Root gravitropism is regulated by a transient lateral auxin gradient controlled by a tipping-point mechanism

DII-VENUS Reveals Dynamic Changes in Root Auxin Distribution After a Gravity Stimulus.

The DII-VENUS reporter was initially used to qualitatively monitor gravity-induced changes in auxin distribution and response in root tissues. Confocal imaging revealed that 60 min after a 90° gravity stimulus the DII-VENUS signal was reduced in lateral root cap (LRC) and epidermal cells on the lower side of root (Fig. 1C and SI Appendix, Fig. S1). This reduction included LRC cells positioned on the lower side of gravity-sensing columella cells and extending back to the start of the elongation zone. Conversely, on the upper side of the root the DII-VENUS signal was maintained in LRC and epidermal cells, including several cells in the elongation zone. In contrast to the asymmetric DII-VENUS pattern observed in LRC and epidermis, the reporter signal remained symmetric in inner root tissues. Hence, gravity-induced changes were most apparent in the LRC and epidermal cells, consistent with our earlier functional studies (21). One hundred twenty minutes after the gravity stimulus (when root tips had reorientated through ∼46°), the DII-VENUS signal in LRC and epidermal cells on the lower side of the root started to increase, whereas on the upper side of the root the reporter signal remained high in LRC and epidermal cells (Fig. 1C). By 240 min (tip angle ∼ 69°), the DII-VENUS signal in LRC and epidermal cells was equal on both sides of the root. This dynamic pattern contrasted the lack of asymmetry observed either using a stabilized version of DII-VENUS (mDII-VENUS) (Fig. 1C and SI Appendix, Fig. S2) or using GUS translational fusions to TIR1/AFB1-3 receptor sequences (22) (SI Appendix, Fig. S3). Hence, the asymmetric DII-VENUS signal appears to result from gravity-induced changes in auxin distribution, rather than sensitivity.

Auxin-Induced DII-VENUS Degradation Is Dose Dependent.

Although our confocal analysis using DII-VENUS gave qualitative information about the gravity-induced auxin dynamics, we still lacked important knowledge of the timescales and fold changes that characterize hormone redistribution. To address these issues, we first determined the quantitative relationship between auxin abundance and reporter signal. Transgenic roots expressing DII-VENUS were exposed to a range of concentrations of exogenous auxin (1 nM to 1 μM) and the reduction in the YFP reporter signal was quantified using time-lapse confocal microscopy (Fig. 2A). This study highlighted a clear relationship between auxin abundance and reporter signal (Fig. 2A). The dose–response experiments demonstrated that DII-VENUS is most sensitive in a physiological range of auxin concentrations (22). Our measurements revealed that even small changes in auxin concentration from 1 to 5 nM are able to trigger DII-VENUS degradation (Fig. 2A). Hence, DII-VENUS can be considered a quantitative reporter for auxin abundance.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

(A) Quantification of the dose- and time-dependent degradation of the DII-VENUS signal by exogenous auxin concentrations (IAA, 1–1,000 nM); points are means ± 1 SE of at least eight independent replicates (∼100 nuclei quantified in each replicate); solid lines show the fitted model results. (B) Schematic model of the network of interactions that relate the DII-VENUS reporter to auxin. (C) The equations that describe the model dynamics: a system of four ODEs coupled to an equation representing the conservation of TIR1/AFB receptors. The total concentration of TIR1/AFB receptors is denoted by [TIR1]_T and the remaining parameters are labeled in the schematic in B. (D) The reduced model equations: a single ODE that depends on four groupings of the model parameters and the auxin influx rate. Parameter grouping represents the ratio between the auxin influx during auxin treatment and the basal (steady-state) auxin influx, α₀.

Parameterized Mathematical Model Captures the Relationship Between Auxin and DII-VENUS.

To precisely determine the relationship between DII-VENUS and auxin concentration, we developed a mathematical model of DII-VENUS degradation in response to auxin (Fig. 2B). Assuming that the total number of TIR1/AFB receptors is constant (data in SI Appendix, Fig. S3), the network dynamics can be described by a system of four ordinary differential equations (ODEs) (Fig. 2C and SI Appendix). This model shows how the DII-VENUS dynamics depend on the association and dissociation rates of the various complexes; total concentration of TIR1/AFB; influx, efflux, and decay of auxin; and the production and decay of the DII-VENUS signal (Fig. 2 B and C). To model the dose–response experiment, we supposed that the exogenous auxin causes an increase in the auxin influx rate. The precise relationship between auxin dose and influx rates cannot be easily deduced, however, and we therefore treated the auxin influx rates as unknown parameters.

We fitted the model parameters using the measurements from the dose–response experiment (Fig. 2A) (23). Although the model captured the DII-VENUS dynamics observed (SI Appendix, Fig. S15), these data do not suffice to constrain the parameter values tightly. This finding motivated the development of a simplified version of the model. By assuming that, on the timescale of DII-VENUS degradation, the concentrations of auxin, TIR1/AFB, and their complexes can be approximated by quasi–steady-state expressions, we reduced the full model (Fig. 2C) to a single ODE that describes how the DII-VENUS dynamics depend on the auxin influx and four parameter groupings (Fig. 2D) (SI Appendix). The values of these groupings were well constrained by the dose–response data, and the corresponding model simulations captured the key features of the dose–response dynamics (compare solid lines with individual data points in Fig. 2A). Although we obtained a slightly better agreement between the full model and the data, these differences with the simplified model are relatively minor and do not affect the results of the analysis, as stated in SI Appendix. From reliability analysis (24), we found that the parameter values obtained are reliable at the 95% confidence level. The estimate of the value of parameter suggested that ∼6% of the SCF-TIR1 is bound with DII-VENUS, implying that the reporter has minimal impact on the rate at which auxin-SCF-TIR1 complexes bind with other Aux/IAA proteins, and hence should not interfere with auxin perception and gravitropic processes. This result was consistent with the observation that the root gravitropic bending rate of the DII-VENUS transgenic line was equivalent to wild type (9).

Gravity Stimulus Triggers Auxin Redistribution to the Lower Half of the Root Tip.

Having captured the quantitative and temporal relationship between DII-VENUS and auxin levels in our network model (Fig. 2B), we used live imaging to follow dynamic changes in the DII-VENUS distribution at the root apex during a root gravitropic response. To image DII-VENUS throughout a gravitropic response we developed a confocal microscope capable of scanning through vertically growing roots (Materials and Methods). Gravity-induced dynamic changes in the DII-VENUS signal on either side of the root apex were detected by quantifying the reporter in ∼100 root cell nuclei in the upper and lower regions of root tips (SI Appendix, Fig. S4). For vertically growing root tips, no significant asymmetry in the DII-VENUS signal was observed when quantified over a 2-h period (SI Appendix, Fig. S4). However, consistent with our observations using an inverted microscope (Fig. 1C), when seedlings were provided with a 90° gravity stimulus, rapid changes in the ratio of the DII-VENUS signal in cells on the upper vs. lower side of the root could be detected (Fig. 3A and SI Appendix, Figs. S5 and S6). This DII-VENUS asymmetry continued to increase, reaching approximately twofold difference at the midpoint of the bending response before being lost as the root tip angle returned to vertical (Fig. 3A and SI Appendix, Figs. S5 and S6).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

(A) DII-VENUS reporter roots show an asymmetric distribution following a gravity stimulus, peaking in the middle of the response. The angle of the root tip is shown on each image following a 90° reorientation. The ratios of fluorescence intensity between the upper and lower sides of the roots are also shown. (Scale bar: 50 μm.) (B) The fold change in DII-VENUS ratio (upper/lower tissue signals) plotted against time following a 90° gravitropic stimulus. Black crosses show experimental data points (n = 96), and lines show the fitted simulation results: Red, blue, and green lines are, respectively, the signal from the lower side, the signal from the upper side, and the ratio. (C) Dynamic changes in auxin distribution between upper and lower root tissues following a gravity stimulus (t = 0). (D) Transcript profiling of several auxin-inducible genes in root apical tissues minutes after a 90° gravitropic stimulus. Data are mean ± 1 SE of four technical and four biological replicates, representing ∼180 roots. (E) The auxin redistribution in terms of the angle of reorientation following a gravity stimulus.

Lateral Auxin Gradient Is Formed Within Minutes of a Gravity Stimulus.

Determining the auxin redistribution dynamics that created the observed gravity-induced changes in the DII-VENUS distribution (Fig. 3B) required the parameterized mathematical model (Fig. 2D). A key question is whether gravity induces a change in the total amount of auxin in the root; however, no change was detected in root tips 0, 2, and 4 h after a gravity stimulus (SI Appendix, Fig. S7). Thus, assuming the total auxin flux from the root tip is constant, we simulated the DII-VENUS dynamics corresponding to a wide range of auxin redistribution dynamics and used parameter estimation to determine which auxin dynamics fitted the DII-VENUS measurements (SI Appendix). These simulations revealed that auxin is redistributed from cells on the upper to the lower side of the root tip within 5 min of a gravity stimulus (Fig. 3C), a timescale that is in close agreement with experimental measurements of auxin-induced asymmetric changes in root pH and calcium following a gravity simulus (25). In contrast, root bending is first detected 10–20 min after a gravity stimulus (9). Hence, root bending occurs after lateral auxin gradient formation, consistent with the hypothesis that auxin redistribution regulates organ curvature (8, 20).

Gravity-Induced Lateral Auxin Gradient Is Dependent on Statolith Sedimentation.

Specialized starch-filled organelles termed statoliths move within 5 min of a gravity stimulus in gravity-sensing columella cells (26, 27). It has been proposed that statolith sedimentation onto the lower side of these cells triggers the formation of the lateral auxin gradient (28, 29). To establish a link between statolith sedimentation and auxin redistribution, we monitored gravity-induced changes in DII-VENUS in the starchless pgm-1 (30) mutant that exhibits impaired gravitropic responses. Confocal imaging revealed that gravity-induced DII-VENUS gradient formation was severely disrupted in pgm-1 (Fig. 4B and SI Appendix, Figs. S8 and S9). Epidermal and LRC cells failed to form a gradient of DII-VENUS 2 h after a gravity stimulus when the difference in reporter signal between upper and lower wild-type root tip tissues response is at a maximum (compare Fig. 4A with 4B). Our results demonstrate a clear link between statolith sedimentation and auxin redistribution.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

DII-VENUS signal 2 h after a gravity stimulus in root tissues of (A) wild type and (B) starchless mutant pgm-1. (Scale bar: 50 μm.) (C–E) Statoliths visualized within columella cells of root tips at angles of (C) 0°, (D) 40°, and (E) 90° to the horizontal, with schematics of “tipping point” model where statoliths within columella cells redirect auxin distribution at the root tip (denoted by arrows).

Hormone Redistribution Triggers Auxin-Responsive Gene Expression.

The modeling revealed that during a gravitropic response, auxin levels increase by approximately twofold on the lower side of the root (Fig. 3C). This fold change in auxin contrasts with predictions that differences of an order of magnitude are necessary to trigger a gravitropic response and asymmetric root growth (31). Instead, our results imply that there is a redirection of auxin flux from both sides of the apex to tissues on the lower side of the root. This auxin redistribution appeared sufficient to induce changes in auxin-responsive gene expression (Fig. 3D). RT-qPCR profiling of several auxin-inducible transcripts (IAA1, IAA2, and ARF19) (32) using RNA extracted from microdissected root tips (Materials and Methods) revealed that the abundance of these transcripts was elevated after a gravity stimulus (Fig. 3D). Thus, in agreement with our dose–response measurements, these results revealed that small fold changes in auxin concentration are able to trigger Aux/IAA degradation (Fig. 2A). We conclude that gravity-induced fold changes in IAA appear sufficient to induce auxin and gravitropic responses.

Lateral Auxin Gradient Forms Transiently and Dissipates After the Bending Midpoint.

Unexpectedly, using the model to interpret the observed DII-VENUS dynamics revealed that auxin asymmetry was maintained for only a short time (∼100 min) relative to the overall duration of the root gravitropic bending response (∼600 min) (Fig. 3C). After ∼100 min, auxin distribution became symmetric with equal fluxes to the lower and upper root sides (Fig. 3C). Our RT-qPCR expression profiling provided independent validation of this result. We observed that the abundance of auxin-responsive transcripts was elevated minutes after a gravity stimulus and then returned to lower levels after 1–2 h (Fig. 3D), consistent with the lateral auxin gradient existing for a finite period.

Statoliths Function as Tilt Switches That Redirect Auxin Flow at the Root Tip.

How does auxin asymmetry return to normal? Plotting gravity-induced changes in auxin asymmetry vs. root tip angle revealed that loss of auxin asymmetry occurred at a root tip angle of ∼40° following a 90° gravistimulus (Fig. 3E). This result may represent the angle statoliths move within columella cells at the midpoint of a root gravitropic response (33). Live imaging revealed that the initial 90° gravity stimulus causes statoliths to be displaced from the rootward (apical) end of columella cells to the new lower cell wall (26). However, at a root angle of ∼40° statoliths return to the rootward face of columella cells (Fig. 4 C–E). Statoliths therefore appear to function like tilt switches. The rapid reduction in auxin asymmetry after roots reach an angle of ∼40° (Fig. 3E) is likely to reflect the redirection of auxin flux at the root apex following statolith repositioning (Fig. 4 C–E). Consistent with this mechanism, the lateral auxin gradient could be maintained for at least 8 h by repositioning gravity-stimulated root tips above the midpoint following a 90° gravistimulus (SI Appendix, Fig. S10). Conversely, reorienting gravity-stimulated root tips below the midpoint led to premature loss of the lateral auxin gradient (SI Appendix, Fig. S11). We conclude that roots use a “tipping point” mechanism that reverses asymmetric auxin flow at the midpoint of root bending.