Disruption of the Vacuolar Calcium-ATPases in Arabidopsis Results in the Activation of a Salicylic Acid-Dependent Programmed Cell Death Pathway

: Calcium (Ca 2+ ) signals regulate many aspects of plant development, including a programmed cell death pathway that protects plants from pathogens (Hypersensitive Response, HR). Cytosolic Ca 2+ -signals result from a combined action of Ca 2+ -influx through channels, and Ca 2+ efflux through pumps and cotransporters. Plants utilize calmodulin activated Ca 2+ pumps (ACA, Autoinhibited Ca 2+ -ATPase) at the plasma membrane (PM), endoplasmic reticulum (ER) and vacuole. Here we show that a double knockout mutation of the vacuolar Ca 2+ pumps ACA4 and ACA11 in Arabidopsis results in a high frequency of HR-like lesions. The appearance of macrolesions could be suppressed by growing plants with increased levels (> 15 mM) of various anions, providing a method for conditional suppression. By removing plants from a conditional suppression, lesion initials were found to originate primarily in leaf mesophyll cells, as detected by aniline blue staining. Initiation and spread of lesions could also be suppressed by disrupting the production or accumulation of salicylic acid (SA), as shown by combining aca4/11 mutations with a sid2 mutation or expression of a SA degradation enzyme (NahG). This indicates that the loss of the vacuolar calcium pumps by itself does not cause a catastrophic defect in ion homeostasis, but rather potentiates the activation of a SA-dependent PCD pathway. Together these results provide evidence linking the activity of the vacuolar Ca 2+ pumps to the control of a SA-dependent programmed cell death pathway in plants.

Evidence that ACAs can modulate biotic and abiotic stress response pathways has recently been obtained from experiments with moss and tobacco. In the moss Physcomitrella patens, a knockout of a gene encoding the vacuolar ACA (PCA1) resulted in an increased sensitivity to a NaCl stress, which was correlated with a NaCl triggered cytosolic Ca 2+ elevation that was higher in magnitude and longer in duration (Qudeimat et al., 2008). In tobacco, RNAi silencing of NbCA1 resulted in an accelerated pathogen triggered PCD response (Zhu et al., 2010).
An ER location for NbCA1 was proposed based on transient expression of a GFP tagged pump.
The NbCA1-RNAi plants also showed elicitor triggered Ca 2+ signals that were higher in magnitude and longer in duration. These two examples confirm that ACAs can function to modulate the dynamics of Ca 2+ signals triggered by multiple environmental signals, as expected for a Ca 2+ /calmodulin activated Ca 2+ pump.
In controls, WT plants before and after transfer from anion supplemented conditions showed the same low frequency of micro-lesions (callose staining) and low background levels of ROS ( Fig. 4). By contrast, aca4/11 plants even before a transfer showed patches of elevated levels of ROS (Fig. 4A, t = 0h, and Fig. SI1) as well as a detectably higher frequency of microlesions ( Fig. 4B). This indicates that while the anion supplement prevented any macrolesion expansion, it only partially inhibited lesion initiation. However, by 54 hours after transfer from suppression conditions, the surface area covered by patches of ROS had increased more than 2-fold ( Fig 4A). In addition, the number of lesions increased 2.5 fold (Fig. 4B), with a typical lesion increasing in surface area by more than 4-fold from 48 to 72 hours ( Fig 4C). In most cases, lesions grew to occupy the entire leaf surface within 3 to 4 additional days. Thus, this analysis indicates that a period between 30h-54h after transfer from lesion suppression conditions provides the earliest time at which a transition from micro to macro-lesion formation can be visualized.
To identify the cell types in which lesions originate, we mapped the locations of singlecell sized callose deposits (micro-lesions) during a 72 hour lesion induction experiment ( Fig. 5 and Fig. SI2). Under nutritionally suppressed conditions, the micro-lesions detected in aca4/11 appeared evenly distributed among the internal tissues of the leaf (parenchyma, mesophyll and vessels), whereas none were observed at the epidermis (t=0 in Fig 4b, Fig. 5). Within 40h after transfer to lesion permissive conditions, the number of micro-lesions increased, primarily in locations corresponding to mesophyll cells. In a distribution analysis of 117 lesions, 85 microlesions were classified as a mesophyll origin, 25 as parenchyma, and only 7 as epidermal.
However, since this staining assay was destructive, it did not allow us to observe individual micro-lesions as they developed into macro-lesions. Nevertheless, this distribution analysis provides strong evidence that lesions in aca4/11 preferentially initiate in mesophyll cells.
Within the group of lesion mimic mutants (LMMs), some of them such as vad1 ( aca4/11, the distribution of lesion-initials was evaluated relative to vessels. A lesion was considered vessel-associated if was directly adjacent or within one cell layer (~25 μm). Between 18h and 30h after lesion induction only 12% of the initials were located in the vicinity of a vessel ( Fig SI3). This suggests that micro-lesions initiate independently from a potential signal spreading through the vascular system. Together, this lesion mapping study suggest that lesion initials arise predominately in mesophyll cells due to a stimulus that is intrinsic to the region surrounding the initial. For total Ca 2+ levels, there were no significant differences between WT and aca4/11, although both sets of plants showed a 25% increase in Ca 2+ when transferred from nutrient suppressed to lesion permissive conditions. Therefore our results indicate that ACA4 and 11 are not required for leaves to achieve normal Ca 2+ storage levels. This is consistent with a hypothesis that Ca 2+ /proton exchangers (CAX) rather than Ca 2+ pumps have a primary role in Ca 2+ loading into plant vacuoles (Hirschi, 1999;Kim et al., 2006).
Cland NO 3 levels decrease more rapidly in aca4/11 when switched to lesion triggering conditions-The relative concentrations were also determined for each of the 3 anions used here to suppress aca4/11 lesions (Fig. 7). In plants grown using a 15 mM KCl supplement for lesion suppression (Fig. 7, left panels), chloride content in leaves of both WT and aca4-1/11-1 were elevated nearly 10-fold compared to plants grown with a standard hydroponic solution. This elevated level was maintained in wild type plants during the first 30 hours after transfer to our standard hydroponic conditions. In contrast, the mutants showed a 32% loss of Clduring this same period. A similar pattern of anion loss by the mutant was observed for plants transferred from a condition of nutrient suppression using a 15 mM NH 4 NO 3 supplement (Fig. 7, middle panels).
When plants were transferred from supplemented to standard hydroponic solution, only the mutant showed a relatively rapid decrease (22% ) in NO 3 during this first 30h period.
In contrast to suppression by KCl and NH 4 NO 3 , suppression by 15 mM KH 2 PO 4 was not accompanied by any detectable changes in free concentration of the corresponding anion (i.e., PO 4 3-, see Fig. 7, right panels). However, it is important to note that our assay was limited to measuring the free concentration of PO 4 3-, and did not account for other forms of P. Since free PO 4 3levels are expected to be tightly regulated, any difference between the mutant and wild type may have been masked by a rapid homeostasis mechanism that converts PO 4 3to other forms, such as phytate (Loewus & Murthy, 2000).
Despite the inherent difficulty in accounting for the fate of free PO 4 3during these suppression/induction experiments, the relatively rapid loss of NO 3 and Clin mutants upon moving plants to lesion triggering conditions indicates that homeostasis controls for at least some anions are perturbed by the aca4/11 mutations.
SA Signaling is activated in aca4/11 mutant-In plants, SA can function as a signaling molecule to trigger defense responses, including a programmed cell death pathway (Lorrain et al., 2003). To determine if the lesions associated to aca4-1/11-1 involved an SA signal, we examined aca4/11 mutants harboring a sid2-5 mutation that disrupts SA biosynthesis (Nawrath & Métraux, 1999), as well as a NahG transgene that encodes an enzyme that increases the degradation rate of SA (Gaffney et al., 1993;Delaney et al., 1994). Both strategies resulted in suppression of lesions To confirm that endogenous SA levels were up-regulated in aca4/11 mutants, SA was measured in plants before and 30h after moving hydroponically grown plants to lesion inducing conditions. At both time points, mutants showed a 2-fold higher level of SA compared to wild type ( Fig. SI5). It is noteworthy that the SA levels were not significantly reduced when growing plants under suppressed conditions with high anion supplements (see Fig. SI5, t=0h). This suggests that the small 2-fold increase in SA is by itself not sufficient for lesion formation, but requires other signaling functions that can somehow be suppressed by factors related to an increase in nutritional supplements.
Pathogen defense responses occur more quickly in aca4/11 mutants -Infection by the bacterial pathogen Pseudomonas syringae pv Tomato DC3000 was used as a system to monitor a pathogen response in aca4/11 plants. The response was evaluated by measuring bacterial growth ( Figure 9A and B), as well as the expression of a defense-related marker gene PR1 ( Fig. 9C) (Uknes et al., 1992). These experiments were done under lesion suppression conditions to avoid having any pre-existing lesions that could potential alter a pathogen attack.
Under lesion-suppression conditions, the SA-inducible PR1 gene showed no detectable expression in any of the plants lines tested ( Figure 9C, t = 0h). Nevertheless, when lesion suppression conditions were removed, and aca4/11 mutants were allowed to develop their SAdependent lesions, an up-regulation of the PR1 marker gene was observed (data not shown).
Although our lesion suppression conditions prevented the formation of spontaneous SAdependent lesions, as well as the up-regulation of an SA-triggered pathogen defense marker gene (e.g. PR1), the actual defense response to a Pseudomonas syringae pathogen attack was significantly faster and more effective in the aca4/11 mutant, as indicated by lower bacterial growth at 2 and 3 days post inoculation (Fig. 9A), as well as a more rapid induction of a PR1 marker gene (by at least 12 hours) (Fig. 9C). This accelerated defense response was dependent upon SA, as shown using the sid2-5 allele to block SA biosynthesis By including the sid2-5 mutation with aca4-3 and 11-5 (Fig. 9B), the aca4/11 dependent inhibition of bacterial growth was reversed, and the faster pathogen-triggered up-regulation of the PR1 gene was abolished (Fig.   9C). A visual indication that an aca4/11 knockout accelerated the defense response was also confirmed by the more rapid development of HR-lesions, which were clearly visible in the aca4/11 mutant at 54 hours post inoculation, but not yet apparent in the WT control ( Figure 9D).
These pathogen-triggered lesions were morphologically indistinguishable from the spontaneous lesions originally documented as the characteristic feature of the aca4/11 lesion mimic phenotype (See Fig. 2 Nevertheless, additional studies will be required to determine if the vacuolar TPC in Arabidopsis can function alone or in conjunction with other putative Ca 2+ channels to trigger an aca4/11dependent PCD pathway. The aca4/11 mutant can be classified as having a "lesion mimic phenotype" (LMM), since the lesions have features consistent with a classical hypersensitive response (HR), but can initiate in a sterile environment without a pathogen trigger (data not shown). LMMs are often classified as either lesion-initiation or lesion-spreading mutants (Lorrain et al., 2003). The aca4/11 mutation is unusual since both initiation and lesion spreading appear to be enhanced.
When lesion suppression conditions were removed (Fig. 4), lesion initiation appeared to increase, and macro-lesions grew rapidly to cover most of the leaf surface within a week.
Since HR-lesions are considered to be a form of AL-PCD, the aca4/11 lesions can also be classified as a form of AL-PCD. An easily visualized feature of HR-lesions is an increase in callose synthesis at lesion initials. This requires a reprogramming of the cellular machinery, and was observed as a feature of aca4/11 lesions ( Fig. 4 and 5). This supports the contention that aca4/11-lesions develop as part of a PCD, as opposed to a spontaneous and rapid cellular necrosis.
Propagation of aca4/11-lesions involves an SA-dependent PCD pathway-Two genetic lines of evidence indicate that aca4/11-dependent lesions propagate through a salicylic acid (SA) dependent programmed cell death pathway ( Fig. 7 and SI5). First, lesions were suppressed by a sid2 mutation. The sid2-5 mutation used here disrupts the isochorismate synthase gene ICS1 (Wildermuth et al., 2001) and blocks the production of SA (Nawrath & Métraux, 1999 14 lesions were suppressed by expression of a NahG transgene. NahG encodes a bacterial salicylate hydroxylase that degrades SA into catechol (Gaffney et al., 1993;Delaney et al., 1994). This ability to block SA signaling and suppress aca4/11 lesions confirms that lesion development results from a defect in regulating a specific PCD signal transduction pathway, as opposed to an uncontrolled cell death resulting from catastrophic defect in Ca 2+ homeostasis or vacuolar degeneration.
Lesion spread can be suppressed by anion supplements-Interestingly, growth conditions were found that could separate lesion initiation from its uncontrolled spreading (Fig. 3). When aniline blue (Fig. 4, 5, and SI2). However, these lesions did not spread, indicating that the anion supplements functioned primarily to suppress a second distinct phase of lesion development (i.e. spreading). While the mechanism underlying anion suppression is not clear, suppression by NO 3 and Cldid correlate with an increase in their concentrations in rosette leaves, followed by a more rapid loss compared to wild type when transferred to unsupplemented growth conditions (Fig. 7). This supports a model in which the ionic environment at the site of lesion initiation and propagation can be altered to regulate a PCD pathway, potentially through changing ion conductance properties of either the PM or vacuole.
Multiple studies have implicated non-specific anion transporters in membrane depolarization events associated with many ion signaling pathways, including Ca 2+ signals and PCD (Ward et al., 1995;Errakhi et al., 2008b, a). For example, an anion efflux in tobacco leaf suspension cells was observed as an early response to a fungal elicitor cryptogein (Pugin et al., 1997). A pharmacological inhibition of this anion release was also observed to prevent the development of an HR in tobacco leaves (Wendehenne et al., 2002).
Using plants that were transferred from anion suppression to lesion inducing conditions, the location of spontaneous lesion initials was found to be predominately in mesophyll cells, without any correlation to being near or far from vascular elements ( Fig. 5 and SI4). Since microlesions were never seen to appear in cell types of the root (i.e. no aniline blue stained necrotic lesions), it is possible that lesion initiation and propagation are related to physiological triggers associated with photosynthetic pathways, as implicated in several examples of PCD triggered by abiotic stress (Gadjev et al., 2008).
It is noteworthy that ROS and SA levels in both anion-suppressed and non-suppressed plants were approximately 2-fold higher than controls (see Fig. 4 and SI5). Since the anion supplements did not block ROS and SA production, but did prevent the accumulation of PR1 ( 9, t = 0h), this suggests that the mechanism for anion suppression is at a point downstream of an initial signaling pathway that generates increased levels of SA or ROS, and upstream of changes in a transcriptional response that upregulates PR1 mRNA levels A loss of aca4 and 11 potentiates an accelerated defense response to P. syringae-An enhanced defense response against a bacterial pathogen, Pseudomonas syringae, was observed here for aca4/11 mutants (Fig. 9A). The defense response was tested under conditions in which spontaneous lesions in the aca4/11 mutants were suppressed by anion supplements. The initial expression levels for an SA-up-regulated PR1 marker were undetectable in both mutants and WT controls under these conditions, although aca4/11 mutants already showed a moderated elevation in SA (Fig. SI5). However, following a pathogen inoculation, the aca4/11 mutants showed a more rapid induction of the PR1 gene, with significant expression within 12 hours ( Fig 9C). The enhanced resistance and more rapid induction of a PR1 gene marker were both SA-dependent, as indicated using a sid2 mutation to block SA production ( Fig. 9B and C). These results suggest that even under lesion suppressed conditions, the loss of aca4/11 results in a physiologically altered plant that is pre-conditioned to a more rapid defense response, and therefore confirms that ACA4 and 11 act as suppressors of a PCD pathway. The observation here that a loss of ACA4 and ACA11 increases the frequency of SAdependent lesions is significant because it supports a new model in which the vacuole participates in modulating certain Ca 2+ signals that can trigger PCD. Future research will be needed to visualize the Ca 2+ signals that are altered by the loss of ACA4 and 11, and to understand the "upstream" factors that initiate those signals, and the immediate "downstream" targets that link these signals to the activation of PCD. While it is known that Ca 2+ efflux through ACAs can be turned on and off (e.g. Hwang et al., 2000), it remains to be determined if this activity is actually regulated as part of a natural mechanism by which a pathogen or abiotic stress might trigger the activation of a PCD pathway.

Evidence for Ca 2+ signals in regulating PCD-
PCD is also involved in many other aspects of plant development, including senescence, sculpting tissues, and the terminal differentiation of tracheids (Jones, 2001;Lam, 2004 and debris pelleted by 10min centrifugation at 10,000g. DNA in the supernatant was recovered by 66% ispopropanol precipitation. Touch-down PCR (from 66°C to 60°C in -0,3°C steps, and then 14 additional cycles with an annealing temperature of 60°C) were performed in a 25 μ l reaction using ExTAQ DNA polymerase (Takara, Japan) following the manufacturer's protocol.
Oligonucleotides used for the reaction, at a final concentration of 0.2 μ M, can be found in the legend of Fig. 1.

Inoculation of plants with P. syringae-The virulent Pseudomonas syringae pv tomato DC3000
were grown at 28°C on King B's medium (40 g/l Proteose Peptone 3; 20 g/l Glycerin; 10 ml/l MgSO 4 (10% m/v); 10 ml/l K 2 HPO 4 (10% m/v)) supplemented with the appropriate antibiotics: 50 mg/mL of rifampicin. To examine the growth of the bacteria, 3-to-4-week-old plants were sprayed with a bacterial suspension containing 5.10 6 colony-forming units per mL in 10 mM MgCl 2 solution with 0.04% Silwet L-77. Bacterial growth was measured at 0,2 and 3 days after infiltration by extracting bacteria from leaf discs (0.6 cm2 discs per leaf) and plating a series of dilutions on the medium supplemented with appropriate antibiotics.
For the ACA11p::ACA11-GFP construct (ps#1657), a 2,127-bp sequence upstream of the ATG start codon for ACA11 was PCR-amplified from Arabidopsis and replaced the 35S promoter of 35S-ACA11-GFP (construction ps#1658). The DNA sequence of each construct is provided as a supplemental file (Fig SI6).
Northern Blot analysis-Total RNA was isolated from leaves using LiClphenol/chloroform extraction method (Chomczynski & Sacchi, 2006). Total RNA (10ug) was separated on 1.5 % agarose-formaldehyde gel and blotted onto nylon membranes. The membranes were hybridized with [α-32 P] dATP-labeled gene specific probes for 16 hr at 65°C and washed for 10 min twice with 2 X SSC (0,15 M NaCl, 15 mM trisodium-citrate), once with 1 X SSC,and10 min with 0.5 X SSC, 1%(W/V) SDS at 65°C. Detection of callose in leaves was performed essentially as described (Dietrich et al., 1994). The 2 largest leaves were fixed for 2h in 10% formaldehyde, 5% acetic acid, 45% ethanol, cleared for 2 min in boiling alcoholic lactophenol (95%ethanol:lactophenol, 2:1) and stained overnight in a solution of 150 mM K 2 HPO 4 , pH 9.5 with 0.01% aniline blue. Leaves were rinsed in distilled water before observation. Callose deposition was observed with an Olympus FV1000 confocal microscope with 405nm excitation and 440-480nm emission window. Six Z-sections (640μm x 640μm) spanning the whole leaf thickness were taken per leaf. Lesions surface and number were obtained from an analysis of every section in the stack. and processed by one of two procedures. Samples were either ground in liquid nitrogen and resuspended in 3ml chloroform, or frozen samples were ground directly in 3ml HPLC grade chloroform using a mixer mill (Retsch, Newtown, PA, USA). Samples were then incubated in 15ml polypropylene tubes for 1h at 50°C. Ultrapure water (5ml) was added and samples were incubated for an additional hour. Tubes were centrifuged for 15min at 2900g to clear debris from the aqueous phase. The supernatant was analyzed by anion-exchange chromatography. Aliquots of the supernatant (10μl) were run on a Dionex high-performance liquid chromatography (Dionex, USA) through a Dionex AS11-HC column with a gradient of 1 mM to 60 mM NaOH over 40 min.

Ion Concentration Measurements-
The column was at room temperature with a flow rate of 0.27 ml.min −1 . Anions were detected by suppressed conductivity method and NO 3 was specifically detected by absorbance at 210 nm.
Peaks were identified using pure anion salt standards purchased from Sigma.        Pictures are shown of aca4-1/11-1 with and without a sid2-5 mutation 10 days after transfer into hydroponic solution. A similar phenotype is observed with NahG expressed in aca4-3/11-5. Refer to figure 2A and B for comparison to aca4/11 mutants.   An SA dependent pathogen defense response is accelerated by an aca4/11 knockout.

Figure legends
Leaves of 4-week-old plants (wild-type Col-0, sid2, sid2/aca4/11 and aca4/11) grown under lesion-suppressed conditions (50 mM KH2PO4) were sprayed with suspensions of P. syringae DC3000 (OD600=0.01). A and B -Bacterial growth determinations were performed at the times indicated. Data points are the average of three replicate samples (±SD). C -Total RNA was isolated from leaves harvested at the indicated time points after bacterial inoculation. The panel section labelled PR1 shows an autoradiogram of a Northern blot probed for the defense related gene PR1. The section labelled rRNA shows a control for equal loading of RNA, as visualized by ethidium bromide staining of rRNAs. The Northern blot shown is representative of two independent experiments showing equivalent results. D -Representative leaves from the different plant lines assayed are shown 54h after bacterial inoculation. Note the fully developed lesions in the aca4/11 leaf, whereas the other leaves show only yellow chlorotic patches.