Identification of a Plant Receptor for Extracellular ATP

See allHide authors and affiliations

Science  17 Jan 2014:
Vol. 343, Issue 6168, pp. 290-294
DOI: 10.1126/science.343.6168.290

This article has a correction. Please see:

ATP Receptor in Arabidopsis

As well as its role as an intracellular energy source, extracellular adenosine triphosphate (ATP) has diverse functions as a signaling molecule. ATP receptors have been identified in animal cells, but searches based on structural homology have not identified ATP receptors in plants. Choi et al. (p. 290) have now identified an ATP receptor in the model plant Arabidopsis thaliana by tracking down the cause of mutations that leave mutant plants unresponsive to ATP signals. The receptor identified carries an intracellular kinase domain and an extracellular lectin domain.


Extracellular adenosine 5′-triphosphate (ATP) is an essential signaling molecule that is perceived in mammals by plasma membrane P2-type purinoceptors. Similar ATP receptors do not exist in plants, although extracellular ATP has been shown to play critical roles in plant growth, development, and stress responses. Here, we identify an ATP-insensitive Arabidopsis mutant, dorn1 (Does not Respond to Nucleotides 1), defective in lectin receptor kinase I.9 (Arabidopsis Information Resource accession code At5g60300). DORN1 binds ATP with high affinity (dissociation constant of 45.7 ± 3.1 nanomolar) and is required for ATP-induced calcium response, mitogen-activated protein kinase activation, and gene expression. Ectopic expression of DORN1 increased the plant response to physical wounding. We propose that DORN1 is essential for perception of extracellular ATP and likely plays a variety of roles in plant stress resistance.

Adenosine 5′-triphosphate (ATP), the universal energy currency in all organisms, also acts as an extracellular signaling molecule (1). The first report that ATP could play an extracellular signaling role was viewed with skepticism (2). This skepticism disappeared when the plasma membrane receptors for extracellular ATP were identified in mammals (3, 4). These receptors are of two general types: ligand-gated ion channel P2X and G protein–coupled P2Y receptors (5, 6). The various members of these two receptor classes show distinct ligand binding affinity. Spatial and temporal expression patterns of these receptors support their postulated roles in the regulation of a broad range of mammalian physiology, including neurotransmission, muscle contraction, inflammation, and cell growth and death (5).

In contrast, relatively little is known about the role of ATP as an extracellular signal in plants, and some skepticism exists as to the role this molecule may play in plants. However, a number of papers have suggested a role for extracellular ATP in plant growth (7, 8), development (9, 10), and stress responses (1113). Furthermore, the presence of extracellular ATP was visualized at actively growing root hair tips (7). ATP is actively released from plant cells in response to abiotic stresses (14), fungal elicitors (7, 15), and mechanical stimuli (16). Some plant cellular responses to ATP are similar to those seen in mammals (17), for example, cytoplasmic calcium increase, production of reactive oxygen species and nitric oxide, and the role of ecto-apyrases that regulate extracellular ATP homeostasis. However, efforts to identify plant ATP receptors through their genomic sequence homology to animal purinoceptors failed to identify any suitable candidate proteins.

To identify genes involved in extracellular ATP recognition in plants, we performed a forward genetic screen for Arabidopsis thaliana mutants impaired in their ability to respond to ATP treatment. Exogenously applied ATP triggers cytoplasmic calcium influx in wild-type Arabidopsis seedlings expressing the calcium reporter protein, aequorin (18, 19). We screened 50,000 ethylmethanesulfonate (EMS)–mutagenized seedlings for the ATP-induced calcium response (fig. S1) and identified two mutants lacking a cytoplasmic calcium response to ATP addition (Fig. 1A and fig. S2). As described below, these mutants also failed to respond to a variety of other nucleotides, except pyrimidine nucleotides, that is, cytosine triphosphate (CTP) (Fig. 2 and fig. S2). The two mutants were found to be allelic (fig. S3), and, therefore, we named them Does not Respond to Nucleotides 1-1 (dorn1-1) and dorn1-2. Addition of ATP to dorn1-1 plants failed to trigger phosphorylation of mitogen-activated protein kinase 3 (MPK3) and MPK6 as seen in wild-type plants (Fig. 1B).

Fig. 1 The dorn1 mutants are insensitive to extracellular ATP.

(A) The dorn1 mutants are defective in ATP-induced calcium response. The bar graph indicates the integrated calcium response to ATP addition for 400 s. Asterisks indicate significant differences between wild-type and mutants (means ± SEs, n = 6, *P < 0.01). (B) The dorn1-1 mutant did not phosphorylate MPK3 and MPK6 in response to 100 μM of ATP. Phosphorylation of MPK3 and MPK6 was detected by using antibody against phospho-p44/p42 mitogen-activated protein kinase. The nonspecific band was used (*) as a loading control. (C) Microarray analysis of ATP-induced gene expression in wild-type (Col-0) and the dorn1-1 mutant plants. Red dots show the genes significantly altered 30 min after treatment with 100 μM ATP (false discovery rate < 0.05, fold change ≥ 1.5). The data were based on three replicates for each treatment. Normalized probe intensities (log2 scale) were compared for ATP treatment (y axis) and buffer control (x axis). Black hatched line indicates the same intensity between two treatments, whereas black dash lines indicate 1.5-fold change.

Fig. 2 The dorn1 mutants show defects in the calcium responses to various nucleotides.

(A) Calcium responses to adenine nucleotides and its slowly hydrolyzable derivatives were absent in the dorn1 mutants. (B) Calcium responses to CTP were normal in the dorn1 mutants. (C and D) Biotic (C) and abiotic stress reagents (D, 300 mM NaCl, 4% D-glucose and 300 mM mannitol)–induced calcium responses in the dorn1 mutants were comparable to those of wild type (Col-0). All data represented as means ± SEs, n = 6 (*P < 0.05, **P < 0.01, ***P < 0.001).

To define the transcriptional response to ATP, we performed microarray analysis in wild-type and dorn1-1 mutant seedlings after the addition of ATP. In wild-type plants, 332 and 242 genes were significantly up- and down-regulated in response to ATP, respectively (Fig. 1C). However, none of these genes responded to ATP in the dorn1-1 mutant. Quantitative reverse transcriptase–polymerase chain reaction (qRT-PCR) confirmed the ATP-induced gene expression of WRKY domain transcription factor 40 (WRKY40) and calcium-dependent protein kinase 28 (CPK28) in the wild-type but not in the mutant (fig. S4). The Gene-Ontology enrichment test showed that stress-responsive genes and signal transduction–related genes were overrepresented in the list of up-regulated genes, whereas cell-wall organization genes were specifically down-regulated (fig. S5).

We identified the gene responsible for the ATP-insensitive phenotype in the dorn1 mutants by map-based cloning and whole-genome sequencing. Map-based cloning initially showed genetic linkage to a 615-kb interval at the bottom of chromosome 5 (fig. S6). We compared whole-genome sequences in wild-type, dorn1-1, and dorn1-2 and found independent point mutations in the gene, At5g60300, predicted to encode a lectin receptor kinase–I.9 (LecRK-I.9). LecRK-I.9 consists of an extracellular legume-type lectin domain, a single transmembrane domain, and an intracellular kinase domain (fig. S7). The mutations in dorn1-1 and dorn1-2 resulted in the conversion of amino acid Asp572 and Asp525 to Asn in subdomain X and in the DxWxxG motif (where D indicates Asp; x, any amino acid; W, Trp; and G, Gly) in subdomain IX, respectively, each of which is important for phosphorylation of the activation loop or stabilization of the catalytic loop of the kinase (20, 21). In vitro kinase assays demonstrated that the kinase domain of dorn1-1 and dorn1-2 is completely inactive (fig. S8). The identification of the LecRK-I.9 gene allowed isolation of a dorn1 transferred DNA (T-DNA) insertion (Arabidopsis Biological Resource Center accession code Salk_042209; dorn1-3), which has reduced transcript levels (fig. S9). The phenotype of the dorn1-3 mutant was identical to that of dorn1-1 and dorn1-2 in that the mutant did not show an increase in intracellular calcium concentration or gene induction upon the addition of ATP (Fig. 1A and fig. S10). In addition, ectopic expression of the wild-type version of LecRK-I.9 in dorn1-2 plants reinstated normal calcium influx in response to ATP (fig. S11).

If DORN1 encodes the essential receptor for extracellular ATP, then dorn1 mutants should not show defects in signaling processes other than the response to ATP. Indeed, the intracellular calcium responses of the dorn1 mutants to various biotic (flg22, chitin, elf26, and pep1) and abiotic (NaCl, mannitol, and cold water) calcium elicitors were identical to those of wild-type plants (Fig. 2). The T-DNA insertion line, dorn1-3, consistently showed a normal response to various calcium elicitors despite some variation observed in EMS mutants, possibly because of the secondary mutations.

Some animal P2Y receptors recognize other nucleotides, for example, adenosine diphosphate (ADP) and uridine triphosphate (UTP), in addition to ATP (5, 6). Therefore, we tested the calcium response of dorn1 mutants to other nucleotides. All of the dorn1 mutant seedlings showed strong defects in the calcium response to ADP and slowly hydrolyzable forms, ATPγS and ADPβS, and also to other purine nucleotides, guanosine triphosphate (GTP) and UTP. In contrast, the response to a pyrimidine nucleotide, CTP, was comparable to the level of wild-type plants (Fig. 2). We also tested the nucleotide-induced calcium responses in the DORN1 ectopic expression line (oxDORN1). These plants showed a much stronger response to most of nucleotides tested, for example, ~20-fold higher response to ATP in comparison to wild type. The relative calcium responses of the oxDORN1 line showed a rank order of ligand potency of ATP > GTP > inosine triphosphate (ITP) > thymidine triphosphate (TTP) = UTP > CTP (fig. S12), identical to that of wild-type plants (18). Our observation that the dorn1 mutants retain sensitivity to CTP (Fig. 2 and fig. S2) is consistent with evidence for an independent recognition system for CTP (18). Collectively, our results suggest that DORN1 is a key player in the recognition of various nucleotides with a preference for purine nucleotides over pyrimidine nucleotides.

Localization of the LecRK-I.9 protein to the plasma membrane (22, 23) is consistent with a role in the recognition of an extracellular ligand. Therefore, we measured ATP binding to the extracellular domain of the DORN1 recombinant protein by using radiolabeled ATP (Fig. 3A). The DORN1 protein exhibited a typical saturation curve for ATP binding with relatively high affinity [dissociation constant = 45.7 ± 3.1 nM, maximum binding capacity = 488.0 ± 6.3 pmol/mg of protein]. If DORN1 is the receptor responsible for the plant response to nucleotides, then the protein should also bind to those nucleotides shown to be biologically active with the same relative specificity as the cellular response. We assessed the ability of various unlabeled nucleotides to compete with binding of radiolabeled ATP. The results showed that unlabeled ATP and ADP were the best competitors for binding with radiolabeled ATP, with less competition by ITP, GTP, and UTP and little or no competition by other nucleotides, e.g., CTP and adenosine (Fig. 3B). This ligand specificity reflects the order of magnitude of the calcium response in the wild-type Arabidopsis (18) and the substrate potency seen with the transgenic line ectopically expressing DORN1 (Fig. 2 and fig. S12). Collectively, our results demonstrate that DORN1 is a nucleotide-binding protein with preferred affinity for ATP. Next, we conducted an in planta ATP-binding assay using Arabidopsis protoplasts (Fig. 3C). The extracellular domain including transmembrane domain of DORN1 was cross-linked with biotinylated 8-azido-ATP and was precipitated by streptavidin beads. Biotinylated 8-azido-ATP cross-linking to DORN1 was abolished by competition with unlabeled ATP (Fig. 3C), whereas CTP had no effect (Fig. 3D). Thus, DORN1 binds extracellular ATP at the cell surface.

Fig. 3 DORN1 binds ATP.

(A) Saturation binding assay for DORN1. The extracellular domain of DORN1 was incubated with the indicated concentrations of radiolabeled ATP for 30 min. Bound radiolabeled ATP was separated from free ATP by gel filtration chromatography. Data were plotted as a specific binding with SE of three replicates. (B) Competitive binding assay for DORN1. Samples containing 25 nM radiolabeled ATP in the presence of 10 nM to 10 mM of the unlabeled nucleotides were assayed for specific binding of labeled ATP. Inhibition constant (Ki) values were calculated after the data points were fitted by the one-site competition model using GraphPad Prism 5, (C and D) DORN1 binds ATP in planta. Protoplasts expressing hemagglutinin (HA)–tagged extracellular domain of DORN1 with its transmembrane domain were ultraviolet cross-linked with biotinylated 8-azido-ATP followed by precipitation with streptavidin beads. DORN1 was detected with antibody against HA. Unlabeled 1 mM ATP (C) or CTP (D) was used as a competitor.

Lectin domains are typically associated with proteins that bind carbohydrates or oligosaccharides (24). However, the legume-type lectin domain found in DORN1 lacks the conserved Ca2+ and Mn2+ binding residues that are critical for monosaccharide binding (25). Early studies of legume lectins made note of their ability to bind adenine, a component of ATP (26, 27). However, adenine was not able to compete with ATP for binding to DORN1 (Fig. 3B). Therefore, the exact ATP binding site in DORN1 remains to be determined.

Although ATP can be released from plant cells by vesicle fusion with the plasma membrane (7, 16), the levels released in this way appear to be quite low, that is, nanomolar levels (16). In contrast, ATP concentrations as high as 40 μM were recorded in the fluid released at sites of physical wounding (13). Therefore, we compared the transcriptional response to ATP (Fig. 1C) to the response to wounding (28). Sixty percent of the ATP-induced genes were also induced by wounding, with 90% of these genes responding very early to wounding (Fig. 4A). For example, out of 332 ATP-induced genes, 112 were up-regulated 15 min after wounding (Fig. 4B). Expression of selected genes after wounding was much higher in the oxDORN1 plants and notably reduced in the dorn1-3 plants (Fig. 4C). This is consistent with the gene expression pattern when ATP was applied to these plants (Fig. 4D). The data suggest that ATP is released during physical damage as a danger signal, which is then recognized by the DORN1 receptor.

Fig. 4 ATP and wounding induce a common set of genes.

(A and B) ATP up-regulated genes (Fig. 1C) were compared with transcriptome data from wounded rosette leaves (28). (A) Hierarchical clustering of ATP-induced genes with genes regulated over a time course after wounding of rosettes. Note that the majority of ATP-induced genes respond early to wounding. FC, fold change. (B) Venn diagram shows the number of ATP up-regulated genes and wound-induced genes (15 min). (C and D) Gene expression of coregulated genes by ATP and wounding in dorn1 and oxDORN1 plants. Locally wounded rosettes (C) or ATP-treated whole seedlings (D) were subjected to qRT-PCR analysis. Histograms show means with SEs as values relative to those of untreated controls. Symbols show statistically significant differences (P < 0.05) between wild type (Col-0) and dorn1-3 (#) or oxDORN1 (*).

Genomic sequence–based surveys for canonical P2X and P2Y receptors in plants failed to detect any likely candidates for an ATP receptor. This is now explained by the fact that the ATP receptor in Arabidopsis, DORN1, has a very different molecular structure from the known animal receptors. Given this unique structure, we propose the name P2K (K for kinase) for this previously unknown family of purinoreceptors, with DORN1 (P2K1) being the founding member. The P2K family is likely to be plant-specific because animals appear to lack lectin receptor kinases (29). On the basis of the phylogenetic occurrence of lectin receptor kinases, P2K receptors may be found in primitive plants (e.g., mosses) through higher plants but not in green algae. Indeed, at least one P2X receptor–like receptor has been identified in green algae (30). Given the variety of roles that extracellular ATP has been shown to play in animal systems, it is likely that further biological functions will also be found for this interesting signal molecule in plants.

Supplementary Materials

Materials and Methods

Figs. S1 to S12

Table S1

References (3142)

References and Notes

  1. Acknowledgments: We thank M. R. Knight (Durham University) for providing aequorin-expressing transgenic Arabidopsis; S. C. Peck and J. C. Anderson (University of Missouri) for technical advice on charge-coupled device camera imaging; G. A. Weisman (University of Missouri) for critical comments on the manuscript; and M. R. Dixon (University of Dallas), A. J. Witte (University of Missouri), and Y. Tanaka (Columbia, MO) for support during the mutant screening. This work was supported by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences of the U.S. Department of Energy through grant DE-FG02-08ER15309 and the Next-Generation BioGreen 21 Program Systems and Synthetic Agrobiotech Center, Rural Development Administration, Republic of Korea (grant no. PJ009068) to G.S. The accession number of the microarray discussed in this manuscript is GSE52610 in Gene Expression Omnibus. J.C., K.T., and G.S. are co-inventors on a provisional patent application titled, “Novel extracellular ATP receptor, lectin receptor kinase I.9 and its application to develop plants more stress resistant to bacterial disease and wound damage.” The authors declare that there are no competing financial interests.
View Abstract

Navigate This Article