Pattern dynamics in adaxial-abaxial specific gene expression are modulated by a plastid retrograde signal during Arabidopsis thaliana leaf development

Abstract

The maintenance and reformation of gene expression domains are the basis for the morphogenic processes of multicellular systems. In a leaf primordium of Arabidopsis thaliana, the expression of FILAMENTOUS FLOWER (FIL) and the activity of the microRNA miR165/166 are specific to the abaxial side. This miR165/166 activity restricts the target gene expression to the adaxial side. The adaxial and abaxial specific gene expressions are crucial for the wide expansion of leaf lamina. The FIL-expression and the miR165/166-free domains are almost mutually exclusive, and they have been considered to be maintained during leaf development. However, we found here that the position of the boundary between the two domains gradually shifts from the adaxial side to the abaxial side. The cell lineage analysis revealed that this boundary shifting was associated with a sequential gene expression switch from the FIL-expressing (miR165/166 active) to the miR165/166-free (non-FIL-expressing) states. Our genetic analyses using the enlarged fil expression domain2 (enf2) mutant and chemical treatment experiments revealed that impairment in the plastid (chloroplast) gene expression machinery retards this boundary shifting and inhibits the lamina expansion. Furthermore, these developmental effects caused by the abnormal plastids were not observed in the genomes uncoupled1 (gun1) mutant background. This study characterizes the dynamic nature of the adaxial-abaxial specification process in leaf primordia and reveals that the dynamic process is affected by the GUN1-dependent retrograde signal in response to the failure of plastid gene expression. These findings advance our understanding on the molecular mechanism linking the plastid function to the leaf morphogenic processes.

Figure 1. A simple mutual repression of genes mediated by mobile factors easily shifts the boundary between gene expression domains.

(A) Schematic illustration of the model of mutual repression and mobility. (B) The phase plane of the mutual repression system. Blue and red lines indicate nullclines of equations (1) and (2), respectively, without the diffusion terms. The filled circles with magenta and green indicate the stable steady states, named the AD-expressing state and the AB-expressing state, respectively. The open circle is an unstable steady state. (C–E) Simulation results of the mathematical model using three parameter sets: the symmetric parameter set between AD and AB (C); the asymmetric parameter sets (D, E). The parameter values are set to be p₁ = p₂ = 0.1, r₁ = r₂ = 2.0, d₁ = d₂ = 0.1, c₁ = c₂ = 2.0 and D_AD = D_AB = 0.1 (C), except for r₂ = 1.8 (D), r₁ = 1.8 (E).

Figure 2. The boundary between the FIL-expression and miR165/166-free domains shifts during leaf development.

(A–F) Confocal images of longitudinal (A–C) and transverse (D–F) sections showing FILpro:GFP (green) and 35Spro:miYFP-W (magenta) marker expression patterns at different stages: 50-µm-long (A), 200-µm-long (B) and 300-µm-long (C–F). Lower schematic illustrations represent each boxed region in (A–C). (D–F), A series of sections from a leaf of approximately 300 µm in height. The approximate heights of the observation plane from the leaf base are indicated in (D–F). Arrowheads indicate the distal (A–C) and marginal (D–F) tip cells. (G–R) Confocal (G, Q, R) and stereoscopic (H–P) images showing VENUS expression patterns (yellow and yellow-green) of the FILpro:CRE-GR 35Spro:loxP-Ter-loxP-VENUS system in the third leaves of 12-day-old plants. The timing of DEX treatment for CRE/loxP recombination is indicated at the bottom left of each panel. The confocal imaging planes are a transverse section of a shoot apex (G) and third leaves (Q, R). The red color represents chlorophyll fluorescence. “+” marks the meristem center in all figures. Scale bars represent 50 µm (A–G), 1 mm (H–P) and 100 µm (Q, R). ad, adaxial side; ab, abaxial side.

Figure 3. The enf2 mutant shows slow FMB shifting and an abaxialized leaf phenotype.

(A–E) Confocal images of transverse sections showing FILpro:GFP (green) and 35Spro:miYFP-W (magenta) marker expression in the wild-type (A) and enf2 (B–E) leaf primordia. The arrowheads indicate the provascular cells. (C–E), A series of sections from a leaf of approximately 300 µm in height. The approximate heights of the observation plane from the leaf base are indicated. The comparable WT data are Figure 2D–F. (F) FIL-expression area sizes (%, y-axis) at different stages (grouped by section area sizes, x-axis) of the wild-type and enf2 leaf primordia. Bars indicate standard errors. n.s., not significantly different; *, significantly different (p<0.05, t-test) between the wild type and enf2. (G, H) Seedlings of the wild type and enf2. (I, J) Scanning electron microscope images of leaf sections from the wild type and enf2. Scale bars represent 50 µm (A–E, I, J) and 1 mm (G, H). WT, wild type.

Figure 4. FMB shifting is quicker in the phb-1d/+ mutant than in the wild type.

(A–E) Confocal images of longitudinal (A) and transverse (B–E) sections showing FILpro:GFP (green) and 35Spro:miYFP-W (magenta) marker expression in phb-1d/+ leaf primordia. (C–E), A series of transverse sections from a leaf of approximately 300 µm in height. The comparable WT data are Figure 2D–F. (F, G) VENUS expression pattern (yellow and yellow-green) of FILpro:CRE-GR 35Spro:loxP-Ter-loxP-VENUS in phb-1d/+. A transverse section of leaf primordia (F) and a stereoscopic image of mature leaf (G). (H, I) Seedlings of the wild type and phb-1d/+. (J, K) Scanning electron microscope images of leaf sections in the wild type and phb-1d/+. Scale bars represent 50 µm (A–F), 1 mm (G–I) and 100 µm (J, K).

Figure 5. The ENF2 gene encodes a plastid-targeted protein expressed throughout the shoot apex.

(A) Schematic representation of the ENF2 gene. Boxes represent exons. The untranslated regions, the putative transit peptide and the PotD/F homology domain are highlighted in gray, green and blue, respectively. The mutations in the enf2-1 and enf2-2 alleles are indicated. LB, left boarder of T-DNA. Blue, pink and orange arrows indicate the primers for RT-PCR analysis (B). (B) RT-PCR using the primers represented as the blue and orange arrows (left), the blue and pink arrows (right). (C–E) Confocal images showing VENUS-ENF2 (green, C), chlorophyll (magenta, D) fluorescence and both (E) in petiole cells. (F, G) Confocal images showing a transverse (F) and longitudinal (G) sections of a shoot apex expressing ENF2pro:VENUS-ENF2 (green). Scale bars represent 50 µm (C–G).

Figure 6. Inhibition of plastid gene expression machinery retards FMB shifting and leads to narrow lamina formation.

(A–D) Seedlings of lincomycin- and erythromycin- treated wild type, and untreated flv mutant and wild type. (E–H) Confocal images of transverse sections showing FILpro:GFP (green) and 35Spro:miYFP-W (magenta) marker expression in leaf primordium of each above plant. The arrowheads indicate the provascular cells. Scale bars represent 1 mm (A–D) and 50 µm (E–H). Lin, lincomycin; Ert, erythromycin.

Figure 7. The plastid effects on FMB shifting and lamina expansion depend on the GUN1 gene.

(A–C) Seedlings of lincomycin-treated gun1, untreated enf2 gun1 and untreated gun1. (D–F) Confocal images of transverse sections showing FILpro:GFP (green) and 35Spro:miYFP-W (magenta) marker expression in leaf primordium of each above plant. The arrowheads indicate the provascular cells. Scale bars represent 1 mm (A–C) and 50 µm (D–F).

Figure 8. Model for FMB regulation by GUN1-dependent retrograde signal.

Most leaf cells express FIL and have miR165/166 activity just after leaf initiation. However, during the early developmental stages, the FIL-expressing and miR165/166-active cells switch the nuclear gene expression state to that expressing PHB-like genes, thus FMB shifts. When plastid gene expression machinery is functional (A), the gene expression switch in nuclei progress smoothly regardless whether GUN1 is functional or not. The pace of this gene expression switch is important for the full lamina expansion. When the plastid gene expression machinery is impaired (B), the GUN1-dependent retrograde signal affects the nuclei to delay or stop the gene expression switch. This plastid effect contributes to prevent the wide lamina expansion. Possibly, the GUN1-dependent retrograde signal regulates also other nuclear genes to repair the plastid condition. When the plastid gene expression machinery is impaired and the plant is devoid of the GUN1-dependent retrograde signal (C), the switch in nuclear gene expression progress normally and lamina expands despite the absence of photosynthetic activity.