A new temperature-insensitive allele of the Arabidopsis AXR6/CUL1 locus derived from a missense mutation in the C-terminal RBX1 binding region

Abstract

We isolated a new recessive allele at the AUXIN RESISTANT6/CULLIN1 (AXR6/CUL1) locus, axr6–101, from an EMS-mutagenized population of Arabidopsis thaliana, the Landsberg erecta ecotype. axr6–101 is auxin resistant and semi-dwarf similar to the other recessive axr6 mutants. The axr6–101 phenotype is caused by the E716K substitution of the CUL1 protein, which is likely to affect its ability to bind to the C-terminal RING domain of RING-box 1 (RBX1). The previously reported allele of AXR6, cul1–7, is caused by a substitution at T510 that binds to the N-terminal β-strand of RBX1. Although cul1–7 shows temperature-sensitive phenotype, the axr6–101 phenotype is largely unaffected by temperature. axr6–101 may provide an important genetic resource for study of the structure−function relationship of the CUL1 protein.

Abbreviations

AXR

auxin resistant

CUL

cullin

2,4-D

2,4-dichlorophenoxyacetic acid

2,4-DB

2,4-dichlorophenoxybutyric acid

Ler

Landsberg erecta

RBX1

RING-box 1

CULLIN1 (CUL1) is a scaffolding protein of the SCF-type ubiquitin ligase E3,^1,2 and consists of 738 amino acid residues in Arabidopsis. At its N-terminus, it binds to a substrate-recognizing subunit through a SKP1-like protein, ASK1, in Arabidopsis. In addition, it binds to RING-finger protein RING-box 1 (RBX1) at its C-terminus, which recruits the ubiquitin-charged ubiquitin conjugating enzyme E2. There are a large number of substrate-recognizing F-box proteins,³ including the auxin F-box receptors, TIR1/AFB1–4.⁴ Thus, while null alleles of cul1 are not viable,⁵ missense alleles of cul1 show multiple phenotypic alterations, including auxin-insensitivity due to disruption of auxin signaling through the altered SCF^TIR1/AFB1-4 To date, 5 alleles that contain missense mutations at the CUL1 locus have been reported. Interestingly, their phenotypes vary widely. The first reported alleles, auxin resistant (axr) 6–1 and axr6–2, were caused by missense mutations at F111 and are dominant.^6,7 They are seedling lethal because embryogenesis is affected. In contrast, cul1–6, harboring a L115F substitution, is recessive.⁸ Furthermore, axr6–3, featuring a substitution of E159K, is also recessive, but is temperature-sensitive.⁹ Defects observed in axr6–3 mutants are dramatically enhanced at elevated temperatures. For example, the primary stem of axr6–3 is about half as long as the wild type at 20°C; however, it is shorter than 1 cm at 28°C; auxin insensitivity is highly elevated at 28°C at the seedling stage. F111, L115, and E159 are located in the region that binds to ASK1.^1,8,9 To date, cul1–7 is the only reported missense mutation in the C-terminal region that results in a T510I substitution.¹⁰ In the C-terminal region, CUL1 interacts with RBX1.¹ cul1–7 is recessive and temperature-sensitive, like axr6–3. Besides these missense mutations, 2 recessive and viable alleles of CUL1, icu13¹¹ and cul1–494,¹² have been reported, which are caused by missplicing. It is not known whether they are temperature-sensitive or not.

Previously, we isolated a few Arabidopsis mutants that can grow in the presence of 2,4-dichlorophenoxybutyric acid (2,4-DB), but not in the presence of 2,4-dichlorophenoxyacetic acid (2,4-D), to study glyoxysomal fatty acid β-oxidation.¹³ In this screen, we carried out 2 successive screenings with 2,4-DB and 2,4-D. As a result, we found a recessive mutant that was resistant to 2,4-D as well as 2,4-DB. Here we report that this auxin-resistant mutant is a new, recessive allele of AXR6/CUL1, which we named axr6–101. We found that axr6–101 has a E716K substitution in the CUL1 protein, and that its defects are not temperature-sensitive, in contrast to cul1–7. axr6–101 may be an important genetic resource to study the structure−function relationship of the CUL1 protein.

The axr6–101 mutant line was isolated from an EMS-mutagenized Arabidopsis population (ecotype Landsberg erecta (Ler)) as a mutant that germinated and grew on growth medium containing 0.05 μg ml⁻¹ 2,4-D (0.23 μM).¹³ This mutant was recessive: when root growth was examined on media with 0.08 μM 2,4-D, 11 of the 41 F₂ seedlings obtained by a cross between our mutant and Ler exhibited 2,4-D-resistant root growth. For positional cloning, a mapping population of 332 phenotypically mutant plants derived from a cross between our mutant and the Columbia (Col) ecotype was used for linkage analysis using simple sequence length polymorphic (SSLP) and cleaved-amplified polymorphic sequence (CAPS) markers. Our mutation mapped to a 116-kb interval between 2 SSLP markers on top of chromosome 4, which were generated using 2 insertions/deletions of the Cereon Arabidopsis polymorphism collection,¹⁴ CER458376 and CER458010 (Table S1). This interval encompassed 36 annotated genes, one of which was AXR6. Furthermore, no recombination was found at the CAPS marker that was generated using a single nucleotide polymorphism, CER441938, which is located in the 3′ non-coding region of AXR6 (Table S1). We sequenced the AXR6 transcription unit in our mutant, finding a G -> A transition in the last exon that produced the E716K amino acid substitution. This mutation was detected by CAPS using a pair of oligonucleotide primers (5′-GTGGTGATGAAACTCATTGG-3′ and 5′-CCTCTCCAAATAATCTCTGGT-3′) and the restriction enzyme MnlI.

To verify that our mutation was an allele of AXR6, we carried out a complementation test between our mutant and axr6–1.⁶ axr6–1 is semi-dominant and seedling-lethal when homozygous. Therefore, we examined F₁ plants obtained from a cross between our mutant and a heterozygote for axr6–1. Of 17 F₁ plants examined, 8 plants grew normally, 8 were seedling-lethal, and one did not germinate, indicating that our mutant was recessive and did not complement axr6–1. Thus, we named it axr6–101. Taken together, these results show that our mutant is a new recessive allele of AXR6.

Next, we examined the phenotype of axr6–101 using the twice-backcrossed line. Root growth showed 2,4-D resistance with the half maximal inhibitory concentration (IC50) of about 0.3 μM (Fig. 1A). When the IC50 was compared, the auxin resistance of axr6–101 was similar to that of axr6–3,⁹ and was slightly stronger than that of cul1–7 ¹⁰ at 23 – 24°C. axr6–101 had wrinkled rosette leaves like axr6–1 heterozygotes.⁶ Mature axr6–101 plants were ∼20% as tall as Ler at 23°C (Fig. 1B). Flower morphology was altered in axr6–101, as is reported for cul1–6.⁸ The number of floral organs was often reduced: the most common flowers of axr6–101 consisted of 4 sepals, no petals, 4 stamens, and 1 gynoecium, in contrast to Ler flowers showing 4 sepals, 4 petals, 6 stamens, and 1 gynoecium (Table S2). Although fused flower organs such as sepal−petal fusion and petal−anther fusion are observed in cul1–6,⁸ no fusion of floral organs was observed in axr6–101 at any of the temperatures tested. At the seedling stage, the formation of lateral roots was inhibited (Fig. S1). Furthermore, the gravitropic response of the roots was compromised (Fig. S2). However, the defects of the hypocotyls were rather small in terms of gravitropism (Fig. S3) and phototropism (Fig. S4).

Figure 1.

axr6–101 is auxin-resistant and semi-dwarf at the mature stage. (A) Root length was measured in Ler (open circle) or axr6–101 (closed circle) seedlings grown on agar medium containing the indicated concentrations of 2,4-D for 6 d under continuous white-light condition at 23°C. The data represent the mean and SD of 12 seedlings. (B) Ler (left) and axr6–101 (right) plants grown for 5.5 weeks at 20 (left panel), 23 (middle panel), and 28°C (right panel) under continuous white light. Plants were grown in 5.5-cm-square pots.

Then, we raised axr6–101 plants at 20 and 28°C and investigated whether the phenotype was temperature-sensitive. The axr6–101 plants were small at 20°C; however, the size was not reduced further at 28°C (Fig. 1B). In cul1–7, the morphology of the etiolated hypocotyls is more affected at higher temperatures.¹⁰ At 28°C, cul1–7 seedlings exhibit a de-etiolated phenotype in the dark, with short hypocotyls (∼20% as long as wild-type hypocotyls) and open hooks and cotyledons. For etiolated axr6–101 seedlings, growth of hypocotyls and roots was reduced to ∼70 − 80% of the mean size achieved by Ler. However, the relative inhibition of growth was not affected in either organ by an elevated temperature (Fig. 2A). Ler was affected more readily by temperature in the hook opening characteristic. As a result, at 20°C more axr6–101 hypocotyls had open hooks than those of Ler, but at 28°C more open hooks were observed in Ler (Fig. 2B). Further, most etiolated axr6–101 seedlings had closed cotyledons, similar to Ler seedlings, a feature that did not change between 20 and 28°C (Fig. S5). Finally, we investigated the gravitropic growth orientation of roots and hypocotyls in etiolated seedlings. SD of the growth angle can be a measure of growth orientation: a larger SD indicates a more random growth orientation, which reflects weaker gravitropic control of growth direction.¹⁵ SD of the growth angle of axr6–101 roots was clearly larger than that of Ler at both temperatures, indicating gravitropic malfunction in axr6–101 roots. However, the extent of the defects was similar at both 20 and 28°C (Fig. 2C). In the case of the hypocotyl, the SD was similar between Ler and axr6–101 at 20°C; however, a larger SD was observed in axr6–101 at 28°C, suggesting that gravitropic defects are promoted at higher temperatures. In conclusion, the defects of axr6–101 were largely unaffected by elevated temperatures, in contrast to cul1–7 and axr6–3.

Figure 2.

Phenotype of axr6–101 is largely independent of growth temperature. (A) Growth of roots and hypocotyls. The length of each organ was examined in 4-day-old etiolated Ler and axr6–101 seedlings at 20 and 28°C, and the relative length (%) of axr6–101 to Ler is shown in each organ for 30 − 36 seedlings. The relative length of each organ was not significantly different between 20 and 28°C (P > 0.195 in t-test). (B) Hook opening. Hook structure was examined in etiolated hypocotyls grown as above. The data represent the mean and SD of 4 measurements, in which 9 − 22 seedlings were used. (C) Gravitropic growth orientation of the root and hypocotyl. The growth angle was measured in 4-day-old etiolated seedlings. The data represent the mean and SD of 5 − 6 experiments, in which 9 − 41 seedlings were measured. The measurements of each organ from axr6–101 were not significantly different between 20 and 28°C (P > 0.076). The SD of the hypocotyl was larger for axr6–101 than that of Ler at 28°C (P = 0.0036).

axr6–101 has an amino-acid substitution, E716K, in the C-terminal region of CUL1 where it binds to RBX1. RBX1 consists of an N-terminal β-strand and a C-terminal RING domain. Amino acid residues of human CUL1 (HsCUL1) that interact with RBX1 have been identified by the crystallographic study of the RBX1−HsCUL1 complex.¹ Most residues bind to the N-terminal β-strand of RBX1, and the remaining bind to the C-terminal RING domain. E716 of Arabidopsis CUL1 (AtCUL1) corresponds to D754 in HsCUL1, and the adjacent residue of HsCUL1, I755, is involved in binding to the RING domain. Furthermore, E716 is located near K682 of AtCUL1 (corresponding to K720 in HsCUL1), which is the neddylation site of CUL1. Neddylation of CUL1 is necessary for full activity of the SCF ligase E3 and auxin signaling.^2,16 Though E716 may not interact directly with RBX1, it is highly conserved and mostly occupied by E or D in the CUL1 sequences of other organisms. In contrast, cul1–7 is defined by a T510I substitution. T510 corresponds to S541 of HsCUL1 that interacts directly with the N-terminal β-strand of RBX1.^1,10 Thus, axr6–101 is an allele of AtCUL1 that harbors an amino-acid substitution that could affect binding to the RING domain of RBX1. This structural difference may be the reason for the phenotypic difference between axr6–101 and cul1–7. However, axr6–101 was obtained on the Ler genetic background,¹⁷ whereas cul1–7 and most of other axr6/cul1 alleles have a Col background. Therefore, the different genetic background should be considered when comparing the phenotype of axr6–101 with that of other alleles. Further study of axr6–101 may result in a deeper understanding of CUL1 function in auxin signaling.

Funding

This work was supported in part by a Grant-in-Aid for Scientific Research from Japan Society for the Promotion of Science to M.H. (15K07117), and Grants-in-Aid from Ministry of Education, Culture, Sports, Science and Technology, Japan to K.T.Y. (19060008).

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgment

We thank the Arabidopsis Biological Resource Center (Ohio State University) for axr6–1 seeds.

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website.