WRKY6 transcription factor restricts arsenate uptake and transposon activation in Arabidopsis

Abstract

Stress constantly challenges plant adaptation to the environment. Of all stress types, arsenic was a major threat during the early evolution of plants. The most prevalent chemical form of arsenic is arsenate, whose similarity to phosphate renders it easily incorporated into cells via the phosphate transporters. Here, we found that arsenate stress provokes a notable transposon burst in plants, in coordination with arsenate/phosphate transporter repression, which immediately restricts arsenate uptake. This repression was accompanied by delocalization of the phosphate transporter from the plasma membrane. When arsenate was removed, the system rapidly restored transcriptional expression and membrane localization of the transporter. We identify WRKY6 as an arsenate-responsive transcription factor that mediates arsenate/phosphate transporter gene expression and restricts arsenate-induced transposon activation. Plants therefore have a dual WRKY-dependent signaling mechanism that modulates arsenate uptake and transposon expression, providing a coordinated strategy for arsenate tolerance and transposon gene silencing.

Figure 1.

As(V) Represses and Delocalizes the Pi Transporter PHT1;1. (A) Diagram of the 2-kb PHT1;1 promoter region fused to the luciferase reporter gene (PHT1;1-LUC; top). Analysis of LUC activity (right panel) in transgenic plants expressing PHT1;1-LUC grown on 1 mM phosphate medium for 7 d, transferred to -Pi medium for 2 d, and finally to -Pi medium supplemented with 30 μM As(V) or 1 mM Pi for 16 h (left panel). (B) Kinetic study of LUC activity in response to 30 μM As(V) or 1 mM Pi in PHT1;1-LUC–expressing plants and in control transgenic lines CCA1-LUC and CAB-LUC. Values show mean ± sd. (C) Kinetic study of LUC activity in response to different As(V) concentrations in PHT1;1-LUC–expressing plants. Values show mean ± sd. (D) and (E) Confocal analysis of PHT1;1-GFP–expressing Arabidopsis root epidermal cells. (D) Five-day-old plants grown in -Pi medium were transferred to fresh -Pi medium or to medium containing 1 mM Pi (+Pi) or 30 μM As(V) [+As(V)] for 1.5 h in the dark. Roots were stained with propidium iodide (column 1) or with the endocytic tracer FM4-64 (columns 2 to 4). Arrowhead indicates colocalization of FM4-64 with PHT1;1-GFP in endosomes. Bars = 10 μm. (E) Five-day-old PHT1;1-GFP–expressing plants grown in -Pi medium were incubated (24 h) (-Pi) or exposed to 30 μM As(V) for 3 h [+As(V) 3 h] and 24 h [+As(V) 24 h] in the dark.

Figure 2.

PHT1;1 Expression and Its Membrane Localization Depend on As(V) Concentration. (A) Kinetic study of LUC activity in response to pulses of 30 μM Pi (±Pi; blue line), 30 μM As(V) [±As(V); green line], or to a continuous concentration of 30 μM As(V) (red line) in PHT1;1-LUC–expressing plants. Duration of each pulse and gaps between them were 1.5 h; during the gap, samples were washed with buffer to remove Pi and As(V) from the medium. Values show mean ± sd. (B) Analysis of PHT1;1-GFP localization after two pulses of 30 μM As(V) [±As(V)] in PHT1;1-GFP–expressing Arabidopsis root cells. Duration of each pulse and gaps between them were 1.5 h in the conditions as in (A). Bars = 10 μm.

Figure 3.

The ARE Contributes to Downregulation of the Pi Transporter in Response to As(V). (A) Diagram showing the relative sizes of the three classes of downregulated genes in response to 30 μM As(V) and 30 μM Pi identified in a microarray analysis. As(V) > Pi, genes preferentially downregulated by As(V); As(V) < Pi, genes preferentially downregulated by Pi; As(V) = Pi, genes downregulated equally in response to both. For microarray analysis, wild-type plants were grown on Johnson medium with 1 mM Pi for 7 d, transferred to -Pi for 2 d, and finally to -Pi medium supplemented with 30 μM As(V) or 30 μM Pi (8 h). (B) Analysis of ARE frequency in 0.5 kb of the promoter regions of genes in the As(V)>Pi and As(V)<Pi classes. The table shows observed and expected ARE numbers assuming random distribution. Significant overrepresentation is highlighted (χ² test, P < 0.05). (C) Diagram showing wild-type (PHT1;1p-I/II) and mutated versions (PHT1;1p-I/IIm, PHT1;1p-Im/II, and PHT1;1p-Im/IIm) of the PHT1;1 promoter region fused to luciferase (PHT1;1-LUC). The ARE in the PHT1;1 promoter (I and II; black) was mutated sequentially (Im and IIm; diagonal stripes) by PCR site-directed mutagenesis. (D) Kinetic analysis of LUC activity in transgenic plants expressing the PHT1;1p-I/II, PHT1;1p-I/IIm, PHT1;1p-Im/II, or PHT1;1p-Im/IIm constructs, in response to 30 μM As(V) (1.5 h). Values represent data from analysis of 12 independent lines of each construct; mean ± sd.

Figure 4.

WRKY6 Responds to As(V) and Represses the Pi Transporter PHT1;1. (A) Kinetic study of PHT1;1 and WRKY6 expression by qRT-PCR in wild-type plants exposed to 30 μM As(V). Values show mean ± sd. (B) qRT-PCR expression analysis of WRKY6 in wild-type plants in response to 30 μM As(V) pulses [±As(V)] or in response to 30 μM Pi pulses (±Pi); duration of each pulse and gap was 1.5 h. Values show mean ± sd. (C) qRT-PCR of PHT1;1 transcript in wild-type plants (Col-0), in the WRKY6-GFP–overexpressing line (OXWRKY6), and in wrky6-TDNA line grown in +Pi medium for 7 d, transferred to -Pi for 2 d and then to -Pi medium alone or with 20 μM As(V) (1.5 h). In the case of WRKY6-overexpressing lines, values show data from analysis of 10 independent lines. Values show mean ± sd. (D) ChIP assay of WRKY6-GFP seedlings and PHT1;1 promoter PCR amplification analysis. qPCR of ARE-containing fragments of the PHT1;1 promoter. Enrichment was calculated relative to wild-type plants. ACT8 was used as negative control. Values show mean ± sd. *P < 0.05 (Student’s t test).

Figure 5.

The ARE Mediates WRKY6 Repression of the Pi Transporter PHT1;1 and Confers the As(V) Tolerance Phenotype. (A) Kinetic analysis of transient LUC activity in N. benthamiana leaf discs agroinfiltrated with PHT1;1-LUC wild type or the mutated versions alone or with a WRKY6-GFP–overexpressing construct. Leaf discs were incubated in medium with 30 μM As(V) (1.5 h). Values show mean ± sd. (B) ChIP assay of WRKY6-GFP followed by qPCR of the PHT1;1 promoter. ChIP assays were performed in N. benthamiana leaf discs agroinfiltrated with PHT1;1-LUC wild type and the PHT1;1-LUC mutated version (PHT1;1p-Im/IIm), with WRKY6-GFP or GFP-overexpressing constructs. Values represent the x-fold enrichment of WRKY6-bound DNA of the PHT1;1 promoter in immunoprecipitated samples relative to total input DNA. ARE (PHT1;1p-I/II-LUC) or mutated ARE-containing fragments (PHT1;1p-Im/IIm-LUC) in the PHT1;1 promoter were amplified by qPCR using specific primers. Values show mean ± sd. *P < 0.01 (Student’s t test). Values show mean ± sd. (C) As(V) tolerance phenotype of wild-type (Col-0), OXWRKY6-GFP–overexpressing line (OXWRKY6), and wrky6-TDNA plants grown on 15 μM Pi supplemented with 15 μM As(V) for 7 d. (D) Intracellular arsenic concentration in Col-0, OXWRKY6, and wrky6-TDNA plants exposed to 5 μM As(V) (1 h). Values show mean ± sd. *p < 0.05 (Student’s t test). DW, dry weight.

Figure 6.

WRKY6 Restricts Transposon Expression. (A) Venn diagram showing relative proportion of As(V)-downregulated genes [Col+As(V) versus Col-As(V)] and Pi starvation-upregulated genes (Col-Pi versus Col+Pi) and degree of overlap. For comparison, only genes of the 12x135K platform (NimbleGen) that coincide with those in the ATH1 platform (Affymetrix) were considered (twofold [2x], FDR < 0.1, and twofold [2x], FDR <0.05, respectively). (B) Comparative transcriptome analysis of As(V)-responsive transposons in wild-type plants and in the WRKY6-GFP–overexpressing line (OXWRKY6). Total number and percentage is indicated of transposons induced and repressed in wild-type [Col+As(V) versus Col-As(V)] and in WRKY6-GFP–overexpressing plants [OXWRKY6+As(V) versus Col+As(V)] in response to As(V). Cutoff values for analysis of As(V)-responsive transposons were twofold (2x), FDR < 0.1, and for OXWRKY6-responsive transposons, 1.5-fold (1.5x), FDR < 0.2. The table shows the size of observed overlap and expected size, assuming random distribution. Significant overlaps are in bold (χ² test, P < 0.05). (C) qRT-PCR time-course analysis of transposon expression in wild-type plants exposed to 30 μM As(V). (D) qRT-PCR analysis of transposon expression in wild-type (Col-0), OXWRKY6, and wrky6-TDNA plants in response to 30 μM As(V) (1.5 h). Values show mean ± sd; *P < 0.05 (Student’s t test).

Figure 7.

WRKY6 Interacts with the W-Box in the Promoter Region of As(V)-Responsive Transposable Elements. (A) and (B) Kinetic analysis of LUC activity in N. benthamiana leaves agroinfiltrated with the promoter region of a transposable element At5g35030 fused to LUC (At5g35030p-LUC) (A) or At5g35030p-LUC in which the W-box was mutated (At5g35030pMUT-LUC) (B), alone or with a WRKY6-GFP–overexpressing construct. Leaf discs were incubated in medium with 30 μM As(V) (1.5 h). (C) ChIP assay of WRKY6-GFP followed by qPCR of the At5g35030 promoter. ChIP assays were performed in N. benthamiana leaf discs agroinfiltrated with At5g35030p-LUC and the At5g35030pMUT-LUC version, with WRKY6-GFP– or GFP-overexpressing constructs. Values represent x-fold enrichment of WRKY6-bound DNA of the At5g35030 promoter in immunoprecipitated samples relative to total input DNA. W-box–containing fragments (At5g35030p-LUC) or mutated W-box–containing fragments (At5g35030pMUT-LUC) of the At5g35030 promoter were qPCR amplified using specific primers. *P < 0.01 (Student’s t test). (D) ChIP assay of WRKY6-GFP–overexpressing plants, followed by qPCR of the W-box–bearing fragments of transposon promoters. Enrichment was calculated relative to wild-type plants. ACT8 was used as negative control. *P < 0.05 (Student’s t test). Values show mean ± sd.

Figure 8.

Activation of As(V)-Responsive Transposable Elements Is Independent of Changes in H3K9me2 and H3Ac Levels. ChIP analysis of As(V)-induced transposable elements in Arabidopsis Col-0 plants grown on 1 mM Pi medium for 7 d, transferred to -Pi medium for 2 d and finally to liquid -Pi medium, alone or supplemented with 30 μM As(V). The assay was performed using antibodies specific for H3K9me2 (A) or H3Ac (B), histone marks associated with repressive and active transcription, respectively. Levels of histone modifications in the genomic regions of As(V)-responsive transposable elements are represented relative to Ta3 in the case of H3K9me2 or to TUB8 for H3Ac. Values show mean ± sd.