A network of transcriptional repressors mediates auxin response specificity

The regulation of signalling capacity plays a pivotal role in setting developmental patterns in both plants and animals (1). The hormone auxin is a key signal for plant growth and development that acts through the AUXIN RESPONSE FACTOR (ARF) transcription factors (2). A subset of these ARFs comprises transcriptional activators of target genes in response to auxin, and are essential for regulating auxin signalling throughout the plant lifecycle (3). While ARF activators show tissue-specific expression patterns, it is unknown how their expression patterns are established. Chromatin modifications and accessibility studies revealed the chromatin of loci encoding ARF activators is constitutively open for transcription. Using a high-throughput yeast one-hybrid (Y1H) approach, we discovered a network of transcriptional regulators of ARF activator genes from Arabidopsis thaliana. Expression analyses demonstrated that the majority of these regulators act as repressors of ARF transcription in planta. Our observations support a scenario where the default configuration of open chromatin enables a network of transcriptional repressors to shape the expression pattern of ARF activators and provide specificity in auxin signalling output throughout development.

eukaryotes such networks typically incorporate a combination of transcriptional activators and repressors. All five ARF ClassA are encoded by loci with a similar structure; coding sequences are interrupted by 11-14 introns, and, in the case of ARF5, 7 and 19, the first intron is 2-3 times bigger than other introns (Fig. S1a). To identify the regulatory sequences establishing ARF patterns, we tested the role of both upstream sequences (17) and the first intron (18) in determining ARF ClassA expression by comparing patterns from transcriptional reporter lines using either sequences 3-5 kb 5' of the ATG and 3' up to the end of the first intron or the 5' sequences alone (Fig. 1k). A difference between the two reports was seen only for ARF7: the transcriptional reporters that included the first intron showed strong expression in the RAM ( Fig. 1c), but this pattern was dramatically reduced when the sequence encompassing the first intron is missing (Fig. 1l, Fig. S1). We therefore conclude that the 3' sequence including the intron contains regulatory information required for ARF7 expression. In the root, the expression patterns of ARF ClassA members differed from previously published patterns obtained with shorter 2 kb (5' of ATG) promoters ( Fig. 1 a-e, (15)). In the shoot, we also found ARF5 and 6 reporters recapitulate the patterns observed with RNA in situ hybridization more accurately than those with 2 kb promoters ( Fig. 1 f-j, Fig. S2 l-p, (16)).
To identify the factors which determine these specific ARF expression patterns, we first analyzed the chromatin status of each ARF ClassA locus by assaying for H3K27me3 and H3K4me3 chromatin modifications, as these are implicated in repressing and promoting gene expression, respectively (19). Meta-analysis of published datasets covering a range of tissues and developmental stages (including the RAM and SAM) revealed H3K27me3 is largely absent from ARF5, 6, 7, 8 and 19 loci, whereas H3K4me3 is detected at these loci ( Fig. 2 a- Table 1). Accessible regulatory regions also characterize ARF ClassA loci in the majority of tissues (Fig. 2 a-b, Fig. S3, Supplementary Table 1). These properties suggest that the chromatin configuration of ARF ClassA loci allows them to be actively transcribed throughout development.
To identify individual TFs that could regulate spatiotemporal transcription of ARF ClassA , we used a semi-automated, enhanced Y1H (eY1H) assay (23) with baits covering the long promoter sequences (including the first intron) identical to those from the transcriptional reporter lines described above. To do so, an existing prey collection enriched in root-stele expressed TFs was expanded to include TFs acting in the SAM and hormone signalling pathways. This extended the library's capacity to identify putative regulators in both the root and shoot tissues, plus crosstalk between hormonal pathways. In total, this eY1H screen identified 42 novel putative transcriptional regulators of ARF ClassA . Mapping this gene regulatory network revealed that individual ARF ClassA are regulated by distinct sets of largely non-overlapping TFs, with only 4 binding to multiple ARF ClassA sequences (Fig. 2c). This network is overrepresented by members of the WRKY, ZFP, AP2/ERF and SPL TF families and contain proteins that mediate either root-or shoot-specific responses; 50% of the identified TFs are expressed in both shoots and roots while 24% and 14% are expressed specifically in roots or shoots respectively (Fig. S4 a and b). The majority of the TFs from the network are involved in development, but many putative regulators of ARF8 are associated with biotic and abiotic stress (Fig. S4c, Supplementary Table 2). This suggests ARF8 may act as an environmental hub to mediate auxin responsiveness.
To validate the ARF ClassA regulatory network, we first used DNA affinity purification sequencing (DAP-Seq) datasets to analyse direct binding in vitro of the putative regulator TFs on ARF regulatory regions (24). DAP-Seq data are available for 17 TFs and 9 (53%) of them show at least one specific peak on ARF ClassA promoters (Fig. S5 a and b). These analyses demonstrate binding of a significant subset of the regulatory TFs to ARF ClassA promoters. A literature search identified binding sites for a further 9 TFs absent from the DAP-seq dataset, supporting a model of direct regulation of ARF ClassA by these TFs (Fig. S5c, Supplementary Table 3). We next tested the activity of each TF in the ARF ClassA regulatory network systematically by employing transient expression assays in Arabidopsis protoplasts ( Fig. 3a and b, Fig. S6). Each interaction was analysed in at least 4 independent experiments using either the TF alone or a fusion of the TF to a VP16 transactivation domain (Supplementary Table.  To confirm these results in planta, we identified 25 mutants in TFs from the regulatory network representing regulators of all five ARF ClassA genes (Supplementary Table 5). We then measured the expression of target ARF ClassA genes using qRT-PCR in whole-root and shoot tissues ( Fig. S7a  The low-sensitivity of expression analysis on whole tissues prompted us to determine at higher spatial resolution how these mutant alleles affect ARF ClassA expression. We focused on ARF7 and crossed the pARF7::mVenus transcriptional reporter into a number of mutants to examine the spatial domain of ARF7 expression in more detail. For two of the regulator mutant backgrounds in which we had not observed changes in ARF7 mRNA levels using qRT-PCR (crf10 and wrky38), we observed a significant expansion of pARF7::mVenus expression pattern in the RAM ( Fig. S7c and d). For nf-yb13 we also observed enhanced expression in the RAM similar to qRT-PCR assays (Fig. 4a). This confirms in planta that these three TFs are repressors of ARF ClassA and provides an illustration of how repressors can shape ARF expression pattern.
To further confirm a role for this network in ARF ClassA tissue-specific repression in planta, the 25 mutants were scored for defects in auxin-regulated processes during root and shoot development: root length on medium supplemented with or without IAA, root gravitropic response, and shoot growth including leaf number, rosette size, shoot length, number of axillary and cauline shoots during vegetative and reproductive growth stages ( Table 7). 28% of the mutants showed root phenotypes in processes associated with auxin-regulation, 12% showed auxin-related shoot phenotypes and 52% had both root and shoot phenotypes. For leaf number, rosette size, root elongation and gravitropic response, a majority of the mutants affected showed opposite effects on development from loss of function mutants in loci known to promote auxin signalling (13,25,26,27). This supports a negative regulation of auxin responses by the corresponding TFs. Mutation of single genes in the ARF ClassA regulatory network can therefore affect auxin-dependent development substantially, demonstrating the functional importance of individual nodes of this regulatory network.
Collectively our results reveal a mechanism through which auxin response is mediated not only through asymmetry in the distribution of the hormone itself, but also through coordinated TF-mediated repression of ARF ClassA expression. Consistent with this scenario, we often observed that members of the ARF ClassA and their regulatory repressors are expressed in complementary domains in the root (Fig. S10).
Despite a general role of Polycomb-mediated gene repression in tissue-specific expression (28), the absence of H3K27me3 at ARF ClassA loci indicates that their regulation does not rely on this epigenetic mechanism. This may be because such a system would not allow the rapid changes in auxin signalling output required to dampen auxin responses and limit them both spatially and temporally. Instead, we have uncovered a system based on the use of transcriptional repressors to modulate expression of constitutively active loci, which offers a more flexible mode of regulation that can constantly adjust auxin responsiveness. Our network is unlike transcriptional regulation networks previously defined in eukaryotes that involve both transcriptional activators and repressors (29). Instead, our network resembles the early scenario for transcriptional regulation via repressors proposed by Jacob and Monod (30), indicating that there is a biological context for the concept of controlling the expression of key developmental regulators almost entirely via transcriptional repression.

Plant material
T-DNA insertion mutants in transcription factors and the arf8-1 mutant were obtained from NASC. T-DNA accession numbers are listed in Supplementary Table 5.
The fragments were inserted into pDONR P4-P1R and recombined with 3x mVenus-N7 pDONR211, OCS terminator pDONR P2R-P3 and pK7m34GW destination vector to yield pARF -intron ::mVenus shorter transcriptional reporter lines. The constructs were transformed into Agrobacterium tumefaciens C58pMP90 strain by electroporation and then transformed into Col-0 plants by the floral dip method (Clough and Bent 1998).

Root microscopy
For root microscopy plants were grown on half-strength Murashige and Skoog (MS) medium supplemented with 1% sucrose and 1% agar. The seedlings were grown in 24h light conditions and imaged at 5 or 6 days in light. Plant cell membranes were visualized by staining with 15 µg/ml propidium iodide solution. Roots were examined using a TCS-SP5 confocal microscope (Leica) with excitation at 514 nm and emission at 526-560 nm for mVenus and 605-745 nm for propidium iodide.

Shoot microscopy
Plants were grown in 8h light/16h dark conditions for 6 weeks and then transferred to 16h light/8h dark conditions for 2 weeks to induce bolting. The bolted shoots were dissected under a stereomicroscope and planted into an Apex Culture Medium (half-strength MS medium supplemented with 1% sucrose, 0.8% agarose, 1x vitamin solution (myo-Inositol 100 mg/L, nicotinic acid 1 mg/L, pyridoxine hydrochloride 1 mg/L, thiamine hydrochloride 10 mg/L, glycine 2 mg/L)), for overnight incubation at 16h light/8h dark conditions. Before microscopy cell membranes were visualized by staining the shoot apexes with 100 µg/ml propidium iodide solution. The shoot apices were examined using a TCS-SP5 confocal microscope (Leica) with excitation at 514 nm and emission at 526-560 nm for mVenus and 605-745 nm for propidium iodide.

Y1H assay
The yeast one-hybrid assay was conducted according to Gaudinier et al. 2011 (23). The overall ARF promoters screened correspond in length and content to the ones used in the construction of the transcriptional reporter lines: pARF5 -5418 bp to + 134 bp, pARF6 -3255 bp to +197 bp, pARF7 -2973 bp to + 374 bp, pARF8 -5091 bp to + 42 bp, pARF19 -4906 bp to + 452 bp. pARF5, pARF8 and pARF19 promoters were divided in two fragments and each was screened separately.

Transient expression in Arabidopsis thaliana protoplasts
In the transient protoplast assay, a reporter and an effector plasmid were co-transformed (Fig.   S6). To produce the reporter plasmid, the promoter fragment of the respective ARF corresponding to the one used in the eY1H assay and the ARF transcriptional reporter lines described above were amplified by PCR and recombined with a pDONR P4-P1R plasmid. For the ARF8 promoter a short part of the 35S promoter (-107 to +1) was inserted at position -115 bp. Separately, a construct containing NLS followed by mVenus coding sequence and an octopine synthase (OCS) terminator was cloned into pDONR 211 plasmid. Thirdly, a construct containing the promoter of RPS5a (promoter of the ribosomal protein S5A) driving TagBFP followed by a NLS signal and a nosT terminator were recombined with the pDONR P2R-P3 plasmid. These three plasmids were recombined with a multisite Gateway to yield the final reporter plasmid pARF-NLS-mVenus-term-pRPS5a-TagBFP-NLS-term. An alternative reporter plasmid contained a shorter ARF promoter fragment which contained sequences upstream and lacked sequences downstream of the start codon (corresponding to the transcriptional reporter lines with shorter promoters described above).
To produce the effector plasmid, the RPS5a was cloned into pDONR P4-P1R plasmid. The A Zeiss 710 LSM confocal microscope was used for imaging the protoplasts (Fig. S6). The sequential scanning was performed with mVenus (excitation at 514, emission at 520-559), TagBFP (excitation at 405 and emission at 423-491), mCherry (excitation at 561, emission at 598-636) and bright-field channels. Z-stacks of several protoplasts were taken. The data was analysed using ImageJ software. The image with the best focus for each protoplast was selected from the z-stack. The nucleus was selected and the mean fluorescence was measured as illustrated in Fig. S6. For the statistical analysis an unpaired t-test was conducted with p ˂

Expression analysis with qRT-PCR
Wild-type and mutant seedlings were grown in 24h light conditions on 1/2 MS plates containing 1% sucrose and 1% agar for 7 days. The whole root and the whole shoot parts of the seedlings were collected separately. For one root sample, roots from 30 seedlings grown on the same plate were pooled together. For one shoot sample, 8 shoots from seedlings grown on the same plate were pooled together. Three independent replicates per genotype were collected. RNA was extracted using Spectrum Plant Total RNA kit (Sigma-Aldrich). The DNA was removed using TURBO DNA-free kit (Invitrogen). The cDNA was produced using SuperScript VILO cDNA Synthesis kit (Thermo Fisher) with 500 ng RNA. The cDNA was diluted 1:100 before use. The qRT-PCR was performed using Applied Biosystems Fast SYBR Green Master Mix. Expression of TUB4 gene was used as standard. The statistical analysis was performed with Mann-Whitney test with p ˂ 0.1 considered as statistically significant.

Expression analysis of crosses between ARF transcriptional reporter lines and TF mutants
Mutants of the regulatory transcription factors were crossed with pARF7::mVenus transcriptional reporter line described above. The crosses were selected for the presence of a homozygous pARF7::mVenus reporter construct. The F3 generation wild-type and mutant plants were compared. The plants were grown in 16h light / 8h dark conditions. Root microscopy was performed as described above.

Shoot phenotype analysis of the TF mutants
25 T-DNA insertion mutants and the wild-type Col-0 were grown in 8h light/16h dark conditions on soil for 43 days. Leaf number was counted every 3 days starting from day 24.
Rosette diameter was measured at 43 days. After 43 days of growth in the above conditions, the plants were transferred to 16h light/8h dark conditions to induce bolting. The following parameters were measured at 21 and 27 days in the 16h light/8h dark conditions: length of the main stem, number of cauline branches growing from the main stem, number of axillary branches growing from rosette (the main stem not included). The number of replicates per genotype was 12 plants. For the statistical analysis an unpaired t-test was conducted with p ≤ 0.05 considered as statistically significant.

Root phenotype analysis of the TF mutants
For root length measurement and for gravitropic analysis plants were grown on ½ MS medium supplemented with 1% agar in 12h light/12h dark conditions. For root length analysis, plants were grown either on medium lacking IAA or supplemented with 10 µM IAA.
Images were taken at 5 and 15 days in light and the root length was measured. The number of replicates per genotype was at least 26 plants without IAA and 15 plants with IAA. For the gravitropic response, plants were grown for 5 days, then turned at a 90°C angle and the images of the root gravitropic growth were taken every 1 hour for the next 12h hours in the dark with the infrared camera. The number of replicates per genotype was at least 26 plants.
Statistical analysis was done with one-way analysis of variance (ANOVA) test with p ≤ 0.05 considered as statistically significant.

In silico analysis
Expression of TFs in the root and the shoot apical meristems was analysed using cell-specific Overrepresentation of TF gene families was analysed for families represented by two or more members in the network. The number of gene family members in the network was compared to the total number of genes from this family in the TF library.
Involvement of TFs in specific developmental processes (development, biotic and abiotic stress) was analysed based on literature description.

Chromatin state analysis
Binary data on H3K27me3-and H3K4me3 marked genes and chromatin accessibility regions were retrieved from multiple datasets covering a range of tissues and developmental stages.