Control of bud activation by an auxin transport switch

Abstract

In many plant species only a small proportion of buds yield branches. Both the timing and extent of bud activation are tightly regulated to produce specific branching architectures. For example, the primary shoot apex can inhibit the activation of lateral buds. This process is termed apical dominance and is dependent on the plant hormone auxin moving down the main stem in the polar auxin transport stream. We use a computational model and mathematical analysis to show that apical dominance can be explained in terms of an auxin transport switch established by the temporal precedence between competing auxin sources. Our model suggests a mechanistic basis for the indirect action of auxin in bud inhibition and captures the effects of diverse genetic and physiological manipulations. In particular, the model explains the surprising observation that highly branched Arabidopsis phenotypes can exhibit either high or low auxin transport.

Fig. 1.

Pathways of auxin transport in a branching structure simulated using a canalization model (Eqs. 1 and 2). (A) Schematic representation of a metamer. Edge length of the blue square is proportional to auxin concentration; width of the red rectangle is proportional to the concentration of PIN proteins in the corresponding metamer face, and width of the lines entering or leaving a metamer is proportional to auxin flux. (B–F) Stable states in a branching structure consisting of 2 potential sources of auxin in the terminal and lateral positions, a branching node, and a sink in the proximal position. (B–C) Stable states of the system with a single auxin source in either terminal (B) or lateral (C) metamer. (D–F) Stable states of the system with 2 auxin sources. The terminal source has been established before (D), after (E), or near simultaneously (F) with the lateral source.

Fig. 2.

Apical dominance. (A–C) Schematic representations of an apex in the vegetative state (A), in the flowering state (B), and of the root (C). Other aspects are represented as in Fig. 1A. (D–F) Selected stages of the simulation of acropetal bud activation. The simulated plant initially consists of a vegetative apex and root (D). As the plant grows, auxin from the apex is transported basipetally and inhibits lateral buds close to the apex (E). Auxin levels decrease with distance from the growing apex; this decrease eventually switches lateral buds to the active state, producing an acropetal activation sequence (F). (G–I) Simulation of decapitation experiments. After decapitation of a growing plant (G), the lateral apex closest to the decapitation site is activated and becomes dominant (H). Several buds close to the decapitation site are activated in the case of overcompensation (I). (J) GUS expression (red arrowheads) detected with the chromogenic substrate X-Gluc, driven by the DR5 auxin-responsive promoter in an Arabidopsis stem section immediately below a growing lateral shoot, prepared as in ref. . Lateral shoots vascularize into 2 adjacent vascular bundles in the main stem (avb) (19, 22) lt: leaf trace.

Fig. 3.

Simulation models of basipetal and convergent patterns of bud activation controlled by auxin transport. Schematic representations are as in Figs. 1 and 2. (A–E) Selected stages of basipetal bud activation. During vegetative growth beginning with the initial structure (A), the main apex creates a sequence of metamers with associated lateral buds (B). The flow of auxin from the apex inhibits the outgrowth of lateral apices. Upon transition to flowering (C), the supply of auxin to the topmost metamer decreases, resulting in the activation of its lateral bud (transition from state D to C in Fig. 1). Auxin produced by these buds is exported into the main stem, inhibiting the bud below. After transition of the topmost bud to the flowering state, the next lateral bud becomes activated (D). The resulting relay process continues until it is stopped by the residual supply of auxin from the floral apices (E). (F) Convergent activation pattern resulting from a combination of this process with acropetal activation (Fig. 2 D–F).

Fig. 4.

The axr1 phenotype. (A) Simulation model. Compared to the WT (Fig. 3E), auxin concentrations and bud activation increase, but PIN levels are unaffected. (B) Auxin mass (pg) exported from the basal end of 2 cm stem segments excised from the apical and basal region of the bolting stems of glasshouse-grown WT (Col) and axr1–3 mutant (axr1) Arabidopsis plants. Error bars: SEM, n ≥ 4.

Fig. 5.

Arabidopsis max4 and tir3 mutants and their interactions. (A) Aerial phenotype of 6-week-old Columbia WT, max4–1, tir3–101, and max4–1,tir3–101 plants. (B) Mean auxin mass (pg) exported from 2 cm stem segments excised from the apical, medial, and basal region of the bolting stems of WT (Col) and max4–1 mutant (max4) plants. For the “Col no apices” samples, all visible shoot apices above the point of stem excision were removed 24 h before the stem segments were excised. The auxin was collected from the basal end of the segments, except in the “inverted” segment samples, where it was collected from the apical end. Error bars: SEM, n ≥ 4. (C) Mean radiolabelled auxin (cpm) transported along excised stem segments of the genotypes shown in (A). Error bars: SEM, n = 10. (D) Simulation of max mutant. PIN concentrations, auxin fluxes and auxin concentrations are increased with respect to the WT (Fig. 3E). (E) Simulation of tir3 mutant. PIN concentrations, auxin fluxes, and auxin concentrations are decreased with respect to the WT (Fig. 3E). (F) Simulation of max, tir3 double mutant. PIN concentrations, auxin fluxes, and auxin concentrations are increased with respect to the WT (Fig. 3E), but decreased compared to max (D).

Fig. 6.

Bud activation at the cellular level: simulations (A–F) and in planta data (G and H). (A) Schematic representation of a cell. Auxin concentration is shown as the intensity of blue, PIN as the intensity of red, and the dominant direction and magnitude of auxin flux is indicated by a black arrow. (B–F) Selected stages of the simulation. A longitudinal section of a stem with 2 buds is represented schematically as a grid of cells, approximating the shape of the section. Source cells are outlined in green; the sink is outlined in red. In the initial state there is small background production of auxin in every cell, and a single sink cell is present at the base (B). Following the placement of an auxin source at the top of the main stem, a vascular strand running through the stem emerges (C). Subsequent placement of auxin sources in the 2 buds does not trigger formation of lateral vasculature (D) until the auxin source at the top of the main stem is removed (E). Removal of the source in the upper bud causes the lower bud to be activated (F). (G and H) PIN1 localization in Arabidopsis axillary bud stems. Hand sections through the stems of inhibited (G) and active (H) axillary buds carrying the PIN1 protein fused to GFP under the control of the native PIN1 promoter. Arrows indicate an example of basally localized GFP in a file of cells. (Scale bar: 50 μm.)