Research ArticlePLANT SCIENCES

An auxin-regulable oscillatory circuit drives the root clock in Arabidopsis

See allHide authors and affiliations

Science Advances  01 Jan 2021:
Vol. 7, no. 1, eabd4722
DOI: 10.1126/sciadv.abd4722


In Arabidopsis, the root clock regulates the spacing of lateral organs along the primary root through oscillating gene expression. The core molecular mechanism that drives the root clock periodicity and how it is modified by exogenous cues such as auxin and gravity remain unknown. We identified the key elements of the oscillator (AUXIN RESPONSE FACTOR 7, its auxin-sensitive inhibitor IAA18/POTENT, and auxin) that form a negative regulatory loop circuit in the oscillation zone. Through multilevel computer modeling fitted to experimental data, we explain how gene expression oscillations coordinate with cell division and growth to create the periodic pattern of organ spacing. Furthermore, gravistimulation experiments based on the model predictions show that external auxin stimuli can lead to entrainment of the root clock. Our work demonstrates the mechanism underlying a robust biological clock and how it can respond to external stimuli.


Both plants and animals can regulate patterning through developmental clocks that involve oscillating gene expression (1). In Arabidopsis thaliana, the root clock determines organ spacing along the primary root axis by establishing prebranch sites (PBS) through oscillating gene expression approximately every 6 hours (2). However, periodicity of the root clock can vary under specific environmental conditions or by supplementation with the phytohormone auxin (24). Oscillations in gene expression occur as propagating waves in the oscillation zone (OZ) in two opposite phases: in-phase and antiphase based on expression of the DR5::Luciferase auxin response reporter (2). When expression of in-phase genes is activated in the OZ, the expression of antiphase genes is repressed and vice versa.

At the cellular level, PBS formation is interpreted as priming of pericycle cells to specify lateral root (LR) founder cells, which is a prepatterning stage (5). Subsequently, LRs originate from PBS (2) through division of founder cells (1, 59).

Programmed cell death has been proposed to release auxin into the OZ to control periodicity of the root clock (4). In addition, auxin derived from the lateral root cap (LRC) and the Aux/IAA factor IAA28 regulate DR5::Luciferase oscillations and PBS formation (3, 4, 10). However, how periodicity of the in-phase and antiphase oscillations is established in the OZ to determine PBS spacing is not resolved.


To gain further insight into the root clock mechanism, we performed a mutagenesis screen in plants carrying DR5::Luciferase in combination with markers for subsequent LR organogenesis [pWOX5::ER-YFP and pSCR::ER-GFP (11)]. From this screen, we identified a mutant with increased expression of DR5::Luciferase in the OZ as well as PBS distributed throughout the primary root axis without regular spacing (Fig. 1A). Time course analyses of in-phase gene oscillations in the mutant showed a persistent signal in the OZ (movies S1 and S2) without the typical 6-hour oscillatory behavior, thereby causing abnormal PBS spacing (Fig. 1, B and C). Introgression of a specific marker for LR founder cells (12) in this mutant showed that most pericycle cells had been specified as founder cells (Fig. 1D). Because of this increased capacity to form PBS and LR founder cells, we named this mutant potent. In summary, root clock function is impaired in potent, which results in more priming and abnormal LR prepatterning events.

Fig. 1 The IAA18/POTENT factor regulates the root clock oscillations and LR priming.

(A and B) DR5::Luciferase oscillations lack the typical oscillatory behavior causing abnormal PBS spacing in iaa18/potent mutant. Arrows: PBS. (C) Kymograph showing increased PBS production in iaa18/potent roots. Asterisks: PBS. Note that not all PBS are maintained over time. (D and E) iaa18/potent allele overproduces founder cells (FC) as shown by (D) confocal microscopy and (E) the quantification of the FC marker pSKP2B0.5::ER-3mCherry in a iaa18/potent (IAA18P102L) estradiol-inducible line (pER8::iaa18/potent). Arrowheads: FC limits. (F) Altered positioning of LRs in iaa18/potent mutant. (G) Quantification of LR formation in iaa18/potent, in iaa18/potent allele (IAA18P102L) expressed under its regulatory regions, and in an alternative mutant allele (crane2). (A to D) Seven days post imbibition (dpi). (E to G) Seedlings were treated at 4 dpi. Scale bars, 5 mm (A), 1 mm (B), 50 μm (D), and 0.1 mm (F). *P < 0.001 by general linear model (GLM)/least significant difference (LSD). n per sample: ≥10 (E) and ≥20 (G). Error bars, SD.

Mapping of the potent mutation showed a change of cytosine to thymine in the coding sequence of the IAA18 gene, causing the substitution of proline 102 by leucine in the DII domain (fig. S1, A to D, and data file S1). Mutations in conserved DII domain prolines have been associated with auxin insensitivity and increased stability of the Aux/IAA proteins, resulting in dominant mutations (fig. S1E) (1315). The expression of the iaa18/potent allele (IAA18P102L) under an estradiol-inducible promoter showed increased numbers of LR founder cells with increasing doses of estradiol (Fig. 1E), thus confirming the regulation of priming by IAA18/POTENT activity. Aux/IAA proteins are thought to act redundantly during development (13, 16). It is possible that IAA18/POTENT may function redundantly with other Aux/IAAs; however, IAA26, the closest homolog of IAA18/POTENT, is not expressed at detectable levels in the OZ (17).

Next, we determined that founder cells in iaa18/potent mutant did not undergo further development (fig. S2A), explaining the absence of LRs in this mutant (fig. S2, B and C). Other Aux/IAA proteins regulate LR founder cell division (11, 18, 19), so we reasoned that IAA18/POTENT might function redundantly during LR initiation. Using an auxin concentration (0.25 μM) that does not alter prepatterning in the wild type (fig. S3A) but promotes LR initiation, we observed production of LRs in iaa18/potent roots (Fig. 1F). This confirms that LRs can be initiated through alternative Aux/IAA combinations even in the presence of the iaa18/potent mutated protein. Furthermore, we observed increased production of LRs with reduced or no spacing in iaa18/potent mutant as well as when iaa18/potent (IAA18P102L) was expressed under its native regulatory regions and in crane2 (Fig. 1E), which harbors an alternative mutation in the DII domain of IAA18 (fig. S1E) (20).

When we treated the iaa18/potent (IAA18P102L) estradiol-inducible line with low estradiol doses and with an auxin concentration over the threshold (0.25 μM) that induces founder cell specification, we observed an additive effect of iaa18/potent allele with auxin in founder cell specification and LR formation (fig. S3B). In contrast, higher estradiol doses prevented development of the numerous induced founder cells (fig. S3C), thus inhibiting LR initiation (fig. S3B). These results indicate that priming and LR founder specification are highly sensitive to IAA18/POTENT levels, placing IAA18/POTENT in a central role in prepatterning.

Consistent with this role, we found that the wild-type IAA18/POTENT-Luciferase protein was present in the OZ and PBS, while in the iaa18/potent allele (IAA18P102L-Luciferase), the protein was primarily stabilized in the OZ and shootward regions where priming is typically observed (fig. S4A). IAA18/POTENT-YFP displayed both nuclear and cytoplasmic localization in the endodermis and pericycle cells of the OZ, whereas the iaa18/potent mutation (IAA18P102L-YFP) caused accumulation in the nucleus (fig. S4, B and C). Supplementation with auxin caused degradation of wild-type IAA18/POTENT-YFP, whereas no change in iaa18/potent (IAA18P102L-YFP) protein levels was observed (fig. S4, D and E), consistent with observations for other Aux/IAA DII domain mutants (14, 15). The absence of the DR5::Luciferase oscillatory behavior in iaa18/potent mutant is, therefore, associated with the reduced capacity of this factor to be degraded by auxin.

Aux/IAA proteins are inhibitors of auxin response factors (ARFs) (21). ARF7 is a member of the ARF family that has been shown to be involved in PBS formation (2), root hydropatterning (22), and LR initiation (23). Furthermore, ARF7 transcripts tend to oscillate in antiphase to DR5::Luciferase (2). Analysis of ARF7 under its native regulatory regions using recombineering (24) showed a predominant nuclear localization coinciding, at the maximum ARF7 levels, with the beginning of the OZ (fig. S5). Nucleocytoplasmic partitioning of ARF7 has been associated with auxin responsiveness (25), suggesting that transcriptional regulation mediated by ARF7 would be active primarily in the OZ. In addition, we observed that ARF7 levels fluctuated in the OZ over time (fig. S5). Time course analyses of DR5::Luciferase in a loss-of-function arf7-1 mutant showed enhanced signal in the OZ (Fig. 2A) and absence of the characteristic oscillations (Fig. 2B and movie S3), resulting in more PBS formation (Fig. 2C). Furthermore, introgression of the LR founder cell marker in arf7-1 mutant showed that many pericycle cells were specified as founder cells, indicating more priming events as compared with the control (Fig. 2D). Many of these founder cells in arf7-1 mutant did not develop into LRs, even upon 0.25 μM auxin supplementation, which is consistent with a requirement for the transcriptional activating function of ARF7 during LR initiation (23), and the observation that not all PBS in arf7-1 roots are maintained over time (Fig. 2C). In conclusion, the root clock function in arf7-1 mutant is impaired in the OZ, resulting in altered prepatterning, which phenocopied that of iaa18/potent mutant.

Fig. 2 ARF7 and IAA18/POTENT jointly regulate the root clock.

(A and B) Altered DR5::Luciferase oscillations and PBS spacing are observed in the loss-of-function mutant arf7-1. Arrows: PBS. Dashed line: Polynomial regression. (C) Kymograph showing increased PBS production in arf7-1 roots. Asterisks: PBS. Note that not all PBS are maintained over time. (D) Confocal microscopy images of the FC marker pSKP2B0.5::ER-3xmCherry in arf7-1 roots show FC overspecification. Arrowheads: FC limits. (E) Split-luciferase signal reconstitution shows heterodimerization of AUXIN RESPONSE FACTOR 7 (ARF7) and IAA18/POTENT or iaa18/potent (IAA18P102L) in the OZ and shootward priming regions. (F) Pull-down of ARF7 and IAA18/POTENT or iaa18/potent (IAA18P102L) expressed in Escherichia coli. (G and H) RNA-seq reveals (G) common differentially expressed genes (DEG) in arf7-1 and iaa18/potent mutants and (H) directional regulation of in-phase (activated) and antiphase (repressed) genes. (I and J) Quantification of DR5::Luciferase and pLBD6::Luciferase in the OZ. LBD16 is an in-phase DEG. Experiments were performed at 7 dpi. Scale bars, 2 mm (A), 1 mm (C and E), and 50 μm (D). *P < 0.05, **P < 0.001 by GLM/LSD. n per sample: ≥20 (I) and ≥10 (J). Error bars, SD.

As both iaa18/potent and arf7-1 roots showed similar alterations in LR prepatterning and ARF7 and IAA18/POTENT interacted in yeast (fig. S6A), we hypothesized that they could form heterodimers in Arabidopsis. To test this hypothesis, we fused these proteins to the half moieties of split-luciferase and expressed them under their own promoters. Reconstitution of split-luciferase has been shown to report functional interactions spatially (26). We observed reconstitution of the luciferase signal demonstrating IAA18/POTENT and ARF7 dimerization in the OZ and shootward priming regions (Fig. 2E). No luciferase reconstitution was observed in controls (fig. S6, B to D). Luciferase reconstitution between iaa18/potent (IAA18P102L) and ARF7 showed increased signal in the OZ and priming regions, as compared with wild-type IAA18/POTENT, indicating heterodimer accumulation in iaa18/potent roots (Fig. 2E and fig. S6E). This accumulation correlates with an increase in DR5::Luciferase expression in the OZ and absence of the oscillatory behavior, suggesting that heterodimer formation derepresses the oscillations, probably because IAA18/POTENT inhibits ARF7 activity in the OZ.

ARF7 function during hydropatterning requires posttranslational modification by SUMOylation, which occurs in the presence of auxin and mediates ARF7 interaction with IAA3 (22). SUMOylation of ARF7 does not appear to be required for interaction with IAA14/SOLITARY-ROOT during LR initiation (22). Because bacteria lack SUMOylation, we tested the interaction using proteins produced in bacteria and found that IAA18/POTENT (or iaa18/potent-IAA18P102L) and ARF7 interaction does not require SUMOylation (Fig. 2F).

To understand regulation of the oscillation phases by IAA18/POTENT and ARF7, we performed RNA sequencing (RNA-seq) on samples from the OZ taken at the minimum of DR5::Luciferase expression for wild type and, simultaneously, for arf7-1 and iaa18/potent mutants regardless of DR5::Luciferase expression to avoid bias (fig. S7A) (9). Analysis of differentially expressed genes in arf7-1 and iaa18/potent mutants (compared with wild type; data file S2) showed large overlap, with ~65% deregulated genes in common (Fig. 2G). Of the deregulated oscillating genes, we found that most of the in-phase genes were activated in arf7-1 and iaa18/potent mutants, whereas most antiphase genes were repressed (Fig. 2H). When we overexpressed ARF7, we observed repression of in-phase genes in the OZ, which would return to levels similar to wild type by the introgression of iaa18/potent allele (Fig. 2I). These results reveal that ARF7 acts as a repressor in the regulation of in-phase oscillating genes, and this function is modulated by IAA18/POTENT.

To investigate upstream regulation of IAA18/POTENT and IAA28 function, we crossed iaa28-1 allele, which harbors a mutation in the DII domain causing auxin insensitivity (10), with iaa18/potent allele and introgressed markers. We observed that increased expression of DR5::Luciferase in the OZ caused by iaa18/potent mutant did not occur in iaa28-1 mutant (Fig. 2J), suggesting epistasis. When we investigated expression of DR5::Luciferase in the double-mutant arf7-1 iaa28-1, we observed expression levels similar to those in iaa28-1 mutant. We conclude that ARF7 repression of in-phase genes and its inhibition by IAA18/POTENT require activation by auxin. We interpret these results as the existence of two separate signaling mechanisms for auxin that determine two antagonistic actions. IAA28 signaling, which comes first and activates auxin responses, would define the intensity or amplitude of the oscillations. Next, IAA18 signaling would act by repressing gene expression to define the oscillations and their periodicity through negative feedback.

In our transcriptomic analysis, we also investigated categories of genes involved in root clock function (fig. S7B). Several of these such as cell wall remodeling and vesicle trafficking have been previously shown to participate in the root clock (9). We found up-regulation of the auxin biosynthetic branch shared with glucosinolate production in arf7-1 and iaa18/potent mutants (fig. S7C). Indole-3-butyric acid conversion to auxin in the LRC has been associated with root clock function and would require auxin transport to the epidermis (27). However, it is unknown whether local auxin biosynthesis is required for the root clock mechanism. We investigated auxin content using the auxin biosensor R2D2, in which the ratio between the red fluorescent protein and yellow fluorescent protein (YFP) is a proxy for auxin concentration (28). We found high variability in the OZ of control plants. Notably, arf7-1 and iaa18/potent mutants had more auxin in the OZ, while the ARF7 overexpression line had lower levels (Fig. 3A). LRC-derived auxin has been associated with gene expression oscillations through the turnover of DR5 stripes (6). We found that DR5 stripes in iaa18/potent mutant disappeared at the same rate as control plants, although there appear to be more stripes in iaa18/potent mutant (Fig. 3B). These results indicate that IAA18/POTENT and ARF7 are unlikely to regulate LRC-derived auxin, although modulation of auxin levels in the OZ by these factors could be part of the mechanism of the root clock.

Fig. 3 ARF7 and IAA18/POTENT regulate auxin, and IAA18/POTENT levels are regulated by auxin.

(A) Quantification of auxin in the OZ epidermis of arf7-1 and iaa18/potent roots using the R2D2 biosensor. (B) Stripes of DR5::VENUS-N7 in control and iaa18/potent mutant. Scale bars, 50 μm. (C) IAA18/POTENT expression in the OZ is activated by auxin. (D) IAA18/POTENT levels in the OZ change in response to auxin supplementation. Note that IAA18/POTENT levels are the result of synthesis and degradation. (E) ARF7 expression and ARF7 protein levels in the OZ do not respond to auxin supplementation. Experiments were performed at 7 dpi. *P < 0.05 by generalized linear model (GzLM) (A), GLM/LSD (C and D), and Student’s t test. (E). n per sample: ≥10 (A), ≥5 (C and D), and € ≥ 12. Error bars, boxplot whiskers (A) and SD (C to E).

As auxin levels change in the OZ, we asked whether auxin feeds back on IAA18/POTENT and ARF7 expression or their protein levels. When we used the minimum local auxin concentration (0.01 μM) capable of mimicking an oscillation (2), we did not observe changes in iaa18/potent levels (Fig. 3C), whereas IAA18/POTENT-YFP protein levels decreased after 3 hours (Fig. 3D). Next, we tested an auxin concentration (1 μM) that induces founder cell specification. We observed increased transcription starting at 1 hour, while IAA18/POTENT-YFP protein levels decreased after 3 hours (Fig. 3, C and D). Intriguingly, both 0.01 and 1 μM auxin treatments produced a similar decrease in IAA18/POTENT-YFP levels at 3 hours, indicating the existence of a buffering mechanism. In contrast, ARF7 transcription and its protein levels did not change when the higher auxin concentration was tested (Fig. 3E).

To understand the dynamics of IAA18/POTENT and ARF7 circuit in pattern formation (Fig. 4A), we built a computer model of LR priming leading to PBS formation based on parameters inferred from the experimental data (see data file S3 for details). In addition to using experimentally derived model parameters, we performed a sensitivity analysis and found that the ARF7-IAA18 circuit was most sensitive to ARF7 synthesis rates as well as auxin turnover. These findings are consistent with the importance of ARF7 in the regulation of the oscillatory behavior and PBS formation as indicated by experimental observations. Also, the predicted importance of auxin homeostasis in the ARF7-IAA18 circuit can explain why the excess of auxin in the OZ can override the system and cause induction of LRs (29). The model tracks the growth of xylem pole pericycle or pericycle in time and space as priming occurs in these tissues (Fig. 4, B to D) (6, 25, 27). We found that the model requires an additional hypothetical factor (F) in phase or activated by auxin to explain the observed buffering of IAA18/POTENT protein degradation in the OZ when roots were treated with different auxin concentrations (fig. S8). Predictions suggest that this new component modulates ARF7-IAA18/POTENT heterodimerization, creating an auxin-dependent feedback loop, which promotes auxin response (Fig. 4A).

Fig. 4 Multilevel computer model of LR priming predicts ARF7-IAA18/POTENT circuit dynamics in response to a changing environment.

(A) Schematic representation of the IAA18/POTENT-ARF7 circuit. (B) Snapshots of model simulations for the wild-type scenario (left), compared to experimental DR5::Luciferase assays (right). M, meristematic root zone; E, elongation root zone; D, differentiation root zone. The pseudocolored ellipse reflects DR5 expression and depicts priming sites and PBS. (C) Computer simulations of iaa18/potent mutant (left) compared to experimental DR5::Luciferase assays in iaa18/potent mutant (right). (D) Color map for DR5 levels in (E) and (F). a.u., arbitrary units. (E) Kymographs of simulated wild type (WT) (top) and iaa18/potent mutant (bottom) with oscillating ARF7 mRNA shown on the top bar. DR5 levels are shown across the OZ (~cells 20 to 35). QC, quiescent centre. (F) Same as in (E) but with the addition of periodic pulses of external auxin every 3 hours. Note the entrainment of wild type to external auxin stimuli that was absent in iaa18/potent mutant simulation. (G) Frequency of distance between PBS priming events observed across all simulated scenarios. (H) Distribution of oscillatory period between successive priming events. The DR5 threshold for priming was kept the same (4.35) in all simulations.

The model simulations showed a dynamic wave of DR5/in-phase genes originating at the basal meristem and moving shootward with a wavelength of approximately 15 cells (Fig. 4, B and E, and movie S4), which coincides with the region previously described as the OZ. This demonstrates that the propagating waves of gene expression are an emergent property of this system and carry positional information. Oscillations appeared in the simulations every 5 to 6 hours (Fig. 4H), which is in agreement with experimental observations (2). When maxima in DR5/in-phase gene expression were associated with LR priming in the model, as in the experimental observations (2), the simulations created a correct pattern of PBS spacing (Fig. 4, B and G). These findings suggest that in-phase genes need to be activated in the root region with low cell division and active growth to create a pattern. In simulated iaa18/potent mutant, LR priming occurred continuously, causing fusion of consequent PBS recapitulating the iaa18/potent mutant phenotype (Fig. 4, C, E, and G, and movie S5).

We also tested whether the introduction of experimentally derived ARF7 expression profiles in our model would change its behavior, and we found that the model was robust to fluctuations in ARF7 expression (fig. S9A). Thus, despite the ARF7 oscillations being noisy, the ARF7-IAA18 circuit shows buffering capacity and produces periodically originating PBS with a similar frequency as in the experimental observations (fig. S9B).

Biological clocks can be entrained by external or environmental cues (30). As a periodic auxin stimulus has been proposed to regulate the root clock periodicity (3, 6), we performed computer model simulations to test whether a periodic external auxin input into the OZ would entrain the root clock. Simulations of auxin influx into the pericycle showed that the effect on priming is specific to the OZ (cell numbers 20 to 35 from the quiescent center; see data file S3 for details), and thus, influx of auxin at other locations has little or no effect on LR priming frequency. We used 0.01 μM auxin in the model simulations, as experimentally, this concentration mimics oscillations of DR5 (2). Pulses of auxin every 3 hours led to entrainment in a wild type–like scenario that resulted in increased frequency of priming and reduced PBS spacing (Fig. 4, F to H). However, no entrainment was observed in iaa18/potent mutant simulations (Fig. 4, F to H). Gravistimulation has been shown to induce DR5 oscillations and priming (2, 31) and involves auxin transport, which may lead to increased auxin levels in the pericycle (31, 32), reduction in IAA18/POTENT, and, thus, an alteration of clock circuit dynamics. Therefore, we used gravistimulation to test these model predictions. Gravistimulated wild-type roots displayed an increased frequency in PBS formation (fig. S10, A and B), whereas areas with fused PBS were observed in iaa18/potent and arf7 mutants (fig. S10C). Thus, these results are in good agreement with the model predictions, confirming a reduced capacity of iaa18/potent and arf7 mutants to respond to external changes. Last, our model simulations demonstrate that auxin inputs that are in-phase to DR5 amplify the oscillations, resulting in more PBS formation, whereas antiphase auxin inputs did not have a large impact on priming (fig. S9, C and D). This result is in agreement with previous experimental results (2) and suggests that external auxin would require coordination with the OZ-located ARF7-IAA18/POTENT circuit to create a pattern. From these observations, we propose a mechanism in which auxin reaching the OZ is interpreted by the ARF7-IAA18 oscillator, which controls auxin through ARF7 in a feedback-dependent manner. Therefore, the oscillations are not merely a readout of external auxin transport or accumulation in the OZ but rather an intrinsic property of a regulable developmental clock.

Biological oscillators such as circadian rhythms and the segmentation clock require negative feedback loops to create robust cyclic patterns (3336). The key aspects of our circuit that are necessary for sustainable oscillations are the repressive activity of ARF7 on the in-phase genes and on auxin levels and, predicted by the model, an auxin feedback on ARF7-IAA18/POTENT dimerization, which negatively affects ARF7 activity. Our oscillatory circuit demonstrates central properties of biological clocks as it can buffer molecular noise, creating a periodic pattern and, in addition adapt to external signals, being entrained by persistent fluctuations in hormone levels. When this circuit was implemented in a multilevel model, it generated waves of gene expression traveling across growth domains, a specific characteristic of the root clock, which is shared with the segmentation clock (1). Our research demonstrates how an oscillator can be positioned in a growing organ to create a robust pattern of periodic organogenesis.


Plant material and growth conditions

A. thaliana seeds were surface sterilized with gas chlorine (1% HCl) in a confined environment for 2 hours. After stratification for 1 to 2 days at 4°C, plants were grown on 12 × 12–cm plates with Murashige and Skoog basal medium [MS (2.2 g/liter), 0.05% MES, 1% sucrose, 1% plant agar] in a walk-in custom-made chamber with 16/8-hour light/dark photoperiod at 21° to 23°C. Plants were analyzed at 7 days post imbibition (dpi), except when indicated. Lines previously reported and used in this study were crane2 (20), iaa28-1 (37), the ratiometric R2D2 auxin sensor (28) DR5rev::VENUS-N7 (38), pWOX5::ER-YFP (39), pSCR::ER-GFP (40), and DR5::Luciferase, pARF7::Luciferase and pLBD16::Luciferase (2). The triply marked line DR5::Luciferase pWOX5::ER-YFP pSCR::ER-GFP was generated by crossing the respective parental lines, followed by selection of plants carrying the homozygous markers by means of a fluorescence scope or by luciferase assays. arf7-1 mutant corresponds to the SALK_040394 line.

Mutagenesis screening and mapping

Seeds from the triply marked line DR5::Luciferase pWOX5::ER-YFP pSCR::ER-GFP were mutagenized using ethyl methanesulfonate solution [50 mM EMS and 0.1 M sodium phosphate buffer (pH 5.5)] for 12 hours. About 50 seeds from 2000 independent M2 lines (N = 100,000 seeds) were screened for impairment in LR formation using a fluorescence scope or in luciferase assays. Potent was crossed six times with the parental reporter line in Col-0 ecotype and maintained in heterozygosity as it is a dominant mutation. To map the mutation, potent was then crossed five more times with Ler ecotype. Approximately 200 seeds generated in the fifth cross were sown, and seedlings with and without potent phenotype were collected and frozen at −80°C. Samples were ground with liquid nitrogen using mortar and pestle and incubated with extraction buffer [0.055 M cetyl trimethyl ammonium bromide (CTAB), 1.4 M NaCl, 0.02 M EDTA, and 0.1 M tris-HCl (pH 8.0)] at 60°C for 30 min. Samples were cleaned with equal volume of chloroform and the supernatant precipitated with isopropanol and washed with 70% EtOH before their resuspension in deoxyribonuclease (DNAse)–free water. Samples were sent for sequencing using HiSeq 2000 System [100–base pair (bp) single-end sequencing]. Expression Omnibus accession number for DNA sequences is GSE149996. The SNPtrack pipeline (41) was used to identify single variant/nucleotide polymorphisms associated to Ler ecotype as well as all the heterozygous mutations present in the potent sample. Homozygous Ler polymorphisms in the potent sample were used to define Col-0 DNA islands associated to the potent phenotype. A heterozygous mutation present in the potent sample within the Col-0 DNA islands identified a dominant mutation in the DII domain of the IAA18 gene. This mutation was not found in the sample without the potent phenotype.

Plasmid construction and plant transformation

Transcriptional and translational fusion lines were generated with the three-fragment Invitrogen Gateway System (Carlsbad, California, United States). pDONR plasmids were first generated through BP reaction and then recombined into dpGreenBarT or dpGreenBarT plasmids through LR recombination. To generate pDONR p4p1 carrying the promoter of IAA18, a 3.5-kb promoter region from the start codon of IAA18 was amplified from Col-0 DNA by polymerase chain reaction (PCR) with the primers 5′-GGGGACAACTTTGTATAGAAAAGTTGCACGACGCCAGTGAAATAGTGT-3′ and 5′-GGGGACTGCTTTTTTGTACAAACTTGGTAGGATTTTTTTTAGAGGAACTACAGAA-3′. To generate pDONR p4p1 carrying the short version of the SKP2B promoter, a 0.5-kb promoter region from the start codon of SKP2B was amplified from Col-0 DNA by PCR with the primers 5′-ACAACTTTGTATAGAAAAGTTGAAGCTTTAAAAAATTAACGGATTAGT-3′ and 5′-ACTGCTTTTTTGTACAAACTTGCCTTGAAGCGGTTTCTTTGAT-3′. To generate the pDONR 221 genomic or coding sequence gene versions of IAA18 and potent, the corresponding regions were amplified from Col-0/potent DNA or complementary DNA (cDNA), respectively, by PCR using the primers 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTGCATGGAGGGTTATTCAAGAAAC-3′ and 5′-GGGG-ACCACTTTGTACAAGAAAGCTGGGTATCTTCTCATTTTCTCTTGCTTAC-3′. To generate pDONR 221 IAA28, the coding sequence of IAA28 was amplified from cDNA using the primers 5′-ACA AGT TTG TAC AAA AAA GCA GGC TCC ATG GAA GAA GAA AAG AGA TTG GAG C-3′ and 5′-AC CAC TTT GTA CAA GAA AGC TGG GTC TTC CTT GCC ATG TTT TCT AGG-3′. To generate pDONR 221 ARF7, the coding sequence of ARF7 was amplified from cDNA using the primers 5′-ACAAGTTTGTACAAAAAAGCAGGCTTTATGAAAGCTCCTTCATCAAATG-3′ and 5′-ACCACTTTGTACAAGAAAGCTGGGTTTCACCGGTTAAACGAAGTGG-3′ for the version with a stop codon and 5′-ACCACTTTGTACAAGAAAGCTGGGTTCCGGTTAAACGAAGTGCTG-3′ for the version without a stop codon. To generate pDONR 221 NLS, we used the primers 5′-ACAAGTTTGTACAAAAAAGCAGGCTGCATGGAGCAGAAGCTGATCTC-3′ and 5′-ACCACTTTGTACAAGAAAGCTGGGTAGAATCCTCGAGCGAATTCAT-3′. To generate pDONR 221 ER, we used the primers 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTGCATGAAGACTAATCTTTTTCTC-3′ and 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTAGGCCGAGGATAATGATAGGA-3′. To generate pDONR p2p3 N-Luciferase, we used the primers 5′-ACAGCTTTCTTGTACAAAGTGGAA-ATGGAAGACGCCAAAAACATAAAG-3′ and 5′-ACAACTTTGTATAATAAAGTTG-TTATCCATCCTTGTCAATCAAGGC-3′, and to generate pDONR p2p3 C-Luciferase, we used the primers 5′-ACAGCTTTCTTGTACAAAGTGGAA-TCCGGTTATGTAAACAATCCGGA-3′ and 5′-ACAACTTTGTATAATAAAGTTG-TTACACGGCGATCTTTCCGC-3′. pDONR p4p1 pARF7 and pDONR p2p3 Luciferase (2) and pDONR p4p1 35S and pDONR p2p3 3xYFP (42) were also used. The final constructs were performed using the IAA18/potent coding sequence versions except for pIAA18::IAA18/potent-3xYFP and split-luciferase fusions, which were made using the genomic versions. The construct pER8:potent was generated through traditional LR Gateway recombination using pDONR p221 potent (see above) and the pER8 destination plasmid (43). The construct RecARF7:AraYPET:ARF7 was generated through recombineering (24) by fusing the AraYPET fluorescent protein gene into the N-terminal part of ARF7 being carried in the plasmid JATY61O18. The primers used were 5′-TCAGATTATTTATTGGGTTTATTCTTCAGAGAAAGTAAAGTTGAGTGATCGGAGGTGGAGGTGGAGCT-3′ and 5′-AAACCTTCAACAGGATTAGGAGAAACTCCATTTGATGAAGGAGCTTTCATGGCCCCAGCGGCC-′. All constructs were sanger sequenced and transformed into Col-0 or indicated background by floral dip using Agrobacterium strain GV3101. pDEST22/pDEST32 constructs for yeast two hybrid assays and pMal-p2/pGEX-2T constructs for IAA18/potent and ARF7 expression in bacteria were generated through traditional LR Gateway recombination.

Chemical treatments and quantification assays

LR number during growth was quantified on the basis of morphology of emerged LRs or, when indicated, using the markers pWOX5::ER-YFP and pSCR::ER-GFP in a Leica M205FA fluorescence scope. Mutant or reporter line seedlings were incubated at 4 dpi on MS plates containing 0.25 μM indole-acetic acid (IAA), and the number of LRs was counted at 24, 48, or 72 hours upon treatment. Local auxin treatments upon the OZ were performed at 7 dpi using a 0- to 1-μl micropipette as indicated in (2). Mutant or reporter line seedlings were incubated at 4 dpi (or 6 dpi if indicated) on MS plates containing 0.1, 0.5, or 10 μM estradiol or the same estradiol concentrations plus 1 μM IAA for 3 days. To determine the dynamics of ARF7, IAA18, or potent protein levels, seedlings from corresponding translational fluorescent lines were added 0.01, 1, or 5 μM IAA or 5 μM IAA plus 5 μM MG132 and mounted on slides for observation under a microscope laser confocal at 1, 2, and 3 hours. Mock or control seedlings were added the equivalent volume or the solvent (dimethyl sulfoxide or ethanol) used to dilute IAA, estradiol, or MG132. To determine the dynamics of ARF7 transcripts, seedlings from the corresponding Luciferase line were incubated on MS plates containing 1 μM IAA or control MS plates for 1 hour, followed by luciferase assays The number of founder cells in mutant lines carrying pSKP2B0.5:ER-3xmCherry was determine through the quantification of the number of mCherry-marked cells under confocal laser microscopy.

Luciferase imaging and expression analysis

Plates were sprayed with 1 ml of 2.5 mM potassium luciferine (Gold Biotechnology, St. Louis, Mo.,, cat. no: LUCK-1) and then imaged using a Lumazone CA Automated Chemiluminescence System (Roper Bioscience), NightOwl II (Berthold), or Flumazone (Leica M205FA adapted with Hamamatsu EMCCD X2 camera). Time course analyses were taken using MetaMorph Microscopy Automation Software in a sequence of one bright-field image followed by a 3-min dark interval and then a chemiluminescence image with a 6-min exposure every 20 min for 24 hours. Luciferase expression movies were made by combining the frames, normally three frames/s, using MetaMorph Image Analysis Software. Expression was measured by selecting the region of interest (ROI) and quantifying the analog-digital units per pixel using the MetaMorph Image Analysis Software. When indicated, the luciferase measurements are referred to as the percent change with respect to its own control. The number of PBS was determined using the DR5::Luciferase reporter through the quantification of the number of sites with high expression relative to the adjacent regions along the primary root.

Confocal laser microscopy

For confocal laser microscopy, we used a Leica SP8 microscopy with the Leica Application Suite (Las AF Lite) X software or a vertical Zeiss LSM 880 with the ZEN 2.3 SP1 software. Roots were stained with propidium iodide (PI) as indicated. To investigate the expression of the different transcriptional or translational fluorescent protein fusions, we used the standard settings for the corresponding green fluorescent protein (GFP), YFP, or mCherry tags. For R2D2 measurements, all images were taken in the SP8 confocal on counting mode. Venus was excited at 488 nm and detected at 496 to 547 nm, while tdTomato was excited at 561 nm and detected at 565 to 615 nm. In addition, samples were stained with PI, which was excited at 561 nm and detected at 631 to 727 nm. The ratio between red/yellow (mDII/DII) was calculated measuring the mean value of the first six cells through the ROI tool of the Las AF Lite X software.

RNA-seq analyses of the OZ

In the experimental design, the OZ of the control was taken at the minimum expression levels of the oscillation based on the DR5::Luciferase reporter and simultaneously for potent and arf7-1, without considering the reporter expression, to avoid bias. DR5::Luciferase expression was followed using a Lumazone CA Automated Chemiluminescence System (Roper Bioscience). OZs were dissected by using an ophthalmological blade under a dissecting scope and immediately frozen in UltraPure RNA-free water (Invitrogen, Carlsbad, California, United States) using liquid nitrogen. Samples were ground in the same tube with an adapted pestle in the presence of liquid nitrogen, then 0.2 ml of RNAzol (Sigma-Aldrich, Deisenhofen, Germany) was added, and incubated for 15 min at room temperature. Following centrifugation (12,000g for 15 min at 4°C), 1.25 μl of 4-bromoanisole (BAM) (Sigma-Aldrich, Deisenhofen, Germany) was added to the supernatant, incubated for 5 min at room temperature, and centrifuged (12,000g for 10 min at 4°C). One microliter of Glycoblue (15 mg/ml; Thermo Fisher Scientific, Waltham, Massachusetts, USA) was then added to the supernatant, mixed with 300 μl of isopropanol, and precipitated overnight at 4°C. Samples were centrifuged (12,000g for 30 min at 4°C), and the pellet was rinsed with 70% ethanol and air dried. The pellet was resuspended in 8 μl of RNA-free water (Invitrogen, Carlsbad, California, United States) and used for library preparation following the protocol described by Picelli et al. (44). Libraries were sequenced using HiSeq2000 System (50-bp single-end sequencing). Expression Omnibus accession number for RNA-seq data is GSE149995. Analysis of differentially expressed genes was performed using TopHat and Cufflinks software (45). Gene Ontology enrichment analysis was performed using Chip-Enrich (42, 43), and the derived data were represented with the software MeV 4.9 (MultiEexperiment Viewer) (

Pull-down assays and Western blot analyses

We used the vectors pMal-p2 and pGEX-2T, carrying ARF7, IAA18, or potent genes to express protein fusions to the maltose binding protein (MBP) and/or glutathione S-transferase (GST) epitopes in Escherichia coli. Expression in E. coli was induced upon 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) supplementation, followed by protein extraction through sonication (3× 30-s pulse/30-s pause) in immunoprecipitation (IP) buffer [0.1 M Hepes, 0.3 M KCl, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 1 mM Triton X-100]. Amylose Magnetic Beads (NEB) were used for the purification of MBP-ARF7 and MBP control using recommendations from the manufacturer. Bound MBP-ARF7 or MBP was incubated with GST-IAA18, GST-potent, or GST protein extracts at 4°C with agitation for 3 hours. Next, beads were washed five times with IP buffer and used for Western blot assays using anti-MBP or anti-GST (Santa Cruz).

Yeast two-hybrid assay

Saccharomyces cerevisiae strain Hf7c cultures were grown at 28°C in standard or minimal growth media. For direct interaction testing, paired baits (ARF7) and preys (IAA18) in pDEST22/pDEST32 were cotransformed into Hf7c cells. Colonies were selected in solid media containing X-Gal. All experiments were performed in triplicate. Positive and negative controls were IAA18 bait paired with ARF7 prey, and bait or prey paired with the opposing empty vector, respectively. Bait-prey interactions were scored according to β-galactosidase activity.

Statistical analysis

Statistical differences were detected using SPSS Statistics 21 software (IBM). Wald χ2 was used to analyze homoscedasticity or heteroscedasticity among samples. Homoscedastic groups were analyzed using univariate general linear model (GLM) with least significant difference (LSD) post hoc, whereas heteroscedastic groups were analyzed using generalized linear model (GzLM). For analyses with two homoscedastic samples, we performed Student’s t test analysis. Significant differences were collected with 5% level of significance.


Supplementary material for this article is available at

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: Funding: This work was funded by the Ministerio de Economía y Competitividad of Spain (MINECO) and/or the ERDF (BFU2016-80315-P to M.A.M.-R., BIO2017-82209-R to J.C.d.P., and TIN2016-81079-R to A.R.-P.), the Comunidad de Madrid and/or ERDF and ESF (2017-T1/BIO-5654 to K.W. and S2017/BMD-3691 to A.R.-P.), the Howard Hughes Medical Institute and the NIH (R35-GM131725 to P.N.B.), the Fonds Wetenschappelijk Onderzoek (FWO Flanders) (G022516N, G020918N, and G024118N to T.B.), and the “Severo Ochoa Program for Centres of Excellence in R&D” from the Agencia Estatal de Investigacion of Spain [SEV-2016-0672 (2017–2021)] to K.W., P.P.-G., and M.A.M.-R. through CBGP. M.M. was supported by a postdoctoral contract associated to SEV-2016-0672, E.B.-A. by Ayudante de Investigacion contract PEJ-2017-AI/BIO-7360 from the Comunidad de Madrid, A.S.-C. and L.S.-R. by FPI contracts from MINECO (BES-2014-068852 and BES-2017-080155, respectively), J.C. by a Juan de la Cierva contract from MINECO (FJCI-2016-28607), P.P.-G. by a Juan de la Cierva contract from MINECO (FJCI-2015-24905) and Programa Atraccion Talento from Comunidad Madrid (2017-T2/BIO-3453), A.S. by a Torres Quevedo contract from MINECO (PTQ-15-07915), and K.W. by program PGC2018-093387-A-I00 from the Ministerio de Ciencia e Innovacion (MICIU). Author contributions: J.P.-R., E.B.-A., G.W., A.S.-C., H.D.G., J.C., P.P.-G., I.G., A.S., and L.S.-R. contributed to the acquisition and analyses of data. M.A.M.-R., J.P.-R., E.B.-A., G.W., J.C.d.P., P.N.B., H.D.G., and T.B. contributed to the conception and design of the experiments. M.R., M.M., K.W., and A.R.-P. performed the modeling and simulations. M.A.M.-R., K.W., M.R., M.M., J.P.-R., and E.B.-A. wrote the manuscript. P.N.B., J.C.d.P., and T.B. revised the manuscript. Competing interests: P.N.B. is the co-founder and Chair of the Scientific Advisory Board of Hi Fidelity Genetics Inc., a company that works on crop root growth. The authors declare that they have no other competing interests. Data and materials availability: Materials will be provided through material transfer agreements. Expression Omnibus accession number for DNA mapping sequences is GSE149996 and for RNA-seq data is GSE149995. All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Stay Connected to Science Advances

Navigate This Article