Leaf senescence signaling: The Ca 2+ -conducting Arabidopsis cyclic nucleotide gated channel2 acts through nitric oxide to repress senescence programming

Ca 2+ and nitric oxide (NO) are essential components involved in plant senescence signaling cascades. In other signaling pathways, NO generation can be dependent on cytosolic Ca 2+ . The Arabidopsis ( Arabidopsis thaliana ) mutant dnd1 lacks a plasma membrane-localized cation channel (CNGC2). We recently demonstrated that this channel affects plant response to pathogens through a signaling cascade involving Ca 2+ modulation of NO generation; the pathogen response phenotype of dnd1 can be complemented by application of an NO donor. At present, the interrelationship between Ca 2+ and NO generation in plant cells during leaf senescence remains unclear. Here, we use dnd1 plants to present genetic evidence consistent with the hypothesis that Ca 2+ uptake and NO production play pivotal roles in plant leaf senescence. Leaf Ca 2+ accumulation is reduced in dnd1 leaves compared to wild type. Early senescence-associated phenotypes (such as loss of chlorophyll, expression level of senescence associated genes, H 2 O 2 generation, lipid peroxidation, tissue necrosis, and salicylic acid levels) were more prominent in dnd1 leaves compared to wild type. Application of a Ca 2+ channel blocker hastened senescence of detached wild type leaves maintained in the dark; increasing the rate of chlorophyll loss, expression of a senescence associated gene, and lipid peroxidation. Pharmacological manipulation of Ca 2+ signaling provides evidence consistent with genetic studies of the relationship between Ca 2+ signaling and senescence with the dnd1 mutant. Basal levels of NO in dnd1 leaf tissue were lower than that in leaves of wild type plants. Application of an NO donor effectively rescues many dnd1 senescence related phenotypes. Our work demonstrates that the CNGC2 channel is involved in Ca 2+ uptake during plant development beyond its role in pathogen defense response signaling. Work presented here suggests that this function of CNGC2 may impact downstream 'basal' NO production in addition to its role (also linked to NO signaling) in pathogen defense responses, and that this NO generation acts as a negative regulator during plant leaf senescence signaling. Ca 2+ uptake and leaf senescence. leaf

In animal cells, all three NOS isoforms require Ca 2+ /CaM as a cofactor (Alderton et al., 2001;Stuehr, 1999;Nathan and Xie, 1994). Notably, animal NOS contains a CaM binding domain (Stuehr, 1999). It is unclear whether Ca 2+ /CaM can directly modulate plant NOS, or if Ca 2+ /CaM impacts plant leaf development/senescence through (either direct or indirect) effects on NO generation. However, recent studies from our lab suggest that Ca 2+ /CaM acts as an activator of NOS activity in plant innate immune response signaling (Ma et al., 2008;Ali et al., 2007).
However, the AtNOA1 loss-of-function mutant does display reduced levels of NO generation and several groups have used the NO donor sodium nitroprusside (SNP) to reverse some low-NO related phenotypes in Atnoa1 plants (Zhao et al., 2007;Bright et al., 2006;Guo et al., 2003). Importantly, plant endogenous NO-deficiency- (Mishina et al., 2007;Guo and Crawford, 2005) or abscisic acid-(ABA)/MeJA (Hung andKao, 2004, 2003) induced early senescence can be successfully rescued by application of exogenous NO. Addition of NO donor can delay gibberellin (GA)-elicited PCD in barley (Hordeum vulgare) aleurone layers as well (Beligni et al., 2002).
It has been suggested that salicylic acid (SA), a critical pathogen defense metabolite, can be increased in natural (Mishina et al., 2007;Morris et al., 2000) and transgenic NOD induced senescent Arabidopsis leaves (Mishina et al., 2007(Mishina et al., ). et al., 2003Sunkar et al., 2000). Arabidopsis 'defense, no death' (dnd1) mutant plants have a null mutation in the gene encoding the plasma membrane localized Ca 2+conducting CNGC2 channel. This mutant also displays no hypersensitive response to infection by some pathogens (Ali et al., 2007;Clough et al., 2000). In addition to involvement in pathogen mediated Ca 2+ signaling, CNGC2 has been suggested to participate in the process of leaf development/senescence (Köhler et al., 2001). dnd1 mutant plants have high levels of SA and expression of PR1 (Yu et al., 1998) and spontaneous necrotic lesions appear conditionally in dnd1 leaves (Jirage et al., 2001;Clough et al., 2000). Endogenous H 2 O 2 levels in dnd1 mutants are increased from WT levels (Mateo et al., 2006). Reactive oxygen species molecules such as H 2 O 2 are critical to the PCD/senescence processes of plants (Zimmermann et al., 2006;Guo and Crawford, 2005;Hung and Kao, 2004;Navabpour et al., 2003;Overmyer et al., 2003). Here, we use the dnd1 mutant to evaluate the relationship between leaf Ca 2+ uptake during plant growth and leaf senescence. Our results identify NO, as affected by leaf Ca 2+ level, to be an important negative regulator of leaf senescence initiation. Ca 2+ -mediated NO production during leaf development could control senescence associated gene (SAG) expression as well as the production of molecules (such as SA and H 2 O 2 ) that act as signals during the initiation of leaf senescence programs.

Leaf Ca 2+ accumulation is reduced in the dnd1 mutant
Ca 2+ conducting ion channels facilitate transitory cytosolic Ca 2+ 'spikes' as components of numerous signaling pathways in plant cells. They are also thought to play a role in Ca 2+ nutrition; i.e., uptake and translocation of Ca 2+ within the plant (Hetherington and Brownlee, 2004;White and Broadley, 2003). Patch clamp studies of plant cells also indicate that nonselective, weakly voltage gated (i.e. potentially ligand gated) channels contribute to Ca 2+ uptake in plants (Demidchik and Maathuis, 2007;White and Broadley, 2003;White et al., 2002). However, current reviews indicate no gene product has yet been associated with Ca 2+ uptake into plants or accumulation in leaves (Maathuis, 2009). Therefore, it is currently unclear from prior work which cation channels contribute to Ca 2+ uptake into plants and accumulation in leaves during plant growth and development (in contrast to temporary influx of Ca 2+ associated with signaling).
Of the 57 known cation conducting channels in Arabidopsis, 20 members are CNGCs. CNGCs are candidates for specific gene products involved in the plant Ca 2+ uptake pathway (Demidchik and Maathuis, 2007;White and Broadley, 2003;White et al., 2002). The cytosolic secondary messenger cAMP activates inward Ca 2+ current through the plasma membrane in WT leaf cells while this current is absent in leaves of the CNGC2 loss-of-function mutant dnd1 (Ali et al., 2007;Lemtiri-Chlieh and Berkowitz, 2004). Pathogen recognition results in activation of CNGC-dependent inward Ca 2+ current and downstream NO production during plant innate immune response signaling cascades (Ma et al., 2009a;Ali et al., 2007). These studies link CNGC channels with inward Ca 2+ flux associated with signaling.
Here, we first investigated whether CNGC2 plays a role in Ca 2+ uptake from a plant nutrition perspective by comparing leaf Ca 2+ content in WT and dnd1 plants grown with different Ca 2+ levels in their growth medium. We found that the CNGC2 null mutation in dnd1 plants has effects on long-term Ca 2+ acquisition ( Figure 1) at a range of external Ca 2+ , either when we measured total shoot Ca 2+ levels in plants grown on solid (agar) medium (experiments 1 and 2), or grown hydroponically on liquid nutrient solution (experiment 3 and 4). In these experiments (Figure 1), we observed a decrease in Ca 2+ content of leaves of dnd1 plants as compared to WT plants under a number of different growth conditions (solid and liquid media, and varying external Ca 2+ ). These results are consistent with a hypothesis that conductance through cation channels formed by CNGC2 contributes to Ca 2+ nutrition of the plant. It appears that CNGC2 channels contribute to Ca 2+ acquisition by leaves as part of the normal growth and development of the plant in addition to the Ca 2+ conductance associated with innate immune signaling described by Ali et al. (2007) and Ma et al. (2009a).
The work shown in Figure 1 is presented because the results are consistent with function of CNGC2 as a Ca 2+ uptake pathway in leaves during growth and development (i.e. in addition to its role in immune responses). We do not assert here that the levels of leaf Ca 2+ found in WT and dnd1 plants as shown in Figure 1 are specifically causal to differences in senescence programming in WT and mutant plants. Rather, we speculated that since CNGC2 is functioning as a Ca 2+ uptake pathway during normal growth of WT plants, this channel may be responding to as-yet-unidentified signals during growth that impact the onset of senescence in leaves. CNGC2 loss-of-function in dnd1 plants could impact this signaling.

of-function mutant (dnd1) displays early senescence phenotypes
Results ( Figure 1) indicating that CNGC2 provides a pathway for Ca 2+ uptake in plants beyond that related to immune signaling led us to investigate the link between CNGC2 and leaf senescence. We observed that dnd1 plants show early leaf senescence phenotypes compared to WT plants ( Figure 2). In Figure 2A (Mishina et al., 2007).
Previous studies have reported that spontaneous necrotic lesions can appear in dnd1 leaves (Jirage et al., 2001;Clough et al., 2000). Here, our observation further expands our understanding of the dnd1 mutant phenotype. Subjecting detached young leaves to darkness also revealed that dnd1 leaves achieve senescence faster than WT leaves ( Figure 2B and C). With leaves from WT plants, Gd 3+ , a Ca 2+ channel blocker, hastens senescence, mimicking the early senescence phenotype of dnd1 leaves ( Figure 2B and C). Results in Figure 2A-C are consistent with the hypothesis that the presence of a functional Ca 2+ uptake pathway may act to defer senescence during development.  Figure 2D). We also found increased PR1 transcript levels in dnd1 plants ( Figure 2D), similar to what was reported by Yu et al. (1998). Furthermore, measurements of lipid peroxidation (i.e., quantified by monitoring malondialdehyde (MDA) levels), which increases during senescence (Wingler et al., 2004;Berger et al., 2001;Buchanan-Wollaston, 1997;Dhindsa et al., 1981) indicate that dnd1 leaves have a higher level of lipid peroxidation than WT leaves ( Figure 2E). A number of different experimental approaches as delineated in Figure 2, then, indicate that senescence 'programs' are activated in leaves of dnd1 plants as compared to leaves of WT plants. Figure 2B and C indicate that with regard to the rapid senescence that occurs in detached leaves kept in the dark, application of Gd 3+ , a channel blocker that prevents inward currents through plasma membrane Ca 2+ conducting channels (Tegg et al., 2005), mimics the phenotype shown by dnd1. Further studies were undertaken to determine if the effect of Gd 3+ on senescence programming could be observed at the biochemical and gene expression level. Expression of SAG12 is increased during senescence of detached leaves kept in the dark, as well as in planta in leaves undergoing natural senescence; not all SAG genes upregulated in planta show a similar response to darkness in detached leaves (Weaver et al. 1998). Quantitative real-time PCR (qPCR) evaluation of SAG12 expression in detached WT leaves kept in the dark demonstrated that the hastening of senescence by Gd 3+ could also be observed at the level of gene expression ( Figure 2F). SAG12 expression was greater at days 1-3 after detachment ( Figure 2F) in WT leaves when they were exposed to Gd 3+ . Further studies indicated that hastening of senescence by Gd 3+ could be observed at the biochemical level as well. With

dnd1 has a lower endogenous NO level in leaves compared to WT plant
The hypothesis underlying the work presented in this report is that impairment of a Ca 2+ uptake pathway functional during growth and development of dnd1 plants leads to loss of signal that represses senescence programming in leaves. We speculate that CNGC2-dependent Ca 2+ uptake affects leaf senescence programming through an intermediary step of NO synthesis. Thus, in dnd1 plants with a null mutation in CNGC2, downstream Ca 2+ -dependent activation of NO generation may be impaired. Reduced NO generation could lead to a derepression of leaf senescence programming. The basis for this conjecture is as follows. A) Prior studies (Mishina et al., 2007;Guo and Crawford, 2005;Corpas et al., 2004;Chou and Kao, 1992;Poovaiah and Leopold, 1973) discussed above suggest that both leaf Ca 2+ and NO repress leaf senescence. B) High levels of SA and expression of PR1 are associated with treatments that reduce NO production in leaves and induce early senescence (Mishina et al., 2007;Ülker, et al., 2007;Morris et al., 2000); SA level and PR1 expression are elevated in leaves of dnd1 plants (Yu et al., 1998). C) Senescing leaves have reduced endogenous NO level (Corpas et al., 2004). D) Recent studies from this lab (Ali et al., 2007) indicate that impairment of innate immune signaling in dnd1 plants occurs due to a lack of Ca 2+ -dependent NO generation. We measured the endogenous NO level in WT and dnd1 leaves. As shown in Figure 3, dnd1 plants have lower endogenous levels of leaf NO as compared to WT.
Endogenous basal NO (in contrast to NO generation in response to a specific signal) in leaves of any Arabidopsis genotype has not been previously reported. In prior studies from this lab (Ali et al., 2007), basal level of NO in guard cells was monitored in epidermal peels of WT and dnd1 plant leaves. No significant differences were noted although the level in dnd1 guard cells was slightly higher. The difference between these prior results focusing on guard cells and the work reported here may be due to the measurement of total leaf NO in the work shown in Figure 3.

Nitric oxide donor application rescues dnd1 senescence associated phenotypes
With regard to impaired pathogen response signaling in dnd1 plants, application of the NO donor SNP to these mutant plants complemented the dnd1 pathogen response phenotype (Ali et al., 2007). We therefore tested the model that is the focus of the work presented here about involvement of NO in Ca 2+ repression of senescence programming by examining the effect of SNP (an NO donor) application on some senescence-related phenotypes in leaves of dnd1 plants. As is the case with other NO donors used with plants, NO release from SNP requires light (Floryszak-Wieczorek et al., 2006). Therefore, our evaluation of SNP effects on senescence in dnd1 plants focused on phenotypes occurring in the light.
As mentioned above, prior studies have shown that PR1 expression and SA levels are elevated in dnd1 plants as compared to WT plants (Yu et al., 1998). We confirmed these results (results not shown, also see Figure 2D). We also find that SNP treatment reduced the high constitutive PR1 transcriptional level in dnd1 ( Figure 4A). In addition, SNP application to dnd1 plants also (modestly) lowered the constitutively high level of SA accumulation ( Figure 4B). SNP application had no significant effect on either Although only qualitative effects of SNP treatment can be discerned from the images shown in Figures 5A and 5B, it appears that the effect of exogenous NO supply on dnd1 plants, i.e. reduced H 2 O 2 and necrosis, is not induced by SNP in leaves of WT plants.
SNP application also was found to have an effect on the high level of lipid peroxidation in dnd1 plants ( Figure 2E). As shown in Figure 5C, application of the NO donor SNP reduced the high level of lipid peroxidation in dnd1 leaves while SNP had no effect on the level of lipid peroxidation in leaves of WT plants.
The expression level of the SAGs At5g10760, WRKY70, RLK5 and PR1 is increased in (attached) leaves of dnd1 plants ( Figure 2D). Application of the NO donor SNP to dnd1 plants reduced the expression level (monitored using RT-PCR) of these SAGs ( Figure 5D). Analysis using qPCR also demonstrated that exogenous NO application reduced expression of PR1 in leaves of dnd1 plants; SNP had no significant effect on PR1 expression in leaves of WT plants (Supplemental Figure 3). Results shown in Figure 5D and Supplemental Figure 3 are from experiments performed on young seedlings grown on agar medium enclosed in sealed boxes. As discussed above, in an experiment performed on mature plants with fully expanded leaves, we also found that a SNP treatment (in this case, mature leaves of plants were sprayed with aqueous solutions containing SNP) reduced PR1 expression (monitored in this experiment using Northern analysis) in leaves of dnd1 plants ( Figure 4A).
In this report, we have shown that a null mutation in CNGC2, a Ca 2+ -conducting plasma membrane cation channel, results in a reduction in leaf Ca 2+ levels during growth and development of dnd1 plants. We have associated the loss of function of this Ca 2+ uptake pathway in dnd1 plants with a number of senescence-related phenotypes that are complemented by exogenous application of the NO donor SNP. In Figure 6 Presumably, growth of seedlings in sealed, illuminated containers would allow for greater buildup of gaseous NO around (and in) plant tissue (as compared with spraying SNP on leaves or adding the NO donor to irrigation solution). Under these conditions, we note that increasing SNP in growth medium (to 100 µM) is lethal to WT seedlings while dnd1 seedlings are less affected (Figure 6). At 50 µM SNP, there is a visible difference between WT (more wilted appearance) and dnd1 seedlings (less affected) as well ( Figure   6). This sensitivity of WT Arabidopsis seedlings to SNP addition to the growth medium we report here is similar to that shown by He et al. (2004). Current work notes the paucity of easily discerned plant phenotypes of CNGC loss-of-function mutants (Frietsch et al., 2007). Thus, our identification here of an NO-related phenotype of the dnd1 mutant could provide the basis for further study of CNGC-related effects of Ca 2+ uptake inhibition on growth and development. One possible mechanism underlying the dnd1 phenotype shown in Figure 6 is as follows. Lower endogenous levels of NO present in dnd1 plants during growth (Figure 3) could allow for tissue in this mutant to be less sensitive (i.e., in an additive sense) to addition of exogenous NO.

DISCUSSION
Evidence presented in this manuscript depicts a model linking the function of the Ca 2+ conducting channel CNGC2 with downstream NO production and senescence programming. Prior work has shown that transitory Ca 2+ uptake into plant cells through CNGC2 initiated during pathogen defense signaling cascades involves downstream NO production. Here, we show that the same channel contributes to Ca 2+ uptake into the leaf during growth and development of the plant beyond this previously reported role in pathogen defense responses. The Ca 2+ uptake capability provided by CNGC2 apparently impacts NO generation during growth and development as well; null mutation of CNGC2 results in reduced endogenous NO level in dnd1 plants as compared to WT plants. We find that exogenous application of an NO donor to dnd1 plants reverses a number of senescence-related phenotypes. This link, between CNGC2-mediated uptake of Ca 2+ into leaf tissue and NO, presumably mediates leaf senescence development as a negative regulator. Therefore, this work provides new information about the molecular mechanism of plant senescence signaling.
We provide new experimental evidence indicating that dnd1 mutants have reduced NO production during growth and development of the plant and associate this reduction in 'basal' NO level with the absence of a Ca 2+ uptake pathway which is operative during growth of the plant. Guo and Crawford (2005) also attributed the complementation of the early senescence phenotype of atnoa1 mutants by SNP to a reversal (by the NO donor) of depressed NO generation during leaf senescence in these plants; they did not monitor the basal level of NO in these mutants. Mishina et al. (2007) also attributed the induction of early senescence by an NO-degrading treatment (NOD expression) to presumed changes in the basal level of this signaling molecule in leaves. In the work reported here, we observe that dnd1 plants has a reduced endogenous level in leaves compared to WT, which supports the early senescence phenotypes in dnd1 leaves that we found in this study.
Previous studies suggest that NO not only functions as a senescence signaling regulator (Mishina et al., 2007;Guo and Crawford, 2005) but also inhibits SA elevation during this process (Mishina et al., 2007). Work presented here is consistent with that model. Application of SNP to dnd1 plants reduces their high SA level, as well as the transcript level of a marker gene (PR1) for SA generation.
In summary, our studies provide new genetic information linking Ca 2+ uptake through CNGC2 channels and accumulation in leaves during the course of plant growth and development as a component of leaf senescence signaling. NO is also proposed to be involved in this signaling cascade. Results presented here are consistent with NO action as a negative regulator during the developmental progression to leaf senescence.

Dark-induced leaf senescence
The darkness-induced leaf senescence assay we used was a method adapted from

Lipid peroxidation
The extent of lipid peroxidation in leaves was evaluated by measuring MDA formation as described by Guo and Crawford (2005). For these studies, 3-week-old plants were used except for the experiment shown in Supplemental Figure 1; in this case plants were 5-week-old. In brief, each leaf sample was either ground in chilled extraction buffer (containing 0.25 % (w/v) thiobarbituric acid in 10 % (w/v) trichloroacetic acid (Fisher Scientific)) directly or first ground in liquid nitrogen followed by resuspension in extraction buffer. The homogenates were incubated in a water bath (90 o C) for 20 min.
Heated homogenates were then equilibrated to room temperature, and centrifuged (12,000 X g for 15 min). Supernatants were decanted for optical density measurement at

Endogenous NO in leaves
The method used to measure endogenous NO in leaves was adapted from that described by Lamattina and colleagues (Martin et al., 2009;Graziano and Lamattina, 2007). NO was monitored using the NO-specific

Leaf salicylic acid
Eight-week-old plants were treated with an NO donor by spraying leaves with water containing 0 or 100 µM SNP daily for 2 d as described in Beligni and Lamattina (1999). SA and its glucoside were quantified using gas chromatography-mass spectrometry. 100 mg of frozen leaf tissue were extracted twice with 800 μ L acetone: 50 mM citric acid (     ANOVA evaluation of means separation between -SNP and +SNP treatments indicated a significant difference for dnd1 leaves (p < 0.05; indicated with a '*' above bar representing the '+ SNP' treatment) and no significant difference for WT leaves. (D) Effect of (50 µM) SNP on SAG transcript accumulation in dnd1 leaves was analyzed by semi-quantitative RT-PCR. This experiment was repeated three times. tubulin is shown as a loading control. Leaf 5 of seedlings was used for this assay.   i.e. leaves begin turning yellow at the tip (highlighted by arrows). Insert shows an enlarged image of a portion from one leaf of the dnd1 plant. (B) Leaves detached from WT plants (top and center rows) and dnd1 plants (bottom row) were incubated on water, or water containing 100 µM Gd 3+ (center row). From left to right, lanes 1-4 show leaves after 0, 3, 4, and 5 d in darkness, respectively. Counting from the first true leaf, leaves 3-8 of seedlings were used for this assay. (C) Images are shown of individual detached leaf (in this case, leaves shown are either leaf 1 or 2) of the seedlings used for the experiment shown in (B). Lanes 1-3 correspond to 0, 3, and 4 d in darkness treatment, respectively. (D) RT-PCR products generated with primers corresponding to SAG genes, and RNA prepared from leaves (the entire rosette) of WT or dnd1 plants as template were subjected to agarose gel electrophoresis. A band corresponding to tubulin is shown as a loading control. (E) MDA levels in (rosette) leaves of WT and dnd1 plants (expressed per unit leaf fresh weight (FW)). Results are presented as mean (n = 3) + SE. ANOVA evaluation of means separation indicated differences between WT and dnd1 MDA levels were significant at p < 0.01. Similar results were found when the experiment was repeated three times. (F) Quantitative real-time PCR analysis of SAG12 transcript accumulation (relative to tubulin transcript) in WT detached leaves (leaves 3-5) left in the dark (0-4 d) on water or Gd 3+ . Results are shown as means + SE (n = 3). ANOVA analysis was used to evaluate means separation at each time point. Significant differences (at p < 0.01) are indicated by a '**' above symbols. A '**' above a bar representing NO levels in dnd1 plants indicates the difference from the level found in WT leaves is significant at p < 0.01 for an individual experiment. ANOVA (paired T-Test) evaluation of means separation between WT and dnd1 genotypes for the pooled values from all three experiments indicated the genotype differences were significant at p < 0.01. levels in leaves of WT (left panels) and dnd1 (right panels) plants were detected using 3,3'-diaminobenzidine (DAB) staining. Leaves are shown from plants treated with water or with SNP are shown in upper and bottom panels, respectively. (B) Necrosis, monitored using Trypan blue staining, in leaf tissue of WT and dnd1 plants treated with water (-SNP) or 100 µM SNP (+SNP). Dead cells become blue after staining. This experiment was repeated three times. (C) Lipid peroxidation (MDA level) in leaf tissue of WT and dnd1 plants treated with water (-SNP) (dark bars) or 100 µM SNP (+SNP) (light bars). Results are presented as means (n = 3) + SE. ANOVA evaluation of means separation between -SNP and +SNP treatments indicated a significant difference for dnd1 leaves (p < 0.05; indicated with a '*' above bar representing the '+ SNP' treatment) and no significant difference for WT leaves. (D) Effect of (50 µM) SNP on SAG transcript accumulation in dnd1 leaves was analyzed by semiquantitative RT-PCR. This experiment was repeated three times.