miRNAs mediate SnRK1-dependent energy signaling in Arabidopsis

Abstract

The SnRK1 protein kinase, the plant ortholog of mammalian AMPK and yeast Snf1, is activated by the energy depletion caused by adverse environmental conditions. Upon activation, SnRK1 triggers extensive transcriptional changes to restore homeostasis and promote stress tolerance and survival partly through the inhibition of anabolism and the activation of catabolism. Despite the identification of a few bZIP transcription factors as downstream effectors, the mechanisms underlying gene regulation, and in particular gene repression by SnRK1, remain mostly unknown. microRNAs (miRNAs) are 20-24 nt RNAs that regulate gene expression post-transcriptionally by driving the cleavage and/or translation attenuation of complementary mRNA targets. In addition to their role in plant development, mounting evidence implicates miRNAs in the response to environmental stress. Given the involvement of miRNAs in stress responses and the fact that some of the SnRK1-regulated genes are miRNA targets, we postulated that miRNAs drive part of the transcriptional reprogramming triggered by SnRK1. By comparing the transcriptional response to energy deprivation between WT and dcl1-9, a mutant deficient in miRNA biogenesis, we identified 831 starvation genes misregulated in the dcl1-9 mutant, out of which 155 are validated or predicted miRNA targets. Functional clustering analysis revealed that the main cellular processes potentially co-regulated by SnRK1 and miRNAs are translation and organelle function and uncover TCP transcription factors as one of the most highly enriched functional clusters. TCP repression during energy deprivation was impaired in miR319 knockdown (MIM319) plants, demonstrating the involvement of miR319 in the stress-dependent regulation of TCPs. Altogether, our data indicates that miRNAs are components of the SnRK1 signaling cascade contributing to the regulation of specific mRNA targets and possibly tuning down particular cellular processes during the stress response.

Keywords: Arabidopsis; DCL1; SnRK1; energy signaling; miRNA; stress.

Figure 1

The dcl1–9 mutant is partially compromised in the overall transcriptional reprogramming induced by starvation. (A) The dark-triggered transcriptional response encompasses more differences in WT than in dcl1–9 plants. Principal component analysis of normalized microarray data was performed using PARTEK®. The X axis represents the distance between control and dark-treated samples, the Y axis represents the distance between WT and dcl1–9 plants and the Z axis represents the distance between the 3 independent biological replicates. (B) Dark treatment triggers a “starvation response” in WT and dcl1–9, which is partially deficient in dcl1–9. Venn diagrams illustrate the intersection between genes activated (UP) or repressed (DOWN) similarly by SnRK1 activation and various starvation conditions [core starvation genes, core SGs, Table S4 in Baena-Gonzalez et al. (2007)], and genes significantly activated or repressed by darkness in WT and dcl1–9 leaves.

Figure 2

Reduced accumulation of MIR transcripts in response to darkness and SnRK1 activation relies partly on the energy status. (A) The energy status contributes to the decline of MIR transcripts in dark-treated leaves. Values represent fold-repression of MIR transcripts in the dark (D) and dark+sugar (D+S) relative to the light control. (B) SnRK1.1 activation in mesophyll protoplasts causes a reduction in MIR transcript levels. Values represent relative transcript levels upon transient overexpression of SnRK1.1 or control DNA. The induction of the SnRK1 marker gene DIN6 serves as control of SnRK1 activation by darkness (A) and SnRK1.1 overexpression (B). Relative mRNA levels were assessed by qRT-PCR, error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, paired t-test.

Figure 3

Different mechanisms underlie the dark-triggered decline of MIR transcripts. (A) Misregulation of specific MIR genes in the dcl1–9 mutant. Fold-reduction of MIR transcripts in response to darkness in WT and dcl1–9 leaves corresponds to the 90% confidence lower bound of fold-change, as calculated with dChip from the microarray hybridization data obtained from detached leaves incubated in the light or in darkness. (B) The dark-triggered reduction in MIR824A was confirmed by qRT-PCR. Values denote the fold-repression of MIR824A in the dark relative to the light in WT and dcl1–9 leaves. (C) The activity of MIR161 and MIR775A putative promoters is reduced by darkness and SnRK1.1 overexpression. LUC activity was measured as readout of promoter activity using the indicated proMIR::LUC fusion constructs. Activation of proDIN6::LUC is a positive control for activation of the SnRK1 pathway. LUC activities were normalized to GUS activities generated by the co-transfected UBQ10::GUS construct that served as an internal transfection control. Error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, unpaired t-test (B) or ratio t-test (C).

Figure 4

Identification and functional analysis of potential miRNA-regulated starvation genes. (A) Pipeline for the identification of starvation genes (SGs) misregulated in the dcl1–9 mutant. The list of genes with significant differential expression in response to darkness in the WT (“Dark-regulated”) was intersected with that of genes regulated by SnRK1 (Baena-Gonzalez et al., 2007). The response of the overlapping genes (SGs) was examined in the dcl1–9 mutant and those exhibiting at least 15% misregulation in dcl1–9 were selected as SGs misregulated in dcl1–9. (B,C) Functional clustering analysis of SGs misregulated in dcl1–9 reveals 18 and 15 enriched clusters (enrichment score ≥ 1.3) for the repressed (B) and induced (C) genes, respectively. (D,E) Functional clustering analysis of SGs misregulated in dcl1–9 that are validated or predicted miRNA targets reveals 5 and 4 enriched clusters (enrichment score ≥ 1.3) for the repressed (B) and induced (C) genes, respectively. TS, Supplementary Table.

Figure 5

Repression of TCPs by energy deprivation requires miRNA function. (A) Repression of TCPs and Hsp70-15 in dark-treated leaves is dependent on the energy status. Values represent fold-repression in the dark (D) and dark+sugar (D+S) relative to the light control. (B) SnRK1.1 overexpression in mesophyll protoplasts causes a reduction in TCP and Hsp70-15 levels. Values represent relative transcript levels upon transient overexpression of SnRK1.1 or control DNA. (C) Repression of TCPs and Hsp70-15 by energy deprivation is partly compromised in the dcl1–9 mutant. (D) Repression of TCPs but not of Hsp70-15 by energy deprivation is partly compromised in MIM319 plants. Values in (C) and (D) denote the fold-repression of transcripts in dark-treated as compared to light-treated leaves in the indicated genotypes. Relative mRNA levels were assessed by qRT-PCR, error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, paired (A,B) or unpaired t-test (C,D).

Figure 6

miRNAs are a novel component of the starvation response. Stress-derived energy deficiency activates the SnRK1 protein kinase which leads to a major transcriptional reprogramming partly via transcription factors and partly via miRNAs. SnRK1 activity may impact miRNA function at several levels, including regulation of MIR promoter activity. Components of the miRNA pathway may also influence the SnRK1 transcriptome e.g., through changes in transcript stability. miRNAs contribute to SnRK1 signaling mainly through repression of gene expression, targeting TCPs and Hsp70-15, and possibly impacting major cellular processes like translation.

Figure A1

MIR398C expression is not significantly affected by glucose. Dark treatment of leaves results in MIR398C repression, but this repression is not significantly altered by the addition of glucose. Values represent fold-repression in the dark (D) and dark+sugar (D+S) relative to the light control. Relative mRNA levels were assessed by qRT-PCR, error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, paired t-test.

Figure A2

MIR gene induction in response to darkness is independent of the energy status and SnRK1 activity. (A) Increase of MIR transcript levels in dark-treated leaves is independent of the energy status. Values represent fold-induction in the dark (D) and dark+sugar (D+S) relative to the light control. (B) SnRK1.1 activation in mesophyll protoplasts does not significantly alter MIR172A transcript levels. Values represent relative transcript levels upon transient overexpression of SnRK1.1 or control DNA. Relative mRNA levels were assessed by qRT-PCR, error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, paired t-test.

Figure A3

MIR159B repression by darkness is not DCL1-dependent. The dark-triggered reduction in MIR159B was measured by qRT-PCR from an independent set of three experiments. Values denote the fold-repression of MIR159B in the dark relative to the light in WT and dcl1–9 leaves. Relative mRNA levels were assessed by qRT-PCR, error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, unpaired t-test

Figure A4

Promoter analysis of MIR genes. PlantPAN (Plant Promoter Analysis Navigator) cis-motif analyses reveal abundant ATHB1 and ATHB5 (indicated with arrows) binding sites in the MIR161 (A) and to a lesser extent in the MIR775A (B) promoters.

Figure A5

SnRK1 activation is not impaired in the dcl1–9 mutant. SnRK1 activation through transient overexpression of SnRK1.1 was comparable in protoplasts of WT and dcl1–9 plants. The activation of the proDIN6::LUC and proDIN1::LUC reporters was used as readout of SnRK1 activity. LUC activities were normalized to GUS activities generated by the co-transfected UBQ10::GUS construct that served as an internal transfection control. Values represent the fold-induction of GUS-normalized LUC activities in response to SnRK1.1 overexpression relative to transfection with control DNA. Error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, paired t-test.

Figure A6

SnRK1 signaling is not generally compromised in MIM319 plants. Dark-treated leaves of MIM319 plants show a normal regulation of most SnRK1 marker genes (Baena-Gonzalez et al., 2007), with the exception of DIN6 and DIN1, whose induction is compromised. Values denote the fold-change of transcripts in the dark as compared to light in control plants expressing the empty vector and in MIM319 plants. Relative mRNA levels were assessed by qRT-PCR, error bars represent the standard error of the mean (SEM) from at least three independent experiments. p-values, unpaired t-test.