Characterization of a novel temperature-sensitive allele of the CUL1/AXR6 subunit of SCF ubiquitin-ligases


(fax +1 612 625 1738; e-mail


Selective protein degradation by the ubiquitin-proteasome pathway has emerged as a key regulatory mechanism in a wide variety of cellular processes. The selective components of this pathway are the E3 ubiquitin-ligases which act downstream of the ubiquitin-activating and -conjugating enzymes to identify specific substrates for ubiquitinylation. SCF-type ubiquitin-ligases are the most abundant class of E3 enzymes in Arabidopsis. In a genetic screen for enhancers of the tir1-1 auxin response defect, we identified eta1/axr6-3, a recessive and temperature-sensitive mutation in the CUL1 core component of the SCFTIR1 complex. The axr6-3 mutation interferes with Skp1 binding, thus preventing SCF complex assembly. axr6-3 displays a pleiotropic phenotype with defects in numerous SCF-regulated pathways including auxin signaling, jasmonate signaling, flower development, and photomorphogenesis. We used axr6-3 as a tool for identifying pathways likely to be regulated by SCF-mediated proteolysis and propose new roles for SCF regulation of the far-red light/phyA and sugar signaling pathways. The recessive inheritance and the temperature-sensitive nature of the pleiotropically acting axr6-3 mutation opens promising possibilities for the identification and investigation of SCF-regulated pathways in Arabidopsis.


Proteolysis has emerged as an important regulatory mechanism in plant growth and development. The highly conserved ubiquitin/26S proteasome-dependent pathway is the primary proteolytic system in eukaryotes (Hershko and Ciechanover, 1998). Ubiquitinylation involves the covalent attachment of a multi-ubiquitin chain to specific proteins. The ubiquitin chain serves as a degradation tag, leading to substrate proteolysis via the 26S-proteasome (Pickart, 2001). This pathway degrades a wide range of proteins but is highly specific, and controls a large number of cellular events due to the diversity of the participating enzymes. In plants, ubiquitin-mediated proteolysis plays a key role in several developmental pathways, including hormonal responses, photomorphogenesis, and circadian rhythms (Hellmann and Estelle, 2002). Compared with Saccharomyces cerevisiae, Caenorhabditis elegans and Drosophila, Arabidopsis contains an over-representation of ubiquitin conjugases and ligases (Bachmair et al., 2001), suggesting the importance of regulated proteolysis in plant developmental processes.

Ubiquitinylation involves a three-step reaction where the E1-activated ubiquitin moiety is transferred to the substrate protein by the E2 conjugating and E3 ligating enzymes (Pickart, 2001). Individual E3 complexes exhibit differential affinity for selected substrate proteins, thereby conferring specificity to the ubiquitin pathway. The SCF E3 superfamily comprises a major class of ubiquitin-ligases in plants. The SCF complex is composed of Skp1, Cullin, an F-box protein, and the small RING protein Rbx1. The cullin subunit acts as a large scaffold protein that ensures optimal presentation of the substrate to E2 (Zheng et al., 2002a). The Skp1 protein binds to the F-box domain of the F-box protein and mediates interaction between the F-box protein and the N-terminal domain of cullin (Schulman et al., 2000; Zheng et al., 2002a). The C-terminal domains of F-box proteins commonly contain protein–protein interaction motifs such as LRR, WD40 and Kelch repeats, which may confer binding specificity to different substrates. One of the best-characterized SCF ubiquitin-ligases in plants is the SCFTIR1 complex, which mediates response to the hormone auxin. In the SCFTIR1 complex, the TIR1 gene encodes an F-box protein that interacts with the Skp1-like proteins ASK1 or ASK2 and the cullin, CUL1 (Gray et al., 1999). In response to the hormone, SCFTIR1 targets members of the Aux/IAA family of transcriptional regulators for ubiquitin-mediated degradation (Gray et al., 2001).

With approximately 700 F-box proteins in Arabidopsis (Gagne et al., 2002; Risseeuw et al., 2003), the target protein-recruiting F-box component is highly variable within the complex. The Skp1-like proteins are represented by 21 ASK genes (Risseeuw et al., 2003), and the gene family encoding cullins contains at least 11 predicted members (Shen et al., 2002). Only five of these contain open reading frames encoding apparently intact canonical C- and N-terminal domains, with CUL1 and CUL2a suggested to be the dominant participants in SCF complex formation in Arabidopsis (Risseeuw et al., 2003). Although CUL2a interacts with Skp1/F-box protein modules in yeast two-hybrid assays (Risseeuw et al., 2003), only CUL1 has been shown to be part of SCF complexes in planta (Gray et al., 1999; del Pozo et al., 2002; Wang et al., 2003; Xu et al., 2002). Hence, CUL1 may be a common subunit of most SCF complexes. Mutations in this pleiotropically acting gene hold the potential for identifying and studying pathways regulated by SCF-mediated ubiquitinylation. Homozygous CUL1 T-DNA insertion mutants (cul1-1cul1-4) result in embryonic arrest at the zygote stage (Shen et al., 2002) and therefore, have limited use for study and identification of additional SCF complexes later in development. Hellmann et al. (2003) identified two mutant CUL1 alleles, axr6-1 and −2, that confer reduced auxin response. Both point mutations are semi-dominant and affect the same residue near the N-terminus of CUL1, a region known to interact with the SKP1/F-box protein module. The axr6-1 and −2 mutations are homozygous seedling lethal. These two mutations confer deleterious effects on the SCF-dependent pathways regulating auxin response and floral development (Hobbie et al., 2000; Ni et al., 2004), however, their semi-dominant and homozygous lethal nature complicates genetic and biochemical analysis.

In a screen for enhancers of the tir1-1 auxin-resistance phenotype we previously identified ETA3/SGT1b and ETA2/CAND1 as regulators of SCFTIR1 (Chuang et al., 2004; Gray et al., 2003). Here, we report the identification and characterization of eta1/axr6-3, a novel recessive, temperature-sensitive allele of CUL1. We show the capability of eta1/axr6-3 as a tool for the analysis of multiple pathways known or suggested to involve SCF regulation, and identify additional pathways likely to be regulated by SCF-mediated ubiquitinylation.


Identification and isolation of the eta1/axr6-3 mutation

The enhancer of tir1-1 auxin resistance (eta1) mutant was isolated in a tir1-1 enhancer screen designed to identify additional genes required for SCFTIR1-mediated auxin response. This screen identified numerous loci including ETA3/SGT1b (Gray et al., 2003) and ETA2/CAND1 (Chuang et al., 2004), two components of the SCFTIR1 pathway. The eta1 tir1-1 M2 plant was backcrossed to tir1-1, and auxin response of the F2 progeny was assessed by examining root growth on media containing 0.25 μm 2,4-D. 115 of 490 of the F2 seedlings exhibited auxin-resistant root growth, indicating that eta1 is a recessive mutation of a single locus (3:1; χ2 = 0.48).

Analysis of the F2 progeny from a Columbia backcross revealed that eta1 single mutants were resistant to applied auxin. The eta1 mutants also exhibited a striking dwarf phenotype that was largely unaffected by the tir1-1 mutation. Surprisingly, we found that growth at elevated temperatures dramatically enhanced the eta1 phenotype (Figure 1). Similarly, temperature had a profound effect on the eta1 auxin response defect, but did not alter the dose–response curves of wild type or axr1-12 seedlings (Figure 2a). These findings suggested that eta1 is a temperature-sensitive mutation. Also consistent with a reduction in auxin response, eta1 seedlings developed fewer lateral roots than the wild type on unsupplemented nutrient medium, and the eta1 mutation enhanced the tir1-1 lateral root defect (Figure 2b). In addition to impaired auxin response, eta1 plants exhibit reduced apical dominance, delayed senescence, reduced male fertility, and aberrant flower development (data not shown).

Figure 1.

Temperature-sensitive growth of the eta1 mutant. Eight day-old wild type and eta1 seedlings grown on ATS nutrient medium at 20°C were transferred to soil and grown an additional 38 days at 20, 24, or 28°C. All size bars = 1 cm.

Figure 2.

eta1 confers reduced auxin response.
(a) Inhibition of root elongation by increasing concentrations of the synthetic auxin 2,4-D. Five-day-old seedlings grown on ATS at 20°C were transferred to medium containing 2,4-D and grown for four additional days at the indicated temperature. Data points are averages from 10 seedlings. Standard deviations for all data points were ≤12% of the mean.
(b) Lateral root (LR) initiation was assessed in 13-day-old seedlings grown on unsupplemented ATS nutrient medium at 20 or 28°C. Error bars indicate standard deviation from the mean (n = 10).
(c) Eight-day-old seedlings grown at 20°C on ATS nutrient medium. The genotypes of axr6-3 axr1-12 double mutant seedlings were confirmed by sequencing PCR products.

A map-based cloning approach was used to isolate the eta1 mutation. Linkage studies placed eta1 in an approximately 300 kb interval at the top of chromosome 4. Among the genes in this region was AXR6, encoding the CUL1 core component of the SCFTIR1 complex (Hellmann et al., 2003). We therefore sequenced AXR6 in the mutant and detected a missense mutation resulting in replacement of glutamic acid 159 with a lysine. Further confirmation that the eta1 mutation affects CUL1 was obtained by complementation tests with the cul1-3 null allele (Hellmann et al., 2003) of AXR6 (data not shown). We therefore renamed the eta1 mutant as axr6-3.

All aspects of the eta1/axr6-3 mutant phenotype are recessive, even when seedlings are grown at elevated temperatures (data not shown). In contrast, the previously characterized axr6-1 and −2 alleles of CUL1 confer a semi-dominant auxin response defect, with homozygous mutants exhibiting a seedling-lethal phenotype. When crossed with the semi-dominant axr6-1 allele, the axr6-3/axr6-1 F1 progeny resembled axr6-1/AXR6 plants, although a slight increase in the level of auxin-resistant root growth was observed (data not shown).

The axr1-12 mutation also confers a severe auxin response defect. Previous studies have demonstrated that AXR1 functions in the conjugation of the ubiquitin-like protein RUB to CUL1, which is required for optimal SCFTIR1 activity (Gray et al., 2001, 2002; del Pozo and Estelle, 1999). Homozygous axr6-3 axr1-12 double mutants exhibited severe developmental defects resulting in lethality early in seedling development similar to the homozygous phenotype of axr6-1 or −2 seedlings (Hobbie et al., 2000) (Figure 2c). The finding that the axr1-12 mutation enhances the recessive axr6-3 phenotype to the extent that the double mutant combination phenocopies homozygous axr6-1 and −2 plants suggests that the two semi-dominant axr6 mutations are strong dominant-negative alleles.

The axr6-3 mutation disrupts the SCFTIR1 complex

The SCFTIR1 complex is known to target Aux/IAA proteins for proteasomal degradation. We therefore examined whether and how the axr6-3 mutation affected SCFTIR1 ubiquitin-ligase activity. Aux/IAA protein stability was examined using AXR2/IAA17 pulse-chase assays (Gray et al., 2001). AXR2 protein was immunoprecipitated from 35S-methionine-labeled extracts prepared from seedlings immediately after the labeling period or following a 15-min chase with unlabeled methionine. Steady-state levels of AXR2 protein as well as AXR2 stability were increased in the axr6-3 mutant compared with the wild type (Figure 3a). The average AXR2 half-life and standard deviation determined from three independent experiments was 11.8 ± 1.6 min in wild type versus 63.1 ± 10.3 min in axr6-3.

Figure 3.

Effects of the axr6-3 mutation on the SCFTIR1 complex.
(a) AXR2 protein was immunoprecipitated from wild type and axr6-3 seedlings labeled with 35S-Met. Precipitations were performed immediately after labeling (t = 0) or following a 15-min chase with medium containing 1 mm Met and 100 μg ml−1 cycloheximide (t = 15).
(b) Col and axr6-3 seedlings carrying the HS::AXR3NT-GUS transgene were heat-shocked to induce expression of the transgene and stained for β-glucuronidase activity immediately (t = 0) or following a 20-min incubation at room temperature in media containing 10 μm IAA (t = 20). Seedlings shown are representative of the approximately 50 analyzed for each condition.
(c) CUL1 Western blot analysis of 40 μg of crude extract prepared from 8-day-old Col, axr6-3, axr6-1, and axr6-2 seedlings. Seedlings were grown at 20°C continually or shifted to 28°C for 24 h where indicated. Asterisk indicates RUB-modified CUL1. Relative amounts of total CUL1 protein are indicated below the lanes. Bottom panel shows a cross-reacting protein used as a loading control.
(d) Anti-c-myc antibody was used to immunoprecipitate the SCFTIR1 complex from extracts prepared from tir1-1 and tir1-1 axr6-3 seedlings expressing a TIR1-myc fusion gene. Immunoprecipitates were immunoblotted with anti-ASK1 and anti-CUL1 antibodies. Seedlings were grown at 20°C continually or shifted to 28°C for 24 h where indicated.
(e) Serial dilutions of yeast strain YPB2 carrying the indicated two-hybrid plasmids. Cells were plated onto 3-AT selection medium (right) or control medium (left) and incubated at room temperature.
(f) Anti-CAND1/ETA2 antibody was used to immunoprecipitate CAND1 from Col, cand1-null, and axr6-3 seedling extracts. Precipitates were immunoblotted with α-CAND1 and α-CUL1 antisera. Asterisk indicates RUB-modified CUL1.

Results examining Aux/IAA stability using the HS::AXR3NT-GUS reporter construct (Gray et al., 2001) were also consistent with axr6-3 mutants exhibiting a defect in SCFTIR1-mediated ubiquitinylation of Aux/IAA proteins. Wild type and axr6-3 seedlings containing the HS::AXR3NT-GUS reporter were heat-shocked to induce expression of the transgene followed by incubation at room temperature in the absence or presence of exogenous auxin and subsequently stained to detect GUS expression levels. Immediately after the heat shock period, axr6-3 seedlings exhibited a slight increase in AXR3NT-GUS staining compared with wild-type controls. However, following a 30-min incubation in the presence of 10 μm IAA, the axr6-3 mutant showed dramatically higher levels of AXR3NT-GUS staining than wild type (Figure 3b).

We next examined whether the axr6-3 mutation affected CUL1 levels or RUB modification of CUL1. Crude extracts were prepared from seedlings grown continually at 20°C or shifted to 28°C for 24 h prior to extraction and immunoblotted with α-CUL1 antisera. When grown at 20°C, total CUL1 levels were unaffected by the axr6-3 mutation (Figure 3c). Consistent with the possibility that axr6-3 is a temperature-sensitive allele, mutant seedlings grown at high temperature exhibited a two- to threefold reduction in CUL1 steady-state levels. This effect on CUL1 abundance is contrary to that of axr6-1 and −2, both of which result in increased CUL1 levels (Hellmann et al., 2003).

As the degradation of Aux/IAA proteins was affected by the axr6-3 mutation, we next looked at the assembly of the SCF complex. c-myc epitope-tagged TIR1 was immunoprecipitated from tir1-1[TIR1myc] and axr6-3 tir1-1[TIR1myc] seedling extracts and the precipitates immunoblotted with α-ASK1 and α-CUL1 antisera. While the ASK1 protein co-immunoprecipitated with TIR1myc from both extracts, CUL1 was dramatically reduced in precipitates from axr6-3 seedlings grown at 20°C and was absent from the precipitates of axr6-3 seedlings shifted to 28°C for 24 h (Figure 3d).

Numerous molecular and structural studies on SCF complexes have demonstrated that CUL1 interaction with the F-box protein is dependent upon both proteins binding to the SKP1 subunit (Zheng et al., 2002a). The finding that the axr6-3 mutation dramatically diminishes the amount of CUL1 present in TIR1-myc precipitates suggested that the mutation may affect the ability of CUL1 to interact with SKP1. We therefore examined the affect of axr6-3 on SKP1 binding in a yeast two-hybrid assay. Whereas wild-type CUL1 interacts with ASK1 in this assay, the axr6-3 mutant derivative does not, as indicated by the inability to activate the PGAL-HIS3 reporter gene and confer resistance to 3-amino-1′,2′,4′-triazole (Figure 3e). Together, these findings demonstrate that the axr6-3 point mutation interferes with Aux/IAA protein degradation by preventing assembly of the SCFTIR1 complex.

CUL1 also interacts with the ETA2/CAND1 protein, which regulates SCF complex assembly (Chuang et al., 2004; Feng et al., 2004). In contrast to its effect on SCF assembly, the axr6-3 mutation did not prevent CUL1 from associating with CAND1 (Figure 3f) in co-immunoprecipitation assays.

Exploring the effects of the axr6-3 mutation on other SCF pathways

As CUL1 is a core component of the SCF, the axr6-3 mutation provides a novel tool for identifying and studying SCF-regulated pathways in Arabidopsis. In contrast to the semi-dominant, homozygous lethal axr6-1 and −2 alleles, the recessive nature of the axr6-3 allele simplifies both genetic and biochemical studies of SCF function. To determine whether axr6-3 is a capable tool for the identification of SCF-regulated pathways, we first examined additional pathways known, or suggested to be regulated by SCF-mediated ubiquitinylation.

Response to the phytohormone methyl jasmonate (MeJA) requires the SCFCOI1 ubiquitin-ligase complex (Xu et al., 2002). axr6-3 seedlings exhibited increased resistance to root growth inhibition by MeJA (Figure 4a). Consistent with this physiological assay, axr6-3 seedlings also exhibited reduced expression of the jasmonic acid-induced gene, VSP1 (Figure 4b). As observed for the auxin pathway, growth at elevated temperature exacerbated the axr6-3 defect in both JA response assays.

Figure 4.

The axr6-3 mutation confers reduced jasmonate response.
(a) Wild type and axr6-3 seedlings were grown for 8 days at 20 and 28°C on nutrient medium supplemented with MeJA. Error bars indicate standard deviations from the mean (n = 10).
(b) RT-PCR analysis of VSP1 transcript abundance in wild type and axr6-3 seedlings. Seedlings were grown for 5 days in liquid ATS nutrient medium at 20°C, then shifted to 20 or 28°C for 2 days, and induced with 100 μm MeJA for 8 h. Values represent the mean ± SD of three experiments. Inset: VSP1 Northern blot of total RNA prepared from the 28°C samples.

We also examined the effects of the axr6-3 mutation on SCFUFO and SCFAtSKP2, which regulate floral development and AtE2Fc stability, respectively (del Pozo et al., 2002; Wang et al., 2003). We found that the axr6-3 mutation acts as an enhancer of the ufo-6 floral organ defect and confers a modest increase in AtE2Fc stability (data not shown), consistent with previous findings that CUL1 is a component of these SCF complexes. Two additional plant hormones known to be regulated by SCF-mediated proteolysis are ethylene and GA (Dill et al., 2004; Guo and Ecker, 2003; Potuschak et al., 2003; Sasaki et al., 2003). Surprisingly, we were unable to detect a significant difference between axr6-3 and wild type in physiological assays for these two hormones (data not shown). Furthermore, GA-mediated degradation of the DELLA protein RGA was unaffected by the axr6-3 mutation suggesting that the SCFSLY1 ubiquitin-ligase remains functional in the mutant.

axr6-3 confers far-red light hypersensitivity and delayed phyA degradation kinetics

The EID1 and AFR F-box proteins have both been implicated in regulation of far-red light signaling. EID1 (empfindlicher im dunkelroten Licht) encodes an F-box protein that interacts with the Arabidopsis Skp1 homologs ASK1 and ASK2. eid1 mutants exhibit hypersensitivity to far-red and weak red light, leading to the suggestion that EID1 acts by targeting positive components of the phyA-signaling pathway for ubiquitin-dependent proteolysis (Dieterle et al., 2001). In contrast, RNAi lines with diminished expression of another F-box gene – attenuated far-red response (AFR) – display phenotypes consistent with reduced phyA-mediated light signaling, and SCFAFR is predicted to mediate the turnover of a repressor of phyA signaling, possibly to prepare the plant to receive light signals at dawn (Harmon and Kay, 2003).

We examined the response of axr6-3 seedlings to several light conditions. No differences between axr6-3 and wild type could be found for blue and red light over a wide range of fluences in hypocotyl elongation assays (data not shown). Under far-red light, however, axr6-3 mutants exhibited a strong hypersensitive phenotype with open cotyledons and reduced hypocotyl growth, while wild-type seedlings remained largely etiolated (Figure 5a,b). As axr6-3 seedlings exhibited an etiolated phenotype in the dark and normal responses to blue and red wavelengths, this phenotype is strictly far-red light dependent. Double mutant analysis with phyA confirmed that the far-red light hypersensitivity of axr6-3 requires phyA (data not shown).

Figure 5.

axr6-3 confers hypersensitivity to far-red light.
(a) Phenotype of wild type and axr6-3 seedlings grown for 4 days in the dark or under continuous far-red light (0.016 μmol m−2 sec−1).
(b) Fluence rate response curves for the inhibition of hypocotyl growth in continuous far-red light. Hypocotyl length was analyzed using 4-day-old seedlings. The relative lengths of the hypocotyls were determined in relation to the lengths of etiolated seedlings for each line. Error bars indicate standard deviations from the mean (n = 12).
(c) phyA degradation in wild type and axr6-3 seedlings. Four-day-old dark-grown Col and axr6-3 seedlings were treated with red light (20 μmol m−2 sec−1) for the indicated time and phyA levels were assessed by Western blot using anti-phyA antibody. A loading control is shown in the bottom panel.
(d) NIH Image was used to quantitate relative phyA protein levels following red light irradiation. Values represent the mean ± SD of three independent experiments. Seedlings were grown at 23°C for all of these experiments.

As phyA is known to be subject to ubiquitin-mediated turnover, we compared the rate of phyA degradation in wild-type Col plants with axr6-3 mutants after irradiation with red light. Under our conditions (20 μmol m−2 sec−1), 3 h irradiation of 4-day-old dark-grown seedlings with red light reduced phyA levels in the wild type to approximately 10% of the dark control, and no phyA was detectable after 5 h (Figure 5c). Irradiation of axr6-3 seedlings also resulted in a reduction in phyA levels, however, the rate at which this occurred was significantly slower than wild type, supporting a role for SCF-mediated proteolysis in the regulation of phyA stability (Figure 5c,d).

axr6-3 confers hypersensitivity to sugars and environmental pH

Having demonstrated defects in several pathways known to involve SCF-mediated protein degradation, we next examined the axr6-3 mutant in several additional assays to identify additional pathways that might employ SCF-mediated ubiquitinylation. No effects could be identified in response to the plant hormones cytokinin, abscisic acid, and salicylic acid. Furthermore, no differences between axr6-3 and wild type could be found in response to NaCl, heavy metals, and osmotic or oxidative stress (data not shown).

Numerous mutants implicated in sugar response display insensitivity to high levels of sucrose or glucose on germination and seedling establishment (Rolland et al., 2002). The axr6-3 mutant, however, displayed increased sensitivity to sugar-mediated inhibition of seedling establishment. While germination was not affected, emergence of true leaves on increasing concentrations of glucose was progressively reduced compared with the wild type (Figure 6a). Root growth assays also revealed a dramatic difference between wild type and axr6-3 on both glucose and sucrose medium (Figure 6b,c). Similar to other aspects of the axr6-3 phenotype, growth at elevated temperatures exacerbated the mutant defect. For example, while axr6-3 seedlings exhibited comparable glucose sensitivity to wild type up to concentrations of 2% when grown at 20°C, at 28°C axr6-3 was significantly inhibited at concentrations as low as 0.5% (Figure 6b). Neither mannitol nor sorbitol (data not shown) differentially affected the growth of axr6-3 and wild-type seedlings, suggesting that a specific carbohydrate-mediated signal, rather than an osmotic signal, is behind the enhanced response of the axr6-3 mutant.

Figure 6.

The sugar hypersensitive phenotype of axr6-3.
(a) True leaf formation of wild type and axr6-3 seedlings grown for 14 days at 20°C on nutrient medium supplemented with the indicated concentration of glucose (n = 100 seedlings/data point).
(b) Root growth assay of seedlings grown for 8 days at 20 or 28°C on nutrient medium supplemented with the indicated concentration of glucose. Error bars indicate standard deviations from the mean (n = 10).
(c) Root growth assay of seedlings grown for 8 days at 20 or 28°C on nutrient medium supplemented with the indicated concentration of sucrose. Error bars indicate standard deviations from the mean (n = 10).
(d) Increased starch levels in axr6-3. Wild type and axr6-3 seedlings were grown for 7 days at 20°C under continuous light on ATS nutrient medium. Prior to iodine staining seedlings were incubated for 17 h in the dark in order to empty starch sinks and incubated for the indicated time under light to induce starch synthesis. Depicted seedlings are representative of several independent experiments. Bar = 1 mm.
(e) RT-PCR analysis of transcript levels in 10-day-old wild type and axr6-3 seedlings. Seedlings were grown under constant light on nutrient medium supplemented with the indicated concentrations of sucrose. Transcript abundance was normalized to an Actin2 standard. Values represent the mean ± SD of three experiments.

Consistent with the sugar hypersensitivity phenotype, starch levels were substantially higher in axr6-3 plants than in wild type. After 17 h darkness, complete depletion of starch stored in the leaves was observed, as indicated by yellow, unstained leaves after qualitative starch measurement by iodine staining (Figure 6d). Following extended exposure to light, the wild type gradually fills its starch sinks – as can be seen by darkly stained leaves following iodine staining – until saturation. In contrast, the starch levels in the axr6-3 mutant seem to be saturated as early as 2 h after exposure to light.

The increased sugar response could also be shown for the expression of sugar-induced genes, as shown by real-time semiquantitative PCR in 7-day-old seedlings grown under constant light on medium containing 0, 0.1, or 0.2 m sucrose (Figure 6e). The cytosolic beta-amylase (Atβ-amylase), the large subunit of ADP-glucose pyrophosphorylase (ApL3), and chalcone synthase (CHS) all showed increased transcript levels in axr6-3 compared with the wild type, suggesting disturbed regulation of sugar signaling in axr6-3.

Lastly, we also detected a difference between wild type and axr6-3 plants in their response to environmental pH. Extreme pH of the surrounding environment induces proton stress and can alter the solubility of metal ions affecting their uptake into the plant. A root growth assay on nutrient medium with low (3.5, 4.5) or high (10) pH, showed significantly decreased root elongation for axr6-3 compared with the wild type. While these differences were striking at high temperature (28°C) (Figure 7), the difference was not significant at lower temperatures (data not shown). Very little is known about how plants respond to suboptimal environmental pH conditions. Young et al. (1998) demonstrated that expression of a constitutively active derivative of the AHA3 plasma membrane H+-ATPase confers increased pH tolerance to Arabidopsis. It seems possible that some aspect of proton pump activity is subject to SCF-mediated regulation, however, further study is required to elucidate how plants cope with suboptimal pH conditions.

Figure 7.

axr6-3 confers hypersensitivity to environmental pH. Root growth assay of wild type and axr6-3 seedlings grown for 8 days at 28°C on nutrient medium of different pH. Values are normalized to growth on ATS plates, pH 5.7 (n = 10). Error bars indicate standard deviation from the mean (n = 10).


SCF ubiquitin-ligases have emerged as key regulators of numerous plant signal transduction pathways. We have identified a novel recessive allele of the CUL1/AXR6 subunit of SCF complexes. axr6-3 confers several phenotypes consistent with reduced auxin response, including auxin-resistant root elongation, diminished apical dominance, and reduced lateral root development. Further analysis indicates that these phenotypes are the result of reduced SCFTIR1 ubiquitin-ligase activity as Aux/IAA stability is dramatically increased in the axr6-3 mutant.

In contrast to axr6-3, the previously characterized axr6-1 and −2 alleles are both semi-dominant alleles that confer seedling lethality when homozygous. Both alleles cause a missense mutation of phenylalanine 111 in the N-terminal Skp1-binding domain of CUL1 and result in increased levels of CUL1 protein (Hellmann et al., 2003). The axr6-3 mutation affects a highly conserved glutamic acid residue 48 amino acid C-terminal to this site. Based on the crystal structure of the human SCFSKP2 complex, the eta1 mutation is in between the first and second helices of the second cullin repeat and should not affect interactions with Skp1, which contacts helices 2 and 5 of the first cullin repeat (Zheng et al., 2002a). However, we find that the axr6-3 mutation does indeed diminish Skp1 binding in a yeast two-hybrid assay. Consistent with this possibility, the axr6-3 mutation results in a reduction in the amount of CUL1 that co-immunoprecipitates with TIR1 from plant extracts suggesting that the mutant cullin cannot function as an SCF scaffolding protein. The possibility that the mutation has a global destabilizing affect on CUL1 structure seems unlikely as the interaction with ETA2/CAND1 is unaffected by the axr6-3 mutation. CAND1 binding requires both the N-terminal Skp1-binding domain of CUL1 as well as the C-terminal RUB/NEDD8 modification domain (Goldenberg et al., 2004; Zheng et al., 2002b). One likely possibility is that the axr6-3 mutation causes a local structural change that interferes with Skp1 binding without altering the interaction with CAND1.

axr6-3 as a probe for SCF-regulated pathways

As a core component of SCF complexes, CUL1 is predicted to regulate many processes in the plant. Consistent with this possibility, we find that the axr6-3 mutation affects numerous developmental pathways. In contrast to the previously characterized seedling-lethal alleles of AXR6, the axr6-3 mutation is recessive, thus simplifying both genetic and biochemical studies. Furthermore, the axr6-3 mutation exhibits temperature sensitivity. All of the phenotypes associated with the mutation that we have examined are significantly more severe when plants are grown at elevated temperatures. This is likely due to thermo-lability of the mutant protein as we consistently observed a two- to threefold reduction in CUL1 protein levels in extracts prepared from axr6-3 plants shifted to high temperature. The ability to modulate the severity of the mutant phenotype by temperature might facilitate the study of SCF regulation of essential processes such as the cell cycle and embryogenesis.

We have demonstrated that the axr6-3 mutant is defective in several SCF-regulated pathways including auxin and jasmonate response and flower development. We have also found a dramatic effect on circadian rhythms, likely due to defects in SCF complexes containing the ZTL family of F-box proteins (W.M. Gray, Frank G. Harmon and Steve A. Kay, personal communication). While these findings clearly demonstrate the axr6-3 mutant is a useful probe for identifying SCF-regulated pathways, it should also be noted that we failed to detect defects in ethylene and GA signaling, both of which have been shown to involve SCF components. This is particularly surprising given that the SLY1 and EBF1/2 F-box proteins interact with CUL1 in co-immunoprecipitation and pull-down assays, respectively (Fu et al., 2004; Potuschak et al., 2003). Our findings raise the possibility that a related cullin such as CUL2a might be capable of substituting for CUL1 in the SCFSLY1 and SCFEBF complexes.

Far-red light signaling

As early as 1987 Shanklin et al. (1987) provided the first evidence for the possible involvement of the ubiquitin/26S proteasome pathway in the degradation of phyA by identifying the ubiquitin modification of phyA soon after Pfr formation. Seo et al. (2004) have recently shown that the RING motif-containing E3 ligase COP1 ubiquitinylates phyA in vitro. Furthermore, phyA degradation is decreased in cop1 mutants suggesting that COP1 acts as part of an E3 ligase to regulate phyA signaling by targeting the photoreceptor for degradation.

We find that axr6-3 mutants are hypersensitive to far-red light and exhibit increased phyA stability. Two F-box proteins, EID1 and AFR (Dieterle et al., 2001; Harmon and Kay, 2003), have been implicated in SCF-mediated regulation of phyA signaling. However, phyA degradation kinetics are not altered in either eid1 or afr mutants, suggesting that these F-box proteins are not directly involved in the control of phyA stability. Instead, EID1 has been suggested to target other positively acting downstream components of the phyA signaling pathway for degradation (Dieterle et al., 2001), while AFR has been proposed to mediate the turnover of a downstream repressor of phyA signaling (Harmon and Kay, 2003). These findings, together with our studies on axr6-3, suggest that an additional CUL1-containing SCF complex exists that is capable of ubiquitinylating phyA upon Pfr formation. Alternatively, the COP1 ubiquitin-ligase could potentially be subject to regulation by an SCF complex involving CUL1.

Potential involvement of an SCF complex in sugar signaling

Sugars are often considered similar to hormones in that they are synthesized in one part of the plant and then transported to other parts of the plant where they affect gene expression, time of flowering, early seedling development and other developmental processes. We have demonstrated a hypersensitive sugar response in the axr6-3 mutant in several physiological and molecular assays.

Baier et al. (2004) recently described the high sugar response (hsr) mutants that affect the regulation of sugar-induced and sugar-repressed processes controlling gene expression, growth, and development in Arabidopsis. Like axr6-3, the hsr mutants exhibit increased sugar sensitivity in seedling establishment, gene expression, and starch accumulation assays and appear to act independently of ABA and ethylene. It is therefore possible that the HSR gene products and CUL1 act in the same pathway to regulate sugar signaling.

A simple model for explaining the sugar hypersensitivity of axr6-3 is that an SCF complex targets a positive regulator of sugar signaling for proteolysis. To our knowledge, no SCF component has previously been implicated in sugar signaling in Arabidopsis. Nonetheless, there are several links between sugar signaling and SCF-mediated ubiquitinylation. In yeast, SCFGRR1 is a central component in glucose signaling and is tightly linked to the function of hexokinase (HXK) (Ozcan and Johnston, 1999). The role of HXK as a sugar sensor in plants has recently been demonstrated (Moore et al., 2003) and GRR1-like F-box proteins are encoded by the Arabidopsis genome (Thelander et al., 2002). Furthermore, the yeast Snf1 Ser/Thr protein kinase is one of the major components in yeast sugar signaling and is required for derepression of a large number of glucose-regulated genes (Carlson, 1999). A comparable role as global regulator of carbon metabolism seems to be the case for plants (Halford and Hardie, 1998). The Arabidopsis pleiotropic regulatory locus (prl1) mutant exhibits transcriptional derepression of sugar-regulated genes and a sugar hypersensitive growth phenotype (Nemeth et al., 1998) comparable to axr6-3. PRL1 inhibits Arabidopsis Snf1-related protein kinases (SnRKs), and competes with the SCF subunit ASK1 for the same binding sites within the SnRKs, which have been proposed to regulate docking of SCF complexes with the 26S proteasome (Farras et al., 2001). We are currently investigating the possibility that AXR6 and PRL1 act in a common pathway to regulate sugar responses.

With approximately 700 predicted F-box proteins in the Arabidopsis genome, it is likely that SCF ubiquitin-ligases regulate a myriad of processes throughout plant growth and development. However, fewer than 20 F-box proteins have been assigned to specific pathways. The axr6-3 mutant provides a novel tool for identifying SCF-regulated pathways and investigating SCF function in plant growth and development. Furthermore, the temperature-sensitive nature of the mutation should be helpful in studying essential processes and may be useful for proteomic approaches designed to identify SCF substrates.

Experimental procedures

Plant materials and growth conditions

All Arabidopsis thaliana lines employed in this study are in the Col ecotype. Seedlings were grown under sterile conditions on ATS nutrient medium (Lincoln et al., 1990) under long-day lighting conditions unless indicated otherwise. Conditions for the mutagenesis and screen for eta mutants have been previously described (Gray et al., 2003).

A total of 240 auxin-resistant F2 seedlings from a cross between axr6-3 and Ler were used to map the eta1/axr6-3 mutation using cleaved-amplified polymorphic sequence (CAPS) and simple sequence length (SSLP) polymorphic markers. The mutation was initially mapped to an interval between ciw5 and ciw6 ( Additional markers were generated using the Cereon Arabidopsis polymorphism collection (Jander et al., 2002). Markers defining our final mapping interval were CER457975 (5′-CTAACCCTAGTTCTAATCTTC-3′ and 5′-CTGTCCGATGACGAGAAGGG-3′), which amplifies 104 and 98 bp fragments from Col and Ler, respectively, and GA1.1 (

Yeast two-hybrid assays

An EcoRI–XhoI fragment containing the CUL1 coding sequence was cloned in-frame with the DNA-binding domain of GAL4 of the pBridge vector (Clontech Laboratories, Inc., Mountain View, CA, USA). The axr6-3 mutation was introduced using the Quickchange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). The ASK1 coding sequence was cloned as an XmaI–BamHI fragment into pGADT7 (Clontech Laboratories, Inc.). Constructs were introduced into yeast strain YPB2 (Kohalmi et al., 1998). Liquid cultures were grown to saturation in synthetic complete media lacking tryptophan and leucine. Cells were harvested by centrifugation, washed once with sterile water, and serial dilutions plated onto synthetic complete medium lacking tryptophan, leucine, and histidine, supplemented with 15 mm 3-amino-1′,2′,4′-triazole to assess the ability of the clones to activate the GAL1UAS-HIS3 reporter gene.

Antibodies, co-immunoprecipitation, and Western blot analysis

All antibodies used in this study have been described previously (Chuang et al., 2004; Gray et al., 1999; del Pozo et al., 2002; Silverstone et al., 2001). phyA antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Protein gel blot, AXR2 pulse-chase, and TIR1myc co-immunoprecipitation analyses were performed as described by Gray et al. (1999, 2001). ETA2-CUL1 co-immunoprecipitations were performed using seedling extracts in PBS buffer that had been chemically cross-linked using 3,3-dithio bis(sulfo-succinimidylpropionate) as previously described (Rancour et al., 2004). Where indicated, quantification of immunoblots was performed using NIH Image with ECL exposures on pre-flashed autoradiography film.

Glucuronidase histochemical staining

The HS::AXR3NT-GUS transgene (Gray et al., 2001) was crossed into the axr6-3 mutant. Six-day-old Col and axr6-3 seedlings homozygous for the reporter construct were heat-shocked for 2 h at 37°C to induce expression of the transgene. Seedlings were then stained immediately or transferred to 20°C medium and incubated for 20 or 40 min before staining for β-glucuronidase activity (Stomp, 1991). Indole-3-acetic acid (10 μm) was added to the 20°C medium where indicated.

Northern blot analysis and RT-PCR

For the MeJA assay, plants were grown for 5 days in liquid ATS at 20°C under constant shaking and then transferred to 28°C for another 2 days before induction with 50 μm MeJA or EtOH for 8 h. For the sucrose assays, plants were grown under 24 h light for 5 days at 20°C on ATS plates containing the indicated sucrose concentrations and were then shifted to 28°C and grown an additional 2 days. Total RNA was extracted from seedlings using the RNeasy plant kits (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions including DNase treatment.

Northern blots were performed with 10 μg total RNA/sample using standard techniques. For RT-PCR analysis, first-strand cDNA synthesis was performed using 2 μg RNA and M-MLV-RTase (Promega, Madison, WI, USA) following the manufacturer's instructions. cDNA samples were diluted 20-fold. Gene-specific primers were designed to allow detection of amplification products from contaminant genomic DNA using primers spanning one or more introns. The specific primers were Act2-F: GAGAAGATGACTCAGATC, Act2-R: ATCCTTCCTGATATCGAC, β-Amy-F: CGG AGAAGGGGAAGTTTTTC, β-Amy-R: AATCTCATGCCCGTACTTCG, CHS-F: GGCTCAGAGAGCTGATGGAC, CHS-R: TTAGGGACTTCGACCACCAC, ApL3-F: CAACTGGCATCGATCTGAAA, ApL3-R: CTCCCAAGAAACATCCGTGT, VSP1-F: ACGTCCAGTCTTCGGCATCC, VSP1-R: CTTAAAAACCCTTCCAG.

Real-time PCR and control reactions were performed on the LightCycler System (Roche Applied Sciences, Indianapolis, IN, USA) using the REDExtract-N-Amp Plant PCR Kit (Sigma, St Louis, MO, USA) supplemented with SYBR Green. PCR cycling conditions comprised an initial denaturation step at 95°C for 3 min, followed by 45–50 cycles of 95°C for 1 sec, 50°C for 8 sec, and 72°C for 28 sec. To eliminate the detection of non-specific products, fluorescence was acquired during each cycle at 82°C, which is above the melting temperature of primer dimers. Background subtraction and determination of the maximum second derivative (Rasmussen, 2001) for each amplification curve was performed using the LightCycler 3.5 analysis software (Roche Applied Sciences). The quantitative real-time RT-PCR results of each primer pair were normalized to the Actin2 transcript as an internal standard. Transcript abundance of the sucrose- or MeJA-treated Col and mutant samples was calculated in relation to the respective Col untreated sample (transcript abundance Col untreated = 1). Relative transcript levels were visualized by reducing the cycle number until the rate of PCR product generation was in the early exponential stage of amplification (β-Amy: 30 cycles, ApL3: 29 cycles; CHS: 25 cycles).

Starch staining

Iodine staining of starch levels was conducted as previously described (Yu et al., 2001). Briefly, 8-day-old seedlings were incubated in 80% ethanol at 70°C for 15 min to clear the leaves. Ethanol was removed and the seedlings incubated in iodine stain (0.2 g I2, 1 g KI in 100 ml H2O) for 15 min, rinsed with water, and subsequently documented.


We would like to thank Drs Carlos del Pozo, Crisanto Gutierrez, and Tai-ping Sun for providing antibodies, Jeff Esch for assistance with the RT-PCR analysis, Min Ni for providing monochromatic light sources, and Frank Harmon and Steve Kay for communicating unpublished results. This work was supported by National Institutes of Health grant GM067203, the McKnight Foundation (W.M.G.) and a Japanese Society for the Promotion of Science Fellowship for Young Scientists (H.I.).