Regulation of flowering time and floral organ identity by a MicroRNA and its APETALA2-like target genes

Abstract

MicroRNAs (miRNAs) are approximately 21-nucleotide noncoding RNAs that have been identified in both animals and plants. Although in animals there is direct evidence implicating particular miRNAs in the control of developmental timing, to date it is not known whether plant miRNAs also play a role in regulating temporal transitions. Through an activation-tagging approach, we demonstrate that miRNA 172 (miR172) causes early flowering and disrupts the specification of floral organ identity when overexpressed in Arabidopsis. miR172 normally is expressed in a temporal manner, consistent with its proposed role in flowering time control. The regulatory target of miR172 is a subfamily of APETALA2 (AP2) transcription factor genes. We present evidence that miR172 downregulates these target genes by a translational mechanism rather than by RNA cleavage. Gain-of-function and loss-of-function analyses indicate that two of the AP2-like target genes normally act as floral repressors, supporting the notion that miR172 regulates flowering time by downregulating AP2-like target genes.

Figure 1.

EAT Overexpression Phenotype. (A) Wild-type (Columbia [Col-0] ecotype) plant, 3.5 weeks old. (B) eat-D plant, 3.5 weeks old. (C) Wild-type flower. (D) eat-D flower. Note the absence of second-whorl organs (petals). The arrow indicates a sepal with ovules along the margins and stigmatic papillae at the tip. The eat-D phenotype is virtually identical to that of strong ap2 mutant alleles. (E) Cauline leaf margin from a 35S-EAT plant. Arrows indicate bundles of stigmatic papillae projecting from the margin. (F) Solitary gynoecium (arrow) emerging from the axil of a cauline leaf of a 35S-EAT plant.

Figure 2.

EAT Contains a miRNA That Is Complementary to AP2. (A) Location of EAT on chromosome 5. The T-DNA insertion and the orientation of the 35S enhancers are indicated. The 21-nucleotide sequence corresponding to miR172a-2 is shown below the EAT gene. (B) Structure of the EAT precursor transcript. The polyadenylated and 5′ capped, unspliced transcript is shown, along with the positions of two introns. Two alternatively spliced transcripts, one with both introns removed and the other with intron 2 spliced out, also accumulate in wild-type plants. The location of the miR172a-2 miRNA is indicated with an asterisk. (C) The putative 21-nucleotide miR172a-2/AP2 RNA duplex is shown below the gene structure of AP2. Two mismatches are indicated in boldface type. The boxed area delineates the coding region (ATG to stop) of the AP2 transcript, and the black boxes show the locations of the two AP2 domains. (D) Alignment of the AP2 21-nucleotide target site (black bar at top) and surrounding sequence with three other Arabidopsis AP2 family members and two maize AP2 genes (IDS1 and GL15). Nucleotide positions of the sequences within each cDNA sequence are indicated at right. Black boxes indicate nucleotides that are highly similar to other family members, and gray boxes show nucleotides that are less well conserved. (E) Alignment of miR172a-2 miRNA (black bar at top) and surrounding sequence with other miR172 sequences from Arabidopsis (miR172a-1, miR172b-1, miR172b-2, and miR172c), tomato (AI484737), soybean (BI320499), potato (BQ114970), and rice (AP003277, AP004048, AP005247, and ctg7420). Black and gray boxes are as described for (D).

Figure 3.

miR172 Expression. (A) RNA gel blot of 50 μg of total RNA from wild-type Col-0 (lanes 3 and 7) and eat-D (lanes 4 and 8). RNA was isolated from the aboveground parts of plants that had already bolted. Blots were probed with sense (lanes 1 to 4) or antisense (lanes 5 to 8) oligonucleotides to miR172a-2 miRNA. One hundred picograms of sense oligonucleotide (lanes 2 and 6) or antisense oligonucleotide (lanes 1 and 5) was loaded as a hybridization control. Nucleotide length markers are indicated at left. The arrowhead shows the putative hairpin precursor of miR172a-2. Ethidium bromide–stained 5S RNA served as a loading control for lanes 3, 4, 7, and 8. (B) S1 nuclease mapping of miR172a-2 miRNA. A 5′ end labeled probe was undigested (lane 1) or digested after hybridization to total RNA from wild-type (lane 2), eat-D (lane 3), or tRNA (lane 4) plants. RNA samples in lanes 2 and 3 were isolated from aboveground parts of plants that had already bolted. Nucleotide length markers are indicated at left. (C) RNA gel blot of 30 μg of total RNA from wild-type Col-0 (Col), luminidependens (ld), or constans (co) probed with antisense oligonucleotide to miR172. Plants were harvested at 16 days after germination. 5S RNA, loading control. (D) RNA gel blot analysis of 30 μg of total RNA from wild-type seedlings harvested at 2, 5, 9, 12, and 16 days after germination or from mature leaves (L) and floral buds (F). Probes for miR172 and 5S RNA are indicated at left. (E) Quantification of miR172 transcript abundance (normalized to 5S RNA) with respect to developmental time (i.e., days after germination). The normalized miR172 RNA level at 16 days after germination was set arbitrarily to 1.

Figure 4.

miR172 Downregulates AP2. (A) At top and bottom left are RNA gel blots of 1 μg of poly(A⁺) RNA from the wild type (wt) and eat-D. RNA from each genotype was isolated from floral buds. The blots were hybridized with probes specific for the AP2-like genes indicated below or at left. At bottom right are immunoblots of proteins from wild-type (wt) and eat-D floral buds probed with AP2 antibody. AP2 is severely reduced in eat-D. RbcL, large subunit of ribulose-1,5-bisphosphate carboxylase as a loading control. (B) Target sites for miR172 in three AP2-like transcripts, with the 5′ ends of RNA cleavage products that were identified by 5′ RACE PCR indicated by arrows. The numbers of 5′ RACE clones sequenced that correspond to each cleavage product are indicated.

Figure 5.

Activation-Tagged and Loss-of-Function Alleles of AP2-Like Target Genes. (A) Location of the T-DNA insert in toe1-1D. The 35S enhancers are located ∼5 kb upstream of TOE1 (At2g28550). (B) Semiquantitative RT-PCR analysis of TOE1 and ACT11 expression in the wild type versus toe1-1D. RNA from each genotype was isolated from floral buds. Inclusion of reverse transcriptase (RT) is indicated by + or − above the gels. (C) eat-D suppresses the phenotype of toe1-1D. Flowering-time phenotypes (expressed as rosette leaf number) of F1 plants derived from the cross of a toe1-1D heterozygote to an eat-D homozygote. Approximately 50% of the F1 plants contained the toe1-1D allele, based on PCR genotyping; these plants were early flowering, indicating the suppression of toe1-1D (which overexpresses TOE1) by eat-D (which overexpresses miR172). In this experiment, wild-type control plants flowered with ∼11 leaves. (D) Locations of the T-DNA inserts in toe1-2 and toe2-1. Exons are indicated by solid bars, and the target sequences for miR172 are indicated with asterisks. Brackets delineate the two AP2 domains contained in each gene.

Figure 6.

Model Depicting the Temporal Regulation of Flowering by miR172. In this model, the temporal expression of miR172 causes temporal downregulation of the AP2-like target genes at the level of translation, which triggers flowering once the target proteins decrease below a critical threshold (dotted line). The onset of floral initiation is indicated by the black bar.