AtNUDT7 was reported to be a negative regulator of EDS1-mediated immunity in Arabidopsis. However, the underlying molecular and genetic mechanism of the AtNUDT7-regulated defense pathway remains elusive. Here we report that AtNUDT7 and its closest paralog AtNUDT6 function as novel negative regulators of SNC1, a TIR-NB-LRR-type R gene. SNC1 is upregulated at transcriptional and possibly post-transcriptional levels in nudt6-2 nudt7. The nudt6-2 nudt7 double mutant exhibits autoimmune phenotypes that are modulated by temperature and fully dependent on EDS1. The nudt6-2 nudt7 mutation causes EDS1 nuclear accumulation shortly after the establishment of autoimmunity caused by the temperature shift. We found that a low ammonium/nitrate ratio in growth media leads to a higher level of nitrite-dependent nitric oxide (NO) production in nudt6-2 nudt7, and NO acts in a positive feedback loop with EDS1 to promote the autoimmunity. The low ammonium/nitrate ratio also enhances autoimmunity in snc1-1 and cpr1, two other autoimmune mutants in Arabidopsis. Our study indicates that Arabidopsis senses the ammonium/nitrate ratio as an input signal to determine the amplitude of the EDS1-mediated defense response, probably through the modulation of NO production.
The plant innate immune system is bipartite. One branch of this system is based on the detection of pathogen-associated molecular patterns (PAMPs) by pattern recognition receptors (PRRs) in plants (Boller and Felix, 2009), which elicits the so-called PAMP-triggered immunity (PTI). To dampen the host PTI, pathogens deliver a repertoire of effectors into host cells. However, plants evolved another layer of surveillance system to specifically recognize the presence of effectors through resistance (R) proteins, culminating in a much stronger defense response termed effector-triggered immunity (ETI), often characterized by a hypersensitive response (HR) on the infection site (Chisholm et al., 2006; Jones and Dangl, 2006).
In Arabidopsis, ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1) is essential for basal resistance to some virulent pathogens (Parker et al., 1996; Feys et al., 2001, 2005) and also for ETI mediated by Toll/Interleukin-1 receptor-nucleotide binding-leucine rich repeat (TIR-NB-LRR)-type R proteins (Wiermer et al., 2005). It has long been established that EDS1 functions downstream of activated TIR-NB-LRR receptors as a signal transducer (Zhang et al., 2003; Wiermer et al., 2005; Mestre and Baulcombe, 2006; Wirthmueller et al., 2007). However, two recent studies revealed a direct physical interaction of EDS1 with two effectors (AvrRps4 and HopA1), and also with their corresponding TIR-NB-LRR proteins (RPS4 and RPS6) (Bhattacharjee et al., 2011; Heidrich et al., 2011), suggesting that EDS1 could be the target of some pathogen effectors and might be guarded by the corresponding R proteins (McDowell, 2011).
The onset and amplitude of TIR-NB-LRR-type R gene-mediated resistance needs to be tightly controlled to ensure optimum growth and development. At the transcriptional level, elevated or deregulated expression of some R genes leads to their over-activation and plant autoimmunity (Stokes et al., 2002; Li et al., 2007; Wirthmueller et al., 2007; Yi and Richards, 2009). Some point mutations of R proteins were found to increase their activity, such as snc1-1 (Zhang et al., 2003), chs2 (Huang et al., 2010) and chs3 (Yang et al., 2010). Recently, a group of genes were identified as negative regulators of SNC1, including BON1 (Yang and Hua, 2004; Yang et al., 2006b), BAP1 (Yang et al., 2006a, 2007), BIR1 (Gao et al., 2009; Wang et al., 2011), SRFR1 (Kim et al., 2010; Li et al., 2010), MKP1 (Bartels et al., 2009) and CPR1 (Cheng et al., 2011; Gou et al., 2012). Mutants of these genes share several common features: (i) in the Col-0 background and at 22°C they exhibit constitutive defense responses without pathogen perception, (ii) the constitutively elevated defense response is abolished when plants are grown at 28°C, (iii) the constitutive defense response is partially or fully mediated by SNC1 and EDS1.
AtNUDT7 is also a negative regulator of the EDS1-mediated defense response (Bartsch et al., 2006; Straus et al., 2010). The loss of function of AtNUDT7 leads to enhanced resistance to some virulent and avirulent pathogens, and elevated sensitivity to the oxidative stress imposed by paraquat (Jambunathan and Mahalingam, 2006; Ge et al., 2007). The phenotypes of nudt7 are conditioned by the composition of the soil used to grow the plants (Jambunathan et al., 2010; Straus et al., 2010). AtNUDT7 is a member of the FGFTNE clan of the Nudix hydrolases encoded by the Arabidopsis genome, which also includes AtNUDT6 (Gunawardana et al., 2009). AtNUDT6 and AtNUDT7 exhibit ADP-ribose/NADH pyrophosphohydrolase activity in vitro (Ogawa et al., 2005). Recently, the avirulence effector Avr3b in Phytophthora sojae was reported to be a secreted ADP-ribose/NADH pyrophosphorylase, and functions as a modulator of host immunity (Dong et al., 2011).
In this study, we found that AtNUDT6 shares overlapping functions with AtNUDT7 as negative regulators of R gene-mediated immunity. In addition, we found that the ammonium/nitrate ratio regulates the autoimmunity in nudt6-2 nudt7, and also in cpr1 and snc1-1, probably by modulating the production of NO through nitrate reductase. Our findings suggest that plants sense the ammonium/nitrate ratio as an input signal in modulating the EDS1-mediated immunity.
AtNUDT6 functions redundantly withAtNUDT7 as negative regulators of plant innate immunity
We previously reported that the loss of function of AtNUDT7 leads to hyperactivation of defense-related genes and elevated resistance to both virulent and avirulent strains of Pseudomonas syringae (Ge et al., 2007). AtNUDT6 is the closest paralog of AtNUDT7. An insertion line (SALK_084842) was obtained with a T-DNA at the 5′ untranslated region (5′-UTR) of AtNUDT6 (Figure 1a). This mutant was referred to as nudt6-2, to distinguish it from the allele characterized in a previous study (Ishikawa et al., 2010).
The expression of AtNUDT6 and AtNUDT7 in Col-0, nudt6-2 and nudt7 was measured by real-time quantitative RT-PCR (qRT-PCR). In Col-0, AtNUDT6 was expressed at a very low level in uninfected plants, but was strongly induced (by around 600-fold) 6 h after inoculation with the avirulent P. syringae pv. tomato (Pst) avrRpm1. In nudt6-2, however, the AtNUDT6 transcript level was around 30% of that found in Col-0 mock-treated leaves, and remained unaltered upon pathogen inoculation (Figure 1b). Therefore, nudt6-2 is a knock-down and deregulated line of AtNUDT6. Interestingly, expression of AtNUDT7 was around 50% higher in nudt6-2 than in Col-0 in both mock- and Pst avrRpm1-inoculated leaves. In addition, the expression level of AtNUDT6 in nudt7 was around 20-fold and twofold of that in Col-0 in mock and Pst avrRpm1-treated samples, respectively. These results indicate that disruption of either gene would enhance the expression of the other.
Unlike nudt7, which often displays slight growth retardation under our growth conditions, nudt6-2 was visually indistinguishable from wild-type plants, with no detectable differences in the expression of defense-associated marker genes PR1 or PR2, or resistance to the virulent strain Pst DC3000 or avirulent strain Pst avrRpm1 (Figure 1c–e). To investigate the potential genetic interaction between AtNUDT6 and AtNUDT7, nudt6-2 was crossed with nudt7. In the F2 population, the nudt6-2 nudt7/ + plants had a slightly smaller stature than nudt6. The nudt6-2/ + nudt7 plants showed more severe growth retardation than nudt7 (Figure 1c). Also, the nudt6-2 nudt7/ + and nudt6-2/ + nudt7 plants constitutively accumulated higher levels of PR1 and PR2 transcripts, and showed elevated resistance against Pst, compared with nudt6-2 and nudt7, respectively (Figure 1d,e). The nudt6-2 nudt7 double mutant was extremely stunted, with curly and deformed leaves (Figures 1c, 2a). The double mutant was too small for measuring its pathogen resistance, but qRT-PCR analysis revealed that it constitutively expressed PR1 and PR2 at the highest levels among all genotypes tested (Figure 1d). Consistent with these results, the knock-down of AtNUDT6 by RNAi in the nudt7 background also led to severely retarded growth and over-accumulation of PR transcripts (Appendix S1; Figure S1). Taken together, these data indicate that AtNUDT6 shares an overlapping function with AtNUDT7 as negative regulators of the defense response.
Cell expansion and division in nudt6-2 nudt7 is temperature dependent
The inhibition of the plant defense response by a moderately high temperature was reported in some mutants, such as bon1-1 (Hua et al., 2001; Alcazar and Parker, 2011). Thus we tested whether growing plants at 28°C would alleviate the phenotypes of nudt6-2 nudt7. At 22°C, nudt6-2 nudt7 grew normal cotyledons but later became visually distinguishable from the wild-type plants after the emergence of the first pair of true leaves, which were curly and severely reduced in size (Figure 2a). At this temperature nudt6-2 nudt7 plants never grew to the reproductive stage, and died within the first 6 weeks of germination. But at 28°C, the double mutant fully restored its stature and grew normally throughout its entire life cycle (Figure 2b). A scanning electron microscope (SEM) was used to observe the abaxial epidermal pavement cells of nudt6-2 nudt7 and Col-0. At 22°C, the pavement cells of nudt6-2 nudt7 were much smaller. We also observed a loss of the interdigitation of adjacent pavement cells in nudt6-2 nudt7, compared with their interlocked arrangement in Col-0. At 28°C, however, both the cell size and the cell shape of nudt6-2 nudt7 plants were restored to that of the wild type (Figure 2c). Because the leaf area and cell area of Col-0 at 22°C was around 60 and 10 times that of nudt6-2 nudt7 (Figure 2d), respectively, the number of pavement cells per leaf in nudt6-2 nudt7 was about one-sixth of that in Col-0, indicating that the reduced leaf size of nudt6-2 nudt7 at 22°C was caused by a reduction in both the number and size of cells.
To determine whether the requirement for AtNUDT6 and AtNUDT7 to maintain normal plant growth at 22°C is specific for some developmental stages or some tissues, plants at different stages were transferred from 28°C to 22°C. If the transfer occurred during vegetative growth, leaves of the double mutant became curly and the expansion of young leaves was arrested (Figure 3a). If the transfer occurred at the bolting stage, elongation of the stems of the double mutant ceased, whereas the development of flowers and siliques were normal, leading to clustered siliques (Figure 3b). Seeds harvested from these siliques were normal in morphology and germination. Conversely, whenever the double mutant was shifted from 22°C to 28°C, new leaves and stems began to grow normally (Figure 3c). No developmental defects were observed in the roots of nudt6-2 nudt7, even at 22°C (Figure S2). These results indicate that the nudt6-2 nudt7 mutation has no obvious effects on reproductive growth and root development.
The autoimmune response in nudt6-2 nudt7 is temperature dependent
Mutants with a constitutive defense response often exhibit spontaneous cell death. Thus we examined the microscopic lesions of Col-0, nudt6-2, nudt7 and nudt6-2 nudt7. As shown in Figure 4a, when plants were grown at 22°C, spontaneous microscopic lesions were not observed in Col-0 or nudt6-2, and only occurred sporadically in nudt7. The double mutant, however, developed microlesions in high densities. For all four genotypes, microlesions were absent when plants were grown at 28°C. If plants grown at 28°C were shifted to 22°C, microlesions began to develop in nudt7 and nudt6-2 nudt7 within 4 days, to an extent comparable with the corresponding plants grown at 22°C. One week after the transfer, patches of dead cells visible to the naked eye began to emerge on young leaves of nudt6-2 nudt7 but not on Col-0 (Figure 4b). There was no significant H2O2 accumulation on leaves of any of the four genotypes grown at 28°C. One day after the temperature shift, H2O2 accumulation was the strongest on nudt6-2 nudt7 leaves and intermediate on nudt7 leaves, but was absent on Col-0 and nudt6-2 leaves (Figure 4c). Similarly, constitutive strong expression of PR1 and PR2 and over-accumulation of SA were detected in nudt6-2 nudt7 plants grown at 22°C or transferred from 28°C to 22°C, but not in plants grown at 28°C (Figure 4d,e). We also found that nudt6-2 nudt7 and Col-0 plants grown at 28°C showed no significant difference in their resistance to Pst DC3000, but a stronger resistance was observed in nudt6-2 nudt7 plants transferred from 28°C to 22°C (Figure 4f). Taken together, we concluded that the moderately high temperature (28°C) could fully suppress the autoimmune response in nudt6-2 nudt7.
The autoimmunity of nudt6-2 nudt7 is partially mediated by SNC1 and fully dependent on EDS1
The phenotype of nudt6-2 nudt7 is reminiscent of bon1-1, which exhibits a constitutively elevated defense response and stunted growth at 22°C, but a normal phenotype at 28°C. The loss of function of BON1 in the Col-0 ecotype activates SNC1, a TIR-NB-LRR-type R gene (Yang and Hua, 2004). We therefore tested whether the activation of SNC1 also contributes to the phenotype observed in nudt6-2 nudt7. nudt6-2 nudt7 was crossed with snc1-11, a loss-of-function mutant of SNC1 (Yang and Hua, 2004), to generate the nudt6-2 nudt7 snc1-11 triple mutant. When grown at 22°C, the stature and PR1 gene expression level of nudt6-2 nudt7 snc1-11 were intermediate between that of Col-0 and nudt6-2 nudt7 (Figure 5a,b). In addition, after being transferred from 28°C to 22°C, the resistance of nudt6-2 nudt7 snc1-11 to Pst DC3000 was weaker than that of nudt6-2 nudt7, but was stronger than that of Col-0 (Figure 5c). These results indicate that the autoimmune response of nudt6-2 nudt7 is partially mediated by SNC1, and could be divided into two branches: an SNC1-dependent pathway and an SNC1-independent pathway.
The SNC1-mediated defense response was modulated by temperature (Zhu et al., 2010). Our results show that the SNC1-independent autoimmune pathway activated in nudt6-2 nudt7 is also temperature dependent. In addition to the fact that it could be fully blocked at 28°C (Figures 2–4), we found that this SNC1-independent pathway was enhanced at 19°C, characterized by reduced stature and higher PR1 expression in nudt6-2 nudt7 snc1-11 plants grown at 19°C, compared with plants grown at 22°C (Figure 5a,b). When nudt6-2 nudt7 snc1-11 at the reproductive stage was transferred from 22°C to 19°C, elongation of the stems was arrested (Figure 5a), similar to nudt6-2 nudt7 transferred from 28°C to 22°C (Figure 3b).
It was reported that eds1-2 is epistatic to nudt7 (Straus et al., 2010). We therefore tested whether defects in nudt6-2 nudt7 also require EDS1. Combining eds1-2 (in the Col background) (Bartsch et al., 2006) with nudt6-2 nudt7 led to the full suppression of nudt6-2 nudt7 dwarfism at 22°C (Figure 5a). Moreover, the nudt6-2 nudt7 eds1-2 mutant was indistinguishable from eds1-2 in basal PR1 gene expression and resistance to Pst DC3000 (Figure 5b,c). Like nudt7 (Ge et al., 2007), nudt6-2 nudt7 plants were also more sensitive to paraquat than the Col-0 plants (Figure S3a). However, there was no significant difference in the sensitivity to paraquat between Col-0, eds1-2 and nudt6-2 nudt7 eds1-2, as indicated by their visible symptoms and similar total chlorophyll content before and after the paraquat treatment (Figure S3b). Taken together, we concluded that eds1-2 is epistatic to both nudt6-2 and nudt7. Interestingly, we also observed the nuclear accumulation of EDS1 in nudt6-2 nudt7 shortly after the temperature shift from 28°C to 22°C, which might be the triggering event for the downstream autoimmune response (Appendix S1; Figure S4).
SNC1 is upregulated at the transcriptional level by a positive feedback regulation from EDS1-mediated immunity in nudt6-2 nudt7
We next investigated how AtNUDT6 and AtNUDT7 regulates SNC1. In nudt6-2 nudt7, the expression level of SNC1 was similar to that in Col-0 at 28°C, but was significantly upregulated (by around threefold) after being transferred from 28°C to 22°C (Figure 6a). It was reported that the upregulation of SNC1 at the transcriptional level is sufficient for the induction of autoimmunity (Stokes et al., 2002; Yi and Richards, 2007). Therefore, the detected transcriptional upregulation of SNC1 contributed at least partially to the autoimmunity in nudt6-2 nudt7 at 22°C. We also observed that the accumulation of SNC1 transcript was fully abolished in the nudt6-2 nudt7 eds1-2 mutant (Figure 6a), suggesting that transcriptional regulation of SNC1 is not a direct effect of AtNUDT6 and AtNUDT7, but the result of positive feedback regulation by the EDS1-mediated immune response. Thus, the regulation of SNC1 at the post-transcriptional level is probably needed to initiate its activation in nudt6-2 nudt7. The feedback upregulation of SNC1 was also reported in bon1-1 (Yang and Hua, 2004). However, the autoimmunity of nudt6-2 nudt7 does not seem to be caused by a suppression of BON1, as the expression of BON1 was highly induced by the nudt6-2 nudt7 mutation at 22°C (Figure 6b).
To investigate whether AtNUDT7 regulates SNC1 in a way similar to CPR1, three Col-0 lines overexpressing AtNUDT7 were crossed with snc1-1. snc1-1 exhibits constitutive disease resistance caused by a point mutation in SNC1 (Zhang et al., 2003). The snc1-1 plants homozygous for the 35S:AtNUDT7 transgene were obtained in the F2 population. These plants were not morphologically distinguishable from snc1-1 plants, nor did they show reduced PR1 gene expression (Figure 6c). This result is in contrast to CPR1, the overexpression of which rescued the autoimmune phenotypes of snc1-1 (Cheng et al., 2011; Gou et al., 2012).
AtNUDT6 and AtNUDT7 are not direct in vivo regulators of NAD(H) or ADP-ribose
Several lines of evidence support a positive feedback loop between NADH and the plant defense response (Ge et al., 2007; Zhang and Mou, 2009; Petriacq et al., 2012). Loss of function of AtNUDT6 or AtNUDT7 led to elevated levels of NADH (Ge et al., 2007; Ishikawa et al., 2009, 2010; Jambunathan et al., 2010). AtNUDT6 and AtNUDT7 could negatively regulate cellular NADH levels directly by their enzymatic activity, or indirectly through suppressing the plant defense response, as illustrated in Figure 7a.
To determine which hypothesis in Figure 7a is true, NADH was quantified in Col-0, nudt6-2 nudt7, eds1-2 and nudt6-2 nudt7 eds1-2 grown at 28°C or transferred from 28°C to 22°C. No difference in NADH content was detected between these plants grown at 28°C. The temperature shift increased the level of NADH in nudt6-2 nudt7, but not in other genotypes, indicating that the activation of defense response in the double mutant was needed to elevate the NADH level. Notably, nudt6-2 nudt7 eds1-2 did not accumulate more NADH than eds1-2 before or after the temperature shift (Figure 7b), or even after the expression of AtNUDT6 and AtNUDT7 was greatly induced by pathogen inoculation in eds1-2 (Figure 7c,d). Taken together, our data support an indirect role of AtNUDT6 and AtNUDT7 in regulating cellular NADH levels. Similarly, nudt6-2 nudt7 eds1-2 plants did not accumulate more NAD+ or ADP-ribose than eds1-2 plants, even when AtNUDT6 and AtNUDT7 were induced in eds1-2 by pathogen invasion (Figure S5), eliminating the possibility that AtNUDT6 and AtNUDT7 directly regulate the cellular levels of NAD+ or ADP-ribose.
The ammonium/nitrate ratio in the growth medium modulates the autoimmune response of nudt6-2 nudt7
Although soil-grown nudt6-2 nudt7 is severely stunted at 22°C, the double mutant was visually indistinguishable from Col-0 when grown on solid MS medium containing 3% sucrose (Figure S6a). We hypothesized that certain ingredients in the MS medium might suppress the growth stunting of nudt6-2 nudt7. We first reduced the concentration of sucrose to 0.5 or 0.2%, which caused slower growth of both Col-0 and nudt6-2 nudt7 plants to a similar extent (Figure S6b), eliminating the role of sucrose as an anti-dwarfism factor specific for nudt6-2 nudt7 in the MS medium. This is also the case for growth media with a reduced concentration (2% of the concentration in the MS medium) of phosphorus, potassium, calcium, magnesium or sulfur (Figure S6a).
Inorganic nitrogen is available in two forms in the MS medium: nitrate and ammonium. Nitrate is normally the dominant nitrogen source in soil because of microbial nitrification (Marschner and Marschner, 2012). The MS medium contains 21 mm ammonium and 39 mm nitrate. Reducing the nitrogen level by reducing both nitrate and ammonium proportionally (Figure 8c–e) or increasing the ammonium/nitrate ratio (Figure 8f) did not make nudt6-2 nudt7 distinguishable from Col-0. Note that nitrogen deficiency (Figure 8e) or ammonium toxicity (Britto and Kronzucker, 2002) (Figure 8f) caused chlorosis in all genotypes tested. However, when the ammonium level was reduced to 5% of that in the standard MS medium (Figure 8i), or when the ammonium/nitrate ratio was further reduced (Figure 8g, h, j), nudt6-2 nudt7 plants exhibited extremely stunted growth. Consistent with this, increasing the ammonium/nitrate ratio in soil or in liquid culture alleviated the stunted growth phenotype of nudt6-2 nudt7 at 22°C (Figure S7).
The low ammonium/nitrate medium also caused a much higher level of constitutive expression of PR1 in nudt6-2 nudt7 compared with the other media tested (Figure 8k). Besides, the dwarf phenotype of nudt6-2 nudt7 caused by the low ammonium/nitrate media was fully dependent on EDS1 (Figure 8l). These results indicate that the low ammonium/nitrate ratio enhances the autoimmunity of nudt6-2 nudt7. Interestingly, snc1-1 and cpr1, two other autoimmune mutants that display constitutive disease resistance and growth retardation when grown in soil (Li et al., 2007; Cheng et al., 2011; Gou et al., 2012), had a near-normal stature on the standard MS medium, but were severely stunted on the MS medium with a low ammonium/nitrate ratio (Figure 8l). Taken together, our findings indicate that the low ammonium/nitrate ratio could be a common factor contributing to the autoimmune phenotypes in these mutants.
The ammonium/nitrate ratio modulates cellular levels of nitrite and NO
In plant cells, nitrate is reduced by nitrate reductase (NR) to nitrite, which is further reduced to ammonium by nitrite reductase (NiR) (Crawford and Forde, 2002). Ammonium is then assimilated into amino acids. NR is also known to catalyze the formation of NO in plants by transferring an electron from NAD(P)H to nitrite (Dean and Harper, 1988; Leitner et al., 2009). Arabidopsis utilizes ammonium preferentially when provided with a mixed nitrogen source (Gazzarrini et al., 1999). When fed with nitrate as the sole nitrogen source, Arabidopsis may accumulate more nitrite as the nitrate → nitrite → ammonium transition cannot be circumvented. Indeed, a significant increase in nitrite content was observed within 8 h of plants being transferred from the MS medium to the MS/−NH4+ medium (Figure 9a). Using the NO-specific fluorescent probe 4-amino-5-methylamino-2′,7′-difluorofluorescein diacetate (DAF-FM-DA), we observed elevated levels of NO in nudt6-2 nudt7 within 8 h of the transfer from MS medium to MS/−NH4+ medium (Figure 9b,c). The NO-associated fluorescence was mainly localized in guard cells, a pattern reported previously by Bright et al. (2006). Such NO accumulation was not observed in Col-0, eds1-2 or nudt6-2 nudt7 eds1-2, suggesting that it depends on the activation of the EDS1-mediated immunity. The observed fluorescence in nudt6-2 nudt7 could be diminished or enhanced by treating nudt6-2 nudt7 with the NO scavenger 2-(4-Carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (cPTIO) or the NO donor sodium nitroprusside (SNP), respectively, and was fully abolished if the plant material was not loaded with DAF-FM-DA (Figure 9b,d), indicating that the observed fluorescence was NO-specific.
Nitrite-dependent NO production could be catalyzed by NR (Rockel et al., 2002) or by the mitochondrial electron transport chain (Gupta and Igamberdiev, 2011), or occurs non-enzymatically in the apoplast (Bethke et al., 2004). Tungstate, an inhibitor of NR, abolished the NO accumulation in nudt6-2 nudt7 (Figure 9b,d), indicating that NR was responsible for the observed NO production. To further investigate whether NO contributes to the autoimmunity of nudt6-2 nudt7, plants grown on MS medium were treated with SNP. PR1 expression was induced significantly in SNP-treated nudt6-2 nudt7, but only slightly in SNP-treated Col-0, eds1-2 and nudt6-2 nudt7 eds1-2. Moreover, the scavenging of NO generated in the nudt6-2 nudt7 plants transferred to MS/−NH4+ medium greatly reduced the amplitude of PR1 upregulation (Figure 9e). These results indicate that NO acts as an amplifier of the EDS1 signaling pathway in nudt6-2 nudt7, in addition to the previous finding that NO accumulation in nudt6-2 nudt7 is EDS1-dependent.
In this study, we found that unlike nudt7, the nudt6-2 mutation alone does not lead to observable alterations in growth, development or defense response. However, the nudt6-2 nudt7 mutation triggers a much stronger autoimmune response than the nudt7 mutation. Such a synergistic effect of nudt6-2 and nudt7 mutation indicates that AtNUDT6 and AtNUDT7 function redundantly as negative regulators of defense responses, in contrast to a previous report that AtNUDT6 is a positive regulator of plant defense response (Ishikawa et al., 2010).
Autoimmune phenotypes of nudt6-2 nudt7 were fully dependent on EDS1, consistent with the previous reports that eds1-2 is epistatic to nudt7 (Bartsch et al., 2006; Straus et al., 2010). We also found that nudt6-2 nudt7 eds1-2 was indistinguishable from eds1-2 in sensitivity to oxidative stress imposed by paraquat. Several other abiotic stresses were also tested, including high salinity, osmotic stress, UV-radiation, high temperature, chilling, drought, genotoxic stress and endoplasmic reticulum (ER) stress. In all these tests, no differences between eds1-2 and nudt6-2 nudt7 eds1-2 were observed. These results suggest that AtNUDT6 and AtNUDT7 might function as specific regulators of the EDS1-mediated defense pathway.
The autoimmunity activated in nudt6-2 nudt7 is mediated by an SNC1-dependent pathway and an SNC1-independent pathway, with both pathways sensitive to temperature and dependent on EDS1. As all TIR-NB-LRRs tested to date genetically converge on EDS1 (Wiermer et al., 2005), and NB-LRR proteins were suggested to be the target of temperature modulation (Zhu et al., 2010), it is highly possible that the temperature-sensitive, SNC1-independent and EDS1-dependent pathway in nudt6-2 nudt7 is caused by the or activation of one more other TIR-NB-LRRs (Figure 10).
In this study, we also found that AtNUDT6 and AtNUDT7 do not directly regulate cellular levels of NAD(H) or ADP-ribose in vivo. Moreover, the MJ1149 transgene encoding a highly specific ADP-ribose pyrophosphohydrolase (Sheikh et al., 1998) could not alleviate the phenotypes of nudt6-2 nudt7, indicating that ADP-ribose accumulation is not a factor contributing to the phenotypes of nudt6-2 nudt7. These results raised the possibility that AtNUDT6 and AtNUDT7 exert their functions not by modulating NAD(H) or ADP-ribose levels in vivo, but by interacting with and modulating the activities of other proteins, a process probably modulated by their substrates (such as NADH or ADP-ribose). Such examples include GTPases and R proteins, which function as signaling components, rather than modulators of cellular GTP and ATP levels, respectively (Li and Zhang, 2004; Takken and Tameling, 2009). Interestingly, AtNUDT7 was reported to interact with 14-3-3 proteins (Olejnik et al., 2009), RACK1A and the gamma subunit of G proteins (Olejnik et al., 2011), but the physiological effects of these interactions are still unknown. Some point mutations in AtNUDT7 may change its quaternary structures and hamper its interaction with other proteins (Olejnik et al., 2009), but the physiological consequences of these point mutations have not been assessed.
The phenotypes of nudt6-2 nudt7 are reminiscent of several mutants extensively studied recently, including bon1 (Yang and Hua, 2004; Yang et al., 2006b), bap1 (Yang et al., 2006a, 2007), bir1 (Gao et al., 2009; Wang et al., 2011), srfr1 (Kim et al., 2010; Li et al., 2010), mkp1 (Bartels et al., 2009) and cpr1 (Cheng et al., 2011; Gou et al., 2012). Activation of SNC1 in these mutants, albeit a common factor contributing to autoimmunity, is probably mechanistically distinct (Gou and Hua, 2012). CPR1 is a component of the SKP1-CULLIN1-F-box (SCF) complex, which negatively regulates the accumulation of R proteins such as SNC1 and RPS2 (Cheng et al., 2011; Gou et al., 2012). SRFR1 also negatively modulates the accumulation of several R proteins including SNC1 (Kim et al., 2010; Li et al., 2010). The detailed mechanism on how BON1, BAP1, BIR1, MKP1, AtNUDT6 and AtNUDT7 negatively regulate SNC1 is unclear. It is likely that they negatively modulate R genes at the transcriptional or post-transcriptional level. It is also possible that they represent ‘gardees’ or ‘decoys’ (van der Hoorn and Kamoun, 2008), the presence of which are under the surveillance of R proteins. One such example is the Arabidopsis RIN4 gene. The presence of RIN4 is guarded by RPS2, and the knock-out of RIN4 mimics the elimination of RIN4 by the effector AvrRpt2, thus triggering the RPS2-mediated autoimmunity (Mackey et al., 2003; Belkhadir et al., 2004). Another possibility is that these negative regulators of SNC1 are not ‘guardees’ or ‘decoys’ themselves, but regulators of a ‘guardee’ or ‘decoy’ of SNC1. It was reported that SNC1 physically interacts with EDS1 (Bhattacharjee et al., 2011). This raised the hypothesis that EDS1 might be a guardee of SNC1, in addition to its role as a TIR-NB-LRR signal transducer (McDowell, 2011). The loss of function of these negative regulators of SNC1 might lead to alterations of EDS1, which activates SNC1.
Improper activation of defense responses almost always has a fitness cost, characterized by a reduced biomass and seed yield (Tian et al., 2003; Heidel et al., 2004). We found that the inhibition of cell expansion and division by autoimmunity in nudt6-2 nudt7 is specific for leaves and stems, but not in roots, flowers or siliques, probably because in these organs autoimmunity is either not established or is uncoupled from growth inhibition. It was reported that AtNUDT6 and AtNUDT7 are both expressed in leaves and stems, but only AtNUDT7 transcript could be detected in roots (Ogawa et al., 2005). Further work is needed to dissect the physiological significance of their organ-specific expression patterns. Accumulating evidence has revealed a role of auxin in plant defense responses (Kazan and Manners, 2009). In this study, we found that in nudt6-2 nudt7, the interdigitation of leaf pavement cells was normal at 28°C, but was compromised at 22°C. Recently it was reported that the interdigitation of adjacent pavement cells is dependent on the Rho GTPase-mediated mechanisms activated by auxin (Fu et al., 2005; Xu et al., 2010).
The role of nitrate and ammonium as essential macronutrients and signaling molecules has been extensively studied (Krouk et al., 2010; Vidal et al., 2010; Castaings et al., 2011), but their impact on plant defense response is still obscure. We found that a low ammonium/nitrate ratio in the growth medium contributes to the autoimmune response of nudt6-2 nudt7, snc1-1 and cpr1, suggesting that it might be a common phenomenon for mutants with R gene over-activation. A low ammonium/nitrate ratio results in the accumulation of nitrite and NO in nudt6-2 nudt7, and our results support a positive amplification loop between NO and the EDS1 signaling pathway. But it should be noted that NO might not be the only route through which the ammonium/nitrate ratio exerts its effect on EDS1-dependent immunity, considering that nitrate and ammonium controls many aspects of plant growth and development.
Our findings that a low ammonium/nitrate ratio induces the autoimmune response of the three constitutive disease resistance mutants characterized in this study raise the possibility that plants might monitor the cellular nitrate and ammonium levels as a signal of pathogen invasion. Plant pathogens were found to preferentially use readily assimilated nitrogen sources such as ammonium and glutamine (Marzluf, 1997). Depletion of cellular ammonium by a pathogen is expected to result in a low ammonium/nitrate ratio in plant cells and, in turn, higher levels of nitrite and NO, which could be sensed by plants as a signal of pathogen infection and positively regulate plant defense responses. The mutations in nudt6-2 nudt7, snc1-1 and cpr1, and perhaps also in other constitutive disease resistance mutants, might sensitize the defense pathway, which is readily activated by a low ammonium/nitrate ratio, even in the absence of pathogen infection.
Plant growth and inoculation with pathogens
Plants were grown at 22°C or 28°C with 50% humidity and a light intensity of 125 m m−2 s−1 under a 14 h light/10 h dark photoperiod. Plants used in the in planta bacteria growth assay or in the growth media were grown under a 10 h light/14 h dark photoperiod for a longer vegetative growth phase. The solid growth media used in this study contains various ingredients of inorganic salts (Table S2), 3% sucrose and 0.8% agar, unless otherwise stated. The liquid growth media have the same recipe of inorganic salts, except that they were supplemented with 5 g L−1 sucrose and 1 × Vitamin Solution (Sigma-Aldrich, http://www.sigmaaldrich.com). The pH was adjusted to 5.7 for all the media mentioned above.
Growth of P. syringae and the in planta bacteria growth assay were conducted as described by Ge et al. (2007). For the induction of the hypersensitive response, 1 × 107 cfu ml−1Pst avrRpm1 or Pst avrRpt2 suspended in water was infiltrated into half leaves of 4-week-old plants.
3,3′-diaminobenzidine staining and trypan blue staining
3,3′-Diaminobenzidine (DAB) staining and trypan blue staining were carried out as described previously (Ge et al., 2007).
Quantification of H2O2, NAD(H), ADP-ribose, free salicylic acid, chlorophyll and nitrite
H2O2, NAD(H) and ADP-ribose were measured as described previously (Ge et al., 2007). Free salicylic acid (SA) (Pan et al., 2010), chlorophyll (Ni et al., 2009) and nitrite (Sugiura et al., 2007) were measured as reported before.
Pharmacological study of NO
The detection and quantification of NO in leaves was performed as described by Wang et al. (2010), with minor modification. Two-week-old plants grown on vertical MS medium were transferred to liquid MS medium or liquid MS/−NH4+ medium to grow for the time specified (Figure 9b,c), before being loaded with DAF-FM-DA. Fluorescence was observed using the FV1000 confocal microscope (Olympus, http://www.olympus.com), with the 488 nm argon laser line, 515–560 nm bandpass detection for NO and 650–700 nm bandpass detection for chlorophyll. For NO quantification, the pixel values were measured using ImageJ from areas of 1.6 × 105 μm2 randomly selected on each leaf. The mean and SD values were each calculated from 15 independent leaf samples.
For NO observation in cPTIO-, SNP- or tungstate-treated nudt6-2 nudt7, MS medium-grown nudt6-2 nudt7 seedlings were transferred to liquid MS/−NH4+ medium to grow for 8 h. The liquid MS/−NH4+ medium was supplemented with 100 μm SNP or 500 μm cPTIO 15 min before loading leaf samples with DAF-FM-DA or with 100 μm tungstate 2 h before loading with DAF-FM-DA.
To investigate the effect of NO on PR1 gene expression, 2-week-old plants grown vertically on MS medium were transferred to liquid MS medium (with or without 100 μm SNP) or liquid MS/−NH4+ medium (with or without 200 μm cPTIO) to grow for 4 h. Then SNP and cPTIO was removed by transferring plants to corresponding fresh liquid medium. Samples were harvested 2 h later and PR1 expression was measured by qRT-PCR.
We thank Dr Jian Hua at Cornell University for the snc1-11 seeds, Dr Jane E. Parker at the Max Planck Institute and Dr Shuhua Yang at China Agricultural University for the eds1 seeds, Dr Yuelin Zhang at National Institute of Biological Sciences for the snc1-1 seeds, Dr Hao Chen and Dr Jing Li at the Donald Danforth Plant Science Center for helpful discussions and help with SA measurements. This work was supported by Research Grants Council of Hong Kong (grant no. HKBU1/CRF/10 and HKBU261910 to YX) and by the strategic development fund of Hong Kong Baptist University (grant no. SDF090910P03 and SD-HKBU1/CRF/10TOPUP to YX and JZ).