The role of amino acid metabolism in signaling and metabolic adaptation to stress-induced energy deficiency in plants

Abstract

The adaptation of plant metabolism to stress-induced energy deficiency involves profound changes in amino acid metabolism. Anabolic reactions are suppressed, whereas respiratory pathways that use amino acids as alternative substrates are activated. This review highlights recent progress in unraveling the stress-induced amino acid oxidation pathways, their regulation, and the role of amino acids as signaling molecules. We present an updated map of the degradation pathways for lysine and the branched-chain amino acids. The regulation of amino acid metabolism during energy deprivation, including the coordinated induction of several catabolic pathways, is mediated by the balance between TOR and SnRK signaling. Recent findings indicate that some amino acids might act as nutrient signals in TOR activation and thus promote a shift from catabolic to anabolic pathways. The metabolism of the sulfur-containing amino acid cysteine is highly interconnected with TOR and SnRK signaling. Mechanistic details have recently been elucidated for cysteine signaling during the abscisic acid-dependent drought response. Local cysteine synthesis triggers abscisic acid production and, in addition, cysteine degradation produces the gaseous messenger hydrogen sulfide, which promotes stomatal closure via protein persulfidation. Amino acid signaling in plants is still an emerging topic with potential for fundamental discoveries.

Abiotic stress, alternative respiration, amino acid metabolism, branched-chain amino acid degradation, energy deficiency, lysine degradation, signaling, SnRK, TOR

Introduction

Plants are confronted with a broad spectrum of unfavorable environmental conditions ranging from drought to flooding, high salinity, or extreme temperatures. Often, these abiotic stresses occur not individually but in combination (e.g. dry heat), and they might even coincide with attacks by various pathogens. Being sessile, plants had to develop a sophisticated metabolic defense system to differentiate between the various threats, correctly assess them, and adapt their defense strategy accordingly. Many aspects of this complex regulatory system are still unknown. Several strategies to cope with the different challenges of abiotic stress require major adaptations in amino acid metabolism (Fig. 1). The proteome has to be modified to shift from growth to defense, including increased synthesis of proteins required for stress tolerance and damage control at the expense of a decrease in the photosynthetic apparatus and anabolic enzymes (Less et al. 2011; Gururani et al., 2015). This shift inevitably leads to large fluxes through the free amino acid pool and changes in its composition. At the same time, de novo nitrogen assimilation and amino acid synthesis are often restricted by low transpiration rates and limited availability of energy (Araus et al., 2020). Still, a sufficient supply of the 20 proteinogenic amino acids is required for the synthesis of stress-relevant proteins.

Fig. 1.

Major functions of amino acid metabolism during the abiotic stress response in plants. During abiotic stress, amino acids are required as precursors for stress-induced proteins (1) and secondary metabolites (2), and as osmoprotectants (3), substrates for mitochondrial ATP production (4), and signaling molecules (5). This review mainly focuses on aspects 4 and 5.

In addition to proteins, the spectrum of defense compounds derived from amino acids includes a highly diverse set of secondary metabolites with often complex structures. Several classes of secondary metabolites have been reported to accumulate under diverse stressful environmental conditions but their exact functions are largely not clear (Box 1). Osmotic stress during severe drought or high salinity leads to a loss in cell turgor, which can be counteracted by increasing the amount of non-toxic small molecules. Thus, a low water potential strongly induces the synthesis of compatible solutes to maintain cell turgor and stabilize membrane and protein integrity (Singh et al., 2015). Proline acts as a major osmoprotectant in plants (Szabados and Savouré, 2010). During severe drought it constitutes 20% of the total amino acid pool and thus ties up valuable resources but also acts as a nitrogen store (Heinemann et al., 2021). Proline accumulation in the cytosol might also be relevant for balancing the increased vacuolar osmolarity due to autophagic degradation of macromolecules during stress (Signorelli et al., 2019).

Box 1.

Amino acids as precursors for secondary metabolites during abiotic stress

The role of secondary metabolites derived from amino acid metabolism in abiotic stress tolerance is a large field with many open questions. Most importantly, specific functions of these highly complex and diverse molecules, beyond ROS scavenging, have to be addressed.

Flavonoids derived from phenylalanine or tyrosine strongly accumulate under various abiotic stress conditions such as UV, temperature, salt, and drought (Nabavi et al., 2020; Falcone Ferreyra et al., 2012). Flavonoids are ubiquitous in the plant kingdom, and estimations suggest that ~20% of the total carbon flux accounts for this pathway (Haslam, 1993). A total of 54 different flavonoids have been identified in Arabidopsis, among them 11 anthocyanins, which cause the characteristic purple color that indicates suboptimal conditions (Saito et al., 2013). Their postulated functions in stress defense are ROS scavenging and the storage of resources. However, since the process of synthesis of aromatic amino acids and their subsequent conversion to secondary metabolites is quite complex and expensive in terms of energy requirement, there might be additional, more specific, benefits during stress that are yet to be discovered.

Polyamines are synthesized from arginine and strongly increase during abiotic stress (reviewed by Alcázar et al., 2010). The overexpression of enzymes involved in polyamine synthesis pathway leads to higher stress tolerance indicating a protective role. However, their function is largely unknown and possibly involves signaling.

Glucosinolates are synthesized from a range of amino acids, including methionine, tryptophan, and phenylalanine specifically in Brassicaceae (reviewed by Halkier and Gershenzon, 2006). They are well known for their contribution to herbivore tolerance, since breakdown upon tissue damage leads to the production of toxic and highly reactive compounds. However, auxin signaling maintains the expression of enzymes involved in glucosinolate biosynthesis during drought, indicating an additional function of these secondary metabolites in abiotic stress resistance. Indeed, increased levels of aliphatic glucosinolates improve drought tolerance in Arabidopsis, and a breakdown product, possibly isothiocyanate, promotes stomatal closure independently of ABA signaling (Salehin et al., 2019).

A major problem in coping with diverse stressful conditions is energy deprivation. Resources are scarce due to restricted photosynthetic activity and have to be diverted from growth into defense and stress tolerance, leading to a general decrease in cellular energy levels (Biswal et al., 2011; Gururani et al., 2015). Stress may also require prioritizing the growth of specific organs over others and thus cause a local energy deficit. For example, increasing root growth at the expense of flowers or leaves may improve the water supply sufficiently for the plant to survive limited periods of severe drought. Metabolic adaptation to energy deprivation utilizes amino acids as signaling molecules and alternative substrates, and these aspects will be the focus of this short review.

Plants cope with energy shortage by activating respiratory pathways that use amino acids as alternative substrates

When carbohydrate stores are depleted, plants induce catabolic pathways (autophagy, lipid and protein degradation) to provide alternative substrates for ATP production. The induction of the respective pathways during different stress conditions associated with energy deprivation, such as drought or low light conditions, has repeatedly been demonstrated on a transcript and a protein level (Less and Galili, 2008; Angelovici et al., 2013). Mutants with defects in the autophagy apparatus show poor growth and early senescence when exposed to carbon and/or nitrogen starvation, drought, or high salinity, demonstrating the physiological relevance of this bulk degradation pathway for energy homeostasis during stress (Liu and Bassham, 2012; Izumi et al., 2013; Hirota et al., 2018; Tang and Bassham, 2018). Severe dehydration can lead to a substantial loss in protein mass, indicating that the demand for the resources tied up in proteins is high (Heinemann et al., 2021). Amino acid degradation produces tricarboxylic acid cycle intermediates or precursors and thus contributes to the production of substrates for mitochondrial respiration. Notably, the oxidation of amino acids with a complex structure (branched-chain and aromatic amino acids) provides comparable amounts of energy for ATP synthesis to those provided by glucose (Hildebrandt et al., 2015). The catabolic pathways of branched-chain amino acids (BCAAs), lysine, and proline are even physically connected to the mitochondrial respiratory chain, since individual reaction steps transfer electrons into the ubiquinone pool. However, not all the enzymes required for amino acid degradation in plants are known yet, and the mechanisms of their regulation remain partially elusive, hampering progress in fully understanding—let alone exploiting—their role in stress tolerance.

Recently, there has been some clear progress with respect to lysine and BCAA catabolism (Fig. 2; see also Box 2). These amino acids can feed electrons into the mitochondrial respiratory chain via the electron-transfer flavoprotein (ETF)-ubiquinone oxidoreductase complex (ETFQO) (Ishizaki et al., 2005). The initial four reaction steps in lysine catabolism leading to 2-oxoadipate are conserved between animals and plants (Fig. 2, left). An unusual oxidative decarboxylation and hydroxylation reaction is subsequently performed by a recently identified hydroxyglutarate synthase, producing d-2-hydroxyglutarate, which can then be oxidized to 2-oxoglutarate by d-2-hydroxyglutarate dehydrogenase, transferring electrons to the ETF/ETFQO system (Thompson et al., 2020). Lysine can also be converted to the immune signal N-hydroxypipecolic acid via three recently identified reaction steps (Ding et al., 2016; Chen et al., 2018; Hartmann and Zeier, 2018).

Fig. 2.

An update on the catabolic pathways for lysine and branched-chain amino acids in plants. Recently identified reaction steps are shown in blue (see also Box 2). Published information about the previously known steps is summarized in Hildebrandt et al. (2015). Previously identified enzymes: 1, lysine-ketoglutarate reductase/saccharopine dehydrogenase (AT4G33150); 2, aldehyde dehydrogenase 7B4 (AT1G54100); 3, GABA transaminase (AT3G22200); 5, d-2-hydroxyglutarate dehydrogenase (AT4G36400); 6, electron transfer flavoprotein (AT1G50940, AT5G43430); 7, electron-transfer flavoprotein:ubiquinone oxidoreductase (AT2G43400); 11, BCAA transaminase (AT1G10060, AT1G10070, AT3G49680, AT3G19710, AT5G65780, AT1G50110, AT1G50090); 12, branched-chain alpha-keto acid dehydrogenase (AT5G09300, AT1G21400, AT3G13450, AT1G55510, AT3G06850); 13, isovaleryl-CoA-dehydrogenase (AT3G45300); 14, methylcrotonyl-CoA carboxylase (AT1G03090, AT4G34030); 16, hydroxmethylglutaryl-CoA lyase (AT2G26800); ?, Unknown reaction steps. Metabolites: 2-AD, 2-aminoadipate; 2-HG, d-2-hydroxyglutarate; 2-MAA, 2-methylacetoacetyl-CoA; 2-MBC, 2-methylbutanoyl-CoA; 2-MHB, 2-methyl-3-hydroxybutyryl-CoA; 2-MPC, 2-methylpropanoyl-CoA; 2-OA, 2-oxoadipate; 2-OG, 2-oxoglutarate; 3-HIB, 3-hydroxyisobutyrate; 3-HIBC, 3-hydroxyisobutyryl-CoA; 3-HMG, 3-hydroxymethylglutaryl-CoA; 3-MCC, 3-methylcrotonyl-CoA; 3-MGC, 3-methylglutaconyl-CoA; 3-MOB, 3-methyl-2-oxobutanoate; 3-MOP, 3-methyl-2-oxopentanoate; 4-MOP, 4-methyl-2-oxopentanoate; AC, acrylyl-CoA; ADS, 2-aminoadipate-6-semialdehyde; ATA, acetoacetate; ATC, acetoacetyl-CoA; HP, hydroxypropionate; HPC, hydroxypropionyl-CoA; IC, isovaleryl-CoA; MAC, methylacrylyl-CoA; MMS, methylmalonate semialdehyde; MSA, malonate semialdehyde; PDC, l-Δ1-piperideine-2-carboxylate; Pip, l-pipecolate; SP, saccharopine; TC, tiglyl-CoA. mETC, mitochondrial electron transport chain.

Box 2.

Key developments in understanding the role of amino acid metabolism in signaling and metabolic adaptation to stress-induced energy deficiency

• Completing knowledge on the lysine catabolic pathway in plants

Thompson et al. (2020) identified the last missing step in the lysine degradation pathway in plants, which is catalyzed by a hydroxyglutarate synthase (AT1G07040; Fig. 2). This enzyme, an iron (II)-dependent oxygenase, uses an unusual catalytic mechanism to perform the oxidative decarboxylation and hydroxylation of 2-oxoadipate to d-2-hydroxyglutarate. It strongly responds to both abiotic and biotic stress on a transcriptional level.

• New insights into branched-chain amino acid catabolism in plants

Three recent studies improved our understanding of branched-chain amino acid metabolism substantially (Fig. 2). Using comparative genomics, Latimer et al. (2018) identified a 3-methylglutaconyl-CoA hydratase (AT4G16800) catalyzing the dehydration of 3-hydroxymethylglutaryl-CoA to 3-methylglutaconyl-CoA during leucine degradation. Schertl et al. (2017) characterized a 3-hydroxybutyrate dehydrogenase (AT4G20930), which can also metabolize 3-hydroxypropionate and is thus involved in both valine and isoleucine degradation. Two additional enzymes required for the degradation of valine, a 3-hydroxyisobutyryl-CoA hydrolase (AT4G31810) and a methylmalonate-semialdehyde dehydrogenase (AT2G14170), were identified and functionally characterized by Gipson et al. (2017).

• The molecular mechanism of activating alternative mitochondrial respiration pathways during energy deprivation

SnRK1 kinases act as central metabolic regulators during adaptation to low-energy stress. Pedrotti et al. (2018) demonstrated that in response to extended darkness SnRK1 phosphorylates group C bZIP transcription factors, leading to the formation of a complex containing SnRK1, C bZIP, and an additional S1 bZIP transcription factor. This complex coordinates the strong induction of genes involved in the oxidation of branched-chain amino acids, which serve as alternative respiratory substrates during carbohydrate starvation (Figs 3, 4).

• Amino acids activate TOR signaling in plants

Liu et al. (2021) demonstrated that 15 of the 20 proteinogenic amino acids activate TOR signaling in the leaf primordia of inorganic-nitrogen-starved seedlings, albeit with different capacities. Both inorganic nitrogen and amino acids require the plant-specific small GTPase ROP2 for TOR induction. O’Leary et al. (2020) performed long-term measurements of oxygen consumption rates in Arabidopsis leaves and discovered a regulatory mechanism based on the balance between individual amino acid levels. Proline and alanine, which accumulate to high concentrations during abiotic stress conditions, stimulate respiration via transcriptional up-regulation of their respective degradation pathways. Other amino acids, such as isoleucine and methionine, block this induction by activating TOR signaling.

• Cysteine triggers ABA production and stomatal closure during drought stress

Batool et al. (2018) described a new role of cysteine signaling in the physiological water limitation response in Arabidopsis. During soil drying, sulfate is transported to the guard cells via the xylem and incorporated into cysteine. Increased cysteine concentrations stimulate ABA biosynthesis in the leaves by activating two enzymes in the synthesis pathway (NCED3 and AAO3). Zhou et al. (2021), Chen et al. (2020), and Shen et al. (2020) report additional downstream regulatory steps in cysteine-induced stomatal closure mediated by post-translational activation of the protein kinase SnRK2.6, the NADPH oxidase RbohD, and the transcription factor ABI4 (ABA-INSENSITIVE 4) via persulfidation (Fig. 3).

The catabolism of the three BCAAs to acetyl-CoA combines shared reaction steps that use intermediates from leucine, valine, and isoleucine degradation as a substrate with individual, amino-acid-specific steps (Fig. 2, right). Although several enzymes involved in BCAA degradation have been identified during the past 3 years, the pathway is still not complete (see Box 2, Fig. 2). The identification of the last missing steps is hampered by the large number of candidates derived from homology searches, which additionally have multiple isoforms that might be redundant.

The degradation of BCAAs and lysine is thought to be particularly relevant during stress-induced energy deficiency, since knockout lines for several reaction steps can be distinguished from the wild type by their shorter survival time in complete darkness (Ishizaki et al., 2005, 2006; Araújo et al., 2010; Peng et al., 2015; Hirota et al., 2018). They may also show decreased drought tolerance under specific growth conditions (Pires et al., 2016), which, however, seems to be a less robust phenotype, since it was not observable under our experimental drought setup (Heinemann et al., 2021). The defects in seed development and/or germination reported for several lines may be caused by the accumulation of toxic intermediates such as methacrylyl-CoA during embryogenesis or by starvation prior to the full establishment of photosynthesis (Ding et al., 2012; Angelovici et al., 2013; Peng et al., 2015; Gipson et al., 2017). In addition, the production of storage lipids during seed filling seems to be compromised by defects in BCAA catabolism, but the reason for this effect is not clear yet (Gipson et al., 2017).

The adaptation of amino acid metabolism to energy deprivation is mediated by the balance between TOR and SnRK signaling

Plants need to monitor their nutritional status and integrate this information with environmental stress signals in order to react appropriately and find the best balance between growth and defense to guarantee survival and reproduction. Adapting amino acid catabolism to energy requirements is one of these tasks. The protein kinase complexes SnRK1 (Snf1-related protein kinase 1) and TOR (target of rapamycin) are to a large extent evolutionarily conserved in eukaryotes and act as central metabolic regulators (for recent reviews, see Broeckx et al. 2016; Baena-González and Hanson, 2017; Shi et al., 2018; Wurzinger et al., 2018; Caldana et al., 2019; Jamsheer et al., 2019; Margalha et al., 2019; Rodriguez et al., 2019; Ryabova et al., 2019; Wu et al., 2019; Fu et al., 2020; Pacheco et al., 2021; Sharma et al., 2021). In general, TOR promotes growth when the supply of nutrients is sufficient, whereas SnRK1 acts antagonistically and restores energy homeostasis during stress (Fig. 3, Box 3).

Fig. 3.

The role of amino acids in metabolic regulation by TOR and SnRK signaling during abiotic stress. Metabolic adaptation to the environmental conditions is achieved by the antagonistic protein kinases TOR and SnRK. TOR signaling induces amino acid synthesis pathways and represses autophagy as well as amino acid catabolism in response to optimal growth conditions. Individual amino acids act as TOR activators (Cao et al., 2019; O’Leary et al., 2020; Liu et al., 2021). Under unfavorable conditions, SnRKs are activated and repress TOR signaling. During energy deficiency, SnRK1 induces amino acid degradation pathways via bZIP transcription factors (Pedrotti et al., 2018). Increased cysteine synthesis in the guard cells contributes to the activation of ABA signaling during drought (Batool et al., 2018; Rajab et al., 2019). Cysteine degradation produces the gaseous signaling molecule hydrogen sulfide (H₂S), which persulfidates and thus activates SnRK2.6, ABI4, and the NADPH oxidase RbohD, ultimately leading to stomatal closure and other physiological stress responses (Chen et al., 2020; Shen et al., 2020; Zhou et al., 2021). In addition, inhibition of PP2C by ABA results in SnRK1 activation and thus promotes SnRK1 signaling during stress (Rodrigues et al., 2013). Dashed arrows indicate indirect or postulated interactions. Intersections with amino acid metabolism are highlighted in red. ABI4, abscisic acid insensitive 4; bZIP, basic leucine zipper transcription factor; Des1, l-cysteine desulfhydrase 1; PYL, pyrabactin resistance 1-like (ABA receptor); PP2C, protein phosphatase type 2C; RbohD, respiratory burst oxidase homolog protein D; ROP2, rho-related protein from plants 2; ROS, reactive oxygen species; S6K1, S6 kinase 1; SnRK1, Snf1-related protein kinase 1; SnRK2, Snf1-related protein kinase 2; -SSH, persulfidation of cysteine residue; TOR, target of rapamycin.

Box 3.

TOR and SnRK1 signaling in plants

TOR (target of rapamycin)

Two structurally and functionally distinct TOR protein complexes with several regulatory partners have been identified in eukaryotes. Plants contain only one of them, the TORC1 complex consisting of the TOR kinase and the regulatory subunits RAPTOR (regulatory-associated protein of TOR) and LST8 (lethal with SEC13 protein 8) (reviewed by Fu et al., 2020; Caldana et al., 2019; Wu et al., 2019).

A key function of TORC1 in plants is integrating different signals indicating growth-promoting environmental conditions such as nutrient and light availability to induce anabolic processes. TOR signaling is involved in the regulation of several plant-specific processes, such as the glyoxylate cycle, cell wall synthesis, and nitrogen and sulfur assimilation (Wu et al., 2019). During activation of the stem cells at the shoot apical meristem, plant TOR kinase integrates light availability with the universal TOR-activating signal, glucose, via auxin signaling (Li et al., 2017; Pfeiffer et al. 2016). In contrast, sugar energy signaling is sufficient for TOR activation in the root apex (Li et al., 2017).

SnRK1 (Snf1-related protein kinase 1)

The SNF1/AMPK/SnRK1 family forms heterotrimeric holoenzymes containing a catalytic α-subunit and non-catalytic β- and γ-subunits. Arabidopsis expresses two isoforms of the SnRK1 catalytic α-subunit (AKIN10 and AKIN11), three β-subunits, and one γ-subunit (reviewed by Wurzinger et al., 2018; Emanuelle et al., 2016).

SnRK1 kinase activity is induced by starvation conditions and activates an energy-saving program including autophagy and catabolic pathways. A distinct pattern of regulation of plant SnRK1 compared with its mammalian and yeast counterparts, AMPK and SNF1, allows adaptation to plant-specific metabolic requirements. SnRK1 activity is repressed by high-energy sugar-phosphates such as trehalose-6-phosphate, whereas SNF1 and AMPK sense the adenylate charge (Baena-González and Lunn, 2020; Emanuelle et al., 2016). The plant SnRK1 has a unique autophosphorylation activity and thus is constitutively active unless repressed by high energy status. Low-energy stress triggers nuclear translocation of the catalytic subunit, leading to induced target gene expression (Ramon et al., 2019). There is also evidence indicating direct redox-regulation of conserved cysteine residues in the kinase subunit in response to ROS signaling (Broeckx, 2018; Wurzinger et al., 2017).

Reciprocal regulation by TOR and SnRK1 signaling adapts plant protein and amino acid metabolism to the availability of energy and nutrients (Fig. 3). Growth-promoting conditions lead to increased TOR activity, which induces amino acid synthesis and the reinitiation of translation of specific mRNAs, but represses autophagy and amino acid degradation (van Leene et al., 2019; Schepetilnikov et al., 2013; Xiong et al., 2013). In contrast, energy deprivation activates SnRK1 signaling, which represses TOR and induces autophagy (Nukarinen et al., 2016; Huang et al., 2019). Recent evidence suggests that the coordinated induction of key enzymes in amino acid catabolic pathways in response to carbohydrate starvation is also mediated by SnRK1 signaling as part of a general energy-saving program (Pedrotti et al., 2018; Dietrich et al., 2011). Transcriptional induction of a core set of eight enzymes involved in BCAA, proline, and tyrosine degradation requires the formation of a ternary complex between SnRK1 and a heterodimer of two basic leucine zipper (bZIP) transcription factors. A group C bZIP is phosphorylated by SnRK1 and acts as a bridge to an additional bZIP of the S1 group, which in turn directly controls the transcription of the genes encoding amino acid catabolic enzymes via binding to the G-box promoter elements (Pedrotti et al., 2018; see also Box 2). The mechanism of induction is not known in such detail for the other amino acid degradation pathways. However, the transcriptional response of plants with induced or inhibited TOR and SnRK1 pathways (Fig. 4) indicates that reciprocal regulation by these kinases may adapt the general direction of amino acid metabolism (de novo synthesis versus degradation) to the environmental conditions.

Fig. 4.

Regulation of amino acid catabolic pathways in response to the energy status by TOR and SnRK1 signaling in Arabidopsis thaliana. Enzymes involved in amino acid degradation are regulated by the antagonistic action of TOR and SnRK1 signaling. TOR induction (dataset 1) as well as knockdown of SnRK1 (dataset 2) or downstream transcription factors of the bZIP group (dataset 3) generally lead to a transcriptional repression of amino acid catabolic enzymes (indicated by blue colors). The opposite effect, that is, an induction of amino acid catabolism (indicated by red colors), is caused by TOR inhibition (dataset 4) or SnRK1 overexpression (dataset 5), and also becomes apparent in datasets of plants subjected to stress conditions that are associated with energy deficiency [extended darkness (dataset 6) and drought (dataset 7)]. Gene IDs and, in addition, the numbers used in Fig. 2 for lysine and BCAA catabolic enzymes are indicated at the top. The colored squares represent log₂-fold changes in mRNA abundance (treatment versus control) compiled from the following datasets: 1, TOR induction by treatment with glucose (Xiong et al., 2013); 2 and 3, knockdown lines for SnRK1 (snrk1α1/α2) and S1 bZIPs (bZIP1/2/11/44/53) after 6 h of extended darkness (Pedrotti et al., 2018); 4, inhibition of TOR by AZD8055 (Dong et al., 2015); 5, overexpression of the SnRK1 subunit KIN10 (Baena-González et al., 2007); 6, extended darkness (means from six time points, 2–48 h; Usadel et al., 2008); 7, drought (means from three experiments, 5–10 days of dehydration; Perera et al., 2008; Ludwików et al., 2009; Pandey et al., 2013). KD, knockdown; OE, overexpression, SnRK1, Snf1-related protein kinase 1; TOR, target of rapamycin.

Amino acids serve as signaling molecules to coordinate growth and stress responses

Amino acids are well known as signaling molecules in animals. Glutamate, aspartate, γ-aminobutyric acid (GABA), and glycine act as neurotransmitters (Hyman et al., 2005). One of the mammalian TOR complexes (mTORC1) is activated by leucine, arginine, and S-adenosylmethionine (SAM) derived from methionine via complex sensing mechanisms (Kim and Guan, 2019). Amino acids are an adequate signal of good nutritional status in humans, who are not able to synthesize all 20 proteinogenic amino acids but have to acquire some essential amino acids (aromatic amino acids and BCAAs, histidine, lysine, methionine, and threonine) in their diet. In plants, the sulfur and nitrogen status is used as a TOR input signal of good nutrient availability instead (Dong et al., 2017; Liu et al., 2021). Equivalents of the mammalian amino acid sensors and TOR activation pathways have not been identified in plants so far (Shi et al. 2018). However, some recent findings indicate that the plant TOR pathway can nevertheless be induced by specific amino acids via the small GTPase ROP2 (Rho-related protein from plants 2) (Liu et al., 2021). In addition to nitrate and ammonium, a total of 15 proteinogenic amino acids were able to restore TOR activity in the leaf primordia of nitrogen-starved Arabidopsis seedlings, with glutamine, alanine, glycine, and cysteine being the strongest activators (Fig. 3). Proline and alanine accumulate during osmotic stress and hypoxia, respectively. They transcriptionally induce their own degradation pathways to allow rapid removal after stress release. This induction can be blocked by increased contents of some other amino acids (e.g. methionine and isoleucine) in a TOR-dependent manner. O’Leary et al. (2020) confirmed the activation of the Arabidopsis TOR kinase by isoleucine and glutamine using an S6K phosphorylation assay. In addition, over-accumulations of BCAAs were reported to trigger rearrangements of the actin cytoskeleton during vacuole morphogenesis in Arabidopsis (Cao et al., 2019). This effect was blocked by TOR silencing but independent of the adapter protein RAPTOR (regulatory-associated protein of TOR). The mechanism and additional components mediating these signaling events still need to be discovered.

Additional input parameters are required to fully assess and integrate the environmental conditions. The plant, for example, needs to distinguish between osmotic stress and optimal growth conditions, which both lead to high levels of glucose and free amino acids. SnRK2 (Snf1-related protein kinase 2) kinases allow an additional fine-tuning during osmotic stress in plants, in particular, the inhibition of TOR signaling in the presence of sugar. TOR kinase phosphorylates abscisic acid (ABA) receptors and thus represses stress responses under growth-promoting conditions and during stress release. SnRK2s in turn phosphorylate RAPTOR, a regulatory component in the TOR complex, to inhibit TOR activity and prevent growth during environmental stress (Fig. 3) (Wang et al., 2018). The metabolism of the sulfur-containing amino acid cysteine is highly interconnected with TOR and SnRK signaling, and thus might act as an additional adjusting screw (Fig. 3). The cellular cysteine concentration is generally kept at a low micromolar level, and this can probably be achieved by a coordinated regulation of the different synthesis and degradation pathways (Hildebrandt et al., 2015). Isoforms of the hetero-oligomeric cysteine synthase complex consisting of O-acetylserine(thiol)lyase and serine acetyltransferase subunits are present in chloroplasts, mitochondria, and the cytosol (Heeg et al., 2008). In addition, several catabolic routes have been identified in the different subcellular compartments (Hildebrandt et al., 2015; Gotor et al., 2019). Limitation of the sulfur precursor for cysteine synthesis inhibits TOR activity indirectly by down-regulating glucose metabolism (Dong et al., 2017). In contrast, a decreased flux of cysteine into glutathione synthesis leads to TOR activation and stimulates protein translation (Speiser et al., 2018).

Recently, there have been several new insights into the role of cysteine as a signaling molecule during drought (Fig. 3; see also Box 2). Sulfate acts as a mobile signal to report a low soil water content to the leaves. This signal is translated into increased sulfate reduction and cysteine synthesis in guard cell chloroplasts, which in turn induces ABA production (Batool et al., 2018; Chen et al., 2019; Rajab et al., 2019). ABA then triggers the expression of DES1, a cytosolic cysteine desulfurase catalyzing the degradation of cysteine to pyruvate and the gaseous signaling molecule hydrogen sulfide (H₂S) (Chen et al., 2020). H₂S regulates the activity and/or stability of multiple proteins via persulfidation of regulatory cysteine residues (Gotor et al., 2019). In the guard cells, two central players in ABA-induced stomatal closure have been shown to be regulated by this post-translational modification. Persulfidation of the NADPH oxidase RbohD leads to increased reactive oxygen species (ROS) signaling, and persulfidation of two cysteine residues exposed on the surface close to an activation loop promotes the activity of the protein kinase SnRK2.6 and its interaction with ABA response element-binding factor2 (ABF2), a transcription factor downstream of ABA signaling (Chen et al., 2020; Shen et al., 2020). The transcription factor ABI4 (ABSCISIC ACID INSENSITIVE 4) is also activated by persulfidation (Zhou et al., 2021). In addition, ABA-dependent persulfidation of the cysteine protease ATG4 seems to be involved in the regulation of autophagy (Laureano-Marín et al., 2020). However, H₂S is also a potent inhibitor of cytochrome c oxidase (Birke et al. 2012). Thus, increased sulfide concentrations would lead to a down-regulation of mitochondrial respiration, and efficient detoxification mechanisms are required in the mitochondria to maintain or restore ATP production during stress (Birke et al., 2015). The high level of interconnection between cysteine metabolism and ABA signaling is not restricted to the drought response in stomata. Endogenous ABA levels are decreased in germinating seedlings of several sulfur-assimilation mutants, and these lines are also hypersensitive to salt stress. ABA in turn induces the transcription of sulfate transporters and other genes associated with sulfur metabolism (Cao et al. 2014).

Several additional amino acids, such as methionine, lysine, histidine, and BCAAs, are normally present in low concentrations and would thus be suitable as signaling molecules. Imbalances in the homeostasis of specific amino acids (e.g. glutamine, phenylalanine, and cysteine) elicit a strong immune reaction, indicating a potential signaling function in the biotic stress response (Pilot et al., 2004; Liu et al., 2010; Álvarez et al., 2012; Pajerowska-Mukhtar et al., 2012). Ca²⁺ influx mediated by glutamate receptor-like channels is involved in the regulation of several physiological processes in plants, such as pollen tube growth, root meristem proliferation, wound responses, and modulating ABA sensitivity during seed germination (Wudick et al., 2018; Qiu et al., 2020; Grenzi et al. 2021). These Ca²⁺-channels are not specific for l-glutamate but can also be activated by d-serine and the l-enantiomers of several other amino acids (Michard et al., 2011; Alfieri et al., 2020). Additional potential amino-acid-sensing mechanisms in plants are the GCN2 (general control non-derepressible 2) kinase pathway, which is activated by uncharged tRNAs, and the plastid-localized PII protein involved in the regulation of arginine and fatty acid biosynthesis (reviewed by Gent and Forde, 2017). There are also indications that proline might act as a signaling molecule to regulate specific aspects of the stress response and plant development (reviewed by Szabados and Savouré, 2010). Experimentally interfering with proline catabolism, for example, leads to developmental defects in seeds, leaves, and inflorescences (Nanjo et al., 1999; Székely, et al., 2008). Proline might directly induce the expression of a set of genes that are relevant during recovery from drought by interaction with specific transcription factors (Oono et al., 2003; Satoh et al., 2004; Weltmeier et al., 2006). In addition, local accumulation of proline during pathogen infection can act as an apoptotic signal and trigger a hypersensitive response via its degradation intermediate pyrroline-5-carboxylate (Fabro et al., 2004). Research on amino acid signaling in plants is still at an early stage and has much potential for fundamental discoveries.

Conclusion

In the past few years, great progress has been made in understanding the role of amino acid metabolism in signaling and stress-induced energy deficiency. However, several key questions remain unanswered with respect to basic metabolic events as well as their potential to be utilized in agriculture. How do plants correctly interpret the different metabolic signals and balance the levels of individual amino acids to fulfill their diverse roles? Which amino acids act as signals, and how does amino acid signaling work mechanistically? How can plants sense amino acid levels? Do any particular pathways or enzymes act as bottlenecks during stress tolerance that might be addressed to improve the performance of crops under increasingly stressful growth conditions? What is the role of amino acids during growth–defense trade-offs, and can they contribute to uncoupling these processes, if necessary? Why do attempts to increase the seed contents of essential amino acids often lead to severe growth phenotypes, and how can this effect be reduced? What happens to the nitrogen released during amino acid catabolism? Thus, research addressing amino acid metabolism in plants has the potential to advance the fields of stress physiology as well as plant energy biology in the near future.

Acknowledgements

We thank Hans-Peter Braun for critical reading of the manuscript. Research in TMH’s group is supported by the Deutsche Forschungsgemeinschaft (HI 1471).