Damage Signaling by Extracellular Nucleotides: A Role for Cyclic Nucleotides in Elevating Cytosolic Free Calcium?

Extracellular ATP (eATP) is now held to be a constitutive damage-associated molecular pattern (DAMP) that is released by wounding, herbivory or pathogen attack. The concentration of eATP must be tightly regulated as either depletion or overload leads to cell death. In Arabidopsis thaliana, sensing of eATP is by two plasma membrane legume-like lectin serine–threonine receptor kinases (P2K1 and P2K2), although other receptors are postulated. The transcriptional response to eATP is dominated by wound- and defense-response genes. Wounding and pathogen attack can involve the cyclic nucleotides cyclic AMP (cAMP) and cyclic GMP (cGMP) which, in common with eATP, can increase cytosolic-free Ca2+ as a second messenger. This perspective on DAMP signaling by eATP considers the possibility that the eATP pathway involves production of cyclic nucleotides to promote opening of cyclic nucleotide-gated channels and so elevates cytosolic-free Ca2+. In silico analysis of P2K1 and P2K2 reveals putative adenylyl and guanylyl kinase sequences that are the hallmarks of “moonlighting” receptors capable of cAMP and cGMP production. Further, an Arabidopsis loss of function cngc mutant was found to have an impaired increase in cytosolic-free Ca2+ in response to eATP. A link between eATP, cyclic nucleotides, and Ca2+ signaling therefore appears credible.


INTRODUCTION
Extracellular ATP (eATP) is now recognized as a plant cell regulator, with the ability to affect growth, development, and stress responses (reviewed by Matthus et al., 2019a). With resting levels thought to be in the nanomolar range, eATP concentration must be tightly regulated as either depletion or overload leads to cell death (Chivasa et al., 2005;Sun et al., 2012;Deng et al., 2015). This extracellular nucleotide is now thought to be involved in the response to wounding and microbial attack, such that it is classified as a constitutive damage-associated molecular pattern (DAMP; Choi et al., 2014b;Tanaka and Heil, 2021). It is readily envisaged that a breach of the plasma membrane would allow release of cytosolic ATP to the apoplast. Increased eATP concentration has indeed been measured from Arabidopsis thaliana roots, plus Arabidopsis and kidney bean (Phaseolus vulgaris L) leaves in response to mechanical wounding Frontiers in Plant Science | www.frontiersin.org 2 December 2021 | Volume 12 | Article 788514 (Song et al., 2006;Weerasinghe et al., 2009;Dark et al., 2011;Wang et al., 2019a). Wounding can also result in accumulation of extracellular glutamate, which in turn can cause eATP accumulation (Dark et al., 2011;Toyota et al., 2018). Non-damaging mechanical disturbance generated by touch can lead to increased eATP of roots (Weerasinghe et al., 2009), suggesting that microbial growth that does not breach the plasma membrane could cause eATP increase by putting pressure on walls. If the plasma membrane were to remain intact, studies suggest that ATP release mechanisms include exocytosis and a diversity of transport proteins (Kim et al., 2006;Choi et al., 2014b;Matthus et al., 2019a). Microbial presence has indeed been found to influence eATP levels, which in turn may affect colonization. The early phase of root colonization of both Arabidopsis and barley (Hordeum vulgare) by the endophytic fungus Serendipita indica is associated with significant eATP increase (Nizam et al., 2019). However, secretion of an eATP hydrolyzing ecto-5'-nucleotidase by the fungus lowers eATP and allows greater colonization (Nizam et al., 2019). Arabidopsis roots expressing the S. indica nucleotidase became more susceptible to colonization by the pathogenic fungus Colletotrichum incanum, suggestive of a role for eATP depletion (Nizam et al., 2019). The presence of Pseudomonas syringae (and to a lesser extent flg22) results in increased eATP levels in Arabidopsis leaf, particularly around guard cells. The high eATP causes stomatal closure, limiting bacterial entry (Chen et al., 2017). The mycotoxin beauvericin (BEA) increases eATP of wheat (Triticum aestivum) coleoptiles and causes cell death (Šrobárova et al., 2009). BEA is most notably produced by Fusarium species and acts in wheat head blight (Šrobárova et al., 2009). As a Ca 2+ ionophore (Caloni et al., 2020), BEA could act to increase Ca 2+ influx into cells. In Arabidopsis roots, promoting Ca 2+ influx results in increased eATP most likely by stimulating exocytotic ATP release (Kim et al., 2006), which may help explain BEA's ability to elevate wheat eATP. In contrast, the mycotoxin Fumonisin B1 (FB1) causes eATP depletion in Arabidopsis leading to cell death (Chivasa et al., 2005). Delineating how cells sense and respond to varying eATP levels is therefore critical to understanding the role of eATP in regulating what may be a continuum between growth, immunity, and death (Choi et al., 2014b).
While receptor characterization and analyses of receptor mutants have established eATP as a DAMP, the signaling systems downstream of receptors remain less well understood. A transient increase in cytosolic-free Ca 2+ ([Ca 2+ ] cyt ) as a second messenger is one of the earliest detectable changes downstream of eATP (Demidchik et al., 2003;Choi et al., 2014a;Behera et al., 2018). Mechanical wounding and insect feeding also increase [Ca 2+ ] cyt (Hander et al., 2019;Malabarba et al., 2021), suggesting the involvement of eATP. It can be noted in passing that saliva from insect herbivores can contain eATP hydrolyzing enzymes such as ecto-apyrase, suggesting that eATP is an important target to control (Su et al., 2012;Wu et al., 2012). Studies on how eATP increases [Ca 2+ ] cyt in Arabidopsis have focused on plasma membrane Ca 2+ -permeable channels. Pharmacological block of such channels prevents [Ca 2+ ] cyt elevation (Demidchik et al., 2003;Behera et al., 2018) and adversely affects transcription of jasmonate-dependent genes involved in wounding and defense (Tripathi et al., 2018). Patch clamp electrophysiological analyses of root epidermis have revealed eATP-and DORN1/P2K1-dependent Ca 2+ influx channel conductances that would be competent to elevate [Ca 2+ ] cyt (Demidchik et al., 2009;Zhu et al., 2017;Wang et al., 2018Wang et al., , 2019b. eATP-activated Ca 2+ influx conductances have also been found in tobacco pollen tube and Vicia faba guard cell plasma membranes (Wang et al., 2014;Wu et al., 2018). The identity of the underpinning Ca 2+ channels in all of these cell types remains under investigation. To date, the plasma membrane NADPH oxidase RBOHC (respiratory burst oxidase homologue C; Demidchik et al., 2009) and the heterotrimeric G protein alpha subunit (GPA1; Zhu et al., 2017) are implicated in regulating Arabidopsis root eATP-activated channels. Annexin1 (which supports a hydroxyl radical-activated Ca 2+ channel activity; Laohavisit et al., 2012) is implicated in eATP-induced [Ca 2+ ] cyt increase of Arabidopsis roots and transcriptional regulation of wound-and defense-response genes, but its mode of action has not been determined (Mohammad-Sidik et al., 2021). It also underpins wounding-and herbivory-induced [Ca 2+ ] cyt increases (Malabarba et al., 2021). Annexin4 supports an eATP-induced [Ca 2+ ] cyt increase when expressed in Xenopus oocytes, but its in planta activity has not yet been reported (Ma et al., 2019). Recently, a patch clamp study on Arabidopsis pollen grains identified two cyclic nucleotide-gated channel (CNGC) subunits (CNGC2 and CNGC4) as underpinning a plasma membrane eATP-activated Ca 2+ influx conductance that may be relevant to germination (Wu et al., 2021). The conductance was lost in the dorn1 loss of function mutants suggesting the channels act downstream of this receptor although it was not checked whether CNGC2 and CNGC4 expression were downregulated in dorn1. The CNGC2 and CNGC4 channel subunits form a heterotetrameric plasma membrane Ca 2+ influx channel in flg22 pathogen-associated molecular pattern-triggered immunity (PAMP; Tian et al., 2019), indicating a possible common point in DAMP and PAMP signaling. Furthermore, the results of Wu et al. (2021) imply that the cyclic nucleotides 3',5'-cyclic adenosine monophosphate (cAMP) or 3',5'-cyclic guanosine monophosphate (cGMP) act downstream of eATP and DORN/P2K1. Although eATP has long been known to increase animal cell cAMP (Choi and Kim, 1997;Castro et al., 1998), there appear to be no reports to date of eATP's having an effect on cAMP or cGMP levels in plants, but these nucleotides can increase [Ca 2+ ] cyt through CNGC activation (reviewed by Jarratt- Barnham et al., 2021). This perspective will consider the possibility that cyclic nucleotides and CNGCs act in eATP-DAMP [Ca 2+ ] cyt signaling.

CYCLIC NUCLEOTIDE PRODUCTION BY MOONLIGHTING PROTEINS IN RESPONSE TO PATHOGENS AND WOUNDING
cAMP and cGMP are involved in growth and development, abiotic stress responses, stomatal kinetics, control of photorespiration and photosynthesis, and immunity (reviewed by Jarratt- Barnham et al., 2021). Production of cAMP and cGMP is now held to be by adenylyl cyclase (AC) and guanylyl cyclase (GC) centers embedded in a diverse range of soluble and membrane-integral proteins (Ludidi and Gehring, 2003;Ruzvidzo et al., 2019;Al-Younis et al., 2021;Turek and Irving, 2021). There is clear evidence supporting cytosolic cyclic nucleotide production by a "moonlighting" function of members of the receptor-like kinase (RLK) superfamily, relevant to response to pathogens, where the catalytic center is present in the cytosolic kinase domain (Qi et al., 2010;Kwezi et al., 2011;Turek and Irving, 2021). For example in Arabidopsis, LLRAC1 (leucine-rich repeat adenylyl cyclase 1) is implicated in defense against P. syringe and the biotrophic fungus Golovinomyces orontii (Bianchet et al., 2019). Wall-associated kinase-like (WAKL) proteins in Arabidopsis, wheat and Brassica napus are implicated in cGMP production and pathogen defense (reviewed by Turek and Irving, 2021). Most recently, expression of rice WAKL21.2 (which has GC activity) in Arabidopsis resulted in transcription of SA-related defense genes and conferred resistance to P. syringae (Malukani et al., 2020).
In terms of DAMP signaling, wounding was found to result in a rapid, fivefold increase in Arabidopsis leaf cAMP and cGMP by van Damme et al. (2014). In this instance, the 2',5' isomers were detected but the mode of production and downstream consequences remain unknown. Larval oral secretions from the moth Spodoptera littoralis can increase cAMP content of Arabidopsis leaves within 3 min, which may relate to herbivory damage (Kumar et al., 2020). An initial decreased transcription followed by longer-term increased transcription of Hippeastrum hybridum Guanylyl Cyclase 1 (HpGC1, most likely encoding a soluble enzyme) was found in wounded and infected scales (fungal infection by Peyronellaea curtisii). Although a causal relationship was not demonstrated, HpGC1 activity may relate to downstream production of cGMP in the challenged scales (Świeżawska et al., 2015). Since then, triphosphate tunnel metalloenzyme 3 (TTM3) from Brachypodium distachyon has been found to have both ATP hydrolyzing and AC activities, with upregulation of transcription by wounding (Świeżawska et al., 2019). There remains a significant knowledge gap when it comes to whether eATP as a DAMP (or in its role as growth regulator) has an effect on cAMP or cGMP levels. However, increasing cAMP through expression of an AC increases transcription of wound-related genes  which could place eATP (as a DAMP governing woundrelated transcription) upstream of cAMP production. A key question is whether the eATP receptors could themselves generate cAMP or cGMP.  (Xu et al., 2018b; http://gcpred/acpred/; Figure 1A) does reveal possible sequences. Two overlapping sequences are present in DORN1/P2K1 ( Figure 1B) that harbor an aspartic acid residue ([D]) at position 3 that should confer ATP-binding specificity, but the sequences are extracellular, lying close to the transmembrane domain and beyond the residues already implicated in eATP binding in the lectin domain (Nguyen et al., 2016;Cho et al., 2017; Figure 1C). The first sequence starting at S259 has an aspartic acid residue predicted for cation binding that should follow the 14 amino acid catalytic center motif (Al-Younis et al., 2021) but after an interval of 2 amino acids. The second sequence (starting at R267) lacks a cation-binding residue but overlaps at position 4 with the AC catalytic center motif used recently by Al-Younis et al.  Figure 1A). Curiously, this second sequence contains the PHPR found previously to be an RGD-binding sequence, relevant to the RGD-containing IPI-O protein secreted by P. infestans (Gouget et al., 2006). Potentially, therefore, DORN1/P2K1 has more than one way to bind eATP. P2K2 harbors a putative AC sequence in its cytosolic serine-threonine kinase domain and this sequence conserves an aspartic acid residue for cation binding following the catalytic center motif (Figure 1D) (Xu et al., 2018a; http://gcpred; Figure 1A) revealed two intracellular sequences in DORN1/P2K1, with one harbored in its serinethreonine kinase domain ( Figure 1C). For P2K2, two sequences were found in the extracellular lectin domain and one in the intracellular serine-threonine kinase domain ( Figure 1D). (Wong and Gehring, 2013) and is present in the Pharbitis nil GC1 enzyme which has been shown to catalyze cGMP production in vitro (Szmidt-Jaworska et al., 2009 (Ludidi and Gehring, 2003;Wong et al., 2018). Both eATP receptors therefore appear to have the requisite sequences to generate cyclic nucleotides, and this now requires experimental validation.

CNGC2 IS IMPLICATED IN COTYLEDON eATP-INDUCED [Ca 2+ ] CYT INCREASE
While there remains a question mark between eATP and cyclic nucleotides, critically there is a precedent of a DAMP-cytosolic cGMP-CNGC2 pathway. Wounding of Arabidopsis increases [Ca 2+ ] cyt resulting in activation of metacaspase 4. The latter then catalyzes production of plant elicitor peptide1 (Pep1) from its inactive precursor PROPEP1 (Hander et al., 2019). Pep1 can be perceived as a DAMP in a neighboring cell by its cognate plasma membrane receptors PEPR1 and PEPR2. Both of these LRR RLKs are predicted to have cytosolic GC activity, with PEPR1 confirmed experimentally. The cytosolic cGMP activates CNGC2 to effect a [Ca 2+ ] cyt increase and downstream defense transcriptional response (Qi et al., 2010). eATP is now reported to act upstream of CNGC2. Wu et al. (2021) reported that eATP (0.1 mM) failed to elicit a plasma membrane Ca 2+ -permeable influx conductance in Arabidopsis pollen grain protoplasts of a cngc2 mutant. This concentration of eATP stimulated pollen germination in the wild type but not cngc2 (Wu et al., 2021), but it remains unknown whether activation of CNGC2 by eATP would increase [Ca 2+ ] cyt . Here, the cngc2 loss of function mutant defense not death 1 (dnd1; which constitutively expresses cytosolic (apo)aequorin as a luminescent [Ca 2+ ] cyt reporter, Qi et al., 2010) was tested for impaired eATP-induced [Ca 2+ ] cyt increase. The dnd1 mutant of CNGC2 has a single point mutation causing a stop codon in the third exon that would result in a truncated protein lacking the pore region for ion conduction. dnd1 has a dwarf rosette phenotype that was also observed here (Supplementary Figure S1). Addition of control solution alone to single, excised cotyledons of Col-0 and dnd1 only induced the transient "touch response" caused by mechanical disturbance, which did not differ between genotypes (Figures 2A,B). In contrast, addition of 0.1 mM eATP caused a monophasic increase in [Ca 2+ ] cyt after the touch response ( Figure 2C). A similar pattern was observed in a previous study by Mohammad-Sidik et al. (2021). The response of dnd1 to 0.1 mM eATP was significantly lower than the Col-0 wild type, both for the maximum [Ca 2+ ] cyt increase and estimated total [Ca 2+ ] cyt mobilized [estimated as area under the curve (AUC), Figure 2D]. Increasing eATP to 1 mM (Figures 2E,F) also revealed a significant impairment in the response of dnd1 (Figure 2G)  (Xu et al., 2018b). "X" denotes any amino acid, and "{}" denotes the number of undetermined amino acids. Centre: The AC catalytic center sequence used by Al-Younis et al. (2021). Bottom: The GCPred GC prediction tool sequence (Xu et al., 2018a). (B) The two extracellular, overlapping putative AC sequences from DORN1/P2K1 shown separately for clarity. (C) Putative AC and GC sequences of DORN1/P2K1. Overlapping AC sequences are highlighted in light gray. The extracellular lectin domain is underlined with a wave, and the intracellular serine-threonine kinase domain is denoted with a straight underline. The transmembrane domain is annotated in ochre. Amino acids found to be important for kinase activity are bold in yellow on a gray background (Choi et al., 2014b;Nguyen et al., 2016;Cho et al., 2017). (D) Putative AC and GC sequences of P2K2. The AC sequence is highlighted in light gray. Domains are identified as for (B). Amino acids found to be important for kinase activity are bold in yellow on a gray background (Pham et al., 2020).
although clearly at both eATP concentrations loss of CNGC2 function allows the majority of the [Ca 2+ ] cyt increase to occur.
Overall these results show that CNGC2 could be involved in an eATP-induced [Ca 2+ ] cyt increase in leaf tissue. As with flg22, ATP causes pronounced acidification of solutions, and so in this study, pH of eATP test solutions was maintained by buffers to avoid acidic pH artefactual responses which can impair interpretation (Westphal et al., 2019). [Ca 2+ ] cyt values were calculated as described by Matthus et al. (2019b).   Turek and Irving, 2021). Could the plant plasma membrane harbor a receptor for extracellular cAMP in close proximity to those for eATP? Two decades ago, there were no signs of a receptor for eATP in the Arabidopsis genome, but if a measurable output for sensing extracellular cyclic nucleotides could be established, then there is a way forward through forward genetic screens. Similar questions and future directions can also be posed for the extracellular putative cGMP synthesis sites of P2K2. The finding here that a small but significant component of the eATP-induced [Ca 2+ ] cyt elevation requires CNGC2 (at eATP levels consistent with wounding) helps fill a gap in our knowledge on how [Ca 2+ ] cyt elevation is achieved and helps indirectly make the case for a role for cAMP or cGMP in DAMP signaling by eATP (summarized in Supplementary Figure S2). However, it remains feasible that CNGC2 could be phosphorylated by DORN1/P2K1 or P2K2 (if downstream of these) to promote Ca 2+ influx. CNGC2 is already known to act downstream of PEPR1 and PEPR2 in DAMP signaling (Qi et al., 2010) and downstream of flg22 in PAMP signaling (Tian et al., 2019). CNGC2 therefore could be a critical component in coordinating the response to pathogens, which requires both damage and PAMPs to elicit the strongest immunogenic transcriptional outcome (Zhou et al., 2020). CNGC2 has recently been reported to be involved in auxin signaling (Chakraborty et al., 2021) so it could possibly play a part in the restorative growth response of cells surrounding damage sites as these sustain an increase in auxin as a longer-term response (Hoermayer et al., 2020). Much will depend on which cells harbor CNGC2, what its possible sub-unit partners in heterotetrameric channels are to help generate specific [Ca 2+ ] cyt signals and their position in the membrane relative to receptors. There are another nineteen CNGCs in Arabidopsis that now warrant consideration as components of eATP DAMP signaling, although CNGC14 appears not to be involved in the root's [Ca 2+ ] cyt elevation by eATP (Shih et al., 2015).

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
JS and KW designed and performed the aequorin experiments. YN and LW analyzed the data. LW designed, performed, and analyzed the biomass determinations. JD analyzed sequences, conceived the project, and wrote the manuscript with contributions from the other authors. All authors contributed to the article and approved the submitted version.

FUNDING
Funding was from the UKRI BBSRC (BB/K009869/1), the framework of the 3rd call of the ERA-NET for Coordinating Action in Plant Sciences with funding from the BBSRC (BB/ S004637/1), the University of Cambridge and Jiangsu University.

ACKNOWLEDGMENTS
We thank G. Berkowitz for the dnd1 line. We apologize to authors whose work we have not cited due to manuscript length restriction.