Arabidopsis thaliana CYCLIC NUCLEOTIDE‐GATED CHANNEL2 mediates extracellular ATP signal transduction in root epidermis

Summary Damage can be signalled by extracellular ATP (eATP) using plasma membrane (PM) receptors to effect cytosolic free calcium ion ([Ca2+]cyt) increase as a second messenger. The downstream PM Ca2+ channels remain enigmatic. Here, the Arabidopsis thaliana Ca2+ channel subunit CYCLIC NUCLEOTIDE‐GATED CHANNEL2 (CNGC2) was identified as a critical component linking eATP receptors to downstream [Ca2+]cyt signalling in roots. Extracellular ATP‐induced changes in single epidermal cell PM voltage and conductance were measured electrophysiologically, changes in root [Ca2+]cyt were measured with aequorin, and root transcriptional changes were determined by quantitative real‐time PCR. Two cngc2 loss‐of‐function mutants were used: cngc2‐3 and defence not death1 (which expresses cytosolic aequorin). Extracellular ATP‐induced transient depolarization of Arabidopsis root elongation zone epidermal PM voltage was Ca2+ dependent, requiring CNGC2 but not CNGC4 (its channel co‐subunit in immunity signalling). Activation of PM Ca2+ influx currents also required CNGC2. The eATP‐induced [Ca2+]cyt increase and transcriptional response in cngc2 roots were significantly impaired. CYCLIC NUCLEOTIDE‐GATED CHANNEL2 is required for eATP‐induced epidermal Ca2+ influx, causing depolarization leading to [Ca2+]cyt increase and damage‐related transcriptional response.

Arabidopsis has a family of 20 CNGC subunits, with members contributing to [Ca 2+ ] cyt signatures evoked by abiotic stress, pathogen attack, and hormones (Jarratt- Barnham et al., 2021). Because eATP accumulates during pathogen infection and acts as a damage-associated molecular pattern (DAMP) that drives a transcriptional response through P2K1/DORN1 (Choi et al., 2014;Jewell et al., 2019;Kumar et al., 2020), CNGCs involved in pathogen sensing could also be acting in the eATP pathway. CYCLIC NUCLEOTIDE-GATED CHANNEL2 is a key candidate for testing, as it operates in root signalling (Chakraborty et al., 2021), it is involved in both DAMP and pathogen-associated molecular pattern (PAMP) signalling, and it generates a PM hyperpolarization-activated Ca 2+ -permeable channel conductance (Qi et al., 2010;Tian et al., 2019). Cyclic Nucleotide-Gated Channel2's closest paralogue, CNGC4, can interact with CNGC2, and these two subunits are hypothesized to form a heteromeric channel in PAMP signalling (Chin et al., 2013;Tian et al., 2019). Cyclic Nucleotide-Gated Channel2 and CNGC4 could potentially work together in the eATP pathway.
Here, two Arabidopsis cngc2 loss of function mutants were used: cngc2-3 and defence not death1 (dnd1; which expresses cytosolic aequorin). Extracellular ATP-induced depolarization of PM voltage has been used as a diagnostic of PM Ca 2+ channel activity in single epidermal and cortical root cells. Results show an absolute requirement for CNGC2 but not CNGC4 in the epidermis. Patch clamp electrophysiological analysis of eATP-induced PM Ca 2+ influx conductance of epidermal cells confirmed an absolute requirement for CNGC2. Both root eATP-induced [Ca 2+ ] cyt signature and transcriptional response were impaired by loss of CNGC2 function.

Membrane potential measurements
Plasma membrane potential E m of root elongation zone cells was measured using a glass microelectrode. A plant was fixed in a plexiglass chamber and immersed in assay solution (10 ml) containing 2 mM calcium chloride (CaCl 2 ; with or without 5 mM ethylene glycol-bis(b-aminoethyl ether)-N,N,N 0 ,N 0tetraacetic acid) (EGTA) or with or without 0.5 mM lanthanum chloride (LaCl 3 )), 0.1 mM potassium chloride (KCl), 1 mM MES-Tris (pH 6.0) for at least 30 min before impalement. Microelectrode construction, recording circuitry, and impalement are described in Methods S3. After observing a stable E m (> 6 min), eATP (ATP magnesium salt (MgATP) or ATP disodium salt (Na 2 ATP); Sigma) was added to the chamber (final concentration 300 µM in the assay medium, pH 6.0). In controls, magnesium sulphate (MgSO 4 ) or sodium sulphate (Na 2 SO 4 ) was added.

Patch clamp recordings
Protoplasts were isolated from root elongation zone epidermis, with origin confirmed using the N9093 epidermal-specific green fluorescent protein reporter line as described by Wang et al. (2019). Details of isolation, patch clamp solutions, and protocols are in Methods S4.

Cytosolic free calcium ion measurement
Excised primary roots of Col-0 and dnd1 expressing cytosolic (apo) aequorin were used for luminescence-based quantification of [Ca 2+ ] cyt . Roots were placed individually into a 96-well plate (one root per well) and incubated overnight at room temperature in darkness with 10 µM coelenterazine in 100 µl of buffer: 2 mM CaCl 2 , 0.1 mM KCl, 1 mM MES-Tris (pH 5.6). CaCl 2 was included to maintain a similar level to that of the growth medium. Samples were washed with coelenterazine-free buffer and left to recover for at least 20 min in darkness. A FLUOstar Optima plate reader (BMG Labtech, Ortenberg, Germany) was used to record luminescence as described in Matthus et al. (2019b). [Ca 2+ ] cyt was calculated as described by Knight et al., 1997.

Analysis of gene expression
Total RNA was extracted from roots (frozen in liquid nitrogen) using the RNAeasy Plant Mini Kit (Qiagen) and subjected to DNase I treatment (RNAse-free DNAse kit; Qiagen). Complementary DNA (cDNA) was synthesized using the QuantiTect Reverse Transcription Kit (Qiagen). Quantitative real-time (qRT)-PCR was performed in a Rotor-Gene 3000 thermocycler with the Rotor-Gene TM SYBR ® Green PCR Kit (Qiagen). UBQ10 and  Table S1. Further details are in Methods S5.

Statistical analysis
Data normality was first analysed with the Shapiro-Wilk test in R. Student's t-test or Tukey's honestly significant difference was used for parametric data comparison, whereas the Mann-Whitney U test was used to compare the nonparametric data.

Results
AtCNGC2 mediates the extracellular-ATP-induced depolarization of root epidermal plasma membrane voltage and does not require AtCNGC4 The stable resting membrane voltage E m of a single Col-0 root elongation zone epidermal cell ( Fig. 1a) was significantly but transiently depolarized by 300 µM eATP (Fig. 1b). This concentration of eATP was found previously to activate a PM Ca 2+ influx conductance in this cell type . Mean maximal depolarization from À118.9 AE 4.8 to À69.2 AE 7.6 mV ( Fig. 1c,d;  (Figs 1g,h, S1d). The loss-offunction cngc2-3 mutant (T-DNA insert in second exon) and the complemented cngc2-3,CNGC2::CNGC2 mutant ( Fig. S2a-c) were then analysed. Expression levels of P2K1/DORN1 and the coreceptor P2K2 were normal in cngc2-3 roots, indicating that eATP perception itself would be unimpaired (Fig. S2d). There were no significant differences in resting E m between genotypes (Table S2). In contrast to Col-0, 300 µM eATP failed to depolarize cngc2-3 E m (Fig. 1b-d; Table S2). Complementation fully restored the mutant's E m response to eATP (depolarization and recovery time) (Fig. 1b-e), but maximum E m depolarization occurred sooner than in Col-0 ( Fig. 1f). This may reflect the approximately doubled abundance of CNGC2 transcript in the complemented mutant, although this was not statistically significant ( Fig. S2e). To verify the cngc2-3 results, the CNGC2 dnd1 mutant (Fig. S3a-c) was also tested. This has a single point mutation causing a stop codon in the third exon and expresses cytosolic aequorin (Qi et al., 2010). Resting dnd1 E m was not significantly different to those of other genotypes and was unaffected by eATP treatment (Fig. S3df; Table S2). These results show that the eATP-induced and Ca 2+dependent PM E m response is reliant on CNGC2. Elongation zone epidermal cells of the dorn1-3 loss-of-function mutant, the dorn1-1 kinase mutant, and the p2k2 mutant all retained a small but significant depolarization of E m when challenged with 300 µM eATP ( Fig. S4a-d; Table S2). CNGC2 transcript levels were normal in both dorn1-3 and p2k2 mutant roots, so their lowered response is most likely due to loss of receptor function rather than channel function (Fig. S4e). The dorn1-3p2k2 double mutant (p2k1p2k2) also sustained a small but significant depolarization of E m when challenged with 300 µM eATP, but this was not significantly different to that caused by the Na 2 SO 4 control ( Fig. S5a-c; Table S2; P = 0.74). Under control conditions, the p2k1p2k2 mutant had a significantly more negative E m (À143.9 AE 4.3 mV; n = 10) than its paired Col-0 wild-type (À129.9 AE 4.6 (n = 5); P = 0.005), and this may help explain why sodium ions (Na + ) caused a depolarization in this mutant but not in Col-0. Overall, the results suggest that the two receptors working together are sufficient to initiate the eATP-induced depolarization of E m and that CNGC2 is an absolute requirement in this cell type.
Cyclic Nucleotide-Gated Channel2 has been shown to interact with CNGC4 in immune signalling (Chin et al., 2013;Tian et al., 2019). Here, the root elongation zone epidermis of the cngc4-5 lossof-function mutant ( Table S2). These results show that CNGC2 controls the PM E m response to eATP without the need for CNGC4.

Plasma membrane calcium-ion currents induced by extracellular ATP in Col-0 root epidermal protoplasts require CNGC2
Whole-cell currents across the PM of root elongation zone epidermal protoplasts Wang et al. (2019) of Col-0 and cngc2-3 were recorded. No significant differences in control currents or reversal potential were found between genotypes (mean AE SE reversal potential: Col-0 À59 AE 16.3 mV, n = 4; cngc2-3 À35 AE 8.9, n = 4). For Col-0, 300 µM eATP activated wholecell inward current upon membrane hyperpolarization, but not outward current upon membrane depolarization (Fig. 2a). No effect of Na + as the salt control was found in previous trials (Wang et al., 2018. Analysis of the reversal potential of eATPactivated currents (average control (no ATP) currents were subtracted from average eATP-activated currents (Wang et al., 2013)) revealed an approximate value of +22 mV (n = 4), far from the equilibrium potentials of potassium ions (K + ; À79 mV) and chloride ions (À28 mV) and indicating Ca 2+ permeability. Extracellular ATP-activated inward current was significantly inhibited by 100 µM gadolinium ions (Gd 3+ ), a plant Ca 2+ channel blocker that is effective against CNGC2 (Demidchik et al., 2009;Wang et al., 2018Wang et al., , 2019Tian et al., 2019;Fig. 2a). These results suggest that Ca 2+ influx across the PM contributed to the eATP-activated current in Col-0. As Gd 3+ is an effective blocker of a variety of PM Ca 2+ -permeable channels (Demidchik et al., 2002(Demidchik et al., , 2009Wang et al., 2018Wang et al., , 2019 it is likely that it also blocked Ca 2+ -permeable channels that were not activated by eATP, causing the significant reduction in inward current in the presence of both eATP and Gd 3+ to below the control value. The eATP-activated Ca 2+ inward current was absent from dorn1-3 PM (Fig. S7) Col-0 epidermal cells (c. À120 mV) the eATP-activated current would deliver Ca 2+ to the cytosol, which would both elevate [Ca 2+ ] cyt and initiate depolarization. It can be inferred that some eATP-activated Ca 2+ influx should have occurred in membrane voltage trials at the less negative E m caused by EGTA (À85.2 AE À5.4 mV; Figs 1(g,h), S1c) but this was not observed, further supporting the role of Ca 2+ influx in eATP-induced depolarization of E m . In contrast to Col-0, PM whole-cell currents of cngc2-3 (either inward or outward) failed to respond to 300 µM eATP (Fig. 2b). Gd 3+ (100 µM) blocked inward and outward currents in the presence of eATP, but these currents were not investigated further (Fig. 2b). Thus, the results strongly suggest that the eATPactivated inward current in Col-0 would be due to the hyperpolarization-activated Ca 2+ influx through CNGC2, helping to explain how eATP failed to depolarize the E m of the cngc2 mutants.

Extracellular-ATP-induced cytosolic free calcium ion increase in roots is impaired in dnd1
The requirement for CNGC2 in eATP-activated epidermal PM depolarization and Ca 2+ influx conductance should manifest in impaired eATP-induced [Ca 2+ ] cyt elevation in the dnd1 mutant, which expresses cytosolic (apo)aequorin as a bioluminescent [Ca 2+ ] cyt reporter. The typical monophasic [Ca 2+ ] cyt increase ('touch response') after sodium chloride (NaCl) addition (control for mechanostimulation and cation effect of Na 2 ATP) was observed in individual roots of Col-0 and dnd1. The amplitude of the touch peak and total [Ca 2+ ] cyt mobilized did not differ significantly between genotypes (Fig. 3a). By contrast, 300 µM eATP caused a biphasic [Ca 2+ ] cyt increase (after the touch response) in both Col-0 and dnd1 roots (Fig. 3b), confirming that this part of the [Ca 2+ ] cyt signature was caused by eATP. This biphasic signature ('peak 1' and 'peak 2') was observed in previous studies on Arabidopsis roots and seedlings using aequorin (Demidchik et al., 2003;Tanaka et al., 2010;Matthus et al., 2019a,b;Mohammad-Sidik et al., 2021) and also root tips using YC3.6 (Tanaka et al., 2010). dnd1 roots were significantly impaired in the amplitude of both of the eATP-induced [Ca 2+ ] cyt peaks and also total [Ca 2+ ] cyt mobilized (Fig. 3d). Significant impairment was also observed at 100 µM and 1 mM eATP (Fig. S8). Since P2K1/DORN1 governs the eATP-induced [Ca 2+ ] cyt signature in Arabidopsis roots (Matthus et al., 2019a), impairment of the [Ca 2+ ] cyt response in dnd1 helps place CNGC2 downstream of that eATP receptor, consistent with the electrophysiological data presented here. relationships of Col-0 before (À), after (+) ATP and in 100 µM gadolinium ions (Gd 3+ ; dark blue; the calcium channel blocker was applied after eATP treatment) (n = 4). Right panel: comparison of the inward currents at À190 mV (solid bars) and the outward currents at +50 mV (hollow bars) before and after eATP addition and in the presence of Gd 3+ . Gd 3+ block of control inward currents is also evident. (b) As (a), but for cngc2-3 protoplasts. The mutant did not respond to eATP even with an extended observation period (10 min). Data are means AE SE (n = 4; *, P < 0.05; ns, not significant).
New Phytologist (2022)  Root cortical plasma membrane depolarization does not require CNGC2 but may require CNGC4 The residual eATP-induced [Ca 2+ ] cyt increase seen in dnd1 roots suggests CNGC2-independent Ca 2+ influx pathways in other cells, such as the cortex. Cortical cells also increase [Ca 2+ ] cyt in response to eATP (Krogman et al., 2020). Cyclic Nucleotide-Gated Chan-nel2 redundancy was investigated by measuring elongation zone cortical cell E m . Resting Col-0 cortical cell E m was À131.6 AE 9.1 mV ( Fig. S9a; Table S2), which was not significantly different to the epidermis. Application of eATP (300 µM) to the root transiently and significantly depolarized the cortical PM ( Fig. S9a; Table S2). There was no significant difference between cortex and epidermis in terms of the maximum depolarization amplitude, the time to reach the maximum depolarization, or recovery time. The E m of elongation zone cortical cells in the two CNGC2 mutants was then investigated. Unlike the null response of epidermal cells of cngc2-3 and dnd1, addition of eATP to the root triggered cortical E m depolarization in both mutants (Fig. S9b,c; Table S2). No significant difference in the PM E m before (no ATP added) or after ATP (ATP added) was observed between Col-0 and these two mutants (Fig. S9e), indicating that CNGC2 is not involved in this cell type. The cngc4-5 mutant still supported a significant depolarization of cortical E m when eATP was added to the root (Fig. S9d,e; Table S2), but this was significantly smaller than that found previously in its epidermal cells (cortex, 21.6 AE 6.8 mV; epidermis, 62.2 AE 8.8 mV; P = 0.012). This indicates a CNGC4-dependent pathway in the cortex. The residual depolarization in the cngc4-5 implies involvement of other CNGCs (but not CNGC2) or other transport systems (Fig. S9f). Together, the results help explain the residual eATP-induced [Ca 2+ ] cyt increase in dnd1 roots; CNGC2 does not operate in all other cells.

Discussion
Effects of eATP on plants were reported almost half a century ago (Jaffe, 1973), yet relatively few components of eATP signalling pathways have been identified. A forward genetic screen based on eATP's ability to increase [Ca 2+ ] cyt led to the identification of the first angiosperm eATP receptor, P2K1/DORN1 (Choi et al., 2014). Here, eATP's ability to depolarize root PM E m (Lew & Dearnaley, 2000) was used in a targeted gene approach. Depolarization can arise from Ca 2+ influx across the PM (Dindas et al., 2018), and eATP causes a rapid [Ca 2+ ] cyt increase in roots that could initiate depolarization (Waadt et al., 2020) as a multiconductance process . Here, eATP-induced depolarization required extracellular Ca 2+ (Figs 1g,h, S1c,d), showing its reliance on Ca 2+ influx. Thus, the unresponsiveness of cngc2 mutant root elongation zone epidermal PM to eATP ( Fig. 1) is consistent with its lack of eATP-induced PM Ca 2+ influx currents (Fig. 2) and reveals CNGC2 as a necessary component for initiating depolarization downstream of P2K1/DORN1/P2K2 in young epidermal root cells (Fig. 4e).
Cyclic Nucleotide-Gated Channel2 works together with CNGC4 in PAMP signalling, acting as a heterotrimeric Ca 2+ channel in the flagellin 22 pathway (Chin et al., 2013;Tian et al., 2019). During the course of this study, Wu et al. (2021) reported that Arabidopsis pollen grain PM has an eATP-activated Ca 2+ influx conductance, measured using whole-cell patch clamp electrophysiology. This conductance was impaired in both a single mutant of CNGC2 and a single mutant of CNGC4, suggesting that these two channel subunits might work together to facilitate germination. Whether CNGC2 and CNGC4 underpin eATP-induced [Ca 2+ ] cyt elevation and transcription in pollen remains untested. Here, with eATP as a potential DAMP, CNGC2 could be acting either as a homotetramer or a heterotetramer (that includes CNGC4) in the root epidermis, but in either event it is the obligate component of the depolarization response given CNGC4's redundancy (Fig. S5e-g; Table S2). If a heterotetramer included CNGC4 (which is expressed at almost half the level of CNGC2 in the epidermis; Dinneny et al., 2008), that CNGC4 subunit could be replaced. This is in contrast to CNGC4's pivotal role in the PAMP signalling CNGC2/4 heterotetramer, where CNGC4 is the phosphorylation target of the BIK1 kinase (Tian et al., 2019).
A residual [Ca 2+ ] cyt signature and a transcriptional response were still observed in CNGC2 mutants, showing that other channels are involved in the root's overall response to eATP that now need to be identified. The results here from the cortex implicate a role for CNGC4 (Figs 4e, S9f). Annexin1 is implicated at whole root level, but its mode of action is not yet determined (Mohammad-Sidik et al., 2021). Extracellular ATP's upregulation of defence-related and wound-response genes MPK3, WRKY40, CPK28, and MC7 is P2K1/DORN1 dependent (Choi et al., 2014;Jewell et al., 2019) and was significantly impaired here in cngc2-3 (Fig. 4). Metacaspase 7 expression can be upregulated by the necrotrophic fungus Alternaria brassicicola (Kwon & Hwang, 2013). Its CNGC2-dependent upregulation by eATP may relate specifically to DAMP signalling following ATP release by damaged cells. Wounded root cells not only release ATP (Dark et al., 2011) that could act as a DAMP for their neighbours but also release another DAMP, the peptide PLANT ELICITOR PEPTIDE 1 (PEP1; Hander et al., 2019). This is perceived in neighbouring cells by the cognate PM receptors PEP1 RECEPTOR 1 (PEPR1) and PEPR2 that relay to CNGC2 to cause [Ca 2+ ] cyt elevation (Qi et al., 2010). PEPR2 is coexpressed with P2K1/DORN1 (Tripathi et al., 2017). Extracellular ATP also upregulates PEPR1 and PEPR2 transcription (Jewell et al., 2019), so CNGC2 could be a common component in these DAMP pathways to facilitate the adaptive response.

Supporting Information
Additional Supporting Information may be found online in the Supporting Information section at the end of the article.          (CNGC2) is not required for extracellular ATP (eATP)-induced depolarization of primary root elongation zone cortical plasma membrane potential but CNGC4 is involved.
Methods S2 Growth conditions. Methods S3 Membrane voltage measurement.

Methods S4 Patch clamp recordings.
Methods S5 Quantitative real-time PCR analysis of gene expression.
Table S1 Primers used for genotyping transfer DNA mutant lines and quantitative real-time PCR.