ARF2 represses expression of plant GRF transcription factors in a complementary mechanism to microRNA miR396

Abstract

Members of the GROWTH REGULATING FACTOR (GRF) family of transcription factors play key roles in the promotion of plant growth and development. Many GRFs are post-transcriptionally repressed by microRNA (miRNA) miR396, an evolutionarily conserved small RNA, which restricts their expression to proliferative tissue. We performed a comprehensive analysis of the GRF family in eudicot plants and found that in many species all the GRFs have a miR396-binding site. Yet, we also identified GRFs with mutations in the sequence recognized by miR396, suggesting a partial or complete release of their post-transcriptional repression. Interestingly, Brassicaceae species share a group of GRFs that lack miR396 regulation, including Arabidopsis GRF5 and GRF6. We show that instead of miR396-mediated post-transcriptional regulation, the spatiotemporal control of GRF5 is achieved through evolutionarily conserved promoter sequences, and that AUXIN RESPONSE FACTOR 2 (ARF2) binds to such conserved sequences to repress GRF5 expression. Furthermore, we demonstrate that the unchecked expression of GRF5 in arf2 mutants is responsible for the increased cell number of arf2 leaves. The results describe a switch in the repression mechanisms that control the expression of GRFs and mechanistically link the control of leaf growth by miR396, GRFs, and ARF2 transcription factors.

Figure 1

Analysis of the miR396 target site in GRF transcription factors. A, Tree representation of 36 species with publicly available genome sequences (Adapted from Phytozome v12.1); each box denotes a gene encoding a GRF transcription factor. Blue boxes represent GRFs harboring the consensus site for miR396, as indicated in (B); light blue boxes indicate a GRF harboring a miR396-binding site with three changes or less with respect to the consensus sequence, whereas brown boxes show GRFs with more than three changes. B, Scheme representing a typical GRF gene. The consensus target site for miR396 (including the amino acids encoded) and the interaction with Arabidopsis miR396a is shown below. Mfe ratio: mfe ratio between each site and a perfectly complementary site. C, Interaction of miR396 with GRFs that have deviations from the miR396-consensus site (three changes or less). The amino acids encoded by the miR396-binding sites are indicated on the left. Genes with the same changes in the miR396-binding site are presented together. Gorai004G204600 (G. raimondii); Manes05G043700, Manes12G117600 (M. esculenta); Solyc08g79800, Solyc09g009200 (S. lycopersicum); DCAR_011951, DCAR_008567, DCAR_013369 (D. carota); SapurV1A.0045s0430 (S. purpurea); Potri.003G118100, Potri.019G042300 (P. trichocarpa), Cucsa.141640 (C. sativus); MDP0000842815 (M. domestica); Kaladp0094s0052 (K. laxiflora); XP_09133081(Bra) (B. rapa); and evm.model.supercontig_25.118 (C. papaya).

Figure 2

Conservation and function of GRF5. A, Nucleotide and amino acid sequences of the remnants of a miR396-binding site in GRF5- and GRF6-like sequences in Brassicaceae. Red letters in the nucleotide sequences indicate bases that change in all the Brassicaceae GRFs lacking a miR396-binding sequence. Violet and green letters indicate positions with changes in the GRF5 and GRF6 clade, respectively. See Supplemental Table S1 for details about the species analyzed. Mfe: mfe of the interaction with Arabidopsis miR396a. Mfe ratio: mfe ratio between each site and a perfectly complementary site. B–E, Rosettes of wild-type (B), grf3 (C), grf5 (D), and grf3 grf5 (E) plants. Bar = 1 cm. F, Area and pictures of fully expanded first leaf of wild-type, grf3, grf5, and grf3 grf5 plants. Different letters indicate significant differences as determined by ANOVA followed by Tukey’s multiple comparison test (P < 0.05; n ≥ 12). Leaves were digitally extracted for comparison. Bar = 1 cm. G–L, Effects of miR396 over-expression on wild-type (G, J), grf5 (H, K), and grf3 grf5 (I, L) in 14-d-old seedlings. Bar = 0.5 cm.

Figure 3

Transcriptional regulation of GRF5. A and B, GUS staining showing promoter activity of GRF5 (pGRF5:GUS) (A) and GRF3 (pGRF3:GUS) (B) in 14-d-old seedlings, inflorescences, and flowers (from left). Bar = 1 mm. C and D, VISTA plot of pair-wise comparisons of Arabidopsis GRF5 (C) and GRF3 (D) with orthologs from Brassicaceae species. The analyzed regions correspond to 2-kb sequences upstream of the ATG. See Supplemental Table S1 for details about the species analyzed. E, Alignment of a conserved promoter region present in nine Brassicaceae species. ARF2-binding sites predicted by a FIMO analysis (P < 0.01) with the ARF2-binding motifs highlighted in blue.

Figure 4

ARF2 regulates GRF5 expression. A–C, Rosettes of wild-type (A), p35S:GRF5 (B), and arf2 (C) 25-d-old plants. Rosettes were digitally extracted for comparison. Bar = 1 cm. D, Binding of ARF2 at the GRF5 promoter. Data are means ± SEM of three biological replicates. Asterisks indicate significant differences from the wild type as determined by Student’s t test (P < 0.05). At the right, GRF5 gene schematic representation. Pink box indicates localization of putative ARF2-binding sites in the GRF5 promoter, as detailed in Figure 3, E. The black arrowheads show the position of the primers used in the ChIP-qPCR experiments, with A and B indicating amplicons as shown to the left. E and F, GUS staining of typical 14-d-old transgenic plants harboring reporters for wild-type GRF5 promoter (E) and a mutant version with a deletion in the putative ARF2-binding sites (F). Figures of the seedlings in their entirely shows composite images assembled from multiple photos. Bar = 1 mm. G, Relative expression levels of GRF5 in arf2 and wild-type third leaves of 12-d-old plants. The data shown are mean ± SEM of three biological replicates. Asterisks indicate significant differences from the wild type as determined by Student’s t test (P < 0.05). H, Relative expression levels of ARF2 in p35S:GRF5 and wild-type third leaves of 12-d-old plants. The data shown are mean ± SEM of three biological replicates. Asterisks indicate significant differences from the wild type as determined by Student’s t test (P < 0.05).

Figure 5

Transcriptional regulation of Thecc1EG029130, a GRF5-like transcription factor that harbors a miR396-binding site, in Arabidopsis plants. A, Phylogenetic tree of the 9 Arabidopsis GRFs and the 10 GRFs from T. cacao. Arabidopsis GRF5 and GRF6 (highlighted in violet and green, respectively) are the only GRFs lacking the consensus miR396-binding site. B, Scheme of the GRF5-like gene Thecc1EG029130 (highlighted in gray in the phylogenetic tree (A)) with detailed localization and sequence of miR396 target site. C and D, GUS staining showing promoter activity of Arabidopsis GRF5 (C) and Thecc1EG029130 (D) in typical 14-d-old seedlings and inflorescences. Bar = 1 mm.

Figure 6

Genetic interaction between arf2 and grf5. A–D, Rosettes of 25-d-old wild-type (A), grf5 (B), arf2 (C), and arf2 grf5 (D) plants. Rosettes were digitally extracted for comparison. Bar = 1 cm. E, Area and pictures of fully expanded first leaf of wild-type, arf2, grf5, and arf2 grf5 plants. Different letters indicate significant differences as determined by ANOVA followed by Tukey’s multiple comparison test (P < 0.05; n = 15). Leaves were digitally extracted for comparison. Bar = 1 cm. F and G, Cell size (F) and estimation of palisade cell number per leaf (G) in wild-type, arf2, grf5, and arf2 grf5 plants. Different letters indicate significant differences as determined by ANOVA followed by Tukey’s multiple comparison test (P < 0.05).

Figure 7

Feedback loops in the ARF2–miR396–GRF regulatory network. A and B, Typical GUS staining of third leaves in 14-d-old transgenic plants harboring reporters for pMIR396b:GUS in wild-type (A) and arf2 mutant backgrounds (B). Bar = 0.5 mm. C, Relative expression levels of miR396 in arf2 third leaves of 12-d-old plants. The data shown are mean ± SEM of three biological replicates. Asterisks indicate significant differences from the wild type as determined by Student’s t test (P < 0.05). D, Relative expression levels of GRF1, 2, 3, 4, and 6 in arf2 third leaves of 12-d-old plants. The data shown are mean ± SEM of three biological replicates. Asterisks indicate significant differences from the wild type as determined by Student’s t test (P < 0.05). E–H, Rosette of 25-d-old wild-type (E), arf2 (F), rGRF3 (G), and arf2 rGRF3 (H) plants. Rosettes were digitally extracted for comparison. Bar = 1 cm. I, Area and pictures of fully expanded first leaf of wild-type, arf2, rGRF3, and arf2 rGRF3 plants. Different letters indicate significant differences as determined by ANOVA followed by Tukey’s multiple comparison test (P < 0.05; n ≥ 15). Leaves were digitally extracted for comparison. Bar = 1 cm. J–K, Cell size (J) and estimation of palisade cell number per leaf (K) in the first leaves of wild-type, arf2, rGRF3, and arf2 rGRF3 plants. Different letters indicate significant differences as determined by ANOVA followed by Tukey’s multiple comparison test (P < 0.05). L, Relative expression levels of GRF5 in wild-type, rGRF3, arf2, and arf2 rGRF3 plants. The data shown are mean ± SEM of three biological replicates. Different letters indicate significant differences as determined by ANOVA followed by Tukey’s multiple comparison test (P < 0.05). M, Relative expression levels of miR396 in rGRF3 third leaves of 12-d-old plants. The data shown are mean ± SEM of three biological replicates. Asterisks indicate significant differences from the wild type as determined by Student’s t test (P < 0.05). N, Relative expression levels of GRF1-6 in p35S:GRF5 third leaves of 12-d-old plants. The data shown are mean ± SEM of three biological replicates. Asterisks indicate significant differences from the wild type as determined by Student’s t test (P < 0.05).

Figure 8

A proposed model for the regulation of Arabidopsis growth by the ARF2–miR396–GRF regulatory network. GRFs are repressed by diverse mechanisms: while GRF1–4 are post-transcriptionally repressed by miR396, GRF5 is repressed transcriptionally by ARF2. These transcription factors interact with the co-activator GIF1 and promote cell proliferation in leaves. In addition, ARF2 levels have a mild effect on miR396 and GRF1–4.