Auxin-induced, SCF^TIR1-mediated poly-ubiquitination marks AUX/IAA proteins for degradation

Summary

The plant hormone auxin (indole-3-acetic acid or IAA) regulates plant development by inducing rapid cellular responses and changes in gene expression. Auxin promotes the degradation of Aux/IAA transcriptional repressors, thereby allowing auxin response factors (ARFs) to activate the transcription of auxin-responsive genes. Auxin enhances the binding of Aux/IAA proteins to the receptor TIR1, which is an F-box protein that is part of the E3 ubiquitin ligase complex SCF^TIR1. Binding of Aux/IAA proteins leads to degradation via the 26S proteasome, but evidence for SCF^TIR1-mediated poly-ubiquitination of Aux/IAA proteins is lacking. Here we used an Arabidopsis cell suspension-based protoplast system to find evidence for SCF^TIR1-mediated ubiquitination of the Aux/IAA proteins SHY2/IAA3 and BDL/IAA12. Each of these proteins showed a distinct abundance and repressor activity when expressed in this cell system. Moreover, the amount of endogenous TIR1 protein appeared to be rate-limiting for a proper auxin response measured by the co-transfected DR5::GUS reporter construct. Co-transfection with 35S::TIR1 led to auxin-dependent degradation, and excess of 35S::TIR1 even led to degradation of Aux/IAAs in the absence of auxin treatment. Expression of the mutant tir1-1 protein or the related F-box protein COI1, which is involved in jasmonate signaling, had no effect on Aux/IAA degradation. Our results show that SHY2/IAA3 and BDL/IAA12 are poly-ubiquitinated and degraded in response to increased auxin or TIR1 levels. In conclusion, our data provide experimental support for the model that SCF^TIR1-dependent poly-ubiquitination of Aux/IAA proteins marks these proteins for degradation by the 26S proteasome, leading to activation of auxin-responsive gene expression.

Introduction

The plant hormone auxin (indole-3-acetic acid or IAA) plays an essential role in a large variety of developmental processes throughout a plant’s life cycle. Auxin is transported in a polar cell-to-cell manner, and this transport directs cell division and growth by generating dynamic auxin gradients in tissues and organs. At the cellular level, auxin concentrations are translated into a gene expression response by the complex and dynamic interaction between two major families of transcriptional regulators: the auxin response factors (ARFs) and the labile Aux/IAA proteins (Guilfoyle et al., 1998; Ulmasov et al., 1999; Tiwari et al., 2001). The ARF proteins bind to specific sequences in the promoters of auxin-responsive genes through their N-terminal DNA-binding domain, and either activate or repress transcription. At the C-terminus they share the conserved domains III and IV with the Aux/IAA proteins, through which they homo- or heterodimerize with other ARFs or with the Aux/IAA proteins (Guilfoyle et al., 1998).

The Aux/IAA proteins are short-lived transcriptional regulators that repress transcription controlled by auxin-responsive elements (AuxREs) by heterodimerizing with ARFs. The repressor activity of these proteins is located in the N-terminal domain I, whereas their stability is regulated by the central domain II. Auxin binds to the receptor TIR1 or the related auxin signaling F-box (AFB) proteins that are part of the E3 ubiquitin ligase complex SCF^TIR1/AFB. Auxin enhances the affinity of TIR/AFB for domain II of Aux/IAAs (Gray et al., 2001; Dharmasiri et al., 2005a,b; Kepinski and Leyser, 2005; Tan et al., 2007). Mutations in either the Aux/IAA domain II or in one of the SCF components lead to auxin-resistant phenotypes that are mostly due to the stabilization of the Aux/IAA repressors (Worley et al., 2000; Ouellet et al., 2001). Moreover, treatment of plants with proteasome inhibitors leads to the accumulation of these proteins, indicating that levels of Aux/IAA proteins are controlled by the 26S proteasome (Gray et al., 2001; Ramos et al., 2001; Tian et al., 2003). Taken together, this information leads to a model in which auxin-enhanced binding of TIR1/AFB to domain II of Aux/IAAs results in ubiquitination of these proteins, which marks them for degradation by the 26S proteasome. However, there is no direct experimental evidence for the SCF^TIR1/AFB-mediated ubiquitination of Aux/IAA proteins. Recent proteome-wide screens using a multidimensional protein identification technology were not able to detect Aux/IAAs among the ubiquitinated proteins in Arabidopsis (Maor et al., 2007; Manzano et al., 2008), suggesting that more direct methods may be necessary to detect the ubiquitinated Aux/IAAs.

Although implicated in a large variety of cellular responses, protein ubiquitination of plant transcriptional regulators has been demonstrated for only a few targets, including SLENDER RICE1 (SLR1) by SCF^GID2 in gibberellin responses (Sasaki et al., 2003), LONG HYPOCOTYL IN FAR RED (HRF1), LONG AFTER FAR-RED LIGHT1 (LAF1) and LONG HYPOCOTYL 5 (HY5) by CONSTITUTIVE PHOTOMORPHOGENIC1 (COP1) during photomorphogenesis (Xie et al., 2002; Saijo et al., 2003; Jang et al., 2005) and LEAFY (LFY) by SCF^UFO (Chae et al., 2008) in floral development. Targeted proteolysis by SCF-mediated ubiquitination has been implied for the transcription factor ETHYLENE INSENSITIVE3 (EIN3) by SCF^EBF1/EBF2 in ethylene signaling (Guo and Ecker, 2003; Potuschak et al., 2003), for JASMONATE-ZIM-DOMAIN (JAZ) repressors by SCF^COI1 in jasmonate signaling (Chini et al., 2007; Thines et al., 2007) and for many other proteins in diverse processes, but experimental evidence for ubiquitination of these proteins and for the specificity of the SCF-ligases for their targets is lacking.

SHORT HYPOCOTYL 2 (SHY2/IAA3) and BODENLOS (BDL/IAA12) are distantly related Aux/IAA proteins that regulate auxin responses at different stages of a plant’s life cycle. Both have been shown to interact with TIR1 and to be stabilized by treatment with proteasome inhibitors. The mutant proteins shy2-2 and bdl carry the same proline to serine mutations in their domain II, leading to stabilized products that show no or only residual interaction with TIR1/AFB (Tian et al., 2003; Dharmasiri et al., 2005b). BODENLOS has been described as an interactor/regulator of the ARF MONOPTEROS (MP/ARF5) acting on embryo patterning (Hamann et al., 2002). SHORT HYPOCOTYL 2 is also able to interact with MP/ARF5 and to inhibit its activity, but the expression patterns of these 2 proteins do not overlap in planta. Recent studies have connected the action of SHY2/IAA3 to modulation of NON-PHOTOTROPHIC HYPOCOTYL4 (NPH4/ARF7)- and ARF19-regulated hypocotyl growth (Weijers et al., 2005).

The Aux/IAAs are short-lived proteins with hardly detectable endogenous levels (Abel et al., 1994; Ramos et al., 2001; Tian et al., 2003). This instability and rarity makes the study of their targeted degradation difficult to perform in planta. For a long time now, plant cell protoplasts have been used to assess hormone responses due to their amenability for transformation and their responsiveness to diverse stimuli (Abel and Theologis, 1998; Sheen, 2001). Much of the information gathered on the mechanisms of regulation of Aux/IAA stability has been performed using transiently transformed protoplasts (Ramos et al., 2001; Tiwari et al., 2001, 2004), and other important components of the auxin signaling pathway have been functionally characterized in this system including ARFs, SCF^TIR1 and RAC GTPases (Guilfoyle et al., 1998; Tao et al., 2005; Wang et al., 2005).

In this study, we used Arabidopsis cell suspension protoplasts to demonstrate that auxin-enhanced TIR1-mediated ubiquitination of SHY2/IAA3 and BDL/IAA12 marks these proteins for degradation and leads to auxin-responsive gene expression. Our results show that auxin sensitivity of Arabidopsis protoplasts is strictly connected to the relative abundance of the TIR1 protein and Aux/IAA proteins, and that an excess of TIR1 leads to depletion of Aux/IAAs even in the absence of exogenous auxin. The mutant protein tir1-1 and the related F-box protein COI1, which is involved in jasmonate signaling, had no effect on Aux/IAA stability or ubiquitination, corroborating the specificity and importance of TIR1 in the process.

Results and discussion

Auxin-responsive gene expression in Arabidopsis cell suspension protoplasts is repressed by Aux/IAAs

Leaf protoplasts have been used extensively to study auxin signaling and the regulation of auxin-responsive gene expression by ARF and Aux/IAA proteins (Ulmasov et al., 1997, 1999; Abel and Theologis, 1998; Kovtun et al., 1998). For our experiments we obtained protoplasts from Arabidopsis cell suspension cultures because of their continuous availability, the easy isolation procedure and the high transformation efficiencies obtained. To establish the experimental conditions under which the repressive effect of Aux/IAA proteins on auxin-responsive gene expression could be detected in these cells, we transfected the auxin-responsive reporter construct DR5::GUS alone or in the presence of plasmids encoding hemagglutinin (HA)-tagged Aux/IAA proteins. We selected BDL/IAA12 or SHY2/IAA3 for our studies, because they are well-characterized, but distantly related, and they are involved in different developmental processes and are therefore representative of the other Aux/IAAs (Tian and Reed, 1999; Tian et al., 2002, 2003; Weijers et al., 2005).

In order to identify working parameters for assaying transcriptional responses to auxin in protoplasts, we transfected 1 million cells with 10 μg of the auxin-responsive DR5::GUS reporter plasmid alone. GUS expression was induced by auxin in a concentration-dependent manner (Figure 1a), and a maximum response was obtained with 1 μm IAA. Co-transfection with 1 μg of the 35S::HA-BDL/IAA12 effector plasmid led to a significant reduction of this response, and when 5 or 10 μg effector plasmid was co-transfected the repression of the DR5 promoter was saturated in that its activity remained at 30%, even when the cells were treated with 1 or 10 μm IAA (Figure 1a). Similar results were obtained with NAA (data not shown). Based on these results (Figure 1a), a reporter:effector plasmid ratio of 10:1 (in microgram) and an auxin concentration of 1 μm IAA were used to study the repression activity on the DR5 promoter. Repression activity on the DR5 promoter was shown by both SHY2/IAA3 and BDL/IAA12. The amount of BDL/IAA12 plasmid transfected resulted in a stronger repression than transfection with the same amount of SHY2/IAA3 plasmid. The auxin response was almost completely repressed by the stabilized mutant versions shy2-2 and bdl (Figure 1b).

Figure 1.  Auxin-responsive DR5::GUS reporter gene expression in Arabidopsis protoplasts is repressed by the auxin/indole-3-acetic acid (Aux/IAA) proteins BODENLOS (BDL/IAA12) and SHORT HYPOCOTYL 2 (SHY2/IAA3). (a) Relative GUS/luciferase (LUC) activity (arbitrary units) in transfected Arabidopsis protoplasts treated for 8 h with increasing concentrations of IAA. The numbers on the x-axis indicate the reporter (DR5::GUS) to effector (35S::HA-BDL/IAA12) ratio in microgram plasmid DNA per transformation. (b) Effect of co-transfecting Aux/IAA effector plasmids expressing SHY2/IAA3, BDL/IAA12 or the stabilized mutant proteins shy2-2 or bdl on the activity of the DR5::GUS reporter. Reporter and effector plasmids were co-transformed in a reporter to effector ratio of 10:1 μg plasmid DNA. Protoplasts were treated for 8 h with DMSO or 1 μm IAA. Letters indicate significantly different groups based on one-way anova (P < 0.05).

Cellular auxin and TIR1 levels are interdependent parameters in Aux/IAA degradation

Although clearly active in the repression of transcription, the transiently expressed Aux/IAA proteins in the previous experiments could not be detected on western blots (data not shown). To be able to correlate the level of transcriptional inhibition caused by the HA-tagged Aux/IAA proteins with the degree of protein turnover, we designed a different experimental set-up. Protoplasts were transfected with 20 μg of the effector plasmids 35S::HA-SHY2/IAA3 or 35S::HA-BDL/IAA12 and after auxin treatment total protein extracts were analyzed on western blots using antibodies against the HA-epitope. Both HA-BDL/IAA12 and HA-SHY2/IAA3 were clearly detectable, but under these conditions the latter failed to show a clear enhanced turnover after auxin treatment (Figure 2a, left panel). In fact, levels of IAA3 protein increased during incubation, indicating that the de novo production was higher than the turnover rate. It is interesting to note that although the HA-SHY2/IAA3 protein was more abundant in Figure 2a, BDL/IAA12 had higher repression activity in the DR5 promoter assays in Figure 1b. The facts that excessive amounts of repressor construct led to saturated repression of the DR5 promoter, which could not be overcome by the addition of higher auxin concentrations (Figure 1a), and that the increased amount of SHY2/IAA3 protein failed to be degraded following auxin treatment (Figure 2a, left panel), led us to hypothesize that some component of the auxin-responsive protein degradation machinery in protoplasts was rate-limiting. To test if the amount of auxin receptor TIR1 was rate-limiting, we co-transfected the cells with 20 μg Aux/IAA effector construct and increasing amounts of 35S::FLAG-TIR1-c-Myc. Co-transfection of 4 μg TIR1 plasmid led to enhanced auxin-dependent turnover of SHY2/IAA3 (Figure 2a, IAA3/TIR panels). When 10 or 20 μg of the TIR1 plasmid was added, the basal levels of SHY2/IAA3 became very low, while at 20 μg the effect of auxin could not be visualized due to detection limitations (Figure 2a, IAA3/TIR 20 μg panel). The stability of the mutant HA-shy2-2 protein was not affected by auxin without co-transfected TIR1, but with high amounts of TIR1 even the turnover of the stabilized HA-shy2-2 started to become evident (Figure 2a, shy2-2/TIR1 panels). For shy2-2 it has been reported that the mutant protein retains part of its TIR1-binding activity in the presence of auxin (Tian et al., 2003). This may explain the enhanced turnover of the shy2-2 protein in the presence of additional TIR1 and exogenous auxin.

Figure 2.  Auxin-induced TIR1-dependent degradation of auxin/indole-3-acetic acid (Aux/IAA) proteins leads to auxin-responsive gene expression in Arabidopsis protoplasts. (a) Western blot detection of transiently expressed hemagglutinin (HA)-BODENLOS (BDL/IAA12), HA-SHORT HYPOCOTYL 2 (SHY2/IAA3), HA-bdl and HA-shy2-2 in Arabidopsis protoplasts co-transfected with control plasmid 35S::FLAG-GFP and increasing amounts of plasmid 35S::FLAG-TIR1c-myc. One million transfected protoplasts were split in two halves and each was treated either with 0.1% DMSO (–) or 1 μm NAA (+) and samples were harvested on the indicated time points (in hours). The labels on top indicate the Aux/IAA construct and the amount of TIR1 plasmid. The panels indicated with ‘α-HA’ represent the detection of HA-tagged Aux/IAA with anti-HA antibodies and ‘α-FLAG input’ indicates detection of FLAG-GFP with anti-FLAG antibodies. Numbers on the bottom of each lane indicate the HA versus FLAG signal density ratio relative to time point zero in each transformation. (b) Correlation between auxin-responsive DR5::GUS expression and TIR1-dependent Aux/IAA degradation. Protoplasts were transfected with 10 μg DR5::GUS, 10 μg 35S::RLuc, 10 μg 35S::FLAG-GFP and where indicated with 20 μg 35S::HA-Aux/IAA effector plasmid. The graph in (b) shows relative GUS/luciferase (LUC) activity (in arbitrary units) in protoplasts after 4 h treatment with DMSO or 1 μm NAA, in the absence (No TIR1) or presence (TIR1 added) of 10 μg 35S::FLAG-TIR1-c-Myc. Letters indicate significantly different groups based on one-way anova (P < 0.05). (c) Western blot analysis showing the level of HA-tagged Aux/IAA protein in the NAA-treated samples in (b). Addition of 35S::TIR1 is indicated (−/+). Detection of the FLAG-GFP control (α-FLAG) is used to demonstrate comparable transformation efficiencies and loading.

TIR1-dependent Aux/IAA degradation coincides with auxin-responsive gene expression

To correlate the TIR1-dependent turnover of Aux/IAAs with auxin-responsive gene expression, we repeated the DR5::GUS trans-activation experiments, but now using 20 μg of effector plasmid, and measuring both GUS activity and protein levels. Again, in these experiments the BDL/IAA12 protein showed equal or even stronger repression activity on DR5::GUS expression as SHY2/IAA3 (Figure 2b), whereas it accumulated to a much lower level than SHY2/IAA3 (Figure 2c). In the absence of additional TIR1, auxin treatment led to a weak activation of the DR5 promoter (Figure 2b, ‘No TIR1’ panel, and 2c). Co-transfection of 10 μg TIR1 plasmid enhanced the responsiveness of the DR5 element (Figure 2b, ‘TIR1 added’ panel) and this effect correlated with an increased turnover of the Aux/IAA repressors (Figure 2c). TIR1 alone resulted in an increase of GUS activity in DMSO-treated cells without co-transfected Aux/IAAs (Figure 2b, first bar in the ‘TIR1 added’ panel). This effect is probably due to the degradation of the endogenous pool of Aux/IAAs. As expected, the levels of the mutant shy2-2 and bdl proteins were less sensitive to overexpression of TIR1. The slight increase in DR5::GUS activity observed after auxin treatment in the samples co-transfected with shy2-2 and TIR1 might reflect the residual binding of the mutant shy2-2 protein to TIR1 (Tian et al., 2003) and its enhanced turnover rate at higher TIR1 levels (compare with Figure 2a, shy2-2/TIR1 20 μg panel). Additionally, comparative analysis of BDL/IAA12 and SHY2/IAA3 indicated that optimized ARF and Aux/IAA interaction pairs are active in specific auxin-regulated developmental processes (Weijers et al., 2005). It is therefore likely that the specific ARFs that are responsible for the activation of the DR5 element in cell suspension protoplasts interact more efficiently with BDL/IAA12 or endogenous Aux/IAAs than with SHY2/IAA3. This may also explain the stronger repressor activity of BDL/IAA12.

The observation that excessive amounts of TIR1 led to reduction of detectable Aux/IAAs even in the absence of auxin treatment is in agreement with the effect of overexpressing TIR1 in plants, which mimics auxin treatment and causes auxin hypersensitivity (Gray et al., 1999). It also corroborates the finding that in in vitro pull-down experiments Aux/IAA proteins do interact with TIR1 in the absence of auxin, albeit at low efficiency (Dharmasiri et al., 2003, 2005a,b; Kepinski and Leyser, 2005; Tan et al., 2007), and that Aux/IAA degradation and auxin-responsive gene expression are severely affected in the tir afb2 afb3 triple mutant (Dharmasiri et al., 2005b). All these data indicate that TIR1 and AFB protein levels are important determinants of the auxin responsiveness of cells.

TIR1 differentially regulates SHY2/IAA3 and BDL/IAA12 abundance

Interestingly, there is a clear and significant difference in the abundance and repressor activity of the two wild-type Aux/IAA proteins, with BDL/IAA12 being the stronger but less abundant repressor (Figure 2b). To assess the roles of auxin and TIR1 in these different behaviors of the two Aux/IAA proteins, we tested the auxin-induced degradation of SHY2/IAA3 and BDL/IAA12 with increasing TIR1 levels (Figure 3a). The HA-BDL/IAA12 levels were auxin-sensitive when cells were transfected with 20 μg of 35S::HA-BDL/IAA12 effector plasmid alone. The observed variation between experiments in the effect of auxin treatment on IAA12 turnover (compare Figures 2a and 3b) possibly relates to differences in endogenous TIR1 levels. In the same experiment, auxin treatment did not lead to a clear reduction in HA-SHY2/IAA3 levels, corroborating our observation that SHY2/IAA3 has a longer half-life than BDL/IAA12 (Figure 2b). Co-transfection of the effector plasmids with increasing amounts of 35S::FLAG-TIR1-c-Myc made the levels of both proteins more sensitive to auxin treatment (Figure 3a). As a control, effector plasmids were also transfected with plasmids expressing mutant versions of TIR1 (tir1-1 [G147→D]), ΔF-TIR1 lacking the F-box, or ΔF-tir1-1 carrying both mutations (Figure 3b), or the related F-box protein COI1, which is involved in jasmonate signaling (Figure 3c). Neither the mutant versions of TIR1 (Figure 3b) nor COI1 (Figure 3c) affected the abundance of the Aux/IAA proteins (compare with 20/10 HA-IAA/TIR1 treatments in Figure 3a), corroborating the specificity of TIR1 in Aux/IAA degradation.

Figure 3.  TIR1-dependent degradation of transiently expressed SHORT HYPOCOTYL 2 (SHY2/IAA3) and BODENLOS (BDL/IAA12) in Arabidopsis protoplasts. (a) Western blot analysis of total extracts from protoplasts co-transfected with increasing amounts of 35S::FLAG-TIR1-c-Myc, 35S::FLAG-GFP control plasmid and 35S::HA-SHY2/IAA3 (upper panel) or 35S::HA-BDL/IAA12 (lower panel). Transfected protoplasts were either untreated (0), or treated for 1 h with DMSO (−) or 1 μm NAA (+). The numbers on top indicate the 35S::HA-AuxIAA to 35S::FLAG-TIR1-c-Myc plasmid ratio in microgram DNA. Blots were probed with anti-hemagglutinin (HA) or anti-FLAG antibodies. (b) As in (a), following co-transfection with 10 μg 35S::FLAG-tir1-1-c-Myc, 35S::FLAG-ΔF-TIR1-c-Myc or 35S::FLAG-ΔF-tir1-1-c-Myc. (c) As in (a), following co-transfection with 10 μg 35S::FLAG-COI1. (d) Western blot detection of FLAG-tagged TIR1, tir1-1 and COI1 immunoprecipitated using anti-FLAG antibodies from pooled extracts from five transformations of 10⁶ protoplasts with 20 μg DNA of 35S::FLAG-TIR1-c-Myc, 35S::FLAG-tir1-1-c-Myc or 35S::FLAG-COI1, respectively. The position of the FLAG-tagged TIR1-c-Myc and tir1-1-c-Myc proteins (79 kDa) is indicated by an arrowhead, the position of FLAG-COI1 (70 kDa) with a star. The lane marked UNT contained protein from untransformed protoplasts. An unspecific band recognized by the anti-FLAG antibody in all samples (circle) shows equal loading.

In these experiments TIR1, tir1-1 and COI1 were not detectable by western blot analysis of total protein extracts using the anti-FLAG antibody, even though their expression was driven by the strong 35S promoter. This observation suggests that TIR1 and COI1 themselves are short-lived proteins, which is in line with the observation that several F-box proteins, including TIR1 and COI1, are targets for ubiquitination (Maor et al., 2007; Jurado et al., 2008; Stuttmann et al., 2009). To demonstrate that the F-box proteins were expressed to similar levels in our model system, we transfected five times more cells than usual (5 × 10⁶) and immunoprecipitated the FLAG-tagged F-box proteins from total cell extracts. Western blot analysis of the concentrated eluates showed that the FLAG-tagged versions of TIR1, tir1-1 and COI1 were expressed at similar levels (Figure 3d). As expected, transfection with double the amount of plasmid led to the production of more TIR1 protein (Figure S1). Treatment with MG132 for 4 h did not lead to elevated TIR1 levels nor to the appearance of additional modified bands (see Figure S1 in Supporting Information), which is in contrast to the conclusion by Stuttmann et al. (2009) that the protein is a target of the 26S proteasome.

TIR1 was shown to act as an auxin receptor together with other F-box family members ABF1, ABF2 and ABF3 (Dharmasiri et al., 2005a,b; Kepinski and Leyser, 2005). Interestingly, a tir1 afb1 afb2 afb3 quadruple loss-of-function mutant shows a variable phenotype, but several of these mutant plants are able to flower and set seed, suggesting further functional redundancy. We have not tested the effect of AFB1, AFB2 or AFB3 in our system, but it is likely that they will show similar effects as TIR1 overexpression, since they were found to physically interact with GST-tagged BDL/IAA12 in an auxin-dependent manner in pull-down assays, and the BDL/IAA12 protein was stabilized in the tir1-1 afb2 afb3 triple mutant (Dharmasiri et al., 2005b). The related F-box proteins AFB5 and AFB4 have not been studied in detail yet, but the specific resistance of afb5 mutants to picolinate auxin analogs indicates that AFB5 is involved in the response pathway to these herbicides (Walsh et al., 2006).

COI1 is the closest relative of the TIR/AFB auxin receptors in the F-box family tree. The similarity in sequence and the ability of different proteins from this clade to associate with the same SCF components (Gray et al., 1999; Xu et al., 2002) raised the possibility that there is cross-recognition of targets among related F-box proteins and that COI1 may also be actively involved in Aux/IAA protein degradation. The results in Figure 3(c) clearly show that the presence of overexpressed FLAG-COI1 did not affect the stability of either SHY2/IAA3 or BDL/IAA12. Based on our results, we conclude that COI1 is not involved in the process of Aux/IAA proteolysis.

SCF^TIR1-mediated ubiquitination marks Aux/IAA proteins for degradation

The previous experiments demonstrated that the cell suspension protoplast system reproduces the in planta action of Aux/IAA proteins and the receptor F-box protein TIR1 in auxin responses. The same system was used to establish whether auxin- and TIR1-enhanced degradation of Aux/IAA proteins is connected to SCF^TIR1-mediated ubiquitination. A plasmid encoding HA-tagged versions of BDL/IAA12 or SHY2/IAA3 together with a plasmid carrying TIR1 or its mutant version tir1-1 were transfected into cells. Transformations of 1 million protoplasts were performed in triplicate and one of the samples was pre-treated for 1 h with the proteasome inhibitor MG132. Samples were subsequently treated for one additional hour with either DMSO or 1 μm NAA. Five per cent of the protein input was analyzed by western blotting using anti-HA antibodies, and the remaining sample was immunoprecipitated with anti-HA antibodies conjugated to agarose beads. The recovered proteins were analyzed by western blotting using anti-ubiquitin and anti-HA antibodies.

The relative amounts of Aux/IAA proteins recovered by immunoprecipitation correlated well with those present in the input extracts (Figure 4, α-HA: 5% input versus IP-HA). Samples from protoplasts that were treated with auxin and MG132 showed additional anti-HA reactive bands migrating slower than the unmodified SHY2/IAA3 and BDL/IAA12 monomers (Figure 4a,b, IP-HA α-HA, +/+ lanes). Anti-ubiquitin antibodies detected bands of sizes larger than 40 kDa (IP-HA α-Ub panels) that overlapped with the additional bands detected with anti-HA antibodies (marked with black arrowheads), and thus represent poly-ubiquitinated versions of Aux/IAA proteins. The combined auxin and MG132 treatment enhanced the amount of detectable ubiquitinated Aux/IAA proteins. Interestingly, when TIR1 was co-expressed, auxin treatment resulted in an increase of the ubiquitinated signal for both HA-SHY/IAA3 and HA-BDL/IAA12 (Figure 4a,b, +/+ lanes), corroborating our previous observation that SCF^TIR1-directed degradation of Aux/IAA proteins is dependent on a fine balance between auxin and TIR1 levels (Figure 3a,b). As expected, the co-expression of the mutant tir1-1 protein had no effect on the rate of turnover or the ubiquitination level of SHY2/IAA3 (Figure 4a, tir1-1 lanes). As the goal of this experiment was to detect the rare ubiquitinated Aux/IAAs forms, significantly higher amounts of protein were loaded on the gel, and this prevented us from observing the auxin-induced turnover of Aux/IAA proteins in the absence of co-transfected 35S::FLAG-TIR1-c-Myc.

Figure 4.  Auxin/indole-3-acetic acid (Aux/IAA) proteins are poly-ubiquitinated and degraded in an auxin- and TIR1-dependent manner. (a, b) Western blot analysis of anti-hemagglutinin immunoprecipitations (IP-HA) or total extracts (5% input) from Arabidopsis protoplasts transfected with (a) 35S::FLAG-GFP and 35S::HA-SHY2/IAA3 with either 10 μg of 35S::FLAG-TIR1-c-Myc (TIR1) or 35S::FLAG-tir1-1-c-Myc (tir1-1) or (b) 35S::HA-BDL/IAA12 and 35S::FLAG-GFP co-transfected with 10 μg 35S::FLAG-TIR1-c-Myc. Each transformation was performed in triplicate. One sample was treated for 1 h with 0.1% DMSO (−/−), the second sample for 1 h with 1 μm NAA (+/−), and the third replicate was pre-treated for 1 h with MG132 followed by 1 h treatment with 1 μm NAA (+/+). Five per cent of the total protein extract of each transfection was analysed as input (5% input). The remaining sample was used in an immunoprecipitation with anti-HA antibodies conjugated to agarose beads and these samples (IP-HA) were first probed with anti-ubiquitin antibodies (α-Ub) to detect ubiquitinated proteins, and subsequently with anti-HA antibodies (α-HA) to detect HA-tagged proteins. The bands corresponding to unmodified HA-IAA3 and HA-IAA12 monomers are indicated. The poly-ubiquitinated forms of IAA12 and IAA3, visible as higher molecular size bands/smears in the α-Ub and α-HA blots, are indicated by Poly-Ub. Arrowheads mark the bands overlapping in the α-Ub and α-HA blots. The arrow indicates the position of a putative homodimer of the HA-tagged Aux/IAA proteins.

One constant observation in the western blots with anti-HA antibodies was the presence of two distinct bands: one at ∼30 kDa corresponding to the size of unmodified Aux/IAA monomers (indicated in Figure 4 on the left of each panel by HA IAA3 or HA IAA12) and a fainter band around 60 kDa in size (Figure 4, indicated by arrows, α-HA panels). This band was not detected with the anti-ubiquitin antibodies, and although we cannot exclude other post-translational modifications such as phosphorylation or sumoylation, based on the size shift we believe this band to represent Aux/IAA homodimers. The 60-kDa band was also detected even when protein samples were prepared and gel separated under strong denaturing conditions (boiled in Laemmli loading buffer prior to urea-SDS–PAGE, data not shown), suggesting a covalent coupling, or a denaturation-resistant association of the Aux/IAA proteins. One possibility is that the dimers are stabilized by intermolecular disulfide bonds, which is a common mechanism in redox control of transcription factor activity (Benezra, 1994; Zheng et al., 1998; Mou et al., 2003). However, the fact that the band is not dissolved by the thiol reducing agent 2-mercaptoethanol in the Laemmli buffer suggests that the Aux/IAA dimers are stabilized by another mechanism. The 60-kDa band putatively representing Aux/IAA homodimers almost disappeared in TIR1 co-transfected samples, suggesting that an additional consequence of Aux/IAA degradation is the dissolution of Aux/IAA dimers. Hypothetically, the dimerized forms might be more accessible for interaction with SCF^TIR1 than the DNA-ARFs associated ones and hence be more easily degraded.

Repressor poly-ubiquitination: a paradigm for plant hormone signaling pathways

Our results indicate that auxin-responsive gene expression in Arabidopsis protoplasts depends on a fine-tuning of the intracellular concentrations of different elements that participate in the auxin perception pathway. Auxin sensitivity and Aux/IAA stability are directly correlated with intracellular levels of TIR1. Previous reports showed that auxin responses in seedlings are enhanced by TIR1 overexpression (Gray et al., 1999) and repressed by TIR1/AFB loss of function (Dharmasiri et al., 2005b). In addition, our experiments directly correlate the TIR1-enhanced auxin response with the increased turnover of Aux/IAA proteins, and suggest that TIR1, when present at sufficiently high levels, sensitizes the protoplast cells to endogenous auxin levels and can mediate Aux/IAA degradation without the need for exogenous auxin application. We demonstrate that the SHY2/IAA3 and BDL/IAA12 proteins behave differently in protoplasts, the latter being less stable but more active in the repression of the auxin response in this system. The simplest explanation is that IAA12 is a more efficient repressor than IAA3, but we cannot rule out that the effect is indirect through the efficient interaction of overexpressed IAA12 with the SCF^TIR1/AFB complexes, which sequesters these complexes and thereby stabilizes endogenous Aux/IAAs.

It is well established that several plant hormonal signaling pathways act through proteasomal degradation of transcriptional repressors, and that the hormones show similar roles in enhancing the association of SCF E3 ligase complexes with their targets (e.g. auxin/TIR1/Aux/IAA, JA-Ile/COI1/JAZ and GA/GID/DELLA.) (Dharmasiri et al., 2005a; Kepinski and Leyser, 2005; Griffiths et al., 2006; Thines et al., 2007). Our data provide experimental support for the model that hormone-responsive gene expression is mediated by hormone-enhanced poly-ubiquitination and subsequent proteolytic degradation of repressor proteins by the 26S proteasome.

Experimental procedures

Protoplast isolation and transformation

Arabidopsis thaliana Col-0 cell suspension cultures were used for protoplast preparations. Culture maintenance, protoplast isolation and transfections were performed as previously described (Schirawski et al., 2000) with minor modifications. Four-to-six-day-old cultures were diluted five-fold in auxin-free cell medium (30 g L⁻¹ saccharose, 3.2 g L⁻¹ Gamborg’s B5 basal medium with mineral organics, adjusted to pH 5.8 with KOH and sterilized by autoclaving), incubated overnight and used for protoplast isolation in auxin-free solutions. Transfected cells were kept at 25°C in the dark for 16 h before treatments. Where necessary, additional DNA of plasmid pART7 (Gleave, 1992) was added, to equalize the amount of DNA for each transformation.

DNA constructs

For the auxin-responsive GUS assays, a DR5::GUS reporter construct (Benjamins et al., 2001) was used. A plasmid carrying the Renilla reniformis luciferase (LUC) gene under the control of the CaMV 35S promoter was co-transfected as a control for transformation efficiency (De Sutter et al., 2005). All effector plasmids are based on pART7 carrying the CaMV 35S promoter and OCS terminator. GATEWAY^® destination cassettes (Invitrogen, http://www.invitrogen.com) derived from pEarleyGate 201 and 202 (Earley et al., 2006) were transferred into pART7 to generate plasmids pART7-HA and pART7-FLAG for the expression of, respectively, N-terminally HA- or FLAG- tagged proteins in plant cells.

The N-terminally HA-tagged cDNAs of SHY2/IAA3 and shy2-2 (P69→S) were cloned from pACT2-SHY2 and pACT2-shy2-2 (kindly provided by Jason Reed, University of North Carolina) using XhoI/XbaI sites into pART7, generating 35S::HA-SHY2/IAA3 and 35S::HA-shy2-2. The BDL/IAA12 cDNA was excised with BamHI/XbaI from pET16H-BDL (Weijers et al., 2006) and introduced into pENTR 3C (Invitrogen, http://www.invitrogen.com/), and the resulting entry clone was used to create 35S::HA-BDL/IAA12 via LR recombination in pART7-HA (C.S. Galvan-Ampudia and RO, unpublished) to generate 35S::HA-BDL/IAA12. The bodenlos (P75→S) mutation (Hamann et al., 2002) was introduced in this plasmid using the Quickchange Site-directed Mutagenesis kit (Stratagene, http://www.stratagene.com/) resulting in 35S::HA:bdl.

The entry clone for TIR1myc (Gray et al., 1999) was used for generating the tir1-1(G147→D) mutation by Quickchange site-directed mutagenesis. The deletions of the F-box motif in TIR1 and tir1-1, which removes the first 50 amino acids from the original sequence substituting I50 for an alternative M as a start codon, were generated via PCR with primers: 5′-GAATTCATGGGGAACTGCTACGCCGTGAG-3′ and 5′-GCGGATCCCTAAAACCTCATTGTTGAGTC-3′. The COI1 cDNA was amplified from a leaf cDNA library with the primers 5′-CGAGCTCAAAATGGAGGATCCTGATATCAAG-3′ and 5′-GGGGTACCGACTGACTCTATGTAATCTCC-3′ and cloned into pENTR2B. Entry clones were used in a LR reaction with pART7-FLAG, generating 35S::FLAG-GFP, 35S::FLAG-TIR1myc, 35S::FLAG-tir1-1myc and 35S::FLAG-COI1.

GUS and LUC assays

In the DR5::GUS transactivation assays 10⁶ protoplasts were transfected with 10 μg of the DR5::GUS reporter construct and 2 μg of 35S:Rluc (p2rL7; De Sutter et al., 2005) for experimental normalization. The amounts of DNA of the effector constructs varied per experiment and are indicated in the figure legends. All transformations contained 10 μg of 35S::FLAG-GFP as a control for transformation efficiency, and were split into two portions containing 5 × 10⁵ protoplasts in a total volume of 2.5 ml of protoplast medium. After 16 h the samples were treated for 4 h either with 1 μm IAA or the same volume of the solvent DMSO. Treated cells were collected by centrifugation at 80 g for 1 min and the pellets were frozen in liquid nitrogen for GUS (van der Fits and Memelink, 1997) and LUC measurements (Dyer et al., 2000). Triplicate transfections were assayed and mean GUS/LUC relative activities were analyzed by one-way anova using spss 15.0 software.

Immunoblotting and immunoprecipitation

For the Aux/IAA degradation/ubiquitination assays, 10⁶ protoplasts were transfected with 20 μg 35S::HA-Aux/IAA construct and 10 μg of 35S::FLAG-GFP. Depending on the experiment, plasmids encoding FLAG-tagged TIR1myc, tir1-1myc or COI1 were co-transfected in the amounts indicated in the figure legends. Treated protoplasts were resuspended by vortexing in cold extraction buffer (PBS, 1× Roche complete protease inhibitor cocktail, http://www.roche.com/) containing 1% Triton X-100, and the lysate was cleared by centrifugation at 20 000 g for 10 min. Total protein was quantified by Bradford assay (Bio-Rad, http://www.bio-rad.com/) and 20 μg was mixed with sample buffer and separated on 15% SDS–PAGE minigels. The PAGE-separated proteins were blotted onto nitrocellulose membranes, blocked with non-fat dry milk and incubated with the horseradish peroxidase (HRP)-conjugated antibodies anti-HA High Affinity 3F10 (Roche) and anti-FLAG M2 (Sigma, http://www.sigma-aldrich.com/). For detection of ubiquitinated proteins, 10⁶ transformed cells were resuspended in 100 μl extraction buffer containing 1% Triton X-100, 5 mm EDTA, 10 mm 1,10-phenanthroline monohydrate (OPA, Sigma) and 10 μm MG132 (Sigma). The lysate was cleared by centrifugation at 20 000 g for 10 min, and 5 μl was western-analyzed as 5% input control. The remaining total extract was diluted to a final volume of 900 μl with extraction buffer without Triton X-100 to bring the Triton concentration to 0.1%. This extract was then mixed with 40 μl of Anti-HA affinity matrix (Roche) and incubated for 2 h at 4°C. The matrix was pelleted and washed 3× in Extraction Buffer, mixed with sample buffer, and the eluted proteins were separated on 12% SDS–PAGE minigels. Polyvinylidene fluoride (PVDF) membranes containing transferred proteins were blocked with Qiagen blocking reagent (http://www.qiagen.com) and probed with 1000-fold diluted HRP-Anti-Ub P4D1 antibodies (Santa Cruz, http://www.scbt.com/). After chemiluminescent detection (LumiGLO, Cell Signaling, http://www.cellsignal.com/) the blots were stripped and reprobed with anti-HA High Affinity 3F10 antibodies (Roche). When necessary, quantification of the signal was performed on scanned X-ray films using a Bio-Rad GS-800 calibrated densitometer.

Acknowledgements

We thank Marc Estelle (UC San Diego, California, USA) and Dolf Weijers (Wageningen University, NL) for sharing materials and Jason Reed (University of North Carolina, Chapel Hill, USA) for providing pACT2:SHY2/IAA3 and pACT2:shy2-2. FSM was partially supported through a van der Leeuw grant from the Netherlands Organisation for Scientific Research (NWO) awarded to JM.

Accession numbers: BDL/IAA12 (At1g04550), SHY2/IAA3 (At1g04240), TIR1 (At3g62980), COI1 (At2g39940)