors (Figure 1; Underhill et al., 1973). The sulfation is catalyzed by sulfotransferases (SOTs) which transfer activated sulfate from 3′-phosphoadenosine 5′-phosphosulfate (PAPS) to a hydroxylated substrate. The PAPS is synthesized from sulfate in two ATP-dependent steps (Figure 1). Sulfate is first adenylated by ATP sulfurylase (ATPS) to adenosine 5′-phosphosulfate (APS), which is then further phosphorylated by APS kinase (APK) to PAPS. The importance of PAPS for GS synthesis was recently demonstrated in Arabidopsis by the finding of very low levels of GS in plants disrupted in two of the four APK genes (Mugford et al., 2009). However, APS is also an intermediate in primary sulfate assimilation, where it is reduced by APS reductase (APR) to sulfite, and after further reduction to sulfide it is incorporated into O-acetylserine (OAS) to form cysteine (Figure 1; for recent review see Kopriva, 2006). Genes involved in PAPS synthesis as well as in APS reduction were reported to be affected by mis-expression of the GS MYB factors based on microarray analyses (Sønderby et al., 2007; Malitsky et al., 2008).
We here demonstrate that at least some of the ATPS, APK, and APR genes are directly regulated by the R2R3-MYB transcription factors involved in the regulation of core GS biosynthesis. Using trans-activation assays and analysis of plants disrupted in or overexpressing these MYB factors, evidence is presented that several genes involved in primary sulfur metabolism are members of the glucosinolate biosynthesis network.
Regulation of ATPS and APK by MYB transcription factors
To test the hypothesis that genes involved in the provision of activated sulfate for GS synthesis are under the same control as those involved in core GS biosynthesis, we used the co-transformation assay developed to identify regulatory genes in the GS synthesis network (Berger et al., 2007). Since disruption of APK1 and APK2 isoforms of APS kinase reduced foliar GS levels by 90% (Mugford et al., 2009), the APK genes were the first to be tested. Reporter constructs consisting of promoters of the four APK genes fused to the uidA (GUS) reporter gene were simultaneously expressed with constructs overexpressing the aliphatic GS controlling MYB28, MYB76, and MYB29 (Gigolashvili et al., 2007a) and MYB51, MYB122, and MYB34, regulating indolic GSs, in cultured Arabidopsis thaliana cells. It was revealed by GUS activity staining that all six effectors strongly activated expression from APK1 and APK2 promoters (Figure 2a). On the other hand, APK3-derived expression seemed to be only weakly activated, while the APK4 promoter did not trigger any GUS activity. No obvious differences between the trans-activation potentials of the two MYB factor groups on APK were observed.
To test whether ATPS, the other enzyme necessary for PAPS synthesis, is also directly linked to the regulation of GSs, we isolated promoters of the four A. thaliana ATPS genes and used them in the same trans-activation assays as APK. The GUS-dependent blue staining was observed only in cells expressing the reporter from ATPS1 and ATPS3 promoters (Figure 2b). All six MYB factors were capable of activating the promoters; however, the intensity of the staining was not equal. The factors associated with aliphatic GSs, MYB28, MYB76, and MYB29, appeared to induce a stronger reaction with the ATPS1 promoter than with ATPS3, while the opposite was true for all three indolic GS transcription factors (Figure 2b). Thus, ATPS1 seems to be strongly associated with the control of synthesis of aliphatic GSs, whereas ATPS3 may be linked to regulation of indolic GSs.
Overexpression of MYB factors affects mRNA levels of APK and ATPS
The trans-activation assays revealed that the six MYB transcription factors are capable of interacting with the promoters of APK1, APK2, ATPS1, and ATPS3, and partly with APK3. To test whether this is also the case in vivo, we compared by quantitative RT-PCR (qPCR) the steady-state mRNA levels of the APK and ATPS genes in leaves of plants constitutively overexpressing the MYB factors and plants in which single MYB factor genes were disrupted (Gigolashvili et al., 2007a,b, 2008). The 35S promoter-driven expression of the MYB factors leads to 17-, 110-, and 7-fold increases in mRNA encoding MYB28, MYB76, and MYB29, respectively, compared with the wild type (Col-0), and to 40-, 8-, and 26-fold increases in transcripts for MYB51, MYB122, and MYB34. The transcript levels of APK1 and -2 isoforms were greatly increased in the leaves of all overexpressing plants compared with the wild type, except APK1 in MYB76_ox plants (Figure 3a,b). The mRNA levels of APK3 were significantly higher only in the transgenic plants expressing MYB28 and MYB51, the major regulators of aliphatic and indolic GS biosynthesis, respectively, but the increase was 3.5- and 2.5-fold, respectively, not as dramatic as with APK1-induced 16- and 9-fold increases in these plants, for example. In contrast, APK4 expression was not significantly increased (Figure 3a,b). To test whether any of the MYB factors is essential for the expression of APK genes we also analyzed steady-state mRNA levels in mutants of the individual MYB genes. No disruption in any of the MYB genes resulted in a reduction in APK transcript levels; on contrary, APK3 was induced in myb34 mutants (Figure 3c,d).
The results of qPCR for members of the ATPS gene family also agreed well with the results of the trans-activation assays (Figure 4). Overexpression of all six MYB factors resulted in the accumulation of ATPS1 and ATPS3 transcripts compared with wild-type plants, while the mRNA levels for ATPS2 and ATPS4 were not affected (Figure 4a,b). Corresponding to the results of the trans-activation assay, ATPS1 transcript was significantly more induced in MYB28_ox, MYB76_ox, and MYB29_ox plants than in those overexpressing the indolic GS-associated factors MYB51, MYB122, and MYB34. On the contrary, ATPS3 mRNA was increased 8–12-fold by overexpression of the latter group but only 5–7-fold in the plants overexpressing aliphatic GS-associated factors. Analysis of the MYB factor mutants again did not reveal any large effects on ATPS transcript levels, except induction of ATPS4, which was apparent in all six mutants but significant only in myb122 and myb34 (Figure 4c,d). These results thus reveal that the genes involved in the provision of PAPS are part of the GS biosynthesis network controlled by the MYB factors.
Regulation of primary sulfate assimilation by MYB factors
Because previous microarray analyses predicted that genes involved in primary sulfate reduction may also be regulated by the MYB factors (Sønderby et al., 2007; Malitsky et al., 2008) we extended the analysis to the APR gene family, encoding the key enzyme of the pathway (Vauclare et al., 2002; Loudet et al., 2007). The trans-activation assays revealed that all six MYB factors activated expression of GUS directed from promoters of all three APR genes (Figure 5). APR1 was strongly activated by all MYB factors, while APR2 seemed to be activated by MYB122 and MYB76 more strongly than by the others. The qPCR analysis of MYB-overexpressing plants confirmed that APR1 and APR3 transcripts accumulated to a greater extent in all six transgenic lines analyzed (Figure 6a,b). Also in agreement with the results of the trans-activation assay, APR2 was elevated in MYB122_ox, MYB28_ox, and MYB76_ox plants. In the sulfate assimilation pathway, the sulfite produced by APR is further reduced to sulfide by sulfite reductase (Figure 1). The mRNA for the single-copy sulfite reductase (SiR) gene was significantly increased in plants overexpressing the indolic GS regulators MYB51, MYB122, and MYB34, as well as the aliphatic GS regulator MYB28 (Figure 6a,b). Steady-state levels of APR and SiR mRNA were not different in lines with disrupted genes encoding the MYB factors regulating aliphatic GSs, myb28_RNAi, myb76, and myb29, and wild-type plants. On the other hand, disruption of the expression of the three MYB factors of the indolic GS clade resulted in a strong elevation of APR2 mRNA, while transcripts of APR1 and APR3 were increased in myb51 and myb122, respectively (Figure 6c,d).
Adenosine 5′-phosphosulfate reductase (APR) and other components of the sulfate assimilation pathway are also regulated at the post-transcriptional level (Koprivova et al., 2008; Kawashima et al., 2009). To test whether the observed changes in transcript levels are really reflected at the activity level, we measured the activities of the enzymes ATPS and APR. Both enzyme activities were significantly elevated in MYB51_ox and myb51 plants compared with the wild type (Figure 7a,b). In addition, APR activity was reduced in plants with reduced expression of MYB28 (Figure 7b). The foliar concentration of glutathione, the major product of sulfate assimilation, was reduced in MYB28_ox plants and increased in myb51 compared with wild-type plants (Figure 7d). Figure 7(e) shows the changes in GS levels caused by the expression of, or mutation in, the two MYB factors. In addition, we have determined whether overexpression and disruption of MYB51 affects the flux through the sulfate assimilation pathway, to assess the biological significance of increased transcript levels and enzyme activities in these plants. Indeed, in both MYB51_ox and myb51 plants the incorporation of 35S from [35S]sulfate to thiols and proteins is significantly increased (Figure 8). Thus, modulation of expression of GS-controlling MYB factors does indeed affect primary sulfur metabolism.
Although the significance of the sulfate group for GS function has been recognized (Ratzka et al., 2002), apart from simple plant nutrition studies (Falk et al., 2007; Schonhof et al., 2007) little is known about the interconnection of GS synthesis and sulfate assimilation. Glucosinolate synthesis was shown to be coordinately repressed by sulfur starvation, dependent on the SLIM1 factor, which on the other hand activates sulfate uptake and assimilation in these conditions (Hirai et al., 2005; Maruyama-Nakashita et al., 2006). Microarray analyses indicated that at least some genes involved in sulfate assimilation and PAPS synthesis might be regulated by MYB factors controlling core GS synthesis (Sønderby et al., 2007; Malitsky et al., 2008); however, no direct evidence has been presented so far.
Very recently, a reduced capacity to synthesize PAPS was shown to significantly affect the accumulation of GS and its biosynthetic pathway (Mugford et al., 2009). Simultaneous disruption of two of the four APK isoforms in A. thaliana, APK1 and APK2, resulted in an approximately 90% decrease in GS levels in the leaves and seeds. Both aliphatic and indolic GSs were affected (Mugford et al., 2009). In the leaves of apk1 apk2 plants, the desulfo-GS precursors accumulated to very high levels and transcripts of most genes involved in GS synthesis were induced compared with wild-type plants. Based on these data it was hypothesized that the different APK isoforms provide PAPS for different sets of sulfotransferases in synthesis of different sulfated metabolites. This would imply that the different genes would be under different regulatory circuits, with APK1 and APK2 most probably co-expressed and co-regulated with the GS synthesis genes. Indeed, two independent web-based tools, ‘Expression angler’ from the Bio-Array Resource for Arabidopsis Functional Genomics (BAR) (Toufighi et al., 2005) and the ‘Drawing gene networks and searching co-expressed genes’ tool at the ATTED-II database (http://atted.jp/top_tool.shtml) (Obayashi et al., 2009) indicated that expression of APK1 and APK2 is co-regulated with the GS biosynthetic genes, while expression of APK3 and APK4 is not (data not shown). However, these were only in silico data, and the hypothesis that APK1 and APK2 are regulated by transcription factors controlling GS synthesis had to be tested by a direct experimental approach. Therefore, we turned to the trans-activation assay (Berger et al., 2007), which was employed previously to identify the targets of R2R3-MYB factors (Gigolashvili et al., 2007a,b, 2008). The analysis indicated that APK1 and APK2 are regulated by all six MYB factors, while the trans-activation of APK3 was much weaker and APK4 did not seem to be activated by the MYB effectors (Figure 2). These observations were confirmed by measuring the levels of APK transcripts in leaves of plants overexpressing the MYB factors (Figure 3). Also the microarray analysis of MYB factors overexpressing plants revealed increased mRNA levels of APK1 and APK2 but not the other APK isoforms (Malitsky et al., 2008). This is especially important, since in these transgenic plants the MYB factors were not controlled by the constitutive 35S promoter as in this study but by a late 650 promoter (Malitsky et al., 2008). Thus, pleiotropic effects due to high transcript and protein accumulation of the trans factors expressed from the 35S promoter can probably be excluded. We therefore conclude that APK1 and APK2 are part of the GS synthesis network controlled by the MYB factors (Figure 9). APK3 and APK4 are capable of providing at least some PAPS for GS synthesis, as apk1 apk2 plants still contained about 10% of normal GS levels (Mugford et al., 2009), but seem to be only loosely, if at all, associated with this network.
Since two APK isoforms were indeed co-regulated with GS biosynthetic genes, we subjected the other gene family involved in PAPS synthesis, ATPS, to the same analysis. ATPS catalyses the entry step of sulfate into both primary and secondary metabolism. It is regulated in a demand-driven manner; when the demand for reduced sulfur is low, for example due to nitrogen deficiency or in the presence of reduced sulfur compounds, the enzyme activity is decreased (Reuveny et al., 1980; Lappartient and Touraine, 1996; Lappartient et al., 1999). When the demand for reduced sulfur is high, such as during sulfur deficiency or abiotic stress (Reuveny et al., 1980; Farago and Brunold, 1990;Heiss et al., 1999; Harada et al., 2000), the activity is increased. The links of ATPS with abiotic and biotic stress (Heiss et al., 1999; Rausch and Wachter, 2005) and the importance of PAPS for GS synthesis (Mugford et al., 2009) indicated that at least some ATPS isoforms might also be co-regulated with genes controlling GS biosynthesis. This has indeed been the case for ATPS1 and ATPS3 in the in silico co-transcriptional analyses, similar to APK1 and APK2 (data not shown). Similar to APK, both direct approaches – the trans-activation assays and analysis of MYB factor-overexpressing plants – revealed that two ATPS isoforms, ATPS1 and ATPS3, are strongly regulated by the MYB effectors while the other two are not. However, while the trans-activation assays suggested that ATPS1 is much more strongly activated by the aliphatic GS-controlling factors than by the indolic GS ones, with the opposite being true for ATPS3, the differences in the transgenic plants were less marked (Figures 2 and 4). Possibly, other unknown factors modulate the binding of the MYB factors to ATPS promoters in the plant so that they are not as discriminating as in the trans-activation assay. Again, the data correspond well with the microarray analysis by Malitsky et al. (2008).
Mugford et al. (2009) showed that limitation of PAPS synthesis and GS content result in an increased synthesis of thiols through the reductive part of the sulfate assimilation pathway. In addition, genes involved in primary sulfate reduction were found by microarray analyses to be induced in plants overexpressing several of the studied MYB factors (Sønderby et al., 2007; Malitsky et al., 2008). Since APR has a very high control over the pathway (Vauclare et al., 2002; Loudet et al., 2007), we tested the possibility that the APR genes are also under the control of the MYB factors (Figures 5, 6). Indeed, the trans-activation assays clearly indicate the potential of the MYB effectors to regulate APR. The results of the trans-activation assays again agreed well with qPCR data for the MYB-overexpressing plants, showing that APR1 and APR3 are under the control of all the MYB factors while APR2 is regulated only by some. Sulfite reductase was also induced in MYB-overexpressing plants; the indolic GS controlling factors had a higher effect than the aliphatic ones (Figure 6). Thus, the expression of genes involved in sulfate assimilation is affected by the GSs controlling MYB factors.
To complement knowledge about the regulation of sulfate assimilation by the GS-associated MYB factors we also compared the expression of these genes in plants in which the expression of single MYB factors was disrupted by T-DNA insertions or RNAi. Such disruption had very little effect on mRNA levels of APK, ATPS, and SiR genes, especially on the isoforms shown to be highly activated by the overexpressed factors. There are several explanations for this phenomenon. Firstly, the genes are regulated by multiple MYB factors (Figures 2–4), therefore disruption of a single MYB gene does not have a large effect on the accumulation of mRNA of the target gene. This, however, does not seem to be the case since, for example, the genes involved in synthesis of aliphatic GSs are activated by all three MYB factors – MYB28, MYB76, and MYB29 – but their transcripts are still significantly reduced in MYB28_RNAi plants (Gigolashvili et al., 2007b, 2008). Secondly, since ATPS and APK have other functions apart from GS biosynthesis, their expression is likely to be controlled by multiple trans factors. The GS-associated MYB effectors might thus be important to rapidly induce the transcription of these genes when demand for GSs increases, but not to keep their basal mRNA levels. Interestingly, APR was much more strongly affected by the MYB disruptions. Whereas the APR transcript levels were not affected by disruption of the aliphatic GS MYBs, mutants in the indolic group showed highly increased levels of APR2 and in some cases also of APR1 and APR3.
How can disruption and overexpression of the same transcription factor result in the same response of the target gene? The key enzyme in the control of sulfate assimilation is APR (Vauclare et al., 2002), and the APR2 isoform in Arabidopsis contributes most to the total APR activity (Loudet et al., 2007). Overexpression of the MYB factors leads to an increased demand for reduced sulfur to incorporate in the thioglucoside bond and for the Met-derived side chains of aliphatic GS, which consequently induces APR. In the MYB_ox plants, increased GS production of approximately 10 nmol (mg dry weight)−1 compared with the wild type (Figure 7) requires 17 nmol (mg dry weight)−1 of reduced S (10 nmol for the thioglycoside bond and 7 nmol for the approximately 70% Met-derived GSs). This corresponds to a S demand of approximately 1.7 μmol S (g fresh weight)−1 that is about sixfold higher than the amount of sulfur in the GSH pool. To sustain such an increased rate of sulfate reduction the capacity of the assimilatory pathway has to be increased, which seems to be achieved by co-regulation of the key genes, such as APR and SiR, with GS synthesis via the MYB factors. Disruption of the MYB factors on the other hand results in a reduction of GS accumulation (Gigolashvili et al., 2007a,b, 2008; Hirai et al., 2007; Sønderby et al., 2007). The indolic GSs are very important for the protection of plants against insects, but they also have a broad antifungal and antibacterial activity in the pen2 pathway (Kim and Jander, 2007; Schlaeppi et al., 2008; Bednarek et al., 2009; Clay et al., 2009). Therefore, when synthesis of the indolic GSs is compromised, e.g. in myb51 plants, reduced sulfur compounds may substitute for GSs in defense against pests (Rausch and Wachter, 2005). Indeed, reduced synthesis of GSs in apk1 apk2 plants resulted in increased thiol synthesis (Mugford et al., 2009). Therefore, the increase in APR2 mRNA may reflect the increased requirement for sulfate reduction to synthesize the alternative defense compounds, which are especially needed in the absence of indolic GSs. Indeed, GSH accumulates in the myb51 plants, similar to apk1 apk2 plants. The regulatory circuit to increase levels of APR transcript in these conditions, however, can be completely different from the two groups of R2R3-MYB factors and thus unaffected by the disruption of MYB51, MYB34, or MY122 genes. APR is often regulated differently from other genes in the pathway, e.g. in the sulfur deficiency response it is not part of the SLIM1 regulatory network (Maruyama-Nakashita et al., 2006). Indeed, APR is upregulated by many hormones involved in stress signaling, including jasmonate, salicylate, ethylene, and nitric oxide (Koprivova et al., 2008). Signaling by these compounds may be triggered by the low levels of indolic GS in myb51 plants, in a similar way that the synthesis of GSs was induced in apk1 apk2 plants due to low levels of GSs.
Importantly, the changes in transcript levels of the sulfate assimilation genes were at least also partly translated to changes in enzyme activities and thiol levels (Figure 7). Both overexpression and disruption of MYB51 had the same effect on ATPS and APR activity. For both enzymes, mRNA of at least one isoform was increased in the corresponding plant material. Correspondingly, the increased activities indeed resulted in higher rate of sulfate reduction in both genotypes. However, in MYB51_ox plants the increased flux through sulfate assimilation did not result in higher GSH levels. The extra cysteine and GSH synthesized (Figure 8) were most probably immediately used as a sulfur source for the increased GS synthesis in these plants and not for increasing the size of the GSH pool. On the other hand, in myb51 plants the increase in flux also resulted in increased glutathione levels, as the extra reduced sulfur was not used for GSs. Increase in GSH correlating with decrease in GS levels was observed previously in apk1 apk2 plants (Mugford et al., 2009). The effects of mis-expression of MYB28 on APR and ATPS activity are not so easy to explain. Although overexpression of this trans factor increased transcript levels for ATPS1, ATPS3, and the three APR genes, the enzyme activities were not affected. Thus, it appears that a post-transcriptional and/or post-translational regulation prevented the increase in the enzyme activities. Similar phenomena were observed in the regulation of APR by salt, where in several signaling mutants APR transcript levels were increased but the activity was not (Koprivova et al., 2008). It is also possible that the enzymes were directly inhibited by intermediates in aliphatic GS synthesis, which during methionine elongation reactions increases the concentration of reduced sulfur compounds in the plastids. Indeed, APR is known to be redox regulated in a post-translational manner, to which these intermediates may contribute (Bick et al., 2001). Because APR activity was not induced, the extra S used for increased GS synthesis depleted the GSH pool and the GSH content decreased in MYB28_ox plants (Figure 7). Disruption of MYB28 had very little effect on primary sulfate assimilation despite a dramatic reduction of GS content in the MYB28_RNAi plants. It is feasible to argue that the aliphatic GSs are less important for plant defense than indolic GSs, and therefore the low rate of their synthesis does not induce the synthesis of alternative compounds such as GSH. Thus, there is no need for an increase in flux through sulfate assimilation and APR and ATPS transcripts are not affected. The reduction in APR activity in MYB28_RNAi compared with Col-0 might actually be caused by a very high activity measured in these particular control plants compared with other Col-0 material. Regulation of APR takes place on many levels and is known to respond strongly to small alterations in growth conditions (Kopriva, 2006).
Our analysis of the control of sulfate assimilation by the GS-biosynthesis-regulating MYB factors thus resulted in two significant findings. We could successfully place four new enzymes, namely APK, ATPS, APR, and SiR, into the GS biosynthesis network (Figure 9). At least the synthesis of activated sulfate, the first step of which is simultaneously a part of primary sulfate assimilation, has to be considered as an integral part of the GS network. In addition, we have significantly improved our knowledge of molecular mechanisms of regulation of primary sulfate reduction. Six novel transcription factors have been revealed to regulate this pathway. While gel shift assays or chromatin immunoprecipitation are necessary to unequivocally prove direct binding of the effectors to the corresponding promoters, the effects of the MYB factors on ATPS and APR enzyme activities as well as the stimulation of flux through the pathway corroborate the results of the trans-activation assays and expression analysis. Despite a sound understanding of the regulation of plant sulfate assimilation at the physiological level and responses of enzyme activities and gene expression to various stimuli, only one transcription factor, SLIM1, has been identified to directly control the pathway (Maruyama-Nakashita et al., 2006). The identification of MYB factors controlling GS biosynthesis also as regulators of ATPS and APR thus shows the complexity of regulatory interactions between primary and secondary sulfur metabolism.
Plant material and growth conditions
The construction and selection of plants overexpressing the MYB factor was described previously (Gigolashvili et al., 2007a,b, 2008). For the expression analysis the plants were grown in soil in short-day conditions (8-h light, 16-h dark) at 22–25°C and 40% humidity. Overexpression mutants used for the qPCR and biochemical analysis had only moderately increased levels of corresponding transgenes. Gain-of-function mutants of indolic GS regulators used for the analysis were HIG1-1D, Pro35S:MYB122-11, and Pro35S:ATR1-17 lines (called here MYB51_ox, MYB122_ox, and MYB34_ox), whereas gain-of-function mutants of aliphatic GS regulators were Pro35S:MYB28-15, Pro35S:MYB76-23, and Pro35S:MYB29-6 (MYB28_ox, MYB76_ox, and MYB29_ox). Loss-of-function mutants used in this work were MYB28-RNAi-10 (accumulating approximately 20% of WT transcript levels), myb76 (hag2), myb29 (hag3), and myb51 (hig1-1) described previously (Gigolashvili et al., 2007a,b, 2008), and myb122 and myb34 knockout plants (SALK collection) isolated recently (T. Gigolashvili, unpublished data).
Arabidopsis Col-0 suspension culture for the trans-activation assays was grown in 50 ml of A. thaliana (AT) medium [4.3 g L−1 MS basal salt medium (Duchefa, http://www.duchefa.com/), 1 mg L−1 2,4-dichlorophenoxyacetic acid (2,4-D), 4 ml of a vitamin B5 mixture (Sigma, http://www.sigmaaldrich.com), and 30 g L−1 sucrose pH 5.8]. Suspension cell culture was diluted weekly to 1:4 or 1:5 with fresh AT medium and gently agitated at 150 rpm in the dark at 22°C.
In order to find which genes are co-expressed with the different ATPS and APK isoforms the web-based data-mining tool ‘Expression angler’ from BAR was utilized (Toufighi et al., 2005). Existing microarray data were compared with a chosen query gene, and Pearson correlation coefficients were calculated to identify genes with similar expression and response patterns. Parameters were set to return the 25 best hits from the AtGenExpress global stress expression data set (Kilian et al., 2007) for APK and from the complete NASC data set for ATPS.
The co-expression analysis based on the publicly available 1388 microarray data of AtGenExpress using Pearson’s correlation coefficients between all combinations of 22 263 Arabidopsis genes was performed using the ‘Drawing gene networks and searching co-expressed genes’ tool at the ATTED-II database (http://atted.jp/top_tool.shtml) (Obayashi et al., 2009). The gene set used to build the network was based on ‘Glucosinolate metabolism’ in the Kazusa Plant Pathway Viewer (http://kpv.kazusa.or.jp/kappa-view/) (Tokimatsu et al., 2005) and the ATPS and APK gene families.
Cloning of promoters of ATPS, APK, and APR genes and trans-activation assays
To generate reporter constructs, the promoter regions of the ATPS [5 kbp (ATPS1 and ATPS4) and 2.5 kbp (ATPS2 and ATPS3) upstream of the ATG], APK [3 kbp (APK1 and APK3) and 3.5 kbp (APK2 and APK4) upstream of the ATG], and APR [3 kbp upstream of the ATG] genes were amplified from genomic DNA of A. thaliana with specific primers (Table S1 in Supporting Information) and cloned into the Gateway entry vector (Invitrogen, http://www.invitrogen.com). The promoter sequences were then subcloned into the binary plant transformation vector pGWB3i, resulting in translational fusions with the gus reporter gene. As effectors, the constructs Pro35S:MYB28, Pro35S:MYB76, Pro35S:MYB29, Pro35S:MYB51, Pro35S:MYB122, and Pro35S:MYB34 were used as described previously (Gigolashvili et al., 2007a,b, 2008). The reporter and effector constructs were used to transform the supervirulent Agrobacterium strain LBA4404.pBBR1MCS.virGN54D (kindly provided by Dr J. Memelink, University of Leiden, Netherlands). For transient expression assays in the cell culture, agrobacteria containing the effector constructs, the anti-silencing 19-K protein, or one of the reporter constructs were taken from fresh yeast extract broth (YEB) plates, grown overnight, and resuspended in 1 ml AT medium. The agrobacteria were mixed in a 1:1:1 ratio, and 75 μl of this suspension was added to 3 ml cultured A. thaliana root cells, which were then grown for 3–5 or 7 days in the dark and subsequently used for GUS activity measurements or staining (Berger et al., 2007).
RNA extraction and expression analysis
The expression of ATPS, APK, APR, and SiR genes was analysed by real-time quantitative RT-PCR using the fluorescent intercalating dye SYBR Green in a GeneAmp 5700 sequence detection system (Applied Biosystems, http://www3.appliedbiosystems.com/). The Arabidopsis ACTIN2 gene was used as a standard. First, total RNA was isolated from 5-week-old rosette leaves using TRIsure (Bioline, http://www.bioline.com/) and reverse-transcribed into cDNA, using the FirstStrand cDNA Synthesis SSII kit (Bioline) according to the manufacturer’s instructions. Subsequently, the cDNA was used as a template in real-time PCR experiments with gene-specific primers (for primer sequences see Table S2). Real-time PCR was performed using the SYBR Green master mix system (Applied Biosystems) according to the manufacturer’s instructions. The Ct, defined as the PCR cycle at which a statistically significant increase of reporter fluorescence is detected, was used as a measure of the transcript level of the target gene. Relative quantification of the expression levels was performed using the comparative Ct method (manufacturer’s instructions, bulletin 2, Applied Biosystems). Three independent RNA preparations from independently grown plants were analyzed with two technical replicates for the qPCR.
The activity of APS reductase was determined as the production of [35S]sulfite, assayed as acid volatile radioactivity formed in the presence of [35S]APS and dithioerythritol as reductant (Koprivova et al., 2008). The ATPS was measured as the APS and pyrophosphate-dependent formation of ATP (Cumming et al., 2007). The protein concentrations were determined according to Bradford with bovine serum albumin as a standard.
HPLC analysis of low molecular weight thiols
The analysis of cysteine and GSH was performed as described (Koprivova et al., 2008). Twenty to thirty milligrams of leaf material was ground in liquid nitrogen and extracted in a 10-fold volume of 0.1 m HCl. After centrifugation at 15 000 g, 25 μl of supernatant was neutralized by 25 μl of 0.1 m NaOH and 1 μl of 100 mm dithiothreitol was added to reduce disulfides. After 15 min at 37°C, 35 μl water, 10 μl 1 m 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS) pH 8.0, and 5 μl of 100 mm monobromobimane (Thiolyte® MB; Calbiochem, http://www.merck-chemicals.co.uk) were added and derivatization of thiols was allowed to proceed for 15 min at 37°C in the dark. The reaction was stopped and the conjugates were stabilized by the addition of 100 μl of 9% acetic acid. Bimane conjugates were separated by HPLC (Spherisorb® ODS2; 250 × 4.6 mm, 5 μm; Waters, http://www.waters.com/) using 10% (v/v) methanol, 0.25% (v/v) acetic acid (pH 9.3) as solvent A and 90% (v/v) methanol, 0.25% (v/v) acetic acid (pH 9.3) as solvent B. The elution protocol employed a linear gradient from 96 to 82% A in B within 20 min, and the flow rate was kept constant at 1 ml min−1. Bimane derivates were detected fluorimetrically (474 detector, Waters) with excitation at 390 nm and emission at 480 nm.
Determination of flux through sulfate assimilation
The flux through sulfate assimilation was measured as incorporation of 35S from [35S] sulfate to thiols and proteins essentially as described in Kopriva et al. (1999) and Vauclare et al. (2002). Col-0, MYB51_ox, and myb51 plants were grown for 14 days on vertical MS-phytogel plates. The plants were transferred into 24-well plates containing 2 ml of MS nutrient solution adjusted to a sulfate concentration of 0.2 mm and supplemented with 5.6 μCi [35S]sulfate (Hartmann Analytic, http://www.hartmann-analytic.com/) to specific activity of 1860 kBq (nmol sulfate)−1 and incubated in the light for 4 h. After the incubation the seedlings were washed three times with 2 ml of cold non-radioactive nutrient solution, carefully blotted with paper tissue, weighed, transferred into 1.5 ml tubes, and frozen in liquid nitrogen. The plant tissue was extracted 1:10 (w/v) in 0.1 m HCl. Ten microliters of the extract was added to 1 ml of Optiphase HiSafe3 scintillation cocktail (Perkin Elmer, http://www.perkinelmer.com/) and the radioactivity was measured in a scintillation counter (Beckmann, http://www.beckmancoulter.com/) to determine sulfate uptake. To measure 35S incorporation into proteins, these were precipitated from 100 μl of the extract with 25 μl 100% trichloroacetic acid (TCA) as described in Kopriva et al. (1999). After 15 min on ice the precipitate was collected by centrifugation 15 000 g, washed once in 100 μl 1% TCA and once in 200 μl ethanol (EtOH) and dissolved in 100 μl 0.1 m NaOH. The radioactivity was determined after the addition of 1 ml scintillation cocktail in the scintillation counter.
To determine the radioactivity in thiols 100 μl of the extract was mixed with 100 μl 0.1 m NaOH and 2 μl 0.1 m DTT and incubated in the dark at 37°C for 15 min. Afterwards 23 μl of 1 m TRIS pH 8.0 and 10 μl 100 mm monobromobimane were added, mixed, and incubated in the dark at 37°C for 15 min. Then 22.5 μl of 50% acetic acid was added, mixed and centrifuged for 15 min at 15 000 g. Two hundred microliters of the solution was transferred into HPLC vials. Standard thiol analysis was performed as described above with an injection volume of 100 μl. The HPLC was connected to a fraction collector and fractions of 0.8 ml were collected in 6 ml scintillation vials. The radioactivity in these fractions was determined in a scintillation counter after the addition of 2 ml scintillation solution.
This research was supported by the Deutsche Forschungsgemeinschaft and the UK Biotechnology and Biological Sciences Research Council (grant BB/D009596/1). We would like to thank Dr Bok-Rye Lee for critical reading of the manuscript.