Molecular steps in the immune signaling pathway evoked by plant elicitor peptides: Ca2+-dependent protein kinases, nitric oxide, and reactive oxygen species are downstream from the early Ca2+ signal.

Endogenous plant elicitor peptides (Peps) can act to facilitate immune signaling and pathogen defense responses. Binding of these peptides to the Arabidopsis (Arabidopsis thaliana) plasma membrane-localized Pep receptors (PEPRs) leads to cytosolic Ca(2+) elevation, an early event in a signaling cascade that activates immune responses. This immune response includes the amplification of signaling evoked by direct perception of pathogen-associated molecular patterns by plant cells under assault. Work included in this report further characterizes the Pep immune response and identifies new molecular steps in the signal transduction cascade. The PEPR coreceptor BRASSINOSTEROID-INSENSITIVE1 Associated Kinase1 contributes to generation of the Pep-activated Ca(2+) signal and leads to increased defense gene expression and resistance to a virulent bacterial pathogen. Ca(2+)-dependent protein kinases (CPKs) decode the Ca(2+) signal, also facilitating defense gene expression and enhanced resistance to the pathogen. Nitric oxide and reduced nicotinamide adenine dinucleotide phosphate oxidase-dependent reactive oxygen species generation (due to the function of Respiratory Burst Oxidase Homolog proteins D and F) are also involved downstream from the Ca(2+) signal in the Pep immune defense signal transduction cascade, as is the case with BRASSINOSTEROID-INSENSITIVE1 Associated Kinase1 and CPK5, CPK6, and CPK11. These steps of the pathogen defense response are required for maximal Pep immune activation that limits growth of a virulent bacterial pathogen in the plant. We find a synergism between function of the PEPR and Flagellin Sensing2 receptors in terms of both nitric oxide and reactive oxygen species generation. Presented results are also consistent with the involvement of the secondary messenger cyclic GMP and a cyclic GMP-activated Ca(2+)-conducting channel in the Pep immune signaling pathway.

Ca 2+ elevation within the cell cytosol is an early and critical step in the signal transduction cascades that lead to plant innate immunity against infectious pathogens (Lecourieux et al., 2006). One level of immune defense responses is initiated upon plant cell perception of evolutionarily conserved components of microbial invaders, pathogen (or microbe)-associated molecular patterns (PAMPs). Binding of PAMPs to plant cell pattern recognition receptors (PRRs) typically initiates a signaling cascade that evokes basal immune responses in plants. A well-characterized PAMP:receptor protein system in plants is the bacterial motility organ protein flagellin (typically studied by use of the active epitope peptide flg22) and its cognate PRR Flagellin Sensing2 (FLS2; Aslam et al., 2008Aslam et al., , 2009Jeworutzki et al., 2010).
Another component of immune responses evoked by perception of PAMPs such as flg22 is the increased expression of some members of a family of Plant Elicitor Peptides (Pep1-Pep6). These peptide signals are referred to as elicitor peptides because they evoke defense responses in plants (Yamaguchi and Huffaker, 2011). They are also commonly included with such endogenous plant compounds as oligogalacturonides (structural components of plant cell walls that are released by plant cells under physical assault by pathogenic microbes) as plant compounds that are generated/ released during the process of pathogenic infection and involved in activating defense programs (Lecourieux et al., 2006). Oligogalacturonides and Peps are collectively referred to as damage-associated molecular patterns (DAMPs; Boller and Felix, 2009). However, Peps are not necessarily only present at a site of infection due to host cell necrosis; they may be present at the site of infection without having been released from damaged cells. Thus, reference to Peps as DAMPs could be an incorrect categorization. However, because here we present an analysis of Pep peptide signaling within the context of what is known about PAMP signaling, we refer to these endogenous plant elicitors as DAMPs to distinguish them here as plant-derived signals that bind to PRRs and initiate defense signaling cascades.
Peps present initially at the site of infection, along with those released from damaged cells as well as subsequently synthesized peptides, and act (in a fashion similar to PAMPs) as extracellular signals that induce immune responses upon binding to plasma membranelocalized PRRs (Pep Receptors or PEPRs [Yamaguchi and Huffaker, 2011]). Pep binding to the PEPR1 or PEPR2 receptor activates a cyclic GMP (cGMP)-dependent Ca 2+ -conducting ion channel (cyclic nucleotide gated channel2 [CNGC2]), leading to pathogen-associated cytosolic Ca 2+ elevation (Qi et al., 2010;Ma et al., 2012). Some evidence suggests that PEPRs have a cytosolic guanylyl cyclase (GC) domain that is responsible for cGMP production (Kwezi et al., 2007(Kwezi et al., , 2011. Recent work from this lab supported this hypothesis; affinity-purified recombinant PEPR1-GC domain generated cGMP in vitro (Qi et al., 2010). Ma et al. (2012) extended this work by demonstrating that exogenous application of Pep1, a ligand for the PEPRs, led to in vivo cytosolic cGMP production in Arabidopsis (Arabidopsis thaliana) plants expressing a fluorescence indicator of cGMP generation. Furthermore, mutation of functional residues in the PEPR1-GC domain abolished Ca 2+ -dependent immune signaling activated by the Pep ligand.
Ca 2+ -dependent protein kinases (CPKs) are Ca 2+ sensors that trigger downstream responses as components of plant immune defense systems (Lee and Rudd, 2002;Boudsocq and Sheen, 2013). In Arabidopsis, the cpk5 cpk6 cpk11 triple null mutant shows decreased flg22-induced gene expression, decreased flg22induced ROS generation, and impairment of plant resistance to pathogens compared with the wild type and single CPK mutants (Boudsocq et al., 2010). CPKs regulate ROS production through phosphorylation of Respiratory Burst Oxidase Homolog proteins RBOHD and RBOHF (Boudsocq and Sheen, 2013); RBOH proteins are components of an NADPH oxidase that generates superoxide and concomitantly H 2 O 2 . Both PAMPs and DAMPs can induce ROS generation (Qi et al., 2010;Schulze et al., 2010). The absence of RBOHD alone or both RBOHD and RBOHF does not influence DAMP-or PAMP-triggered early-stage membrane polarization and cytosolic Ca 2+ elevation . Thus, it can be concluded that RBOH-dependent ROS generation is downstream from Ca 2+ signaling during pathogen defense responses (Segonzac and Zipfel, 2011). It should be noted, though, that Ranf et al. (2011) concluded that there were some feedback effects of ROS on Ca 2+ signaling during pathogen responses. PAMP-triggered defense responses are impaired in rbohD/rbohF double mutant plants (Torres et al., 2002(Torres et al., , 2005, suggesting that Ca 2+ -dependent ROS generation plays an important role in the signaling leading to innate immunity. Not much is currently known about plant innate immune cascade steps that link RBOH-dependent ROS production to defense gene activation and enhanced immunity. Nitric oxide (NO) has been identified as a defense signal in plant disease resistance. An NO scavenger diminishes hypersensitive response (HR) to avirulent Pseudomonas syringae pv tomato (Pst), promoting bacterial growth and disease progression (Delledonne et al., 1998). Durner et al. (1998) showed that NO treatment enhances cGMP levels and plant defense gene expression; these effects are impaired by treatment of plants either with an NO scavenger or an inhibitor of NO synthase (NOS) thought to be responsible for NO generation during immune responses in plants. More recently, inoculation of Pst was shown to elevate leaf NO levels . Application of lipopolysaccharide (LPS), a PAMP that is a cell wall component of gram-positive bacteria, can trigger NO generation in Arabidopsis guard cells (Ali et al., 2007). Impairment of PAMP or avirulent pathogen-mediated NO generation in these studies occurred in the defense no death1 (dnd1; a mutant allele of CNGC2) and cml24-4 (loss of function of calmodulin [CaM]-like [CML] protein) mutants due to defective Ca 2+ transport and NOS activation, respectively. Although NO has been identified as a plant immune defense signal for over a decade (Durner et al., 1998), its induction and function in plant immune responses downstream from flg22 or DAMPs such as Pep is still unknown.
We recently proposed a model for the molecular steps linking perception of an extracellular pathogen-related DAMP (see discussion above about this term as it relates to Peps) ligand (Pep peptides) to Ca 2+ -mediated immune responses (Ma et al., 2012). This prior study focused on the PEPR receptors. Here, we expand the characterization of the Pep/PEPR signaling pathway. In this study, we used changes in mitogen-activated protein kinase (MPK3) and transcription factor WRKY33 transcript levels as read-outs to (1) further examine the contribution of the PEPR-GC domain to the signaling cascade, (2) evaluate the contribution of other proteins to the generation of the Ca 2+ signal, and (3) delineate signaling steps downstream from Pep-induced Ca 2+ elevation that mediate immune responses in Arabidopsis. We conclude that Pep-dependent gene expression is induced by this elicitor signal due to cGMP production (possibly by the PEPR receptor), a coreceptor may contribute to PEPR defense responses, and CPKs, NO production, and ROS (H 2 O 2 ) generation contribute to downstream gene expression in this immune response signaling cascade.

RESULTS
PAMP-dependent activation of MPKs and WRKY transcription factors have been studied extensively (Wan et al., 2008;Park et al., 2012;Rasmussen et al., 2012); both of these classes of signal transduction proteins are involved in downstream responses to cytosolic Ca 2+ elevation induced by PAMPs. Similarly, Pep peptides can trigger cytosolic Ca 2+ elevations in Arabidopsis wild-type plants (Qi et al., 2010;Ranf et al., 2011), leading to quick transcriptional activation or protein phosphorylation of downstream MPKs and WRKYs (Qi et al., 2010;Yamaguchi et al., 2010;Ranf et al., 2011;Ma et al., 2012). To understand what molecules are involved in the Pep-associated immune responses downstream from the critical early signal of cytosolic Ca 2+ elevation, we investigated the alteration of MPK3 and WRKY33 transcript levels in response to Pep perception using a range of loss-of-function Arabidopsis mutants and by application of different pharmacological treatments to wild-type plants.

CNGC2 and BAK1 Contribute to Pep-Induced Ca 2+ Generation and Downstream Signaling
All studies of Pep signaling included in this report used the Pep3 peptide (of the six known Pep peptides in Arabidopsis; Huffaker et al., 2006). Both Pep1 and Pep3 evoke pathogen defense responses (Huffaker et al., 2006;Huffaker and Ryan, 2007;Qi et al., 2010;Yamaguchi et al., 2010;Ma et al., 2012). However, Pep3 expression is increased by exposure of Arabidopsis plants to pathogens (and specifically P. syringae) as well as by PAMPs (including flg22), while Pep1 expression is not affected by these treatments (Huffaker et al., 2006). Prior studies from this lab also focused primarily on Pep3 as an activating ligand for study of immune signaling (Qi et al., 2010;Ma et al., 2012). In this prior work, Pep3-dependent cytosolic Ca 2+ elevation was strongly impaired in the dnd1 mutant, which lacks a functional CNGC2 channel. Here, we extend the study of CNGC2 involvement in Pep signaling leading to defense gene expression. As shown in Figure 1A, the Pep-dependent increase in expression of MPK3 found in wild-type seedlings is greatly reduced in the dnd1 mutant. This result indicates that CNGC2dependent Ca 2+ signaling is important for Pep activation of defense genes.
BRASSINOSTEROID-INSENSITIVE1 Associated Kinase1 (BAK1) is a coreceptor for both FLS2 and BRASSINOSTEROID-INSENSITIVE1, which are involved in flagellin and brassinosteroid signaling, respectively (Russinova et al., 2004;Heese et al., 2007). BAK1 was recently found to physically interact with PEPRs (PEPR1 and PEPR2) in vitro . In vivo evidence indicates that BAK1 and PEPRs can phosphorylate each other after Pep1 treatment (Schulze et al., 2010). Thus, BAK1 may function as a coreceptor with PEPRs and contribute to Pep immune signaling leading to defense gene expression. Consistent with this possibility, we found that, similar to the dnd1 mutant, Pepdependent MPK3 expression is substantially impaired in the bak1-4 null (Chinchilla et al., 2007) mutant (Fig. 1A). We find that BAK1 contributes to the generation of a Ca 2+ signal in response to both the PAMP flg22 (Fig. 1B) and Pep (Fig. 1C); however, Pep-induced Ca 2+ elevation is only modestly affected in the bak1 mutant Figure 1. BAK1 and CNGC2 contribute to Pep signaling. A, Pep3dependent MPK3 expression in wild-type, bak1, and dnd1 seedlings (see "Materials and Methods" for concentrations of all ligands and inhibitors used for gene expression studies). Asterisks above a bar indicate that Pep-dependent MPK3 expression in the mutant genotype was significantly different (P , 0.05) than the level in wild-type seedlings. Similar statistical analyses are shown for results in Figures 3, 4, and 5. B, Change in cytosolic Ca 2+ in wild-type-aeq and bak1-aeq leaves induced by flg22 (1 mM). C, Change in cytosolic Ca 2+ in wildtype-aeq and bak1-aeq leaves induced by Pep3 (20 nM). For B and C, cytosolic Ca 2+ level was recorded before and after ligand addition at time 0 and used to calculate the change in cytosolic Ca 2+ from the minimum level recorded at or near time 0. Results are presented as means 6 SE (replicate no. in parentheses) calculated every min. Additional statistical analyses of means separation (ANOVA using Student's t tests) were undertaken on data recorded between 2 and 4 min. Asterisks indicate significant difference (P , 0.05) between the wild type and mutants. The null mutant bak1-4 plants were used for the experiment shown in A. The bak1-3 mutant allele was used to generate aeq-expressing plants; therefore, the experiments shown in B and C used this genotype. See "Materials and Methods" for information regarding the bak1-3 and bak1-4 alleles. (Ranf et al., 2011). Nonetheless, Pep-dependent MPK3 expression is substantially inhibited in the bak1 mutant (Fig. 1A). We conclude from the results shown in Figure 1 that both BAK1 and CNGC2 contribute to Pep-induced Ca 2+ elevation and therefore are required for maximal downstream immune signaling by the elicitor peptide.
H 2 O 2 and NO Production Is Impaired in fls2 and pepr1 Mutants NO and ROS defense molecules including H 2 O 2 are generated downstream from Ca 2+ in immune signaling cascades (Levine et al., 1994;Delledonne et al., 1998;Ali et al., 2007;Ma et al., 2008Ma et al., , 2009Moreau et al., 2010;Torres, 2010), and H 2 O 2 generation is required for flg22mediated immune responses (Torres, 2010). Furthermore, chelation of apoplastic Ca 2+ in Arabidopsis leaves by the extracellular polysaccharide alginate derived from P. syringae strongly inhibits flg22-dependent H 2 O 2 generation (Aslam et al., 2008), indicating that flg22induced H 2 O 2 generation is Ca 2+ dependent. In prior studies from this lab using leaf tissue, an interdependence of flg22 and Pep signaling was noted; maximal signaling by either ligand required the presence of receptors for both the PAMP (FLS2) and the Pep elicitor (PEPR1/PEPR2; Ma et al., 2012). Included in this work were studies demonstrating that, when monitored on a leaf tissue basis, the maximal extent of flg22-dependent ROS generation required functional PEPR1/PEPR2. We speculated that this interdependence of PAMP and DAMP signaling might be due to an amplification of the different signals as perception of pathogen infection is spread from cell to cell, i.e. an enhancement of a 'warning alarm' initiated by PAMP perception through the activation of Pep signaling. Pep1-dependent ROS generation has been demonstrated on a whole-leaf (Huffaker et al., 2006) and leaf tissue (Roux et al., 2011) basis, and furthermore, Huffaker et al. (2006) have shown that preventing ROS generation impairs Pep1dependent defense gene expression in detached, intact leaves. (As noted above, we used Pep3 in all studies.) In the work reported here, we used a different experimental system to study interdependent flg22 and Pep immune responses. We also evaluated Pep effects on the defense signal NO for the first time. Ligand (either flg22 or Pep) effects on H 2 O 2 and NO generation were monitored in individual leaf cells of wild-type, fls2 null mutant, or pepr1 null mutant seedlings (Fig. 2). Early immune signaling events such as membrane depolarization (associated with cytosolic Ca 2+ elevation) induced by flg22 require the presence of a functional FLS2 receptor . Thus, it is not surprising that flg22-dependent H 2 O 2 generation was inhibited in cells from fls2 plants compared with the wild type ( Fig. 2A). However, maximal H 2 O 2 generation in response to Pep also required the presence of both PEPR1 and FLS2 receptors in these single-cell assays (Fig. 2B). This interdependence of PAMP flg22 and DAMP Pep signaling on their respective receptors was Figure 2. FLS2 and PEPR1 receptor involvement in flg22-and Pepdependent NO and H 2 O 2 generation in leaf guard cells. Ligands (either Pep3 or flg22) were added to reaction buffer containing epidermal peels that had been previously incubated in NO-and H 2 O 2 -specific fluorescent dyes. Peels were prepared from wild-type, fls2, and pepr1 plants. A, Flg22-dependent H 2 O 2 generation. B, Pep-dependent H 2 O 2 generation. C, Flg22-dependent NO generation. D, Pep-dependent NO generation. Mean maximal fluorescence of guard cells (occurring approximately 5-10 min after ligand addition) in a peel was ascertained exactly as described previously (Ma et al., 2009) and averaged to generate one biological replicate. Signals shown (as bars) are means (6SE) of three replicate peels prepared from leaves of different plants; fluorescence of approximately six guard cells in an epidermal peel was averaged for one replicate. The maximum fluorescence intensity that could be recorded from the peels was 256 fluorescence units per pixel. ANOVA analysis was used to evaluate means separation between the values recorded from wild-type tissue and values recorded from peels prepared from mutant genotype plants. Significant differences (at P , 0.05) between fluorescence recorded from a mutant genotype and the wild type are indicated by an asterisk above the bar representing the mean value recorded from peels obtained from mutant plants. In addition to the quantitative analyses (A-D), fluorescence images recorded from epidermal peels loaded with NO-specific (E) and H 2 O 2 -specific (F) dyes are also shown. Fluorescence recorded from individual guard cell pairs in epidermal peels prepared from wild-type, fls2, and pepr1 plants are shown; images are representative of data used to generate the quantitative analyses. also evidenced with regard to NO generation (Fig. 2,C and D).
NO has been referred to as the concert master of the HR and plant innate immunity (Dangl, 1998), and PAMP-dependent NO generation requires the critical and early Ca 2+ elevation in the signaling pathway (Delledonne et al., 1998;Ali et al., 2007). Thus, the results with H 2 O 2 and NO shown in Figure 2 suggest that the functional interaction between FLS2 and PEPR signaling with regard to cytosolic Ca 2+ elevation impacts downstream steps of the immune response. The interdependent immune responses we identify here occur when signaling is monitored in individual cells (Fig. 2) and assayed minutes after addition of ligand. This suggests that the functional interaction between the PAMP (flg22) and DAMP elicitor peptide (Pep) signals and the cognate receptors is not just occurring due to amplification of defense responses within the leaf as signals are spread from cell to cell or due to increased expression of PROPEPs upon application of flg22 over several hours (Huffaker and Ryan, 2007). The mechanistic basis for this interdependence of PAMP and plant DAMP signaling at the cellular level is not delineated here; however, it is also not without precedent. Human innate immune responses have been also recently found to include the synergistic amplification of flagellin signaling by an endogenous DAMP peptide; the signaling pathways downstream from the cell membrane flagellin and LL-37 peptide receptors include steps that interact and amplify the immune responses evoked by the individual ligands (Nijnik et al., 2012).

Immune Pathway Steps That Affect Pep-Induced Ca 2+ -Dependent Signaling
Protein phosphorylation and MPK cascades are involved in PAMP signaling Park et al., 2012;Rasmussen et al., 2012); little is known at present about involvement of protein phosphorylation in DAMP signaling (Schulze et al., 2010). Boudsocq et al. (2010) identified some specific CPK isoforms that act as Ca 2+ sensors in plant immune signaling cascades and demonstrated that the cpk5 cpk6 cpk11 triple mutant showed compromised immune response to flg22. In their detailed and thorough analysis of CPK involvement in flg22 immune signaling, Boudsocq et al. (2010) also demonstrated that flg22 application to Arabidopsis plants increased endogenous CPK activity and that flg22-dependent increases in defense gene expression were impaired in the cpk5 cpk6 cpk11 triple mutant and by the kinase inhibitor K252a. Here, we find that Pep immune signaling is also compromised in this genotype, as evaluated by monitoring MPK3 and WRKY33 expression (Figs. 3A and 4A). These results suggest that CPK-dependent phosphorylation is involved in the Pep signaling cascade. We also evaluated the contribution of protein phosphorylation as a step in the Pep/ PEPR immune signaling pathway by investigating the effect of the kinase inhibitor K252a on Pep-dependent MPK3 and WRKY33 expression (Figs. 3B and 4B). Results with the kinase inhibitor are consistent with the effect of the cpk5 cpk6 cpk11 mutant on Pep signaling; protein phosphorylation is clearly involved in Pep/ PEPR immune signaling in a fashion similar to that linked to flg22 signaling.
It has been proposed that NO and H 2 O 2 are involved in many common signaling pathways (for review, see Moreau et al., 2010), and both are involved in pathogen defense signaling as delineated earlier. Both signaling molecules are also generated in response to flg22 and Pep (Fig. 2), suggesting their involvement in Pep/PEPR immune signaling. Results of our gene expression analyses support this contention. NO Associated1 is a gene associated with NO generation; noa1 null mutants . Genotype and treatment effects on Pep3-dependent MPK3 expression. A, MPK3 expression was monitored in wild-type seedlings (with and without exposure to Pep3 for 30 min) and in seedlings with mutations affecting CPK signaling (cpk5 cpk6 cpk11), NO generation (noa1), or ROS production (rbohD/rbohF). Results shown are means 6 SE (n = 4) of transcript levels normalized to the level in wild-type seedlings in the absence of Pep3. Asterisks indicate significant differences (P , 0.05) between the wild type and mutants with Pep treatment. B, MPK3 expression in wild-type seedlings treated with protein kinase inhibitor (k252a), an NO scavenger (2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide [cPTIO]), or a GC inhibitor (LY83583). Seedlings were pretreated with these compounds prior to the addition of Pep3 for 30 min. MPK3 expression level was normalized to that in the control without inhibitor or Pep added. Results shown are means of three biological replications 6 SE. Asterisks indicate significant differences (P , 0.05) between the levels of expression with inhibitor (and Pep3 added) compared with the control (with Pep added). exhibit less NO production than wild-type plants and lack some NO-dependent stress responses (Guo et al., 2003;Moreau et al., 2008). NO has been recognized to be involved in plant disease resistance for over a decade; this mobile and reactive molecule regulates gene expression in both biotic and abiotic stress signaling pathways (Guo et al., 2003;Moreau et al., 2010). We find a similar level of inhibition (compared with wild-type plants) of Pep-dependent MPK3 and WRKY33 expression in the noa1 mutant as observed in the cpk5 cpk6 cpk11 mutant (Figs. 3A and 4A). We also find that application of the NO scavenger cPTIO inhibits Pepdependent defense gene expression (Figs. 3B and 4B). This pharmacological approach supports the work with the NO mutant; NO generated from Pep binding to its receptor (Fig. 2) is likely an intermediary step in the downstream signaling pathway of this peptide elicitor.
Binding of Pep to its receptor leads to ROS generation as well ( Fig. 2; Ma et al., 2012). Pep-evoked ROS generation occurs downstream from cytosolic Ca 2+ elevation in this signaling cascade. We used the rbohD/rbohF double mutant to evaluate whether Pep-dependent ROS generation is an important step in the immune signaling initiated by the plant elicitor peptide. A significant reduction of Pep-dependent MPK3 and WRKY33 expression compared with the wild type was observed in this mutant (Figs. 3A and 4A). The level of Pep-dependent WRKY33 expression in the rbohD/rbohF, noa1, and cpk5 cpk6 cpk11 mutants were similar in the experiment shown in Figure 4A.
The primary goal of our studies was to delineate steps in the Pep/PEPR signaling pathway. However, we note that both Pep and flg22 evoke NO and ROS generation in a manner that is interdependent on their respective PRR receptors (Fig. 2). We did some further work within this context to investigate if some of the steps involved in the Pep signaling pathway were similarly involved in flg22-dependent gene expression. As noted above, the involvement of CPK5, CPK6, and CPK11 in flg22 signaling leading to defense gene expression is well characterized (Boudsocq et al., 2010). The involvement of NO and ROS in the signaling pathway leading to flg22dependent defense gene expression is less clarified. Dubiella et al. (2013) recently showed that flg22dependent ROS generation is impaired in the rbohD single mutant, but localized (at the site of treatment) defense gene expression was similar in flg22-treated leaves of wild-type and rbohD mutant plants. These authors concluded that RBOHD-dependent ROS generation was not involved in localized activation of defense genes but rather acted in cell-to-cell communication to activate defense genes in tissues distal to the site of PAMP perception. Macho et al. (2012) also found no effect (at the site of PAMP application) of RBOHD null mutation on flg22-dependent defense gene expression. Here, we extended these studies by examining flg22-dependent WRKY33 and MPK3 expression in the rbohD/rbohF double mutant (Fig. 5). In contrast to a marked difference in Pep-dependent gene expression in the wild type and rbohD/rbohF mutants, flg22 signaling, at least in terms of MPK3 and WRKY33 expression, is not reduced from the wild type in the double mutant.
The involvement of NO in PAMP signaling beyond the studies mentioned above linking LPS application to NO generation is not clarified. Mur et al. (2008) provided indirect evidence that NO is involved in signaling associated with HR response to avirulent pathogens but does not influence some aspects of PAMP-triggered immunity (associated with flg22 application) evaluated in their study. Here, we find that flg22-dependent gene expression was impaired in the noa1 mutant (Fig. 5)   . Genotype and treatment effects on Pep3-dependent WRKY33 expression. A, WRKY33 expression was monitored in wildtype seedlings (with and without exposure to Pep3 for 30 min) and in seedlings with mutations affecting CPK signaling (cpk5 cpk6 cpk11), NO generation (noa1), ROS production (rbohD/rbohF), or CNGC2 channel function (dnd1). Results shown are means 6 SE (n = 4) of transcript levels normalized to the level in wild-type seedlings in the absence of Pep. Asterisks indicate significant differences (P , 0.05) between the wild type and mutants with Pep treatment. B, WRKY33 expression in wild-type seedlings treated with protein kinase inhibitor (k252a), an NO scavenger (cPTIO), or a GC inhibitor (LY83583). Seedlings were pretreated with these compounds prior to addition of Pep3 for 30 min. WRKY33 expression level was normalized to that in the control without inhibitor or Pep added. Results shown are means of three biological replications 6 SE. Asterisks indicate significant differences (P , 0.05) between the levels of expression with inhibitor (and Pep added) compared with the control (with Pep added).

Proteins Involved in CNGC Activation Affect Defense Gene Expression
CaM physically interacts with CNGCs; the CaM binding site on plant CNGCs is in a cytosolic region of the channels and overlaps with the cyclic nucleotidebinding domain (Köhler et al., 1999;Köhler and Neuhaus, 2000;Hua et al., 2003;Ali et al., 2006;Abdel-Hamid et al., 2010). CaM binding diminishes cyclic nucleotidedependent activation of CNGCs (Hua et al., 2003) and reduces CNGC-dependent Ca 2+ influx into cells (Ali et al., 2006). Exposure of Arabidopsis protoplasts to W7, a CaM antagonist, can result in spontaneous CNGC current (Ali et al., 2007;Ma et al., 2008). Prior exposure to W7 increased the cytosolic Ca 2+ elevation associated with inoculation of Arabidopsis leaves with Pst harboring the avirulence gene AvrRpt2 , and prior exposure to W7 increased the PAMP (LPS)-dependent inward current through Ca 2+ -conducting channels in patch-clamp studies of Arabidopsis protoplasts (Ali et al., 2007). These W7 effects were attributed to the prevention of CaM blockage of CNGC channel currents. In the presence of a CaM antagonist, CNGC channels might activate at basal levels of cytosolic cyclic nucleotides (i.e. in the absence of a signal that would cause elevation of these secondary messenger molecules). Consistent with these previous observations, incubation of seedlings with W7 alone significantly increased both MPK3 and WRKY33 transcript levels (Fig. 6) in the absence of added Pep. Interpretation of the results with W7 shown in Figure 6 should be made with a degree of caution as well as within a context due to the following points. W7 is not a specific inhibitor of CaM, but rather also binds to and inhibits other proteins with CML domains such as CPKs (although the concentrations required to fully inhibit CPKs are substantially greater than we used here; Syam Prakash and Jayabaskaran, 2006). Of course, CPKs are involved in Pep signaling (Figs. 3 and 4). Thus, W7 effects on CPKs should be considered when extrapolating from the effects of W7 (added in the absence of Pep) on Pepresponsive genes such as WRKY33 and MPK3 shown in Figure 6. However, genetic and pharmacological inhibition of CPK activity reduces Pep-dependent expression of pathogen defense genes (Figs. 3 and 4). Therefore, the increases in WRKY33 and MPK3 transcript levels caused by W7 shown in Figure 6 are not consistent with these responses being mediated by W7 effects on CPKs. The effects of W7 shown here are consistent with the aforementioned effects of W7 on CaM, CNGCs, and Ca 2+ currents associated with pathogen signaling mentioned above. Nonetheless, interpretation of effects of inhibitors should always be made with due caution regarding their nonspecificity. In the specific case of W7, general nonspecific effects of this compound can be checked by comparing effects with the W7 analog N-(6-aminohexyl)-1-naphthalene sulphonamide (W5; Syam Prakash and Jayabaskaran, 2006). W5 has reduced affinity for CaM but does mimic the nonspecific effects of W7 on biological systems, such as interaction with membranes. In further experiments, we found that W5 (at 50 mM, the concentration at which W7 was used) had no significant effect on transcript levels of either MPK3 or WRKY33 (data not shown). These results are consistent with the possibility that W7 acts to increase expression of Pep-responsive pathogen defense genes by mimicking the effect of Pep on CNGCs.  ) on WRKY33 and MPK3 expression. Wild-type seedlings were treated with W7 or IBMX for 30 min prior to measurement of WRKY33 or MPK3 expression. Results shown are means 6 SE (n = 3) of transcript levels normalized to the level in wild-type seedlings without treatment. Asterisks indicate significant differences (P , 0.05) between the control (water) and treatment with W7 or IBMX. (Li et al., 1994), (2) altering PDE activity in plant cells through IBMX application increases signaling dependent on cytosolic levels of cyclic nucleotide (Li et al., 1994;Volotovski et al., 1998;Ma et al., 2009), (3) the resting state level of cyclic nucleotides in plants cells is influenced by PDE activity (Kurosaki and Kaburaki, 1995), and (4) alterations in the level of cyclic nucleotides in plants during signaling cascades are associated with alterations in PDE activity (Cooke et al., 1994;Assmann, 1995). This body of research suggests that the turnover of cyclic nucleotides in plant cells is rapid enough to be influenced by effects of IBMX application on PDE activity and that application of a PDE inhibitor such as IBMX could affect basal levels of cyclic nucleotides and initiate signaling downstream from this secondary messenger molecule.

IBMX is a cyclic nucleotide phosphodiesterase (PDE) inhibitor that prevents the breakdown of cyclic nucleotides in plant and animal cells (Ma et al., 2009). A number of prior studies have shown that (1) PDE activity is relatively high in leaf mesophyll cells of plants
In a fashion similar to W7, exposure of seedlings to IBMX increased WRKY33 expression in the absence of the activating Pep ligand (Fig. 6), possibly due to the increased cAMP or cGMP levels in the cytosol. The results shown in Figure 6 suggest that the activation of CNGCs (by blocking CaM effects or the inhibition of cyclic nucleotide hydrolysis) increases defense gene expression. Conversely, the inhibition of cyclic nucleotide generation should attenuate Pep-induced defense gene expression. Exposure of seedlings to the guanylyl cyclase inhibitor LY83583 impaired Pep-dependent increases in MPK3 and WRKY33 expression (Figs. 3B and 4B). These results with the CaM antagonist and IBMX in the absence of Pep and the GC inhibitor in the presence of Pep are consistent with the genetic evidence in Figure 1 (with the dnd1 mutant) and suggest that CNGCs and cyclic nucleotides may play important roles in the Pep/PEPR signaling pathway leading to defense gene activation.

Pep-Triggered Immunity against Pst DC3000 Is Compromised in cpk and bak1 Mutants
Prior exposure of plants to the PAMP flg22 or Pep can lead to an immune response; subsequent bacterial inoculation of plants pretreated with either peptide leads to a reduction (compared with untreated plants) in growth of the virulent pathogen Pst DC3000 (Zipfel et al., 2004;Ma et al., 2008;Boudsocq et al., 2010;Yamaguchi et al., 2010;Ma et al., 2012). Here, we extend the characterization of the immune responses triggered by the elicitor peptide Pep signal to determine if some of the steps involved in Pep-dependent increases in expression of MPK3 and WRKY33 are critical to the ultimate end point of the signal transduction cascade, i.e. the development of an immune response. Compared with wild-type plants, the cpk5 cpk6 cpk11 and bak1 mutants demonstrate reduced defense gene expression upon Pep treatment (Figs. 1A, 3A, and 4A), suggesting that the immunization response might be compromised by the impairment of the Pep signal transduction cascade in these plants. We examined this hypothesis in the experiment shown in Figure 7A by monitoring growth of Pst DC3000 in these mutant genotypes. Application of Pep led to an immunization response in wild-type plants; Pst DC3000 bacteria recovered from Pep-treated plants 3 d after inoculation was reduced by 75% compared with water-treated wild-type plants (Fig. 7A). The extent to which Pep pretreatment restricted bacterial growth was impaired in the cpk5 cpk6 cpk11 and bak1 mutants compared with the level of immunization displayed by wild-type Figure 7. Evaluation of Pep3-induced immunity to a virulent pathogen in leaves of wild-type plants and in genotypes that have mutations in genes that affect Pep-dependent expression of pathogen defense genes. A, Proliferation of (virulent) Pst DC3000 on leaves of wild-type, cpk5 cpk6 cpk11, and bak1-4 mutant plants pretreated with water (control) or Pep3. Plants were pretreated 1 d prior to inoculation with Pst, and bacterial growth was evaluated 3 d after inoculation. Results shown (note log scale of ordinate axis) are mean values of Pst recovered from leaves (n = 4) 6 SE on day 3. Measurement of bacteria recovered from leaves on day 0 (immediately after inoculation) indicated that there was no significant difference between any of the treatments (data not shown). B, In another experiment, effect of Pep3 pretreatment on suppression of Pst growth in rbohD/rbohF and noa1 plants was compared with the effect of Pep on Pst growth in wild-type plants. Experimental design and analysis were similar to the work shown in A. plants (Fig. 7A). In a subsequent experiment (Fig. 7B) comparing effects of Pep immunization in wild-type plants with the rbohD/rbohF and noa1 mutants, we observed similar trends. In this experiment (Fig. 7B), Pep pretreatment reduced bacterial levels in wild-type plants by 94%. There is no difference between the water and Pep treatments in rbohD/rbohF and noa1 mutant plants.

DISCUSSION
The desired physiological outcome of immune signaling in the face of a pathogenic challenge is an enhancement of host system fitness (as opposed to the fate of individual cells at the infection site that may or may not survive a skirmish; they could undergo programmed cell death to limit infection spread). Studies included in this report identified some specific steps involved in the Arabidopsis Peps signal transduction cascade and demonstrated in several cases (BAK1, CPK, NO, and ROS) that these molecules contribute to the positive effect the signaling pathway has on plant immunity. As indicated earlier, many of the steps in the Pep signal transduction cascade downstream from generation of the early Ca 2+ signal are also involved in the evocation of defense programs upon perception of PAMPs such as flagellin. We show here that, on an individual cell basis, there is an interdependence of flg22 and Pep signaling; maximal generation of NO and ROS The work presented in this report expands upon prior studies of plant immune responses to support a signaling cascade involving the following steps. a, Pep peptides bind to the extracellular leucine-rich repeat domain of the PEPR receptors; BAK1 contributes to Pep/PEPR signaling, probably as a coreceptor. b, There is a functional interaction between PEPR and FLS2 facilitated by an unknown mechanism (indicated by a question mark); BAK1 is also a coreceptor for FLS2. c, Pep perception by PEPR results in an increase in cytosolic cGMP, possibly due to a putative guanylyl cyclase activity of PEPR. d, Pep-dependent rise in cytosolic cGMP activates the cell membrane Ca 2+ -conducting CNGC2 channel. e, CNGC2-mediated cytosolic Ca 2+ elevation results in Ca 2+ binding to CPKs, CaM, and/or CML proteins. f, Ca 2+ activation of CPKs results in increased expression of defense genes and activation of RBOHD and RBOHF, leading to the generation of ROS, including superoxide and H 2 O 2 . g, ROS can act as an antimicrobial compound or cell-to-cell signal and also has effects on defense gene expression. h, Ca 2+ /CaM or a CML protein can lead to NO generation by an unidentified NOS-type enzyme . NO can act as a cell-to-cell signal and also modulate defense gene expression. i, Ca 2+ /CaM or a CML protein can feedback on CNGC currents and prevent further activation by cyclic nucleotides (Fig. 5; Hua et al., 2003). j, Flagellin/FLS2 interaction leads to immune responses separately from Pep/PEPR signaling (but as noted above, there is a functional interaction between the two pathways). Experimental results included in this report probed steps of the Pep/ PEPR immune signaling cascade using the following pharmacological agents and Arabidopsis genotypes. 1, The dnd1 mutant lacks a fulllength CNGC2 polypeptide, which impairs Ca 2+ influx into the cell. 2, The bak1-3 mutant was used to evaluate BAK1 involvement in flg22and Pep-dependent Ca 2+ elevation in aeq-expressing plants. The bak1-4 mutant was used to evaluate BAK1 involvement in Pep-dependent gene expression and bacterial growth suppression. Mutation of BAK1 compromised Pep-dependent Ca 2+ elevation, gene expression, and immunization against growth of virulent bacteria. See "Materials and Methods" for information about bak1-3 and bak1-4 phenotypes. Effects of bak1-3 and bak1-4 mutations on flg22-dependent gene expression and immune signaling were reported by Ranf et al. (2011). 3, Null mutation of FLS2 affected both flg22-and Pep-dependent ROS and NO generation. 4, Null mutation of PEPR1 impaired both pep-and flg22-dependent ROS and NO generation. 5, Pep-dependent gene expression was impaired in the cpk5 cpk6 cpk11 triple null mutant (and by kinase inhibitors). Null mutation of these CPK genes also compromised the ability of Pep to immunize the plant against growth of virulent bacteria. The cpk5 cpk6 cpk11 triple null mutant also affects flg22 defense signaling. 6, Pep-dependent defense gene expression and immunization against virulent bacterial growth was compromised in the noa1 mutant, which lacks the ability to generate NO in response to external stimuli. Defense gene expression downstream from flg22 perception was affected by the noa1 mutation in a fashion similar to effects on Pep/PEPR signaling. 7, Pep-dependent defense gene expression and Pep-induced immunization of the plant against growth of virulent bacteria were compromised in the rbohD/rbohF double mutant. In contrast to Pep/PEPR signaling, the rbohD/rbohF double mutation did not affect flg22-dependent defense gene expression. 8, Application of W7 leads to increased expression of pathogen defense gene expression without Pep application; this could be due to W7 antagonism of CaM function, leading to increased CNGC currents at ambient levels of cyclic nucleotide. 9, Application of the GC inhibitor LY83538 impairs Pep-dependent defense gene expression. 10, Application of the cyclic nucleotide PDE inhibitor IBMX results in increased expression of WRKY33 without Pep application; this might be due to an increase in ambient cell cGMP because breakdown of the cyclic nucleotide is impaired. 11, Application of the NO scavenger cPTIO impairs Pep-dependent defense gene expression. 12, Application of the kinase inhibitor K252a impaired Pep-dependent defense gene expression, possibly due to inhibition of CPKs or other kinases involved in the signaling pathway. The kinase inhibitor K252a also blocks flg22 signaling (Boudsocq et al., 2010;Jeworutzki et al., 2010). by either peptide requires both cognate receptors. The molecular basis for this is not delineated. However, this interdependence of PAMP and DAMP signaling on a cellular level also occurs in animal cells. In the case of at least some animal DAMP peptides and flagellin (as discussed above), the cell signaling cascades responding to the perception of these peptides share many molecular steps and act synergistically (Nijnik et al., 2012). Our work with the double NADPH peroxidase mutant rbohD/rbohF notes a difference in the involvement of these ROS-generating proteins in Pep/PEPR and flagellin/FLS2 signaling leading to defense gene expression.
Results in this report indicate that BAK1 and the Ca 2+ channel CNGC2 contribute to Pep-dependent generation of the Ca 2+ signal. However, the impact of BAK1 loss of function on Pep-dependent cytosolic Ca 2+ elevation is modest, perhaps less severe than corresponding effects on flg22-dependent Ca 2+ elevation (Fig. 1). Nonetheless, BAK1 loss of function did impair the immune response stimulated by application of exogenous Pep (Fig. 7).
Within the context of prior work in the area of innate immune signaling associated with cytosolic Ca 2+ elevation upon pathogen perception, we conclude that CPKs are involved in transmitting the early Ca 2+ signal to downstream steps such as defense gene expression. Use of a kinase inhibitor supported our conclusions based on use of the cpk5 cpk6 cpk11 mutant. However, it should be noted that some evidence suggests that protein phosphorylation might also be involved in early steps of flagellin and Pep peptide signaling, in addition to the downstream role of CPKs in the signal transduction pathway Schulze et al., 2010). The kinase inhibitor studies here would not discriminate between these different signaling steps.
We speculate, based on numerous prior studies with these pharmacological agents as well as work included in this report, that the cyclic nucleotide PDE inhibitor IBMX, the CaM antagonist W7, and the GC inhibitor LY83583 affected the Pep signaling pathway upstream from Pep-induced cytosolic Ca 2+ elevation. The results of these studies suggested the following trend: inhibition of cGMP generation (LY83583) impaired the Pep signaling pathway, and agents that possibly raise cytosolic cGMP (IBMX) or might allow activation of a Ca 2+ channel in the absence of cGMP elevation (W7) mimic the effect of Pep and initiate the signaling pathway in the absence of added exogenous ligand. (We note, however, that application of IBMX resulted in a significant increase in WRKY33 expression but did not increase MPK3 expression [ Fig. 6].) The work presented in this report focused on innate immune signaling downstream from Pep3 (referred to here, as well as in the literature, as a DAMP), a PAMP (flg22), and a virulent pathogen that, when inoculated onto plants, evokes defense pathways dependent on the PAMP and DAMP.
In summary, the work presented in this report extends our knowledge of the signal transduction cascade evoked by Peps. These peptides, and molecules involved in the signal transduction pathway evoked by these signals, contribute to defense responses of plants to microbial pathogens. The Pep/PEPR pathogen defense signal transduction pathway developed from the work presented in this report, as well as from prior studies from this lab and others, is represented in the model presented in Figure 8.

Plant Materials
In addition to wild-type Columbia plants, the following mutant genotypes (all previously characterized) were used in this work: the cpk5 cpk6 cpk11 triple mutant (Boudsocq et al., 2010), noa1 (Durner et al., 1998), rbohD/rbohF double mutant (Torres et al., 2002), dnd1 (Ali et al., 2007, and two alleles of bak1, bak1-3 (Chinchilla et al., 2007) andbak1-4 (Gou et al., 2012). The bak1-3 genotype was originally characterized as lacking expression of any BAK1 mRNA (Chinchilla et al., 2007). We used the bak1-3 genotype to generate bak1 mutant plants expressing the Ca 2+ reporter protein apoaequorin (see below). However, recent studies (Gou et al., 2012) have presented a more nuanced analysis of this BAK1 allele; although the transcript level of BAK1 in not detectable in bak1-3, it is a leaky mutant, and in the bak1-like1 genetic background, bak1-3 does not display null phenotypes. Gou et al. (2012) presented an analysis of the bak1-4 allele and confirmed that it is a true bak1 null mutant. We therefore used bak1-4 mutant plants for subsequent experiments. We conclude that the use of both bak1 alleles in our work is not a problem based on the following point. Ranf et al. (2011) demonstrated that the level of cytosolic Ca 2+ elevation in response to both flg22 and Pep is similar (and not significantly different [P , 0.05]) in leaves of bak1-3 and bak1-4 plants.
Seeds of Arabidopsis (Arabidopsis thaliana) genotypes were surface sterilized and spread on petri dishes containing one-half-strength Murashige and Skoog medium (Caisson), 2.6 mM MES (adjusted to pH 5.7 with Tris), 1% (w/v) Suc, and 0.8% (w/v) agar (Ma et al., 2012). These seeds were germinated and grown in a growth chamber for 1 week with a day (80-100 mmol m -2 s -1 illumination)/ night cycle of 12 h/12 h at 25°C. The seedlings were then transferred to pots containing LP5 potting mix with starter fertilizer (Sun Gro) and were grown in a growth chamber at 16-h light (100 mmol m -2 s -1 illumination)/8-h dark (72% relative humidity) and 22°C (see below for liquid-grown seedlings used for gene expression analysis). Seeds were incubated at 4°C in the dark for 2 d prior to use to facilitate stratification and increase germination. During growth, plants were irrigated with Jack's Professional Peat-Lite 20:10:20 (N-P 2 O 5 -K 2 O) solution (at 2 g L -1 ) one to two times to provide supplementary fertilizer.

RNA Isolation and Complementary DNA Synthesis
Growth of seedlings in one-half-strength strength Murashige and Skoog liquid medium was performed according to Yamaguchi et al. (2010). Liquidgrown 10-d-old wild-type and mutant genotype seedlings (one seedling grown in a separate tube with 3 mL medium was used as a genotype and treatment replicate) were used for experiments. Wild-type seedlings were treated with 50 mM W7 (Enzo Life Sciences; for use of W7, see Ali et al., 2007) or 1 mM IBMX (Enzo Life Sciences; for use of IBMX, see Ma et al., 2009) in the absence of Pep for 5 min prior to collection for analysis of gene expression. In other experiments, seedlings were treated with 2 mM k252a (Sigma; Zottini et al., 2007), 500 mM (cPTIO (Enzo Life Sciences; Zottini et al., 2007), 20 nM LY83583 (Enzo Life Sciences), or water as a control for 5 min with shaking at 180 rpm, followed by addition of either 20 nm Pep (this concentration was used for all experiments) or water for a further 30 min with shaking prior to collection. For comparison of gene expression in wild-type and mutant genotypes, seedlings were treated for 30 min with 20 nM Pep (or water as a control) prior to collection. After treatments, seedlings were collected and frozen immediately in liquid nitrogen and processed for RNA extraction or stored at -80°C for future use. Total RNA from a whole liquid-grown seedling was isolated using the Plant RNA Extraction Kit (Macherey-Nagel) according to the manufacturer's manual.
After preparation, 500 ng of total RNA was used for reverse transcription using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) following the manufacturer's protocol. The synthesized complementary DNA was made into a 1:5 dilution, and 1 mL of the diluted complementary DNA was used for the subsequent quantitative real-time PCR.

Cytosolic Ca 2+ Measurements
This method was adapted from Qi et al. (2010) with slight modifications (for complete details of how aequorin [aeq]-expressing lines were generated, leaves were handled, and how luminometer recordings were used to calculate cytosolic Ca 2+ levels, see Qi et al., 2010). For the work reported here, we used a line of bak1-3 mutant plants expressing aeq by crossing the mutant with a wild-type line expressing aeq. The T3 progeny of the cross were checked for absence of a full-length BAK1 gene and optimized for aeq expression using methods described previously (Qi et al., 2010).
Pep3 (20 nM) and flg22 (1 mM) ligand effects on cytosolic Ca 2+ levels were evaluated in whole detached leaves of 3-to 4-week-old plants expressing cytosol-localized Ca 2+ -dependent chemiluminescent apoaequorin protein reconstituted with coelenterazine-cp. The concentration of flg22 used in our studies matched that typically used to compare the effects of this PAMP with other PAMPs (Ranf et al., 2011). Flg22 concentrations as low as 0.1 nM have been shown to cause cytosolic Ca 2+ elevations in Arabidopsis (Aslam et al., 2009;Jeworutzki et al., 2010). However, Aslam et al. (2009) found that increasing the concentration of flg22 even up to 0.5 mM (higher levels were not checked in this study) caused increasingly greater Ca 2+ elevations, supporting the logic of using such a high concentration of ligand. We used 20 nM Pep3 here because prior work from this lab (Qi et al., 2010) indicated that repeatable, PEPR1-dependent elevations in cytosolic Ca 2+ could be obtained from Arabidopsis leaves with this concentration of the elicitor. Other studies of Pep signaling used similar concentrations of activating ligand . Although the concentrations of activating ligands used here suggest different relative potencies, we did not investigate this point in our work.
Ligands were added at time 0, and signals shown are averages of biological replicates from different plants. A single leaf was cut at the petiole for each treatment replicate. Detached leaves were vacuum infiltrated with 10 mM coelenterazine-cp in 500 mL control buffer (in a 2-mL centrifuge tube) containing 1 mM KCl, 1 mM CaCl 2 $6H 2 O, and 10 mM MgCl 2 $6H 2 O for 10 s and kept in darkness at room temperature for 1 to 2 h to allow coelenterazine incorporation into the leaves. For luminescence recordings, the centrifuge tube was placed into a luminometer (TD-20/20; Turner Designs) and kept in the dark for 2 to 3 min for recovery from handling and dissipation of touchinduced Ca 2+ spikes; ligands were added to leaves only after they displayed a stable level of luminescence. Ligands were added (at 23 final concentration) in 500 mL of control buffer. Discharge of the remaining aeq after luminescence measurements was done by adding 800 mL 30% ethanol containing 2 M CaCl 2 $6H 2 O. SE was calculated every minute.

NO and H 2 O 2 Production in Guard Cells
The method for NO detection in guard cells using the NO-specific fluorescent dye diaminofluoresceins-2 diacetate (Invitrogen) was adapted from Ali et al. (2007). The method for H 2 O 2 detection in guard cells using the H 2 O 2 -specific fluorescent dye 29,79-dichlorodihydrofluorescein diacetate (Invitrogen) was adapted from Gerber et al. (2004). Rosette leaves of 4-to 6-week-old wild-type, fls2, and pepr1 plants were detached from plants and used to make 2-mm-wide epidermal peels. Epidermal peels were incubated in buffer A (10 mM KCl, 25 mM MES-KOH, and 0.1 mM CaCl 2 $6H 2 O, at pH 6.15) for 1 to 2 h. For assays of NO generation, epidermal peels were transferred to 3 mL of buffer A containing 50 mM diaminofluoresceins-2 diacetate and incubated for 30 min to load cells with NO-specific dye. For assays of H 2 O 2 generation, peels were alternatively placed in 3 mL buffer A, which contained 50 mM 29,79-dichlorodihydrofluorescein diacetate for 30 min. After loading NO-and H 2 O 2 -specific dye in the tissue, peels were washed three times in buffer A and incubated in 3 mL of buffer A (for H 2 O 2 assays) or buffer B (for NO assays) containing flg22 (1 mM) or Pep (20 nM) for 5 to 10 min. The epidermal peels were placed underneath a cover slip on a microscope slide with several drops of buffer (A or B as appropriate, with ligands added). Both NO-and H 2 O 2 -dependent fluorescence were monitored over time.
Mean maximal fluorescence of guard cells (occurring approximately 5 to 10 min after ligand addition) in a peel was ascertained exactly as described previously (Ma et al., 2009). Fluorescence and bright-field images were captured using an inverted Olympus IX70 microscope and green fluorescence protein excitation and emission filters. Digitized images were acquired using a MagnaFire CCD camera and software. Maximal fluorescence measurements of approximately six guard cells in an epidermal peel were averaged for one replicate. Representative guard cell fluorescence images recorded from leaf epidermal strips prepared from wild-type, fls2, and pepr1 plants are shown in Figure 2. Captured images at the maximum fluorescence intensity for treatment replicates were used to calculate data represented as means. Quantitative analyses of the NO-and H 2 O 2 -dependent fluorescence in guard cell pairs were undertaken using ImageJ software as described in Ali et al. (2007). The digitized image showing maximum fluorescence for guard cell pairs from an epidermal peel represented a genotype replicate; a minimum of three epidermal peels were analyzed for each genotype.

Growth of Virulent Pseudomonas syringae in Plants
The following method for preexposure of plants to Pep and subsequent inoculation of leaves with virulent pathogen was adapted from Yamaguchi et al. (2010). Pep (or water as a control) was syringe injected into (attached) rosette leaves from wild-type, cpk5 cpk6 cpk11, bak1, rbohD/rbohF, or noa1 plants 24 h prior to bacterial inoculation. Peptides were delivered to the leaves from a blunt-end 1-mL syringe in approximately 0.5 mL water. Pseudomonas syringae pv tomato DC3000 was grown at 28°C on low-salt (1 g L -1 NaCl) Luria-Bertani (Fisher Scientific) 2% (w/v) agar medium containing 100 mg L -1 rifampicin for about 2 d. Colonies from these plates were used for liquid overnight cultures (medium composition was the same as the plates except no agar was used). Bacteria were isolated from overnight liquid cultures and resuspended in sterile 10 mM MgCl 2 $6H 2 O to a concentration of 3 3 10 8 colony forming units mL -1 . The resuspended bacteria were applied (in 0.01% [v/v] Silwet L-77 surfactant) by spraying on leaves (previously treated with water or peptides) from 4-to 5-week-old plants. After spray inoculation, plants were covered with a plastic lid and placed in a growth chamber for 3 d. Leaf tissue was collected for measurement of bacterial titer on the day of inoculation (day 0) and 3 d post inoculation. For each biological replicate, 0.7-cm-diameter leaf discs were cut from leaves of two different plants and ground together in 100 mL of 10 mM MgCl 2 $6H 2 O in a 1.5-mL tube. The samples were thoroughly vortexed in 900 mL of water and diluted 1:10 serially. Then 10 mL of the samples (and serial dilutions) were spread on plates containing low-salt Luria-Bertani agar medium containing 100 mg L -1 rifampicin. Plates were left at 25°C for 2 d; after this incubation period, colonies on plates were counted.