Abscisic Acid Signaling Through Cyclic ADP-Ribose in Plants

See allHide authors and affiliations

Science  19 Dec 1997:
Vol. 278, Issue 5346, pp. 2126-2130
DOI: 10.1126/science.278.5346.2126


Abscisic acid (ABA) is the primary hormone that mediates plant responses to stresses such as cold, drought, and salinity. Single-cell microinjection experiments in tomato were used to identify possible intermediates involved in ABA signal transduction. Cyclic ADP–ribose (cADPR) was identified as a signaling molecule in the ABA response and was shown to exert its effects by way of calcium. Bioassay experiments showed that the amounts of cADPR in Arabidopsis thalianaplants increased in response to ABA treatment and before ABA-induced gene expression.

Plants endure environmental challenges such as drought, salinity, or cold by adjusting rather than escaping. These responses are mediated by ABA (1), which through unknown signals affects the regulation of many genes (2-7). One signaling intermediary is calcium (Ca2+) (8). ABA-mediated increases in guard cell Ca2+ levels lead to stomatal closure (9). Ca2+ can also induce the expression of an ABA-responsive gene in maize protoplasts (10).

Three regulators of Ca2+ levels are inositol (1,4,5)-trisphosphate (IP3) (11), cyclic adenosine 5′-diphosphate ribose (cADPR), and nicotinic acid adenine dinucleotide phosphate (NAADP+) (12-15). The receptor for IP3 is known (11), whereas those for cADPR and NAADP+are not (15). A putative receptor for cADPR is the ryanodine receptor (RyR) (15). NAADP+, a metabolite of nicotinamide adenine dinucleotide phosphate (NADP+) identified in vitro, regulates a third Ca2+ channel that appears to be distinct from the IP3 and cADPR receptors (14). cADPR can be produced by ADP–ribosyl cyclase or by CD38, a lymphocyte protein, both of which use nicotinamide adenine dinucleotide (NAD+) as a precursor (16).

Both IP3 and cADPR elicit Ca2+ release from beet storage root vacuoles (17), and a RyR-like activity, sensitive to cADPR, has been detected in beet microsomes (18). Here we demonstrate that cADPR is a likely in vivo intermediate of ABA signal transduction that exerts its effects by way of intracellular Ca2+ release.

We used microinjection to screen for compounds that may be involved in ABA responses. We studied the Arabidopsis genesrd29A (also termed Iti78 and cor78) (5), a desiccation-responsive gene (3), andkin2 (also termed cor6.6) (5), a cold-inducible gene (2). Both genes are rapidly induced by ABA, without requiring new protein synthesis.

We microinjected 7- to 10-day-old etiolated hypocotyls of the phytochrome-deficient tomato mutant aurea (19,20) with rd29A-GUS, kin2-GUS(21), and potential agonists and antagonists of ABA signal transduction. The aurea mutant was chosen because it has been used in previous investigations on phytochrome signaling pathways and therefore provided convenient controls. ABA was sufficient to induce gene expression from both constructs (Table1). Neither rd29A-GUS norkin2-GUS was activated in the absence of ABA; both were activated in response to ABA, but not to phytochrome A (phyA). By contrast, cab-GUS, a light-responsive gene, was activated by phyA but not ABA. Moreover, phyA induction of cab-GUS was negatively regulated by ABA, as reported previously (22). These results recapitulated specific activation of the reporter genes by their appropriate physiological signals in our system.

Table 1

Specificity of phyA and ABA signaling pathways. Subepidermal cells of aurea hypocotyls were injected with either rd29A-GUS, kin2-GUS, orcab-GUS (19), with or without oat phytochrome A (phyA). Pipette concentrations were ∼10,000 molecules/pl for phyA and 1 mg/ml for plasmid DNA (except for CaMV35S-GUS, 0.5 mg/ml). The injected volume was ∼ 1 pl and the cell volume was estimated to be 160 pl (19).After injections, seedlings were incubated in KNOP salt medium (19) with or without ABA (50 μM) (cis-,trans-abscisic acid; Sigma) for 2 days. The results are given as the number of cells expressing GUS as a percentage of the total number of cells injected. Because the plant vacuolar volume is about 90% of the cell volume, the frequency of a micropipette hitting the cytoplasm rather than the vacuole is low. Using the constitutive 35S promoter (CaMV35S-GUS) (26) as a control, we found that 5 to 10% of the injected cells expressed GUS activity. Therefore, we regard values that fall within 5 to 10% as the maximum that can be obtained in our microinjection system. No quantitative significance is attached to numbers within this range.

View this table:

rd29A-GUS and kin2-GUS can be activated by coinjected Ca2+. ABA induction of these two genes is blocked by EGTA (3 mM), a Ca2+ chelator (9) (Table 2). cADPR, an endogenous regulator of Ca2+, triggers Ca2+ release from sea urchin microsomes (15) and from beet cell vacuoles in beet storage roots (17). To determine if cADPR could substitute for ABA or Ca2+, we coinjected cADPR with rd29A-GUS orkin2-GUS (Table 2). At 1 μM (estimated intracellular concentration) (19), cADPR replaced ABA or Ca2+in stimulating the two ABA-regulated genes. Lower (0.2 μM) or higher (2 μM) concentrations of cADPR were not as effective (23). Similar concentration dependence is often observed for ion channels, where low concentrations of agonists are below the threshold needed for activity and high concentrations lead to desensitization.

Table 2

cADPR can substitute for ABA or Ca2+ in the activation of rd29A-GUS and kin2-GUS.Aurea hypocotyls were injected with combinations of cADPR, 8-NH2-cADPR (Molecular Probes), Ca2+,rd29A-GUS, kin2-GUS, cab-GUS, and CaMV35S-GUS as indicated. Pipette concentrations of cADPR, 8-NH2-cADPR, and Ca2+ were all 160 μM. Injected seedlings were subsequently incubated in KNOP medium in the presence or absence of ABA (50 μM) or in the presence or absence of EGTA (EGTA/AM, 3 mM) for 2 days. Δ: Heated cADPR or heated 8-NH2-cADPR (25). EGTA/AM: Membrane-permeable version (Calbiochem-Novabiochem, La Jolla, California).

View this table:

EGTA blocked cADPR induction of rd29A andkin2 expression, consistent with the role of cADPR in releasing Ca2+ from internal stores (Table 2). A structural analog of cADPR and specific inhibitor of cADPR activity, 8-amino-cADPR (8-NH2-cADPR) (24), blocked the stimulation ofrd29A-GUS and kin2-GUS by ABA (Table 2). Heated cADPR or heated 8-NH2-cADPR (100°C, 10 min) was inactive (Table 2). Upon heat treatment, cADPR is hydrolyzed to ADP-ribose which is ineffective as a Ca2+-releasing agent (25).

Because cab gene is activated by phyA by way of Ca2+ (19), we examined whether cADPR could substitute for these two inducers. In contrast to rd29A andkin2, cab-GUS expression was induced by Ca2+ but not cADPR (Table 2). These results suggest that cADPR may elicit a unique cytoplasmic Ca2+ wave leading to a specific activation of ABA-responsive but not phyA-responsive genes. The expression of cab-GUS and CaMV35S-GUS(26) was not inhibited by 8-NH2-cADPR, its heated derivative, or EGTA, thus ruling out a general cytotoxic effect (Table 2).

Microinjected Aplysia ADP–ribosyl cyclase, which synthesizes cADPR (16), activates reporter gene expression in the absence of ABA (Table 3). The effect was inhibited by coinjection of 8-NH2-cADPR.Neurospora NADase, an enzyme that degrades NAD+, the precursor of cADPR (27), inhibited ABA induction of the two reporter genes (Table 3). Coinjection of cADPR (intracellular concentration 1 μM), which is insensitive to NADase, overcame the inhibition. Injection of bovine serum albumin (BSA) had no effect (Table 3).

Table 3

ADP–ribosyl cyclase activates, NADase inactivates,rd29A-GUS and kin2-GUS. AplysiaADP–ribosyl cyclase (ADPR cyclase) (7000 molecules; Sigma),Neurospora NADase (3000 molecules; Sigma), and BSA (7000 molecules; Boehringer Mannheim) were coinjected with eitherrd29A-GUS or kin2-GUS. Pipette concentrations of cADPR and 8-NH2-cADPR were 160 μM. After injections, some of the seedlings were incubated in KNOP medium with ABA (50 μM).

View this table:

To examine the effect of protein phosphorylation on ABA signal transduction, we injected seedlings and treated them with protein kinase inhibitors K252a (2 μM) (28) and staurosporine (29) (100 nM). The response of reporter genes to ABA, cADPR, and Ca2+ was abolished by K252a and diminished by staurosporine (Table 4). The phosphatase inhibitor okadaic acid (OA) (28) (100 nM) stimulated reporter gene expression in the absence of ABA (Table5). This activation was inhibited by 8-NH2-cADPR, EGTA (3 mM), and K252a (2 μM), indicating that an OA-sensitive phosphatase may be upstream of cADPR, Ca2+ pools, and kinases in ABA signaling.

Table 4

Effects of kinase inhibitors on cADPR- and ABA-induced gene induction. Pipette concentrations of Ca2+and cADPR were 160 μM. Injected seedlings were incubated in KNOP medium with either staurosporine (St.) (100 nM, Calbiochem-Novabiochem) or K252a (2 μM; BIOMOL Research Laboratories) for 2 days in the presence or absence of ABA (50 μM).

View this table:
Table 5

Okadaic acid (OA) mimics ABA responsiveness. Hypocotyls were injected with reporter genes with or without 8-NH2-cADPR (pipette concentration, 160 μM). After injections, seedlings were incubated in KNOP medium containing okadaic acid (OA, 100 nM, BIOMOL Research Laboratories) for 2 days. K252a (2 μM) and EGTA (EGTA/AM, 3 mM) were added to the incubation medium.

View this table:

Thus, cADPR-mediated induction of ABA-responsive gene expression may be positively regulated by protein kinases and negatively regulated by protein phosphatases. ABA-mediated guard cell closure in Vicia faba shows a similar response to OA and K252a, althoughArabidopsis guard cells responded differently to the same inhibitors (7, 28).

IP3-triggered release of Ca2+ has been implicated in guard cell closure (30). We examined whether IP3 was able to induce expression of ABA-responsive genes. Indeed, rd29A and kin2 were activated by IP3, and the activation was blocked by heparin (Table6), a specific inhibitor of IP3 receptor (11). By contrast, ABA activation of the two reporter genes was not affected by heparin, and the same inhibitor had no effect on cADPR-induced gene expression (Table 6). Because phospholipase C mRNA itself is induced by ABA treatment (31), the results suggest that IP3 may be involved in the secondary rather than primary ABA response. Control experiments demonstrated that IP3 was unable to activatecab-GUS (Table 6), thus ruling out its role in phyA responses.

Table 6

IP3 may not be involved in the primary response of ABA. Pipette concentrations were as follows: IP3, 320 μM; cADPR, 160 μM; 8-NH2-cADPR, 160 μM; and heparin (Sigma), 4.8 mM. Injected seedlings were subsequently incubated in KNOP medium with or without ABA (50 μM) or with or without EGTA (EGTA/AM, 3 mM) for 2 days.

View this table:

We examined whether GTP-binding proteins (G proteins) also participate in ABA signal transduction. GDP-β-S, an antagonist of heterotrimeric G proteins and inhibitor of chs and cab gene expression (19), had no effect on rd29A-GUS orkin2-GUS expression (23) over 2 days. Likewise, a G-protein activator, GTP-γ-S, was unable to induce reporter gene expression in the absence of ABA. These results demonstrate that G proteins are not directly involved in ABA signal transduction.

To assess cADPR levels in vivo, we used transgenic Arabidopsis thaliana plants carrying a kin2-luciferase transgene (32) that were treated with ABA. After assessing reporter gene activity (33), we harvested plants and measured amounts of endogenous kin2 mRNA (34) (Fig.1A) and cADPR-triggered Ca2+-release activity (12, 34) (Fig.1B). Gene expression peaked after 4 to 8 hours of exposure to ABA (Fig.1A). Levels of LUC reporter gene activity and endogenouskin2 mRNA were similar throughout the time course. cADPR levels, as measured by the Ca2+-release assays, increased after ABA treatment and before changes in transcription of ABA-responsive genes.

Figure 1

cADPR levels during ABA induction. (A) Gene induction of kin2 after ABA treatment (50 μM) was monitored by Northern blotting (kin2 mRNA) and by in vivo imaging of plants harboring the kin2-LUCtransgene (kin2-LUC activity). Results are expressed in normalized units per gram of fresh weight, either density for mRNA or counts for reporter gene activity, and are averages of two experiments. For clarity, only the positive standard deviation is shown. (B) Ca2+-release activity of the extracts as measured during ABA induction. Ca2+-release activities are expressed in fluo3 fluorescence units induced by cADPR extracted for 1 g of fresh tissue and are the averages of three experiments. For clarity, only the positive standard deviation is shown. The double bars on graphs (A) and (B) indicate an interruption in the xaxis.

Using microinjection into tomato aurea (20) hypocotyls, we identified potential intermediates in ABA signal transduction. Agonists that could replace ABA in the microinjection experiments included OA, cADPR, and Ca2+. cADPR has been described in the chloroplast-containing unicellular protistEuglena gracilis (35). cADPR levels in E. gracilis oscillated during the cell cycle, with maximal levels detected immediately before cell division. Recently, two serine-threonine phosphatases, ABI1 and ABI2, have been reported to be directly or indirectly involved in ABA signal transduction in A. thaliana (6). These phosphatases are insensitive to OA, and their relation to the ABA signaling pathway described here remains to be elucidated. In barley aleurone protoplasts, ABA is thought to activate mitogen-activated protein kinase by way of a tyrosine phosphatase (36). This step appears to be a prerequisite for ABA induction of rab16 gene expression.

We propose that cADPR is a second messenger responsible for initiating the cascade of Ca2+ increases and subsequent Ca2+- dependent phosphorylation and dephosphorylation during ABA signal transduction.

  • * Present address: Laboratoire de Physiologie Cellulaire Végétale, Departement de Biologie Moleculaire et Cellulaire, Universite Joseph Fourier et CEA-Grenoble, 17 rue des Martyrs, F-38054 Grenoble Cedex 9, France.

  • To whom correspondence should be addressed. E-mail: chua{at}


Stay Connected to Science

Navigate This Article