Fig 1.
A mutually inhibitory interaction between auxin and cytokinin positions the auxin signalling maximum along the xylem axis.
(A) A schematic of the Arabidopsis root tip. (B) Localisation of AHP6 mRNA via in situ hybridization. (C) The cytokinin signalling domain reported by pARR5::GUS. (D) The auxin signalling domain reported by pIAA2::GUS. (E–G) Expression patterns of (E) PIN1; (F) PIN3; and (G) PIN7 fused with GFP. (H) Schematic of the parsimonious model showing the mutually inhibitory interactions between auxin and cytokinin which are proposed to generate complementary signalling domains. The network is active in all cells but shown in only two, representative of the procambium/pericycle and the xylem axis; the faint elements are proposed to be downregulated in the corresponding domain. (I) Simulation of a longitudinal section of a root shows uneven auxin accumulation throughout the stele, with outer cells (pericycle) accumulating higher levels of auxin (parameter settings and PIN localisation as described in [5]). Arrowheads mark the protoxylem position in cross sections; scale bars represent 10μm. In the model schematic, yellow arrows represent PIN transporters and pink arrows represent passive auxin influx.
Table 1.
General parameters.
Table 2.
Signalling parameters.
Table 3.
Metabolic parameters.
Table 4.
Auxin-driven cytokinin biosynthesis parameters.
Table 5.
Quantitative RT-PCR primers.
Fig 2.
Expression and regulation of the auxin transporters.
(A) Geometric and realistic cross sections used for the simulations. Cells were assigned a “cell type” (shown according to the colour code in Fig 1A) that determined the type and localisation of auxin transporters, in line with experimental data. (B) Localisation of the influx transporter. (C) Localisation of the efflux transporter in the outer layers. (D–F) localisation of PIN1, PIN3 and PIN7 in the stele of simulated roots, in accordance with experimental observations [3]. Membranes marked in blue have the potential to express a transporter; lighter colour indicates a lower maximum expression level. (G, H) Response curves for hormonal (G) upregulation and (H) downregulation of the auxin transporters. The expression of PIN3, which is simply upregulated by auxin, is Eup([auxin]); since PIN1 and PIN7 are upregulated by cytokinin and repressed by auxin (via AHP6), their expression level is calculated as Eup([cytokinin])*Edown([auxin]). (I) A schematic of the model used in the simulations.
Fig 3.
Expression pattern of the auxin importers.
(A–D) Longitudinal sections showing (A) AUX1-VENUS-YFP, (B) LAX1-VENUS-YFP, (C) LAX2-VENUS-YFP, and (D) LAX3-VENUS-YFP expression in the root tip. (E–K) Transverse sections with colour channels merged (E–H) and only the yellow channel (I–L) showing (E, I) AUX1, (F, J) LAX1, (G, K) LAX2, and (H, L) LAX3 expression in the root tip. Propidium iodide staining is shown in cyan.
Fig 4.
Regulation of the transporters required for the correct auxin pattern.
(A–D) The ‘DR5-like’ output of (A) static simulations with unregulated PIN expression in (B) wild type; (C) wol; or (D) cytokinin-treated roots. (E–G) The ‘DR5-like’ output of (E) ‘dynamic’ simulations, in which the PINs are regulated, in (F) wild type; (G) wol; or (H) cytokinin-treated roots. (I–K) The pattern of the auxin signalling reporter IAA2::GUS in (I) wild type; (J) wol; and (K) cytokinin-treated roots. Scale bars in cross sections represent 10μm. In the schematics, yellow arrows represent PIN transporters, pink arrows represent auxin influx via diffusive permeability, and red arrows represent an auxin influx transporter.
Fig 5.
Auxin leaks out of the stele via the apoplast.
(A) A diagram showing the orientations used in depicting the fluxes. Angular position is measured counter-clockwise from the line labelled 0°. Negative radial fluxes are inwards; positive are outwards. Negative angular fluxes are clockwise; positive are counter-clockwise. (B) The auxin flux integrated across the cells in each tissue layer. (C, D) Heatmap of auxin levels in the stele of (C) static and (D) dynamic simulations of wol. (E, F) Zoomed-in quadrant of the heatmap of the (E) static and (F) dynamic simulations of wol to facilitate comparison of apoplastic auxin levels.
Fig 6.
Active auxin import is required for correct protoxylem formation.
(A) Auxin concentration was plotted for the cells of the xylem axis, marked with a white line. (B) Plot of the auxin concentration along the xylem axis in simulations of wild type (black), pin7 (green), and the importer mutant (blue). px = protoxylem; mx = metaxylem. (C, D) Frequency of protoxylem defects in auxin importer mutants (C) without and (D) with a 48 h treatment of 10nm 6-benzylaminopurine. ‘Defective’ protoxylem has breaks in one or both strands; ‘normal’ protoxylem is continuous. Asterisks indicate a significant difference from wild type (Col-0). (E) Chloral hydrate cleared aux1 roots with long gaps in one (top) or both (bottom) protoxylem strands. Yellow arrows indicate protoxylem; white arrows indicate missing protoxylem. (F–H) Expression of DR5::GFP in the (F, G) aux1lax1lax2 mutant and (H) wild type. Scale bar: 10μm.
Fig 7.
Severely reduced cytokinin movement is required to establish a cytokinin gradient across the Arabidopsis root via diffusion from a local biosynthesis source in the xylem axis.
To establish a cytokinin gradient without altering the diffusion coefficient of cytokinin, we introduced bidirectional membrane permeability. Decreasing the membrane permeability (moving right within rows) concentrates cytokinin in the source cells (here, the xylem axis), but no morphogen gradient is established, because the diffusion within the apoplast evens out the distribution within the rest of the root. Decreasing the diffusion coefficient (downwards in columns) in the apoplast transforms the step-wise cytokinin distribution pattern into a true gradient. Parameter values are given in Table 6.
Table 6.
Cytokinin gradient parameters.
Fig 8.
Cytokinin travels rapidly in the Arabidopsis root.
(A–D) Confocal images showing expression of the fluorescent TCS::GFP cytokinin response marker following a (A) 0 h, (B) 2 h, (C) 6 h, and (D) 24 h cytokinin treatment. (E) Graph quantifying the intensity of TCS::GFP across the diameter of the Arabidopsis root following the cytokinin treatments. The scale bar is 50μm; the yellow boxes highlight the region within 48μm of the QC; error bars mark the 95% confidence interval.
Fig 9.
Apolar PIN1 localisation has only a minor effect on auxin accumulation in the xylem axis.
(A) Polar (left) and apolar (right) subcellular localisation patterns of PIN1. (B, C) ‘DR5-like’ output showing the auxin pattern along xylem axis in simulations of roots with (B) polar and (C) apolar PIN1 localisation. (D) Plot showing the auxin concentration in cells along the xylem axis of simulations with polar (black) and apolar (red) PIN1.
Fig 10.
An auxin circuit is generated in the stele by polar localisation of PIN1.
(A, B) A heatmap of the auxin flux in the stele of plants with polar (A) and apolar (B) PIN1 localisation with arrows depicting the overall flux pattern. The schematic models in the lower panels only show PIN1 localisation, since PIN3, PIN7, and the auxin importer are apolarly localised, as in all other simulations.
Fig 11.
Polar localisation of PIN1 favours AUX1 activation in the xylem-pole pericycle cells.
A 120-second pulse of AUX1 activation was provided to the pericycle cell marked with an asterisk (A, D, G, J). In simulations with apolar PIN1 and PIN7, this resulted in continued AUX1 activity and auxin accumulation in a specific pericycle cell, regardless of which cell received the pulse (B, E, H, K). By contrast, the pulse only led to persistent activation in xylem-pole pericycle cells in simulations with polar PIN1 and PIN7 (C, F, I, L). Auxin levels are depicted using a ‘DR5-like’ scale.
Fig 12.
Xylem-pole pericycle cells compete for auxin accumulation in simulations of geometric cross sections.
A simultaneous 120-second pulse of AUX1 activation in two pericycle cells marked with asterisks (A, D, G, J) results in auxin accumulation in only one cell. In simulations with apolar PIN1 and PIN7, the same cell accumulates auxin (B, E, H, K), whereas the identity of the winning cell varies in simulations with polar PIN1 and PIN7 (C, F, I, L).
Fig 13.
Xylem-pole pericycle cells compete for AUX1 activation in simulations of realistic cross sections.
A simultaneous 120-second pulse of AUX1 activation in two pericycle cells marked with asterisks (A, D, G, J) results in auxin accumulation in only one cell. In simulations with apolar PIN1 and PIN7, the same cell accumulates auxin (B, E, H, K), whereas the identity of the winning cell varies in simulations with polar PIN1 and PIN7 (C, F, I, L). Auxin levels are depicted using a ‘DR5-like’ scale.
Fig 14.
The spatial scale of the cytokinin gradients.
The spatial steady state profile of cytokinin in a source-diffusion-decay model, for different degradation rates. The lower panel is zoomed in (50x) to the dotted area within the upper panel; the root sections are drawn to scale for comparison with the gradient. Dotted lines mark the characteristic length of each gradient. The blue line shows the cytokinin distribution with degradation rate and diffusion coefficient similar to auxin; the red and green lines represent gradients that could be informative at the transverse scale of the root meristematic zone.