A genetic screen for mutants defective in IAA1-LUC degradation in Arabidopsis thaliana reveals an important requirement for TOPOISOMERASE6B in auxin physiology

Abstract

Many plant growth and developmental processes are modulated by the hormone auxin. Auxin-modulated proteolysis of Aux/IAAs, a family of transcriptional repressors, represents a major mode of auxin action. Auxin facilitates the interaction of Aux/IAAs with TIR1/AFB F-box proteins, promoting their ubiquitination by the SCF^TIR1/AFB ubiquitin E3 ligase leading to subsequent degradation by the 26S proteasome. To identify new genes regulating Aux/IAA proteolysis in Arabidopsis thaliana, we took a genetic approach, identifying individuals with altered degradation of an IAA1-luciferase fusion protein (IAA1-LUC). A mutant with 2-fold slower IAA1-LUC degradation rate compared with wild-type was isolated. Positional cloning identified the mutant as an allele of TOPOISOMERASE6B, named top6b-7. TOP6B encodes a subunit of a plant and archea-specific enzyme regulating endoreduplication, DNA damage repair and transcription in plants. T-DNA insertion alleles (top6b-8 and top6b-9) were also analyzed. top6b-7 seedlings are less sensitive to exogenous auxin than wild-type siblings in primary root growth assays, and experiments with DR5:GUS. Additionally, top6b-7 seedlings have a 40% reduction in the amount of endogenous IAA. These data suggest that increased IAA1-LUC half-life in top6b-7 probably results from a combination of both lower endogenous IAA levels and reduced sensitivity to auxin.

Keywords:

• Arabidopsis thaliana
• Aux/IAA
• auxin
• DR5:GUS
• indole-3-acetic acid
• luciferase fusions
• protein degradation
• TOPOISOMERASE6B

Abbreviations

LUC	=	luciferase
top6b	=	TOPOISOMERASE6B
2	=	4-D, 2, 4-dichlorophenoxyacetic acid
IAA1-LUC	=	IAA1-luciferase fusion protein
IAA	=	indole-3-acetic acid
UPS	=	ubiquitin proteasome system

Introduction

Auxins, a class of phytohormone typified by the major endogenous auxin indole-3-acetic acid (IAA), regulate many aspects of plant growth and development. Polar localization of auxin transporters in plasma membranes and localized auxin biosynthesis together establish auxin gradients within plant organs and tissues.¹ Such gradients have developmental consequences for patterning and phyllotaxy of new leaf primordia within the shoot apical meristem, lateral root initiation from the pericycle, and the overall establishment of plant organ polarity.² How such gradients are established and maintained over the developmental history of the plant and how cells sense and transduce differential auxin concentrations have been areas of intense study for several decades. By taking forward genetic, molecular genetic, and biochemical approaches many molecular components of auxin transport and signal transduction have been discovered.³

Exogenous application of auxin rapidly modulates the transcription of a suite of genes, and this regulation is mediated through two families of transcription factors.⁴ Auxin/IAA (Aux/IAA) proteins act as short-lived transcriptional regulators, of which there are 29 members in Arabidopsis thaliana.⁵ These nuclear-localized proteins interact with and affect the activity of a family of transcription factors called Auxin Response Factors (ARFs), whose gene family size is 23 members in Arabidopsis.^6,7 The current model of auxin-mediated transcription proposes that when relieved from interaction with Aux/IAA proteins, ARFs modulate transcription of primary response genes that function in establishment and maintenance of patterning in plant organs.⁵ A recent report describing the interaction of an Aux/IAA with a heat shock transcription factor in sunflower embryos suggests that Aux/IAA may regulate ARF independent processes, at least for one auxin-regulated response.⁸

Our understanding of how auxin changes the activity of Aux/IAA proteins developed from the observation that auxin application stimulates rapid ubiquitin-mediated degradation of several endogenous Aux/IAA proteins or Aux/IAA fusion proteins in transgenic plants.^9-11 This increased Aux/IAA degradation rate is proposed to reduce the extent of Aux/IAA-ARF interactions, affecting transcription mediated by ARF proteins. However, not all Aux/IAA family members have equivalent degradation rates under high auxin conditions or show auxin-regulated proteolysis,¹² suggesting that auxin's effect on composition and concentration of Aux/IAA proteins is complex. Auxin modulates Aux/IAA ubiquitination by directly facilitating interaction of Aux/IAA proteins with the TIR1 family of F-box proteins and conversely, Aux/IAA proteins stabilize auxin's interaction with the F box protein.^13-15 F-box proteins are the substrate specificity subunits of SCF-type ubiquitin E3 ligases, a multiple subunit complex that interacts with the substrate and an E2 carrying activated ubiquitin to facilitate substrate ubiquitination. Thus, auxin-dependent enhancement of Aux/IAA interaction with its ubiquitin ligase results in Aux/IAA ubiquitination and subsequent degradation by the proteasome.¹⁶ Recent work on the affinity of the TIR1/AFB proteins with multiple IAA proteins in vitro indicate that there are quantitative differences in their interactions, providing more evidence in support of a complex model for auxin transcriptional control.¹⁷

In an effort to identify other trans-acting factors affecting Aux/IAA degradation, we treated a previously characterized transgenic Arabidopsis line expressing IAA1-LUC^{11,12, 18-20} with a chemical mutagen (EMS) and screened for progeny with higher IAA1-LUC steady-state levels compared with the progenitor line.²¹ We report here on one mutant recovered in this screen that has a 2-fold slower IAA1-LUC degradation rate. Surprisingly, this mutation mapped to the TOPOISOMERASE6B (TOP6B, At3g20780) locus.

TOP6B encodes the B subunit of the plant homolog of the archeabacterial DNA topoisomerase VI (Topo VI), a member of the B subfamily of type II topoisomerases. Type II topoisomerases catalyze ATP-dependent double-strand cleavage and decatenation, relieving tangling and supercoiling of DNA double strands that occur during replication, transcription, and recombination [for a review see refs.^22,23]. Archeabacterial Topo VI, and presumably others in the same subfamily, is an A₂B₂ heterotetramer.²⁴ The A subunits catalyze the double-stranded DNA cleavage reaction.^25,26 The B subunits contain a GHKL-type (gyrase, Hsp90, histidine kinase, MutL) ATPase domain,^24,27 a transducer domain, a helix–turn–helix domain, and a C-terminal domain.^26,28 The transducer domain links the B subunit to the A subunit and also contains a critical lysyl residue that contacts the γ-phosphate of ATP bound in the GHKL-domain.^26,28,29 Direct B-B subunit interaction occurs upon nucleotide binding and is thought to trap the DNA duplexes in the enzyme prior to strand cleavage and passage.

While type IIA topoisomerases can be found across all domains of life including some viruses, the topoisomerase IIB subfamily containing Topo VI homologs was initially thought to be archea-specific. The A subunit is present in eubacteria and metazoa as sporulation-specific protein 11 (SPO11); however, these organisms do not have any obvious B subunit homologs. Surprisingly, genes encoding both A and B subunits were identified in plants. Three and five genes encoding homologs of SPO11 were identified in Arabidopsis and rice, respectively, as well as a single copy of the B subunit, TOP6B, in these organisms.^30-32 AtTOP6B interacted with both AtSPO11–2 and AtSPO11–3 but not AtSPO11–1 in yeast-2-hybrid assay, suggesting that plants, unlike other eukaryotes, possess a topoisomerase VI composed of AtTOP6B and SPO11–2/SPO11–3.³² Genetic evidence indicates that SPO11–1 and SPO11–2 function in meiotic recombination, while SPO11–3 appears to have a distinct function.³³ A number of mutants in these subunits were later identified by forward and reverse genetics giving some insight into the function of TOP6 in plants. Mutants in the A and B subunits were identified in screens searching for root hairless (RHL),^34-36 short dark-grown hypocotyls (HYP),^35,36 and brassinosteroid insensitive (BIN) mutants.³⁷ SPO11–3/BIN5/RHL2 and RHL3/BIN3/HYP6 encode the TOP6A and TOP6B subunits, respectively. Loss of function mutations in either subunit result in a severe pleiotropic phenotype: extreme dwarfism, lack of both root hairs and trichomes, small cells, photomorphogenesis in the dark, and brassinosteroid insensitivity. TOP6 mutants are defective in endoreduplication, as TOP6 appears to be required for production of a DNA content beyond 8C.^35,36,38

In addition to the core A₂B₂ subunits conserved in the TOP6 subclass of topoisomerases, AtTOP6 appears to include two additional plant-specific subunits required for enzyme activity. These subunits were identified from genetic screens as BIN4/MID1 and RHL1.^{35, 39-41} The phenotypes of bin4/mid, rhl1, top6a/spo11–3 and top6b are indistinguishable from one another, suggesting that together they constitute a functional TOP6 in plants.

All together, the current data indicate that the phenotype of TOP6 loss of function plants is complex and results from a number of defects, including defects in endoreduplication, activation of a DNA damage response pathway, a reduction in heterochromatin, and the mis-expression of many genes. The characterization of bin4/mid revealed that TOP6B enzyme complex is not only required for endoreduplication, but also for heterochromatin formation during interphase and loss of TOP6 results in the activation of DNA damage response pathway.^40,42 A microarray analysis of both top6b/bin3 and top6a/bin5 mutants indicated mis-regulation of a large number of genes, including those involved in brassinosteroid and auxin signaling pathways.³⁷ Here, we report the isolation and characterization of a new TOP6B allele based on slowed IAA1-LUC degradation, and demonstrate that loss of TOP6 activity results in an auxin deficiency that affects responses to exogenous auxin.

Results

A genetic screen for mutants in Arabidopsis thaliana with altered IAA1-LUC degradation uncovers a new TOPOISOMERASE6B allele

To uncover novel regulators of Aux/IAA proteolysis, we developed a genetic screen to identify Arabidopsis thaliana mutants with altered Aux/IAA degradation by measuring the abundance and degradation rate of an IAA1-LUC fusion protein expressed from a transgene in individuals from a mutagenized population.²¹ Our hypothesis was that seedlings with higher steady-state LUC activity than the progenitor line could have increased LUC activity from slower proteolysis of the fusion protein. Degradation of specific endogenous Aux/IAA proteins is difficult to measure due to their low abundance and presence of many related family members. IAA-LUC fusion proteins allow sensitive and quantitative determination of proteolytic rates. We previously showed that multiple Aux/IAA-LUC fusion proteins display basal degradation rates and auxin-modulated and proteasome inhibitor-sensitive degradation rates in vivo identical to tested endogenous Aux/IAA proteins.^11,12,18,19 Using this screen, we reported on the isolation of a novel viable allele of CULLIN1, a subunit of the E3 ligase responsible for Aux/IAA ubiquitination, demonstrating that the screen can successfully identify genes regulating Aux/IAA degradation.²¹

Additional screening identified a distinct M₂ seedling with a higher IAA1-LUC steady-state level than other seedlings on the plate (Fig. S1A, arrow). F1 backcross to the progenitor line gave normal LUC activity, indicating the mutation was recessive. F₂ seedlings from the first backcross (BC₁) segregated 3:1 for the mutant morphological phenotype, and all mutant seedlings had higher IAA1-LUC steady-state levels compared with wild type siblings indicating that the high LUC activity and the phenotype (see below) co-segregated (Fig. 1A). To confirm that this mutant was a bona fide degradation mutant, we measured the IAA1-LUC degradation rate in multiple individuals from a BC₂F₂ population.²¹ The half-life of IAA1-LUC in mutant seedlings averaged 34.5 min, statistically slower than the half-life (17.4 min) determined for IAA1-LUC in wild-type siblings (Fig. 1B), confirming this mutant has impaired IAA1-LUC degradation.

Figure 1. top6b-7 has higher levels of IAA1-LUC and a slower IAA1-LUC degradation rate. (A) Confirmation of higher IAA1-LUC steady-state levels and co-segregation with the mutant phenotype. An F₂ population of seeds from a backcross to the progenitor transgenic IAA1-LUC line was re-screened for the accumulation IAA1-LUC. Seedlings with the top6b phenotype (see Fig. 3) were separated from wild type seedlings and imaged after pre-incubation with the substrate luciferin. (B) IAA1-LUC degradation rates in top6b-7 and wild type sibling seedlings. Using a single-seedling degradation assay²¹ IAA1-LUC half-lives (t_1/2) were determined in 7 d-old individual mutant and wild-type siblings in an BC₂F₂ segregating population. In top6b-7 seedlings IAA1-LUC t_1/2 = 34.5 min, n = 59. In wild-type siblings IAA1-LUC t_1/2 = 17.4 min, n = 180. Bars are standard error. Half-lives are statistically different by a Student's t test (P = 1.16*10⁻¹⁴, α = 0.05).

To positionally clone the mutant gene, we made a mapping population by out-crossing an M₃ plant to Landsberg erecta (Ler). Mutant F₂ plants from this cross were used as the mapping population. Additionally, we tested this population for LUC activity, and LUC activity segregated independently from the mutant phenotype indicating that the IAA1-LUC transgene was not linked to the mutation causing the mutant phenotype. Additionally, the 2-fold increase in half-life of IAA1-LUC only occurred in seedlings with a mutant phenotype, indicating that the longer half-life is not due to a cis mutation in the IAA1-LUC transgene. Bulked segregant analysis^43,44 linked the mutation to markers on the long arm of chromosome III (Fig. S2). Using a small mapping population (Fig. 2A, Experiment 1), we were able to map the mutation between SNPs PERL0464591 and PERL0476408, and using a larger mapping population, we mapped the mutation between SNPs PERL0472202 and PERL0473223 (Fig. 2A). This final mapping interval spanned ∼138 kb, and contained 35 open reading frames (www.arabidopsis.org).

Figure 2. Mapping of top6b-7 and identification of the mutation. (A) Positional cloning of top6b-7. For experiment 1, a small mapping population was used to roughly map, and a larger mapping population was used for experiment 2 fine mapping. Markers for mapping are shown with red tick marks. The mutation mapped between PERL0472202 and PERL0473223 SNPS. (B) Identification of mutation in top6b-7. Genomic DNA for At3g20780 was sequenced from 340 bp 5′ of START codon to 270 bp 3′ of the STOP codon. Sequence is shown for exon 1, intron 1, and exon 2 with intron sequence denoted by lowercase type. The G to A mutation in top6b-7 is bold underlined type. (C) Amino acid substitution in top6b-7. The first 100 amino acids of the TOP6B primary sequence are shown. The E36K substitution in top6b-7 is bold underlined type. (D) Multiple sequence alignment of TOP6B N-terminus. TOP6B sequences were retrieved from NCBI, by protein BLAST search using the Arabidopsis protein. Sequences were aligned using ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/), and the sequences near the N-terminus are shown. An arrow indicates the glutamate residue changed to a lysine residue in top6b-7. Symbols for organisms are: Pp, Physcomitrella patens; Sm, Selaginella moellendorffii; Vv, Vitis vinifera; Rc, Ricinus communis; At, Arabidopsis thaliana; Sb, Sorghum bicolor; Os, Oryza sativa; Ps, Picea sitchensis; Cr, Chlamydomonas reinhardtii; Ta, Thermosphaera aggregans and; Si, Sulfolobus islandicus. Ta and Si are included to show divergence of plant TOP6B proteins from those of Archaea. (E) Gene structure of TOP6B and allele locations used in this study. The point mutation in top6b-7 is denoted by a small, black triangle. T-DNA insertions are indicated with large, black triangles with allele designated above them. Exons, UTRs, introns, and deletions are indicated by black boxes, white boxes, lines, and a small triangle, respectively. Scale bars represent 100 bp. Graphics were generated with Exon-Intron Graphic Maker (http://wormweb.org/exonintron).

Next, we selected candidate genes in this region to sequence based on gene ontology annotations associated with the ubiquitin-proteasome system and also based on genes with similar mutant phenotypes. Mutations in TOPOISOMERASE6B (TOP6B, At3g20780.1) had been previously reported by several groups to result in extreme dwarfism,^34-38 similar to the phenotype of the mutant we recovered. We sequenced TOP6B in our mutant and found a G to A substitution in exon 2 (Fig. 2B) predicted to result in an E36K substitution (Fig. 2C). This residue is conserved among the plants orthologs examined (Fig. 2D) and lies in a highly conserved region adjacent to the conserved “Motif B I” of the GHKL-type ATP binding domain.⁴⁵ Curiously, it is not conserved in the archeal proteins (Fig. 2D).

Mutations in TOP6B Cause Severe Growth Defects

Mutations in TOP6B had been identified in previous genetic screens and in studies to characterize TOP6B in plants. These characterized alleles and corresponding references are summarized in Table 1. In order to assign an allele designation for the mutation identified in this study, we assigned the other mutants a top6b allele number based on the order of publication (Table 1). We designated our allele top6b-7, and obtained two additional previously uncharacterized T-DNA alleles, which we named top6b-8 and top6b-9 (Table 1, Fig. 2E). To ensure that the T-DNA insertions in top6b-8 and top6b-9 were correctly annotated, we determined the T-DNA insertion site in these alleles by sequencing the T-DNA specific genotyping PCR fragment with the primer specific to the T-DNA left border. A T-DNA PCR product for top6b-8 could only be obtained with the reverse gene-specific primer suggesting only one insertion at this site, and sequencing this PCR product placed the insertion in exon 12 (Fig. 2E and Fig. S2). For top6b-9, a PCR product could be obtained using either the forward or the reverse gene-specific primer with the T-DNA primer, suggesting the presence of two tandem inverted T-DNA insertions. Sequencing these products revealed the junction of one of the left-borders in exon 4 and the other in intron 4 (Fig. 2E and Fig. S2) with 139 bases of TOP6B sequence separating the left-border junctions. The presence of the sequence between the insertions, however, is uncertain as the right-border junctions were not sequenced.

Table 1. Alleles of TOP6B

Name^a	Allele number^b	Type of Mutation	Ecotype	Reference
rhl3–1	top6b-1	EMS	Col	Schneider et al. 1997; Sugimoto-Shirasu et al. 2005
rhl3–2	top6b-2	Fast Neutron	Ler	Schneider et al. 1997
bin3–1	top6b-3	EMS	Col	Yin et al. 2002
bin3–2	top6b-4	T-DNA	Col	Yin et al. 2002
top6b	top6b-5	T-DNA	Ws	Hartung et al. 2002
hyp6	top6b-6	EMS	Col	Sugimoto-Shirasu et al. 2002
top6b-7	top6b-7	EMS	Col	This Study
salk_024455c	top6b-8	T-DNA	Col	This Study
salk_140704	top6b-9	T-DNA	Col	This Study

^a Name refers to the published alleles of TOP6B. ^bAll alleles were assigned a top6b allele number in order of publication date.

The phenotypes of top6b-7, -8, and -9 plants are all very similar. Plants homozygous for these three alleles are phenotypically indistinguishable from one another when germinated on agar plates. top6b seedlings are small compared with their wild-type siblings and have pale, epinastic, unexpanded cotyledons (Fig. 3A). By 5 weeks, the growth defects in top6b plants are even more striking. The plants are extremely dwarfed with rosettes measuring less than one centimeter in diameter (Fig. 3B). All three alleles have bolted and flowered by 10 weeks of growth (Fig. 3C). The inflorescences of top6b plants are very short compared with those of wild-type plants, but flowers are fertile and produce siliques with a few seeds.

Figure 3. Morphological phenotype of top6b mutants. (A) top6b seedling phenotype. A population of seeds segregating for each allele was plated, cold-treated for 24 h at 4°C, and grown under continuous light for 10 d. Plants with the homozygous phenotype are circled. Seedlings not circled are wild-type segregants. (B) top6b mutant phenotype after 5 wk growth. Seeds segregating for each allele were sown directly on soil, cold-treated for 48 h at 4 °C, then grown 7 d at 16°C under constant. Seedling were selected by phenotype, and transplanted to individual pots for continued growth at 16°C. Scale bars are 1 cm. (C) top6b phenotype after 10 wk of growth. Plants in (B) were grown for an additional 5 wk as described above. Scale bars are 1 cm.

Top6b-7 is Allelic to T-DNA Insertion Alleles of TOP6B

To confirm that the mutation in top6b-7 was responsible for the observed phenotypes, we crossed top6b-7 to the T-DNA alleles for an allelism test. All three alleles segregate 3:1 for a wild-type to mutant phenotypic ratio (wt:m) when a heterozygous parent is selfed, indicating that the three alleles are recessive (Table 2). We crossed TOP6B/top6b-7 IAA1-LUC hmz plants in both directions to TOP6B/top6b-8 and TOP6B/top6b-9 plants and analyzed segregation of mutant phenotypes in the F₁ progeny from these crosses. If the mutations are allelic, then the F₁ progeny from the crosses should segregate 3wt:1 min. If not allelic, then all the F₁ progeny should have a wild-type phenotype. The F₁ progeny from all crosses tested segregated 3:1, indicated by a Chi-square test (Table 2), demonstrating that the mutation in top6b-7 is responsible for the observed morphological phenotype. To confirm that the F₁ mutant individuals were heteroallelic and did not result from a self, several were genotyped for top6b-7 and the respective T-DNA allele and all tested were heteroallelic (data not shown).

Table 2. Allelism tests of top6b-7 morphological phenotype

cross^a	wt^b (observed)	mutant^b (observed)	χ^2 3:1^c	p-value^c
TOP6B/top6b-7♀ TOP6B/top6b-8♂	126	39	0.164	0.6858
TOP6B/top6b-8♀ TOP6B/top6b-7♂	66	26	0.522	0.701
TOP6B/top6b-7♀ TOP6B/top6b-9♂	35	14	0.333	0.5637
TOP6B/top6b-9♀ TOP6B/top6b-7♂	39	13	0	1
TOP6B/top6b-7⊗	172	71	2.306	0.1289
TOP6B/top6b-8⊗	195	48	3.568	0.0589
TOP6B/top6b-9⊗	181	53	0.689	0.4063

^a TOP6B/top6b-7 plants were crossed in both directions to TOP6B/top6b-8 and TOP6B/top6b-9 plants. ^bSeedlings from 1 to 3 F₁ siliques were scored for segregation of the top6b phenotype. ^cAll crosses and selfings fit a 3:1 wild-type:mutant segregation ratio as assessed by a Chi-square goodness-of-fit test, α = 0.05.

To verify that the IAA1-LUC degradation defect is dependent on loss of TOP6B, the IAA1-LUC degradation rate was determined in F₁ heteroallelic seedlings. If the degradation defect resulted from a mutation in a gene closely linked to TOP6B in top6b-7 and not from top6b-7 itself, then given the recessive nature of the decreased LUC degradation rate, none of the resulting heteroalleleic F₁ plants from the allelism test crosses would have the longer half-life characteristic of top6b-7. The IAA1-LUC half-lives in mutant F₁ individuals from all crosses were very similar to that calculated for top6b-7 mutants (Table 3), indicating that a defect in TOP6B is responsible for the longer IAA1-LUC half-life. Longer half-lives were only observed in seedlings with mutant phenotypes. F₁ wild-type seedlings had a shorter half-life similar to that previously calculated for IAA1-LUC in wild-type Col and similar to that calculated in the F₂ wild-type siblings from the self of TOP6B/top6b-7 BC₂F₁ (Table 3). All together, these data indicated that the mutation in top6b-7 is responsible for both the growth defects and for slower IAA1-LUC degradation observed in this line.

Table 3. Allelism test for IAA1-LUC degradation defects

cross	IAA1-LUC half-life ± sd in mutant F1 individuals^a	IAA1-LUC half-life ± sd in wild-type F1 individuals^a
TOP6B/top6b-7♀ TOP6B/top6b-8♂ Silique 1	34.5 ± 5.6 min, n = 10 P = 0.00020	nd
TOP6B/top6b-7♀ TOP6B/top6b-8♂Silique 2	28.1 ± 5.9 min, n = 11 P = 7.7•10^−5	nd
TOP6B/top6b-8♀ TOP6B/top6b-7♂	34.6 ± 5.6 min, n = 6 P = 0.00012	14.75 ± 2.3 min, n = 12 P = 0.0080
TOP6B/top6b-9♀ TOP6B/top6b-7♂	34.6 ± 3.0 min, n = 12 P = 0.00031	nd
TOP6B/top6b-7⊗ BC2 F1	32.1 ± 5.7 min, n = 12 P = 9.4•10^−7	17.4 ± 2.2 min, n = 12

^a Each half-life was compared with the half-life of IAA1-LUC in wild-type siblings originating from the self of TOP6B/top6b-7 by a Student's t test, and the calculated p-value for each comparison is indicated (α = 0.05, 0.0083 with a Bonferroni correction for multiple comparisons). nd, not determined.

Top6b-7 has Altered Auxin Physiology

Because Aux/IAA degradation rates are affected by auxin levels and auxin responsiveness, we undertook experiments to examine if top6b-7 mutants had altered auxin physiology/responses. A classic and well-documented response to exogenous auxin is inhibition of primary root growth in seedlings. We tested the dose-responsiveness of top6b-7 primary roots to growth inhibition by application of the synthetic auxin, 2,4-dichlorophenoxyacetic acid (2,4-D) (Fig. 4). The auxin response of roots from a known auxin-resistant mutant, axr1–30 ^21,46 grown at the same time was included. Plants were plated and grown continuously on the indicated concentration of 2,4-D, and primary root lengths were measured after 7 d of growth. As expected, axr1–30 roots are resistant to root growth inhibition at these concentrations of 2,4-D. Wild-type Col-0 roots also respond as predicted, with primary roots growing less on increasing concentrations of 2,4-D. The Col-0 line is statistically different from axr1–30 by a factorial ANOVA (p-value <0.001), and by comparing slopes axr1–30 is 97% less responsive to auxin than Col-0. The auxin response for top6b-7 is statistically different for Col-0 by a factorial ANOVA (p-value <0.001). By comparing the slope of the dose response curve in this analysis, top6b-7 is 35% less responsive to 2,4-D than Col-0, indicating that response to exogenous auxin is affected in top6b-7 seedlings.

Figure 4. Root growth inhibition in top6b-7 mutants with auxin treatment. Seeds were plated and grown on solid GM agar plates supplemented with the indicated concentration of 2,4-D for 7 d at 20°C with continuous lighting. The results are the combined data from three independent experiments. Values are mean ± 1 SD 59 ≤ n ≤ 68. The equations for linear regression are: y = −4.11(x) + 21.02, R² = 0.0172; y = −129.32(x) + 16.44, R² = 0.964; y = −83.98(x) + 7.91, R² = 0.940 for axr1–30, Col, and top6b-7, respectively.

To determine whether auxin levels were affected in top6b-7, which could affect auxin response, we used LC-MS/MS to measure IAA levels from a segregating population of BC₂ F₂ seedlings. top6b-7 and wild-type seedlings were pooled into three samples each for two independent growth experiments, totaling six samples for each genotype. Wild-type seedlings have more IAA than their top6b-7 siblings in every sample measured (Fig. 5A). Combining the six replicates, IAA levels averaged 10.26 and 6.26 ng/g FW in wild-type and top6b-7 seedlings, respectively (Fig. 5B), and these measurements are statistically different by a Student's t test (p-value = 5.8·10⁻⁵, α = 0.05). IAA levels are therefore reduced approximately 40% in top6–7 seedlings.

Figure 5. IAA levels in top6b-7 mutants. Seeds segregating for top6b-7, were plated on GM, cold treated for 48 h at 4°C, and grown under constant light (50 μmol·m⁻¹·s⁻²) at 20°C. Mutants and wild-type siblings were collected and sent for IAA measurement using LC-MS/MS. (A) IAA measurements are represented from 2 independent growth experiments. For each growth experiment, 3 different biological samples (S1-S3) for each phenotypic class were used for measurements. (B) The mean for the 6 replicates shown in (A) for each phenotypic class is shown, and are 10.39 and 6.26 ng IAA/g FW for top6b-7 and wild-type respectively. Bars are 1 sd. Averages are statistically different by a Student's t-Test (P = 5.84031*10⁻⁵, α = 0.05).

To further investigate endogenous auxin levels and auxin responsiveness in top6b-7, we crossed TOP6B/top6b-7 to a DR5:GUS reporter line (a minimal 35S promoter with synthetic auxin response elements driving expression of β-glucuronidase)⁴⁷ which has been shown to be an excellent sensor of endogenous auxin production in cells and tissues and also is very responsive to the application of exogenous auxin. We examined GUS staining patterns in a population of DR5:GUS seedlings segregating for top6b-7 with and without treatment with exogenous auxin (Fig. 6). Overall, the GUS staining pattern in top6b-7 mock-treated is similar to that of its wild-type siblings, with high GUS activity concentrated in the root apical meristem (white arrowheads), sites of lateral root emergence, and at the cotyledon apex. GUS activity, however, is reduced in top6b-7 shoot apical meristem (black arrowheads), in newly emerging leaves, and in the veins and peripheral cells of cotyledons (insets) (Fig. 6, row 1).

Figure 6. DR5:GUS expression and responsiveness to 2,4-D in top6b-7. Seeds from a TOP6B/top6b-7 DR5:GUS plants were grown for 10 d on solid GM plates at 22°C, under constant light. Seedlings were then transferred to liquid GM and treated with either solvent or 50 μM 2,4-D for 24 h at 22°C, under constant light. Seedlings were stained for GUS activity for 24 h and 1 h for mock and 2,4-D samples respectively. Two representative seedlings are shown for each sample. Black arrowheads denote shoot apices, and white arrowheads denote the root apical meristems. Insets are enlargements of cotyledons to show GUS staining pattern.

To examine DR5:GUS responses to exogenous auxin, we treated seedlings with 2,4-D and stained for GUS activity (Fig. 6, row 2). top6b-7 roots and true leaves are less sensitive to exogenous auxin in inducing GUS activity. All cells of wild-type roots have a continuous staining pattern and are darkly stained; however, under the same staining conditions, the roots of top6b-7 consist of patches of cells that do not stain or stain much less. Additionally, the true leaves of top6b-7 appear less sensitive to exogenous auxin than those of wild-type, with less intense staining. These results support our previous observations and are consistent with lowered endogenous auxin levels affecting responsiveness to exogenous auxin.

Discussion

Proteolysis by the ubiquitin proteasome system (UPS) is a major mode of regulation for plant growth, development, reproduction and responses to the environment. Genome sequencing and annotation efforts have revealed that components of the UPS are greatly expanded in plants compared with other eukaryotes, and in the model plant Arabidopsis thaliana an estimated 6% of genes are predicted to encode UPS components [for a review see ref. ⁴⁸]. Well-characterized substrates of the UPS in plants are the Aux/IAA proteins, whose degradation is facilitated by, and required for, response to the phytohormone auxin. The current model is that Aux/IAAs are ubiquitinated by the E3 ubiquitin ligase SCF^TIR1/AFB, and this modification targets them for degradation by the proteasome. Interaction with SCF^TIR1/AFB requires the binding of auxin in the presence of both TIR1/AFB and the Aux/IAA to facilitate/stabilize the interaction. Mutations in a specific region of the Aux/IAAs, termed the degron, specifically core residues (GWPPV/L/I) of Aux/IAAs, slows proteolysis likely by disrupting their interactions with auxin and TIR1/AFB.^9,49,50 The core degron, however, does not confer the same degradation rate on all Aux/IAA family members,¹² leading to the hypothesis that other unknown factors might be modulating their degradation. Toward the goal of understanding the requirements for Aux/IAA degradation in Arabidopsis, we undertook a forward genetic screen to find mutants with slower degradation of an IAA1-LUC fusion protein.²¹ Here we describe one mutant, top6b-7, isolated from this screen and present evidence that links topoisomerase VI to IAA levels and response.

Previous characterization of topoisomerase VI mutants in Arabidopsis revealed that plant growth requires the activity of this topoisomerase. One of the prominent molecular defects in these mutants is a block of endoreduplication. Endoreduplication, an alternative cell cycle, results in increased ploidy (called endoploidy) by replication of nuclear DNA without subsequent nuclear and cell divisions.^39,51,52 Plant cells undergo extensive endoreduplication as they differentiate and increase in size, and the extent of endoreduplication depends on the species and cell type. In Arabidopsis, cells range from 2C to 32C.⁵³ From analysis of Arabidopsis top6 mutants, this topoisomerase is thought to be specific to endoreduplication and the resolution of topological problems. The loss of TOP6 activity could make it physically impossible to further replicate DNA past 8C, with tangling becoming too extensive for replication to proceed, triggering the DNA damage response pathway that has been reported to be activated in top6 plants.^40,42

The molecular mechanisms of how increased ploidy contributes to cellular growth/expansion remain to be resolved, but it is generally believed that an increase in ploidy allows for greater metabolic activity of larger cells. This idea is supported by evidence that plant biotrophs induce endoreduplication at sites of nutrient exchange to boost the metabolic output of infected cells [for a review see ref. ⁵⁴]. Many plant hormones including ethylene, gibberellic acid, auxins, brassinosteroids, and cytokinins regulate growth and promote cellular expansion, providing an explanation why mutants in components of the TOP6 complex were isolated in a screen for brassinosteroid insensitive mutants, and shown to be slightly resistant to exogenous brassinosteroid in hypocotyl elongation assays.^37,42 Microarray analysis of bin3/top6b and bin5/top6a also revealed these mutants are defective in brassinosteroid-induced gene expression and that many components of both auxin and brassinosteroid signaling pathways are under-expressed.³⁷

Brassinosteroid and auxin signaling pathways are interdependent to promote cell elongation, and both pathways must be functional in order for either hormone to induce growth.⁵⁵ Specifically, ARF2 is required for brassinosteroid responses.⁵⁶ Brassinosteroid application has been reported to stimulate expression of DR5:GUS without altering endogenous IAA levels⁵⁷ indicating the gene targets of the two hormones’ signaling pathways have significant overlap, which confounds interpretation of auxin responsiveness and DR5:GUS expression in top6b-7. DR5:GUS expression in the roots of top6b-7 appeared to be less responsive to exogenous 2,4-D with patches of cells having very light staining to no staining. This result might be explained by mis-regulation of auxin transporters in top6b-7 roots, as brassinosteroid appears to influence expression of the PIN-FORMED (PIN) auxin transporters.⁵⁸ However, slower IAA1-LUC degradation in the mutant reveals an additional auxin-specific defect, because only auxin stimulates Aux/IAA degradation.^12,55

Examination of endogenous IAA levels in top6b-7 revealed a nearly 40% reduction compared with wild-type siblings, indicating that slower IAA1-LUC degradation could in large part result from lower auxin levels. Why IAA levels in top6b-7 are reduced remains an open question, but microarray analysis of bin3/top6b and bin5/top6a/spo1–3 indicated that YUCCA3 expression was ∼3–5-fold reduced in those lines,³⁷ suggesting an effect on auxin synthesis. IAA biosynthetic pathways are complex, not fully defined, and may be tryptophan (Trp)-dependent or –independent.⁵⁹ However, a major pathway for IAA synthesis is a two-step process from Trp to IAA via indole-3-pyruvate, with YUCCA family of proteins catalyzing the last step⁶⁰ [for a review see refs.^{59, 61, 62}]. The YUCCA-dependent step is thought to be rate limiting for IAA synthesis since YUCCA overexpression lines have increased IAA and conjugated IAA, while overexpression of the penultimate enzyme has little effect on auxin levels.^60,63 However, the role of YUCCA3 specifically in regulating IAA levels is yet not known.

The phenotype of top6b is complex, and not simply a result of lower free IAA levels, as the growth defects cannot be rescued by application of auxin or brassinosteroid.³⁷ The increase in IAA1-LUC half-life in the mutant, most probably results from a combination of both a decreased endogenous IAA level and reduced sensitivity to auxin. Root growth inhibition assays with 2,4-D (Fig. 4) demonstrate reduced sensitivity, which is also supported by the DR5:GUS staining in auxin treated top6b seedlings (Fig. 6 row 2). Basal DR5:GUS staining is also reduced in top6b seedlings, a result that could be explained by reduced sensitivity and/or reduced endogenous IAA levels. We, therefore, are left to conclude that a reduction in both sensitivity and IAA levels both contribute to the stabilization of IAA1-LUC in top6b-7. These mutants also have an activated DNA damage check-point⁴² and TOP6b also appears to function in an ROS response pathway binding to promoters of ROS-activated genes,⁶⁴ suggesting direct regulation of gene activity and processes that arrest growth are activated in top6 mutants.

In conclusion, the data presented here demonstrate for the first time the perturbation of auxin physiology in plants with disrupted topoisomerase VI activity as evidenced by 1) slower degradation of IAA1-LUC, 2) decreased levels of endogenous IAA, 3) reduced basal expression of DR5:GUS, and 4) reduced response to exogenous auxin in both root elongation assays and using the auxin responsive promoter DR5:GUS in top6b-7. The mechanism of IAA reduction remains an unanswered question, but it is tempting to speculate that reduced endoploidy of top6 cells decreases expression of IAA biosynthesis genes from either a reduction in copy number or by altering their transcriptional regulation. Alternatively, the activated DNA damage checkpoint in top6 plants could result in low IAA levels as a means to slow/arrest growth until the damage is repaired. The experiments described here provide evidence for a new pathway regulated by TOP6 activity.

Materials and Methods

Plant materials and growth conditions

All plants used in this study were Arabidopsis thaliana, ecotype Columbia-0 (Col) unless stated otherwise. Two T-DNA insertion lines for TOP6B, At3g20780, were obtained from the Arabidopsis Biological Resource Center (ABRC, http://abrc.osu.edu/) and named top6b-8 (SALK_024455c) and top6b-9 (SALK_10704)⁶⁵ [see Table 1]. top6b-7 plants used for phenotypic analysis were all F₂ plants from a second backcross to the progenitor line and all harbor the IAA1-LUC transgene (see genetic screen for details). All soil-grown plants were first cold-treated for 1–2 d, then grown under continuous white light at 16°C. Plants grown on solid agar growth media (GM)²¹ were first surface sterilized as seeds in 30% bleach, 0.1% Trition-X 100 and cold-treated 1–2 d, then grown at 20°C under continuous white light (40–50 μmol·m⁻²·s⁻¹) for the times indicated.

Genetic screen

The genetic screen for IAA1-LUC degradation mutants has been described previously.²¹ In brief, a transgenic line expressing IAA1-LUC was treated with EMS, and M₂ plants were screened for higher luciferase activity using a CCD camera (Princeton Instruments model NTE/CCd-TKD D12990). For this mutant, the higher luciferase activity was confirmed after the first backcross to the progenitor line (BC1 F₂) using a different CCD camera (Andor Technology model DU 434-BV, Andor SOLIS software, South Windsor, CT). The BC1 F₂ and BC2 F₂ generations were used in the single seedling assay (described below). In these experiments, the degradation defect co-segregated with a phenotype, which was later used as a marker for following the mutation.

Positional cloning

A mapping population was generated by crossing an M₃ plant to the Landsberg erecta (Ler) ecotype, and F₁ plants from this cross were selfed to obtain F₂ individuals. F₂ individuals with the top6b parental phenotype were sacrificed for genomic DNA isolation. Bulk segregant analysis⁴⁴ on pooled DNA from approximately 50 mutant F₂ individuals used a series of molecular markers that span the genome.⁴³ From this population, the mutation was linked to simple sequence length polymorphism (SSLP) marker nga162 (www.arabidopsis.org) on chromosome III. All individuals were then genotyped for additional markers that span chromosome III (experiment 1, Fig. 2A). Markers consisted of previously reported and annotated SSLP or cleaved amplified polymorphic sequence (CAPS) markers,^43,66 and CAPS markers derived from annotated single nucleotide polymorphisms (SNPs). SNPs that could be used as CAPS or dCAPS markers were determined using dCAPS finder 2.0 (http://helix.wustl.edu/dcaps/dcaps.html),⁶⁷ and new markers used in this study are listed in Table S1. For experiment 1, DNAs from a total of 71 mutant individuals were tested. Because of problematic genomic DNA preparations, not all markers worked for every individual. 71 individuals were genotyped for nga162, but only 37 individuals for marker ciw11. Individuals that were recombinant with nga162 and ciw11 markers were then genotyped for additional markers spanning the interval between them to narrow the mapping interval. For experiment 2, genomic DNA was prepared from 342 additional individuals. These individuals were genotyped in the interval between PERL0464591 and PERL0476408 with 329 individuals and 336 DNAs working for PERL0464591 and PERL0476408, respectively. Individuals recombinant at these markers, were then genotyped for additional markers within the interval. The mapping interval was narrowed to 138 kb between PERL0472202 and PERL0473223, and candidate genes in this interval were sequenced.

Genotyping top6b alleles

The mutation in top6b-7 creates an NruI restriction site that is missing in the wild-type sequence. We designed primers (5′TAACCCCAGCAACCAATCTC3′ and 5′GTTCCCGATC-GATAAGACCA3′) for PCR that amplify a region of 724 bp spanning the mutation. The PCR product was then digested with NruI and resolved by agarose gel electrophoresis. PCR and digestion of wild-type DNA results in one 724 bp band, top6–7 DNA results in 2 bands of 344 and 380 bp, and TOP6B/top6b-7 DNA results in three bands of 344, 380, and 724 bp. To genotype the T-DNA alleles top6b-8 and top6b-9, we used the gene-specific primers 5′GCCAAATGGCTGTCATTA-CTC3′ and 5′AAAGGATGACCTTCGAAGACC3′ for top6b-8 and 5′TTACCATCTTTGCACCCAGAC3′ and 5′TCGTGA-ATTTGCAGAATCTCC3′ for top6b-9. To genotype for the T-DNA insertion in these lines, we used a primer specific to the left border of the T-DNA, 5′TGGTTCACGTAGTGGGCCATC-G3′ with either gene-specific primer for each allele to test the direction of the insertion. Both gene-specific primers for top6b-9 when used with the T-DNA primer produced a DNA fragment, suggesting two tandem T-DNA insertions in this line. To verify the position of the T-DNA in these alleles, T-DNA-specific PCR products for each allele were sequenced using the T-DNA primer.

IAA1-LUC degradation experiments

BC₂ F₂ seeds segregating for top6b-7 were plated and grown in 96-well flat-bottom plates (Whatman, Clifton, NJ) with 100 μL solid GM for 7 d under constant light at 22°C. The experiments followed a protocol previously described²¹ using cycloheximide assays. Light emission from seedlings was measured every 15 min over a 60 min time course. To calculate IAA1-LUC half-lives in individual seedlings, the RLU (relative light units) at each time-point was normalized to the zero time-point measurement. To linearize the data, the ln(normalized RLU) was calculated. These data (y-axis) were then plotted against their respective time-points (x-axis), and the slope of this line determined for each seedling in the plate. The following formula was used to calculate the half-life: half-life (min) = (slope)/ln(0.5). Seedlings were scored as either mutant or wild-type based on the morphological phenotype, and long half-lives always occurred in seedlings that were visibly different from wild-type. The average half-life from multiple seedlings with wild-type or mutant phenotypes was then calculated for the final respective half-life calculation for each phenotype.

Allelism test

To determine if the mutation identified in top6b-7 was responsible for the observed morphological phenotype and IAA1-LUC degradation defect, we crossed top6b-7 to plants heterozygous for T-DNA alleles top6b-8 and top6b-9. F₁ progeny from these crosses were plated on GM plates and grown 10 d under constant light at 20°C, then scored for segregation of the mutant phenotype. Some individuals with mutant phenotypes were sacrificed for genomic DNA extraction and genotyping for top6b using the strategy described above.

To determine if F₁ mutant individuals that originated from the crosses also had slower IAA1-LUC half-lives, single-seedling degradation assays performed on those individuals. Seedlings were transferred from the GM plates into 96-well plates, treated with 100 μl of 1 mM D-luciferin in liquid GM solvent for 1 h in the dark. After the 1 h incubation, cycloheximide (2 mg/mL stock dissolved in water) was added to final concentration of 200 μg/ml. Light emission from seedlings was measured every 15 min over a 60 min time-course and half-lives calculated as described above.

Auxin response experiments

Seeds for axr1–30, Col, and TOP6B/top6b-7 were surface-sterilized, and grown on 2,4-dichlorophenoxyacetic acid (2,4-D, Sigma) plates at the indicated concentration. All media including the “0” 2,4-D contained the same percentage of solvent as a control. Seedlings were grown at 20°C under constant light for 7 d, then transferred to the surface of agar plates, imaged with flat-bed scanner, and roots measured with Image J v1.36b (http://rsbweb.nih.gov/ij/). Three independent experiments were performed using different batches of 2,4-D plates for each replicate.

IAA measurement

A population of seeds segregating for top6b-7 were plated on solid GM plates, and cold-stratified at 4 °C for 48 h, then grown for 10 d at 20°C under constant light (50 μmol·m⁻²·s⁻¹). Seedlings from multiple plates were pooled together (50–100 mg) into three different samples for each phenotype, wild-type and top6b-7. Seedlings were grown for two independent experiments, with three samples for each phenotype per experiment. For both experiments, tissue was collected mid-morning to control for circadian regulation of IAA biosynthesis. Tissue was flash-frozen in N₂ and sent for IAA quantification at the proteomics and mass spectrometry facility at the Donald Danforth Plant Science Center, St. Louis, MO USA (http://www.danforthcenter.org/science/core_facilities/proteomics_and_mass_spectrometry/). Samples were quantified by LC-MS/MS (20 min run) with d₅-IAA as an internal standard.⁶⁸

DR5:GUS assays

To generate the plant lines used for these assays, pollen from a BC₂ F₂ top6b-7 homozygote was crossed to a previously published and well characterized DR5:GUS line.⁴⁷ F₁ plants were genotyped for top6b-7 to ensure plants originated from the cross. The DR5:GUS T-DNA confers resistance to hygromycin, and F₂ seedlings were selected for hygromycin resistance and transplanted to soil. F₂ individuals were again genotyped for top6b-7, and F₃ seedlings from top6b-7 F2 heterozygotes were tested for segregation of hygromycin resistance to select for lines homozygous for the DR5:GUS transgene. The IAA1-LUC T-DNA also carries a 35S:GUS expression transgene that confers strong constitutive GUS expression; therefore, we also examined the LUC activity in F₃ seedlings segregating for top6b-7 and homozygous for the DR5:GUS T-DNA (by hygromycin resistance) to ensure that the IAA1-LUC transgene was absent in those lines. F₃ lines segregating for top6b-7 and homozygous for DR5:GUS in which the IAA1-LUC transgene had segregated away were used for analysis.

For these assays, seedlings were grown on GM plates for 10 d at 22°C under continuous light. Seedlings were treated with 2,4-D as indicated prior to GUS staining. For GUS staining, tissue was washed twice (20 min) in GUS assay buffer (25 mM Na₂HPO₄/NaH₂PO₄ pH 7.0, 1.25 mM K₃Fe(CN)₆, 1.25 mM K₄Fe(CN)₆, 0.25% Triton-X 100, 0.25 mM EDTA, 20% methanol). After the second wash, the buffer was replaced with 2.4 mM X-Gluc (5-bromo-4-chloro-3-indoxyl-β-D-glucuronide cyclohexylammonium salt, stock dissolved in N,N-dimethylformamide) in GUS assay buffer and incubated at 37°C for the indicated times. After the staining reaction was complete, the seedlings were transferred to 100% ethanol to clear chlorophyll from the tissue.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors gratefully acknowledge support from the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, US. Department of Energy (contracts DE-FG02–03ER15416 and DE-FG02–09ER16077) to JC. In addition, the initial development of the genetic screen was supported in part by a grant from the National Science Foundation (IBN 0212659) to JC. Additional support was from the Paul K. and Ruth R. Stumpf Endowed Professorship in Plant Biochemistry to J. Callis. We wish to acknowledge the UC-Davis Controlled Environment Facility (CEF) for assistance in the propagation of transgenic plants. We would like to thank Joey Lasiter and Gurjeet Bath for help with positional cloning and genotyping, Tania Gonzalez for assistance with auxin dose-response experiments, and Marissa Simon for advice on GUS experiments. We thank former and current members of the Callis laboratory and Dr. Anne Britt (UC-Davis) for helpful discussions.