Repression of flowering by the miR172 target SMZ

Abstract

A small mobile protein, encoded by the FLOWERING LOCUS T (FT) locus, plays a central role in the control of flowering. FT is regulated positively by CONSTANS (CO), the output of the photoperiod pathway, and negatively by FLC, which integrates the effects of prolonged cold exposure. Here, we reveal the mechanisms of regulation by the microRNA miR172 target SCHLAFMUTZE (SMZ), a potent repressor of flowering. Whole-genome mapping of SMZ binding sites demonstrates not only direct regulation of FT, but also of many other flowering time regulators acting both upstream and downstream of FT, indicating an important role of miR172 and its targets in fine tuning the flowering response. A role for the miR172/SMZ module as a rheostat in flowering time is further supported by SMZ binding to several other genes encoding miR172 targets. Finally, we show that the action of SMZ is completely dependent on another floral repressor, FLM, providing the first direct connection between two important classes of flowering time regulators, AP2- and MADS-domain proteins.

Figure 1. Expression pattern of SMZ.

(A) Expression of SMZ is developmentally regulated. High SMZ expression is detected in hypocotyl, cotyledons, and the meristematic region of 7-d-old seedlings (data from AtGenExpress expression atlas , selected samples). SMZ is not detectable in leaves at the rosette stage and flowers, but SMZ mRNA levels increase again as seeds mature. Nomenclature of floral and seed stages according to and , respectively. (B–D) A gSMZ∶GUS reporter confirms the developmental regulation of SMZ in 5-d-old (B), 10-d-old (C), and 15-d-old (D) seedlings. Scale bar indicates 1 mm.

Figure 2. Genetic interactions of SMZ.

(A) SMZ acts as a repressor of flowering under LD photoperiod. smz-D plants flower late under LD, but normally under SD conditions. (B) SMZ and SNZ act redundantly with TOE1 and TOE in controlling floral transition. Loss-of-function alleles of smz-2 and snz-1 significantly (p<0.001) enhance the early flowering of a toe1-2 toe2-1 double mutant. Number of rosette leaves (dark grey) and cauline leaves (light grey) are shown. (C) smz-D delays flowering in 35S::CO. smz-D partially represses early flowering caused by CO overexpression. Early flowering by overexpression of FT is not alleviated by smz-D. (D) SMZ represses flowering independently of FLC. smz-D was introduced into the flc-3 background. Double homozygous plants and controls were grown under LD conditions. Error bars indicate 2× standard error of the mean (SEM).

Figure 3. Repression of flowering by SMZ requires FLM.

Early flowering of flm-3 is epistatic over SMZ and rSMZ overexpression. Constitutive expression of SMZ (B and G) and rSMZ (C and G) in Col-0 background strongly delays the onset of flowering (G) compared to wild-type control (A and G). Expression of SMZ (E and G) and rSMZ (F and G) in a flm-3 mutant background (D and G) results in wild-type–like flowering. (F) rSMZ flm-3 plants display stunted growth, leaf curling, and reduced apical dominance. (G) Loss of SVP does not prevent late flowering by expression of 35S::(r)SMZ. Error bars indicate 2× SEM. Scale bars indicate 1 cm.

Figure 4. Tissue-specific misexpression of SMZ and rSMZ.

(A) Phenotypes of plants expressing either SMZ or rSMZ mRNA under the control of the constitutive 35S promoter, the phloem companion cell-specific SUC2 promoter, or the meristem-specific FD promoter. Magnification of the abnormal shoot and flower morphology observed in FD::rSMZ plants is shown (inset; picture taken at a later time point). (B) Flowering time of SMZ and rSMZ misexpression plants. Data are from T2 plants, except for those lines that did not produce any flowers in T1 (35S:: rSMZ) or that did not produce any fertile flowers (FD::rSMZ) and for which T1 data are shown instead. Error bars indicate 2× SEM.

Figure 5. SMZ represses FT transcription.

The effect of SMZ overexpression on FT induction was analyzed by quantitative real-time PCR in wild type, as well as in flc-3 and flm-3 mutants. (A) smz-D prevents the induction of FT by LD in flc-3 1 and 4 d after plants were shifted to inductive LD conditions. (B) 35S::SMZ prevents induction of FT in Col-0 after three inductive LD. FT expression is restored in a flm-3 loss-of-function background. Plants were initially grown under noninductive SD conditions, and synchronous flowering was induced by shifting plants to inductive LD. (C) FT is precociously expressed in LD-grown toe1 toe2 smz snz (dashed line) when compared to wild type (solid line). Plant tissue for RNA extraction was collected at the peak of FT expression shortly before the end of the day (ZT = 15 in a 16-h LD). Error bars indicate standard deviation of triplicate measurements.

Figure 6. Effects of SMZ overexpression on leaf and meristem transcriptome.

Microarray analysis in leaves (left) and at the shoot apical meristem (right) in flc-3 and smz-D flc-3. Changes in gene expression in response to inductive photoperiod were determined 0, 1, and 4 d after the shift to LD in leaves and 0, 3, 5, and 7 d after the shift to LD at the shoot apex. Median normalized expression intensities are shown (log2). x-axis: days after shift to inductive LD.

Figure 7. SMZ tagged with GFP remains functional and represses flowering.

Transgenic lines constitutively expressing SMZ tagged at the N-terminus with eGFP are late flowering. Flowering times of selected 35S::GFP-SMZ T2 lines. Error bar indicates 2× SEM.

Figure 8. Identification of SMZ targets by ChIP-chip.

(A–D) SMZ binds to the genomic regions of (A) SMZ, (B) SNZ, (C) TOE3, and (D) AP2, suggesting extensive feedback regulation among the miR172 target genes. (E and F) SMZ binds to the promoters of the floral repressors (E) TEM1 and (F) FRI. (G–I) The floral integrator and flower development genes (G) FT, (H) SOC1, and (I) AP1 are directly bound by SMZ. (J) Binding of SMZ to the regulatory regions identified by ChIP-chip was confirmed by quantitative PCR. Bound regions (peaks) are highlighted in grey. All peaks fall within the top 1% of enriched regions (99th percentile), and at least one peak per gene is statistically significant at a FDR <5%. For a complete list of all 434 regions bound by SMZ at a FDR <5%, see Dataset S5.

Figure 9. Genetic Interactions governing photoperiodic flowering.

Photoperiod is perceived in leaves and entrains the circadian clock. CONSTANS (CO) is a major output of the clock in terms of regulating flowering. CO is activating expression of FLOWERING LOCUS T (FT) specifically under LD conditions. This activation of FT is counteracted by floral repressors such as SMZ, which itself is negatively regulated by miR172. SMZ is also repressing related genes in a negative feedback loop. At the shoot apex, SMZ binds to regulators sequences of APETALA1 (AP1) and SUPRESSOR OF CONSTANS OVEREXPRESSION (SOC1), two other known regulators of flowering time and floral development.