The ubiquitin-proteasome system regulates plant hormone signaling


  • Aaron Santner,

    1. Molecular Kinetics, Inc., 6201 La Pas Trail, Suite 160, Indianapolis, IN 46268, USA
    Search for more papers by this author
  • Mark Estelle

    Corresponding author
    1. University of California San Diego, Section of Cell and Developmental Biology, 9500 Gilman Drive, La Jolla, CA 9209, USA
      For correspondence (fax +858 534 7108; e-mail
    Search for more papers by this author

For correspondence (fax +858 534 7108; e-mail


Plants utilize the ubiquitin-proteasome system (UPS) to modulate nearly every aspect of growth and development. Ubiquitin is covalently attached to target proteins through the action of three enzymes known as E1, E2, and E3. The ultimate outcome of this post-translational modification depends on the nature of the ubiquitin linkage and the extent of polyubiquitination. In most cases, ubiquitination results in degradation of the target protein in the 26S proteasome. During the last 10 years it has become clear that the UPS plays a prominent regulatory role in hormone biology. E3 ubiquitin ligases in particular actively participate in hormone perception, de-repression of hormone signaling pathways, degradation of hormone specific transcription factors, and regulation of hormone biosynthesis. It is certain that additional functions will be discovered as more of the nearly 1200 potential E3s in plants are elucidated.


Plants utilize a number of mechanisms to modulate protein level. The most prominent of these is regulated protein degradation via the ubiquitin-proteasome system (UPS). In this system a small protein, ubiquitin (Ub), is covalently attached to target proteins and either regulates their function or marks them for destruction by the multisubunit 26S proteasome (Hershko and Ciechanover, 1998). The consequence of this post-translational modification depends on the extent of polyubiquitination and the position of the ubiquitin linkage in the polyubiquitin chain. Early studies in plants made important contributions to our understanding of the UPS. However, it was not until the Arabidopsis genome became available that we understood the remarkable prominence of this pathway in plant cells. Genomic studies predicted that >5% of the Arabidopsis proteome directly participates in the UPS with potentially thousands of additional proteins serving as targets (Vierstra, 2003; Smalle and Vierstra, 2004). Moreover, genetic studies enabled by genome-based programs such as the Arabidopsis 2010 project, revealed that the UPS impacts nearly every aspect of plant growth and development including the cell-cycle, embryogenesis, senescence, defense, environmental responses, and hormone signaling (Vierstra, 2009).

Plant hormones are comprised of a group of structurally unrelated small molecules that regulate a wide variety of plant processes. The hormones also act to integrate diverse environmental cues with endogenous growth programs. So far ten phytohormones have been identified including auxin, abscisic acid (ABA), cytokinin (CK), gibberellin (GA), ethylene, brassinosteroids (BR), jasmonate (JA), salicylic acid (SA), nitric oxide, and strigolactones (Davies, 1995; Browse, 2005; Vert et al., 2005; Grun et al., 2006; Loake and Grant, 2007; Gomez-Roldan et al., 2008; Umehara et al., 2008). Plants also utilize several peptide hormones to regulate various growth responses (Jun et al., 2008). With the application of biochemical, genetic, and genomic approaches, many aspects of hormone biology have been elucidated, especially in the model flowering plant Arabidopsis thaliana. Most hormones are involved in multiple processes and impact each other through elaborate crosstalk strategies. In elucidating these hormone-signaling pathways, it has become clear that the UPS plays an essential role in hormone perception and response (Dharmasiri et al., 2005a,b; Kepinski and Leyser, 2005; Katsir et al., 2008; Melotto et al., 2008; Mockaitis and Estelle, 2008; Santner and Estelle, 2009; Schwechheimer and Willige, 2009). In this review we will describe the central role of the ubiquitin/26S proteasome pathway in hormone signaling by focusing on the critical roles of E3 ubiquitin ligases in hormone perception and signal transduction (Table 1).

Table 1.   Summary of Arabidopsis E3 ligases that mediate specific hormone responses
NameBiological roleKnown substrates
F-box proteins (approximately 700 in Arabidopsis)
 TIR1Auxin perception and signal transductionAUX/IAA proteins
 AFB1Auxin perception and signal transductionAUX/IAA proteins
 AFB2Auxin perception and signal transductionAUX/IAA proteins
 AFB3Auxin perception and signal transductionAUX/IAA proteins
 AFB4? Auxin perception and signal transductionAUX/IAA proteins
 AFB5? Auxin perception and signal transductionAUX/IAA proteins
 COI1JA-Ile perception and signal transductionJAZ proteins
 SLY1GA signal transductionDELLA proteins
 SNEGA signal transductionDELLA proteins
 MAX2Strigolactone signal transduction?
 ETP1Ethylene signal transductionEIN2
 ETP2Ethylene signal transductionEIN2
 EBF1Ethylene signal transductionEIN3, EIL1
 EBF2Ethylene signal transductionEIN3, EIL1
BTB proteins (approximately 80 in Arabidopsis)
 ETO1Regulation of ethylene biosynthesisType-2 ACSs
 EOL1Regulation of ethylene biosynthesisType-2 ACSs
 EOL2Regulation of ethylene biosynthesisType-2 ACSs
RING finger E3s (approximately 465 in Arabidopsis)
 AIP2ABA signal transductionABI3
 KEGABA signal transductionABI5
 SDIR1ABA response?
 RHA2aABA response?
 AtPUB9ABA response?
 SAUL1 (PUB44)Regulation of ABA biosynthesisAAO3

The ubiquitin proteasome system

Ubiquitin is an evolutionarily conserved 76-amino acid protein that is attached to specific target proteins via the sequential action of three enzymes (Hershko and Ciechanover, 1998). Initially, the E1 or Ub-activating enzyme hydrolyzes ATP to form an E1–Ub intermediate in which the C-terminal glycine of Ub is linked through a thiolester bond to the E1 (Figure 1). Activated Ub is then transferred to a cysteinyl residue on the ubiquitin conjugating enzyme (E2) (Figure 1). The E2–Ub intermediate can directly conjugate ubiquitin to a lysine residue in the substrate protein by first binding to an E3 ubiquitin ligase, or in the case of HECT (homology to E6-AP C terminus)-domain E3s, transfer the activated Ub to the E3, which then transfers it to the substrate (Pickart, 2001). In both scenarios it is the E3 that confers substrate specificity to the pathway and correctly positions the substrate protein for ubiquitin conjugation. Attachment of a single ubiquitin to a substrate protein can modify the localization or activity of the protein (Mukhopadhyay and Riezman, 2007). However, when the process is reiterated resulting in a polyubiquitin chain the substrate protein is more often degraded by the 26S proteasome (Figure 1).

Figure 1.

 The ubiquitin-conjugation pathway begins with Ub activation by the E1. Activated Ub is transferred to the E2 ubiquitin-conjugating enzyme. The E3 ubiquitin ligase binds both the protein substrate and E2–Ub to facilitate the transfer of Ub to the substrate. Polyubiquitination often targets a substrate to the 26S proteasome where it is degraded.

The 26S proteasome is a large, multi-subunit, ATP-dependent protease that unfolds and degrades polyubiquitinated proteins (Voges et al., 1999). It is assembled from a 20S core protease that is cylindrical in shape and capped at each end by a 19S regulatory particle. Each regulatory particle is composed of a lid and a base. The lid is assembled from at least eight subunits that function together to remove the ubiquitin chain from the substrate protein (Fu et al., 2001). The 19S regulatory particle also contains a base component consisting of several subunits that both recognize ubiquitinated substrates and unfold the substrate protein. The unfolded protein is then fed into the central chamber of the 20S core protease. The central chamber is formed by two outer α-subunit rings and two central proteolytic β-subunit rings. Once inside, the unfolded protein is broken down into short peptides (Fu et al., 2001; Yang et al., 2004).

In 2004 the Nobel Prize in Chemistry was awarded to Aaron Ciechanover, Avram Hershko and Irwin Rose for their pioneering biochemical studies utilizing a reticulocyte lysate expression system to discover and characterize ubiquitin and the enzyme activities required to conjugate it to substrates as described above (Wilkinson, 2005). While the importance of this pathway was not fully realized at the time of its discovery, subsequent work has demonstrated how critical the UPS is to eukaryote biology. Genomic studies made possible by the availability of genome sequences for Arabidopsis and other plants have revealed that this system is particularly important in plants. Several groups have independently identified components of the pathway in Arabidopsis. Interestingly, >1300 proteins or approximately 5% of the Arabidopsis proteome has been identified as playing a role in the UPS compared to approximately 150 proteins in Saccharomyces cerevisiae or approximately 2.5% of the proteome (Vierstra, 2003, 2009). Approximately 1200 or 90% of these genes encode components of E3 ubiquitin ligases suggesting that hundreds or even thousands of proteins are regulated by the UPS (Smalle and Vierstra, 2004). The large expansion of E3 ubiquitin ligases indicates that regulated protein turnover has been co-opted to modulate many processes in plants.

Diversity among plant E3 ubiquitin ligases

The E3 ubiquitin ligases comprise a diverse family of proteins or protein complexes that can be distinguished based on the type of interaction domain (RING domain, U-box domain, or HECT domain) used to bind E2 enzymes and whether they act as single subunits or multisubunit complexes (Moon et al., 2004). In the Arabidopsis UPS, the most abundant E2 interaction domain is found in the approximately 465 RING (Really Interesting New Gene) proteins that are characterized by a approximately 70 amino acid motif known as a RING finger (Freemont, 2000). The RING finger is a zinc-binding motif that binds to the E2 Ub-conjugating enzyme during the ubiquitin conjugation cascade. These proteins can function as single subunit E3s or participate as part of a multisubunit E3 complex. U-box E3 ubiquitin ligases comprise a much smaller subfamily consisting of approximately 64 predicted members and are characterized by a approximately 70 amino acid U-box domain (Yee and Goring, 2009). The U-box motif was revealed to be a modified RING-finger domain. While structurally similar to the RING motif, the U-box does not use zinc ions to stabilize its secondary structure. The smallest E3 subfamily in Arabidopsis is the HECT (Homology to E6-AP C Terminus) domain proteins that consist of approximately 20 members (Vierstra, 2003). The HECT domain is a 350 amino acid motif that contains both an ubiquitin-binding site and an E2-binding site (Pickart, 2001). Unlike RING and U-box E3s that indirectly mediate ubiquitin ligation by docking the E2 enzyme and substrate protein appropriately for Ub transfer from the E2 to the substrate, HECT E3s accept activated Ub from the E2 and then transfer it directly to the substrate protein (Figure 2).

Figure 2.

 E3 Ubiquitin ligases in plants. Single subunit E3 are divided into three classes based on the E2 interaction domains. Substrates are green, substrate specificity factors are purple and the E2 interaction domains are yellow. The HECT-type E3s are unique in that they accept Ub before transferring it to the substrate. Three multi-subunit Es are presented. Cullin is pink, RBX1 is yellow, and the substrate specificity factors are purple. The SCF and CUL4-DDB1 also have adapter subunits that are light green.

In plants there are at least three types of multisubunit RING E3s that are comprised of a RING protein RBX1 (RING-Box 1), a cullin scaffold-like protein, and an additional substrate-recognition protein that can bind directly to the cullin or indirectly via adapters proteins (Hershko and Ciechanover, 1998; Moon et al., 2004; Chen et al., 2006). The type of cullin subunit determines which recognition protein is incorporated into the complex. Therefore, cullin-RING multisubunit E3 ubiquitin ligases are categorized depending on the substrate recognition subunit-cullin combination (Figure 2). The three broad categories are SCF (Skp-Cullin-F-box), CUL3-BTB (Broad-complex, Tramtrack, Bric-a-Brac), and CUL4-DDB1 (DNA-DAMAGE BINDING 1). In these complexes, the cullin serves as the scaffold for the protein complex. In Arabidopsis, five canonical cullin proteins (CUL1, CUL2, CUL3A, CUL3B, and CUL4) have been shown to be components of E3 ligase complexes. In Arabidopsis, CUL1 and potentially CUL2 are subunits of SCF complexes (Gray et al., 1999; Risseeuw et al., 2003). CUL3A and CUL3B interact directly with substrate recognition proteins that contain a BTB/POZ domain to form the CUL3-BTB ubiquitin ligases (Figueroa et al., 2005; Gingerich et al., 2005; Weber et al., 2005). CUL4-based ligases utilize DDB1 as an adapter to incorporate WD40-domain containing substrate recognition proteins DWD (DDB1-BINDING WD40 Protein) or DCAF (DDB1 and CUL4-ASSOCIATED Factor) into the complex (Bernhardt et al., 2006). A fourth multisubunit E3 ubiquitin ligase named APC (Anaphase Promoting Complex) is also present in Arabidopsis. The APC consists of 11 subunits. Of these subunits, APC2 shares homology to cullin and APC11 is a RING protein (Capron et al., 2003).

SCF complexes

Among the multisubunit ubiquitin ligases, the SCF group is the most abundant and best characterized. The SCF name is derived from three of its four subunits: SKP1 (ASK in plants), Cullin, and the F-box protein. The fourth subunit is the RING finger protein RBX1 (RING-Box 1) (Hershko and Ciechanover, 1998). In these complexes CUL1 serves as the backbone of the complex with RBX1 binding at the C-terminus and an ASK adapter protein at the N-terminus (Zheng et al., 2002b). The ASK protein also binds to the F-box motif of F-box proteins to form the complete complex (Figure 2). Typically the F-box motif is found at the N-terminus of the protein while a variety of protein–protein interaction domains including but not limited to Leu-rich repeat, KELCH, WD-40, and Arm domains comprise the remainder of the protein and facilitate substrate interactions (Gagne et al., 2002). A recent study comparing all F-box genes from Arabidopsis (692), poplar (337), and rice (779) determined that the F-box superfamily can be divided into 42 families based on domain organization (Xu et al., 2009). Some of the families are highly conserved suggesting that their substrates are the same or similar. Other families have diverged more extensively presumably yielding family members with functional differences (Xu et al., 2009).

SCF complex assembly and hormone signaling

Most hormone signaling proteins identified in plants were first discovered through mutant screens and many of the hormone-resistant mutants identified in this way are disrupted in components of the SCF (Skp1/Cullin/F-box) ubiquitin ligases (E3) or in proteins that regulate SCF activity (Dharmasiri and Estelle, 2002). For example, several cul1/axr6 mutants have been isolated that provide insights into SCF complex function, assembly, and regulation. A series of mutants that disrupts AtCUL1 has been particularly insightful. The first CUL1 mutants, originally named axr6, were identified in a genetic screen for auxin-resistant seedlings (Hobbie et al., 2000; Hellmann et al., 2003). In addition to auxin resistance, the original axr6 mutants exhibited a range of defects in lateral organ initiation, meristem organization, and other processes. In addition null cul1 mutants arrest very early in embryogenesis (Hobbie et al., 2000; Shen et al., 2002; Hellmann et al., 2003). Further characterization of CUL1 has been aided by the discovery of several weak alleles that are viable and fertile as homozygotes. One of these, cul1-7, was identified in a genetic screen for mutations that interfere with the degradation of a known SCFTIR1 substrate reporter protein. The cul1-7 mutation affects the C-terminus of CUL1 thereby interfering with RBX1 interaction. Phenotypic characterization of cul1-7 mutants demonstrated altered auxin responses, gibberellin responses, and light responses (Gilkerson et al., 2009). Two point mutants named axr6-1 and axr6-2 were isolated in a screen for auxin-resistance but were also shown to confer JA-resistance (Ren et al., 2005). In a separate screen for mutations that enhanced the tir1-1 auxin resistance, axr6-3 was found (Quint et al., 2005). Further characterization revealed that the axr6-1, axr6-2, and axr6-3 phenotypes were the result of changes in CUL1 that interfere with active SCF assembly. As a result, seedlings harboring these mutations exhibit defects in auxin response, JA response, flower development, and photomorphogenesis highlighting the importance of CUL1 in a broad range of processes (Quint et al., 2005; Ren et al., 2005; Gilkerson et al., 2009).

Cullin-based E3 complexes are themselves regulated by the attachment or removal of a small ubiquitin-like protein called RUB (Related to Ubiquitin) to the cullin subunit of the complex (Hotton and Callis, 2008). In addition, a 120-kDa HEAT repeat protein called CAND1 (Cullin Associated and Neddylation Dissociated) can also bind to cullin (Goldenberg et al., 2004). Another CUL1 mutant allele, cul1-6, causes various defects in morphology and is hyposensitive to ethylene and various light conditions. The cul1-6 protein product is disrupted in its ability to be modified by RUB and interact with CAND1 (Moon et al., 2007).

Regulation of SCF complexes

The RUB (Nedd8 in animals) family of proteins is highly conserved in eukaryotes and is essential in most organisms including plants. In Arabidopsis, loss of RUB1 or RUB2 does not result in an obvious phenotype (Bostick et al., 2004). However, rub1rub2 double mutants are embryo lethal highlighting the importance of RUB conjugation to plant viability (Bostick et al., 2004). Like the ubiquitin pathway, RUB conjugation is dependent on three enzyme activities. In Arabidopsis, an AXR1/ECR1 dimer serves as the RUB E1 (del Pozo et al., 2002). The RUB E2 is known as RCE1 (RUB Conjugating Enzyme 1) (Dharmasiri et al., 2003b). Interestingly, overexpression of the RBX1 subunit of SCF E3 ligases promotes the attachment of RUB to a specific lysine in the cullin leading the speculation that RBX1 may serve as the RUB E3 ligase (Gray et al., 2002; Dharmasiri et al., 2003b). However, a protein called DCN-1 or Dcn1p was determined to be the RUB E3 ligase in C. elegans and S. cerevisiae respectively (Kurz et al., 2005). There are at least three DCN1-like proteins in Arabidopsis and genetic studies implicate one of these, AAR3, in auxin response regulation (Biswas et al., 2007). While much of the work elucidating this pathway in plants has focused on RUB modification of the CUL1 subunit of SCFs, CUL3a, CUL3b, and CUL4 are also known to be modified by RUB (Dharmasiri et al., 2003b; Figueroa et al., 2005; Gingerich et al., 2005; Chen et al., 2006).

The RUB/Nedd8 conjugation pathway was first identified through genetic studies of auxin response in Arabidopsis (Lincoln et al., 1990). Loss of AXR1 results in a variety of hormone-related phenotypes including reduced sensitivity to auxin, cytokinin, ethylene, epi-brassinolide, and JA in root elongation assays (Leyser et al., 1993; Timpte et al., 1995; Schwechheimer et al., 2002; Tiryaki and Staswick, 2002). In one study, axr1-24 mutant seed was found to be more resistant to the inhibitory effects of abscisic acid during germination than wild-type (Tiryaki and Staswick, 2002). Older axr1 plants exhibit reduced apical dominance and fertility (Lincoln et al., 1990). A double mutant containing axr1 and its closest relative axr1-like (axl1) dies early in development (Dharmasiri et al., 2007). AXR1 is only half of the RUB E1 and must dimerize with ECR1 for activity (del Pozo et al., 1998). Overexpression of a mutant ECR1 protein that is unable to contribute to E1 activity results in a dominant-negative phenotype that resembles axr1 mutants (del Pozo et al., 2002). Furthermore, these transgenic lines have less RUB-modified CUL1 consistent with the role of ECR1 in the RUB-conjugation pathway. The significance of the RUB1 pathway has also been confirmed by studies of the RUB E2 enzyme. A recessive mutation in the promoter of the RCE1 gene causes a substantial reduction in expression. The rce1-1 plants are auxin resistant and bushy similar to axr1 mutant plants. Moreover, RUB-modified CUL1 levels are reduced and the levels of IAA7, a substrate of SCFTIR1, are increased in rce1-1 plants.

The COP9 signalsome (CSN) was originally identified as a repressor of photomorphogenesis but is now known to have a broader role in plant development including catalyzing the removal of RUB from cullins (Wei et al., 2008). The CSN is a multi protein complex with eight subunits that share similarities to the 19S regulatory particle of the 26S proteasome (Serino and Deng, 2003). Loss of a single CSN subunit can destabilize the entire complex resulting in a broad range of phenotypes including lethality in strong csn alleles (Serino and Deng, 2003). Moreover, loss of the CSN results in a shift to completely rubylated CUL1, CUL3 and CUL4 in protein extracts (Gusmaroli et al., 2007). CSN5, a zinc metalloprotease, is the specific subunit known to be responsible for the removal of RUB from CUL1 (Cope et al., 2002). Support for this finding comes from a csn5 antisense line that has an axr1-like morphology but increased levels of RUB-conjugated CUL1 (Schwechheimer et al., 2002). The fact that increased and decreased RUB-modification results in similar phenotypes suggests that RUB conjugation and deconjugation are both required for E3 activity.

CAND1 is another protein that binds cullin proteins. In Arabidopsis, cand1 mutants were identified in genetic screens for sirtinol resistance and as an enhancer of tir1 resistance (Cheng et al., 2004; Chuang et al., 2004). The cand1 plants also have a pleiotropic phenotype with altered responses to auxin and GA (Chuang et al., 2004). CAND1 preferentially binds to unmodified CUL1 preventing the formation of an active SCF (Zheng et al., 2002a; Feng et al., 2004; Zhang et al., 2008). Based on these findings, it is believed that the interactions between CAND1 binding and rubylation and de-rubylation by CSN regulate the pool of active cullin-based ligases to maintain homeostasis under various conditions (Zhang et al., 2008).

In all of the examples listed above, mutation of an SCF subunit or a protein known to be involved in SCF assembly or regulation results in a highly pleiotropic phenotype that is indicative of a role for SCFs in many developmental processes. However, the identification of F-box gene mutations has allowed individual SCF complexes to be assigned functions in specific hormone signaling pathways because the F-box protein determines substrate specificity.

SCFTIR1/AFB and auxin perception

SCFs were first implicated in auxin signaling with the identification of an F-box protein called TIR1. It was later determined that TIR1 belongs to a family of F-box proteins that includes five additional auxin-signaling F-Box proteins (AFBs) but most of the research to date has focused on TIR1. Recessive mutations in TIR1 confer auxin resistance implying that the protein is required for degradation of negative regulators of auxin response (Ruegger et al., 1998). A key event in the characterization of the auxin-signaling pathway was the discovery that members of the Aux/IAA protein family are substrates of SCFTIR1 (Gray et al., 2001). The Aux/IAA proteins are short lived and their degradation is promoted by auxin. Many gain-of-function mutations in Aux/IAA genes have been isolated and in every case the mutations affect residues within a highly conserved region called domain II (Reed, 2001). Biochemical studies demonstrated that Aux/IAA domain II binds TIR1 and that this binding is enhanced by auxin (Gray et al., 2001; Dharmasiri et al., 2003a; Kepinski and Leyser, 2004).

The Aux/IAA proteins control the auxin transcriptional response through their interaction with a second large family of transcription factors called Auxin Response Factors (ARFs). ARFs bind to cis elements in the promoters of auxin-responsive genes and either activate or inhibit gene transcription depending on the type of ARF (Guilfoyle and Hagen, 2007). Aux/IAAs bind to ARFs through shared domains in both proteins called domains III and IV and thereby repress ARF activity (Reed, 2001).

Domain I of most Aux/IAAs contains an ERF-associated amphiphilic repression (EAR) motif that is required for transcription repression (Tiwari et al., 2004). Recently, a protein called TOPLESS (TPL) was shown to associate with domain I of the Aux/IAA protein IAA12 and to function as a transcriptional co-repressor (Szemenyei et al., 2008). These findings suggest that Aux/IAA proteins act as repressors of ARF-mediated transcription by recruiting TPL or related transcriptional co-repressors to the multi-protein complex (Szemenyei et al., 2008). Auxin de-represses transcription by promoting ubiquitination of Aux/IAA proteins through the action of SCFTIR1, thus targeting them for degradation by the 26S proteasome. Without the Aux/IAA proteins, TPL is no longer held in close proximity to ARFs and auxin-regulated genes are de-repressed.

Although a model for auxin-dependent de-repression of transcription was in place, the important question of how auxin promotes the SCFTIR1-Aux/IAA interaction remained. Often substrate-SCF recognition requires phosphorylation of the substrate. However, several studies suggested that this was not true for the SCFTIR1-Aux/IAA interaction. Rather, TIR1 itself was shown to directly and specifically bind biologically active auxins (Dharmasiri et al., 2005a; Kepinski and Leyser, 2005). Structural studies of TIR1 in the presence of auxin and a peptide encompassing domain II revealed that a single pocket formed by the LRR domain of TIR1 binds both auxin and the domain II peptide (Figure 3a; Tan et al., 2007). Auxin rests at the base of this pocket and stabilizes the binding of the domain II peptide to TIR1 (Tan et al., 2007). One important implication of the structure is that both TIR1 and the Aux/IAAs appear to contribute to high-affinity binding of auxin implying that different combinations of TIR1 and Aux/IAAs may have unique auxin-binding characteristics.

Figure 3.

 Function of E3s in plant hormone biology.
(a) Hormone perception. The TIR1 F-box protein binds auxin to promote auxin-regulated transcription through the Ub-mediated degradation of Aux/IAA transcriptional regulators.
(b) Signal de-repression. SCFSLY1 targets DELLA repressor proteins for degradation in response to GA. This action frees the PIF3/4 transcription factors to regulate expression. SCFTIR1 and SCFCOI1 also act de-repress transcription.
(c) Direct regulation of transcription factor levels. SCFEBF1 acts to keep EIN3 levels low in the absence of ethylene.
(d) Regulation of hormone biosynthesis. CUL3-BTBETO1 targets type-2 ACC synthases for degradation to limit ethylene production.

Auxin research has a long history and the discovery that SCFTIR1 functions directly as a receptor was groundbreaking (Figure 3a). Further work with F-box proteins auxin-signaling F-box proteins 1–5 (AFB) and Coronatine Insensitive 1 (COI1) indicates that additional F-box proteins can also function as receptors for small molecules. This discovery presents a new avenue for the development of drugs that target the ubiquitin-proteasome pathway (Tan and Zheng, 2009).

SCFCOI1 and jasmonate perception

Jasmonic acid (JA) and its metabolites, collectively known as jasmonates, are important plant signaling molecules that mediate stress responses and aspects of growth and development (Wasternack, 2007). One of the mutants that helped define the role of JA in plant growth is the Arabidopsis coronatine-insensitive1 (coi1) mutant (Feys et al., 1994). The coi1 mutant is perturbed in every aspect of jasmonate response indicating an essential role for COI1 in JA signal transduction (Feys et al., 1994). COI1 was determined to be an F-box protein and shown to associate with components of SCF complexes including ASK1, CUL1, and RBX1. Based on these observations it was suggested that SCFCOI1 targets repressors of jasmonate-mediated transcription.

Recently, a family of transcriptional regulators called jasmonate ZIM-domain (JAZ) proteins was identified and characterized (Chini et al., 2007; Thines et al., 2007; Yan et al., 2007). JAZ proteins were shown to be degraded in a proteasome-dependent manner following jasmonate treatment, but were stabilized in the coi1-1 background. This suggested that SCFCOI1 was responsible for ubiquitinating the JAZ proteins. In support of this model, members of the JAZ family were shown to interact with COI1 both in vitro and in yeast two-hybrid assays (Chini et al., 2007; Thines et al., 2007). The interaction between COI1 and JAZ proteins are enhanced in the presence of various jasmonates with JA-isoleucine being the most effective. In addition, coronatine was shown to bind to COI1–JAZ complexes with high affinity (Melotto et al., 2008). An independent study used a variety of predictive and biochemical approaches to clearly demonstrate that COI1 does in fact directly bind JA-Ile and coronatine (Yan et al., 2009). Taken together, the data suggest that SCFCOI1 serves as a receptor for JA-Ile/coronatine and targets JAZ proteins for degradation upon hormone binding.

JAZ proteins directly interact with MYC2, a well characterized transcription factor that modulates JA-mediated transcription (Chini et al., 2007). Therefore, SCFCOI1–JA-Ile-mediated degradation of JAZ proteins permits MYC2 to activate or repress downstream target genes in JA signaling cascades. Several of the JAZ genes are themselves up-regulated, indicating that a negative feedback mechanism limits the response after jasmonate perception (Chini et al., 2007).

SCFSLY1/GID2 and gibberellin signaling

The GAs are a large family of growth regulators that exert influence over diverse growth processes including seed development, organ elongation, and the control of flowering time (Yamaguchi, 2008). The GA receptor, Gibberellin-Insensitive Dwarf 1 (GID1), was first identified in a genetic screen for GA signaling mutants in rice (Oryza sativa) (Ueguchi-Tanaka et al., 2005). Three orthologous genes GID1a, GID1b, and GID1c were identified in Arabidopsis (Griffiths et al., 2006; Nakajima et al., 2006). Genetic studies indicate that all GA responses require functional GID1 proteins in both rice and Arabidopsis (Griffiths et al., 2006). Structural studies indicate that the GID1 protein forms a deep binding pocket whose access is controlled by an N-terminal lid (Murase et al., 2008; Shimada et al., 2008). It is speculated that GA binding in the pocket induces the N-terminal lid to fold back over the GA binding pocket (Murase et al., 2008; Shimada et al., 2008). This compact form of GID1 then interacts with DELLA proteins through a surface of the lid and conserved motifs in the DELLA domain (DELLA, VHYNP, and LExLE) (Ueguchi-Tanaka et al., 2005; Griffiths et al., 2006; Willige et al., 2007; Murase et al., 2008).

Mutations within the DELLA domain of DELLA proteins result in a dominant gain-of-function gibberellin-insensitive phenotype. The DELLAs were also found to accumulate in Arabidopsis and rice mutants defective in the F-box protein genes, sleepy1 (sly1) and gibberellin-insensitive dwarf2 (gid2) respectively (McGinnis et al., 2003; Sasaki et al., 2003). The sly1 and gid2 mutants are also GA-insensitive indicating that DELLA proteins are negative regulators of GA-responses. Taken together, these observations suggest that SCFSLY1/GID2 is responsible for regulating DELLA turnover and that the DELLA domain is important for this regulation. In the absence of SLY1, a homologous F-box protein named SNEEZY (SNE) can regulate DELLA protein stability (Strader et al., 2004). Interestingly, the interaction between DELLA proteins and the F-box proteins SLY1/GID2 was enhanced in the presence of GA-bound GID1 (Griffiths et al., 2006). These data suggest that DELLAs are better able to interact with SCFGID2/SLY while in a complex with gibberellin-bound GID1. This interaction ultimately leads to ubiquitination and degradation of the DELLA repressor, thus promoting GA-mediated transcription.

Recent work exploring the integration of light and GA signals has elucidated a model for DELLA-mediated growth regulation. DELLA proteins were shown to directly interact with the DNA-binding domain of two bHLH transcription factors (PIF3 and PIF4) sequestering them in inactive complexes (Feng et al., 2008; de Lucas et al., 2008). GA accumulation destabilizes DELLAs, freeing PIF3 and PIF4 to activate the transcription of their target genes (Figure 3b; de Lucas et al., 2008; Feng et al., 2008). PIF3 and PIF4 are members of a larger subfamily of bHLH proteins that have similar DNA-binding domains. It seems likely that DELLAs regulate the activity of several other transcription factors by this mechanism. Thus, as observed in auxin and JA signaling, GA appears to regulate the abundance of a transcriptional repressor family by promoting ubiquitination through the activity of SCF-type ligases.

SCFMAX2 and strigolactone signaling

It has long been known that auxin synthesized in the apex inhibits the growth of lateral branches. Mutant studies in pea, rice, Arabidopsis, and petunia indicated that axillary bud outgrowth is also inhibited by a separate hormone that originates in the root (Ongaro and Leyser, 2008). Molecular analysis of branching mutants revealed that several of the affected genes encode CAROTENOID CLEAVAGE DIOXYGENASE 7 (CCD7) or CCD8 suggesting that the hormone is a carotenoid or related compound (Ongaro and Leyser, 2008). Strigolactone, a carotenoid derivative, was ultimately shown to be the mysterious branching hormone (Gomez-Roldan et al., 2008; Umehara et al., 2008). The mechanism of strigolactone signaling is unclear but an F-box protein called MAX2/RMS4 is required for the response (Stirnberg et al., 2002; Johnson et al., 2006). This finding strongly suggests that protein degradation is an important component in the strigolactone response pathway.

SCFs and ethylene signaling

Ethylene is a gaseous hormone that regulates several plant processes including hypocotyl and root elongation, apical hook formation, and fruit ripening (Kendrick and Chang, 2008). Ethylene is perceived through five receptors that are related to bacterial two-component histidine kinases (Kendrick and Chang, 2008). In the absence of ethylene the receptors activate CONSTITUTIVE TRIPLE RESPONSE 1 (CTR1), a negative regulator of ethylene signaling. Ethylene perception by the receptors blocks CTR1 activation and allows ETHYLENE INSENSITIVE 2 (EIN2) to positively regulate ethylene signaling through a so far unidentified mechanism (Kendrick and Chang, 2008). EIN2 levels are regulated by two F-box proteins, called EIN2 TARGETING PROTEIN 1 (ETP1) and ETP2 that have been shown to promote degradation of EIN2 in the absence of ethylene (Qiao et al., 2009). Conversely, the expression of ETP1/2 is reduced in the presence of ethylene allowing the accumulation of EIN2. Ethylene signaling is also dependent on two positive transcriptional regulators known as ETHYLENE INSENSITIVE 3 (EIN3) and EIN3-Like (EIL1) that control the transcription of downstream genes (Solano et al., 1998). EIN3 and EIL1 levels are also regulated through the action of at least two related F-box proteins, EIN3-Binding F-box 1 (EBF1) and EBF2 (Figure 3c; Guo and Ecker, 2003; Potuschak et al., 2003). Thus SCF complexes are involved in ethylene signal pathways at multiple points.

CUL3-BTBETO1/EOL1/EOL2 and ethylene biosynthesis

In addition to SCF complexes, many other types of E3 ubiquitin ligases have been directly linked to plant hormone biology. Take for example the role of three CUL3-BTB E3 ubiquitin ligases in regulating ethylene biosynthesis. The rate-limiting step of ethylene biosynthesis from methionine to 1-aminocyclopropane-1-carbolxylic acid (ACC) is catalyzed by a family of ACC synthases (Tsuchisaka et al., 2009). The ACS family has been divided into three groups based on distinct sequences in their C-termini that mediate their stability (Yoshida et al., 2005). In a genetic screen for mutants that overproduce ethylene, two dominant mutants, eto2 and eto3, were identified that had changes in the C-terminal extension of type-2 ACSs (ACS5 and ACS9 respectively) (Woeste et al., 1999; Chae et al., 2003). The mutations markedly increased protein stability thus leading to increased ethylene production. Another important mutant that came from the screen was eto1. ETHYLENE OVERPRODUCER1 (ETO1) was found to be a member of the BTB protein superfamily that interacts with CUL3a and CUL3b (Wang et al., 2004). ETO1 also directly interacts with type-2 ACSs through their C-termini suggesting a role for CUL3-BTBETO1 in targeting type-2 ACSs for degradation in the proteasome (Figure 3d; Yoshida et al., 2005, 2006). The Arabidopsis genome sequence revealed two proteins with significant sequence similarity to ETO1 now called ETO-LIKE 1 (EOL1) and EOL2. Like ETO1, EOL1 has been shown to directly interact with type-2 but not type1 and type-3 ACSs (Yoshida et al., 2006). Analyses of eol1 single mutants and eol1eol2 double mutants do not demonstrate an obvious ethylene-overproduction phenotype. However, both mutations exaggerate the phenotype of eto1 seedlings indicating that ETO1, EOL1, and EOL2 collectively regulate ethylene production by controlling the levels of type-2 ASCs in plants (Christians et al., 2009).

RING-E3s and ABA signaling

ABA is an isoprenoid compound that impacts a number of growth processes including the control of seed dormancy and drought response (Nambara and Marion-Poll, 2005). Although several potential ABA receptors have been described some of these are quite controversial (Santner and Estelle, 2009). At this point the most promising candidate ABA receptors are a pair of G-protein coupled receptors, GTG1 (GPCR-type G protein 1) and GTG2 (Pandey et al., 2009) and a family of START proteins called PYR/PYL/RCAR (Ma et al., 2009; Park et al., 2009). The intermediate signaling steps between ABA perception and response are complex and involve a variety of kinases, phosphatases, and transcription factors (Hirayama and Shinozaki, 2007). Not surprisingly, several ubiquitin ligases have been linked to ABA responses. Two RING E3 ligases, ABI3-Interacting Protein (AIP2) and Keep on Going (KEG), promote normal ABA signaling by regulating the abundance of ABA-responsive transcription factors, namely ABA-Insensitive 3 (ABI3) and ABA-Insensitive 5 (ABI5) (Zhang et al., 2005; Stone et al., 2006). Evidence suggests that ABA increases AIP2 expression thereby increasing the ubiquitination and degradation of ABI3 (Zhang et al., 2005). Conversely, ABA protects ABI5 from ubiquitin-mediated degradation by preventing either the recognition or the ubiquitination of ABI5 by KEG (Stone et al., 2006). keg seedlings have a severe phenotype suggesting a broader role for KEG in ABA signaling that may also extend to other ABA-responsive transcription factors (Stone et al., 2006). The SALT- and DROUGHT-INDUCED RING FINGER1 (SDIR1) is also known to positively regulate ABA signaling (Zhang et al., 2007). SDIR1 overexpression leads to ABA hypersensitivity and a variety of ABA-linked phenotypes that correlate with altered expression of known ABA genes (Zhang et al., 2007). Another RING E3 ligase, RHA2a, has been recently reported to regulate ABA signaling during seed germination and early development but does so independently of the ABI3, ABI4, and ABI5 transcription factors (Bu et al., 2009).

U-Box E3s and hormone responses

There are 64 predicted U-box in the Arabidopsis genome and at least two, AtPUB9 and AtPUB44, have now been linked to specific hormone responses. AtPUB9 is responsive to ABA treatment at the transcript level (Samuel et al., 2008). Subsequently, the AtPUB9 protein was found to move from the nucleus to the plasma membrane in BY-2 cells following ABA treatment (Samuel et al., 2008). It is not yet clear what the biological relevance of AtPUB9 translocation is but it is noteworthy that pub9 mutant lines are hypersensitive to ABA in seed germination assays indicating that AtPUB9 has a role in this process (Samuel et al., 2008).

AtPUB44 is also known as SENESCENCE-ASSOCIATED E3 UBIQUITIN LIGASE 1 (SAUL1) and normally prevents premature senescence. The saul1 mutants exhibit enhanced ABA biosynthesis coinciding with accumulation of Arabidopsis aldehyde oxidase 3 (AAO3) (Raab et al., 2009). A direct interaction between SAUL1 and AAO3 was demonstrated and suggests that SAUL1 targets AAO3 for degradation in order to keep AAO3 levels low and thus ABA levels low until senescence (Raab et al., 2009). Another U-box protein that has been linked to hormone responses is potato PHOR1. PHOR1 was first identified based on its photoperiod-dependent expression pattern (Amador et al., 2001). When PHOR1 expression was knocked down, semi-dwarf plants with higher endogenous GA levels were recovered. Conversely, overexpression of PHOR1 resulted in hypersensitivity to exogenous GA application and resistance to the effects of paclobutrazol and GA biosynthesis inhibitor (Amador et al., 2001).

Future perspectives

During the last 10 years dramatic advances in our understanding of plant hormone signaling have gone hand in hand with a growing appreciation for the importance of the UPS in cellular regulation. It is also important to emphasize that research on plant hormone signaling has had important implications for biomedical research. For example, the RUB/Nedd8 E1 enzyme was first identified in Arabidopsis (Lincoln et al., 1990; Leyser et al., 1993). This enzyme is highly conserved among eukaryotes and recently, inhibitors of the RUB/Nedd8 E1 were shown to have promise as anticancer drugs (Soucy et al., 2009). Further, the TIR1/AFB and COI1 F-box proteins represent a new class of receptors in which hormone perception is directly linked to ubiquitin ligase activity. It seems possible that additional examples of this mechanism will be uncovered as more E3 ubiquitin ligases are characterized. In addition, the discovery that a small molecule promotes E3 binding to its substrate suggests that it may be possible to identify pharmaceuticals that promote degradation of E3 substrates implicated in disease.

During the next decade, we can expect exciting progress in several directions. We can certainly anticipate new insight into the molecular mechanism of E3 ligase function in the context of hormone signaling. In addition, an increasing focus on how genetic and environmental information is integrated will require that we understand how components of the UPS contribute to dynamic and coordinated growth responses. Furthermore, only a small fraction of the E3s encoded in plant genomes have been characterized. Using the powerful resources that are increasingly available to plant biologists, we expect that many of the E3s with unknown function will be found to contribute to growth and development through the mediation of hormone responses.


Work in the authors’ laboratory is supported by grants from the NIH (GM43644), DOE (DE-FG02–02ER15312), and NSF (IOS 0744800).