Perception of Damaged Self in Plants^,

Abstract

Plants use specific receptor proteins on the cell surface to detect host-derived danger signals released in response to attacks by pathogens or herbivores and activate immune responses against them.

Multicellular eukaryotes including plants and animals have evolved highly complex, multilayered immune systems to fight off microbial infections. How the immune systems function is a fundamental question for immunologists. The animal immune system was originally thought to function by distinguishing between “self “and “nonself ”(the Self–Nonself model; Burnet, 1959), and later between “infectious-nonself” and “noninfectious-self” (the Infectious–Nonself model; Janeway, 1989, 1992). In 1994, Matzinger proposed that the immune system is more concerned with “danger” than with “non-self” (the Danger model; Matzinger, 1994, 2002, 2007). The Danger model suggests that the immune system is activated by danger/alarm signals that are sent from both microbial pathogens and damaged host cells. In this model, it is assumed that healthy cells or cells undergoing normal physiological death do not produce danger signals (Matzinger, 2002). Over the years, the Danger model has been supported by the discovery of a large number of endogenous danger signals (Tang et al., 2012; Pouwels et al., 2014; Schaefer, 2014; Hernandez et al., 2016; Yatim et al., 2017; Dinarello, 2018).

Danger signals consist of conserved pathogen-associated molecular patterns (PAMPs) from the microbes and damage-associated molecular patterns (DAMPs) from injured host cells (Matzinger, 2002). Although the term “DAMPs” originally referred to the hydrophobic portions of biological molecules from dead and dying host and pathogen cells, which trigger immunity when exposed (Seong and Matzinger, 2004), it is now generally used to describe danger signals from damaged host cells (Martin, 2016; Yatim et al., 2017). Besides PAMPs and DAMPs, pathogen-derived effectors, which are proteins expressed by pathogens to aid infection of their hosts, and effector-caused perturbations on/in the host cells should also be considered as danger signals, though they were not included in the original model (Matzinger, 2002; Boller and Felix, 2009). PAMPs/DAMPs and extracellular effectors or their disturbances are generally recognized by germline-encoded cell-surface pattern recognition receptors (PRRs; Takeuchi and Akira, 2010), whereas intracellular effectors or their interruptions are often sensed by cytoplasmic nucleotide-binding oligomerization domain-like receptors (Chen et al., 2009).

The plant immune system shares a similar conceptual logic with the animal immune system, though plants lack adaptive immunity (Nürnberger et al., 2004; Haney et al., 2014). A simple coevolutionary model called the “Zigzag” was proposed to describe the molecular events in plant–microbe interactions (Jones and Dangl, 2006). Based on this model, plant cells employ PRRs to detect PAMPs, activating PAMP-triggered immunity (PTI), while adapted pathogens utilize effectors to dampen PTI. Plants in turn exploit nucleotide-binding oligomerization domain-like receptors to sense the presence of effectors, leading to effector-triggered immunity, which usually culminates in a hypersensitive cell death response at the infection site. Natural selection constantly drives the arms race between plants and pathogens, resulting in different levels of pathogen virulence and plant resistance (Jones and Dangl, 2006). The Zigzag model has conceptually stimulated enormous research in the plant–microbe interaction field; however, it did not encompass DAMPs. The recent “Invasion” model included DAMPs and introduced a new term, “invasion patterns,” which essentially refers to the same type of molecules as “danger signals” (Cook et al., 2015). It was suggested that adapting the Danger model for plants would allow the holistic concept of host immunity to be better shared by the entire community of immunologists (Gust et al., 2017). Nevertheless, neither the Zigzag model nor the Invasion model accommodates systemic resistance, including systemic acquired resistance (SAR) and induced systemic resistance (ISR), which are also essential parts of the plant immune system (Durrant and Dong, 2004; Pieterse et al., 2014). SAR and ISR are two forms of induced resistance wherein the plant immune system is primed by a prior localized infection that results in resistance throughout the plant against subsequent challenge by a broad spectrum of pathogens. However, induction of the two forms of systemic resistance is mechanistically distinct. SAR depends on the immune signal molecule salicylic acid (SA), whereas ISR relies on the signaling pathways activated by the plant hormones jasmonic acid (JA) and ethylene (ET; Durrant and Dong, 2004; Pieterse et al., 2014). The SA, JA, and ET response pathways serve as the backbone of the plant immune signaling network (Pieterse et al., 2012).

Compared to the large number of DAMPs that have been identified and characterized in animals, research on DAMPs in plants has only just begun (Rubartelli and Lotze, 2007; Choi and Klessig, 2016; Roh and Sohn, 2018). In the past several years, we have witnessed a marked increase in the number of potential DAMPs in plants, and the number is still growing (Table 1; Duran-Flores and Heil, 2016; Gust et al., 2017; Hou et al., 2019). Moreover, potential receptors for more than a dozen plant DAMPs have been identified (Gust et al., 2017; Hou et al., 2019). Characterization of these receptors is expected to significantly boost DAMP research in plants. While the DAMP field is blooming, the identity of DAMPs is under debate (Martin, 2016). It was argued in animals that a canonical DAMP should only be released from cells during necrosis; act through binding to cell-surface receptors; be upregulated, but not released, in response to PAMP detection or stress stimuli that are likely to presage necrosis; be synergistic with PAMPs in activating robust immune responses; and initiate relatively broad-acting responses in a manner similar to that shown by pathogen components (Martin, 2016). Based on these characteristics, members of the extended IL-1 cytokine family (IL-1α, IL-1β, IL-18, IL-33, IL-36α, IL-36β, and IL-36γ) have been reasoned to be the canonical DAMPs activating the immune system, whereas most other proposed DAMPs, e.g. ATP, uric acid, calreticulin, HMGB1, HSPs, and DNA fragments, likely act through liberating IL-1 family cytokines via promoting necrosis (Martin, 2016).

Table 1. Putative DAMPs proposed in plants.

Abbreviations: CAPE1, CYS-RICH SECRETORY PROTEINS, ANTIGEN5, AND PATHOGENESIS-RELATED1 PROTEINS (CAP)-DERIVED PEPTIDE1; eDNA, extracellular DNA; GmSubPep, G. max Subtilase Peptide; HMGB3, HIGH MOBILITY GROUP BOX3; HypSys, Hyp-rich Systemin; IDLp, INFLORESCENCE DEFICIENT IN ABSCISSION (IDA)-LIKE peptide; INR, INCEPTIN RECEPTOR; N/A, Not Applicable; RGS1, REGULATOR OF G-PROTEIN SIGNALING1; SCOOP12, SER-RICH ENDOGENOUS PEPTIDE12; Zip1, Z. mays immune signaling peptide1.

DAMP	Precursor	N-Terminal Signal Peptide	Degree of Polymerization, Amino Acids, or Bps	Receptor	Plant	References
Primary/Constitutive DAMPs
OGs	Pectin	N/A	10–15	WAK1	Arabidopsis	Côté and Hahn, 1994; He et al., 1996; Brutus et al., 2010
eATP	N/A	N/A	N/A	DORN1	Arabidopsis	Choi et al., 2014a; Roux, 2014
eNAD(P)	N/A	N/A	N/A	LecRK-I.8/VI.2	Arabidopsis	Zhang and Mou, 2009; Wang et al., 2017a, 2019a
Amino acids and glutathione	N/A	N/A	N/A	GLR3.3/3.6	Arabidopsis	Qi et al., 2006; Stephens et al., 2008; Li et al., 2013; Toyota et al., 2018
Extracellular sugars	N/A	N/A	N/A	RGS1	Arabidopsis	Johnston et al., 2007; Bolouri Moghaddam and van den Ende, 2012
HMGB3	N/A	No	N/A	Unknown	Arabidopsis	Choi et al., 2016
Cutin monomers	Cuticle	N/A	N/A	Unknown	Arabidopsis, tomato	Fauth et al., 1998; Buxdorf et al., 2014
Cellooligomers	Cellulose	N/A	2–7	Unknown	Arabidopsis	Souza et al., 2017; Johnson et al., 2018
Xyloglucans	Hemicellulose	N/A	7–9	Unknown	Common grape vine, Arabidopsis	Claverie et al., 2018
Methanol	Pectin	N/A	N/A	Unknown	Arabidopsis	Dixit et al., 2013; Hann et al., 2014; Tran et al., 2018
eDNA fragments	DNA	N/A	<700	Unknown	Pea (Pisum sativum), lima bean (Phaseolus lunatus), maize, common bean	Wen et al., 2009; Barbero et al., 2016; Duran-Flores and Heil, 2018
Secondary/Inducible DAMPs
Systemin	Prosystemin	No	18	SYR1/2	Some Solanaceae species	Pearce et al., 1991; Wang et al., 2018b
HypSys	ProHypSys	Yes	18–20	Unknown	Some Solanaceae species	Pearce et al., 2001a; Pearce, 2011
Peps	PROPEPs	No	23	PEPR1/2	Arabidopsis, maize	Huffaker et al., 2006, 2011, 2013; Yamaguchi et al., 2006, 2010
RALFs	RALF preproproteins	Yes	49	FER	Tobacco, Arabidopsis	Pearce et al., 2001b; Haruta et al., 2008, 2014; Stegmann et al., 2017
PSKs	PSK precursors	Yes	5	PSKR1/2	Asparagus, rice, Arabidopsis	Matsubayashi and Sakagami, 1996; Yang et al., 1999, 2001; Matsubayashi et al., 2002; Amano et al., 2007
PIP1/2	PrePIP1/2	Yes	13/15	RLK7	Arabidopsis	Hou et al., 2014
IDLp	IDLs	Yes	14	HAE/HSL2	Arabidopsis	Stenvik et al., 2008; Butenko et al., 2014; Patharkar et al., 2017; Vie et al., 2017; Wang et al., 2017b
GRIp	GRI	Yes	11	PRK5	Arabidopsis	Wrzaczek et al., 2009, 2015
CAPE1	PR1	Yes	11	Unknown	Tomato, Arabidopsis	Chen et al., 2014
Zip1	PROZIP1	No	17	Unknown	Maize	Ziemann et al., 2018
Inceptins	ATP synthase γ-subunit proteins	No	11	INR	Cowpea (Vigna unguiculata), maize	Schmelz et al., 2006; Steinbrenner et al., 2019
GmSubPep	Subtilase	Yes	12	Unknown	Soybean	Pearce et al., 2010a
GmPep914/890	GmPROPEP914/890	No	8	Unknown	Soybean	Yamaguchi et al., 2011
SCOOP12	PROSCOOP12	Yes	13	Unknown	Arabidopsis	Gully et al., 2019

In plants, the identity of DAMPs has not been vigorously debated. Recently, immunogenic plant factors were roughly divided into two categories, primary and secondary DAMPs, which correspond to constitutive and inducible DAMPs proposed in animals (Gust et al., 2017; Yatim et al., 2017). Primary/constitutive DAMPs are derived from pre-existing structures or molecules, including breakdown products of extracellular matrix and passively released intracellular molecules, while secondary/inducible DAMPs are actively processed and released upon tissue damage and other stimuli (Gust et al., 2017). Although this delineation of primary DAMPs is aligned with the original definition of DAMPs (Matzinger, 2002; Seong and Matzinger, 2004), it is worthwhile to compare the secondary DAMPs with the proposed canonical DAMPs in animals (Martin, 2016). One central argument for members of the extended IL-1 family being the canonical DAMPs in animals is that they do not possess N-terminal signal sequences and are released during necrosis (Martin, 2016). In contrast, precursors of most of the candidate peptide DAMPs in plants carry an N-terminal signal peptide (Table 1), suggesting active release via the conventional secretion pathway. They would, nevertheless, also be passively released upon cell damage during microbial infection and herbivore attack. Thus, besides being DAMPs under pathological conditions, such molecules may function in normal physiological processes.

In this review, we focus on several proposed plant primary and secondary DAMPs and their receptors, which have been shown to physically bind each other. For a complete inventory of potential DAMPs in plants, we refer interested readers to several recent excellent reviews and references therein (Choi and Klessig, 2016; Duran-Flores and Heil, 2016; Gust et al., 2017; Hou et al., 2019). A new item that was recently added to the inventory is the Arabidopsis (Arabidopsis thaliana) SCOOP12 peptide, which is perceived by plants in a BRASSINOSTEROID INSENSITIVE1 (BRI1)-ASSOCIATED KINASE1 (BAK1) coreceptor-dependent manner (Table 1; Gully et al., 2019). We explore potential roles of DAMPs in plant immunity, particularly in SAR. Future perspectives of DAMPs in plants are also discussed.

PRIMARY/CONSTITUTIVE DAMP-RECEPTOR PAIRS

Oligogalacturonides—WALL-ASSOCIATED KINASE1

Oligogalacturonides (OGs) are degradation products of the primary cell wall component pectin, a complex polysaccharide comprising mainly esterified d-GalUA residues in α-(1-4)-chain (Côté and Hahn, 1994; Ferrari et al., 2013; Kohorn, 2016). Pectin is partially degraded by pathogen- or plant-derived enzymes during pathogen infection or herbivore attack, resulting in oligomers of d-galacturonic acids with varying degrees of polymerization (Bishop et al., 1981; Côté and Hahn, 1994; Bergey et al., 1999; An et al., 2005). OGs with a degree of polymerization between 10 and 15 are potent elicitors (Côté and Hahn, 1994; Moscatiello et al., 2006; Ferrari et al., 2007; Denoux et al., 2008), able to induce reactive oxygen species (ROS) production, MAP kinase activation, callose deposition, defense protein accumulation, and resistance to the necrotrophic fungal pathogen Botrytis cinerea in multiple plant species (Hahn et al., 1981; Davis and Hahlbrock, 1987; Broekaert and Peumans, 1988; Bellincampi et al., 2000; Aziz et al., 2004; Denoux et al., 2008; Galletti et al., 2008; Rasul et al., 2012). Short OGs with a degree of polymerization between two and six have also been shown to elicit immune responses, but the effect of short OGs on the expression of immune-related genes appears to be not as strong as that of long OGs (Moloshok et al., 1992; Davidsson et al., 2017).

WALL-ASSOCIATED KINASE (WAK) proteins are proposed receptors of OGs (Kohorn and Kohorn, 2012; Ferrari et al., 2013). WAKs are receptor-like kinases (RLKs), with an extracellular domain containing epidermal growth factor motifs, a transmembrane domain, and an intracellular Ser/Thr kinase domain (He et al., 1996; Anderson et al., 2001). There are five WAK and 21 WAK-LIKE genes in Arabidopsis (Anderson et al., 2001; Verica and He, 2002). Biochemical analyses suggested that WAK1 is tightly associated with pectin (He et al., 1996; Wagner and Kohorn, 2001). The extracellular domains of WAK1 and WAK2 indeed bind pectin in vitro (Kohorn et al., 2009). A recombinant peptide containing amino acids 67 to 254 of the extracellular domain of WAK1 (called “WAK_67–254”) binds polygalacturonic acid (PGA), OGs, pectins, and structurally related alginates (Decreux and Messiaen, 2005). At least five specific amino acids in the extracellular domain of WAK1 are involved in the interaction with PGA (Decreux et al., 2006). Interestingly, binding of WAK_67–254 to PGA, OGs, and alginates depends on Ca²⁺ and ionic conditions that promote formation of Ca²⁺ bridges between oligomers or polymers, resulting in a structure known as an “egg-box dimer,” which significantly enhances binding to WAK1 and induces increased extracellular alkalinization when applied to Arabidopsis cell suspensions (Decreux and Messiaen, 2005; Cabrera et al., 2008).

Multiple lines of genetic evidence strongly support that WAKs are OG receptors and function in plant immune responses. First, a chimeric receptor with the extracellular domain of WAK1 and the kinase domain of ELONGATION FACTOR Tu receptor (EFR) responds to OGs and activates the kinase domain, and conversely, elf18, a polypeptide consisting of the first 18 amino acids at the N terminus of ELONGATION FACTOR Tu, activates a chimeric receptor formed by the EFR ectodomain and the kinase domain of WAK1 and induces the typical responses triggered by OGs (Brutus et al., 2010). Second, pectin- and OG-induced transcription of a number of genes depends on WAK2 in Arabidopsis protoplasts (Kohorn et al., 2009, 2012). Third, pathogen infection and SA treatment induce WAK1 gene expression and the induction depends on NONEXPRESSOR OF PATHOGENESIS-RELATED (PR) GENES1 (NPR1), a key immune regulator (Cao et al., 1997; He et al., 1998). SA also induces the expression of WAK2, WAK3, and WAK5 (He et al., 1999), and WAK1 and WAK2 are wound-inducible as well (Wagner and Kohorn, 2001). Fourth, overexpression of WAK1 enhances tolerance to SA toxicity, and expression of an antisense allele of WAK1 reduces the level of PR1 gene expression induced by the biologically active analog of SA, 2.6-dichloroisonicotinic acid (He et al., 1998). Fifth, a dominant gain-of-function WAK2 allele, WAK2cTAP, exhibits autoimmune phenotypes including ROS accumulation and cell death (Kohorn et al., 2009, 2012). Importantly, the stunted growth phenotype of WAK2cTAP is largely suppressed by mutations in the key immune regulators, ENHANCED DISEASE SUSCEPTIBILITY1, PHYTOALEXIN DEFICIENT4, and MAP KINASE6 (MPK6) genes (Kohorn et al., 2012, 2014), which is reminiscent of autoimmune phenotypes (van Wersch et al., 2016).

Extracellular ATP-DOES NOT RESPOND TO NUCLEOTIDES1

Extracellular ATP (eATP) is one of the best-studied DAMPs in animals. As the energy currency, cellular levels of ATP are normally maintained in the range of 1 to 10 mm. In animals, ATP is constitutively released into the extracellular space through various mechanisms including ATP binding cassette transporters, vesicular exocytosis, gap junctions, and pannexin hemichannels, as well as the P2X7 receptor (Lazarowski et al., 2003; Spray et al., 2006; Suadicani et al., 2006; Zhang et al., 2007). ATP also leaks into the extracellular milieu upon cell lysis or necrosis during tissue damage and inflammation (la Sala et al., 2003). Once in the extracellular milieu, ATP binds to either P2X ligand-gated channels or P2Y G-protein coupled receptors, triggering outside–in signaling including changes in intracellular [Ca²⁺], production of cytokines, and cell death (Hattori and Gouaux, 2012; Jacobson et al., 2015). Depending on the tissue and cell types, eATP signaling acts in both normal physiological and abnormal pathological processes in animals (Trautmann, 2009).

In plants, research with exogenous ATP can be traced back to the 1960s (Jaffe and Galston, 1966). However, it was unclear in the early studies whether the exogenously added ATP functioned as a signal molecule, a precursor, or an energy supply (Jaffe and Galston, 1966; Williamson, 1975; Kamizyo and Tanaka, 1982; Nejidat et al., 1983). Recent studies with the widely used stable ATP analog, adenosine 5′-[γ-thio]triphosphate, suggested that eATP might act as a signal molecule in the apoplast (Jeter et al., 2004; Song et al., 2006; Torres et al., 2008; Clark et al., 2010, 2011). The presence of eATP was proven by directly measuring ATP accumulation in Arabidopsis leaves and roots (Thomas et al., 1999; Demidchik et al., 2003; Deng et al., 2015), and active secretion of ATP in plants was confirmed by feeding Arabidopsis cultures with [³²P]H₃PO₄ and monitoring radiolabeled ATP in the extracellular matrix (Chivasa et al., 2005). Furthermore, the distribution of eATP in plants was directly visualized using luciferase reporters including a cellulose-binding domain-luciferase fusion, an ecto-luciferase, and the infiltration of a luciferase/luciferin mixture (Kim et al., 2006; Chivasa et al., 2009; Clark et al., 2011). These tools allowed discoveries of the dynamics of eATP accumulation in roots, leaves, and around guard cells (Kim et al., 2006; Chivasa et al., 2009; Clark et al., 2011).

The constitutive eATP appears to be essential for plant cell viability. Depletion of basal eATP using the cell-impermeant traps Glc-hexokinase and apyrase triggers cell death in both Arabidopsis cell cultures and whole plants (Chivasa et al., 2005). Competitive exclusion of eATP from its binding sites with nonhydrolyzable ATP analog β,γ-methyleneadenosine 5′-triphosphate also results in cell death in Arabidopsis, maize (Zea mays), bean (Phaseolus vulgaris), and tobacco (Nicotiana tabacum; Chivasa et al., 2005). Interestingly, the programmed cell death-eliciting mycotoxin fumonisin B1-induced cell death in Arabidopsis seems to be mediated by depletion of eATP (Chivasa et al., 2005). Furthermore, environmental stresses induce ATP release (Clark et al., 2011; Sun et al., 2012; Lim et al., 2014; Deng et al., 2015). Although the biological relevance of the increases in endogenous eATP levels remains to be fully elucidated, studies with exogenous ATP and/or adenosine 5′-[γ-thio]triphosphate have shown that eATP induces ROS and nitric oxide production, Ca²⁺ influx, and H⁺ efflux in a G protein α-subunit and RESPIRATORY BURST OXIDASE HOMOLOG (RBOH)-dependent manner (Jeter et al., 2004; Song et al., 2006; Foresi et al., 2007; Wu et al., 2008; Wu and Wu, 2008; Demidchik et al., 2009; Clark et al., 2011; Hao et al., 2012; Sun et al., 2012). Intriguingly, plants appear to respond to eATP in a dose-dependent manner. Low doses of eATP induce stomatal opening, accelerate vesicular trafficking, and stimulate cell elongation, whereas high doses of eATP trigger stomatal closure, inhibit vesicular trafficking, and suppress cell elongation (Clark et al., 2010, 2011, 2013; Wang et al., 2014; Deng et al., 2015). Although depletion of eATP or exclusion of eATP from its binding sites leads to cell death, high doses of eATP also reduce cell viability (Sun et al., 2012; Deng et al., 2015). Currently, the molecular mechanisms underlying such biphasic responses are unknown.

Identification of the eATP receptor DOES NOT RESPOND TO NUCLEOTIDES1 (DORN1) in Arabidopsis is a major breakthrough in eATP biology and provided a key to addressing many questions about eATP (Choi et al., 2014a; Roux, 2014). DORN1 is a legume-type lectin receptor kinase (LecRK), LecRK-I.9, which had previously been shown to recognize RGD (Arg-Gly-Asp) tripeptide motif-containing protein in mediating plasma membrane-cell wall adhesions (Gouget et al., 2006). The extracellular domain of DORN1 binds ATP with a dissociation constant (K_d) of ∼46 nm (Choi et al., 2014a). A point mutation in the DORN1 gene completely blocks exogenous ATP-induced transcriptional changes in Arabidopsis seedlings, indicating that DORN1 is the major, if not the sole, receptor of eATP (Choi et al., 2014a). However, as eATP plays an important role in plant growth, development, and cell viability (Tang et al., 2003; Chivasa et al., 2005; Clark and Roux, 2011; Liu et al., 2012; Yang et al., 2015), but dorn1 mutants do not have obvious growth and developmental defects (Choi et al., 2014a), it has been suggested that there might be other eATP receptors mainly regulating plant growth signaling (Roux, 2014).

It was recently proposed that eATP functions as a DAMP in plants (Choi et al., 2014b; Tanaka et al., 2014). Indeed, eATP levels at the wound sites reach ∼40 μm, well above the concentration needed to induce ROS production and gene expression (Choi et al., 2014a), and reducing eATP levels by overexpressing an apyrase suppresses wound responses (Song et al., 2006; Wang et al., 2019b). Furthermore, ∼60% of the genes induced by exogenous ATP are also induced by wounding (Choi et al., 2014a), and ATP mainly activates JA signaling through MYC transcription factors (Tripathi et al., 2018; Jewell et al., 2019). Therefore, eATP clearly plays an important role in wound responses. Furthermore, exogenous ATP induces resistance to the necrotrophic fungal pathogen B. cinerea in Arabidopsis (Tripathi et al., 2018), suggesting a potential role for eATP in immunity against fungal pathogens. Interestingly, although more than a dozen ATP-induced genes depend on NPR1 (Jewell et al., 2019), eATP and SA antagonize each other (Chivasa et al., 2009). Exogenous ATP reduces basal SA levels, whereas SA treatment triggers collapse of eATP in tobacco leaves (Chivasa et al., 2009). In line with these results, exogenous ATP does not induce apoplastic resistance to Pseudomonas syringae pv. maculicola strain ES4326 in Arabidopsis (Zhang and Mou, 2009). On the other hand, eATP plays an important positive role in stomatal immunity. In Arabidopsis, bacterial infection induces ATP release, particularly around guard cells, and exogenous ATP induces stomatal closure and stomatal resistance against bacterial pathogens in a concentration-dependent manner (Chen et al., 2017). Importantly, exogenous ATP-induced stomatal movement and resistance depend on DORN1 and RBOHD. It was proposed that eATP activates DORN1, which in turn phosphorylates the N terminus of RBOHD, leading to ROS production that induces stomatal closure (Chen et al., 2017).

Extracellular NAD(P)—LecRK-I.8/LecRK-VI.2

It is well known that extracellular NAD (eNAD) and NADP (eNADP) play a significant role in animal immune responses (Billington et al., 2006; Haag et al., 2007; Adriouch et al., 2012). However, whether eNAD(P) is a DAMP in animals remains elusive (Roh and Sohn, 2018). Under normal conditions, intracellular NAD⁺ levels are in the range of 0.2 to 0.5 mm (Cantó et al., 2015), whereas eNAD levels, e.g. in mammalian serum, are ∼0.1 μm (Zocchi et al., 1999; O’Reilly and Niven, 2003). Cell lysis during tissue damage and inflammation presumably can lead to dramatic increases in eNAD(P) levels (Billington et al., 2006). At least three distinct mechanisms perceive eNAD(P) in animals. First, eNAD(P) can be processed by a number of NAD(P)-metabolizing ectoenzymes such as CD38 and CD157, which have ADP-ribosyl cyclase, cADP-ribose hydrolase and NAD-hydrolase activities, into Ca²⁺-mobilizing second messengers cADP-ribose and nicotinic acid adenine dinucleotide phosphate (Ceni et al., 2003; Partida-Sánchez et al., 2003; de Flora et al., 2004; Heidemann et al., 2005; Malavasi et al., 2006). Second, eNAD⁺ is a substrate of the GPI-anchored or secreted ectoenzymes known as mono(ADP-ribosyl)transferases in ADP-ribosylation of plasma membrane signaling proteins (Nemoto et al., 1996; Han et al., 2000; Bannas et al., 2005). Finally, eNAD(P) is a potential agonist of plasma membrane receptors. It has previously been shown that NAD⁺ binds to rat brain synaptic membranes and is a potential inhibitory neurotransmitter (Khalmuradov et al., 1983; Mutafova-Yambolieva et al., 2007). Recent studies have suggested that several purinergic P2X and P2Y receptors function in eNAD(P)-triggered biological responses (Moreschi et al., 2006; Mutafova-Yambolieva et al., 2007; Grahnert et al., 2009; Klein et al., 2009). Nevertheless, binding between NAD(P) and these receptors has not been reported.

In plants, intracellular NAD(P) levels are in the range of 1 to 2 mm (Noctor et al., 2006). We found that, upon wounding and bacterial infection, NAD(P) concentrations in the extracellular washing fluid are comparable to those from infiltration with ∼0.7 and ∼1.2 mm of NAD(P), respectively (Zhang and Mou, 2009). We also showed that treatment of Arabidopsis and citrus plants with 0.2 mm of NAD(P) significantly induces resistance to bacterial pathogens, but not to the necrotrophic fungal pathogen B. cinerea (Zhang and Mou, 2009; Wang et al., 2016; Alferez et al., 2018). Importantly, exogenously applied NAD(P) does not change intracellular NAD(P) homeostasis (Zhang and Mou, 2009), suggesting that it acts in the apoplast. Furthermore, we found that transgenic expression of the human CD38 gene in Arabidopsis reduces eNAD(P) concentrations and partially compromises SAR (Zhang and Mou, 2012). These results together indicate that the eNAD(P) accumulated during pathogen infection is both necessary and sufficient for activation of plant immune responses. In addition, exogenous NAD(P) induces ROS production and changes in cytosolic [Ca²⁺] (Pétriacq et al., 2016a, 2016b). Thus, eNAD(P) is a DAMP in plants.

Using a reverse genetic approach based on exogenous NAD⁺-induced transcriptome changes in Arabidopsis, we have identified two potential eNAD(P) receptors, LecRK-I.8 and LecRK-VI.2, both of which are legume-type LecRKs (Singh et al., 2012; Wang et al., 2017a, 2019a). The LecRK-I.8 and LecRK-VI.2 genes can be induced by exogenous NAD⁺, and both LecRK-I.8 and LecRK-VI.2 are localized in the plasma membrane and have kinase activity (Xin et al., 2009; Singh et al., 2013; Wang et al., 2017a). However, the two receptors are not alike. LecRK-I.8 only binds NAD⁺ (K_d, ∼437 nm), whereas LecRK-VI.2 binds both NAD⁺ and NADP⁺ with a slightly higher affinity for NADP⁺ (Wang et al., 2017a, 2019a). LecRK-VI.2 binds ³²P-NAD⁺ with a K_d of ∼787 nm, and the binding can be effectively competed by unlabeled NAD⁺ (50% inhibition concentration, IC₅₀, 1,887 nm) and NADP⁺ (IC₅₀, 945 nm; Wang et al., 2019a). Consistently, mutations in LecRK-I.8 and LecRK-VI.2 suppress NAD⁺- and NADP⁺-induced immune responses, respectively (Wang et al., 2017a, 2019a). Interestingly, the lecrk-I.8/VI.2 double mutant behaves like lecrk-I.8 for NAD⁺ responses and like lecrk-VI.2 for NADP⁺ responses, indicating that the two receptors function in two separate pathways (Wang et al., 2019a). Importantly, mutations in LecRK-I.8 and LecRK-VI.2 significantly compromise basal immunity and biological induction of SAR, respectively (Wang et al., 2017a, 2019a), indicating that LecRK-I.8 primarily functions in basal immunity, whereas LecRK-VI.2 plays a major role in SAR.

The Leu-rich repeat receptor kinase (LRR-RK) BAK1 is a coreceptor of a group of LRR-RK receptors including BRI1, FLAGELLIN-SENSITIVE2 (FLS2), EFR, and PEP RECEPTOR1 (PEPR1)/PEPR2 (Li et al., 2002a; Nam and Li, 2002; Chinchilla et al., 2007; Heese et al., 2007; Postel et al., 2010; Schulze et al., 2010; Roux et al., 2011). BAK1 is also required for signaling triggered by several other potential DAMPs including the Arabidopsis HMGB3 protein and the SCOOP12 peptide (Choi et al., 2016; Gully et al., 2019). BAK1 and LecRK-VI.2 form a complex in vivo and function in eNAD(P) signaling and SAR (Wang et al., 2019a). The interaction between BAK1 and LecRK-VI.2 appears to be constitutive and independent of eNAD(P), which is different from the inducible associations between BAK1 and LRR-RK receptors. Moreover, the bak1-5 mutation has been shown to impair signaling mediated by the non-RD kinases FLS and EFR, but not that mediated by the RD kinase BRI1 (Schwessinger et al., 2011). Interestingly, although LecRK-VI.2 is an RD kinase, eNAD(P) signaling is significantly inhibited in bak1-5 (Wang et al., 2019a). In addition, it has been shown that C-terminal tags on BAK1 have limited effects on several BR responses, but strongly impact PTI signaling (Ntoukakis et al., 2011). Surprisingly, a BAK1-GFP fusion protein is able to complement the defects of bak1-5 in NADP⁺-induced immune responses and biological induction of SAR (Wang et al., 2019a). Because C-terminally tagged BAK1 fusion proteins are not phosphorylated at S612 upon PAMP treatment (Perraki et al., 2018), it would be interesting to test whether S612 phosphorylation in BAK1 is required for eNAD(P) signaling and SAR.

Interestingly, exogenously added NAD⁺ moves systemically and induces systemic resistance (Wang et al., 2019a), suggesting that eNAD(P) might be an SAR mobile signal. Consistently, high levels of exogenous NAD(P) induces SA accumulation and NADPH oxidase-independent ROS production (Zhang and Mou, 2009; Pétriacq et al., 2016b). Surprisingly, exogenous NAD(P)-induced systemic immunity does not depend on the putative SAR mobile signals pipecolic acid (Pip), N-hydroxy-Pip, azelaic acid (AzA), and glycerol-3-phosphate (G3P), but requires an intact SA signaling pathway (Wang et al., 2019a). Although the role of DEFECTIVE IN INDUCED RESISTANCE1 (DIR1) and ROS in NAD(P)-induced systemic immunity has not been tested, exogenous NAD(P)-induced local resistance and PR gene expression is independent of DIR1 and NADPH oxidase, respectively (Zhang and Mou, 2009; Wang et al., 2019a). It appears that eNAD(P) functions either downstream or independently of the putative SAR mobile signals Pip, N-hydroxy-Pip, AzA, G3P, DIR1, and ROS in both local and systemic resistance. Furthermore, although exogenous eNAD(P) requires SA signaling for immune response activation, SA induces the expression of LecRK-VI.2 in an NPR1-dependent manner (Wang et al., 2019a). In addition, because Pip, ROS, AzA, and G3P form a signaling amplification loop (Wang et al., 2018a), it is possible that ROS produced in the amplification loop causes reversible or irreversible damages to the plasma membrane (Cwiklik and Jungwirth, 2010; Tero et al., 2016), leading to leakage of cellular NAD(P) into the apoplast. Thus, the interplay between eNAD(P) and SA as well as other SAR signal molecules is complicated and deserves further investigation.

Glu—GLU-RECEPTOR3.3/GLU-RECEPTOR3.6

Glu is the most prominent neurotransmitter in the brain and excites postsynaptic neural cells through different types of receptors including ionotropic and metabotropic Glu receptors (Brassai et al., 2015). Ionotropic Glu receptors (iGluRs) are ligand-gated channels that are activated upon Glu binding (Krieger et al., 2019). The Arabidopsis genome encodes 20 GLU-RECEPTORs (GLRs) that are homologous to iGluRs (Chiu et al., 2002). GLRs carry the same signature domains as animal iGluRs, including the “three-plus-one” transmembrane domains and the extracellular ligand-binding domains (Lam et al., 1998; Chiu et al., 1999; Lacombe et al., 2001). Upon herbivore and mechanical damage, Glu is released into the apoplast where it activates GLR3.3 and GLR3.6, triggering long-distance electric and Ca²⁺ signaling as well as JA accumulation and defense gene expression in undamaged leaves (Mousavi et al., 2013; Toyota et al., 2018). At least six amino acids (Glu, Gly, Ala, Ser, Asn, and Cys) and the tripeptide glutathione can also serve as agonists of GLR3.3 and induce membrane depolarization and cytosolic [Ca²⁺] elevation in a GLR3.3-dependent manner (Qi et al., 2006; Stephens et al., 2008; Li et al., 2013). Moreover, seven out of the 20 standard amino acids (Met, Trp, Phe, Leu, Tyr, Asn, and Thr) activate GLR1.4 transiently expressed in Xenopus oocytes to various extents, and Met-induced membrane depolarization in Arabidopsis leaves depends on GLR1.4 (Tapken et al., 2013).

Interestingly, several amino acids have been shown to induce disease resistance in plants. For instance, His induces ET biosynthesis and ET-related defense gene expression as well as resistance to the soil-borne bacterial pathogen Ralstonia solanacearum and the fungal pathogen B. cinerea partially in an ET-dependent manner in tomato (Solanum lycopersicum) and Arabidopsis (Seo et al., 2016). Glu induces several genes of the SA signaling pathway in rice (Oryza sativa) and tomato fruit, and enhances resistance to Magnaporthe oryzae and Alternaria alternata in rice and tomato fruit, respectively (Kadotani et al., 2016; Yang et al., 2017). Surprisingly, other amino acids except Trp and Tyr also improve rice resistance to M. oryzae to various degrees (Kadotani et al., 2016). Furthermore, Cys, Asp, and GSH enhance resistance to P seudomonas syringae pv. tomato (Pst) DC3000 in Arabidopsis (Li et al., 2013). Importantly, Cys- and GSH-induced disease resistance depends on GLR3.3, and mutations of the GLR3.3 gene compromise resistance to Pst DC3000 and Hyaloperonospora arabidopsidis in Arabidopsis (Li et al., 2013; Manzoor et al., 2013), suggesting that GLR3.3 is a potential receptor for Cys and GSH released into the apoplast during pathogen infection.

SECONDARY/ INDUCIBLE DAMP-RECEPTOR PAIRS

Systemin—Systemin Receptor1/Systemin Receptor2

Systemin is the first reported extracellular peptide that induces defense signaling in plants. It was purified from tomato leaf extracts using HPLC based on its proteinase inhibitor (PIN) gene-inducing activity (Pearce et al., 1991). Systemin is an 18-amino acid peptide processed from a 200-amino acid precursor named “prosystemin” (Pearce et al., 1991; Beloshistov et al., 2018). Genes encoding well-conserved prosystemins were identified in the Solanaceae species tomato, potato (Solanum tuberosum), bell pepper (Capsicum annuum), and black nightshade (Solanum nigrum), but not in tobacco (McGurl et al., 1992; Constabel et al., 1998). The tomato prosystemin gene is constitutively expressed throughout the plant except in the roots, and is further induced by wounding (McGurl et al., 1992). The prosystemin protein accumulates in the cytosol and nucleus of vascular parenchyma cells in response to wounding and methyl JA (MeJA) treatment (Narvíez-Vásquez and Ryan, 2004). Prosystemin does not carry an N-terminal signal sequence and, upon cell damage, is expected to passively leak into the apoplast where it is processed by phytaspases and possibly Leu aminopeptidase A (Ryan and Pearce, 1998; Beloshistov et al., 2018). Systemin is highly active. When supplied to the cut stems of young tomato plants, ∼40 fmol of systemin per plant is sufficient to induce half-maximal accumulation of two wound-inducible PINs that break the activity of digestive enzymes in the insect midgut (Green and Ryan, 1972; Pearce et al., 1991). Overexpression of the prosystemin gene leads to constitutive synthesis of the PINs (McGurl et al., 1994).

Although exogenously supplied systemin moves systemically, systemin may not be the mobile signal mediating systemic wound responses. Grafting experiments with tomato JA biosynthesis and recognition mutants indicated that systemic wound signaling requires both biosynthesis of JA at the wound site and recognition of a JA signal in remote tissues, suggesting that JA controls the production of or acts as the mobile wound signal (Li et al., 2002b). It was proposed that systemin promotes systemic wound signaling by augmenting JA biosynthesis in the vascular tissues (Schilmiller and Howe, 2005).

Identification of the receptor of systemin was a daunting task. A 160-kD systemin-binding protein named SR160 was initially purified from plasma membranes of tomato suspension cells using a photoaffinity analog of systemin (Scheer and Ryan, 2002). SR160 turned out to be the tomato homolog of the steroid hormone brassinolide receptor BRI1 (Scheer and Ryan, 2002; Scheer et al., 2003). Later studies indicated that, although SR160 increases binding of systemin to tobacco plasma membranes, it does not mediate systemin-triggered defense responses (Holton et al., 2007; Lanfermeijer et al., 2008; Malinowski et al., 2009). Two distinct LRR-RKs termed SYSTEMIN RECEPTOR1 (SYR1) and SYR2 were recently identified as the bona fide systemin receptors (Wang et al., 2018b). Tobacco leaves expressing SYR1 and SYR2 respond with an EC₅₀ of ∼0.03 and >30 nm systemin based on systemin-induced ROS production, respectively (Wang et al., 2018b). Importantly, systemin is unable to induce production of ET and expression of the PIN gene PIN1 in tomato mutant lines lacking functional SYR1 and SYR2 (Wang et al., 2018b). Surprisingly, mechanical wounding still induces local and systemic expression of the PIN1 gene, even though tomato plants expressing a prosystemin antisense gene accumulate <40% of the wild-type level of PIN1 (McGurl et al., 1992; Wang et al., 2018b). Nevertheless, both the prosystemin antisense lines and the receptor mutant line support significantly better herbivore larval growth than wild type (McGurl et al., 1992; Wang et al., 2018b), demonstrating that systemin signaling contributes to resistance against insect herbivores in tomato.

Plant Elicitor Peptides—Pep Receptor1/Pep Receptor2

The first plant elicitor peptide, Pep1, was isolated as a 23-amino acid peptide from extracts of Arabidopsis leaves, which is derived from the C terminus of a 92-amino acid precursor protein encoded by the PROPEP1 gene (Huffaker et al., 2006). The PROPEP1 protein does not carry an N-terminal signal peptide (Huffaker et al., 2006). It has been shown that PROPEP1 is processed by Ca²⁺-dependent type-II metacaspases in Arabidopsis (Hander et al., 2019; Shen et al., 2019). The Arabidopsis genome carries eight PROPEP genes, PROPEP1 to PROPEP8 (Huffaker et al., 2006; Bartels et al., 2013). PROPEP1, PROPEP2, PROPEP3, PROPEP5, and PROPEP8 are expressed in the roots and slightly in the leaf vasculature, and are inducible by wounding, and expression of PROPEP4 and PROPEP7 is restricted to the root tip and is not inducible by wounding (Bartels et al., 2013). Expression of PROPEP1, PROPEP2, and PROPEP4 is inducible by MeJA, whereas that of PROPEP2 and PROPEP3 is inducible by methyl SA (Huffaker and Ryan, 2007). PROPEP2 and PROPEP3 are also inducible by pathogen attacks and elicitors derived from pathogens (Huffaker et al., 2006). Furthermore, expression of PROPEP1 is strongly induced by Pep1 to Pep3, PROPEP2 and PROPEP3 are strongly induced by Pep1 to Pep6, PROPEP4 and PROPEP5 are weakly inducible, and PROPEP6 is not inducible by the peptides (Huffaker et al., 2006; Huffaker and Ryan, 2007; Yamaguchi et al., 2010). Interestingly, while PROPEP3-YFP is localized in the cytoplasm, PROPEP1-YFP and PROPEP6-YFP are associated with the tonoplast (Bartels et al., 2013). The different gene expression patterns and localization suggest nonredundant roles among the members of the PROPEP family. Based on the responses of PROPEP gene promoters to various stimuli, PROPEP genes were classified into four groups, with PROPEP1 in the first group, PROPEP2 and PROPEP3 in the second group, PROPEP4, PROPEP7, and PROPEP8 in the third group, and PROPEP5 in the fourth group (Safaeizadeh and Boller, 2019). Nevertheless, all Peps, when applied exogenously, activate MPK3 and MPK6, induce ET production, and inhibit seedling growth (Bartels et al., 2013). Exogenous Peps also induce expression of several defense genes including PDF1.2, MPK3, and WRKY33, production of ROS, elevation of cytosolic [Ca²⁺], and resistance to the bacterial pathogen Pst DC3000 (Huffaker et al., 2006; Qi et al., 2010; Yamaguchi et al., 2010). Pep1 also induces resistance against B. cinerea (Liu et al., 2013). Overexpression of PROPEP1 and PROPEP2 in Arabidopsis results in constitutive PDF1.2 expression and/or resistance against a root oomycete pathogen Pythium irregulare (Huffaker et al., 2006).

The first PEPR, an LRR-RK called “PEPR1,” was purified from Arabidopsis suspension cells using a photoaffinity analog of Pep1, ¹²⁵I₁-Tyr-Pep1 (Yamaguchi et al., 2006). ¹²⁵I₁-Tyr-Pep1 is as active as Pep1 and binds to Arabidopsis suspension cells with a K_d of ∼0.25 nm (Yamaguchi et al., 2006). The second PEPR, PEPR2, was identified by phylogenetic analysis and searching for the most closely related gene to PEPR1 (Yamaguchi et al., 2010). Transgenic tobacco cells expressing PEPR1 and PEPR2 bind ¹²⁵I₁-Tyr-Pep1 with K_d values of 0.56 and 1.25 nM, respectively. PEPR1 and PEPR2 also bind Pep2 to Pep6 and Pep2, respectively (Yamaguchi et al., 2010). Both PEPRs carry a guanylyl cyclase (GC) catalytic domain with residues for catalysis being conserved (Qi et al., 2010; Yamaguchi et al., 2010), and the GC activity of PEPR1 has been experimentally demonstrated (Qi et al., 2010). It has been shown that Pep1 induces rapid formation of a heterocomplex containing de novo phosphorylated BAK1 and an ∼160-kD polypeptide that is expected to be PEPR1 (Schulze et al., 2010), while Pep2 induces PEPR1 association with BAK1, BAK1-LIKE1, SOMATIC EMBRYOGENESIS RECEPTOR-LIKE KINASE1 (SERK1), and SERK2 in Nicotiana benthamiana (Yamada et al., 2016). Consistently, the kinase domains of PEPR1 and PERP2 interact with that of BAK1 in yeast (Postel et al., 2010), and disruption of BAK1 sensitizes PEPR signaling (Yamada et al., 2016). The kinase domain of PEPR1 also interacts with and directly phosphorylates the receptor-like cytoplasmic kinase BOTRYTIS-INDUCED KINASE1 (BIK1) and BIK is required for Pep1-induced resistance against B. cinerea (Liu et al., 2013).

Expression of PEPR1 and PEPR2 is inducible by wounding, MeJA, most Peps, and PAMPs such as flg22 (a 22-amino acid peptide corresponding to the N terminus of bacterial flagellin) and elf18 (Yamaguchi et al., 2010). It appears that PEPR1 is inducible in different parts of the plant, whereas PEPR2 induction is restricted to the root (Safaeizadeh and Boller, 2019). Pep-induced expression of defense genes including MPK3 and WRKY33 is partially suppressed in the pepr1 and pepr2 single mutants, and completely blocked in the pepr1 pepr2 double mutant (Yamaguchi et al., 2010). Pep1-induced expression of PR1 and PDF1.2 as well as resistance against Pst DC3000 are also compromised in the double mutant (Yamaguchi et al., 2010). Interestingly, ET-induced expression of defense genes and resistance to B. cinerea are also compromised in the pepr1 pepr2 double mutant (Liu et al., 2013). Furthermore, local application of Pep2 activates both JA and SA signaling pathways and resistance to Colletotrichum higginsianum path-29 strain in systemic leaves, although Pep2 may not be a mobile signal (Ross et al., 2014). In agreement with this result, biological induction of SAR is compromised in the pepr1 pepr2 mutant (Ross et al., 2014).

RAPID ALKALINIZATION FACTORs—FERONIA

RAPID ALKALINIZATION FACTOR (RALF) peptides were first isolated from tobacco, tomato, and alfalfa (Medicago sativa) leaves based on their activity in alkalinating the medium of tobacco suspension cells (Pearce et al., 2001b), and later from sugarcane (Saccharum) leaves using a similar approach (Mingossi et al., 2010). The tobacco RALF is a 49-amino acid peptide located at the C terminus of a 115-amino acid preproprotein. The preproprotein carries an N-terminal signal peptide and the derived RALF peptide contains four cysteines that form two disulfide bridges important for its activity (Pearce et al., 2001b). Later studies indicated that many, but not all, RALF preproproteins are cleaved at a conserved dibasic site RRXL by plant subtilisin-like Ser proteases such as the Arabidopsis SITE-1 PROTEASE/SBT6.1 (Matos et al., 2008; Srivastava et al., 2009; Stegmann et al., 2017). A photoaffinity analog of the tomato RALF peptide, ¹²⁵I-azido-LeRALF, which has biological activity similar to the native LeRALF, binds to tomato suspension cells with a K_d of 0.8 nm (Scheer et al., 2005). A highly conserved YISY motif located at positions 5 through 8 from the N terminus is essential for RALF activity, presumably being required for productive binding to its putative receptor (Pearce et al., 2010b).

RALF proteins have been identified in a large number of plant species that represent a variety of land plant lineages (Cao and Shi, 2012; Murphy and de Smet, 2014). The Arabidopsis genome carries 39 RALF genes (Sharma et al., 2016). Comprehensive analysis of the identified 795 RALF proteins from various plant species revealed four major clades. Clades I, II, and III carry the features important for RALF activity, including the RRXL cleavage site and the YISY motif important for receptor binding, whereas clade IV is highly diverged and lacks these features (Campbell and Turner, 2017). While the mean length of the RALF proteins in clades I, II, and III is 125 amino acids, the clade-IV RALFs have an average length of only 88 amino acids, suggesting that the members in clade IV may not be true RALFs (Campbell and Turner, 2017).

RALF peptides were initially found to suppress root growth of tomato and Arabidopsis seedlings as well as tomato pollen tube growth (Pearce et al., 2001b; Covey et al., 2010). In line with these results, silencing of the tobacco RALF gene leads to increased root growth and abnormal root hair development (Wu et al., 2007), whereas transgenic overexpression of the Arabidopsis RALF1 and RALF23 genes results in dwarf phenotypes (Matos et al., 2008; Srivastava et al., 2009). Moreover, RALF genes are highly expressed in roots, shoots, and flowers (Zhang et al., 2010; Cao and Shi, 2012; Campbell and Turner, 2017). Collectively, these results support a role for RALF peptides in plant growth and development. On the other hand, the fungal pathogen Fusarium oxysporum f. sp. ciceri (Race 1)-induced expression of a RALF-related EST is 5-fold higher in resistant chickpea (Cicer arietinum) plants than in a susceptible variety (Gupta et al., 2010). In Arabidopsis, RALF8 is induced by a combination of water deficit and nematode stress, and overexpression of RALF8 confers susceptibility to drought stress and nematode infection (Atkinson et al., 2013). Moreover, synthetic RALF17 peptide increases resistance to Pst DC3000, while RALF23 reduces resistance to Pst DC3000 (Stegmann et al., 2017). Consistently, overexpression of RALF23 inhibits resistance to Pst DC3000 coronatine-minus, whereas loss of RALF23 enhances resistance to Pst DC3000 coronatine-minus (Stegmann et al., 2017). Interestingly, genomes of 26 species of phytopathogenic fungi encode RALF homologs, and the predicted F. oxysporum RALF appears to contribute to the virulence of the pathogen in tomato plants (Masachis et al., 2016; Thynne et al., 2017). These data together suggest potential involvement of RALFs in plant immunity.

The first RALF receptor FERONIA (FER), a Catharanthus roseus receptor-like kinase1-like (CrRLK1L) receptor, was identified by quantitative phosphoproteomic profiling of RALF1-treated Arabidopsis seedlings (Haruta et al., 2014). The finding that the abundance of FER phosphopeptides increased in RALF1-treated samples led to the hypothesis that FER might be the receptor of RALF1. This hypothesis was supported by reduced RALF1 sensitivity of fer mutants and binding of RALF1 to FER (Haruta et al., 2014). Recent studies have shown that RALF4 and RALF19 bind to other CrRLK1L receptors including ANXUR1, ANXUR2, Buddha’s Paper Seal1, and Buddha’s Paper Seal2, as well as LEU-RICH REPEAT EXTENSIN proteins in regulating pollen tube integrity and sperm release in Arabidopsis (Mecchia et al., 2017; Ge et al., 2017). FER is also a receptor of RALF23 and perhaps RALF33 as well (Stegmann et al., 2017). Interestingly, FER constitutively associates with both FLS2 and BAK1 to act as scaffolds for ligand-induced FLS2-BAK1 complex formation. The constitutive association between BAK1 and FER can be strongly enhanced upon treatment with flg22, whereas binding of RALF23 to FER inhibits flg22/elf18-induced complex formation between FLS2/EFR and BAK1, leading to attenuation of FLS2/EFR-mediated PTI signaling (Stegmann et al., 2017). Furthermore, the GPI-anchored protein LORELEI-like GPI-AP1 (LLG1) constitutively associates with both FER and FLS2 and is required for PTI signaling (Li et al., 2015; Shen et al., 2017). LLG1 and the related LLG2 directly bind RALF23 to nucleate the assembly of a RALF23–LLG1/2–FER heterocomplex (Xiao et al., 2019), suggesting that RALFs may be perceived by distinct CrRLK1L receptor kinase-LLG/LRE heterocomplexes in regulating various biological processes including plant immunity.

Phytosulfokines—Phytosulfokine Receptor1/Phytosulfokine Receptor2

Phytosulfokines (PSKs) are sulfated Tyr-containing pentapeptides with mitogenic activity in vitro. The first PSK was purified from conditioned medium of rapidly growing asparagus (Asparagus officinalis) cell cultures by following its mitogenic activity (Matsubayashi and Sakagami, 1996). Based on the amino acid sequence of the asparagus PSK, rice and Arabidopsis PSK genes were subsequently identified (Yang et al., 1999, 2001; Matsubayashi et al., 2006). PSKs are derived from ∼77- to 89-amino acid prepropeptide precursors through tyrosylprotein sulfotransferase (TPST)-mediated Tyr sulfation and subtilisin-like Ser protease-catalyzed proteolytic cleavage (Srivastava et al., 2008; Komori et al., 2009). The PSK precursors carry N-terminal signal sequences and are sulfated in the Golgi apparatus, secreted, and cleaved in the extracellular milieu (Yang et al., 1999, 2001; Srivastava et al., 2008; Komori et al., 2009).

PSK binds to plasma membrane-enriched fractions with both high and low affinities (K_d values ranging from 1 to 100 nm; Matsubayashi et al., 1997; Matsubayashi and Sakagami, 1999). Photoaffinity cross-linking analysis indicated that the putative receptors for PSK in rice are 120- and 160-kD glycosylated proteins (Matsubayashi and Sakagami, 2000). The first PSK receptor, an LRR-RK, was purified from microsomal fractions of carrot suspension cells using ligand-based affinity chromatography, and the carrot PSK receptor (PSKr) gene encodes both 120- and 150-kD proteins (Matsubayashi et al., 2002). Amino acid homology search revealed that the Arabidopsis genome encodes two PSKRs, PSKR1 and PSKR2 (Matsubayashi et al., 2006; Amano et al., 2007). Structure analysis indicated that PSK interacts with and stabilizes an island domain of PSKR, which enhances PSKR heterodimerization with a SERK coreceptor (Wang et al., 2015). The cytoplasmic domain of PSKR1 has not only kinase activity but also GC activity. Both exogenous PSK treatment and overexpression of PSKR1 increase cGMP levels in protoplasts (Kwezi et al., 2011). Moreover, PSKR1, BAK1, CNGC17, and H⁺-ATPases AHA1 and AHA2 form a complex in mediating PSK-triggered signaling (Ladwig et al., 2015).

PSK was initially shown to induce the proliferation of asparagus suspension cells (Matsubayashi and Sakagami, 1996; Matsubayashi et al., 1997). PSK precursors are constitutively secreted by suspension cells, and overexpression and silencing of PSK genes led to increased and reduced PSK levels in conditioned media of rice transgenic cells, respectively (Yang et al., 1999, 2001). PSK genes are stably expressed not only in suspension cells but also in intact plants (Yang et al., 1999, 2001). Overexpression of PSK genes resulted in enlarged transgenic calli (Yang et al., 2001; Matsubayashi et al., 2006). Similarly, transgenic carrot cells expressing high levels of sense mRNA of the PSK receptor exhibited accelerated proliferation, whereas those expressing antisense showed substantially reduced callus growth (Matsubayashi et al., 2002). Individual cells of the Arabidopsis pskr1-1 mutant gradually lose their potential to form calli as the tissues mature, while PSKR1-overexpressing plants exhibit significantly greater callus-forming potential than wild type (Matsubayashi et al., 2006).

Genes encoding PSK precursors, processing enzymes, and/or receptors are inducible by wounding, elf18, flg22, and B. cinerea (Srivastava et al., 2008; Igarashi et al., 2012; Hou et al., 2014; Zhang et al., 2018), suggesting a potential involvement of PSK-PSKR signaling in plant immunity. Indeed, elf18-triggered immune responses are enhanced in the Arabidopsis pskr1–3 mutant (Igarashi et al., 2012). Mutations of the PSKR1 and TPST genes enhance resistance to Pst DC3000 and increase susceptibility to A. brassicicola, whereas overexpression of PSK2, PSK4, and PSKR1 leads to opposite effects (Mosher et al., 2013). However, overexpression of the rice PSKR1 gene activates SA signaling and enhances resistance to the bacterial pathogen Xanthomonas oryzae pv. oryzicola (Yang et al., 2019). Furthermore, exogenous application of PSK enhances Pst DC3000 growth in the Arabidopsis tpst-1 mutant (Mosher et al., 2013), and increases resistance to Botrytis cinerae in tomato (Zhang et al., 2018). In addition, silencing of the tomato PSKR1 gene enhances susceptibility to B. cinerae (Zhang et al., 2018). Binding of PSK to tomato PSKR1 elevates cytosolic [Ca²⁺], which enhances interaction between calmodulins and auxin biosynthetic YUCCAs, resulting in auxin-dependent immunity against B. cinerae (Zhang et al., 2018).

GRIM REAPER Peptide—POLLEN-SPECIFIC RLK5

GRIM REAPER (GRI) belongs to a small family with six members in Arabidopsis. Its C-terminal Cys-rich domain is highly homologous to STIGMA-SPECIFIC PROTEIN1 that functions in regulation of exudate secretion in the pistils and promotion of pollen tube growth (Verhoeven et al., 2005; Huang et al., 2014). The GRI protein is 169-amino acids long, carries a predicted N-terminal signal peptide (amino acids 1–30), and is secreted into the apoplast (Wrzaczek et al., 2009). As the GRI gene expression in flowers is 1,000-fold higher than in leaves (Wrzaczek et al., 2009), GRI likely plays a role in reproduction. Indeed, a gain-of-function gri mutant and GRI-overexpressing plants exhibit reduced seed content in the siliques (Wrzaczek et al., 2009). Interestingly, the low basal GRI expression in leaves is inducible by ozone exposure and both gri and GRI-overexpressing plants are sensitive to ozone (Wrzaczek et al., 2009). The gri mutant is also resistant to the virulent bacterial pathogen Pst DC3000 (Wrzaczek et al., 2009). These gri phenotypes are likely caused by accumulation of a GRI peptide (GRIp) corresponding to the N-terminal variable region after the signal peptide (amino acids 31–96; Wrzaczek et al., 2015). Exogenous GRIp^31–96 induces superoxide- and SA-dependent ion leakage, an indicator of cell death. GRI is cleaved by an apoplast-localized type II metacaspase METACASPASE9, releasing an 11-amino acid peptide, GRIp^68–78, which is sufficient for induction of ion leakage (Wrzaczek et al., 2015). GRIp-induced ion leakage depends on the atypical LRR-RK, POLLEN-SPECIFIC RECEPTOR-LIKE KINASE5 (PRK5; Wrzaczek et al., 2015). Full-length GRI without the signal peptide and GRIp^31–96 interact with the extracellular domain of PRK5 in vitro. A radiolabeled GRIp, ¹²⁵I-Y-GRIp^68–78, which is active for ion leakage induction, binds to Arabidopsis membrane extracts with a K_d of 1.9 nM. Binding of ¹²⁵I-Y-GRIp^68–78 to membrane extracts is reduced to background levels in prk5 mutants (Wrzaczek et al., 2015). These results support that PRK5 is a receptor of GRIp. However, because the prk5 and mc9 mutations have no significant effects on extracellular superoxide-induced ion leakage and resistance to Pst DC3000 (Wrzaczek et al., 2015), whether GRIp is a bona fide DAMP requires further investigation.

PAMP-INDUCED SECRETED PEPTIDE1—RLK7

Genes encoding PAMP-INDUCED SECRETED PEPTIDE (PIP) precursors named prePIP1, prePIP2, and prePIP3 were identified by searching flg22- and elf18-induced transcription data (Hou et al., 2014). Eleven prePIP homologs were identified in Arabidopsis based on the highly conserved C-terminal sequences. All of the prePIP family members carry a N-terminal signal peptide (Hou et al., 2014; Vie et al., 2015). Orthologs of prePIPs were also identified in multiple other plant species such as soybean (Glycine max), grape (Vitis vinifera), maize, and rice (Hou et al., 2014). The prePIP1 gene is induced not only by PAMPs but also by methyl SA, Pst DC3000, and the fungal pathogen F. oxysporum f. sp. conglutinans strain 699 (Foc 699; Hou et al., 2014). Overexpression of prePIP1 and prePIP2 inhibits root growth and enhances resistance to Foc 699. Synthetic PIP1 and PIP2 comprising the conserved C terminus also inhibit root growth and induce immune responses similar to PTI (Hou et al., 2014). Interestingly, PIP1- and PIP2-mediated root growth inhibition and immune responses are compromised in transferred DNA insertion mutants of the RLK7 gene, which encodes a class XI LRR-RK, suggesting that RLK7 is a potential receptor of these PIPs (Hou et al., 2014). Indeed, RLK7-HA was pulled down with PIP1-biotin–associated streptavidin beads from membrane extracts of transgenic Arabidopsis plants expressing RLK7-HA, and specific binding of radiolabeled ¹²⁵I-Y-PIP1 was detected in homogenates of tobacco leaves transiently expressing RLK7-HA in photoaffinity labeling assays, indicating that PIP1 directly binds to RLK7 (Hou et al., 2014). Moreover, PIP1-induced root growth inhibition and/or ROS production are reduced in the bak1-4 mutant but not in the bik1 mutant, indicating that PIP1-RLK7 signaling is partially dependent on BAK1, but independent of BIK1 (Hou et al., 2014). Finally, both PIP1 and PEP1 induce the expression of PrePIP1, ProPEP1, RLK7, PEPR1, and FLS2, suggesting that PIP1 and PEP1 function cooperatively in amplification of FLS2-initiated immune signaling (Hou et al., 2014).

IDA-LIKE6 Peptide—HAESA/HAE-LIKE2

INFLORESCENCE DEFICIENT IN ABSCISSION (IDA) and IDA-LIKE (IDL) proteins are precursors of peptides that induce floral abscission (Butenko et al., 2003; Stenvik et al., 2008). The Arabidopsis IDA family has nine members (IDA and IDL1–8) characterized by an N-terminal signal peptide, a variable region, and a C-terminal conserved region where the PIP motif is located (Butenko et al., 2003; Stenvik et al., 2008; Vie et al., 2015). Genetic studies suggested that two LRR-RKs, HAESA (HAE) and HAE-LIKE2 (HSL2), are receptors of IDA/IDL-derived peptides (Stenvik et al., 2008). A chemiluminescent acridinium-labeled PIP with a Val residue at the N terminus and hydroxylation of the conserved Pro at position 7 termed “acri-PIPPo” binds to leaf materials of N. benthamiana expressing HSL2ΔKD with a K_d of ∼20 nm (Butenko et al., 2014), demonstrating that HSL2 is a bona fide receptor of IDA/IDL peptides. The IDA and IDL6 genes are upregulated by PAMPs, and IDL6 is also induced by Pst DC3000 (Hou et al., 2014; Wang et al., 2017b). Synthetic IDL6 and IDL7 extended PIP peptides downregulate the expression of a broad range of stress-responsive genes (Vie et al., 2017). Moreover, overexpression of IDL6 enhances susceptibility to Pst DC3000, whereas silencing of IDL6 increases resistance to the bacterial pathogen (Wang et al., 2017b). IDL6 elevates the transcription of Arabidopsis DEHISCENCE ZONE POLYGALACTURONASE2 (ADPG2), which encodes an active polygalacturonase that promotes pectin degradation to facilitate Pst DC3000 infection. Consistent with HAE and HSL2 being receptors of IDL6, IDL6-mediated ADPG2 expression and Pst DC3000 susceptibility are completely suppressed in the hae hsl2 double mutant (Wang et al., 2017b). Interestingly, the IDA-HEA/HSL2 ligand-receptor pair is required for P. syringae type III effector-triggered leaf abscission, which likely represents a new form of plant immunity (Patharkar et al., 2017).

CONCLUSIONS AND FUTURE PERSPECTIVES

A large and compelling body of evidence has accumulated in recent years, which supports an important role for DAMPs in plant immune responses (Fig. 1). Nevertheless, the identity of DAMPs in plants remains to be unambiguously defined. The Danger model postulates that healthy cells or cells undergoing normal physiological death do not generate danger signals (Matzinger, 1994, 2002). It was recently further argued in animals that a canonical DAMP can be upregulated, but not released, in response to PAMP detection or stress stimuli that presumably lead to necrosis (Martin, 2016). In plants, however, it seems that some DAMPs are actively released upon PAMP detection or environmental stresses (Deng et al., 2015; Chen et al., 2017). Release of DAMPs in the absence of cell death appears to be inconsistent with the Danger model. However, before we arrive at such a conclusion, we must consider the following possibilities. First, some DAMPs may play dual functions in plants. For instance, as in animals (Trautmann, 2009), eATP in plants not only acts as a DAMP in wound response, but also plays a major role in growth control (Choi et al., 2014b; Roux, 2014). The constitutive eATP and actively released ATP may be crucial for cell viability and growth changes (Chivasa et al., 2005; Liu et al., 2012; Deng et al., 2015). Second, the amount of DAMPs actively released may not be sufficient for immune activation. For example, in response to cold stress (4°C for 7 d), the concentration of eATP in the extracellular root medium of 7-d–old Arabidopsis seedlings is ∼8 nM, whereas that in the fluid released at the sites of physical wounding is ∼40 μm (Choi et al., 2014a; Deng et al., 2015). The eATP concentration under cold stress is likely too low to activate the eATP receptor DORN1 (K_d ∼46 n m) for wound response (Choi et al., 2014a). These results suggest that DAMPs may induce immune responses in a concentration-dependent manner, or there may be a threshold below which DAMPs do not activate immune response. Third, because plants lack specialized immune cells and adaptive immunity, cell-autonomous immunity may play a more important role in plants than in animals (Randow et al., 2013). Plants might have thus evolved mechanisms to actively release high amounts of DAMPs for activation of cell-autonomous immunity. Clearly, further investigations are required to determine whether sufficient DAMPs can be released in the absence of cell death for immune activation in plants. Regardless, even though the Danger model may need some modifications for the plant immune system, the general principles should be applicable.

Figure 1.

Putative DAMP-receptor pairs and their functions in plant immunity. The cell surface receptor cartoons depict the putative DAMP receptors with their coreceptors or associated proteins. The cartoon for the Glu receptors GLR3.3/GLR 3.6 is based on an animal iGluR, a ligand-gated ion channel formed by four subunits. Each subunit has four domain layers: the extracellular N-terminal domain and ligand-binding domain, the transmembrane domain, and an intracellular C-terminal domain. For the sake of clarity, only two subunits are shown in the cartoon for GLR3.3/3.6. Moreover, although several RALFs including RALF17, RALF23, RALF33, and RALF34 are potential DAMPs that positively or negatively regulate immunity, RALF23 has been shown to bind LLG1/2 and FER to nucleate the assembly of RALF23-LLG1/2-FER heterocomplexes. Thus, only RALF23-LLG1/2-FER–mediated inhibition of PTI is presented here. Note that both LLG1 and FER are required for PTI signaling. In addition, although PIP1-induced–ROS production and root-growth inhibition partially depend on BAK1, whether the PIP1 receptor RLK7 interacts with BAK1 has not been reported. A question mark (?) is thus included in the RLK7/BAK1 cartoon to illustrate the uncertainty. Finally, dashed arrows are used to indicate the immune responses that are induced either by exogenously added DAMPs or by overexpression of the receptors; however, whether these immune responses are induced by the DAMPs through their receptors is unclear. By contrast, solid arrows represent immune responses that are activated by the DAMPs through their receptors. EGF, epidermal growth factor; IDL6p, INFLORESCENCE DEFICIENT IN ABSCISSION-LIKE6 peptide; INR, INCEPTIN RECEPTOR; SOBIR1, SUPPRESSOR OF BIR1-1; TM, transmembrane.

It is expected that multiple DAMPs would be released upon any type of cell damage. However, the combinations of DAMPs following different types of cell damages may be different. For instance, besides primary DAMPs, mechanical damage leads to release of wounding-induced secondary DAMPs such as systemin (Pearce, 2011), whereas pathogen attack results in release of pathogen-induced secondary DAMPs including Peps and PIPs (Huffaker et al., 2006; Hou et al., 2014). Moreover, DAMPs may be released at various times during plant-microbe interaction due to their different subcellular localizations. In this regard, DAMPs derived from the cell wall would be released early, followed by those from the cytoplasm, and finally from the nucleus. Additionally, the half-lives and apoplastic mobility of DAMPs as well as the activities of receptors for DAMPs may differ significantly (Adriouch et al., 2012). Thus, DAMPs should function cooperatively with each other, as well as with PAMPs in a temporal, spatial, and stress-specific manner to generate a peculiar immune response.

Determining the role of DAMPs in plant immune responses is an important but challenging task (see Outstanding Questions). Several studies have investigated the interplays between DAMPs and PAMPs (such as flg22 and elf18) as well as between different DAMPs. Both synergism and antagonism between DAMPs and PAMPs/DAMPs have been observed (Fauth et al., 1998; Stennis et al., 1998; Aslam et al., 2009; Ma et al., 2012; Flury et al., 2013; Tintor et al., 2013; Stegmann et al., 2017). However, because DAMPs are released during pathogen infection or herbivore attack, the context is extremely complex. It would be difficult to sort out the contribution of individual DAMPs to the final specific immune phenotype. Perhaps something similar to the recently proposed PAMP/DAMP combination-based “inflammatory code” could help solve this puzzle (Escamilla-Tilch et al., 2013). Moreover, although evidence supporting a role for DAMPs in effector-triggered immunity is accumulating (Ma et al., 2012; Zhang and Mou, 2012), in-depth investigations are warranted. In addition, several DAMPs have been implicated in systemic responses including SAR (Pearce et al., 1991; Ross et al., 2014; Toyota et al., 2018; Wang et al., 2019a), suggesting that DAMP signaling is an integral component of biological induction of systemic responses. Future research should investigate whether DAMPs move systemically or act through other signal molecules similarly to systemin (Li et al., 2002b; Schilmiller and Howe, 2005), and how the DAMP signal is transduced into the nucleus.

It is worth mentioning that what we currently know about DAMPs is just the tip of the iceberg. Among the countless numbers of intracellular molecules, many could potentially become DAMPs if released into the apoplast. Furthermore, a recent study using a bioinformatics approach identified >1,000 putative secreted peptides in Arabidopsis (Lease and Walker, 2006), not to mention other plant species with larger genomes than Arabidopsis. Many of the putative peptides could potentially function as DAMPs. Identification of potential new DAMPs as well as the processing enzymes and/or receptors for the candidate DAMPs would greatly improve our understanding of plant DAMP signaling and the plant immune system as a whole. It is expected that a deeper understanding of plant DAMPs and the plant immune system could significantly help design new strategies to breed crop varieties with increased resistance against pathogens and/or herbivores.