The plant macronutrient phosphorus is a scarce resource and plant-available phosphate is limiting in most soil types. Generally, a gene regulatory module called the phosphate starvation response (PSR) enables efficient phosphate acquisition by roots and translocation to other organs. Plants growing on moderate to nutrient-rich soils need to co-ordinate availability of different nutrients and repress the highly efficient PSR to adjust phosphate acquisition to the availability of other macro- and micronutrients, and in particular nitrogen. PSR repression is mediated by a small family of single SYG1/Pho81/XPR1 (SPX) domain proteins. The SPX domain binds higher order inositol pyrophosphates that signal cellular phosphorus status and modulate SPX protein interaction with PHOSPHATE STARVATION RESPONSE1 (PHR1), the central transcriptional regulator of PSR. Sequestration by SPX repressors restricts PHR1 access to PSR gene promoters. Here we focus on SPX4 that primarily acts in shoots and sequesters many transcription factors other than PHR1 in the cytosol to control processes beyond the classical PSR, such as nitrate, auxin, and jasmonic acid signalling. Unlike SPX1 and SPX2, SPX4 is subject to proteasomal degradation not only by singular E3 ligases, but also by SCF–CRL complexes. Emerging models for these different layers of control and their consequences for plant acclimation to the environment will be discussed.
Introduction
Nutrient availability varies across soil types, ecosystems, geography, seasons and fluctuates in space and time. Consequently, plants as sessile organisms exhibit acclimation responses to acquire and utilise essential nutrients for growth. The vast majority of agricultural plants are grown in soils with sub-optimal nutrients and require exogenous fertilisation [1,2], where the acquisition of nutrients must be balanced with avoiding the effects of over-accumulation or toxic soil-borne elements [3–5]. Plants in many natural habitats suffer from persistent nutrient limitation, with some species like West Australian Proteaceae able to thrive on severely phosphorus impoverished soils [6] which is often coupled with delayed greening of emerging leaves [7]. These plants lack the ability to down-regulate nutrient uptake and use under sufficient supply, which causes severe phosphate toxicity in mature leaves and can lead to plant death [8]. Their inability to suppress nutrient — and in particular phosphate — uptake would indicate that these ‘extremophiles’ carry elements of a constitutive phosphate starvation response (PSR) with modified functionality of transcriptional repressors such as members of the SYG1/Pho81/XPR1 (SPX) domain protein family. For plants in both natural and agricultural systems, nutrient deficiencies rarely occur in isolation, and research is increasingly focused on the cross-talk and integration of nutrient sensing and signalling pathways [9–11].
The management of resource acquisition and use by plants is dictated by the availability of said resources, as well as physiological, biochemical, and energetic demand. These are not static factors in any organ, or at any stage of development, but are rather in constant feedback with the environment and external stimuli perceived by the plant (Figure 1). Since the energetic capacity of a plant is finite, resource acquisition and use must be hierarchically organised for survival. Throughout plant development, resource requirements are considered against variations in availability and are met by systemic or local transcriptional changes [23,24]. Some key transcription factors (TFs) have been implicated in the transduction of multiple signals and are subject to regulation by signal-specific repressors [9,10,22,25]. Characterisation attempts are often in search of an environmentally derived signal which is paired directly with TF-repressor interaction but have understated the effects of signalling molecules on the efficiency of repressor degradation. Ubiquitination complexes responsible for the degradation of negative regulators have components which are recruited specifically, reversibly, and hierarchically depending on the presence or absence of signalling molecules [11,25–27]. Recent work has shown that the reconstitution of these complexes is often conditional, requiring the integration of multiple signals, and occurs on a continuum of efficiency [25,27]. In combination with the direct effects of signalling molecules on TF-repressor interactions, variations in ubiquitination machinery represent an exponential diversification of plant responses to the environment [28,29]. To understand the complexity of post-translational responses to environmental change, it is necessary to characterise regulatory pathways, understand the signalling molecules, and look at their effects on ubiquitination complex assembly both in isolation, and in combination.
The management of resource acquisition and allocation differs between plant organs and developmental stages.
Plant nutrient status is dependent on resource availability, as well as physiological, biochemical, and energetic demand which changes during ontogeny. The top panel depicts Arabidopsis development from seed germination to seed set (‘fruiting stage’). Hormonal gradients across plant organs are shown as coloured arrows on the right, with auxin (orange) being primarily produced in apical meristems for basipetal transported to the roots, while cytokinins (green) and gibberellic acids (blue) are produced in roots for acropetal transport to the shoot. Requirements for macro (N, nitrogen; P, phosphorus; K, potassium) — and micronutrients (Fe, iron; Zn, zinc; Mn, manganese; Cu, copper; B, boron; Mo, molybdate) change across developmental stages [12,13]: P demand is usually lower during vegetative growth and peaks during flowering. Demand for N and K increases simultaneously with biomass production, with relatively higher K levels required for flowering and fruiting. Microelements Fe, Zn and Mn support seedling establishment and vegetative growth, while B, Cu and Mo become increasingly important during reproductive stages. Phytohormone profiles (GA, gibberellic acid; ABA, abscisic acid; BR, brassinosteroids; JA, jasmonic acid; IAA, indole-3-acetic acid/auxins; CK, cytokinins; ET, ethylene) also change during plant development [14–21]: BR, GA and IAA are important across tissues and plant organs throughout development [14,15,17,20]. ABA specifically inhibits seed germination, while IAA and JA promote germination [19,22]. Active JA pools are then reduced during seedling establishment and vegetative growth before increasing again during reproductive stages [16]. CK is important for seedling establishment with reduced levels found during later reproductive stages [18]. ET promotes later stages of development and in particular fruit ripening [21].
Plant nutrient status is dependent on resource availability, as well as physiological, biochemical, and energetic demand which changes during ontogeny. The top panel depicts Arabidopsis development from seed germination to seed set (‘fruiting stage’). Hormonal gradients across plant organs are shown as coloured arrows on the right, with auxin (orange) being primarily produced in apical meristems for basipetal transported to the roots, while cytokinins (green) and gibberellic acids (blue) are produced in roots for acropetal transport to the shoot. Requirements for macro (N, nitrogen; P, phosphorus; K, potassium) — and micronutrients (Fe, iron; Zn, zinc; Mn, manganese; Cu, copper; B, boron; Mo, molybdate) change across developmental stages [12,13]: P demand is usually lower during vegetative growth and peaks during flowering. Demand for N and K increases simultaneously with biomass production, with relatively higher K levels required for flowering and fruiting. Microelements Fe, Zn and Mn support seedling establishment and vegetative growth, while B, Cu and Mo become increasingly important during reproductive stages. Phytohormone profiles (GA, gibberellic acid; ABA, abscisic acid; BR, brassinosteroids; JA, jasmonic acid; IAA, indole-3-acetic acid/auxins; CK, cytokinins; ET, ethylene) also change during plant development [14–21]: BR, GA and IAA are important across tissues and plant organs throughout development [14,15,17,20]. ABA specifically inhibits seed germination, while IAA and JA promote germination [19,22]. Active JA pools are then reduced during seedling establishment and vegetative growth before increasing again during reproductive stages [16]. CK is important for seedling establishment with reduced levels found during later reproductive stages [18]. ET promotes later stages of development and in particular fruit ripening [21].
Whilst a lot of emphasis has been placed on inositol polyphosphates as signals of cellular ‘P-replete’ status and their impact on the interaction between SPX DOMAIN PROTEIN (SPX) repressors and their most prominent target TF PHOSPHATE STARVATION RESPONSE1 (PHR1), relatively little attention has been paid to the opportunities for cross-talk with well-established hormonal signalling pathways this molecule offers. Inositol polyphosphates have long been recognised as cofactors for SKP1-CULLIN-F-BOX PROTEIN (SCF) complex assembly with TRANSPORT INHIBITOR RESPONSE1 (TIR1) or CORONATINE-INSENSITIVE PROTEIN1 (COI1) that facilitate ubiquitin-mediated degradation of transcriptional repressors such as AUXIN RESPONSE FACTOR (ARF) or JASMONATE-ZIM-DOMAIN PROTEIN (JAZ) proteins. This review will cover the most recent advances in our understanding of nutrient — and in particular phosphate — signalling networks, from signal generation, its perception by regulatory proteins ranging from transcriptional repressors to ubiquitin-proteasome system (UPS) components, to the integration of metabolic and hormonal signals such complex networks offer for achieving an overarching systemic response to environmental or developmental clues. We also discuss recent technological advances that will provide further insight into signals, ubiquitin-proteasome components, and their assembly, and conclude with a list of open questions and challenges for experimental design.
Phosphate sensing, acquisition, and use
Plants require phosphate (Pi) for the biosynthesis of lipids, nucleic acids and ATP as well as energy transfer, the regulation of protein activity, and metabolic processes such as respiration and photosynthesis [13]. Plant responses to limitations in available soil Pi include morphological and biochemical adjustments which either conserve Pi or enhance uptake and transport [13,30,31]. Responses require the regulation of Pi deficiency-associated genes, including phosphate starvation induced (PSI) and phosphate starvation suppressed genes [32]. A key component of the PSR in Arabidopsis thaliana, MYB-CC TF PHOSPHATE STARVATION RESPONSE1 (PHR1) recognises common P1BS motifs (GNATATNC) in the promoter or 5′ untranslated regions of PSR genes [23,33]. The highly conserved 65-bp EZ2 motif contains two of these motifs and is enriched in genes involved in Pi recycling such as phospholipases — but also in regulatory components such as the SPX DOMAIN GENE1 (SPX1) [34]. Although PHR1 is known to regulate the PSR and is degraded in response to environmental cues, such as the accumulation of arsenate [As(V)] [33,35], it is only mildly responsive to changes in cellular Pi concentration at a transcript level [23]. In A. thaliana, many genes associated with the PSR have been defined [23,33,36–41], but the precise mechanisms of Pi sensing and feedback regulation of PHR1 have remained elusive.
Initiation of phosphate acquisition from the rhizosphere depends on the sensing of external and internal phosphorus (P) status. Plants can discern the identity of signalling molecules with a high degree of specificity, and this impacts their responses at a molecular level. Despite their structural similarity, there is a competitive inhibition of phosphite (, Phi) uptake in the presence of Pi, and discrimination against phosphite by phosphate transporter (PHT) proteins not only impairs Phi transport, but also alters subsequent signalling pathways so that they are distinct from Pi sensing events [42,43]. While PHTs are unable to discriminate between Pi and arsenate plants have, for example, recruited F-box proteins, i.e. the substrate-recruiting subunits of E3 ubiquitin ligases, for arsenite detection which then triggers detoxification [35]. Although key upstream regulators, like PHR1, are impervious to changes in Pi at a transcriptional level, the specificity and affinity of their interactions with each other or target promoters may be modified in the presence or absence of signalling molecules to facilitate a range of physiological and molecular responses across tissues, organs, and throughout development. Proteins containing SPX domains have been established as crucial components for the internal sensing of overall P status, the regulation of cellular Pi homeostasis, and the integration of environmental cues [40,44–48]. SPX domain-containing proteins fall into four groups depending on the presence of additional domains, where class I SPX proteins consist exclusively of the SPX domain itself. Most plants have at least three class I SPX genes. Arabidopsis thaliana encodes four SPX genes — with highly similar SPX1 and SPX2 most likely derived from a recent tandem gene duplication event [45]. AtSPX1 and AtSPX2 are transcriptionally induced in P-limited Arabidopsis seedlings and regulate nuclear PHR1 function [38]. Their response to Pi starvation is puzzling given that SPX repressors are only interacting with PHR1 in P-replete condition. Transcript accumulation following Pi withdrawal over time may ‘prime’ the plant for ribosome loading to quickly suppress nuclear PSR when Pi is again plentiful. SPX3 is the least studied paralog in Arabidopsis, most likely due to non-viable knockout plants and severe phosphate-deficiency symptoms in RNAi lines with partial SPX3 suppression [49]. Further research into its function is needed as it is strongly induced in P-limited roots and may play as-yet uncharacterised roles in PSR. In both rice and Arabidopsis, SPX4 genes are not transcriptionally responsive with even a slight suppression observed in P-limited Arabidopsis seedlings [40,49,50]. SPX4 proteins sequester PHR1 in the cytosol and prevent its translocation to the nucleus when Pi is readily available [40,51]. Nuclear sequestration of PHR1 by SPX1 and SPX2 is maintained in roots of P-replete spx4 knockout lines, while impaired PSR induction occurs in P-replete spx4 shoots [40]. This added layer of control in shoots may allow SPX4 to interact with TFs other than PHR1 for the integration of other environmental and developmental clues. This was evident from the transcriptional networks of genes mis regulated in spx4 shoots that were associated not only with ion transport and flavonoid biosynthesis, but also the response to jasmonic acid and regulation of shoot development [40].
The degradation of SPX4 in Pi-limited conditions was disrupted by the application of cycloheximide (CHX), an inhibitor of protein synthesis, indicating that de novo synthesis of an unknown factor is a prerequisite for Pi-dependent SPX4 turnover [40]. It was suggested that the degradation of SPX4 proteins is preceded by the assembly of ubiquitin machinery components, and that the configuration of these components is connected to an even earlier organic P (or Pi) sensing event. Consistent with this observation, many E3 ligase encoding genes are differentially expressed in P-limited Arabidopsis seedlings with F-box, REALLY INTERESTING NEW GENE (RING) finger and U-box family members significantly enriched in the list of PSI genes in shoots [26].
Signalling of plant phosphorus status must be specific and dynamic
A long-standing question in plant nutrient research has been which primary signals confer information about cellular P status. In mammals, yeast, algae and plants, inositol polyphosphates (InsPs) and pyrophosphates (PP–InsPs) are well studied central messengers of energy status [52,53]. Recent work in Arabidopsis has shown that InsP and PP–InsP levels are responsive to changes in Pi availability, and that SPX domains provide binding surfaces for InsPs [54–56]. The vacuolar P storage compound phytic acid (InsP6), as well as two PP–InsPs (5-PP–InsP5/InsP7, 1,5-(PP)2–InsP4/InsP8) were tested for their impact on SPX1–PHR1 complex formation, with InsP8 being most effective in promoting interactions between SPX domains and TF targets [55,57,58]. Ried et al. [57] found that SPX interaction with PHR1 occurs via the plant-specific coiled-coil (CC) domain of this MYB TF. The authors emphasise that critical residues for interaction are conserved across plant MYB–CC proteins which may suggest that SPX proteins are more general PP–InsP-dependent repressors of this TF family [59]. This view is also supported by enrichment of MYB TFs amongst differentially expressed genes in the spx4-1 mutant [40]. The fact that InsP8 and not lower-order InsP have the highest affinity to phosphate binding cluster (PBC) and lysine surface cluster (KSC) binding sites in the SPX domain fits well with the previously observed depletion of InsP8 and accumulation of InsP7 in P-limited shoots which is most likely caused by reduced VIP HOMOLOG1 (VIH1) and VIH2 kinase activity [54,56,57]. Expression of the latter was predominantly observed in shoots [60], so the local reduction in InsP8 would subsequently trigger PHR1 release from SPX4. The PBC and KSC binding sites are also highly conserved across SPX domain containing proteins both within and between plant species which would make InsP8 their preferred target metabolite [45]. As mentioned earlier, on top of this direct metabolic control, SPX4 is also subject to P-status dependent protein turnover and is degraded via the 26S proteasome under Pi limitation [40,51]. Unlike the other SPX class I members, SPX4 proteins from Arabidopsis and rice feature ubiquitination sites albeit these are not conserved between the two species (Figure 2).
SPX4 proteins of Arabidopsis and rice vary greatly in their C-terminal domain structure and predicted ubiquitination sites.
Protein alignment of AtSPX4 (top) and OsSPX4 (bottom) with amino acid sequence consensus in the middle. Annotations include SPX domain and sub-domains (red). Lysine surface clusters associated with inositol phosphate (IP) binding (light blue) are conserved across proteins. SDEL1 E3 ligase interaction domain for OsSPX4 is illustrated in pink (OsSPX4159-208) [61]. Ubiquitination sites are shown in green. Lysine residue K299 in OsSPX4 is not conserved in AtSPX4. Ub sites predicted for AtSPX4 using UbPred [62] are indicated in the top panel.
Protein alignment of AtSPX4 (top) and OsSPX4 (bottom) with amino acid sequence consensus in the middle. Annotations include SPX domain and sub-domains (red). Lysine surface clusters associated with inositol phosphate (IP) binding (light blue) are conserved across proteins. SDEL1 E3 ligase interaction domain for OsSPX4 is illustrated in pink (OsSPX4159-208) [61]. Ubiquitination sites are shown in green. Lysine residue K299 in OsSPX4 is not conserved in AtSPX4. Ub sites predicted for AtSPX4 using UbPred [62] are indicated in the top panel.
The effects of InsPs as signalling molecules are based on the specificity of their interactions [63]. PP–InsPs are signalling molecules in auxin, jasmonate, gibberellin and phosphate signalling pathways, where varying degrees of phosphorylation have been associated with divergent downstream functions [57,60,64,65]. One prerequisite for PP–InsPs signalling is their low abundance with cellular InsP7 concentrations of 1–5 µM and InsP8 concentrations ∼10–20% of those of InsP7 [66]. Another parameter is their localised synthesis from either free InsPs or membrane-associated phosphatidylinositol, the latter creating subcellular microdomains with restricted diffusion of the signal [67]. Inositol polyphosphate kinase INOSITOL-PENTAKISPHOSPHATE 2-KINASE1 (IPK1) was first associated with the abundance of InsP6 in Arabidopsis, where disruption of AtIPK1 expression resulted in a significant increase in the precursors InsP4 and InsP5 [68]. The pyrophosphorylation of InsP6 by INOSITOL 1,3,4-TRISPHOSPHATE 5/6 KINASE1 (ITPK1) and ITPK2 is followed by further phosphorylation of InsP7 to InsP8 by diphosphoinositol pentakisphosphate kinases VIP HOMOLOG1 (VIH1) and VIH2 [60]. Constitutive PSR in itpk1-1 mutants, despite varying Pi availability was a key indicator that Pi sensing in plants is associated with changing InsP levels, and not direct sensing of Pi [58,68]. This is also suggested by a lack of suppression of most PSR genes by the phosphate-mimic phosphite [69]. SPX domains bind PP–InsPs with high affinity, and their association is molecular species selective: InsP8 directly mediated the association between SPX1 and PHR1, but InsP7 could not [55]. InsP8 levels respond in parallel to changes in intracellular Pi, and InsP8 binding to SPX domains is essential for interactions between SPX1 or SPX4 with PHR1 [55,57] and the subsequent attenuation of the PSR. During interactions between SPX domains and PP–InsPs, it is notable that one ligand face remains accessible for additional coordination with target proteins such as PHR1 or other ligands such as plant hormones [58].
Inositol polyphosphates confer signals across multiple pathways
Although the designation of InsP8 as a ‘phosphate signalling molecule’ may be accurate, this fails to capture the complexity and breadth of their potential as signalling molecules. ITPK1 and VIH1/2 exhibit interdependent activity related to the changing abundance of different PP–InsPs, where bifunctional enzymes VIH1/2 will increase either phosphatase activity when Pi is readily available, or kinase activity when there is a reduction in cellular ATP [56]. Despite the absence of a classical phosphatase domain, ITPK1 has recently been found to dephosphorylate a specific isomer of InsP7 in the presence of ADP, allowing for dynamic shifts in the specificity of its enzymatic activity depending on local ATP/ADP ratios in association with changing Pi availability [70]. The perception of abiotic stress at the plasma membrane leads to phospholipid hydrolysis by phospholipases C (PLCs) which releases diacylglycerol and inositol-1,4,5-trisphosphate (InsP3) [71]. In plants, further phosphorylation of InsP3 to InsP6 by the above InsP kinases might activate downstream signalling across several pathways [71,72]. It is possible that the synthesis or degradation of these metabolites is one of the earliest events in many stress signalling pathways, and it is evident that the prevalence of species will impact the specificity of downstream responses.
While phospholipids have long been associated with salicylic acid (SA) signalling [73,74], IPK1 and ITPK1 have only recently been implicated as suppressors of SA dependent immunity [75]. The modulation of InsP/PP–InsP synthesis by these enzymes appears across hormonal pathways, including the VIH2 mediated induction of InsP8 by jasmonate in Arabidopsis shoots, and InsP6 and InsP7 species enhancing the efficiency of auxin signalling by activating the auxin receptor complex [27,60]. Due to parallels in the overall hierarchical interaction of components, regulation by dynamic ubiquitination of transcriptional repressors in the latter two pathways could provide a model for the regulation of Pi homeostasis.
Ubiquitin-tagging of the JAZ repressors for proteasomal degradation is dependent on the recruitment of F-box protein COI1 to the jasmonate (JA) receptor complex, whose activity is moderated by interactions with specific InsP isomers [76]. In vitro competitive binding assays found that the binding affinity of InsP5 [3-OH] to the JAZ1-jasmonate receptor complex was 16-fold higher than its enantiomer InsP5 [1-OH]. Furthermore, it was found that inositol pyrophosphates were bound to the COI1–JAZ1 complex with a higher affinity than either InsP6, or InsP5 [60]. Laha et al. [60] concluded that InsP8 was the most likely in vivo ligand to promote jasmonate perception by F-box protein COI1 and subsequent recruitment of JAZ repressors to the ubiquitination complex.
The degradation of AUXIN/INDOLE-3-ACETIC ACID (AUX/IAA) repressors is mediated by the recruitment of auxin-binding F-box protein TIR1 or one of its five paralogs AUXIN SIGNALLING F-BOX1 (AFB1) to AFB5 to the auxin receptor complex. While a reduction in InsP8 levels in planta was not associated with the disruption of auxin signalling, in vitro competitive binding assays established distinct affinities of the ASK1–TIR1–IAA auxin receptor complex for InsP isomers, where against InsP5 and InsP6, InsP7 showed the highest affinity [27]. The TIR1 F-box protein shows a high degree of relatedness to COI1, with strict conservation of residues associated with the interaction of InsP cofactors, as well as surface pockets for the recognition of hormones and repressor protein substrates [64,77–79].
The intriguing similarity of components raises the question whether a similar receptor complex mediates the degradation of the SPX4 repressor in phosphate signalling, and which F-box proteins are required for its assembly. When Pi is sufficient, PP–InsPs — specifically InsP8 — act as a cofactor which strengthens associations between SPX repressors and PHR1, thereby limiting the induction of PSI genes. When Pi is limiting, InsP8 levels decline and the SPX4 repressor is degraded releasing PHR1 into the nucleus. If receptor complex assembly between SPX4, unknown F-box protein(s) and specific signalling molecules is required, it is possible that the accumulation of lower order InsP or PP–InsP species might initiate assembly of the SPX4 degron in the absence of Pi. Riemer et al. [70] showed that 1/3-InsP7, 5-InsP7, and InsP8 increase specifically in shoots within 12 h of Pi resupply, and that this increase is even stronger in the pho2-1 mutant that hyperaccumulates phosphate. As mentioned earlier, InsP7 levels increased specifically in P-limited seedling shoots where SPX4 also exerts a stronger function [40,54]. What is more, TIR1 has a higher affinity to InsP7 [27] and is transcriptionally activated by Pi limitation in roots [27,80] which then is reinforced by ARF-mediated induction of PHR1 [81]. These findings suggest that a transient spike in InsP7 or other lower order InsPs could be required for SPX4 degradation. Thus, it matters where in the plant, when during stress or development, and by which enzyme or in which subcellular compartment the InsP signal is generated. Characterising ubiquitin machinery components and the specificity of their interactions with signalling molecules and their isomers will be crucial for understanding how different responses might be elicited across pathways.
The ubiquitination machinery allows for further signal diversification
Responses to changing environments are diversified by post-translational protein modifications such as ubiquitination, acetylation, phosphorylation, and addition of SMALL UBIQUITINRELATED MODIFIER proteins (SUMOylation). These responses are often reversible and are highly specific to a given condition, allowing for precise and rapid acclimation with minimal disruption to cellular and developmental processes. The UPS is responsible for the turnover of target substrates and degradation of ‘damaged’ proteins, thereby regulating a wide array of cellular and developmental processes. For target proteins, the consequences of covalent attachment of ubiquitin (Ub) residues are diversified by the so-called ubiquitin code — variations in Ub chain linkage and the number of lysine residues in the target substrate to which ubiquitin is attached [82–84]. Non-canonical protein ubiquitination can induce non-proteolytic modifications impacting target protein activity or receptor recognition — for example during endosomal sorting of plasma membrane target proteins ubiquitinated at lysine 63 (K63) residues. Intracellular signalling is predominantly associated with ubiquitin chains (poly-Ub) linked through lysine 48 (K48), which usually initiates proteolysis of the target proteins via the 26S proteasome [85,86]. Ubiquitin tagging of substrates is preceded by an ATP dependent Ub conjugation cascade involving a ubiquitin activating enzyme (E1), a ubiquitin conjugating enzyme (E2) and a ubiquitin ligase (E3). The diversity of these enzymes facilitates the high substrate specificity needed for dynamic responses to changing conditions, with 2 E1 isoforms, 37 E2 isoforms and over 1400 known E3 ligases in Arabidopsis [87]. E3 ligases may be divided into two types based on whether they contain a RING domain, or a HOMOLOGY TO E6-AP C-TERMINUS (HECT) domain. RING domain containing E3 ligases may be sub-divided into single proteins containing a substrate interaction motif, and multi-subunit E3s, whose components provide further diversity for E2 recognition and substrate specificity. While RING domain containing E3 ligases tend to be associated with specific hormonal or stress pathways, HECT domain containing E3 ligases are often associated with the proteasome itself. Both types of E3 ligases together form stress-specific relays and mediate the handover of substrates between target specific complexes and the proteasome, with HECT-type ligases controlling the continued and effective degradation of these targets [88–90]. Chain trimming by deubiquitinating enzymes at proteasome level helps to fine-tune the response [91,92]. In addition, there is also the unanswered question of which components regulate E3 ligase turnover itself [93].
In plant hormone signalling, ubiquitination either targets TFs or their repressors via proteasomal degradation [94–97]. Less is known about the E3 ligases involved in nutrient signalling. The E2 conjugase PHOSPHATE2 (PHO2) and the RING-type E3 ligase NITROGEN LIMITATION ADAPTATION (NLA) target plasma membrane bound phosphate transporters PHOSPHATE TRANSPORTER1;1 (PHT1;1), PHT1;4 and PHOSPHATE1 (PHO1, a phosphate exporter) for lysosomal degradation in P-replete conditions [24,98–101]. Membrane-localised RING-type E3 ligase ARABIDOPSIS TOXICOS EN LAVADURA8 (ATL8) was proposed as a positive regulator of PSR in roots without providing a mechanistic insight [102]. Ye et al. [103] found that F-box protein PHOSPHATE RESPONSE UBIQUITIN E3 LIGASE1 (PRU1) is part of an SCF complex that targets transcriptional repressor WRKY6 in Pi limiting conditions, thereby increasing PHOSPHATE1 (PHO1) expression and subsequently Pi loading into the xylem. Transcript levels of PRU1 were accumulating over time following Pi withdrawal making it a candidate for SPX4 turnover. RING-containing SPX4 DEGRADATION E3 LIGASE1 (OsSDEL1, Os12g35320) and its homolog OsSDEL2 were identified as regulators of OsSPX4 ubiquitination and turnover [61]. Under P-limitation, the preferential association between OsPHR2 and OsSPX4 was impaired by altered InsP levels liberating OsPHR2 for translocation to the nucleus and allowing SDEL1 to interact with helical domains of OsSPX4 for ubiquitination and subsequent proteasomal degradation. The orthologs of these E3 ligases are yet to be identified in Arabidopsis.
The interplay between nitrogen and P regulatory networks has become a focal point of nutrient signalling research [104]. It has been established that SPX4 also plays a role in nitrate signalling with its turnover mediated by yet another E3 ligase. In the presence of , preferential association between OsSPX4 and membrane-bound high-affinity nitrate transporter and sensor OsNRT1.1B facilitates the ubiquitination of OsSPX4 by the recruitment of RING-type E3 ubiquitin ligase NRT1.1B INTERACTING PROTEIN1 (OsNBIP1) [105]. This work found that OsNBIP1-triggered OsSPX4 degradation was implicated in the liberation of both OsPHR2 and OsNLP3, the master regulator of nitrate-inducible gene expression and ortholog of AtNLP7. Together, these findings would imply involvement of different E3 ligase complexes controlling SPX4 stability in response to different nutrient signals and in different subcellular locations. Whilst the presence of InsPs strengthen the SPX4–PHR1 interaction, potentially obfuscating the ubiquitination of SPX4, it remains unknown if ubiquitination components directly sense lower cellular Pi or lower order InsPs to actively recruit SPX4 to the degradation complex. Given the many parallels to SCF-complex mediated degradation of hormonal repressors, it is tempting to propose that a low P status sensing SCF complex is involved in SPX4 degradation.
Research has established that the degradation of PHR1 via the ubiquitin-26S proteasomal pathway leads to repression of phosphate transporters responsible for arsenate [As(V)] uptake, and that PHR1 protein stability is restored following the vacuolar sequestration of arsenite [As(III)] [106]. In a mechanism which is analogous to selective discrimination by COI1 for structurally similar isomers of InsP in jasmonate signalling [60], plants can moderate the selective uptake of toxic arsenic (As) whilst permitting the acquisition of phosphate, despite structural similarities between the two molecules. ARABIDOPSIS SKP1-LIKE18 (ASK18) was identified from transcriptomic data as the SCF complex associated transcript whose induction by cytosolic As(III) was associated with the degradation of PHR1 [35]. In the presence of cytosolic As(III), another SCF complex component, PHR1 INTERACTOR F-BOX (PHIF1) was found to interact with both ASK18 and PHR1, and phif1 mutants were found to accumulate both phosphate and arsenic at higher levels than wildtype plants. If the specificity for InsP isomers in jasmonate signalling is dictated by the recognition of F-box protein COI1, it is possible that F-box proteins associated with Pi signalling have a similar capacity for signal recognition which integrates and discriminates between cellular concentrations of phosphate, As and InsPs. It is likely that characterisation of F-box proteins, and analysis of their interactions with specific signalling molecules, will reveal their capacity to discriminate between or integrate different metabolites and hormones across signalling pathways.
F-Box proteins recruit target substrates for degradation
F-Box proteins (FBX) are sub-units of SCF complexes, which can trigger signal-dependent degradation of target proteins. Along with CULLIN1 (CUL1) or CUL2, and 1 of 21 ARABIDOPSIS SKP1-LIKE (ASK) proteins, FBX proteins assemble in a SCF complex. FBX can interact with ASK proteins in both cytoplasm and nucleus, which in turn interact with the N-terminus of the CUL1 scaffold. The C-terminus of CUL1 recruits RING-BOX1 (RBX1) and RBX2 E3 ligases, and E2 conjugating (UBC) enzymes for assembly of the modular Cullin-RING E3 ligase (CRL) ubiquitination complex. The identity of each component and their affinity for one another determines the identification and specification of target proteins by the SCF–CRL complex [29,107–109]. However, it has been suggested that the abundance of FBX proteins might be the key metric associated with target selection for protein degradation by their capacity to form — or in their absence, to disassemble — SCF–CRL complexes [110].
In Arabidopsis, over 700 FBX proteins have been reported, vastly greater than numbers found in human or yeast genomes [29]. Across the plant kingdom, the FBX protein family is diversified further by gene duplication, followed by domain shuffling and modified recruitment scaffolds for target proteins. Phylogenetic analyses have also indicated that there is significant variation between species [111]. It has been proposed that within the superfamily of plant FBX proteins, some clades have undergone significant expansion or contraction in a relatively brief evolutionary period, potentially accounting for known differences in ubiquitination across species [112]. There is also divergence in the expression of FBX genes within populations, contributing to observed variations in nutrient stress tolerance within a species [113]. This is often difficult to confirm, since environmentally induced changes in the expression of FBX genes are by nature, highly transitory, and FBX are usually expressed at low levels [112]. In contrast, ASK genes are transcribed more readily across tissues, and exhibit polymorphisms which alter their affinity for FBX proteins, thereby contributing to adaptive diversity in SCF complex assembly across species [107,114]. Given these SCF complexes have the capacity to distinguish between toxic elements and nutrients [103,106], it is likely that combinations of FBX signal recognition and ASK target specificity allow for the differentiation and integration of a broad range of environmental inputs [115].
The ubiquitination of target proteins by SCF–CRL complexes is gated by target protein substrate, Cullin-associated Nedd8-dissociated protein1 (CAND1) mediated FBX exchange, and reversible modifications by ubiquitin-like protein NEDD8, known as neddylation or deneddylation [116–118]. While neddylation facilitates the productive assembly of the CRL complex, deneddylation inhibits CRL assembly, and prevents constitutive proteasomal degradation [116,119]. The removal of NEDD8 is catalysed by the COP9 signalosome (CSN) which further disrupts E3 ligase activity through binding and sequestration of deneddylated CRLs. CAND1 attaches to deneddylated CRL complexes and facilitates the exchange of FBX proteins. In addition, substrates already targeted for ubiquitination are steric inhibitors of CRL complex inactivation by the CSN, and new incoming substrate may dislodge the CSN from the CRL complex, allowing for Cullin neddylation and successive ubiquitination events [120,121]. It has been proposed that CRL–CSN complexes may be assembled with varying affinity, and that low-affinity states would favour CAND1-mediated FBX exchange and complex disassembly, whereas high affinity states would be associated with the degradation of target proteins by neddylated SCF–CRL complexes. A working model was suggested which tied fluctuations between high and low affinity CSN–CRL complexes to the abundance of InsP6 and InsP7, respectively [122], although the generality of this mechanism and its conservation across species has yet to be confirmed.
The role of SCF–CRL complexes in the PSR
Impaired IPK1 or ITPK1 function results in the overaccumulation of neddylated CUL1 and is associated with constitutive PSR in Arabidopsis when Pi is readily available [54,70,123]. In P-limiting conditions, the ubiquitination of substrates by neddylated CRLs is regulated by changes in the activity of InsP-kinases IPK1 and ITPK1 which alter local PP–InsP levels via direct interaction with each other, various CSN components as well as CUL1 and ASK1 [123,124]. Hence, stability and recycling of CRLs by neddylation are related to the effects of changing Pi availability on InsP kinase activity and associated InsP and PP–InsP levels. It was suggested that perception of Pi deficiency, via dissociation of higher order InsP or PP–InsP species from CSN components and resulting neddylation of CRL components allows for substrate degradation [123]. One of the most likely substrates for the latter is SPX4. Further investigation is needed to understand how specific SCF complex components are responsive to Pi availability, and whether association between SPX4 and an interacting FBX protein is impacted by the phosphorylation or dephosphorylation of InsP species or the target protein itself under varying Pi availabilities. Similar to InsP8 induced strengthening of the SPX4–PHR1 interaction P status dependent turnover of SPX4 can be attributed to strengthened or weakened associations between CRL components and the CULLIN1 (CUL1) scaffold in the presence (or absence) of signalling molecules such as arsenic and InsPs, and it is possible that effector molecules generated under a given condition may regulate cross-talk between pathways, either through promotion of SPX4 ubiquitination or its preferential binding to TF targets.
Figure 3 integrates recent findings related to FBX activity in response to substrate availability and proposes mechanisms by which the destabilisation of SCF–CRL complexes, combined with the stabilisation of cytosolic sequestration of SPX4 by PHR1, might prevent the expression of PSR when phosphate is readily available: Pi dependent activity of InsP kinases and phosphatases leads to increased phosphorylation of inositol phosphate species and higher concentrations of InsP8. InsP8 promotes sequestration of SPX4 by PHR1, preventing the nuclear migration of PHR1 and subsequent induction of downstream PSI genes. Cycling of FBX and substrate adaptors is facilitated by the deneddylation of the CRL complex by CSN5 when the substrate (SPX4) is not available in its free form. If CSN–CRL complex stability and CAND1-mediated FBX exchange are tied to fluctuations in InsP species as was recently shown for InsP6 in mice [125], this would provide an additional layer of regulation across pathways.
Model of proposed SCF–CRL components and InsP kinases involved in the stabilisation of SPX4 in shoots when Pi supplies are sufficient (top).
(A) Increased phosphorylation of InsP6 to InsP7 by IP6K-like kinases. Phosphorylation of InsP7 to InsP8 by kinase domain (K) of Vip/PPIP5K-like kinase VIH2. (B) Presence of InsP8 stabilises association between PHR1 and SPX4. One ligand face of SPX4 protein remains accessible for coordination with other proteins. (C) F-box protein (FBX) assembly with CUL1 scaffold is widely varying in efficiency, and assembly cycle will continue until FBX and ASK recruit target substrate. (D) Deneddylation of CRL complex by CSN5 subunit of the COP9 signalosome (CSN) and shuffling of FBX/ASK protein interactors occur to promote new associations with other ASK/FBX combinations until target substrate (SPX4) is available for binding. Representation of SCF–CRL complex will match substrate demands. Model of proposed SCF–CRL components and InsP kinases involved in the degradation of SPX4 in shoots when Pi is limiting (bottom). (E) VIH2 exhibits decreased kinase (K) activity and increased phosphatase (P) activity, lowering levels of InsP8 whilst increasing levels of InsP7 and lower order InsPs. (F) Lower concentration of InsP8 weakens association between SPX4 and PHR1, liberating PHR1 for nuclear migration and expression of PSR genes. Binding of InsP6/7 by an unknown FBX may strengthen its recruitment of SPX4. SPX4-FBX/ASK are subsequently recruited to the CUL scaffold for the ubiquitination of SPX4. (G) Neddylated CUL1 stabilises FBX/ASK protein interactions. Active SCF–CRL complex recruits ubiquitin carrying enzymes E2 conjugase (UBC), E3 ligase (RBX). Continued neddylation is gated by substrate (SPX4). (H) CSN is autoinhibited in the absence of NEDD8.
(A) Increased phosphorylation of InsP6 to InsP7 by IP6K-like kinases. Phosphorylation of InsP7 to InsP8 by kinase domain (K) of Vip/PPIP5K-like kinase VIH2. (B) Presence of InsP8 stabilises association between PHR1 and SPX4. One ligand face of SPX4 protein remains accessible for coordination with other proteins. (C) F-box protein (FBX) assembly with CUL1 scaffold is widely varying in efficiency, and assembly cycle will continue until FBX and ASK recruit target substrate. (D) Deneddylation of CRL complex by CSN5 subunit of the COP9 signalosome (CSN) and shuffling of FBX/ASK protein interactors occur to promote new associations with other ASK/FBX combinations until target substrate (SPX4) is available for binding. Representation of SCF–CRL complex will match substrate demands. Model of proposed SCF–CRL components and InsP kinases involved in the degradation of SPX4 in shoots when Pi is limiting (bottom). (E) VIH2 exhibits decreased kinase (K) activity and increased phosphatase (P) activity, lowering levels of InsP8 whilst increasing levels of InsP7 and lower order InsPs. (F) Lower concentration of InsP8 weakens association between SPX4 and PHR1, liberating PHR1 for nuclear migration and expression of PSR genes. Binding of InsP6/7 by an unknown FBX may strengthen its recruitment of SPX4. SPX4-FBX/ASK are subsequently recruited to the CUL scaffold for the ubiquitination of SPX4. (G) Neddylated CUL1 stabilises FBX/ASK protein interactions. Active SCF–CRL complex recruits ubiquitin carrying enzymes E2 conjugase (UBC), E3 ligase (RBX). Continued neddylation is gated by substrate (SPX4). (H) CSN is autoinhibited in the absence of NEDD8.
Under phosphate limitation (Figure 3, lower panel), lower levels of InsP8, and consequently higher concentrations of lower order InsPs such as InsP7, will release PHR1 from its SPX4 repressor to initiate PSR in the nucleus. What is currently unclear is how SPX4 degradation is then triggered and which E3 ligases are involved in its ubiquitination. We propose that sensing of lower order InsPs by an unknown FBX could facilitate recruitment of SPX4 to an unknown SCF–CRL complex. By elucidating how individual components might respond under decreased phosphate availability, putative candidates might be identified from forward genetic screens and emerging literature more efficiently. Characterisation of these components and their response to varying nutrient availabilities will contribute to the refinement of this working model. The possibility that CSN–CRL complex stability and CAND1-mediated FBX exchange are also tied to fluctuations in InsP species would provide an additional layer of regulation in this pathway that should be investigated further.
An integrated model of PSR and jasmonate signalling
The usefulness of modelling early responses to Pi availability in the context of SCF–CRL mediated repressor degradation, and in relation to changing InsP species is bolstered when the model is viewed in parallel with other nutrient and phytohormonal response pathways. When consolidating recent findings on interactions between TFs PHR1 and MYC2 which co-regulate jasmonate and phosphate responses via repressor turnover, a useful model for cross-talk between pathways emerges (Figure 4).
Model of proposed interactions between MYC2 mediated jasmonate response and PHR1 regulated response to varying Pi availability in shoots.
(A) When both Pi and jasmonate are available, coincidence detection of JA-Ile and InsP8 promotes assembly of SCFCOI1 ubiquitination complex and liberates MYC2 for jasmonate pathway induction. Cytosolic sequestration of PHR1 by SPX4 — promoted by the presence of InsP8 — prevents expression of PSI genes. (B) When Pi is sufficient, InsP8 are abundant, promoting cytosolic associations between SPX4 and PHR1 and preventing induction of PSI genes. In the absence of JA-Ile, SCFCOI1 ubiquitination complex is assembled with low affinity and JAZ repressor proteins inhibit MYC2-mediated jasmonate signalling. (C) When Pi is deficient, lower order InsP species replace InsP8, and SPX4 dissociates from PHR1 for its translocation to the nucleus and PSR. SPX4 is recruited by unknown FBX for ubiquitination. The FBX may require lower order InsP species to facilitate substrate binding. In the absence of JA-Ile, MYC2-mediated JA signalling would not be induced, as JAZ repressors interact with MYC2. This scenario is unlikely given that the PSR triggered by PHR1 also induces JA biosynthesis. (D) When jasmonate is available, phosphate starvation signalling pathways are induced by TF PHR1, but SCFCOI1-mediated JA signalling is not induced in the absence of InsP8. Inhibition of JAZ repressors via recruitment to CUL1 scaffolds under low Pi conditions thus requires further characterisation and may have been replaced by direct competition between PHR1 and JAZ for MYC2 binding for nuclear translocation.
(A) When both Pi and jasmonate are available, coincidence detection of JA-Ile and InsP8 promotes assembly of SCFCOI1 ubiquitination complex and liberates MYC2 for jasmonate pathway induction. Cytosolic sequestration of PHR1 by SPX4 — promoted by the presence of InsP8 — prevents expression of PSI genes. (B) When Pi is sufficient, InsP8 are abundant, promoting cytosolic associations between SPX4 and PHR1 and preventing induction of PSI genes. In the absence of JA-Ile, SCFCOI1 ubiquitination complex is assembled with low affinity and JAZ repressor proteins inhibit MYC2-mediated jasmonate signalling. (C) When Pi is deficient, lower order InsP species replace InsP8, and SPX4 dissociates from PHR1 for its translocation to the nucleus and PSR. SPX4 is recruited by unknown FBX for ubiquitination. The FBX may require lower order InsP species to facilitate substrate binding. In the absence of JA-Ile, MYC2-mediated JA signalling would not be induced, as JAZ repressors interact with MYC2. This scenario is unlikely given that the PSR triggered by PHR1 also induces JA biosynthesis. (D) When jasmonate is available, phosphate starvation signalling pathways are induced by TF PHR1, but SCFCOI1-mediated JA signalling is not induced in the absence of InsP8. Inhibition of JAZ repressors via recruitment to CUL1 scaffolds under low Pi conditions thus requires further characterisation and may have been replaced by direct competition between PHR1 and JAZ for MYC2 binding for nuclear translocation.
Following coincidence perception of JA-Ile and InsP8 when both jasmonate and Pi are readily available, JAZ repressor proteins are recruited by the SCFCOI1 ubiquitination complex and are degraded via the 26S proteasomal pathway (Figure 4A). This degradation relieves the repression of TF MYC2, triggering downstream jasmonate signalling. The induction of PSI genes by PHR1 is inhibited by its association with SPX4 in the cytosol. Under conditions where both jasmonate and InsP8 are available, PHR1 presence in the nucleus is not required to activate MYC2 as JAZ1 repressors are degraded by SCFCOI.
When there is an abundance of InsP8 associated with Pi availability, but bioactive JA conjugate JA-Ile is not present, SPX4 negatively regulates the PSR by its cytosolic sequestration of PHR1, and JAZ repressor proteins inhibit MYC2 mediated jasmonate signalling (Figure 4B) [25,126]. In the absence of JA signalling molecules, the SCFCOI1 ubiquitination complex has a lowered affinity for its target, liberating CRL components to assemble under different configurations.
As Pi availability decreases, InsP8 is dephosphorylated, and the abundance of lower order InsP species increases (Figure 4C). PHR1-mediated PSR is facilitated by the degradation of SPX4, but in the absence of JA-Ile, JAZ repressor proteins maintain their preferential association with MYC2, preventing the induction of jasmonate signalling pathways (Figure 4C). This scenario is rather unlikely, given that Pi limitation promotes jasmonate responses [127].
Upon the detection of both JA-Ile, and lower order InsP species under Pi limiting conditions, regulation of the jasmonate receptor complex is less clear (Figure 4D). The yet-to-be determined SCF complexes responsible for the degradation of SPX4 proteins would assemble with increased affinity in the cytosol, liberating PHR1 for Pi signalling. PHR1 induces jasmonate biosynthesis and responses [127]. MYC2-mediated JA responses, however, would not be induced via JAZ1 degradation in the absence of InsP8 [76]. It is possible that an alternate JA-signalling pathway is induced in P-limited plants, that does not require coincidence detection of JA-Ile and InsP8. In rice, overexpression of OsJAZ11 attenuates PSR while at the same time enhancing Pi uptake by alleviating JA effects on root growth inhibition [128]. Turnover of OsJAZ11 is greatly reduced in P-limited transgenic lines. In wildtype P-limited plants, PHR1 may compete with JAZ1 for binding to MYC2 [25]. And this may require PHR1 to be released by SPX4 in the cytosol. In this context, it is interesting to note that nuclear localisation of some JAZ repressors such as JAZ9, requires MYC2 interaction [129]. In rice, OsPHR2 also binds to the OsMYC2 promoter to boost antibacterial resistance [130] but this was not confirmed in Arabidopsis [25]. OsPHR2 is negatively regulated by nuclear localised OsSPX1 and OsSPX2, and OsPHR2 dimerisation and DNA binding may be inhibited in planta by OsSPX1 in the presence of InsP6 [131,132]. It is not yet clear how interactions with other SPX proteins might be consolidated with the negative regulation of PHR1 by SPX4 in the cytosol, and it is worth noting that interactions with both InsP and higher order PP–InsPs warrant further investigation.
The models presented here are useful to depict how the modular assembly of SCF–CRL complexes might allow for integration of Pi and jasmonate signalling pathways, but do not necessarily represent the breadth and diversity of responses across development and in response to abiotic or biotic stress. Despite a high degree of conservation across JAZ repressor proteins, not all interact with MYC2 or PHR1 in planta [25]. Interestingly, even when MYC2 interactions are conserved, different JAZ repressors have different affinities for FBX and/or ASK proteins [133]. It is also not clear how SCF–CRL complexes involved in other hormonal pathways, in particular SCFSON1–CRL3 mediated NON-EXPRESSER OF PR GENES1 (NPR1) turnover in SA dependent immune response [134], and SCFTIR1–CRL1-mediated AUX/IAA turnover in auxin signalling [27], and their affinity for various InsP species can be integrated into this model.
Modular assembly facilitates cross-talk between pathways
Recent characterisations of cross-talk between nutrient and hormone responses have often featured the assembly of SCF–CRL complexes, and integrated responses of repressor-TF interactions which are analogous to those proposed in Pi and jasmonate signalling pathways in Figure 4. There are opportunities to diversify responses through variations in the abundance, identity or isomer of the signalling molecule, changes in SCF complex components, or in their affinity towards substrates or each other, changes in CRL assembly and E3 ligase recruitment, and changes to CSN activity (Table 1). In this way, nutrient or phytohormone response pathways may be navigated by a series of qualifying steps or biochemical ‘logic gates’, initiated by the accumulation or decumulation of primary signalling molecules. By meeting each qualification, specific combinatorial assembly, or disassembly of ubiquitin proteasomal machinery is carried out, allowing signalling pathways to be induced hierarchically, or in parallel. These complex hierarchical interactions have been modelled for gibberellin (GA)-responsive genes which identified GA interaction with F-box protein GIBBERELLIN INSENSITIVE DWARF1 (GID1) as a key module in signal transduction [135].
. | Potential for diversity . | Consequence of diversity . | References . |
---|---|---|---|
Input signal | ATP : ADP InsPs PP–InsPs Nutrients: Hormones: ABA, BR, ethylene, GA, JA-Ile, IAA, SA | Metabolite phosphorylation/dephosphorylation by IPK1/ITPK1 and VIH1/VIH2 FBX recruitment of TS Affinity of preferential associations with TFs Affinity of CRL complex assembly Efficacy of CSN holoenzyme activity | [9,13,22,55–58,69,76,79,111,121,134] |
SCF assembly | FBX/TS FBX/ASK Inter-species variation Variance across populations | Engagement with TS drives E2/E3 ligase recruitment and CRL assembly Engagement with ASK drives E2/E3 ligase recruitment and CRL assembly Differences in recruitment or binding affinity impact efficacy of stress responses | [29,109–113,115] |
CRL assembly | UBC E3 ligase SCF complex affinity | Combinatorial substitutions allow for components to act in multiple pathways Allows for variations in UPS response across tissues and throughout development | [60,85,94,95,101–104,135] |
CSN activity | Neddylation/deneddylation CSN activity CAND1 activity | Addition of NEDD8 for decreased CRL complex affinity and interchange of SCF Steric obstruction of CSN by TS inhibits deneddylation and promotes TS ubiquitination CAND1 mediated FBX exchange ‘shuffles’ FBX/ASK in the absence of TS | [117–120,122,134] |
. | Potential for diversity . | Consequence of diversity . | References . |
---|---|---|---|
Input signal | ATP : ADP InsPs PP–InsPs Nutrients: Hormones: ABA, BR, ethylene, GA, JA-Ile, IAA, SA | Metabolite phosphorylation/dephosphorylation by IPK1/ITPK1 and VIH1/VIH2 FBX recruitment of TS Affinity of preferential associations with TFs Affinity of CRL complex assembly Efficacy of CSN holoenzyme activity | [9,13,22,55–58,69,76,79,111,121,134] |
SCF assembly | FBX/TS FBX/ASK Inter-species variation Variance across populations | Engagement with TS drives E2/E3 ligase recruitment and CRL assembly Engagement with ASK drives E2/E3 ligase recruitment and CRL assembly Differences in recruitment or binding affinity impact efficacy of stress responses | [29,109–113,115] |
CRL assembly | UBC E3 ligase SCF complex affinity | Combinatorial substitutions allow for components to act in multiple pathways Allows for variations in UPS response across tissues and throughout development | [60,85,94,95,101–104,135] |
CSN activity | Neddylation/deneddylation CSN activity CAND1 activity | Addition of NEDD8 for decreased CRL complex affinity and interchange of SCF Steric obstruction of CSN by TS inhibits deneddylation and promotes TS ubiquitination CAND1 mediated FBX exchange ‘shuffles’ FBX/ASK in the absence of TS | [117–120,122,134] |
ABA, abscisic acid; ASK, ARABIDOPSIS SKP1-LIKE protein; BR, brassinosteriod; CAND1, CULLIN-ASSOCIATED AND NEDDYLATION-DISSOCIATED1; CRL, CULLIN-RING E3 ligase complex; CSN, COP9 signalosome; FBX, F-box protein; GA, gibberellic acid; IAA, indole-3-acetic acid/auxin; JA-Ile, jasmonic acid-isoleucine conjugate; SA, salicylic acid; SCF, SKP1-CULLIN-FBX PROTEIN complex; TS, target substrate; TF, transcription factor; UBC, E2 conjugating enzyme.
Instances of similarly gated responses have recently been characterised in auxin signalling pathways, where ASK and TIR1 recruit AUX/IAA repressors for degradation with high affinity following exposure to InsP6 and 5-InsP7, and coincidence detection of both InsP7 and auxin [27,64]. The efficiency of the auxin receptor complex was found to depend on the generation of specific InsPs and PP–InsPs by inositol kinases, which influence the orientation with which TIR1 interacts with AUX/IAA and allowing TIR1 to discriminate between individual repressors [27]. Parallels may be drawn to jasmonate signalling, and potentially also Pi signalling, through the coincidence detection of multiple signalling molecules, and variations in complex affinity in response to changing InsP species. A direct physical interaction between ITPK1 and TIR1 might serve to prevent other signalling events initiated by InsPs and PP–InsPs [27]. So, it will be important to establish the hierarchy of ITPK1–FBX interactions across SCF–CRL complexes.
In addition to signalling molecules impacting the kinetics of SCF assembly, there is also competition amongst repressor proteins and their TF targets (Figure 4): Auxin dependent TFs ARF10 and ARF16 interact with both JAZ repressor proteins, and ABSCISIC ACID INSENSITIVE5 (ABI5) to regulate JA-mediated ABA responses [136]. The regulation of auxin signalling by TIR1- or AFB1- to AFB5-mediated turnover of AUX/IAA proteins is a qualifying prerequisite for the induction of JA-responsive ABA signalling during seed germination [22,136]. Mei et al. [136] found that ARF10 and ARF16 TFs were positive regulators of ABA signalling, which indicates that the efficacy of JA-dependent ABA signalling is preceded by the de-repression of ARF TFs upon the detection of auxin. Since the efficiency of the auxin receptor complex is also associated with coincidence detection of InsP7 and auxin, and TIR1 interacts directly with ITPK1 — it is possible that TIR1-mediated ubiquitination of AUX/IAA repressors precedes other TF-repressor interactions, or that auxin signalling is a prerequisite for subsequent signalling pathways, and this hierarchy is dictated by transitions between lower and higher order InsPs and PP–InsPs. The involvement of InsPs in the regulation of auxin, jasmonate and Pi sensing via SCF complexes provides a direct mechanism to regulate developmental processes such as lateral root growth to the availability of phosphate.
For the coordination of auxin and phosphate signalling, specifically the formation of lateral roots in response to low Pi, genes encoding MYB–CC domains, including PHR1, were also found to be targets of ARF TFs [81]. The fact that PSR was suppressed in auxin repressor iaa28-1 mutant would also suggest that some ARFs act as suppressors of PHR1 — or related CC–MYC [137]. Huang et al. [81] found that the binding of ARF7 and ARF19 to auxin-response elements of the PHR1 promoter is essential for the induction of PSI genes under low Pi availability, and phosphate uptake was reduced when ARF7 and ARF19 function was abolished. The identification of TIR1, ARF7 and ARF19 as potential targets of PHR1 by chromatin immunoprecipitation sequencing (ChIP-seq) could be indicative of regulatory feedback between ARF and PHR1 TFs and the subsequent degradation of repressor proteins in the proliferation of lateral roots [80,81].
In cross-talk between nitrate and phosphate starvation signalling pathways in rice, OsSPX4 may be ubiquitinated by OsSDEL1/2 under Pi starvation, but also binds to membrane bound transporter OsNRT1.1B and is ubiquitinated by OsNBIP1 upon sensing of [105]. It is possible that the binding affinities for FBX and ASK proteins responsible for the recruitment of E3 ligases in the PSR are responsive to dynamic shifts in PP–InsP species, as is suggested to be the case for cross-talk between Pi and jasmonate signalling. The associations between SPX4 and TF targets NLP7 or PHR1 might strengthen or weaken relative to the binding affinity of SCF–CRL components, allowing for fine-tuned and dynamic prioritisation of vital N and P responses. To reconcile which individual E3 ligase is preferentially assigned to the ubiquitination of SPX4, the environment which promotes interaction with free E3 ligases versus FBX/ASK SCF complex components should be investigated, under varying concentrations of and Pi as well as in response to overlayed hormone treatments.
Methods for quantitative characterisation of a dynamic system
Signalling molecules
The attributes which allow for rapid, dynamic, and reversible responses to environmental changes are by nature difficult to characterise. If the phosphorylation or dephosphorylation of InsP and PP–InsP species is responsible for changing SCF–CRL complex affinity and acts to promote or weaken associations of repressors with TF targets, then a global view of isomer abundance in planta is important. Inositol kinases and phosphatases are implicated in early responses to changing intracellular Pi availability, and are a critical component of cross-talk with auxin, jasmonate, gibberellin and SA responses, so changes to their enzymatic activity, and resulting flux in InsP and PP–InsP pools need to be quantified. Upon Pi starvation, there was a significant increase in the ratio of InsP7 to InsP8 in shoots, but not in roots [54,70], which would have consequences for the InsP-dependent localised assembly of SCF–CRL complexes and highlights the need for validation across organs. Combined approaches for InsP and PP–InsP quantification, from polyacrylamide gel electrophoresis (PAGE), to [3H] inositol labelling and capillary electrophoresis electrospray ionisation mass spectrometry will provide a more comprehensive understanding [70]. The interactome of putative protein targets for InsPs and PP–InsPs, and additional targets of pyrophosphorylation could be identified from the application of chemically synthesised affinity reagents [52]. Since impaired inositol kinase activity results in the perturbation of multiple pathways, efforts to characterise these effects should acknowledge potential interactions between stressors, rather than attributing causation for these effects on a single nutrient (or hormone) pathway.
Interactions between metabolic messengers and their target proteins across tissues can be characterised with increasingly high resolution by developments in spatial and single cell metabolomics, although sensitivity limitations might currently restrict the detection of InsP or PP–InsP species in a single cell [138,139]. Desorption electrospray ionisation mass spectrometry imaging has recently mapped lipids, metabolites and carbohydrates across root tips to detect variations in TCA cycle intermediates across root meristem and differentiation zone, that did not correlate with changes in ATP levels [140]. Similar techniques could be applied to detect gradients for other metabolic messengers across tissues and development.
Ubiquitin components
There have been relatively few characterisations of SCF components, E3 ligases and their respective targets across nutrient and hormonal pathways, owing to the rapid turnover of target proteins which are often only present at low levels, or for brief periods. As is evident by the ubiquitination of SPX4 by free and complex-bound E3 ligases, characterisation may also be obscured by levels of redundancy. Comprehensive libraries of UPS components, like the recently established ubiquitin E3 ligase-encoding open reading frame library (Ub3-ORFeome) for the rice genome [141], should be extended across plant species, and will assist in the documentation of known interactors and the establishment of a UPS interactome. Emerging techniques for the identification of ubiquitinated target proteins, and their respective ubiquitination sites include the coupling of high-resolution mass spectrometry with antibody immunoaffinity techniques which specifically recognise diglycine remnants responsible for lysine modifications at ubiquitination sites [142]. Ubiquitin variants can halt proteolytic activity, e.g. ubR48, carrying arginine instead of lysine in amino acid position 48 of ubiquitin prevents polyubiquitin chain extension [143], and synthetic ubiquitin suicide probes can be used to label deubiquitinase active sites [144]. E3 ligase trapping methods have recently been used in conjunction with the trypsin-resistant tandem ubiquitin-binding entity method, overcoming obfuscation by the activity of multiple E3 ligases degrading a target protein and capturing ubiquitination by E3 ligases with relatively low activity levels [145].
Assembly
The impact of signalling molecules on the stability of associations between repressor proteins and TFs should be characterised in parallel with their effect on the affinity of SCF components for their target substrates. InsP8 has been implicated in binding to SPX domains and partake in the steric obstruction of AtPHR1/OsPHR2 from DNA binding and induction of PSI genes [57,132], but different InsP species have divergent effects on the assembly of downstream SCF–CRL components [125]. The conservation of InsP and PP–InsP binding domains across SPX proteins, and evidence for both cytosolic as well as nuclear localised OsPHR2 repression by OsSPX4 and OsSPX1/2, respectively, are illustrative of the need for in planta characterisations of binding partners, and their interactions in response to localised fluctuations in metabolic messengers. The proliferation of techniques to identify ligand-receptor interactions will prove especially useful when used in combination with predictive artificial intelligence like AlphaFold and will provide valuable insight into the consequence of physical interactions [146,147]. Protein crystallography has already provided great insight into the assembly of FBX/repressor protein co-receptors in the presence of hormones and InsPs [64,148]. This information can then be used to design signalling molecule mimetics to further study complex regulation [149] both in vivo and in vitro.
The mechanism by which SCF–CRL complexes may be prone to assembly or degradation depending on the presence or absence of substrate has been termed as ‘adaptive exchange’ and provides a framework by which plants can adapt rapidly, and reversibly to their environment [116]. The principles of adaptive exchange dictate that the propensity of a neddylated SCF–CRL complex for the degradation of target proteins is gated by the presence of substrate, and that in the absence of substrate, deneddylated SCF complexes must be constantly recycled via CAND1 mediated exchange [121]. The dependence of this system on the availability of substrate (transcriptional repressor), and its ability to interact with metabolites (InsPs, hormones) for altered complex affinity [125] support its capacity to function as a regulator of environmental stress responses in plants. In human medicine, neddylation inhibitors are becoming increasingly attractive for the development of new antitumour drugs [150], and many of them appear to also be effective in plants [151]. Furthermore, SCF–CRL complexes render themselves to targeted engineering for the improvement of crop productivity and will be particularly beneficial for protected cropping systems which offer greater environmental control. Synthetic biology and drug discovery techniques such as targeted protein degradation with small molecule ‘Degraders’ to manipulate protein levels and post- translationally modify proteins directly rather than indirectly via gene knockout (T-DNA insertion, CRISPR/Cas9) or knockdown (RNAi) have been applied in human therapeutics but have not yet been employed in plants [152,153]. E3 Ligases suitable for proteolysis-targeting chimaera (PROTAC) molecules, which utilise the UPS machinery to degrade target proteins, need to be characterised in plants [152]. Békés et al. [152] suggest that the engineering of small molecules which act analogously to auxin and JA to enhance E3 ligase complex efficiency holds great promise for UPS-mediated genetic engineering for crop species. The progression of these technologies will rely on in-depth and foundational characterisation of the ubiquitin-proteasomal machinery during plant development and in response to abiotic or biotic stress, and across plant cell types, tissues, or organs.
Signal integration for launching a successful stress response
The use of models can further our understanding of plant regulatory and metabolic networks. This approach has been used successfully at different scales to model circadian rhythm [154,155], metabolic flux [156,157], root development [158], hormone signalling [135,159,160], and nutrient homeostasis in plants [161–164]. These models further our mechanistic understanding of a given pathway, interaction between different pathways or metabolic signals, as well as enable new scenarios or parameters to be tested without having to generate plant materials and conduct experiments.
Single-cell organisms such as bacteria, algae or yeast, have become increasingly sophisticated model systems for metabolic engineering and exploration of biosynthetic pathways [165]. Whilst still in its infancy, synthetic biology is becoming more attractive to agriculture, and plant breeding, given that molecular tools to manipulate the genetic code, and develop biosensors or stimulants have become more sophisticated [166,167]. An in-depth understanding of signal-SCF-repressor-TF interaction can help develop synthetic operons and logic gates that can modify root development, light, or nutrient perception, as well as the response to hormones, antibiotics, or other chemicals [168–170].
Conclusions
The discovery of higher order inositol polyphosphates as important signals of cellular P status in higher plants has opened up exciting new avenues for investigating interactions between different signalling networks. The importance of InsP species for SCF complex assembly during hormonal signalling has long been established, but their role in the turnover of transcriptional repressors associated with nutrient sensing and signalling is less well understood. Recent scientific advances demonstrate parallels between nutrient and hormonal signalling pathways and the ability of associated SCF–CRL complexes to integrate signals across pathways. This in turn emphasises the importance of localised signal generation, the exact nature and amplitude of the signal as well as the delicate ‘adaptive exchange’ between SCF–CRL complex subunits required to adjust protein levels of key transcriptional regulators to just the right amount required to achieve the most efficient integrated response.
Open questions and future work
While considerable progress has been made in understanding SPX4 interaction with transcriptional regulators and hormonal signalling, and its ubiquitin-dependent protein turnover, there are still a few outstanding questions:
(1) Is SPX4 a repressor of TFs other than PHR1 — and if so, then is it selective to the MYB-CC class of GARP TFs [59]? The immediate answer would be yes, but it is not selective to MYB–CCs as it has been shown to interact with PAP1, a member of the R2R3–MYB family in Arabidopsis [171], and with OsNLP3 which belongs to the RWP–RK class in rice [105].
(2) Is InsP8 binding required for SPX4 interaction with all its TF targets? Availability of pure InsP species as well as analytical tools to manipulate their synthesis and detect them in planta will help shed more light on this.
(3) Are hormonal repressors such as JAZ, DELLA or AUX/IAA competing with SPX4 for binding those target TFs? Evidence is strengthening that this may well be the case [25].
(4) What stimuli other than low phosphate and nitrate trigger the degradation of SPX4? Is there cross-talk with hormones? Apart from Pi limitation enhancing the synthesis of jasmonate and strigolactones [25,127,172], there is first evidence in rice that strigolactones promote SDEL1-dependent SPX4 degradation [172].
(5) What is the role of FBX mediated signal perception and SCF–CRL complexes in SPX4 ubiquitination? And how does signal perception change over developmental time or in response to abiotic or biotic stress?
(6) Given that E3 ligases other than RBX1/2 have an impact on SPX4 turnover, could recruitment of co-E3 ligases to the CRL be required to initiate SPX4 ubiquitination [173].
(7) What factors stabilise SPX4 under P-replete conditions? Deubiquitination enzymes are yet to be discovered. Exciting new drugs for targeted protein degradation are now also available to plant researchers.
(8) Manipulating turnover of transcriptional repressors has an immediate impact on the transcriptome. Therefore, dynamic network analysis to identify TF hubs and how they change in response to stress, hormones and developmental cues will be essential to shedding light on these highly complex regulatory circuits.
Competing Interests
The authors declare that there are no competing interests associated with the manuscript.
Open Access
Open access for this article was enabled by the participation of La Trobe University in an all-inclusive Read & Publish agreement with Portland Press and the Biochemical Society under a transformative agreement with CAUL.
CRediT Author Contribution
Ricarda Jost: Conceptualization, Resources, Supervision, Funding acquisition, Validation, Investigation, Visualization, Writing — original draft, Project administration, Writing — review and editing. Emma Collins: Conceptualization, Formal analysis, Supervision, Validation, Investigation, Visualization, Writing — original draft, Project administration. Huixia Shou: Writing — review and editing. Chuanzao Mao: Conceptualization, Formal analysis, Investigation, Visualization, Writing — original draft, Writing — review and editing. James Whelan: Conceptualization, Resources, Supervision, Funding acquisition, Project administration, Writing — review and editing.