Core Components of Abscisic Acid Signaling and Their Post-translational Modification

Abscisic acid (ABA) is a major phytohormone that regulates plant growth, development, and abiotic/biotic stress responses. Under stress, ABA is synthesized in various plant organs, and it plays roles in diverse adaptive processes, including seed dormancy, growth inhibition, and leaf senescence, by modulating stomatal closure and gene expression. ABA receptor, clade A protein phosphatase 2C (PP2C), and SNF1-related protein kinase 2 (SnRK2) proteins have been identified as core components of ABA signaling, which is initiated via perception of ABA with receptor and subsequent activation or inactivation by phosphorylation/dephosphorylation. The findings of several recent studies have established that the post-translational modification of these components, including phosphorylation and ubiquitination/deubiquitination, play important roles in regulating their activity and stability. In this review, we discuss the functions of the core components of ABA signaling and the regulation of their activities via post-translational modification under normal and stress conditions.


INTRODUCTION
Abscisic acid (ABA) is a major hormone that regulates growth, development, and responses to abiotic/biotic stress throughout the life cycle of plants. Under environmental conditions perceived as unsuitable for plant growth, ABA helps maintain seed dormancy and inhibits germination, whereas during the post-germination stage, this hormone inhibits growth to conserve energy (Chinnusamy et al., 2008;Raghavendra et al., 2010). In addition, ABA plays key roles in stomatal opening and closure, thereby contributing to the regulation of water loss under drought stress (Melotto et al., 2006;Hirayama and Shinozaki, 2007;Cutler et al., 2010;Finkelstein, 2013).
The different components of ABA signaling are modulated to varying extents by post-translational modification (Table 1), with the ubiquitination/deubiquitination and phosphorylation of these components being generally well studied. Ubiquitination of ABA core components can influence ABA signaling and responses both positively and negatively (Coego et al., 2021). Deubiquitination mediated by members of the UBIQUITIN-SPECIFIC PROTEASE (UBP) protein family affects a range of plant physiological processes, including growth, development, immunity, and stress response (Zhou et al., 2017). However, the biochemical mechanisms underlying the involvement of UBPs in ABA signaling have yet to be sufficiently elucidated. A further major category of post-translational modification is phosphorylation; ABA receptors, PP2Cs, and SnRK2s can be positively or negatively regulated by phosphorylation. For example, SnRK2s can undergo autophosphorylation and activation (Yang et al., 2017;Chen et al., 2020). In this review, we focus on the roles of ABA, core ABA signaling pathways, and post-transcriptional regulation of ABA signaling components in plants.

Regulation of Dormancy and Germination
As sessile organisms, plants are particularly responsive to environmental cues with respect to the initiation of germination. Under unfavorable environmental conditions, ABA contributes to the suppression of germination by promoting seed dormancy, as revealed by the reduced seed dormancy of ABA biosynthesis and signaling-impaired mutants (Koornneef et al., 1982;Leon-Kloosterziel et al., 1996;Lefebvre et al., 2006;Nakashima et al., 2009a;Zhao et al., 2018). For example, the Raf-like MAKKKs, Raf10 and Raf11, influence seed dormancy by phosphorylating SnRK2s and ABRE-binding factors (ABFs; Lee et al., 2015;Nguyen et al., 2019). ABA can also influence seed germination by modifying hormonal balance. In contrast to ABA, gibberellic acid (GA) is a phytohormone that promotes seed germination (Debeaujon and Koornneef, 2000;Ogawa et al., 2003), and the activities of these two hormones antagonistically regulate seed dormancy and germination in higher plants (Weiss and Ori, 2007;Golldack et al., 2013).
Notably, compared with wild-type plants, ABA-deficient mutants are characterized by higher levels of GA, and it has accordingly been suggested that ABA plays a role in regulating the GA metabolic pathway . As a key repressor of GA signaling, REPRESSOR OF GA-LIKE 2 (RGL2) plays a negative role in seed germination, as indicated by the lossof-function mutant rgl2, in which ABA concentrations are reduced and germination is promoted (Piskurewicz et al., 2008;Lee et al., 2010). RGL2 also promotes ABI5 expression levels, thereby inducing the ABA-mediated suppression of germination (Liu et al., 2016). Exogenous ABA activates the expression of RGL2 and ABI5 (Piskurewicz et al., 2008), and thus RGL2 is believed to modulate the balance between ABA and GA contents during seed germination.

Regulation of Drought-Responsive Gene Expression
On exposure to drought stress, ABA induces the expression of several genes in mature plants that play roles in the drought stress response. Representative ABA-responsive marker genes in this context include RESPONSIVE TO DESICCATION 29B (RD29B), RESPONSIVE TO ABA18 (RAB18), LATE EMBRYOGENESIS ABUNDANT 1 (EM1), and EM6, the expression of which is induced under conditions of water limitation (Yamaguchi-Shinozaki and Shinozaki, 1994;Mantyla et al., 1995;Uno et al., 2000;Carles et al., 2002;Umezawa et al., 2006;Wang et al., 2018b). The expression of these genes is regulated by a number ABA-inducible transcription factors, including ABA-responsive element (ABRE)-binding proteins (AREBs) and ABFs, which contribute to the regulation of ABA-dependent gene expression under drought stress conditions (Nakashima et al., 2009b;Yoshida et al., 2015). It has also been established that a number drought stress-induced genes are also responsive to exogenous ABA, which is mediated via promoter region ABREs (Nakashima et al., 2009b;Fujita et al., 2011). For example, plants overexpressing (OX) the AREB1/ ABF2, AREB2/ABF4, ABF1, and ABF3 genes are characterized by enhanced ABA sensitivity and drought tolerance (Kang et al., 2002;Fujita et al., 2005). In contrast, areb1, areb2, and abf3 mutants show reduced drought tolerance and ABA sensitivity (Yoshida et al., 2010). The transcripts of other transcription factors, such as AtMYC2 and AtMYB2, have been found to accumulate in response to exogenous ABA, and in AtMYC2-OX and AtMYB2-OX plants, the ABA-inducible gene RESPONSIVE TO DESICCATION 22 (RD22) is upregulated, leading to ABA-sensitive and drought-tolerant phenotypes (Abe et al., 2003). Similarly, expression of the NAC transcription factor RD26 is induced by drought and ABA treatment, with RD26-OX plants being characterized by hypersensitivity to exogenous ABA, whereas contrastingly, RD26 dominant repressor plants exhibit an ABA-insensitive phenotype (Fujita et al., 2004). Collectively, these findings thus indicate that ABA induces ABA-responsive transcription factors, which in turn regulate the expression of ABA-responsive genes, thereby enhancing drought stress tolerance.

Regulation of Leaf Senescence
Exogenous ABA treatment has been widely demonstrated to induce leaf senescence (Gepstein and Thimann, 1980;Pourtau et al., 2004;Raab et al., 2009;Lee et al., 2011). Moreover, it has been observed that as leaves senesce, there is a corresponding increase in endogenous ABA levels (Gepstein and Thimann, 1980;Leon-Kloosterziel et al., 1996;Cheng et al., 2002;Breeze et al., 2011;Yang et al., 2014). In terms of gene expression, ABA signaling-associated genes have been found to be upregulated during leaf senescence and the expression of various senescence-associated genes can also be induced by the application of exogenous ABA (Tan et al., 2003;Parkash et al., 2014). An important factor associated with the induction of leaf senescence is chlorophyll degradation, which is mediated via genes such as STAY-GREEN 1 (SGR1), NON-YELLOW COLORING 1 (NYC1), PHEOPHYYTINASE (PPH), and PHEIDE a OXYGENASE (PaO) (Pruzinska et al., 2005;Ren et al., 2007;Hortensteiner, 2009;Sato et al., 2009b;Schelbert et al., 2009). In this regard, Arabidopsis thaliana NAC-LIKE, ACTIVATED BY AP3/PI (AtNAP) has been found to bind to the promoter of ABSCISIC ALDEHYDE OXIDASE 3 (AAO3) to enhance AAO3 transcription and ABA production, and this accumulation of ABA contributes to the induction of chlorophyll degradationassociated genes . ABA core components have also been established to be involved in chlorophyll degradation. For example, SnRK2s regulate the bZIP transcription factors ABI5 and ABF2/3/4, which directly activate NYC1, PAO, and SENESCENCE-ASSOCIATED GENE 12 (SAG12), which are considered senescence marker genes Zhao et al., 2016b).

Ubiquitination and Deubiquitination
The core components of ABA signaling are modulated by different types of post-translational modification, among which, ubiquitination influences the responses to environmental factors. Ubiquitination is mediated by the activity of a series of enzymes, including ubiquitin-activating enzyme E1, ubiquitin-conjugating enzyme E2, and ubiquitin-protein ligase E3, and the substrate proteins thus ubiquitinated are subsequently degraded by 26S proteasome proteolysis (Stone, 2019). ABA components are regulated by E3 ligase complexes via successional steps (Figure 1). These complexes are classified into four groups, namely, really interesting new gene (RING), cullin-RING ligase (CRLs), U-box, and homologous to E6-AP carboxyl terminus (HECT; Vierstra, 2009), among which, RING, CRLs, and U-box E3 ligases are involved in ABA signaling.
Ring finger of seed longevity 1 (RSL1), a member of the RBR-type E3 ligase family (Marin, 2010;Callis, 2014) interacts with PYR1 and PYL4 in the plasma membrane and subsequently polyubiquitinates these to promote their degradation via the 26S proteasome. Phenotypic analysis has revealed that RNA interference mutants of RSL1 are characterized by an ABA-hypersensitive phenotype, whereas an RSL1 overexpression line was found to show reduced ABA sensitivity during seed germination and early seedling growth (Bueso et al., 2014). These observations indicate that RSL1 plays a negative role in ABA signaling via the degradation of ABA receptors (Bueso et al., 2014). In the same gene family as RSL1, the RING finger ABA-related 1-9 (RFA1 to RFA9) proteins are similarly associated with ABA receptor degradation, among which, RFA1 and RFA4 are localized in the nucleus and promote the nuclear degradation of PYR1 and PYL4 via activity of the nuclear E2 enzyme UBC26 (Fernandez et al., 2020). Compared with wildtype plants, PYR1 and PYL4 were observed to show higher expression in an RFA1/RFA4 double knock-out mutant (Fernandez et al., 2020). The endosomal trafficking pathway functions as an important route for the modification of membrane proteins via the endosomal sorting complex required for transport (ESCRT) machinery. Ubiquitinated PYL4 is localized in the plasma membrane and degraded via a vacuolar degradation pathway mediated by the ESCRT components, in which ESCRT-I components FIVE1, VPS23A, and ALIX promote PYL4 degradation. Knock-down mutations of these ESCRT components are characterized a reduction in the degradation of PYL4 and enhanced ABA responses Yu et al., 2016;Garcia-Leon et al., 2019).

Ubiquitination of Clade A PP2Cs
Plant U-box type E3 ubiquitin ligase (PUB)12/13, Ring domain ligase (RGLG)1/5, CUL3-RING-based E3 ligases, and other E3 ligases collectively play roles in the proteolytic degradation of PP2Cs. PUB12/13 interact with ABI1 and thereby promote the ubiquitination of ABI1 in response to exogenous ABA, which induces the production of the dimeric form of PYR1 or monomeric forms of PYL4/PYL9 (Kong et al., 2015). A pub12 pub13 double mutant was found to be characterized by a high accumulation of ABI1 and an ABA-insensitive phenotype, the latter of which was recovered following the crossing of double mutant plants with those containing an abi1-3 mutation. These findings thus indicate that PUB12 and PUB13 play roles in controlling ABI1 stability (Kong et al., 2015). RGLG1/5 E3 ligases contribute to the degradation of PP2CA. In vivo analysis has revealed that RGLG1 interacts with PP2CA, which is enhanced by ABA, and promotes ubiquitination. In vitro, however, it has been observed that RGLG1 ubiquitinates PP2CA in the absence of either ABA or PYL4 (Wu et al., 2016). In contrast to PP2CA, which is nuclear localized, myristoylated RGLG1 is localized in the plasma membrane. ABA prevents the myristoylation of RGLG1 and induces cycloheximide-insensitive translocation to the nucleus (Belda-Palazon et al., 2019). Thus, RGLG1 plays a positive role in the ABA response by repressing PP2CA (Wu et al., 2016). PP2CA has also been demonstrated to interact with BTB/POZ and MATH domain proteins (BPMs) in plants, based on co-immunoprecipitation analysis performed in conjunction with liquid chromatography-tandem mass spectrometry. Compared with wild-type plants, those overexpressing BPM3 and BPM5 were found to accumulate lower amounts of PP2CA protein following ABA treatment. Contrastingly, in response FIGURE 1 | Degradation of ABA core components by the 26S ubiquitin-proteasome system and endosomal-vacuole pathway. PYR1/PYLs are degraded by RSL1, RFA1/RFA4, and CRL4 DDA1 through ubiquitin-proteasome system. RSL1 ubiquitinates PYR1/PYL4 on the plasma membrane (PM), after that ubiquitinated substrates are degraded by ESCRT components, FIVE1/VPS23A/ALIX via endosomal-vacuole pathway. RFA1-and RFA4/CRL4 DDA1 -mediated ubiquitination processes occur in the cytosol and nucleus, respectively. PP2Cs are degraded by PUB12/13, CRL3 BPM3/5(BTB-MATH) , PIR1/2, and RGLG1/5 through 26S proteasome. PUB12/13 degrade ABI1 by ubiquitination in the PM. RGLG1/5 ubiquitinate PP2CA/ABI2/HAB2, CRL3 BPM3/5(BTB-MATH) ubiquitinate PP2CA/ABI1/HAB1, and PIR1/2 ubiquitinates PP2CA in the nucleus. This figure has been modified from Coego et al. (2021).
to ABA treatment, the levels of PP2CA accumulated in plants harboring a bpm3 bpm5 double knock mutation were observed to be higher than those in the wild-type plants. Consistently, BPM3 and BPM5 overexpressing plants are characterized by enhanced ABA sensitivity, whereas the bpm3 bpm5 mutant shows reduced ABA sensitivity. Consequently, CRL3 BPM complexes appear to play vital roles in ABA responses and signaling . By interacting with UBC27, AIRP3 contributes to ABI1 degradation via E3 ligase activity. The UBC27-AIRP3 ubiquitination complex modulates ABA responses by enhancing ABI1 protein stability. In ubc27 and airp3 mutant plants, the rate of ABI1 degradation is reduced, and with respect to cotyledon greening and stomatal movement, ubc27 and airp3 mutant plants are characterized by ABA-insensitive phenotypes (Pan et al., 2020). Among other RING proteins, PP2CA interacting ring finger protein (PIR)1 and PIR2 have been shown to interact with PP2CA, both in vitro and in vivo, thereby promoting its ubiquitination. pir1 pir2 double knock-out mutants were found to be characterized by higher PP2CA protein levels, thus indicating that PIR1 and PIR2 degrade PP2CA via the 26S proteasome (Baek et al., 2019b).

Regulation of SnRK2s Stability by E3 Ligases
Although E3 ligases have been established to contribute indirectly to the stability of SnRK2s, there has as yet been no conclusive evidence to indicate that E3 ligases directly target SnRK2s. To date, only AtPP2-B11, a component of the SKP1/Cullin/F-box complex, has been shown to interact directly with SnRK2.3, with cell-free degradation assays performed using AtPP2-B11-OE and amiR-AtPP2-B11 cell extracts indicating that AtPP2-B11 negatively influences SnRK2.3 stability. However, there is no evidence to indicate that AtPP2-B11 directly ubiquitinates SnRK2.3 (Cheng et al., 2017). In addition, the E3-ubiquitin ligase HOS15 has been found to interact with SnRK2.6, thereby influencing its stability, and compared with wild-type plants, those harboring a hos15-2 mutation were found to accumulate higher levels of SnRK2.6 (Ali et al., 2019). However, in common with AtPP2-B11, the mechanisms whereby HOS15 regulates SnRK2 stability have yet to be established.

Modulation of ABA Signaling by Ubiquitin-Specific Processing Proteases
With respect to ABA signaling, plant tolerance to high salinity and drought stresses is modulated by UBP24 via the regulation of stomatal closure. Knock-out of the UBP24 gene confers an ABA-hypersensitive phenotype during seedling growth, although contrastingly it reduces ABA sensitivity in the guard cells of mature plants. The latter effect results in a drought-sensitive phenotype, owing to a higher water loss compared with wildtype plants (Zhao et al., 2016a). In pepper plants, CaUBP12 has been observed to suppress the degradation of CaSnRK2.6 (Arabidopsis OST1 homolog; Lim et al., 2021b). CaUBP12 interacts with CaSnRK2.6 both in vitro and in vivo, and cellfree degradation analysis has revealed higher levels of CaSnRK2.6 degradation when incubated with crude extracts of the CaUBP12 knock-down pepper plants, although lower levels in the presence of crude extracts of CaUBP12 overexpressing Arabidopsis and tobacco plants (Lim et al., 2021a,b). These findings thus provide evidence to indicate that deubiquitinase-mediated posttranscriptional modification is implicated in the modulation of ABA signaling.

Phosphorylation of ABA Receptors
Several studies have provided evidence indicating the phosphorylation of ABA receptors. For example, TOR kinase, Arabidopsis Early flowering 1 (EL1)-like casein kinase (AEL), C-terminally encoded peptide receptor 2 (CEPR2), and cytosolic ABA receptor kinase 1 (CARK1) have been established to phosphorylate PYLs (Chen et al., 2018;Wang et al., 2018a;Zhang et al., 2018;Li et al., 2019;Yu et al., 2019). Among these, TOR kinase inactivates PYL1 and PYL4 by phosphorylating the Ser119 and Ser114 residues, respectively. Phosphomimic mutants of PYL1 and PYL4 are characterized by an inability to inhibit ABI1 phosphatase activity. Moreover, ABA has been shown to inhibit TOR kinase activity, thereby enhancing the ABA response. Thus, these observations indicate that TOR kinase functions as a negative regulator of ABA signaling by phosphorylating PYL1 and PYL4 (Wang et al., 2018a). AEL phosphorylates PYR1 and PYL1 at residues Ser109/152 and Ser136/182, respectively, and in doing so, promotes PYR1 and PYL1 degradation. Compared with wild-type plants, PYR1 and PYL1 proteins show greater stability in ael triple mutants. The ABA-and NaCl-sensitive phenotypes in knock-out mutants thus tend to indicate that AEL negatively regulates ABA signaling (Chen et al., 2018). CEPR2 has been found to interact with PYL4 in the plasma membrane, in which it phosphorylates Ser54, thus promoting an accelerated degradation of the PYL4 protein. Consistently, the levels of PYL4 protein in CEPR2 overexpressing and cepr2 mutant plants have been shown to be less and more stable, respectively (Yu et al., 2019). CARK1 phosphorylates PYR1 Thr78, PYL1 Thr105, PYL2 Ser81, and PYL3 Thr101, thereby enhancing ABA responses implicated in the inhibition of seed germination, root growth, and the expression of a number of ABA-responsive genes Li et al., 2019).  (Cai et al., 2014).
FIGURE 2 | Kinase-mediated post-translational modification of core components of the ABA signaling pathway. CARK1, CEPR2, TOR kinase, and AEL phosphorylate PYLs. CARK1 positively regulates ABA signaling by enhancing PYLs activity on the cytoplasm. In contrast, CEPR2 and AEL negatively regulate ABA signaling by promoting degradation of PYLs. TOR kinase also plays negative role in ABA signaling by preventing PYL activation. CEPR2 acts on PM. ABI1 and ABI2 are phosphorylated and enhanced in their phosphatase activity by PR5K2 on the PM. BIN2 and RAF10 phosphorylate SnRK2s enhancing their kinase activity on the cytoplasm and nucleus.

CONCLUSIONS AND PERSPECTIVES
Abscisic acid plays vital roles in plant growth, development, and biotic/abiotic stress responses, particularly with respect to seed dormancy/germination, stomatal opening and closure, and leaf senescence. Moreover, the ABA core signaling pathway and post-translational modification of pathway components have been extensively studied. Ubiquitination, catalyzed by E3 ligases, contributes to the 26S proteasome-mediated degradation of ABA receptor PYR/PYL and clade A PP2C proteins, and numerous E3 ligases have been found to regulate the stability of these target proteins, either via direct or indirect interactions. However, it has yet to be conclusively determined whether E3 ligases directly ubiquitinate SnRK2 proteins, which accordingly limits our current understanding of the E3 ligase-mediated modulation of SnRK2 protein stability. The regulation of target protein stability is also modulated by deubiquitination, although in Arabidopsis, the mechanisms underlying the deubiquitination of ABA signaling core components remain to be ascertained. The involvement of UBPs in stress responses and other developmental processes is well characterized; however, to date there have been few studies that have examined the mechanisms associated with protein deubiquitination. To determine whether target proteins are deubiquitinated by UBPs, it will be necessary to isolate either mono-or poly-ubiquitinated proteins. The phosphorylation of target proteins also contributes to their activity and/or stability. In this regard, several kinases that phosphorylate key ABA signaling factors have been shown to regulate drought stress responses, either positively or negatively. However, only a few kinases involved in the ABA core signalingassociated regulation of drought responses have been examined, and it can be assumed that there are numerous other functional kinases that await identification. Consequently, to gain a better understanding of the post-translational modification of ABA core components, it will be necessary to conduct further studies that focus on identifying and characterizing these undiscovered proteins, which will illustrate potentially novel routes for the regulation of ABA signaling and stress responses.

AUTHOR CONTRIBUTIONS
JL, CL, and SL: conceptualization and writing. JL and CL: data analysis. All authors contributed to the article and approved the submitted version.