Integrated Regulation of Apical Hook Development by Transcriptional Coupling of EIN3/EIL1 and PIFs in Arabidopsis

Abstract

The apical hook protects the meristems of dicot seedlings as they protrude through the soil; multiple factors, including phytohormones and light, mediate apical hook development. HOOKLESS1 (HLS1) plays an indispensable role, as HLS1 mutations cause a hookless phenotype. The ETHYLENE INSENSITIVE3 (EIN3) and EIN3-LIKE1 (EIL1) transcription factors integrate multiple signals (ethylene, gibberellins, and jasmonate) and activate HLS1 expression to enhance hook development. Here, we found that Arabidopsis thaliana PHYTOCHROME INTERACTING FACTOR (PIF) transcription factors act in parallel with EIN3/EIL1 and promote hook curvature by activating HLS1 transcription at a distinct binding motif. EIN3/EIL1 and PIFs can promote hook formation in the absence of the other. Jasmonate represses PIF function to inhibit hook development. Like EIN3 and EIL1, MYC2 interacts with PIF4 and hampers its activity. Acting together, EIN3/EIL1 and PIFs alleviate the negative effects of jasmonate/light and facilitate the positive effects of ethylene/gibberellins. Mutating EIN3/EIL1 and PIFs causes a complete hookless phenotype, marginal HLS1 expression, and insensitivity to upstream signals. Transcriptome profiling revealed that EIN3/EIL1 and PIFs additively and distinctly regulate a wide array of processes, including apical hook development. Together, our findings identify an integrated framework underlying the regulation of apical hook development and show that EIN3/EIL1 and PIFs fine-tune adaptive growth in response to hormone and light signals.

Figure 1.

The ein3 eil1 Mutants Are Responsive to PAC- or JA-Mediated Hook Development. (A) and (B) Three-day-old etiolated seedlings were grown on PAC-gradient medium (A) or JA-gradient medium (B), and the hook phenotype was recorded. Bar = 1 mm. (C) and (D) Quantification of the hook curvature phenotype in (A) and (B), respectively. The values shown indicated means ± se; n ≥ 15. Significance analysis was based on one-way ANOVA along with Bonferroni correction at a significance level of 0.01. Different lowercase letters above the bars indicate a significant difference. Keys show the gradient concentration (μM) of PAC (C) or JA (D), respectively.

Figure 2.

PIFs Promote Hook Development and HLS1 Expression. (A) Three-day-old etiolated seedlings were grown on MS medium, and the hook phenotype was recorded. Bar = 1 mm. (B) Quantification of the hook curvature phenotype in (A). The values shown indicate means ± se; n ≥ 15. Significance analysis was based on one-way ANOVA along with Bonferroni correction at a significance level of 0.01. Different lowercase letters above the bars indicate a significant difference. (C) HLS1 transcript levels of seedlings in (A). HLS1 levels were detected and normalized to PP2AA3. The value for Col-0 was set to 1. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. Significance analysis was based on one-way ANOVA along with Bonferroni correction at a significance level of 0.05. Different lowercase letters above the bars indicate a significant difference. R represents the Pearson’s correlation coefficient of hook angles in (B) and HLS1 levels in (C), with significance of P. (D) Three-day-old etiolated seedlings were grown on β-estrogen gradient medium, and the hook phenotype was recorded. Bar = 1 mm. (E) Quantification of the hook curvature phenotype in (D). The values shown indicate means ± se; n ≥ 15. (F) HLS1 transcript levels of seedlings in (D). HLS1 transcript levels were detected and normalized to PP2AA3. The expression level in the respective zero β-estrogen treatment sample was set to a value of 1. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. Keys showed the gradient concentration of β-estrogen. Statistical significance in (E) and (F) was calculated between β-estrogen treatment and the respective zero treatment for each genotype using two-tailed Student’s t test with asterisks denoting statistical significance (***P < 0.001; **0.001 < P < 0.01; *0.01 < P < 0.05). R represents the Pearson’s correlation coefficient of hook angles in (E) and HLS1 levels in (F), with significance of P.

Figure 3.

PIF4 Directly Binds to an Element in the HLS1 Promoter to Activate HLS1 Transcription. (A) Schematic illustration of the HLS1 promoter region. “ATG” indicates the first codon, and HLS1_EBS indicates the confirmed EIN3 binding sites. HLS1A, HLS1B, and HLS1C contain E-box motifs; HLS1D contains a similar sequence with a G-box motif; and HLS1E corresponds to 5′ end sequences. The nucleotide sequences of candidate binding sites are noted. (B) ChIP-PCR for the in vivo binding of PIF4 with HLS1A. Seedlings were grown on 5 μM β-estrogen medium for 3 d. Cross-linked chromatin was precipitated with anti-MYC antibody, and the eluted DNA was subjected to RT-PCR for the sequences in (A). The enrichment values of the various fragments in iPIF4/pifq were normalized to those in pifq (set to a value of 1). Statistical significance was calculated between relative enrichment of fragment A in iPIF4/pifq and those of other fragments (E, C, B, EBS, D, and 3′ untranslated region [UTR]) in iPIF4/pifq, respectively, using two-tailed Student’s t test with asterisks denoting statistical significance (**0.001 < P < 0.01; *0.01 < P < 0.05). (C) An HLS1 probe containing an E-box motif (CAAATG) was selected from the HLS1A fragment and used for the EMSA experiment. (D) EMSA results showing the in vitro binding of the PIF4 C terminus to the HLS1 promoter. TF-PIF4C (PIF4 201–430 amino acids fused with HIS-tagged trigger factor) was purified in vitro and incubated with 16 fmol DIG-labeled hot probe or 4 pmol unlabeled “cold” or mutant cold (mCold) probe. TF protein (HIS-tagged trigger factor) was used as a negative control. (E) Schematic illustration of effectors and reporters used in the DLR experiment. The E-box motif within the fragment (CAAATG in ProHLS1_WT) was either mutated (ACTTCA, ProHLS1_M) or deleted (ProHLS1_Δ). (F) Transient DLR assays illustrated the requirement for the E-box motif within HLS1_A. Reporter and effector plasmids were combined as indicated then transformed into pifq ein3 eil1 protoplasts. The effects of PIF4 on the three reporters were normalized to the respective empty effectors (set to a value of 1) for comparability. The values shown indicate means ± se; n ≥ 3. Statistical significance was calculated between the various effectors and the empty controls using a Student’s t test with asterisks denoting statistical significance (***P < 0.001).

Figure 4.

EIN3/EIL1 and PIFs Promote Hook Curvature and HLS1 Expression in Parallel. (A) Three-day-old etiolated seedlings were grown on MS medium, and the hook phenotype was recorded. Bar = 1 mm. (B) Quantification of the hook curvature phenotype in (A). The values shown indicate means ± se; n ≥ 15. Significance analysis was based on one-way ANOVA along with Bonferroni correction at a significance level of 0.01. Different lowercase letters above the bars indicate a significant difference. (C) HLS1 transcript levels of the seedlings in (A). HLS1 levels were detected and normalized to PP2AA3. Col-0 was designated as the calibrator and set to a value of 1. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. Statistical significance was calculated between noted samples using two-tailed Student’s t test with asterisks denoting statistical significance (***P < 0.001; **0.001 < P < 0.01; *0.01 < P < 0.05). (D) Three-day-old etiolated seedlings were grown on ACC gradient medium, and the hook angles were quantified. The values shown indicate means ± se; n ≥ 15. Significance analysis was based on one-way ANOVA along with Bonferroni correction at a significance level of 0.01. Different lowercase letters above the bars indicate a significant difference. (E) HLS1 transcript levels of the seedlings in (D). HLS1 levels were detected and normalized to PP2AA3. Col-0 with 0 µm ACC was designated as the calibrator and set to a value of 1. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. Statistical significance was calculated between ACC treatment and zero treatment for Col-0 and pifq, respectively, using two-tailed Student’s t test with asterisks denoting statistical significance (***P < 0.001; **0.001 < P < 0.01; *0.01 < P < 0.05). R represents the Pearson’s correlation coefficient of hook angles in (D) and HLS1 levels in (E), with significance of P. (F) Three-day-old etiolated seedlings were grown on β-estrogen gradient medium, and the hook phenotype was recorded. Bar = 1 mm. (G) Quantification of the hook curvature phenotype in (F). The values shown indicate means ± se; n ≥ 15. (H) HLS1 transcript levels of the seedlings in (F). HLS1 levels were detected and normalized to PP2AA3. The zero β-estrogen treatment control was designated as the calibrator and set to a value of 1. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. Keys show the gradient concentration of β-estrogen. Statistical significance in (G) and (H) was calculated between β-estrogen treatment and the respective zero treatment for each genotype using two-tailed Student’s t test with asterisks denoting statistical significance (***P < 0.001; **0.001 < P < 0.01; *0.01 < P < 0.05). R represents the Pearson’s correlation coefficients of hook angles in (G) and HLS1 levels in (H), with significance of P.

Figure 5.

JA Inhibits the Function of PIF4 Partially through MYC2. (A) Three-day-old etiolated seedlings of pifq were grown on JA gradient medium, and the hook phenotype was recorded. Bar = 1 mm. (B) Quantification of the hook curvature phenotype in (A). The values shown indicate means ± se; n ≥ 15. Angles of Col-0 were taken as control. Significance analysis was based on one-way ANOVA along with Bonferroni correction at a significance level of 0.01. Different lowercase letters above the bars indicate a significant difference. (C) HLS1 transcript levels of Col-0 and pifq seedlings grown on MS or 1 µM JA medium for 3 d. HLS1 levels were detected and normalized to PP2AA3. Col-0 without JA treatment was designated as the calibrator and set to a value of 1. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. (D) Quantified hook angles of iPIF4/pifq ein3 eil1 seedlings grown on MS, 5 µM β-estrogen, or 5 µm β-estrogen plus JA gradient medium (indicated by keys) for 3 d. The values shown indicate means ± se; n ≥ 15. Significance analysis was based on one-way ANOVA along with Bonferroni correction at a significance level of 0.01. Different lowercase letters above the bars indicate a significant difference. (E) HLS1 transcript levels of seedlings in (D). HLS1 levels were detected and normalized to PP2AA3. The mock treatment control was designated as the calibrator and set to a value of 1. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. R represents the Pearson’s correlation coefficients of hook angles in (D) and HLS1 levels in (E), with significance of P. (F) and (G) Transient DLR assays illustrated the negative effect of MYC2 on PIF4-induced HLS1 (F) and PIL1 (G) transcription. Reporter and effector plasmids were combined as indicated and transformed into phyA phyB protoplasts. The values shown indicate means ± se; n ≥ 3. Statistical significance was calculated between the various effectors using two-tailed Student’s t test with asterisks denoting statistical significance (***P < 0.001; *0.01 < P < 0.05). (H) PIL1 transcript levels of 3-d-old seedlings grown on MS or JA medium. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. (I) PIL1 transcript levels of seedlings grown on MS medium for 3 d and transiently treated with different concentrations of JA for 1 h. PIL1 levels were detected and normalized to PP2AA3. The zero JA treatment control was designated as the calibrator and set to a value of 1. Keys show the gradient concentration of JA (µM). The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3. Statistical significance in (C), (E), (H), and (I) was calculated between noted samples using two-tailed Student’s t test with asterisks denoting statistical significance (***P < 0.001; **0.001 < P < 0.01). n.s., not significant.

Figure 6.

MYC2 Interacts with PIFs. (A) and (B) Firefly luciferase complementation assays showing the in vivo interaction between MYC2 and PIFs. MYC2-LUC^c was combined with the indicated PIF-LUCⁿ plasmids and transformed into Col-0 (A) or ein3 eil1 (B) protoplasts. “–” indicates the LUCⁿ or LUC^c control plasmids. LUC activity was recorded after the samples were incubated for 12 to 16 h and mixed with luciferin. Statistical significance was calculated to compare the different PIF + MYC2 combinations with the LUC^c + LUCⁿ control using a Student’s t test with asterisks denoting statistical significance (***P < 0.001). (C) and (D) Pull-down assays showing the in vitro interaction between MYC2 and PIF4 (C) as well as MYC2 and PIF1 (D). TF-PIF4 (PIF4 fused with HIS-tagged trigger factor) and TF-PIF1 were immobilized with Ni-NTA agarose then combined with MBP fusion proteins. After rounds of washing and centrifugation, the precipitated products were subjected to SDS-PAGE and further stained by Coomassie blue or blotted with the respective antibodies. Arrows indicate the corresponding proteins. For consistency, the sample order in (D) was rearranged by assigning two input samples together (lanes 1 and 2) and two pull-down samples together (lanes 3 and 4), in which the original order was lane 1-3-2-4.

Figure 7.

EIN3/EIL1 and PIFs Delay Light-Induced Hook Opening and HLS1 Decrease. (A), (C), and (E) Real-time changes in hook angle upon light exposure. The values shown indicated means ± se; n ≥ 15. (B), (D), and (F) Real-time changes in the hook opening rate upon light exposure. The values shown indicated means ± se; n ≥ 15. (G) Quantified max open rates of the seedlings in (B), (D), and (F). Statistical significance was calculated between different genotypes and Col-0 using two-tailed Student’s t test with asterisks denoting statistical significance (***P < 0.001; **0.001 < P < 0.01). (H) Peak time to reach max open rate (hours after light exposure). N.A., not available. (I) HLS1 transcript levels rapidly decreased upon light irradiation. Seedlings were exposed to light for the indicated amounts of time after growth on MS medium for 2.5 d. HLS1 levels were detected and normalized to PP2AA3. Col-0 at the 0 h time point was designated as the calibrator and set to a value of 1, and then the y axis was rescaled to log₂ values respectively. The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n = 3.

Figure 8.

The pifq ein3 eil1 Sextuple Mutant Exhibits Insensitivity to Multiple Hormones and Light. (A) and (C) Three-day-old etiolated seedlings were grown on ACC gradient (A) or PAC with/without GA gradient (B) medium, and the hook phenotype was recorded. Bar = 1 mm. (B) and (D) Quantification of the hook curvature phenotype in (A) and (C), respectively. The values shown indicate means ± se; n ≥ 15. (E) HLS1 transcript levels in pifq ein3 eil1 seedlings after growth on MS, 1 μM JA, or 1 μM PAC medium for 3 d or after transient treatment with 10 ppm ethylene or white light for 1 h. The HLS1 levels were normalized to PP2AA3 and calibrated to the transcript levels in Col-0 (set to a value of 1). The values shown indicate means of biologically repeated experiments (using different pools of seedlings under noted conditions) with sd; n ≥ 3.

Figure 9.

PIFs and EIN3/EIL1 Additively and Distinctly Regulate Myriad Biological Processes. (A) Venn diagrams showing the pairwise overlap between the HLS1-, PIF-EIN3/EIL1-, EIN3/EIL1-, and PIF-regulated gene sets. “↑” represents activated genes that show decreased transcript levels in both mutants (compared with Col-0); “↓” represents repressed genes that show elevated transcript levels in both mutants (compared with Col-0). Genes showing opposite changes of expression in two compared mutants were excluded from the overlap. Percentage values indicate the percentage of overlapping genes among total regulated genes. P values were calculated using the hypergeometric distribution. (B) Venn diagram showing the overlaps among the HLS1-, PIF-EIN3/EIL1-, EIN3/EIL1-, and PIF-regulated gene sets. Genes sharing accordant transcriptional pattern among mutants were counted as overlaps. Classes 1 to 6 were selected for the GO enrichment analysis in (C). (C) Summary of action for EIN3/EIL1 and PIFs, which function as a transcriptional couple in the regulation of HLS1 and other downstream genes in diverse processes. Representative genes with known functions in various biological processes are listed. Repressed and activated genes (thus, up- and downregulated in corresponding mutants) are shown in black and red, respectively.